Thursday, 12 October 2017

Anti-Orbit Laser Submarines

Laser-equipped nuclear-powered submarines are the perfect last line of defense against an attacking force in orbit.

The situation
You don't win every fight. Eventually, there will come a time in space warfare where a fleet of space warships has defeated all your mobile forces and your immobile defenses. They will bear down on you from above with lasers, missiles and kinetic projectiles and you will have to find a way to prevent their forcing of an unconditional surrender. 

We will refer to the opponents as the 'attackers' and to you as the 'defenders'. The first step to devising an effective defense is to understand the situation.

So what is the situation?
Its an enemy ship.
An attacking warship will start out in high orbit. This is an altitude of 2000km or above. Whether it has just arrived from an interplanetary voyage or has recently destroyed your remaining warships, it will go to high orbit to maximize the effectiveness of its space superiority. Space superiority, borrowing from the term 'air superiority', is when a force has complete dominance over all the orbits around a planet. No space-borne forces can oppose this superiority and no reserve forces can challenge them without being quickly destroyed. 

What does losing space superiority mean for defenders?

The most important consequence is that enemy warships have free reign to change orbits, maneuver into favorable positions and receive re-supplies. Their mobility is unconstrained. 
High Elliptic Orbits and gravity assists from a large moon allow for a huge variety of orbits.
Attackers in high orbit can make optimal use of their laser weaponry. They can get clear lines of sight onto any spot on the surface, and the long distances between objects forces travel times to length with the secondary effect of giving lasers plenty of time to shoot down targets. Laser effectiveness is generally dependent on how far they are from a target and how much 'dwell time' a beam can spend on a target. 

De-orbiting objects from high altitude is inexpensive in terms of deltaV. This works in favour of missiles sent down from orbit by allowing them to use very little propellant to strike ground targets, which makes them lightweight and cheap to send by the hundreds. Additionally, falling towards Earth gives them a big boost to the velocities they achieve before impact. 

The same applies to kinetic projectiles, a fact applied in the Rods From God concept of orbital bombardment. 

Retaliating

So you want to shoot back at the attackers. 

 U.S. Army’s Homing Overlay Experiment
Missiles can do the job. They are currently our only method of delivering weapons into orbital space. Something like an ICBM with an additional stage can reach LEO. Reaching higher orbits will require either a very large rocket, a high Isp engine for the upper stage or a launch system such a laser launch or ram accelerator.

However, each of these solutions have major issues when trying to shoot down an opponent in high orbit.

Laser satellite shooting down a missile slowly climbing to orbit.
Large rockets are easy to target and shoot down. A chemical-propellant rocket that needs to minimize its dry mass to achieve the necessary deltaV capacity will have very good acceleration but will end up being very fragile. Nuclear-thermal or nuclear-powered rockets can be much smaller and more durable, but getting sufficient acceleration out of them implies a very high power requirement, which might make them very expensive to throw at the enemy. 
Maglev track rocket launch system... vulnerable.
Regular launches take tens of minutes and cannot be disguised from the attackers. Shortening this window of vulnerability can be done by using a launch system that powers the rocket or accelerates it externally. However, the infrastructure for the launch systems will in turn become a priority target for the attackers. Massive, hard to hide and immobile, they will receive a lot of firepower. Some launch systems are practically impossible to shield from damage, such as a laser launch system that needs thousands of exposed laser optics, and others reveal their positions as soon as they fire a rocket. Building underground is also a very expensive endeavour when considering that all the work can be undone by a single 'bunker buster'-type weapon.

The logistics of launching missiles against attackers sitting in orbit works against the defenders. The attackers can de-orbit a missile by expending only a few tens to hundreds of meters per second of deltaV. A defender must equip each missile with tens of thousands of meters per second of deltaV. It might be easier to build more missiles and create rocket fuels on the ground at the start of the war, but after an orbital bombardment campaign by attackers with space superiority, it is unlikely to be the case.


Kinetics that can be shot all the way to high orbit need to handle hundreds to thousands of Gs of acceleration, traverse the lower atmosphere at dozens of kilometers per second, survive laser fire for several minutes with minimal capacity to dodge and take out a target with a very short window of interception. This is a tall order! 

Hot hydrogen light gas gun launch system. Can be weaponized.
So, what are the defender's options?

They need to retaliate with something that cannot be shot down, from a platform that can avoid counter-fire and can maintain functionality after infrastructure and services have been disabled world-wide. 


One option that fits the bill is laser submarines. 

Lasers cannot be shot down and hit their target instantly. They can be used so long as electrical power is supplied. Submarines operate underwater, an environment that can hide them until they surface and protects them from high-velocity projectiles and lasers while submerged. The can protect themselves this way for months on end, and if they employ the same life support systems as on spaceships, then it can add up to years. 
For the same reason that today's submarine fleets are considered an unbeatable means of retaliating against a foe after nuclear armageddon has wiped out the homeland, laser submarines will be able to operate and remain dangerous even after orbital attacks destroy all support infrastructure. 

Let us now look at how a submarine can be used to retaliate against attackers in orbit. 

The Challenges
Submarines are already equipped with a high electrical power generation. Large modern nuclear submarines are already able to produce over 100MW for years on end. In a futuristic setting with common space travel and space wars, power generation technology developed for interplanetary travel will allow submarines to produce gigawatts or more. 
Next generation water-cooled nuclear reactor.
The most likely generators for space travel will be nuclear due to their high power density. The biggest limitation to generating power from nuclear reactors is waste heat capacity: It is easy to heat up the reactor core but much harder to remove the heat. Submarines will have an entire ocean as a heatsink so will be able to produce more watts compared to a spacecraft with a reactor of the same mass and volume.
The Bi-modal NTR is a nuclear thermal rocket engine that doubles as power generator.
All of this electrical power can be used to power a laser generator. 

Three elements determine a laser's effectiveness: wavelength, radius of focusing optics and beam power. 

An observatory using a laser guide star.
We have already determined that laser submarines will likely be able to produce more electrical power than a similar laser space warship, so laser submarines will also have the advantage in beam power.

The radius of the focusing optics will depend on the specific arrangement of the laser weapon's components and how they are deployed. We will look into the possible designs down below. 

A free electron laser. Unlike most designs, it can freely switch wavelengths.
The wavelength however is not a variable laser designs have much control over. Submarines operate in an aquatic environment, on top of which is a hundred kilometers of Earth's atmosphere. At the interface of the ocean's surface is sea mist and suspended droplets of water in fog or clouds. The ocean's surface is not flat either, with waves of a few centimeters to a few hundred meters rolling over it endlessly. A beam emitted by a submarine will have to penetrate all of this environment and still travel the hundreds of thousands of kilometers' distance separating it from a target in high orbit. 

The optical properties of water are therefore the determinant factor for which wavelengths the laser should produce.


Here is the absorption spectrum of water:

The lower the 'Relative Absorption value', the less the wavelength's energy is absorbed. We can clearly see that the lowest values are for the 'optical window' that corresponds to the 400-700nm visual spectrum. The highest absorption is for 100nm ultraviolet wavelengths and 3000nm infrared wavelengths. 

Here is the absorption spectrum for our atmosphere:

The atmospheric attenuation of electromagnetic radiation has similar features to that of water: short wavelengths such as X-rays cannot go through while long wavelengths such as radio penetrate easily. 

It might be easier to consider the laser beam as being fired from space and coming down to the surface.


A 100nm ultraviolet laser beam will traverse the vacuum of space with ease, but will stop short of reaching the upper atmosphere. A 400nm blue laser will go through the atmosphere and through 460 meters of water before being reduced to less than 1% of its initial power. A 1000nm infrared laser will lose 20% of its power to the atmosphere and be completely absorbed by half a meter of water. A 100m radio wavelength just bounces off the ionosphere.


While shorter wavelengths are preferred for laser weapons, as they allow a beam to be focused to destructive intensities over longer distances, a laser submarine should use 400nm wavelength lasers to penetrate water and the atmosphere without losing a lot of beam power. 


The equation for how much of a beam's energy is retained after traversing a medium is given by:

  • Percentage transmitted: e^( -1 * Attenuation coefficient * Depth) * 100
The attenuation coefficient is usually given in cm^-1, so the depth should be converted into cm units.

For example, near-infrared light at 800nm wavelength has an attenuation coefficient of 0.01cm^-1. We want to know what percent of a near-infrared laser's energy remains after passing through one meter of water.


One meter equals 100 cm. Per our equation, we find the percentage to be approximately e^(-0.01*100)*100: 36.8%. Just over a third of the beam gets through.


Here is a table of values for how much beam power is lost if the blue wavelength laser submarine fires its weapon at different depths, using an attenuation coefficient of 0.0001cm^-1:

We see that to maintain a good percentage of laser power getting through the water, a laser submarine would have to fire at rather low depths. According the the table above, a 10 meters depth using blue laser light looks like a good compromise: deep enough to escape orbital surveys and strikes, with at least 90% of the laser power going through.

So is a laser being fired at 10 meters depth a good idea?

Factors affecting a solid state laser weapon system.
The table gives an incomplete picture. While laser power being absorbed is an important factor to consider, there is a large number of other elements that affect how effective a laser is. One such element is thermal blooming. That 90% power transmission rate implies that 10% of the laser power is absorbed and goes into heating the water. If the laser power is rated in megawatts, the heat absorbed by the water becomes significant. Hot water has different optical properties compared to colder water - it will work as a lens in reverse, effectively de-focusing the laser. 

Another significant issue is the water/gas interface. When light travels between two mediums of significantly different density, like seawater and atmospheric gasses, it is bent by refraction. Even worse, the sea's surface is constantly disturbed by waves, tides and other movements. Instead of a smooth surface, which angling the laser can compensate for, it is continuously changing and bending light in different and hard to predict direction. Here is a familiar example of the effect: 

This effect is familiar to astronomers trying to gaze at stars through the moving atmosphere, and adaptive optics are used to compensate for the deviations in light traversing the atmosphere.
For a laser weapon, adaptive optics work in reverse. 
Adaptive optics cannot be employed as effectively when used underwater. While the atmosphere's movements are already difficult to detect and correct, trying to effectively measure how light moves through two mediums with a complex and moving interface is much harder. Guide lasers are used for measurement, with the light reflecting off the ionosphere creating an artificial 'guide star' for astronomers to calibrate their instruments. A guide laser placed underwater would submit to the same chaotic disturbances as the weapon laser and would be unusable. They would also have to work much harder. Refraction means that inaccuracies are multiplied once they pass through the water/air interface. 
Interference with sunlight works in both directions.
Finally, there's the problem of reflection. While water is decently able to let higher wavelength visible light penetrate, the massive difference in density between the water and the air (x1000) above it means that it is a good reflector. Laser light would travel from the submarine to the water/air interface, and just be bounced back below the surface. For sea water, less than 6% of the laser light would be reflected at angles below 30 degrees.
While the above table might not give any large figures, remember that the angle is measured against the rippling, swelling and rolling waves. A nominal angle of 10 degrees against the water's surface might transition between -20 and 100 degrees as a wave moves over a laser. This is the difference between 3% and 45% of the laser not going through the ocean's surface.
Submarines can survive this... but it might not be enough to matter.
The principal advantage of staying underwater is that the submarine will be protected from high velocity strikes and retaliatory laser fire. However, if it cannot return fire from this position, then it cannot serve as a perfect last line of defense.

The Solutions


So, based on the previous section, we can affirm that attempting to shoot a laser while underwater provides unequalled benefits but also significant challenges.


We can either tackle the problems a laser submarine faces directly or attempt to circumvent them.


-Long wavelengths

Previously, we considered that blue wavelengths were optimal for shooting while underwater as they were absorbed the least. The reduced absorption would have allowed a submarine to transmit over 90% of the laser power to the ocean's surface from a depth of 10 meters. However, the distortion at the water/air interface rendered this option impractical.

How about using a wavelength that is less efficient at penetrating water, but is less affected by distortion? 

Looking at the absorption spectrum of water, we notice that wavelengths longer than 100 micrometers are absorbed less and less as the wavelength increases. At 1m wavelength, the laser would traverse water as easily as blue light. This corresponds to a frequency of 300MHz. This is the radio band. 

Not coincidentally, frequencies of lower than 300MHz are used to communicate with submarines. At 3 to 30kHz, which is wavelengths of 10 to 100km, can penetrate the seas to a depth of several hundred meters. Using even lower frequencies further increases penetration depth, but would require impractically large lens to focus onto a target in high orbit. Another factor working against longer wavelength radio is that the ionosphere can reflect signals back down to the surface at frequencies below 30MHz.

A 1m wavelength radio laser will be able to traverse the atmosphere mostly undisturbed and go through the ionosphere without refraction. The features of waves are too small to cause it to wobble chaotically at the sea/air interface. The beam would bend coming out of the sea, but it is a single predictable deviation that can be corrected.
Inductive output tube diagram.
A radio-wavelength coherent beam or 'raser' can be generated by a Free Electron laser or inductive output tubes with an efficiency exceeding 70%.

The advantages of a radio-based anti-orbital system is that it allows a submarine to fire upon targets while deep underwater. Even at 20 meters depth, the radio beam would transmit 82% of its power through water and lose less than 1% going through the atmosphere. It is much less affected by small waves and other turbulence in the water, and mostly immune to above-surface weather effects.


There are several downsides however. Such a large wavelength makes it impossible to focus the beam down to destructive intensities without a very large radio dish - this might get impractical when you also want the submarine to move quickly while underwater. Another issue is that the beam won't interact with the target in a consistent manner. 


Lasers, for example, are absorbed by the outermost layers of the materials the target's surface is made of. The heating is concentrated in the 'skin' of the target. Sufficiently intense laser beams heat this skin layer to very high temperatures, causing the material to boil away or even explode. 
Radio beams would use wavelengths a million times longer that do not interact with the target's materials at an atomic level. They are much more sensitive to the conductivity of the materials they are striking. Good conductors such as steel or aluminium efficiently reflect radio waves and are not heated. Good insulators such as ceramics or glass are mostly transparent to radio waves and do no absorb the beam's energy as it passes through them. Radio absorbing materials have to be neither good conductors or insulators, such as 

This is bad news if the targets are space warships with an external metallic hull and an internal structure based on advanced carbon-composite and ceramic materials. Large propellant tanks will let the radio waves pass straight through. Small features of 10cm or smaller are completely invisible to the radio waves too. 


However, there will still be ways to deal damage. 


Openings in the metallic hull would allow radio waves to enter and then bounce around on the internal surface. Like a microwave oven, the trapped radiation will pass through radio transparent materials thousands to millions of times before being fully absorbed. A human is mostly composed of salted water. He or she would absorb about between 0.1 and 1% of a radio beam going through their body. If the radio beam stays inside a 10m diameter hull for just 76 microseconds, 2300 bounces are possible and the percentage of beam energy absorbed rises to 90%. When the beam power is measured in tens to hundreds of megawatts, this has dire consequences for a human crew. 

Another effect is induced current. If even a few watts manage to circulate in microcircuitry, it is enough to short-circuit or even melt down computers, avionics and delicate sensors. RF Shock and Burn is a serious issue for electricians and engineers working on conductive structures near a high frequency radio source. At the power levels radio-laser submarines will pump into targets, induced current is enough to melt steel.

Steel melted by 1-30kHz radio frequencies.
Modern submarines are not powerful enough to compensate for the diffraction of a 1m wavelength radio beam. With 100MW of available electrical power, 30% of which is lost in an inductive output tube and another 10% to seawater and atmospheric absorption, less than 63MW will reach space. Even the largest submarines, such as the Ohio-class SSBN, have a hull diameter (beam) of 13m. Mounting an internal dish to focus a radio beam up to this diameter will create a very low performance laser. 
Targets at 10km distance will receive about 22W/m^2 - a great radio signal, but a terrible weapon. Targets in low orbit and high orbit will receive milliwatts of power. 
A 1296MHz dish. For 300MHz beams, the spaces can be even larger.
What is needed is much more power and an externally mounted radio dish. Thankfully, a 300MHz beam can be focused by a dish with holes up to a tenth of the wavelength in size. A radio dish for this wavelength can is very lightweight and easily collapsible, with conductive spars spaced by 10cm lengths. The spars can be made hollow to have neutral bouyancy, allowing them to support themselves without many structural elements. At 10m depth and below, there are few disturbances in the surrounding water.

Dish diameters of 100 meters or more are envisageable, massing less than 1kg per m^2. Tension wires hold the shape and serve as the mechanisms for adaptive optics to act on the dish. 


A group of submarines using space-grade nuclear reactors might be able to put together 500GW of power with a combined reactor mass of only 2500 tons. Between them, they can hold up a dish 1km in diameter, as follows:

This arrangement allows the submarines to focus a 315GW beam to an intensity of 67kW/m^2 at an altitude of 1000km. For low orbit targets, the intensity is 1.67MW/m^2. These intensities are far from enough to melt or physically damage the structure of a spaceship. However, the beam is large and entirely envelops the target. Any hole through a metallic exterior or any cavity lined by a radar reflector will turn the spaceship into a microwave oven receiving megawatts of heating over time. At high altitudes, the intensity is lower by the targets orbit much slower, giving the Rasar beam time to boil crews to death and melt components directly or indirectly. 

-Interface lens

A variant on the lens used for above.under water filming.
The biggest trouble with optical wavelengths is the chaotic distortions and reflections created by the sea/air interface. They would allow submarines to physically damage targets with relatively small lens and shoot on the move, but aiming though the interface seems impossible...

... unless an interface lens is used.


