Safety Margin for a Artificial Orbital Habitat

Question

I'm always interested in the possibilities of human creating an orbiting space habitat like Stanford Torus or O'neill Cylinder. Those habitats were designed to support certain numbers of colonists.

Now for the question: Based on current safety practice (real world),

What should the safety margin of these habitats be? (Ex: If the habitat's intended pop was 10,000 then with a safety margin of 200% they can safely support up to 20,000 people).

Also, how much redundancy should the colony's systems have? (Ex: If the colony life support system was a 4 tiered redundant system then the life support will only failed if all 4 layers failed).

Answer 1

I would look at it a different way. There are other questions skirting around the edges of the questions you asked which are more valid for very large population biospheres.

A complex system like an orbital habitat will operate in shades of grey, not black and white. Concepts like "safety margins" and "number of redundant systems" are too gross, and the end result will always be an unsustainable habitat, or a habitat that is 10x too expensive and barely operates.

An excellent example of what an artificial habitat has to deal with is to look at our own habitat. Consider the "safety margin" for number of humans it supports? We think about how hard we are pushing the earth as we climb our way towards 10 billion. But what was the safety margin in the first place. A quick search shows that worrying about population is nothing new. In 2nd century Carthage, Tertuilian lamented that the world population of 190 million was too much and would crush the Earth. By those estimates we're looking at a safety margin of 5200%. A better way of looking at it is that we rarely are anywhere near a "carrying capacity." More accurately there is a fluid quality of life degradation as population increases, and we humans eventually set a bar at a minimum quality of life we will endure (we've also made technological advances to push that quality of life issue out).

Likewise, a focus on number of redundant systems leads astray: most systems are not instant-death if they stop working. There's no system which, upon failure, mindlessly opens all the airlocks and vents the habitat. When a system stops working, a clock starts. As time progresses, quality of life diminishes. What matters is not how redundant the systems are, but how gracefully the systems can be repaired ore replaced after a failure occurs. Consider: you only have 2 corneas, one for each eye, yet they are known for being hardy little structures, because the body is configured to make it rather difficult to damage them beyond what they can repair (including a blink reflex, which helps avoid most major cornea trauma.)

EDIT:

As a closing, working with a complex system is always a balance between two extremes. On one hand, you have to predict ahead of time what the system will do. On the other hand, those predictions will be wrong 100% of the time unless you observe, update, and adapt as you go. When designing large systems, if it feels like you need too much expertise in the design phase, try relying more on listening to the system while it operates and making changes as it goes. If it feels like you can't hear her whispers, then you need to make more predictions. Remember Kaylee from Serenity. She knew her stuff, but what made her special is, in her words, "machines just talk to [her]."

Answer 2

Your question is flawed

Safety factors figure into catastrophic (i.e. unmitigatible) failure scenarios, and cannot reasonably be said to apply to the sort of scenario envisioned. When discussing large and self-repairing systems like this, we talk about the limits in terms of it's ability to function "sustainably", to house people "comfortably", etc.

A space station will possess self-diagnostic tools. Parts of the system will be expected to fail at a certain average rate, and the system will possess a sort of additional capacity in order to make up for that. Even in modern times, a freak unexpected shutdown of vital life-support equipment for several days for a station operating beyond capacity should be relatively easy to manage, provided the persons responsible for logistics were competent. Generally speaking, the crew of the station will always have time to respond to overpopulation pressures long before any negative physical health effects would present themselves.

Provided that the crew expects to be operating beyond the old capacity indefinitely, modifications to the ship can be made to further increase the station's capacity. Provided that they only need to do so for 'a while' (possibly as long as several months) space-efficient but expensive consumable reserves can be delved into.

That said, in general, the limiting life support factor for humans in space currently is actually healthy air, which, because technology, is largely the same as water. The ISS produces via recycling and electrolysis ~80% of its expected oxygen needs, and receives the other 20% via shipments from Earth. It also stores 140+ days worth of backup oxygen for it's expected personnel load. Typically, it gets a resupply ship every 90 days.