It is an optical array that floats on the ocean's surface, serving to handle the beam's transition from underwater to atmospheric mediums. Glass can be made to have a refractive index similar to that of water. A laser beam traversing a sea/glass interface would not suffer any distortion. This is the reason why some transparent objects disappear completely while underwater - our eyes cannot make out any distortions that reveal their presence.

The refractive index of the polymer ball is exactly the same as water.
The interface lens can also serve to focus the laser. It can be made much larger than whatever the submarine can carry, as it is not confined by hull dimensions or hydrodynamics. 

One primary advantage of these floating structures is that they can be deployed before firing commences, and each is much cheaper than a submarine. When a target passes overhead in orbit, a laser submarine can rise to firing depth and start shooting. It only needs to equipment to focus the beam from the laser generator to the interface lens, a distance of ten meters or so. The floating lens receives the beam, corrects the angle and aims at the target overhead. If the target is not destroyed, it can trace the laser back to its origin and initiate a retaliatory strike. 


An interface lens is not very mobile and needs to stay on the surface, so it cannot protect itself by diving. There is a good chance it will be destroyed... at which point the laser submarine switches to firing through another interface lens and so on. 


Although the lens will have to be rather large and heavy to receive a laser beam from a wide variety of angles underwater, and re-focus it in a moving target hundreds of kilometers above, which makes it expensive, it must be considered as an expendable asset when compared to cost and size of a nuclear submarine. Also, the lens can be covered by an isothermal sheet and made out of materials transparent to radar, making it hard to detect from orbit until it starts firing.


Using a 10m diameter interface lens, even a modern submarine with 100MW of available output will be able to deal serious damage to targets in low orbit. About 40MW of the submarine's power will reach the target using a diode laser generator at 400nm wavelength, but at an intensity of 133GW/m^2 at 200km altitude. This is enough to rip through 14.8 meters (!) of aluminium per second, or even 6.5m/s of carbon armor. Any target caught by this beam for even a second will be cut in half. At 1000km, its performance is still a respectable 52mm/s through carbon.

Spaceship in Children of a Dead Earth scarred by laser strikes.
More advanced submarines can get away with smaller lens that are harder to counter-attack and still deal devastating damage to high orbit targets. A 10GW laser beam focused through a 4m wide interface lens can blast away targets at a rate of 837mm/s at 1000km. 

Disadvantages of this system is the cost of 'expendable' large adaptive mirrors and possibly the inability to use floating structures in severe storms with large waves rocking the optics. 


-Towed Lens


This fixes the problem with floating optical arrays. The laser is focused by towed apparatus that can be held underwater until firing starts, and then moved around after an initial volley to avoid counter-fire. 

The towed lens can be lighter and cheaper than a fully independent floating lens. Electrical power, computing operations and other functions can be provided by the submarine towing the lens, with only the actuators and suspension retained.

Lasers have been carried through optical fibres with nearly zero losses of beam energy over distance. This is because internal reflection of the beam is done at grazing angles within the optic fibre. In other words, laser beams of megawatts to gigawatts power levels can be transported by optic fibres without any significant losses and no special provisions against heating. 
Optical fibres can carry a laser generated by a submarine to the towed lens, and they can run parallel to the load bearing cables attaching the lens to the submarine. This method of delivering the beam to the lens bypasses the losses and complications of having the beam penetrate several meters of water to reach the surface. 

To these advantages come some downsides. The submarine becomes more vulnerable that if it relies upon fully independent floating lens. If an enemy target locates a floating lens, it might correctly assume that the much more valuable submarine is very close to the lens. When a target spaceship in orbit sends down counter-fire before the submarine is ready, the latter would be forced to cut loose the lens and dive... this breaks the optical fibre cable and renders the lens useless. 

Tactics using towed lens might involve dragging along a fleet of lenses and rotating them to the surface and back. Longer fibre optic cables gives the submarine more freedom to move, while a mix of decoys and intermittent and random firing patterns reduces the disadvantages of the design. 

-Optical phased array floater.

Do away with the submarine!
A specialized vehicle can be built solely for the purpose of hiding a laser weapon underwater and surfacing for short bursts of fire. Since this weapon only attacks from above the water and does not have to worry about hydrodynamics, nifty solutions such as an optical phased array can be used. The laser generator's size is equal to the lens diameter and if it received damage, it will only suffer reduced output. 
Instead of VLS cells, phased array grids?
The more delicate components such as the power generator can remain submerged, only transporting electricity to the phased array on the surface through cables. A reactor embedded in the sea floor can be a very difficult target to locate and destroy.

-Supercavitating platform. 
'Bullet submarines' powered by excessively powerful nuclear reactors.
If submarines have access to gigawatts of power, they can also use it for propulsion.

This level of power output can enable submarines to reach supersonic speeds through supercavitation through water. If they can rise to the surface, fire, dive and relocate in a matter of seconds, then they can evade counter-fire through sheer agility. 
It might be best in this case to mount a set of optical phased array lasers tailored to trans-atmospheric wavelengths to be used once a submarine surfaces. Gigawatts of power means gigawatts of heating: the local atmosphere can be cleared of moisture and mist that distorts the beam most heavily. 

A 1GW laser at 400nm focused by even a relatively small 4m diameter lens can blast past 23.7m/s of aluminium or 10.4m/s of carbon at 200km, and remains deadly at 1000km with 83mm/s of carbon penetration. This allows for short bursts of laser fire to take down any target.
Atmospheric lensing.
Advanced techniques such as thermal lensing using the Kerr effect is being developed in programs such as the Laser Developed Atmospheric Lens (LDALs) by DARPA. LDALs can allow high-power submarines to extend their effective range to tens of thousands of kilometers while reducing the effectiveness of laser counter-fire. 

Conclusion. 

Lasers and underwater environments don't mix well, but there are many solutions to gain the protection a submarine enjoys while attacking instantly and repeatedly with direct energy weapons. 

Once these solutions are applied, a defending planet with deep oceans can hope to maintain an effective last line of defense against invading spaceships.

141 comments:

  1. While submarines are excellent capital ships and can be equipped with a wide array of weaponry, I have never considered the idea of a laser armed submarine, especially one attacking enemy spacecraft! One issue which will make this more difficult is target location and tracking. The submarine will have a very limited view of the atmosphere or space using a periscope mounted camera or imaging array, and certainly would not want to go "active" (much like they avoid active sonar against enemy ships and submarines) to avoid revealing their position.

    One idea that I have played with is to revive project "plumb-bob", where a nuclear device is used to accelerate a large penetrator into orbit against enemy spacecraft. In a way this is a variation of an idea in a previous post, but seems to be somewhat easier to pull off in this setting.

    For those readers who are unfamiliar with the concept, an American underground nuclear test (code named plumb-bob) blew off the heavy steel "cap" which was covering the underground shaft. Calculations indicated the cap could potentially been accelerated at up to 6X escape velocity, but the cap was never recovered (presumably it was vapourized by air friction in flight).

    This can be improved by doing several things.

    Firstly, the shaft should be filled with water, so the intense X-ray radiation of the fireball is absorbed and turned into plasma. The plasma then pushes on and accelerates a properly shaped aerodynamic penetrator which was mounted on top of the device, potentially moving at 10 or more times escape velocity. To maximize the chances of striking the vehicle in orbit, the penetrator could be designed to split into multiple submunitions as it clears the atmosphere (I am imagining long rod penetrators, but other shapes could be imagined).

    In the submarine version of this, the tube housing the nuclear physics package and the penetrator could be hollow, to minimize weight. The submarine deploys them underwater, where they fill with seawater until they are neutrally buoyant and floating just below the surface. The sub escapes at flank speed while the devices receive updates of the position of orbital ships or devices. When they pass overhead, the device is triggered, sending heavy hypervelocity darts screaming skyward.

    As an alternative, they could be planted on the sea bottom like mines. When the time comes, they deploy flotation bags and rise until the tops pierce the surface, and fire upwards at the target as it passes overhead.

    As always, there are many changes and tweaks which could be made, so YMMV

    ReplyDelete
    Replies
    1. Yep, the Nuclear Wang Bullet: https://www.nextbigfuture.com/2010/03/150-kiloton-nuclear-verne-gun.html

      Also some discussion you might find interesting on this blog post: https://forums.spacebattles.com/threads/anti-orbital-laser-submarines.577403/

      Delete
  2. The cooking raser is so evil. LOL
    I pray that the users are not cannibals.

    I never think that a sub can use laser to perform anti-orbital attack, this idea is interesting.

    For the targeting problem, can the sub launch some small UAVs for recon and locating target in spaces and controlling them through some floating antennas?

    ReplyDelete
    Replies
    1. The warships in orbit will be bright and visible, so yes even a small UAV can be used to detect them.

      Delete
  3. A possible option for the RASER system: it should be possible to use phased arrays to focus even RASERs into moderately tight spot sizes. You could have synthetic aperatures of tens to hundreds of km.
    That said, the (mobile) submarine platform might not necessarily be the best option. It should be possible to install numerous localised arrays in semi-deep waters, which can be driven by even more powerful reactors. Any single unit might be sufficient for some applications. The localised arrays could interact with synthetic aperatures of hundreds of meters. Multiple arrays could link to provide the 100 km+ synthetic aperatures. These would be stationary, but quite difficult to detect (accurately enough to target). Also, there are a number of emerging technologies that might yield mobile submarine platforms vulnerable to detection, even from space-based platforms (such as techniques that can measure minute displacements of seawater).

    ReplyDelete
  4. Synthetic apertures, if I understand correctly, move the lens around to cover a larger area and collect as much light over time as a large lens of the same area would have done instantly.

    As a weapon, I don't see how it could be effective. Laser/Raser weapon damage comes from the intensity of the beam. Moving the lens around won't increase the intensity on target at any point.

    Or am I misunderstanding what you mean by synthetic apertures?

    ReplyDelete
    Replies
    1. There are two different types of synthetic apertures. You accurately describe one used for image resolution of non-moving objects. However, the term also applies to interferometric phenomena using multiple instruments spaced over large distances to act as a single lense with the same baseline.
      Interestingly,although the technique is usually used for image resoluton, my understanding is that it has also been applied to reducing laser spot size. Not only do you get the benefit of the combined intensity from multiple beams, but those beams interact in such a way that the combined spot size is smaller than the spot size from a single unit.

      As a side note, as I understand it, the synthetic aperture formed by the moving lens takes advantage of a kind of mathematical interferometry, rather than the "real" interferomtry of simultaneous emissions.

      Delete
    2. That would have pretty impressive consequences on the effectiveness of ground-based anti-space lasers. If I understand correctly, it works like this: https://en.wikipedia.org/wiki/Astronomical_interferometer

      I wonder how efficient the interferometer-laser is compared to a full-sized lens of the same diameter. Is it a simple ratio of area covered?

      Delete
    3. Basic idea, except in reverse. Most importantly, all astronomical interferometry is actually mathematically based, while laser interferometry actually (re)combines the actual photon streams.

      The principal is exactly the same as your average diffraction grating projection. Importantly, even though actual laser interferometry (mostly used for making precise measurements of displacement and surface deviation --mapping) generally involves splitting a single laser beam into two separate processing streams (which are later recombined), the process works just as well (although probably a little more complex) using independent source beams.

      I have been trying to locate my source material for our particular application of laser interferometry, in order to provide more precise answers to your questions. However, I have not been able to find any actual articles, so I might have picked up the information straight from more informal, theoretical, discussions. That said, I will do my best with what I understand of the subject.

      Since laser beams diverge, the pathways of the beams will be of slightly different lengths at different radii. This means that there will be a related shift in phase, dependent upon the distance from centre.
      A second laser will have the same characteristics. In principal, this should result in diffraction patterns forming between areas of the two beams in-phase (constructive interference), and out-of-phase (destructive interference), with one another.

      Now, as I understand it, the destructive interference patterns can not erase the energy from the beams, even if there is no irradiation at these points. This means that all of the combined beam energy must be concentrated at the points of constructive interference.
      Furthermore, again, as I understand it, the interference can not take place outside the scope of the two beams, which means that the interference patterns must be confined to the original spot size(s) of the independent beams.

      Note: there are some factors that might affect the results. For instance, I do not know what effect polarisation might have on the interference patterns. I don't know whether or not there are certain conditions that must be met in order for the laser beams to interfere with one another. In short, to the best of my knowledge, laser interferometry has never been studied on macroscopic scales.
      Likewise, in astronmical interferometry, at least three points are required in order to construct a resolved 3D interferometric image. I don't know whether or not this means that you need at least three lasers to reduce spot size... although I could imagine that two lasers might produce a "bull's-eye" pattern, while a third is required to reduce this to a single reduced spot.

      Delete
    4. I will call in Luke Campbell and Rhys Taylor on this to see if they have anything they can contribute. I know even less of the subject than you, but as far as I can recall, constructive/destructive interference are the operating principles that phased arrays rely upon, not interferometers.

      Delete
    5. Technically, phased arrays are a subset of interferometry. Or, rather, a specific application. In physics, interferometry is any sudy that abstracts information from the analysis of interference patterns formed by waves (the name literally means 'the measurement of interference').
      However, yes, Luke in particular would probably have more of an understanding of the factors involved in long distance laser interferometry of this type.

      Delete
    6. I got some interesting answers. Here's the link: https://plus.google.com/+MatterBeamTSF/posts/PjqegQ6gda6?fscid=z12fyjxyyzmjzxy4b04cgpcgvy2ldpkpddg.1508421596420723

      Delete
  5. Interesting. I've considered submarines as a defense against spaceships before, based on the fact that a submarine is pretty much invulnerable to space-based weapons, while a spacecraft is certainly not invulnerable to ground-based weapons.

    I've only considered missile-armed subs though, and although they might be useful, lasers have a lot of promise since they don't really need resupply. Missiles are more of a one-off strike (or a limited number of volleys, anyway - I'm assuming targets with heavy defensive capabilities) and may not actually need a submarine. Trucks would do well enough - you disperse thousands of those across a continent, and they're difficult enough to find before launch and mostly worthless as targets afterwards.

    In my worldbuilding (concerning a setting with decidedly unscientific FTL and hence interstellar empires) I've figured planetary defenses like this are more common on planets relatively insulated from interstellar politics. The more connected ones know how the game is played, in that it's easiest to just accept your new masters in place of the old ones so that the war stays in space and not on your planet, and not attempt to fight back in a manner where you will most likely lose (because you're up against the resources of an interstellar empire) and where you will take all of the collateral damage. On the flip side the empires lack experience and specialized hardware for dealing with this situation due to its rarity. This is all very setting-specific, of course.

    As a completely unrelated aside, I saw some interesting discussion on my ship designs (the Eastzone and Nikto-ega) in the comments of some earlier posts! I did write a reply to your and Thucydides' thoughts, but I'm pretty sure the Internet ate it.

    ReplyDelete
    Replies
    1. Thanks. Glad to find you here.
      I'm sorry about your comments, I have no record of them going through.

      Being at the bottom of the gravity well and under a tall atmosphere gives submarine-launched missiles a big disadvantage against their spaceborne counterparts.

      I'd be glad to discuss the setting you have. Interstellar propulsion based on fictional principals should be no issue as long as the rest of it makes efforts towards being realistic or at least internally consistent. Feel free to comment here, on other posts or directly by email, whichever way suits you best.

      I can't comment yet on how planetary defenses are set up in your settings as defenses are dependent on the attacks they are designed for, and attacks across interstellar distances are dependent in turn on the specifics of the FTL drive.

      I liked your designs :)

      Delete
    2. Thanks!

      FTL is the only plain impossible thing I have. The engineering I try to have make some amount of sense, but much of it is really pretty arbitrary because it's somewhere between hard and impossible to know how close to physical limits it's plausible for a given technology to get - so I pick some point between current estimates and theoretical limits that gives me the desired effect and that's at least physically plausible and internally consistent.

      Anything in particular you'd be interested in hearing about? I could talk about the wider political and social context (the reason they have big exciting spaceship fights rather than just glassing planets wholesale has to do with this rather than any technology), space combat, spacecraft design in general, or something else... Maybe I should start a blog.

      Delete
    3. As an aside: It appears commenting doesn't work on Chrome or Chromium. I've ran into the same issue on two machines now. Other browsers seem to do fine.

      Delete
    4. If you want to strike for more realism, have the FTL's influence on spaceflight be removed as much as possible. For example, have the FTL deposit spaceships far away enough from planets for 'deep space' to matter.

      A blog might be a good idea if you have a lot of material, but traffic comes with regular posting. An alternative is to prepare a 'draft version' of your setting and post it on forums such as Spacebattles, the worldbuilding reddit and other creative places to get people discussing it. If you post on Google+, Winchell Chung, creator and maintainer of Atomic Rockets, will surely re-share it. A lot of people reading through your work might pick up on some flaws or provide strong suggestions you might have missed.

      If we are to discuss your setting, I can provide commentary on economic/financial topics, and of course, the scientific and especially military technologies in use. I'm certain other readers of this blog are interested too.

      Delete
    5. Yeah, that's pretty much how I've set up my FTL. You can only enter and exit at a certain distance in the outer system. I measure the distance to the FTL "shell" by gravitational acceleration from the star, but that's just to have a more or less consistent standard (I just ignore edge cases that might break the system) for an arbitrary rule intended to connect systems in a way where travel inside them is still all rocketry.

      I should definitely write up some stuff for people to read and comment on.

      Delete
  6. For something with high power density that can be made in the near future see the Dual Fluid Reactor. "2 meter cube for 1 GW"
    https://festkoerper-kernphysik.de/dfr.pdf

    ReplyDelete
    Replies
    1. Very nice find!
      While power per volume (kW/m^3) is of low concern to spacecraft, maybe a bit more so in submarines, when compared to power per mass (kW/kg), the paper has some very relevant information.

      Most interesting is the SSAS method to produce high energy density fuels our of water and nitrogen. This reminds me of the concept I described for converting the vast electrical output of potential fusion reactors or renewable resources into hydrocarbon fuels to make it easy to store, transport and use.

      http://toughsf.blogspot.com/2016/06/saving-planet-with-alcohol-ethanol.html

      Delete
    2. Actually, when dealing with nuclear reactors, power per volume can be VERY important. This is because the volume determines the size of required shielding, the resulting mass of which is a critical concern to any space-based propulsion system. There are also a number of other interrelated factors, such as the amount of resources required to maintain such systems... or the area you have to protect against attack.

      Delete
  7. I have an idea.

    Since the subs use towed laser systems, the lasers can be used or concealed as part of sea-based laser launching network, connected to OTEC powerplants, floating/coastal settlements or facilities.

    They can be repurposed as weapons if battles occur on orbit. In the case that orbital bombardment is imminent, towed lasers are disconnected from the network and the subs pick them up...

    The birds in space won't dare to fly across the sea, as sharks would bite them, with laser teeth.

    ReplyDelete
    Replies
    1. I think you would be interested in reading the conclusions of this blog post's discussion on the SpaceBattles forum:
      https://forums.spacebattles.com/threads/anti-orbital-laser-submarines.577403/

      If I'm reading this right, you want the floating laser lens to be mounted on fixed sea-based installations for use as a laser launch system, but allow them to serve double duty by making them detachable and towable by submarines?

      Delete
    2. Yes, a sea-based and partially mobile version of laser weapon web mentioned in older post.

      If the towed lasers are constructed as weapons openly in first place, it may deter some enemies, but it can't really surprise the enemy.

      And the public or officials may oppose to construct some submerged mirrors that may only be towed undersea for the rest of service life.
      Though it is better that a weapon doesn't need to be use in its life, that is also wasting the potential of the laser. It is not something very specialized like nukes or missile that they can do nothing but blow up the targets.

      So, why don't construct them as sea-based laser launch device at first? Anti-orbital attack is secondary ability, on paper at least.

      Delete
    3. The laser weapon web mentioned in your older post.

      Sorry for the typing miss in the last post.

      Delete
    4. If the laser weapons are made to double as a laser launch system, then there is every argument to reduce costs, ease maintenance and facilitate expansion by installing them on clean, dry land. Many of the features that make the towed floating lens a good weapon system (accuracy, low detectability, ability to operate in rough seas, ability to re-direct beams from moving submarines) are useless for a civilian system.

      If there is a laser launch system, it will be built on ground. But, I concede, there may be promise if the military builds its own towed lenses but makes them capable of exploiting the laser launch infrastructure, such as hooking it up to the civilian laser generators through fibre optics.

      Delete
    5. As the user, the government should pay and build their military-grade towed lens of course.
      Maybe the idea mentioned above is messy, since I typed it immediately after the idea popped up in my mind, I assume that the towed lasers are built and owned by government. They are connected to civilian systems and operated by civilian launch administration(?) in peacetime. The military controls them only in the period of war.

      Delete
    6. @Matter Beam

      Your argument about using land-based systems might not be entirely valid.

      *Land is very seldom "clean". For land systems, it is almost always necessary to clear away and otherwise prepare large swathes of ground. It is problematic enough trying to find patches that are clear, and sufficiently level (or to try to build platforms on sloped land to support facilities). It becomes REALLY messy when you try to incorporate the required infrastructure.
      Sea-based structures do offer some engineering headaches, but this difficulty is often more-than-offset by the number of options for providing the necessary infrastructure.

      *Land platforms are inherently difficult to move, if/when necessary.

      *Building platforms at sea offers a reduced human footprint (again, you do not need to clear away those large swathes of land, which often go hand-in-hand with ecological harm). It also offers greater system flexibility, which is excellent for civilian projects. Another important civilian aspect is that building at sea very seldom runs foul of zoning ordinances, or people complaing of eyesores in their "backyards".

      Etc.

      Delete
    7. In the Laser Launch thread, I was of the opinion that the most realistic path to installing the necessary infrastructure would be to incrementally add laser-beaming modules until you started to have enough lifting capacity to start replacing other launch alternatives, even reusable chemical rockets.

      The small laser beaming units won't have to all be on level ground, just have a clear line of sight to the rocket until it doesn't need the lasers anymore. These requirements are not too hard to meet (maybe not in a mountain valley in Tibet).

      Connecting the laser modules to a power grid is most easily done on land, where overland electric power lines can do the job fine.

      Also, it is very likely that the laser modules will be built for reliability, ease of maintenance and a modular swap in/swap out method of dealing with part failures. This is antithetical to the enclosed and remote floating platform that must be connected by insulated wires and can only be repaired by seamen/engineers that can only work in calm weather. There are many good reasons for building on land, and the result is that we don't have large electrical-power-consuming structures floating on the oceans.

      Moving... well, the launch site will be optimized for the inclination it is supposed to send rockets into, so it won't be a concern. Doubling as a weapon though... it is hard to hide the laser-launch-slash-weapon system that started out a civilian project, so might as well prevent it from being attacked at all or not relying on it to survive the first wave of attacks.

      As for human footprint, well, eh, hard to comment on that as it is very subjective. Maybe we can make the argument that we're saving on thousands of tons of hydrocarbon fuels being burnt as rocket propellant, we can replace them with 'clean air-based propulsion'!

      Delete
  8. Well, if the main weapon is laser, then why use a submarine? A surface naval warship would easily outclass any space warship in laser duel simply due to cooling consideration. Naval ship have the whole ocean to cool her lasers, after all. And she could have much better protection.

    Granted, the surface warship is easier to detect... but not so easy to identify, if we have sufficient naval traffic. Even now, with great sattelite coverage, carrier battlegroups routinely evade detecion just by merging with civilian traffic. It isn't actually simple to identify the naval ship from orbit in realtime, even if weather is good enough.

    And, the surface ship could defend itself actively. It is mobile enough, so just the rock-dropping would not work. And missiles smart enough to detect and home on modern warship... I doubt that they could be build TOO FAST to be impossible to intercept. Again, we are talking about warship, which is localized, not area target. Megaton hit with five km CEP could kill a city, but could not kill a warship.

    ReplyDelete
    Replies
    1. The only thing between a surface warship and a space warship is the atmosphere. Laser attacks will damage them both, with perhaps surface warships having an advantage from using the ocean to cool themselves down.

      Submarines are able to hide underwater until they decide to attack, and when they attack, they can do it while hidden and protected by several meters of water. Instead of just a cooling advantage over space warships, they will have a protection and concealment advantage too. Plus, every attack by a submarine is a surprise attack.

      In the scenario being discussed, there is an enemy in orbit and defenders on the planet. I did not want to complicate the discussion with operational tactics, military strategy, rules of war ect... it is up to the author to integrate the anti-orbit laser submarine into their setting any way they like.

      For example, if the space warships in orbit are from another star system and hate everyone on the planet, they will destroy every moving vehicle they see. Civilian and military surface ships will be sunk - submarines will escape this destruction. Or, maybe we are in a near future setting with space warships of one country declaring war on another country on the same planet. The space warships won't be allowed to just shoot down any civilian ships they find suspicious, as it will provoke other countries to declare war on them in turn. In that case, hiding naval ships in civilian traffic might work... but it might also anger the country those civilian ships belong to.

      The problem with surface warships is that even if they travel very fast and can dodge orbital drops and survive near-misses with nuclear weapons, they cannot hide. Merging with civilian ships works only once. The space warships orbiting overhead will pass over the navy ship again and again, and with lasers you never run out of ammunition. A navy ship can have its defenses stripped by multiple laser passes until it cannot defend itself from a killing strike delivered through missiles from orbit.

      Delete
    2. It is actually rather questionable how long submersible stealth will remain affective, especially for large vessels. There is already discussion in the military about whether or not the current submarine forces, as they exist, should be phased out. In particular, it has been discovered that large submarines produce measurable swells in the ocean surface as they move, and there are existing satellite resources that are capable of detecting these swells (although currently too few in number to be tactically useful, and there is no current way to clearly identify whether a target is an enemy or ally... or a large whale).

      Also, keep in mind that laser armed naval vessels and spacecraft will be on a somewhat equal footing. Neither can hide from the other. You might actually have an overall advantage with the naval force, as there is a greater option for mobility (the z-dimension mobility is not particularly useful here).

      Delete
    3. The primary threat to submarines today is other submarines. In a surface vs orbit situation, the submarines are hidden by several meters of water and a turbulent atmosphere.

      Even if the orbital forces have the ability to detect every swell there is, the submarine and the spaceship are absolutely not on an equal footing. The spaceship cannot loiter, is in full sight and has nothing but its armor to protect itself. It cannot continuously dodge and every move or attack it makes is clear and visible.

      The submarine is hidden until it strikes. Sure, accumulating a number of possible detections in a row will give you a good idea of where the submarine is lurking, but you won't be able to fire back until it decides to attack, and any counterfire will cease to be effective as soon as the submarine withdraws... and even then, the floating lens connected by optical fibres means that the submarine doesn't have to really expose itself to attack.

      Delete
    4. "The spaceship cannot loiter, is in full sight and has nothing but its armor to protect itself."

      What about nuclear depth-charge armed helium steamers?

      Delete
    5. That complicates things slightly. The submarines cannot be a 'last line of defense' if they don't have targets to shoot!

      However, it isn't too bad. It is very difficult for hydrogen steamers to stay hidden in low orbit. They have a bright planet underneath and cold space above. Any sensor looking down from above would spot them. Also, their positions would be immediately revealed if they use lasers. So, the hydrogen steamers would be forced to rely on deploying missiles.

      The problem is... you can't take out a planet with only the missile load of a couple of spaceships. They're sufficient for taking out space stations or enemy fleets... but the entire planetary defense system mounted on the surface? You'd need to reusable effectiveness of a laser.

      Worse still, without lasers, the hydrogen steamers in orbit won't have a quick reaction ability. They can't shoot down a jet fighter that appears over the horizon, or destroy an anti-satellite missiles as it is slowly accelerating into space...

      Delete
    6. *The LASERs traversing from sub-surface to space and from space to sub-surface will have to pass through exactly the same amount of mass.

      *Every second that the space vessel is exposed to the sub, the sub is exposed to the space vessel.

      *In both instances, the LASER beams will be more manoeuverable than the vessels in question. It really does not matter that the sub is more manoeuerable than the space vessel. The real concern will be which asset will have the greater precision.

      *The subs WILL have the advantage of the first shot, so long as it is not in motion. The problem is, once a sub takes that first shot, it leaves an IR trace leading straight back to it. Its displacement rate is far too low to go back into hiding, especially ince that displacement will create trackable swell. The space vessel has greater displacement rate, and actually has excellent manoeuverability, given sufficient propellant (granted, the manoeuverability is largely limited to "great circle" deviations, as other manoeuvres have complex consequences).


      Now, addressing the towed array (I was actually addressing the issue in the context of the sub as the weapons platform): the space vessel doesn't care less about the sub, per se. The space vessel is concerned with the weapon. The sub can be as stealthy as it wants. It doesn't matter because once it starts towing around a large weapons array, it loses pretty much all its stealth (unless that array is the equivalent of a pea-shooter).


      Addressing the issue of a stealth spacecraft: yes, the space vessel has a lot going against it, in terms of maintaining stealth.
      Firing LASERs is NOT one of them.
      Unlike the sub, which will leave an IR trace as soon as the LASER beam interacts with water and/or air, the space-emitted LASER leaves no trace until it enters the atmosphere. Also unlike the sub that has a very low displacement rate, once the space vessel stops firing, it will be several km from that position within a second.

      Delete
    7. The first arguments depend on whether or not laser counter-fire is possible before a submarine dives back below its attack depth. With targeting solutions taking a finite time, as well as adjustments for atmospheric interference, and also supposing that the laser sub's attack is supposed to destroy the target's ability to retaliate or just eliminate it entirely, this is not a possibility to be ignored!

      Subs will leave a trace, but I envision it as two snipers hunting for each other. One sniper (the sub) is hidden and has the second sniper in his crosshairs. The second sniper (the spaceship) is sitting in the open and is waving his gun around. The first sniper reveals his position with a muzzle flash and a bang when he sheets, then has to run away... but will it matter if he hits?

      Towed arrays don't have to be un-stealthy. They can be UUVs even stealthier than the mothership.

      Stealth craft in space cannot use laser weapons for more than a few seconds without revealing themselves, nor can they quickly accelerate out of the way. They are a special case though.

      Delete
  9. Anti-orbital submarine is a fascinating idea. The obvious counter would be to drop nuclear depth charges from orbit, as the area of effect of a nuke is far lager in water than in space. I wonder how much a submarine can be shielded against the shockwave?

    Now, it seems to me that you are underestimating the effectiveness of anti-satellites missiles. There are ASAT today that are launched by jet fighters or from surface ships, and those are far smaller than ICBMs. After all, they only have to reach orbital altitude, not put themselves in orbit. As such, they require far less dV.

    In fact, judging by the size of sounding rockets, it may just be possible to develop a man-portable low-orbit ASAT. Probably with a small solid rocket motor as first stage, a ram-augmented rocket or ramjet as second stage and maybe another solid rocket as third stage, complete with a guidance system and small attitude thrusters. If you can build one of those for less than 20 kg with its casing, then you can make it man-portable.
    Bar that, you can make them fit on a pick-up. ASAT-technical...
    And against high orbit, it could still fit on a truck, like a mobile ICBM. After all, they don't need any payload apart from themselves as kinetic impactors. So they could reach pretty high!

    Now, those would probably be relatively easy to defend against, taking entire minutes to reach their targets. But they are cheap, hidden until they fire, their launch sites can be expendable (get operators to leave the truck vicinity before launch!) or put on deeply immersed submarines...
    Given the cost of building and sending those warships, the defenders may still be at a big advantage. Though between multi-world polities, concentration of force may still counter that.

    Another option is Verne guns with terminal guidance, or Verne-assisted missiles. Those guns don't need to be that gigantic to reach orbital altitude. In fact, you could probably reach a quite high orbit with a spinal-mounted gun on a submarine, or even with a train gun.

    This probably means that an attacker would stay in relatively high orbit, as the higher they are, the better they can defend themselves against projectiles and the fewer of those can reach them.
    Which means that lasers may still have a big place: at those range, even weakened, lasers become more reliable than projectiles.

    ReplyDelete
  10. No defence is going to be 100% effective, but the true advantage of using cheap truck mounted ASATS or submarines mounted laser weapons is to complicate the mission of the attacker.

    If LEO is too dangerous, then the defender is forced into higher orbits, with longer orbital periods for ships or Walker constellations and more gaps opening up in the coverage of the ground below. Sensors will be farther away from targets, so it will be easier to for ground targets to become lost ion the clutter. Weapons will have to be corresponding are powerful ti reach targets at greater ranges, and physical projectiles like nuclear depth charges or "Thor" rods entering the atmosphere will have longer flight times, allowing countermeasures to be deployed or targets to disperse.

    Now all these things can be countered, but each counter adds another step and more resources for the attacker. Can he *really* afford to bring RBoD's with atmosphere penetrating lasers powerful enough to strike ground targets from the Hill Sphere? Can he bring *enough* to provide global coverage? If he can, does that also equate to having enough of the other types of ships needed to actually control the planet, deliver troops or collect tribute (or whatever the actual mission is?)

    So it is well worth putting some resources into this sort of defence.

    ReplyDelete
    Replies
    1. Exactly this. Anti-orbit laser submarines, ground-based ASATs, they don't really have to be perfect or even very effective. They only have to force the enemy into disproportionately devoting resources towards defending against or countering them.

      For example, without AOLS, an attacking fleet can just descend into low orbit after achieving space superiority and start a laser bombardment campaign. Pinpoint strikes against every vehicle, starting fires in forests, wiping out livestock and food supplies, melting railways, blowing up fuel depots and destroying infrastructure. No nation can survive that sort of onslaught for long, even if it managed to keep some defenses hidden away as a last resort.

      With AOLS, low orbit is denied. Laser bombardment becomes less powerful, less accurate and generally less of a threat. If the attacking fleet is warded away far enough, space forces can be launched without fearing being shot down on the launchpad.

      Delete
  11. When I am reading the posts of Next Big Future, I find this.
    Laser subs have a new enemy

    https://www.nextbigfuture.com/2017/10/stealth-fighters-submarines-and-aircraft-carriers-could-become-ordinary-weapons.html#more-137723

    ReplyDelete
    Replies
    1. Interesting read, especially the part where submarines are being detected at short ranges using lasers and LEDs....

      Nontheless, most of the methods mentioned only work when both sides of a war have access to the same ocean. In an orbital vs surface scenario, any attempt by the orbital forces to use HF radar on aircraft or laser light on submarines won't work due to the distances involved or the properties of the mediums (ionosphere, ocean) traversed. If they drop a sensor pod from orbit so that it shares the same environment as the target, it can only be done in a flashy and visible manner including a de-orbit burn, and a hypersonic braking pass through the atmosphere. It will be detected and the submarines will move out of the way.

      The way the sensors might defeat the submarines is if they are dropped en masse all around the world's oceans so that submarines have no-where to hide... but imagine the mass penalty and logistical difficulties of accomplishing that task!