Extending the function of the ISS by ratio to our theoretical space station:

The station can run unsupported at  |  For
80%  capacity                       |  practically indefinitely
100% capacity                       |  11.67 duty cycles
133% capacity                       |  3.89 duty cycles
150% capacity                       |  2.92 duty cycles
200% capacity                       |  1.67 duty cycles
250% capacity                       |  1.17 duty cycles
300% capacity                       |  0.90 duty cycles
500% capacity                       |  0.47 duty cycles
750% capacity                       |  0.29 duty cycles
1000%capacity                       |  0.21 duty cycles

Probably oxygen would stop being the limiting factor at some point in this process, but the function described by this data would still be fairly representative of projects like this in general; operating above capacity mostly just decreases the time the crew has to respond to normal problems encountered in the course of operation and makes the success of supply missions more important. At some point (in our ISS-based model, around the 300% capacity mark) a rescue mission or other extreme action is required as the station cannot survive until the next scheduled resupply on its own. Nonetheless, even in the extreme 1000% capacity scenario the crew has a lot of human-experienced-time to decide what to do about the problems (19 days for the ISS itself).

Answer 3

What should the safety margin of these habitat? Also how much redundancy the colony system should have? These are excellent questions though except in generalities, we can't answer them here.

Safety Margin

A simple, naive answer would be a factor of 2. For example, an oxygen system is capable of handling 10,000 m^3 of air per day but the operational requirements are only for 5,000m^3 per day. If you have two such units, your habitat can tolerate the complete failure of one oxygen system and still operate indefinitely. Granted, you want to get that fixed as quickly as possible but you have enough leeway to fix it properly instead of some rushed hack job.

Redundancy

A simple, naive answer is "Two of everything" and "At least two ways to do everything". You'll see this play out in large computer systems where systems architects are perpetually on the watch for single points of failure and eliminating them as much as possible. Some times, these single points can't be eliminated but in such cases, lots of engineering work goes into making sure the equipment/system is incredibly robust. If the habitat only has a single airlock then that airlock will be incredibly resilient to bad bumps, overspeed docking, torsion moments, vibration, accidental kicks by ingressing astronauts, etc, etc, etc.

Nature of Failure

As other have pointed out, when a system fails, especially in an environment like space, it should fail-safe, ie, the life support system doesn't vent the remaining atmosphere to space when it fails.

Disasters are rarely the result of single large failures. Most often they are the result of a string of small mistakes and failures that individually are easy to counter but collectively cannot. This has proved true in practically every domain where safety is a concern. Big failures, while an omnipresent possibility, don't happen nearly as often as failure cascades where a small failure of one component leads to another failure which leads to another failure until the entire system is compromised and catastrophically fails. A lot of engineering work on the ground will go into designing systems that preclude failure cascades.

In addition to equipment design, training of astronauts is critical. Resilience research has found that it is often the human element that can accommodate failures in the systems. Proper procedures and training need to be in place to fix problems that can't be anticipated or adding the ability to detect a certain error condition would add an unacceptable cost/weight of a particular system.

For further reading about resilience engineering, might I recommend the book "Resilience Engineering: Concepts & Precepts". I've found it a fascinating read about how safety systems degrade and how to treat that degradation.

Answer 4

I'm going to take a stab at this question, but as James has mentioned, this is up to you. I'll take his suggestion and use benchmarks that you can choose to apply or not.

Safety Margin

In the oil industry, on-site safety personnel assist in analyzing work environments and designing processes to control, eliminate, and prevent problems. I use this industry, because I'm familiar with it, but also because it is high-risk, expensive, and nobody wants anything bad to happen.

A good rule of thumb is to have 20% contingency on all equipment for safety purposes on rigs that don't have immediate access. For your purposes, I'd make the outer walls 20% thicker than necessary, etc. throughout your design. Again, this is a design solution, depending on your space colony.

For the number of colonists, I would design a capacity for 20% more as well. If you aim to have 1,000 people, design it for 1,200 and ensure that you have a mechanism for removal of babies or other people to keep it around 1,000 or less.

Redundancy

Redundancy is again your choice. When I was an urban planner in the UAE and they introduced the first Nuclear Power plant, at Baraka, we understood that the rule of thumb for safe nuclear power was "4" for systems such as cooling, etc. I would suggest for your critical items that you have four of everything, because it will take a while for backup supplies.