      Delete
    2. Assuming this is an interplanetary war or interstellar war, what if the offensive fleet captured an asteroid and used it to produce tons of small sensors, then dropping them down?

      The fleet doesn't need to carry that amount of sensors, the factories or 3D printing facilities can produce other materiels.

      ISRU is a must for long range exploration, I think it is the same for interplanetary wars.

      Something off-topic, can I send some of my world-building ideas and questions directly to your Google+?
      It's because some questions are (still?) not discussed at all posts on this blog.

      Delete
    3. The defending planet, even if it is just one country, will vastly outproduce anything the attacking fleet can bring alone. Unless the attacking fleet can haul an entire country's worth of industrial capacity, it should not rely on mass production.

      Instead, the attacking fleet must concentrate its forces into an overwhelming assault that defeats everything the defender has right now. It then removes the defender's ability to compensate for losses over the long term by targeting infrastructure... of forcing a surrender. Not every way is an absolute struggle to the death after all.

      There might be ways to cover an entire ocean with sensors. If you have access to nanotechnology, you can sprinkle the surface with dust-sized robots that work together to sweep the seas. If you have access to terawatts of power, the mass of the sensors needed to cover the area might not be such a big problem either.

      You can contact me on toughsfmatterbeam at gmail.com or write to me on Google Plus.

      Delete
  12. This leads to another question.

    If you have enough resources to blanket the global oceans with sensors *and* a have a fleet capable of conquering the space environment from LEO to the Hill Sphere, why are you bothering to go to war in the first place? That sort of resource base makes you capable of engaging in almost any sort of profitable activity you can contemplate, without referring to other polities or economies at all.

    This is perhaps the biggest "McGuffin" in demi operatic space combat scenarios. A true space civilization haas access to resources and energy on scales we can hardly dream of. If they want to "defeat" an opposing or rival polity, they can outproduce them and blanket them with their economic superiority, much like the United States won the Cold War via economic expansion. The Reagan tax cuts powered economic growth so large that the GDP increase was equal to the *entire* GDP of the Federal Republic of Germany (already considered a major economic power in that time frame). Russia's GDP is about the size of Spain's, and the USSR's wasn't much larger.

    Things like the military build up and the Strategic Defense Initiative were waves of the cape to entice the Soviet Union to push more resources from its already malformed economy into the military (unofficial estimates from that era suggested that up to 25% of the USSR's GDP was consumed by the military, when you add in indirect expenditures and the various secret and black projects, whereas the US military expenditure was @ 5% of the GDP, and even adding in everything that could conceivably be "military" would bring you up to maybe 10% of the US GDP.

    So that leaves things like ideology and religion to instigate the reasons for going to war, and since these are not susceptible to reason (think of the Imperial Japanese Cabinet considering continuing the fight even after the *second* atomic bomb was dropped), you now have the conditions for not just space war, but massively destructive space wars with no mercy, pity or objective other than total destruction of the enemy (think of the 30 Years War or the Albigensian Crusade).

    Of course this scenario also means no space fleets per se, you build massive mass driver canon and fling rocks at enemy moons, planets and free space structures at the greatest velocity that you can manage, and let kinetic energy do the rest.

    ReplyDelete
    Replies
    1. My reading of historical causes for war make the Cold War an exception, rather than the rule. Having one side (USA) completely dominate its opponent (USSR) but decide not to wipe them out is a situation possible only because of the existence of nuclear warheads.

      In the past, massive empires such as Rome and Carthage went to war despite Rome having enough wealth and trade capacity to muscle its competitors out of any market...

      In my opinion, the majority of space war will be small conflicts that have a small number of players and limited battles that try NOT to wipe out their opponent, but to 'continue diplomacy through other means'. They might decide that military action is more cost effective that bargaining or suffering trade losses for the foreseeable future. There's also the fact that information on competitors is always lacking and leads to spurious analysis (see the 'Bomber Ga'), or the enduring myth of the 'short and victorious war'.

      Simply put, war is a messy, complex and human affair that will never follow the clear cut rules of economics. It is not rational to destroy anything, and yet we do it all the time and certainly without ideology and religion being at the forefront.

      Delete
    2. Even there are resources and energy on scales we can hardly dream of, they are still limited in amount and not something can be obtained easily.
      That can still spark up a war if the situation turns ugly enough.

      Delete
    3. My thinking on this matter is that the space environment will be quite different from what we have previously encountered, so the rules will become different.

      On the one hand, time, energy, resources and room (or volume) will be expended by orders of magnitude compared to Earthly conditions. On the other hand, people will be living in artificial environments which are quite fragile (relatively speaking) compared to a planet. You can visualize this by thinking of people living in soap bubbles suspended in space. The final factor is the amount of energy people will be using is also vastly beyond what we are used to seeing or thinking about. Getting from point to point in the Solar System requires changes in velocity measured in kilometres per second, speeds that make paint flecks potentially dangerous weapons (Space Shuttle windows were replaced because of paint chips striking them in orbit).

      So now picture a situation where you are walking through a room filled with floating soap bubbles while carrying a lit welding torch just to do your day to day routine. Any careless action on your part will pop a large number of soap bubbles. Deliberate action on your part could "clear the room" unless someone else with a welding torch (or something more serious) takes steps to oppose you.

      So while the impulses which lead to war will still exist so long as we and our descendants remain human, people living in space colonies, or even planets, will be facing a situation where *everyone* has the ability to harness energies on the order of nuclear weapons. The Chelyabinsk meteor exploded in the atmosphere with an estimated 500Kt of energy, larger than many strategic nuclear warheads. Imagine a cargo pod of similar mass being sent on a one way journey and deliberately aimed at a space colony or the surface of a planet or moon...

      So demi-operatic space war is already operating in conditions at the far edge of our experience and ability to truly comprehend. I suspect that the most common forms of space warfare will not be grand constellations unfolding in the enemy Hill Sphere, but rather carefully planned campaigns of economic manipulation, sabotage, propaganda and cyber attacks, with the occasional direct action by infiltrating teams of saboteurs, assassins or SoF personnel disguised as students, businesspeople, tourists or deep cover agents infiltrated into the polity years before.

      As always, YMMV.

      Delete
    4. My mileage says that for the vast majority of humans living in space, the dangers of high-energy events will be far removed from the environments they live in.

      The largest population centers will be under thick atmospheres or entirely underground. Gigaton impacts aren't so scary when there's meters of ice and stone over your head.

      I don't think large, orbital habitats will become popular. The risks are too high, the running costs too high, the basic energy and travel penalties to moving anything to the station and back amount to a very inefficient habitat when compared to solid surfaces.

      But, there are always people who must go into the void and live for long periods in 'soap bubbles'. They, however, do accept the risks of undertaking the journey, unlike the majority of the population that don't face the same risks when staying at home. It is a fundamental difference.

      We, as a species, always gravitate towards stability. We have lived through times when a variable number of actors were given the power to raise enough energy or forces to bring down large scale destruction, without any sort of limiter or control other than their peers. It can be literal amounts of energy, such as the atom bomb, or comparable dangers in the form of large armies. The Viking, for example, lived for centuries as scattered tribes and Houses that could at any moment march their men to the neighbouring village and lay waste to it, or ride their longboats to foreign shores and bring about sudden, unexpected and unstoppable destruction. But, as time goes on, such power is always held in check, regulated and neutered. The same will happen for cargo pods of the future.

      Another concept I have in mind is that of the inverted pyramid of war. At the widest base at top, we have the 'soft power' conflicts that fit well in cyber-thrillers and spy novels. They influence the small scale wars and skirmishes, by preventing them from happening or triggering them intentionally or not. The latter are much less frequent but much more devastating. For example, a large cyber-warfare campaign precedes a relatively minor movement of Russian forces into neighbouring Ukraine.
      In the third step we have the small, hot conflicts that litter our present-day headlines. They are civil wars, internal strife, asymmetric wars where one side has no fear of actually losing and unofficial wars between nations who don't want to escalate. These wars can be defined as the result of a bunch of smaller scale skirmishes boiling over. Finally, at the lethal tip, we have the major wars. Today, we do everything we can to avoid them. We have the UN, NATO, the EU all struggling to make sure that no-one dares allow a minor conflict to bring forth the full might of the US, Russia or other nuclear-armed nations.

      How does this apply to space warfare?

      Well, your description of carefully planned campaigns and sabotage fits well in the upper tier of wars. They are numerous, but do not threaten anyone with major loss of life. In space, the difference will be that the pyramid has steeper sides. The lethality increases drastically with each tier, and so its frequency diminishes accordingly. I hope this all makes sense.

      Delete
    5. If there is so many people living in space, the space traffic should be closely monitored by many parties.

      A high-speed stray cargo pod that was intended to be a KEV maybe spotted very easily, then the anti-orbital guys will do the rest.

      If the anti-orbital divisions are not composed of blinded Scarif Garrison, that should be no problem. LOL
      If there are so many pods heading to the target, intel agencies may notice some signs of preparation, like someone acquiring pods and fuel for unknown purpose.

      For the BlackOps taking a bigger role, I agree with this. However, the party to deploy BlackOps may needs means to infiltrate the settlement or facility and exact the team after mission.

      Small craft of some kind is still needed, they maybe based on some larger ships. In order to shoot down or chase these enemies, you still need a space fleet.

      Matter Beam, I sent the message to you on Google+, I abandon G+ for a long time and forget many functions, so I am not sure can you receive it.

      Delete
    6. to infiltrate the settlement or facility and extract the team after mission.

      Sorry for the typing miss.

      ===========================

      If building a large space habitat from scratch costs too much, living on moons or planets has to deal with the long term damage from lower gravity, how about putting rotating habitat on asteroid instead?

      Delete
    7. I think spacecraft, at least for civilian purposes, will be a cross between jet airliners, driverless cars and medical equipment.

      They cannot depart without a flight plan, authorization and full tracking from departure to arrival. Every major function will be under automated control, with that control able to be overridden by the Transport Authorities. Like medical equipment, there will be extensive certification, testing and controls performed to make sure that the automated systems and especially the 'kill switch' works as intended. The kill switch, as its name implies, is a remote-activation device that ... ah, helps spread the debris of the wayward spacecraft over a safe area.

      Now, what about malfunctioning spacecraft? Well, just like airports don't allow an airliner to take off if one of its vital sensors failed, the spacecraft won't be allowed to reach lethal velocities without being under the full control of its certified, third-party-signed and continuously monitored AI and/or the transport authority.

      But, there is still the risk of malicious actions made to circumvent all these controls. This will fall under the domain of police and criminal action. A tragic parallel is the multiple vehicular ramming attacks that have occurred recently. How are they prevented? Engine kill switches, barriers and simple investigation and prevention work. In space, this will be forced maneuvers, kill switches, laser defenses and police work.

      Felix, I received and replied to your G+ post.

      Large space habitats shouldn't cost too much, especially with inflatable living spaces. The problem is that getting them to the point where they sustain large populations makes them cost drastically more over time than a surface base. How easy it is to transport supplies to an orbital station changes these assumptions somewhat.

      Delete
    8. The kill switch is a real nightmare if the story takes place at a world that ruled by some oppersive regimes. Even in relatively free society, there is still risk that the kill switch can be abused, killing "undesired" people and cover them as accidents.

      I don't think this kind of regulation would be passed easily.

      A more acceptable version may be "spaceport guards are authorized to shoot down any non-responsive spaceships over a certain velocity that heading to any ships, facilities or the port, after the ships passed a certain perimeter in the space surrounding the port".

      Delete
    9. Its an unpopular solution to the trolley problem. You have a switch which kills a spaceship's crew but prevents the deaths of potentially thousands of people when the spaceship crashes. Governments won't hesitate: kill switches and you sign a disclaimer.

      Of course, there will be a variety of responses. Some friendly ports will try to walk you through troubleshooting or try to remotely control your ship. Others will kill you and then shame your family.

      And... hostage situations in space just got a whole lot sleeker. I can imagine a variant on the Hold The Phone plots (https://en.wikipedia.org/wiki/Phone_Booth_(film)) where the spaceship is held hostage by someone who has access to your kill switch.

      Delete
    10. @Thucydides
      This is a rather narrow view that misses A LOT of subtleties.

      First, there is very seldom a condition of one motivation OR another. Rather, you almost always have a large number of motivations, that are differentially balanced in different persons.

      Second, even ideological/religious motivations are seldom absolute. Total destruction might not be an option if, even though conquest in battle is permitted, there are cultural/religious barriers to destruction. Or perhaps the enemy is only attempting to suppress the possibility of meaningful threats by an unknown culture. There is also the possible incentive of converting a culture to the ideology or religion.

      Third, why do the aliens come? Granted, they might come for the sole purpose of destroying any culture that is different from their own. However, it is far more likely that they have a desire for access to some resource that a solar system provides. If you go full out to destroy a culture without mercy, there is a very strong likelhood that you might endanger access to the very resources you require most.


      @Matter Beam
      "Having one side (USA) completely dominate its opponent (USSR) but decide not to wipe them out is a situation possible only because of the existence of nuclear warheads."

      Sorry, if I understand your meaning, this is simply not historically accurate, depending largely upon what you mean by 'wiping them out'. For example, for a number of centuries, Rome was strong enough that it dominated virtually every culture in the region, within the existing boundaries of the empire at a given time, as well as without. Yet the empire actually seldom resorted to "wiping out" the enemy. Instead, the romans would offer a truce, offering the "enemy" to become part of the empire, with full privileges equal to any other member of the empire. Warriors would then either be levied from the defeated to fight for the empire, or they would be offered the "honour" of fighting in the gladiatorial arenas (where, after a sufficient number of victories, they might regain their freedom). Only if enemy warriors refused both of these options would the romans "wipe out" the enemy forces.

      Delete
    11. @Matter Beam

      I disagree with your assessment of kill switches. Unless you are dealing with an autocracy (Russia, for example), the governments are going to be subject to the will of the people. Quite simply, the public will is unlikely to permit the government to impose kill switches under government control.
      Instead, the public, and government officials representing them, will opt for lines of less intrusive measures. This will start out with measures similar to those used in maritime and aviation; notably, instruction, licensing, and routine recertification and testing. The next level will involve intensive and routine licensing, testing, and recertification of the vessels/craft; as well as regulations regarding transponders and communications. I agree that flight plans will probably have to be filed, at least for any vessel operating (or capable of operating) out of strictly defined and controlled navigational zones.
      After that, there will be frequent patrols with "tug boat" capacities, able to physically force a vessel into a desired orbit. Only if that should fail shall destructive interception be permitted as an option.

      Delete
    12. I completely agree with you in fact. I must have been unclear: kill switches will be a last resort after a multitude of other options are tried first, but there will always be something in place to prevent a single impact from wiping out the fragile homes of thousands of people. People boarding spaceships must consent to this final measure being in place.

      Whether the switch is abused or not is a possibility for a scifi plot was what I was discussing.

      Delete
    13. Murphy's Law: Anything that can be abused will be abused. >;)

      IMO, shoot-after-no-response is enough for the last resort and it may be a little bit easier to be accepted by the public.
      If kill-switch law was passed somehow, I can imagine that some underground shipyards may get rich from removing the switch or limiting its ability. Avionics bypassing the switch may spread like wildfire in black or gray market, just like pirated discs in the past.

      Delete
    14. Not sure about that. Today's governments have the power to shoot down any aircraft they want with patrolling fighter jets, and in sensitive places AA defences. It can even be accepted that shooting a plane down if it has been hijacked is the acceptable response, to avoid a new 9/11.

      A kill-switch is simply a streamlined version of this, and one that will probably be almost never used.

      Delete
    15. As this is a blog, I am painting with broad strokes. Like many other would be writers and creators, I am looking for ways to avoid simply painting historical events on the canvas of space, indeed a carful historian knows and understands how changes in environment and technology make events in different time periods different. The Spanish Armada is appropriate to that particular time and place, while the US Fleet sailing towards Midway would be entirely inappropriate transported back to the 1580's (or Fleet Capitana, the Duke of Medina Sidonia would be equally out of place in 1942). This is not only in terms of the obvious mismatch in technologies, but the war aims of the displaced Americans or Spanish, their logistical requirements and even their understanding of the political, religious and social conditions would be entirely out of place.

      Yet, the Spanish Armada and the Battle of Midway are both essentially naval actions and naval stories. So when we consider some sort of military activity in a Plausible Mid Future (PMF) setting, we need to be careful to recognize that we cannot transpose the Spanish Armada or TF-17 into space. The space warships carefully described by Matterbeam in earlier posts are *not* analogues of the USS Yorktown or the São Martinho, cannot be used as Galleons or Aircraft Carriers or attacked by fireships or dive bombers. Another striking analogue is in Atomic Rockets, where it is pointed out a Victorian imagining air warfare would probably describe flying ironclads trying to "cross the T", and never even imagine an Immelmann turn.

      So my feeling is while people will remain people, the conditions of space will be as different to us as air war is to naval war, so the manner in which conflicts are prosecuted will be quite different as well, even when bounded by the laws of physics as we know them.

      Delete
    16. Thanks Thucydides. If you or others wish to expand on points you do not feel the comment section is appropriate for, I am open to guest blogging.

      Delete
    17. Somew thoughts:

      If lasers are used to ward off spacecraft that the kill-switch doesn't work on (the very definitely last resort), what if craft are legally mandated to carry a sail that the laser can fire at to decelerate the craft? Or people get into life boats before kill-switch use and decelerated by orbital defence lasers. *shrugs*

      Delete
    18. It might be developed into a safety system.
      Lithium bumpers positioned at specific points around the spaceship's center of mass, adjusted for the payload it is carrying. The lithium is vaporized by high-energy pulsed lasers to deflect the spaceship quickly - much more so than relying on a laser sail.
      The bumpers impart a specific deltaV in a specific direction to the spacecraft. This creates 'safety zones' within which a station's lasers can deflect spaceships. The zones will depend on the mass, closing velocity and distance of the spaceship in distress... maybe moving outside of your safety zone before the deceleration burn is illegal.

      It's a neat concept. Might be developed more.

      Delete
    19. @Eth,
      Sorry, no, a kill switch is NOT just a streamlined version of an intercept.

      *Kill switches can be hacked.
      *Kill switches can be triggered accidentally.
      *Kill switches can fail; resulting either in failure of kill, or an unwanted detonation.
      *Kill switches are an invasive tech. That is, they must be installed.
      *Kill switches are not as reliable as intercepts: they can be removed by skilled individuals, kill signals can be jammed with appropriate equipment, etc.
      *If intercepts are as effective, and if kill switches are expected to lmost never be used anyway, then there is little argument justifying their presence.

      Delete
    20. If kill switches can be hacked, then the more complicated flight software is even more vulnerable!
      It is easy to make kill switches less vulnerable to accidental triggering and failure than other vital systems on the spaceship. Kill switches are likely to be integrated into a design by the manufacturer. They don't have to always 'kill' either. If the crew can escape in a pod, they will be easier to recover than trying to divert an entire spaceship.
      Intercepts will likely be another layer of redundancy to the options available for preventing a spaceship from smacking into a habitat.

      Although...

      I can envision a strong alternative. Imagine the 'bug catcher'. The bug catches is massive and every habitat has one. It is a big plate of metal backed by a few meters of ice. In front of the metal is a few hundred layers of whipple shielding sandwiched between ice or water. This plate is positioned between the habitat and an uncontrollable spaceship on a collision course. It catches bugs.

      Delete
    21. "If kill switches can be hacked, then the more complicated flight software is even more vulnerable!"

      Um... no.

      It is not a question of complexity. Kill switches (which have been described here as a REMOTE destruction device, specifically because it has been assumed that the crew could not be relied on to trigger an internal self destruct system) REQUIRE wireless command access. Flight software can, and should, be designed to operate off-line. Flight softaware also typically relies upon multiply-redundant, "voting", systems.


      More later.

      Delete
    22. Sorry, I got interrupted.

      Yes, you can put in a lot of protections on kill switches to avoid many kinds of catastrophic failure. However, it all comes to a bottleneck at the line of communication. If the kill switch "THINKS" it has received a detonation signal, it will detonate. You could put in vote redundancy to avoid signal misinterpretation, or to prevent command misfires. You can require an encoded kill signal, to prevent stray ignals that are too close to a kill signal. However, there is no way to prevent a hacked signal from someone who knows how to break the encoded signal. Also, a stray short circuit at the wrong place would be enough to detonate the kill charge (after any signal processing stage that normally determines a legitimate signal.
      Getting back to the flight software: yes, this is complex. For that very reason, it can be more difficult for someone to successfully hack, even if it were allowing wireless command (and, unlike a kill switch that would require such wireless access, the flight software can be completely isolated from ANY outside access).
      Furthermore, flight software malfunctions are rarely immediately catastrophic. Uncorrected faults might result in a miscalculation for trajectory; easily corrected by outside monitoring, or even manual varification monitoring. Faults might result in a failure to fire engines, or engine misfires; also typically correctable. Even a fault that leads to a nuclear overload will allow time for intervening action, whether this involves employing emergency measures to correct the reactor, and/or to take the reactor offline, or to jettison the reactor to avoid critical damage (remember, nuclear meltdowns do not create nuclear explosions).
      Unlike a faulty kill switch, there is virtually no fault in flight software that could produce an immediately catastrophic situation (unless if it occurs at the worst possible moment, like a downward pitch combined with full throttle when you are trying an aerobreak manoeuvre).

      again more later

      Delete
    23. @Mikkel Haaheim:
      My understanding is that the more complex a system is, the less secure it is overall. Windows OS is massive and has hundreds of vulnerabilities... a pocket calculator is simpler to hack but also simpler to protect against hacking.

      In space, flight software might be a bit different. The main danger is other vehicles. Without a connection to wirelessly-transmitted flight data, you can crash into others spacecraft at extreme relative velocities. 'Offline' might protect you from smashing into an asteroid or missing a space station, but it will force you to rely on your onboard RADAR and passive IR scans to detect external objects. Unlike on Earth, the distances and speeds are too great to make these solutions effective.

      The other option is to equip each civilian spaceship with detection capabilities matching those of warships that need to know where every IR source is and where it is going. It might not be possible, especially if the civilian craft are naturally colder to conserve energy and are made of non-metallic components that don't show up as easily on RADAR.

      My point simply was that all the problems and consequences of a kill switch are likely to be present to a similar degree in other essential systems.

      A kill switch doesn't have to immediately detonate. It could start a countdown. The consequence doesn't have to be explosive either. It can be a procedure that vents the propellant tanks, poisons the reactor and encapsulates the fission duels, sections the structure into smaller pieces and ejects the habitat in a maneuverable 'life-boat'. The objective of a kill switch is to reduce the lethality of an unavoidable impact, and having to deal with millions of small pieces might not be worthwhile compared to simply making the Whipple shields thicker.

      Faulty flight software can create a situation where a kill switch is needed, like a full burn-until-empty maneuver.

      Delete
    24. "the more complex a system is, the less secure it is overall"

      Yes... ...and no. However, mostly yes. There is a universal truth that the more complicated something is, the more that there is to break. Strictly on a fault analysis basis, this means that the more complicated something is, the more vulnerable it is.
      OTOH, there is another universal truth: the more complicated something is, the more difficult it is to control. You could probably hack a flight programme, if it were accessible, but it would be difficult to hack it in such a way to yield a failure with enough control to actually produce a useful result, especially if the flight software had voting redundancy. A worm would make one system fail somewhere, and could even spread to make several systems fail in different redundant programmes. But it would be virtually impossible for a worm or virus to effect multiple voting systems in the same locations... at least before it gets noticed, and all the systems get systematically rebooted with versions without the worm/virus.
      The kill switch uses much easier programming (assuming it uses programming at all). This means that it is easier to protect, and inherently more secure aainst failure. Unfortunately, it also means that it is also that much easier to hack. Yes, there is actually quite a lot of feedback going on here in both directions, because, yes, it is also easier to protect against hacking... up to a point. The easiest hacks would also be easy to anticipate and counter, but hackers can anticipate those early level protections. But, again, the most important aspect here is that the kill switch REQUIRES wireless access to outside sources in order to function.

      Delete
    25. Now, about flight software in space.

      Honestly, space is vast. In transit, vessels are not going to be anywhere near each other. Even at velocities of hundreds of km/s (and it is unlikely they will be travelling at more than a couple dozen km/s, relative to one another), these vessels will generally be several minutes apart. Furthermore, there would be very little probability of being on a collision path. Even if they were trying to hit one another, it would be difficult to produce a collision path.
      Then there is the issue of wirelessly transmitted data. There are means to safely allow data reception, that I will discuss shortly. However, it is unlikely you would even WANT to rely on wireless flight data for navigation, simply because it would take so much more time for a remote sensor to track and transmit the data then it would take for onboard vessels to perform the same task. This is especially true for passive sensors.
      Now, let's say you are on a military vessel, and you want to use your aegis equivalent sensor network. This could be safely incorporated. You begin with your isolated vote redundancy flight software. Then you add a layer of software on yet another system. This system allows for the transmission and reception of data, but does not allow for any execution of data. You get the numbers, in a pre-formatted matrix. Usually, this would be sufficient to prevent any worms/virii from corrupting the system. However, let's say that some hacker manages to find a loophole that allows a worm/virus to "piggyback" on the data. Even if the "open" system is hacked, it still has to vote with the secure systems, which would prevent the hacked system from performing a dangerous or undesired act.

      Delete

    26. Then there is the nature of a "failure", whether a fault or a hack. In a complicated system, this failure could occurvirtually anywhere. However, there are virtually no systems in flight software that can cause devastating consequences before the fault can be located and repaired (or, at the very least, countered). Again, there is only the relatively rare occurrence of a fault happening at the worst possible moment (you make a power dive instead of aerobreak manoeuvre, for example... and, depending upon your available thrust/mass ratio and current altitude, even this might allow you a few minutes to prevent burning up in the atmosphere).
      A fault in the kill switch yields exactly three options: kill does not work when you want it to (it would have been better to send the interceptors); you have an immediate kill, probably when you don't want one; or you have a delayed kill, which may or may not include enough time to evacuate. Something I did not have time to get to before: I was not assuming that "kill" meant that the people would get killed, or that the detonation would be immediate; only that a "kill" would mean the loss of valuable property. Now, the problem is, even if the kill switch is meant to have a delay, that delay would not necessarily be present in a fault or hack.
      A kill switch delay requires a timing mechanism. This is just an extra layer to be hacked. In the end, there is just a signal (usually electric) that results in the ignition of the detonation charge. A hack does not have to go through all the programming phases. it just has to send a signal saying that there is a command to send that detonation signal... or preventing that signal from being sent.
      In a fault, all you need is a little stray electrical energy that detonates the charge (you don't even need a programming fault, you only need a little negligence from the people installing the charge... say a short circuit); or a fault in any layer of programming; or even any kind of sufficient energy surge at the location of the charge itself.
      The problem is, the kill switch charge is inherently lethal. There is no way to ensure a safeguard of the charge itself.


      Your second from last paragraph does bring up a valid point. However, I would not classify a switch that vents propellant (etc) as a "kill" switch. This would normally be refered to as, simply, an emergency cutoff switch. However, there is absolutely no reason why you would need to have a remote emergency cutoff. Such measures would not endanger the crew, nor the expensive vessel, so there is no reason any crewmember would hesitate to use it. Plus, there can be several redundant options available, that can be employed BEFORE doing something that ends a mission, and/or leaves you stranded for a few weeks, such as simply closing the propellant flow valves (long enough to repair any failing systems), or engaging all the control rods.


      A full burn-until-empty manoeuvre is not immediately life threatening. There are many means available to correct the problem, or at least counter it. Manually cutting off propellant flow. Scramming the reactor. Making a hole to vent propellant. Etc.
      In fact, to be quite honest, it is highly unlikely that there is any immediate threat that flight software could pose that having a remote kill switch would be able to counter, and that crew response would not. It takes time for authorities to learn about a situation, and time to collect the information for the vessel in question. You don't want to send a signal that will kill EVERY vessel in transmission range, so the aprropriate kill switch transmission code will have to be found. It will take more time for an authority to notice the problem (or otherwise be made aware of that problem), and then to retrieve and send the appropriate signal than it will take for the commanding officer on the vessel to become aware of the problem, and take immediate remedial action. This would include any on-board "kill" functions.

      Delete
  13. Ah! Something I've thought a good deal about. My own setting has relatively mature bomb-pumped lasers and other Casaba-esque directed nuclear weapons, and therefore submarine defensive weapons are a bit different. Instead of a submarine or a floating lens, ICBM like devices are used. Missiles are either free floating or suspended from a downward cable, though being dropped by a submarine is another option. Targeting data can be either be missile by missile or communicated by fiber optic/acoustic hydrophone (which are both totally invisible to orbit). When the region or planet is under attack, a volley of missiles are launched. They float to the surface before using solid rocket motors. Ideally the initial phase of the launch is over the horizon. The acceleration is extremely high, Sprint style (see link below), to avoid counter attack during boost. You get above the lower atmosphere and shoot at your target.

    I'm curious to hear your thoughts vis-a-vis the benefits and downsides of such a system. Obviously decentralization insulates your defensive elements from attack and makes detection more difficult, but I think the primary issue is money. Bomb pumped lasers require expensive fissile materials or magic pure fusion devices which can't be cheap.

    Sprint Missile launch
    https://www.youtube.com/watch?v=kvZGaMt7UgQ

    ReplyDelete
    Replies
    1. Maybe not so magical anymore, those pure fusion bombs
      Maybe

      http://www.projectrho.com/public_html/rocket/spacegunconvent.php#chemthermonuke

      Delete
    2. On the other hand, this system doesn't require large monolithic investments. A nuclear powered submarine is expensive and requires a great deal of upkeep and specialized skills. Missiles can be bought as needed, serviced easily, deployed without specialized equipment, and have an arbitrary rate of fire. If the planet is under continuous risk of attack, a submarine may make more sense. If the threat is lower or more punctate, missiles are easy install and forget defense systems.

      Delete
    3. Eth, that does look interesting. The biggest downside sounds like the yield-to-weight ratio. How much that matters to this system varies. If enemy forces are always overhead (never below the horizon) and can strike very quickly in response to a detected launch, you need a small, armored, and very high acceleration missile. Increasing bomb mass is going to make that difficult. That can be mitigated by just adding more missiles to your volley (which is easier with cheap devices like these) or bomb-less decoys that distract enemy lasers from your actual missiles.

      If on the other hand enemy forces can't be expected to cover the entire sky at all times, missiles can be a bit more leisurely at launch and pop up over the horizon to snipe at the enemy. It's important to remember that you don't care about reaching orbit with these things, you just want to get away from the opaque soup of the atmosphere. If launch requirements are really extreme, you might be better served by placing launchers in high altitude locations where the atmosphere is thin in both density and thickness. They'll be easier to find and harder to hide when they're installed, but will have a much easier time reaching the target altitude.

      The Sprint missiles I linked to above were only designed for attacking ICBM warheads in the last 1.5-30 km (when decoys aren't possible), so they had to be positioned at whatever altitude the thing they were protecting was. They could reach 30 km in 15s from launch. The graser warhead only needs about 3x that distance and even without advances in technology, the 1980s Excalibur program estimated a missile could pop up in under a minute.

      Delete
    4. @Final Iron God:


      Underwater-launched missiles that can 'Sprint' out of the atmosphere and deliver a devastating strike against orbital targets from a thousand kilometers away or more is pretty much a perfect last line of defense. You dont' have to invest in a huge submarine or expose delicate laser optics - just keep simple missile pods deep underwater.

      Also, you have to consider the cost effectiveness from the point of view of a military planner. The missiles will be expensive and they will consume several kg of fissile material per warhead... but aren't they taking down a space warship that costs thousands of times more with a reasonably good effectiveness?

      Delete
    5. This is where yield becomes an issue. Yes, the space vessels will be expensive, but they could conceivably be designed to survive several hits.
      If I remember the Cassaba proposal, we are talking about a nuclear pumped laser. You might be able to guarantee a puncture that goes through the vessel, but the effect would be similar to shooting a log cabin with armour peircing automatic rifles shells. You could fill the cabin full of (tens of thousands of) holes before it stops functioning like a cabin. You STILL might never hit anyone in the cabin, or make the cabin unusable. Same thing with firing lasers through a properly compartmentalised vessel.

      Delete
    6. It's actually a bit different than that. Excalibur is the nuclear pumped laser concept. A nuclear device energizes rods of lasing media to produce gamma or x-ray beams. Because the pumping device is nuclear, the pulses are necessarily very short and very high energy. You won't so much drill a hole as turn whatever matter you hit into supersonically expanding plasma. It doesn't matter much if your spot is small because even if you just hit a propellant tank, you're dumping millions of megajoules onto the target in 10 nanoseconds (how long the rods have before they are vaporized by the nuke). The material you did hit will conduct the energy for you as a conventional explosion.

      A 1986 review of the program assumed a 1 megaton device that would produce lasers at 2.5% efficiency for a total laser output of 100 million megajoules (1.5 Hiroshima bombs). So the beam is like detonating an atomic bomb on the hull of whatever you hit. The primary issue is not one of over penetration but actually the beam spreading too much. It's more likely to be a wide (for a laser) cone of hot death not a needle. It's pretty likely the beam will be the size of the ship by the time it gets there and deposit its energy across the surface.

      With these devices you don't actually need to kill your target by heating it to vaporization. Gamma and x-rays are ionizing radiation. They will penetrate virtually any material and tend to really mess up delicate things like living cells and computers. Any beam capable of damaging a vessel conventionally is going to more than enough to deliver fatal radiation to its crew and computers. Shielding against this kind of attack requires huge amounts of extremely heavy shielding. Computers can be made more resilient but the crew cannot. You can also trigger Compton scattering, which will create an EMP right there on the surface of the vehicle, further complicating things.

      Granted, yield is still a crucial variable as is divergence. If you're stuck with low yield bombs or really crappy beam divergence, the bomb-pumped laser is a relatively poor weapon. If not, they seriously alter the equation for ship to ship and planet to ship warfare. Fortunately for a defender, the atmosphere will very rapidly absorb high frequency light, so a better bomb-pumped laser is almost entirely in favor of the planet.

      Delete
    7. Right then. Sorry, I did not have the details at hand. Assuming the beam has diverged widely enough, then it would definitely have sufficient yield.

      However, for a lot of these effects, the intensity can actually work against the weapon. If you are incorrect about the level of diffusion (I would be interested in seeing how you made the determination that it would be ship-sized), then the limited time exposure becomes a problem.
      Yes, gamma and x-rays are ionising. However, this is dependent upon the energy actually being transmitted into the material. Now, no matter what, you are going to have a fairly high level of energy being transmitted into the material... but not necessarilly as much as one might think.
      As I understand it, part of the problem is that materials become transparent to wavelengths at sufficiently high energy levels. Now, nothing is ever 100% transparent to any wavelength, which is why you are still going to dump a lot of energy, but you still have an overwhelming amount of energy wasted (until it hits the next object in its path).
      Of course, the second problem is the penetration issue. Once the beam has actually penetrated the target, there is nothing left to absorb the energy in its path, which means nothing new is going to be transmuted.
      Third, with such a short pulse, there might not be time for sufficient indirect irradiation. Gamma and x-rays will transmute material, but only material it actually comes in contact with. Now, yes, you are still going to be lighting up everything you hit, which will likely become extremely radioactive... and THAT material might very well become hot enough to start cooking items around it. BUT... that material is quickly being vapourised, so a lot of it will be lost before it can do its secondary damage; but, yes, you will have some doing a fair amount of secondary irradiation damage as hot vapours expand through a few nearby spaces (until the bulk eventually expels into space). In short, if the pulse is too short, it will not have sufficient exposure time to cook the crew (except for those who are actually in the path of the beam).


      More later.

      Delete
    8. Loving this discussion, it's always great to explore a topic.

      Digging a bit into the literature, found a couple of things of interest. Most of the sources are here: http://www.projectrho.com/public_html/rocket/spacegunconvent.php

      First, with x-rays around 1 nanometer wavelength (the planned wavelength for the Excalibur system), pretty much all matter absorbs it very quickly. The estimate from that Directed Energy Missile Defense in Space (Carter 1984) review suggested that all the energy would be absorbed in easily the first millimeter of an ICBM's aluminium skin. They did note that sufficiently high intensities can generate "bleaching" when all the electrons in the target have been sufficiently displaced (electrons do the absorbing), rendering the target transparent. However, that means you've generated a 100% ionized plasma, so overpenetration only happens when the parts of the target in the beam have been reduced to bare nuclei. They were interested in the effect for tunneling the beam through the upper atmosphere, which might be useful to reduce pop-up height. I couldn't find any solid equations on electron bleaching.

      Second, in response to your question about diffusion and a ship sized beam, that depends a lot on wavelength. At the 1 nm Excalibur design wavelength, diffraction limited beam divergence for a 5 meter rod is 20 microradians. That's pretty awful for a "laser". I couldn't get to the source to get the equations and try some different rod lengths(and I think Atomic Rocket's stated equations are seriously off somehow), but 5 meters seems pretty long for a pop up missile meant to take high g's.

      At 20 microradians and a target distance of 1000 km (the primary example in the blog post) our 1 megaton 2.5% efficiency device delivers a 20 meter wide beam at ~320 billion Joules/m^2 (318.27 GJ/m^2 or 3.18*10^7 J/cm^2), which is pretty much a guaranteed kill via impulsive shock if even a small fraction of the energy is absorbed.

      However, we're trying to pop up to 100km and take our shot, so we need to be farther away. If instead of shooting straight up and firing at our 1100 km high target, we want to do a pop-up, we need to fire from a range of 3700 km (the 1100 km orbiting warship's horizon distance to a 100 km object). Now we've got a beam of a width of 74.72m, for a beam intensity of 22.8 GJ/m^2, which is still absolutely lethal via impulsive shock. At even 10,000 km we're still delivering a 200 meter wide beam at 3.1 GJ/m^2. That divergence is why I think the beam is likely to be pretty "ship-sized", or at least a substantial section of it.

      If a "soft" kill (via crew/computer irradiation) is wanted (or the target is very agile), you can shorten and widen the rods of your device to increase beam divergence. The example device is lethal to unshielded humans for a very, very long distance. It can deliver 30 Grays (death in seconds-minutes) at 13 light seconds and 80 Grays (immediate coma and death) at 8.5 light seconds, and is 50% fatal (4.5 Grays) at 36 light seconds over an area 430 km wide. Aim very carefully, missing and hitting a poorly shielded colony is a war crime.

      Third, gamma rays can be focused far more tightly and potentially could be more efficient at transducing bomb energy into a beam, but are much better at penetrating matter. That's both bad and good for a weapon and might allow their use from orbit against targets inside an atmosphere or really long range anti-crew attacks. The science on generating gamma ray beams is considerably more speculative.

      These devices create a very different environment for space combat in general, with extraordinarily high intensity beams of ionizing radiation the arguable best weapon.

      Delete
    9. Sorry, I seem to get interrupted a lot lately.


      "with x-rays around 1 nanometer wavelength (the planned wavelength for the Excalibur system), pretty much all matter absorbs it very quickly. "

      Yes. This is true. However, there is another factor here that is often overlooked. There is a physical phenomenon in which the intensity of a beam itself actually interferes with absorbtion. At the moment, I forget the name of the phenomenon (I think I have refered to it on this site before), but what happens is that every material has a threshold, dependent upon the wavelength, at which point irradiance intensity itself begins to render that material transparent. Interestingly, it is not a matter of the material absorbing the same amount of energy, and (only) then becoming transparent to any further irradiance; rather, if the intenisty of irradiance is above that threshold, the base absorbtion percentage itself is dramtatically reduced at the outset, with the transparency of the material becoming dependent upon the intensity of irradiance.
      For example, let's say that an imaginary material normally absorbs radiation at a given wavelength at a rate of 90%. So, if it is irradiated by a 100 MW/m^2 source, the material will absorb 90 MW/m^2. However, let's say that the imaginary material has a threshold for that wavelength at 101 MW/m^2, at which point the absorbtion rate drops to 10%. Now, let's next assume that the material is irradiated at a level of 110 MW/m^2. One might magine that the material would absorb the same 90 MW/m^2 for the first 100 MW/m^2, and then drop to 1 MW/m^2 out of the remaining 10MW/m^2 above the threshold, for a total absorbtion of 91 MW/m^2... but this would be wrong. Instead, the material will absorb a total of 11 MW/m^2.
      It is important to note that the actual transparency/absorbtion of a material follows a curve, dependent upon total intensity. The total absorbtion/reflectivity/transparency percentage values are all dependent upon the specific intensity at each specific wavelength. The threshold is a point where there is a sudden drop in absorbtion and increase in transparency (I am uncertain whether or not reflectivity also drops as significantly).
      BTW: no, this phenomenon does NOT mean that all of the electrons have been stripped. The phenomenon is frequently used in "2-way" mirrors, which obviously do not have their electrons stripped into plasma.

      Delete

    10. "divergence for a 5 meter rod is 20 microradians"

      I think that the length of the rod is actually mostly irrelevent. Divergence is generally dependent upon diameter, not length. Sometimes, the important factor is the ratio of diameter to depth (length), but I think this only applies at relatively short ranges.
      That said, assuming that you are taking the divergence value from a document, I don't see any reason to dispute it (the length of the rod might be irrelevent... just indication of interesting specifications, but I would assume that the diameter of the rod is quite small, which would result in a larger divergence value). Given that divergence, I agree that most vessels would be toast... unless it happens to be something the size of a super-star-destroyer.

      I should point out that you actually don't want to focus the beam more tightly, in this case. Focused beams are important if the total yield of the weapon is measured in MJ to perhaps a few GJ, but once you are talking about beams that are already in 100s of GJ/m^2 anyway, a tighter beam just makes it more likely you will waste the beam.
      Note: the effects I am discussing here are similar to terms of "stopping power" of small arms. A 9mm handgun shell can often penetrate armour, but it often will not stop a charging drug addict. A Magnum shell, which has much less energy, might not penetrate armour, but it WILL stop the addict, and WILL slow down even someone wearing anything less thanfull combat armour (even then, it will probably slow the soldier down a little, just from the kinetic force of the impact... but probably not enough to keep that soldier from shooting you).

      Delete
    11. One final note: widening a rod will DEcrease beam divergence, at least at long distances.

      Delete
    12. @Final Iron God:
      Electron bleaching: we can be sure that the maximum ionization of an atom is the upper limit for the energy needed for electron bleaching. Maximum ionization is when every electron, from the outermost to the innermost, jumps away from the nucleus due to the energy they receive.

      Gamma rays are said to be the most penetrating radiation there is, so that seems to conflict with reports of being blocked by a sheet of aluminium. As far as I know, it takes 61cm of water or 5cm of lead to absorb 90% of a gamma ray's energy.

      @Mikkel Haaheim:
      Sorry, I can't contribute here. I haven't looked at the details of nuclear-pumped gamma ray tubes.

      Delete
    13. Sorry, one correction (I think)... thinking this over a bit, I don't think the phenomenon in question is the one responsible for 2-way mirrors (actually, the "deactivation" of the mirror function).

      Delete
    14. Sorry for being so slow, that pesky real life thing is really getting in the way of NUCLEAR SPACE LASERS.

      I’ll be responding essentially most recent to least recent, so bear with me.

      @MatterBeam:
      Gamma rays are extremely penetrating, yes (which may render even a crappy graser a fantastic anti-crew weapon), but it is ~1 nanometer *x-rays* that are extremely well absorbed. The gamma range is ~10 picometers.

      @Mikkel Haaheim:
      The angular divergence of the beam is determined by 2 equations. The first is the geometry of rays in a cylinder: angle=diameter/length. The wider a cylinder and the shorter it is, the greater the angle that a ray can enter one endcap and touch the opposite side of the other. Think of the angle that a drinking straw can lay at in a cup. The wider or shorter the cup, the further out the angle of the straw. The second equation is the classical diffraction equation: angle = 1.22*wavelength/diameter. Combining these two lets you find the ideal width of the rod for your wavelength and rod length. Too wide and the geometry decreases focus, too narrow and diffraction increases the angle of the beam. There are no lenses or mirrors in this apparatus, since they don’t work well at these wavelengths, especially at these energies. At 1 nanometer and a rod length of 5 meters, 0.06 mm is the ideal rod diameter, for a divergence of 20 microradians.

      The effect you’re talking about is saturable absorption, where past a certain energy concentration, the material becomes far more transparent to a wavelength of light. The effect is exploited in visible lasers to produce passive Q-switching. The laser can’t escape the resonating cavity until the beam has built up enough energy. Once it has, the material covering the aperture becomes temporarily very transparent and then after some delay, recovers. This is used to produce a passively pulsing laser, since the pulse needs to reach some limit before it can escape.

      This saturation is due to the light exciting the electron shells. At every wavelength, there is a characteristic layer of the electron orbitals that is preferred for absorption. At low light energies, that shell transfers energy to higher shells, so the outer electron shells are the only ones to become excited.

      If the light intensity is high enough, it can excite the preferred shell of electrons so quickly that the those inner electrons don’t have time to transfer their energy to outer shells. They simply jump to a higher energy level, leaving an vacant band of electron orbitals beneath them. If enough of the atoms reach this point, the absorption of that wavelength of light decreases greatly since they don’t have their favored electron orbital to absorb them anymore. Eventually, these excited electrons pop out (the Auger effect) and take the initial energy they stored with them. Other electrons from higher shells fill the vacancy, allowing the atom to absorb that wavelength again. The Auger electrons heat the material.

      With the right materials, a visible laser (~100s of nanometers) can achieve 90%+ tranmission during the delay period. The intensity needed to saturate one layer of materials designed for this task is in the 10-100 microjoule/cm^2 range. Layer thickness is based on the extinction coefficient of the material for the light. However, most targets won’t be specially designed for the specific wavelength of the laser.

      Delete
    15. Saturable absorption of x-rays is in principle the same, but very different in practice. With visible light, the preferred orbitals are quite low energy and far out, since the wavelength is large. With 1-10 nanometer x-rays, the preferred orbitals are the 1s-2p orbitals, the very closest electrons to the nucleus. They are very, very strongly attracted to the nucleus, so the saturating and Auger energies are tremendous. One recent paper (see below) measured the saturation of x-ray absorption in aluminium targets exposed to extremely short, high intensity x-ray pulses at 13 nanometer wavelength. To achieve 70% transmission, every cm^2 layer of aluminium needed to absorb more than 1000 joules. With a 110 nanometer layer thickness, that’s 1000 J’s into 30 micrograms of aluminum to permit 70% beam transmission. After about 40 femtoseconds, the saturated bands collapsed and the potential energy released by the Auger effect brought the target to a balmy 290,000 Kelvin.

      Our pulse is another order of magnitude smaller wavelength (which will drive even higher energies required for saturation, though once you hit the 1s orbital you usually just sail through without touching the electrons), far longer than the collapse time of saturation (1 nanosecond vs 40 femtoseconds), and even if saturation occurs it will vaporize the matter. Without doing some messy calculus to figure out the maximum depth, our beam of x-rays at 1000 km can saturate only 3.4mm of aluminum to 70%. At 10,000 km, that depth becomes 34 micrometers. I think it is safe to assume that most materials will achieve 100% absorption in the meters to 10’s of meters thickness of a spaceship. Just about every atom above has 1s, 2s, and 2p shells. You might be marginally better off if you build your ship out of lithium, carbon, beryllium, and boron.

      Saturable absorption might actually *help* the beam do damage. If 100% of the energy is absorbed in 100 nanometers, ~50% of that layer will just fly off into space. If the beam goes deeper, more of the energy has no choice but to hit the rest of the ship.

      Delete
    16. If I understand correctly, X-rays will be absorbed at the surface of a material and the rapid expansion of the plasma will create an overwhelimg mechanical pressure that deforms the surroundings.

      Gamma rays will penetrate deeper, distribute their energy and possible have less damaging effects on a material... but more of their energy might reach the vulnerable innards of a spaceship.

      Delete
    17. Precisely, yes. Though it is worth noting that gamma ray lasers is way less understood.

      Delete
  14. Detecting submarines has traditionally been very difficult, but if this article is valid, the Soviet Union may have developed alternatives to sonar: http://www.popularmechanics.com/military/navy-ships/a28724/submarine-sonar-soks/

    And of course the idea of using supercavition for high speed subs is problematic simply because it leaves such a vast acoustic signature. All the enemy needs to do is drop patterns of hydrophones into the ocean and they will be able to pinpoint the sub the second it goes into supercavitation mode.

    One possibility would be to make subs even tougher and capable of diving far deeper. This does not mean they can fire lasers or missiles from that depth (although they could conceivably use some sort of swim out system to release a weapon which then surfaces and fires), but rather they radically increase the volume of underwater space which needs to be searched and controlled.

    ReplyDelete
    Replies
    1. The supercavitation is best used to leave an area after a laser strike reveals the position of the submarine.

      Leaving the area at high speed is necessary as hypervelocity rounds shot down from orbit can reach the area in minutes. If the submarine is not longer in the spot it attacked from, then the forces in orbit have to de-orbit hydrophones, torpedoes and depth charges over a longer timescale.

      At 300m/s, the submarine can clear a full 18km. It then stops, switches to stealthier propulsion and moves on from there. This forces the spaceships in orbit to expend resources that can cover an 18km radius circle within a minute... quite hard to do!

      Delete
    2. The sub first has to accelerate, and then decelerate back to stealth speed, which will make the distance covered less than the 18km. All the enemy really has to be listening for (and since they are not stupid, they will have established a grid of hydrophones the moment it was possible to do so, so they are likely to have the grid active during the engagement).

      So the key is to listen for the "knuckle" of disturbed water as the sub decelerates (and the cavitation bubble collapses) and draw the uncertainty circle entered on *that* point. Not to say that high speed dashes are not a valid solution, but I think it would be a high risk manoeuvre on the part of the commander.

      A possible work around would be that deep diving subs release swim out weapons which can go to supercavitating speeds once near the surface (and perhaps launch several supercavitating decoys as well). That gives the enemy a number of very noisy targets to chase while the real launch platform slinks away a kilometre or more under the surface.

      Delete
    3. Would it not suffice initially to dive as fast as possible, and then move away? That'll protect you from laser and kinetic strike, though not from depth charges or torpedoes, but I imagine those would have to enter the water at a lower speed to survive, and would then move relatively slowly, so you might be able to move out of the way.

      This brings up another point: To reliably engage submarines, spacecraft will probably need specialized anti-submarine weaponry. How strange would it be to have a spacegoing anti-submarine vessel? I'm not quite sure what the weapons would be like, but I'm imagining something between a torpedo and an autonomous attack sub, dropped from orbit. Now we've got submarines armed with lasers and spaceships armed with submarines! This sounds pesky enough technologically and logistically that you might instead try to just suppress the subs and win the war through other means, but if you can't actually gain orbital superiority, that might not be easy.

      Delete
    4. Kiloton atomic depth charges have a kill radius of about a kilometre, so finding the "knuckle" of disturbed water and dropping a MIRV with 4-8 depth charges (RV's with delay fuzes so they enter the water and penetrate to a preset depth before exploding). Since they are mounted on short range rockets, they could have a secondary use as last ditch defensive weapons in space, easing the logistical burden.

      Delete
    5. There are devices today that can listen at a distance, for example by analysing the vibrations of a glass panel. If a similar system is possible for liquid bodies, the attackers could have orbital hydrophones. They would probably not be as good as conventional acoustic hydrophones, but it could be enough to detect the massive signature of a supercavitating vehicle.

      On the other hand, could the defenders shoot down incoming nuclear depth charges?

      Also, if a kiloton atomic depth charge has a kill radius of 1km, how does it scae with energy? What would be the radius of a 500 kt or a 5 Mt atomic depth charge?

      Delete
    6. It is assumed that once the enemy has littered your oceans with hydrophones, then its pretty much game over for the fast-moving laser submarines. Any time they fire, they'll have enemy fire on their backs. This pretty much negates the advantage of having an ocean to protect you.

      There are options, of course. Slow-moving submarines that remain maximally stealthy can evade counter-fire by using one of the towed or fibre optic designs I described. EWAR submarines can track recently splashed down hydrophones and destroy them, or hunt down their space-bound transmissions if they're already in the water.

      Waves chaotically disrupt the sound signal that creates vibrations on the surface of the ocean, so it can't be read from space.

      However, supercavitating submarine will leave a trail of bubbles. These bubbles will rise to the surface and essentially trace the submarine's path. The good news is that bubbles rise slowly, at a rate of 0.1-0.2m/s.

      Following an attack, a submarine is likely to dive down quickly, slow down and travel silently for a few minutes, and then supercavitate away.

      Diving down increases the delay before the bubble trail reaches the surface. At 200m depth, it has at least 16 minutes, more likely 30 minutes, before it reveals itself when it supercavitates. The first high speed run takes out of return fire dropped on its immediate position.

      Subsequent runs would have to alternate between supercavitating and regular speed travel. Each break from supercavitating affords the submarine another half hour of stealth.

      At a first estimate, the effective range of a depth charge will scale by cubed power, as it has to vaporize an increasing volume of seawater to create the displacement from which originates the destructive shockwave and overpressure.

      Delete
    7. While lots of things are possible, I would be more inclined to have a "mothership" submarine running very silent and deep, and use swim out weapons that ascend to the surface to carry out strikes at the enemy. While an ultra deep diving sub is probably very complex and expensive, it is probably only as complex and expensive as a supercavitating sub, and far less likely to be found (particularly when the depths of the oceans have to be included in any search pattern). Maybe I'm just old fashioned, being a fan of WWII fare like "Run Silent, Run Deep" or "Das Boot".

      I would also imagine being in such a submarine fighting enemy spaceships would be about as terrifying a job as you could get, especially since the enemy counterstrikes would almost by definition catch you by surprise (much like a current submariner would be caught by surprise when an air or rocket launched torpedo suddenly hit the water and went active on him......)

      Delete
    8. If the attacking spaceships prove that they are to quickly retaliate against a laser submarine firing on the surface, then AOLS design will tend towards your idea.

      I would actually take it a step further and just have the submarine as a roving nuclear reactor equipped with a laser generator and a forest of fibre optic cables that it raise to the surface with a balloon and bring them back down again by a winch.

      A separate fleet of UUVs stays closer to the surface. It hooks up with a fibre optic cable, surfaces, focuses the laser onto a target, and then manages to escape or gets destroyed. Simple, expensive but effective.

      Note also that if the submarine or the submarine-support fleet has access to detailed information on the location and predicted movement of targets in space, then it is not a stretch to suppose that they have the ability to watch the atmosphere in between. A shower of hot plasma trails indicating a rapid re-entry would give decent warning that torpedoes or depth charges are on their way.

      Delete
    9. So, going over all this....

      Kinetic weapons cannot be used due to the planet's quantitative advantage.

      Laser weapons can only penetrate so much.

      Add in the need for UAVS to see below the clouds, and UAVS/UUVS for the defenders.

      This brings us to one particuler question- how much logistical/repair/resupply equipment can the defender put underground/ under the sea-bed? Otherwise its simply a case of going in, ignoring the subs, and frying surface material to force the defenders to starve/give up.

      I think we need solid numbers on all this (though that might be impossible)

      Delete
    10. I'm not sure 'cannot be used' is the right conclusion. More likely, they have to be used in concentration, but once the main attack is over, then securing the advantage must be conducted by reusable weapons.

      For example, a planetary attack might start with waves of missiles from each side. The attacker wins by using more missiles to the field, even if the defender has a planet's worth of reserve missiles. If the latter are not available immediately, then they only play a role in subsequent battles - if the attacker is successful, there are no subsequent battles.

      Kinetics will then be used to cheaply knock out orbital stations. They can also be necessary for penetrating hardened bunkers.

      Lasers will have to do the rest of the work. Complement the missiles in the initial attack, take out missiles climbing off the ground into orbit, shoot down aircraft, target surface vehicles and methodically destroy surface-based installations. Scouring a planet's worth of defenses would take an inordinate amount of ammunition, even if you've negated your target's ability to retaliate and you can rely on the cheapest and dumbest rocks. The sheer mass of it all makes it impractical. Lasers, however, can melt down anything over time.

      Lasers are effective in space and with the right wavelengths, on the ground and even through the top of the oceans.

      The entire point of the laser submarine is stay outside of the orbital attacker's weapons fire. This means diving deeper than the top of the ocean and why they remain effective as a last line of defense.

      Delete
    11. Just to expand this discussion slightly, in a very SFNal setting, you could substitute laser airships in the lower cloud decks of gas giant planets to defend that turf. The environment and many of the factors involved will have similarities (ever to the point that both summaries and airships use displacement to operate in their fluid mediums).

      This makes doing orbital manoeuvres, aerobraking, scooping the atmosphere for hydrogen remass and other operations in the gas giant setting becomes much more difficult for the attacker, for many of the same reasons discussed above. And in this sort of setting, the attacking fleet may also have to deal with orbital and ground based installations on a multitude of moons orbiting the gas giant as well.

      The fleet admiral will be dealing with fire coming at him from a multitude of directions in both the orbital plane, military installations deliberately orbited "out of plane" (such as in polar orbit) and fire coming from below the atmosphere if the constellations gets too close.

      Delete
    12. A Jupiter-centric setting allows for a lot of interesting ways to use the cloud layers in space warfare. I can imagine submarine equivalents in the lower cloud layers, but also combination gas scoop and rocket warships that orbit at very low altitude and are ready to 'pop up' with high thrust, low Isp engines before diving back and refilling their tanks.

      However, Jupiter itself is not that interesting. No-one live in it really - the points of interest are the moons around Jupiter. The cloud layers only really become relevant if there's a very big gas processing operation to protect, or if the weapon ranges easily exceed the millions of kilometers (this will allow ships hidden around Jupiter to provide cover fire to the Jovian moons).

      Perhaps an extrasolar 'hot' gas giant might be a better fit.

      Delete
    13. Or maybe Venus?

      Delete
    14. Or just placing lasers on the top of Latitude-line Wind Wheel?

      Transmitting energy to the moons at peace, intercepting enemies at war.
      I think terawatt-grade laser is enough to do some serious damage, even the target is a million km away.

      Delete
    15. @Felix: The moons would already have some sort of re-focusing drone mirrors set up to transport the laser beams from the wind wheels to the moons, so you have the basic components of a Laser Weapon Web already in place. Throw in a Terminal Focusing Ship and you can convert any laser web into a weapon.

      @Geoffrey S H:
      That would make for an interesting setting!

      Delete
  15. Massive RBoDs would probably be mounted in or on the moons, as laser thermal rocket or lightsail power plants in peacetime, and to dominate the volume of the system at war.

    Enemy ships or constellations might want to get deep in the gas giant's gravity well to effect a plane change manoeuvre, aerobrake or use momentum transfer to boost their speed. If all they have to worry about is lasers firing from higher up in the gravity well, there may be clever paths to take where moons are occluded or otherwise unable to achieve a firing solution on the fleet. "Submarines" in the cloud decks, using many of the same tactics discussed in the OP would mean the enemy is *still* dealing with a 3D space battle issue, rather than essentially making the battle 2D on the surface plane of the Gas Giant.

    I'm thinking a lot about how operations in space will be different from what we are familiar with here, and the biggest difference is space operations are about time, energy and volume (probably not a startling insight), which has many gross and subtle effects on how things will have to be done. This set of conditions brings about ideas like:

    a. There are no cargo ships, ever. Bulk cargo is shipped as cheaply as possible (just like here) so it is sent on minimum energy orbits in unmanned and unpowered cargo pods, container shipping without the ship, so to speak.

    b. The rocket equation means every gram counts. Ruthlessly stripping away every bit of excess mass means warships will evolve to single purpose platforms like Laserstars, Kineticstars and Firestars (vehicles using nuclear warheads to drive weapons effects), supported by clouds of sensors.

    c. Weapons are designed to dominate or deny volumes of space, while warships and platforms attempt to have enough deltaV to move through and beyond the volumes dominated by the enemy.

    d. People (and eventually AI) are relatively fragile, so do not directly participate in military operations. Since people are ultimately the goals of wars or other means of conquest (think of a commercial empire), conventions and laws will be made to protect them from the violence of war, or even accidental releases of energy (such as off target cargo pods). Deliberately targeting people (either the crews of command and control ships, or habitats in free space and on moons and asteroids) is unacceptable escalation of force, and will invite everyone to band together to exterminate you.

    e. Passenger ships (Fast Packets) will be much different from warships, being robust enough to house their passengers, and also using enough energy to get from place to place as fast as practical. However, since they are quite dangerous due to the massive kinetic energy they have, they are treated with extreme caution and some suspicion wherever they go.

    Somewhat OT, NextBifFuture has reported on a company which is trying to make an inexpensive and robust source of positrons. Cheap antimatter will have a huge impact on any setting it is introduced in:
    https://www.nextbigfuture.com/2017/10/update-on-positron-dynamics-working-towards-antimatter-propelled-cubesat-in-2018.html

    ReplyDelete
    Replies
    1. I can imagine a rather odd situation for A: unmanned & unpowered cargo pods are perfect prey for "pirates", sort of.

      If the cargoes are valuable, pirates may lay some "mass catchers" to intercept the pods and no one can stop the pirates.

      Maybe only the low-value cargoes are transported by the pods, manned transports would carry the high-value ones, or armed drones may escort the pods if you insist unmanned shipping.

      Delete
    2. I agree on B & C, but D...I doubt whether can these conventions work or not.

      Yes, people are the ultimate goal of wars, but for those on the receiving end of weapons, they are targets of subjugation at best, targets of extermination at worst. Many countries just want the land and resource of their enemies, not the people.
      Even highly or fully automated military forces are being used by both sides don't change this.

      Unmanned forces were wiped out? Maybe the people will mount last-ditch attacks on their own, I don't think the unmanned enemy force would stop simply because "killing people with flesh-and-blood spells some really dire consequences".

      If this kind of conventions and laws are somehow made and really enforced with some real retributions, such as unrestricted warfare against the parties violated the laws, how to judge an attack deliberate or not?
      For example, the legally ambiguous Raufoss rounds should not be used against personnel (official stance of Norwegian Government), but in the heat of battle. few soldiers care these regulations.

      If someone (Faction A) violated the convention, like a kinetic round hit an enemy (Faction B) space colony which lost the defense network due to sabotage, people got killed. No one can be sure that the front line commander of A fired the shot intentionally or not.
      And I don't think that A would bring the commander to the court, since A may face extermination if the commander is found guilty. No faction would commit suicide in this way.

      There can be even more dliemmas, for example: only some deadly retributions that mean business may deter even the most reckless factions from attacking people, but those retribution may violated proportionality in the first place.
      Some factions may game this system, conducting some false flag operations, framing their enemies violating the laws first. Then they are justified to do something horrible against their targets...

      Delete
    3. @Thucydides:
      As mentioned just above, the fact that laser transmission will very likely use multiple re-focusing lasers instead of single massive lenses means that there'll be all the components needed to make a Laser Weapon Web out of every Power Transmission Web between the moons. So, trying to find an approach path where the moons are occluded might be possible, but finding one with a Laser Web in place will be impossible.

      A) Cargo pods will be hard to justify in a setting where propulsion is cheap, such as fusion rockets or when everyone has a massive propulsion laser that can be focused on an ablative metal nozzle to produce high Isp exhaust. They are more likely in a setting where transport is expensive... but then you raise the issue of expensive causing the volumes of cargo being sent out becoming smaller and more valuable, potentially valuable enough to justify equipping with their own engines.
      A half-way point would be cargo pods equipped with just enough propulsion to make an insertion burn at their destination. That deltaV can be used to correct for trajectory errors, divert away from a collision or just make a 'middle finger' burn away from pirates. It can always be picked up later at lower cost than losing it entirely to pirates.

      B) Agree. Even if propulsion becomes cheap economically, it will remain a vicious deciding factor in warfare. Nice name 'Firestar'.

      D) Sorry, I have to agree with Felix on this one. The best way to get people to work for you is to threaten them directly, and 'conventions' have not stopped tens of thousands of people being deliberately targeted and killed in modern conflicts. And, akin to the bombing campaigns of WWII, the civilian cost is justified as the industrial base is generally right inside the population centers. I strongly believe future warfare will be just as bloody as today.

      E) I think we are pretty familiar with how to handle things that are very dangerous on a routine basis. Transporting hazardous chemicals, radioactive waste, flying airplanes over cities and so on: can affect a lot of people, cannot be avoided, but we manage it. Also, bonus points for making arguments that space stations will not be large population centers in an inhabited solar system.

      Delete
    4. On the convention thing, I think it is on the opposite.
      A convention works not just because we know adhering it is better to all the parties involved, but also because there is price to pay if it is violated, be it physical or not.

      You may say it is already a threat.

      Delete
  16. The "dismal science" and the rocket equation work against pirates hijacking cargo pods in space. They are on a minimum energy trajectory to ensure they are delivered at the lowest cost possible. Attempting to intercept and move a pod into a new orbit requires vastly more energy (and if you are intercepting a huge pod of water or other bulk material, it has a considerable amount of kinetic energy to overcome already, even at a minimum energy trajectory you need to accelerate it by several kilometres a second). So even if you physically capture a pod, the amount of energy you expend will probably be a large fraction of its worth, and potentially more than the value of the pod if you propose to make a serious change (diverting a pod on its way to Mars to Jupiter will require a considerable change in deltaV to accomplish).

    Would be pirates will more likely be financial operators working the futures market. The pod is set to arrive at a very specific time, but what is the actual value of the pod going to be on that day? If you can somehow purchase it for less than that amount, transfer ownership or manipulate the markets to change the value of "x" before arrival, then you either make a metric crap ton of money, or financially ruin someone (or both).

    The issues of proportionality and response in war was deliberately left a bit ambiguous, both because it is quite possible to place someone on the horns of a dilemma. and because it represents areas which might be interesting to explore in terms of creating stories. In the case of your example of an officer accused of deliberately targeting a habitat, would the enemy polity be satisfied with a trial? Would some third party be required to certify the trial was conducted properly? Would it be a convention to have representatives of the opposing sides to make up the court (Ambassadors and embassy personnel might remain to provide a "point of contact" should one side or the other wish to make overtures, negotiate prisoner exchanges or other activities that "civilized" nations do during wartime), or at least observe to report bak that a fair investigation and trial was conducted?

    And of course there is the issue of third parties. If the Imperial Jovian Navy is being accused of deliberately targeting habitats and either refuses to allow investigations and trials, or it is felt these are being manipulated to ensure no findings of guilt are ever delivered. then the IJN might discover more parties are going to join forces against their operations against the Uranus Space Force. Or third parties may decide the IJN is an existential threat to them, and act accordingly.

    ReplyDelete
    Replies
    1. Piracy in space is something I plan to dedicate a full series to, like Stealth in Space or Interstellar trade.
      A few points I'd like to make right now are:
      -Minimal energy cost trajectories are not always the 'cheapest'. If your jovian cargo pods take 6 years to reach Earth, requiring only 9km/s deltaV, it will soundly be beaten in profitability by cargo that reaches Earth in 1 year despite requiring much more propellant. Propellant is cheap. High Isp, high value cargo like transporting precious metals using ion rockets, will shift the 'profitability' equation sharply towards shorter trip times using more propellant.

      Piracy will happen locally or on similar energy trajectories. You capture the cargo near where it is supposed to be collected. Or, you very slightly divert it to your pirate base. For example, diverting a cargo coming in from Jupiter to either of the inner planets doesn't cost a lot of deltaV. Then, sell the diverted cargo as salvage to anyone willing to pick up the braking burn bill.

      Or, don't bother. Latch on a hijacking drone and ransom the asteroid. Pay the ransom and the drone will detach. Don't pay up, and the drone will manipulate whatever propulsion system is in place for capture to fail.

      There's more to say of course.

      Delete
    2. Unmanned cargo shipping, by pod or by ship, can be the most common way of space shipping, but few things worth for years of waiting.
      It is not scientific exploration that space agencies have various constraint on the mission. Besides, the agencies can still do other kinds of experiments and observations on the long journey.

      Cargo shipping has only one objective: send the cargoes as fast as possible at reasonable price. If a shipping company can transport goods from one point in solar system to another, with a really low price and years of waiting, call this line unattractive is an understatement.

      IMO, unmanned cargo pods may be launched by mass drivers or accelerated by laser/particle beams, in order to reach the destination under a reasonable journey time. If they are fast enough that mass catcher may have difficulty to capture the cargo, it may deter some of the pirates to steal the cargo, or the more determined (company-backed ones or those bend on sabotaging the opponent's business) may choose to do it in Matter Beam's way: steal it at collection spaces or deceleration phase.

      Corporation warfare fighting in the disguise of "piracy", that is interesting.

      Delete
    3. Volume and energy are the names of the game. Using a mass driver to fire a pod on a minimum energy trajectory is the cheapest way to deliver a cargo in a reasonable time (a solar sail is probably cheaper, but the long acceleration time makes it slower).

      Having to wait isn't unknown in the investing world, agricultural commodities are generally bought and sold on the "Futures" market; farmers are negotiating a price before the crop even grows in the ground. People who distill whiskey need to wait 3-12 years before the product is saleable. For free flying pods, the general rule is to establish a "pipeline" of products, which will arrive at a set future date.

      So people living in the space environment may be used to thinking in terms of decades, and in terms of keeping energy costs as low as possible. In fact, I would almost expect that energy costs will be the backbone of the economy, and things will be priced in terms of how many kilowatts are used to produce/transport/use it. This also means the means of producing energy becomes of paramount importance.

      Delete
    4. Liqueurs are worth for years of waiting, they are some of the exceptions. LOL

      Finance is Greek to me.
      You may set price long before harvesting the crops or spend very long time to manufacture the products, but spending very long time to deliver them may force the customer to find alternatives.

      Let's say Jovians really export He3 to inner planets, but they only send He3 pods on a minimum energy trajectory. The customers have to wait for several years, or worse, Inners would never trade with Jovian companies that ship their He3 that slowly. That may even turn away from D-He3 Fusion, using D-T/D-D instead, putting more efforts on p-B fusion.

      Also, sailing slower and unpowered doesn't mean that it can really lower the price, at the end.

      http://www.independent.co.uk/environment/sailing-ships-back-in-vogue-as-a-green-alternative-to-conventional-shipping-a6858411.html

      Sending pods on minimum trajectory save money on energy "only", the shipping companies have to monitor the pods for much longer time, monitoring cost may outweigh energy cost.

      Delete
    5. You are sending bulk cargo, which is inherently "dumb". If the mass driver is short, you may be kicking the payload out at 100g to push it in a minimum energy trajectory, so no people or manufactured goods (unless you use a lot of bubble wrap.....). Monitoring will likely be the job of the Space Guard, simply to ensure it is heading towards the mass catcher at the correct speed and angle, otherwise there is little or nothing to be done.

      In the financial world, fortunes are made or lost on the "futures" market, where you are essentially betting on the prices of commodities years in advance. If you sign a 10 year contract for oil, you are hoping that the price you signed for will be lower than the inflation adjusted price 10 years from now when the last tanker pulls up to the terminal.

      Shipping cargoes like that in space is the same. There is the initial period where payloads are drifting towards the market, but once the "pipeline" is established, then the market will function in the normal way. An interesting analogy is the current spike in the price of whiskey. If distillers believe the demand is going to remain high, they will have to invest in more production, but the product they produce today in 2017 might not even be marketable until 2025 if they age it 8 years. Quite a gamble, if they ramp up production, they may win big, or depress the market by flooding it with large amounts of their product if demand has fallen in the mean time.

      Because space is an alien environment with different factors, the "rules" will be different. I am doing more thinking about this subject and may in due course request a guest blogging spot from our host because there is a great deal to discuss.

      Delete
    6. Some rules may be different, but many are still unchanged.

      Sending goods on a minimum energy trajectory may have another problem: it can't react to changes in market quickly.

      If the demand for a certain commodity in market has surged suddenly, even the reserves are unable to satisfy it, no trader can predict this before. While traders can order more from other planets or moons, the next batch would be arrived at few years later.

      Well-planning may lessen the impact, but in a rapidly-changing world, I suspect this shipping method may see limited use, only commodities that have very stable demand would be transported in this way.

      Delete
    7. @Thucydides:
      I don't believe that minimum energy trajectories are what the future solar system trade will be based on. Energy has a cost, yes, but that cost has to be proportionate to opportunity losses, the loss of flexibility, the greater proportion of your cash stuck in 'dead' inventory, the disadvantages of a loose supply chain, the longer cash turnover, the lower profitability over time, the loss of control over supply, the uncertainty over market conditions upon arrival requiring hedging against market movements and increasing insurance costs...
      and that's all management accounting. I haven't even started on the financial aspect!

      Yes, lower energy is better as it directly saves on the cost of transport. But, absolute lowest energy possible is unlikely due to the factors I mentioned above. How valuable they are compared to each other will determine the actual trajectories employed.

      I can see either end of the spectrum as being possible. At one end, you have The Expanse where transport is so fast and cheap that you have manned ice haulers. On the other end, solar sails are a good idea. Within each possibility, there should be variation between the types of products. Something like nitrogen will be have a demand pretty much consistent with population growth, which is slow and predictable. You can set your 'supply pipeline' up to minimize cost by using unpowered trajectories because you know that even if in ten years your product is just arriving into the market, you are very likely to make a profit. For more volatile materials (volatile in value) such as cutting-edge microprocessors, you cannot afford to take such a risk. You need them delivered fast, while they still have value. Add in a few more categories, such as degradable products, and you can have a solar system trade with the full range of fast to slow transport being viable regardless of how expensive transport is.

      Like today. Shipping is probably the cheapest transport we can get, excluding pipelines. A read of this document on the economics of oil (https://goo.gl/zxW893) shows a $4.8/ton shipping cost through the Suez Canal, excluding charges and after insurance premium is removed. Air freight however costs upwards of $4500/ton, and yet it is still used.

      Maybe because air freight delivers products within two days instead of two months.

      Guest blogging spot is always open. Feel free to contact me by email to discuss in private if you wish.



      Delete
    8. Regardin the cargo aspect:

      Minimum delta-V shipping might be suitable for some bulk items that are shipped on a regular basis. The long travel times really don't matter here so long as the shipments are regular. This would cover such items as "shaving lotion, toothpaste, and toillet paper". In other words, these are non-perishable items that are used on a daily basis, at consumation rates that can be planned for long in advance. Essentially, your supermarket deliveries. You COULD include utilities shipments (the equivalent of petrol and drinking water, etc), but most of this would preferably be supplied ISRU.
      For other cargo shipments, people are not going to want to wait years. Prices will, of course, be determined by velocity classifications. The faster the delivery, the more expensive it will be. However, this probably will not be anywhere near as expensive as one might think.
      In addition to the dry bulk shipments, there will be the regular shipments of perishable items. These will likely have shelflives on the order of a year or two, so there will be regular delivery shipping lanes at significantly higher delta-Vs. The bulk and regularity of these routine shipments will keep costs down. There might be other "specialty" items that have shelflives a little less than a year. There will also be regular traffic of people... business people, people relocating on a semi-permanent basis, rich tourists. These people will not want to spend several months in transit, so you will have the "airliner" class.
      Now, quite often, these vessels will tend to be underbooked. However, for various reasons, the owners will want to send these vessels out with consistent mass loads.
      For long distance shipments, current shipping companies very rarely rely on their own shipping fleets. Instead, for non-prioity shipping, they usually reserve a number of shipping containers (of appropriate size) that can be loaded "last-minute" on a regular transport service. The transport service then tells them how much mass they have available on a given passage, and the shipping company fills their container(s) with that mass of product waiting for delivery (first according to priority, but then according to a first-come-first-serve basis). Since no one is making a special trip, and the mass is not going above the budget, shipping prices remain low.
      That said, sometimes there WILL be a need for emergency "priority express" delivery, so a shipping company serving long distance routes WILL often maintain a modest fleet of vessels or craft necessary. These will be the high-speed clippers, roughly equivalent to the fastest of human transport vessels/craft.

      Delete
    9. @Mikkel Haaheim:
      Good point on the bulk items.
      The 'not going to wait years' is a bit more complicated. Space, for example, has radiation damage randomly damage electronics or micrometeorites penetrate pressurized spaces. This imposes a constant rate of damage on transported goods. Another interesting twist is that a space hauler can modify the propellant to payload ratio on the fly. It can load up half the propellant and make a slower, closer to Hohmann trip. Or, it can go nearly empty, but with a full propellant load, and suffer only a minor penalty compared to the fastest transports on quick trajectories.

      The personal transports might lead to a shipping market that is composed of dual-purpose 'clippers' as you say, with room for passengers that can be quickly modified to fit any cargo.

      The latter are interesting targets for role-play settings.

      Delete
    10. The potential for damage is just another reason that clients are not going to want to wait years for arrival. However, it is possible to arrange the loads to limit such attrition, to an extent.

      Yes, you can send lighter loads. OTOH, there is a certain amount of mass that is fixed (structure, thrusters, support equipment, etc). Sending a reduced payload means that you are decreasing the overall amount of payload that a given amount of propellant can "ship". This means that the shipping will be less efficient. Keep in mind that existing shipping companies could likewise cut unnecessary fuel loads to ship lighter loads. However, in order to maximise efficiency (and, thus, profits), they prefer to ship consistently with maximum loads. Shipping non-priority cargo on a first-come-first-serve basis is a prefered means for ensuring that all shipments (including flights with passengers) are carrying a maximum load.
      BTW: this also reduces the possibility of potentially fatal errors (underestimating the amount of fuel/propellant required, etc).

      Delete
    11. It is not so much that the "dual purpose" transports are modified. All aircraft (and probably future spacecraft) are much more mass limited than they are space limited. You often have flights with a large number of empty seats. But all passenger vessels are equippedwith cargo space (especially for the use of passengers). There is actually MUCH more cargo space available in these craft than are typically used. When there are fewer passengers, the cargo spaces are filled more completely. That is all. Conversions such as you describe are only necessary if you are going to ship oversized items.

      Delete
  17. I wonder if we are talking past each other. I am specifically speaking of bulk and commodity cargos going by unmanned pods on minimum energy trajectories. Things like people, computer parts and cut flowers go by "Fast Packet" (which can by fusion powered, or use augmented solar moth thermal rockets or laser driven sails).

    However, just like here on Earth, bulk commodities are going to be a much larger portion of the economy than cut flowers. Computer chips are an intermediate case, I suspect most polities will be building their own devices via 3D printing (and then the business case is do you build the parts for more 3D printers to expand your economy, or do you import more 3D printers....)

    ReplyDelete
    Replies
    1. It happens! Sorry, I've been very busy lately and I should pay more attention.

      Find more of my stuff on twitter: https://twitter.com/ToughSf

      Delete
    2. The emphasis on using minimum energy trajectory inspires me to have an idea, but Thucydides has already said some of it.

      Let's say an interplanetary chip manufacturer may setup factories on Earth and some asteroids, but they also authorise some companies on certain planets, moons, asteroids or even space colonies produce the latest product by 3D printing or nanofab, in order to release the new model across the system simultaneously.

      The only few things the OEM need are blueprints and rare metals for prodcing key elements, they get other materials by ISRU. This may minimize the problem of slow transportation

      However, I don't know how to deal with the problems of blueprint leaking to the public, even though not everyone can get rare materials to produce the crucial parts easily.

      Delete
    3. @Felix:
      Blueprint leaking isn't too much of a problem if there is any sort of copyright or patent law. You can go ahead and make your microchips out of stolen plans, but you can't sell them to anyone and the blueprint's original owner can sue you for everything you have plus fines and profit loss if you are discovered.

      Its kind of like the copyright situation today, especially with digital media. You can copy a file and distribute it illegally, you might even try to pass it off as your own and try to make money off it, but the risks involved confine this practice to a tiny segment of the market or to places where the copyright rules are not enforced (Russia, China).

      You might end up with the larger habitations in the solar system having enforced rules that protect leaked blueprints from being commercialized and costing potential profit to the blueprint's designer, but with more isolated or vulnerable habitats taking the risk anyway. If the blueprint is for some vital component, like recyclers or avionics, the blueprint manufacturer might simply accept that people's lives might depend on a pirated blueprint and allow it anyway. An example today is Windows OS. You can install it without a key and it will keep your machine running, but it will have limited functionality.

      Delete
    4. Oh, I just forget that the laws can handle this.

      Printing parts for 3D printers...I think only the more isolated settlements may do this
      In the more developed areas, perhaps importing a new printer is much easier and faster than printing and assembling on your own.

      Delete
    5. My maths is very bad, tell me if my calculation is wrong.

      I discussed the shipping mass driver with one of my friends yesterday, when I read Thucydides' reply involving mass driver and checked some estimations on mass driver, a problem emerged.

      I understand that some goods are more fragile and unable to survive under heavy acceleration of several thousands Gee, like that provided by a design from NSS in 1980s.

      http://www.nss.org/settlement/L5news/1980-massdriver.htm

      However, if Jovians want accelerate their cargoes for Earth to 9 km/s at 100G by mass driver, they need to make it 40 to 50 km long.

      Maybe Jovians should install laser or magnetic sails, propelling them with laser/particle beams powered by the Jupiter magnetosphere, the system should be smaller and cheaper than a dozens-km long mass driver.

      And people from other planets will less likely to oppose their construction, a 50 km-long mass driver weaponized above someone's head is quite threatening.

      Delete
    6. Sorry for the late replies recently. I was away.

      The products that are most likely to survive high accelerations are conveniently also the products that are most in need of low-energy transport, such as bulk metals or volatiles.

      If you have room to spare, you can build massive circular coilguns. The acceleration is gradual, up to the breaking point of the wires holding the coilgun tracks 'rim' together.

      Lasers are even more convenient. They can be weaponized just as easily as a coilgun (replace a laser-thermal rocket with a military mirror as target), albeit not on an interplanetary scale.

      Delete
    7. Laser beam stations may raise some eyebrows of neighbors, however, if there are not much mirrors placed in space, the outcry may be fairly limited.
      Also, the mirrors should be quite easy to spot. If the laser network owner deploys so many mirrors, other parties may take actions to stop it.

      However, mass driver has an effectively unlimited range (with a relatively low velocity), it can be difficult for outsiders to distinguish a regular coil maintenance or a military upgrade that make "Big Stick" Super MAC from Halo looks like a toothpick.
      That can trigger system-wide panic, if the relations between planets and moons are tense enough.

      Delete
    8. A combat mirror can be accelerated like any other projectile and kept hidden until it is needed. It can then suddenly throw off its cryogenic shroud, inflate to full diameter and call in a laser, which it then focuses to deadly intensities. I agree however that its reach is much more limited than a mass driver, that can shoot projectiles at interplanetary velocities to anywhere in the solar system.

      Delete
    9. "it can be difficult for outsiders to distinguish a regular coil maintenance or a military upgrade"

      Quite right. Especially since the exact same coil can be militarised simply be using a much smaller mass, with a simple modification to the timing mechanism (subsequent coils will need to "fire" more rapidly, in order to maintain the higher acceleration rate... otherwise, you might end up putting a "break" on the launch.
      The only real difference is how well the hardware is reinforced and protected, and the specifics of the "load" being launched. The load can easily be hidden until the actual launch.

      Delete
    10. Laser weapon webs require many mirrors or lens to attack other planets or intercept incoming fleets effectively, deploying them is relatively obvious.

      Cargoes that somehow decelerated and stopped at certain positions.
      Heat stealth ships lurking around and dropping unknown payloads, even they are supposed to be stealthy, perhaps some very dedicated systems and observers can still track them.
      Mirrors or lens deployed in numbers that far beyond normal launching requirements etc.

      All of these, perhaps even the slightest anomaly, can grab the attention of intelligence agencies. Being paranoiac is part of their job after all.
      Too many actions are needed, this may make the deployment leaky, even compromising the whole operation.

      The kinds of military upgrade for mass driver I was thinking yesterday were adding extra capacitors, radiators and even reactors. If the mass driver is large enough and has some busy traffic, conducting these upgrade may catch less attention.

      Combining with other regular maintenance, like replacing coils that can provide higher acceleration, weaponizing mass driver can be done more secretively, since less large scale movements are needed.

      Maintenance ships traveling back-and-forth, trying to repair a "faulty" mass driver, with cargo ships carrying new components and factory ships recycling old parts and churning out new ones.
      A faulty mass driver may even cheat the foreign intelligence that the big stick over there is anything but interplanetary strategic weapon.

      Until it is too late and the worst happens.

      Delete
    11. No need to add capacitators, radiators, or reactors. "Commercial" mass drivers will be intended to launch extremely heavy loads, typically measured (at the very least) in tons. The energy required to do this is already quite significant. Now, imagine sending shells that mass a few kg through such a system. Instead of launching 1 tonne at 10km/s, it could launch a 10kg shell at 100 km/s, using the same power output. The only thing that needs to be changed is timing mechanism between coil activations, to ensure constant acceleration (acceleration along the same length of launcher will have to be increased in order to achieve the greater max delta-V, which means that successive coils will have to be activated in more rapid sucession).

      Delete
    12. Railguns might be less sensitive to using smaller payloads for higher velocities than coilguns. Coilgun switching speed for the magnets has an upper limit determined by heat and current. If you're pushing a smaller payload at a higher velocity, you are creating and destroying a magnetic field more quickly than during normal operations. The switches have to work faster, and there'll be stronger residual fields behind the the projectiles. This will certainly lead to lower overall efficiency and a 'maximum speed' for the coilgun, regardless of the energy balance.

      Delete
  18. There is one inherent problem with mirrors. The problem is that, unless you are sending the beam back almost to its point of origin (which is counterproductive), you are going to have to expand the mirror along one axis (or, alternatively, sacrifice a great deal of transmission energy). This greatly increases the mass of the mirror, its vulnerability, and the difficulty in handling it.
    A much better approach would be to use a lens. You might have a bit of difficulty designing a lens system that will survive a high power (especially military) laser emission, but the end result would likely bemuch more "doable".

    ReplyDelete
    Replies
    1. We might end up seeing laser webs have zig-zagging paths and rectangular mirrors then.

      Delete
    2. Possibly. But this would be quite inefficient.

      Delete
  19. BTW I'll mention the novel _Antares Dawn_ by Michael McCollum.
    There is a scene in which the major inhabited planet of a solar system is being attacked. The defenders, aside from having lots of space war ships, have a lot of high power lasers set in the oceans & ice caps of the planet. The enormous heat sinks available make it possible for them to be *very* high power & they get used to take out the missiles that get past the defenders space ships.

    ReplyDelete
    Replies
    1. Recently got the audiobook for this, thanks for the recommendation.

      Water is a fantastic heat sink.

      Delete