N+1 contingency is very important when it is applied to areas that a loss would be catastrophic. These sort of situations are luckily few and far between but when they do the initial program of engineering includes some form of redundancy within the system. N+1 means that a process will continue to operate with the failure of an item of equipment and it will not have an effect on the process continuity. Examples of this would include a tunnel extract system - a failure of kit would endanger life. So an N+1 redundacy is built in. It is configured in such a way so it is there and ready to go with minimal amount of fuss. This is shown when you have two systems running side by side, if one fails the other takes over, and an alarm is generated to alert that attention is needed. The process itself is not affected.
REDUNDANCY - Googled from the internet......
Providing secondary components that either become instantly operational or are continuously operational, so that the failure of a primary component will not result in mission failure. There are three levels of redundancy: the first level is called N+1. With N+1 redundancy, one more component (1) than the number required to do the job (N) is provided. If, however, a facility has to shut down those N components for periodic maintenance, it may require the second level of redundancy, called 2N. This consists of two systems, each of which contains N components – in other words, twice the number of components needed to do the job. The third level of redundancy is called N+1x2. This consists of two N+1 systems in parallel. In N+1x2 redundancy, one side can be taken down for maintenance, and the other side will be protected from exposure to the risk of downtime by its own redundant module.
Electrical utilities are often REQUIRED to maintain N-1 redundancy for all transmission, generation, and major distribution components.
The system is obviously required to work properly (i.e maintain standard voltages, acceptable currents, etc) when all N components are available
The N-1 criteria requires that all loads can be restored if any single component fails (i.e. N-1 componets still available). Note that this does not mean no short-term outage should occur, only that the load be quickly (definitions vary on how quick) restorable.
For example, since it can take days to change-out a 300MVA power transformer (assuming a spare is handy), it is usually necessary to have an in-service spare available on-site for every major transformer in the utility's system (or sufficient spare capacity elsewhere in the system to cover for it).
Some critical systems have gone beyond this and enacted N-2 criteria (which as the name suggests, allows for failure of any two components).
In my line of business, data centers, N+1 is important, but many time 2N is more important.
N represents the amount of actual or future load, and N+1 is the ability to have one more "X" to backup the other items. "X" can be any number of items, such as
UPS Modules, generators, HVAC fans, Chiller pumps, etc.
2N means that you have redundant systems running side by side, where one can completely replace the other in function.
Using round numbers, so if the load is 4MW and you have (3) 2MW generators, that is N+1. (4) 2MW generators seperated in most cases into halves, is 2N.
In my electrical system planning days long ago before the new age of driving things till the smoke comes out, we defined "N-1" as being able to withstand the loss of any one item of plant (line, transformer, generator etc) without loss of load or adverse voltage outcomes - ie the system would carry on to all intents as normal. We also planned at "N-2", so that on the loss of a second item of plant, you could recover with some loss of load, or possibly aome adverse voltage outcome. At "N-3" you just made sure the system would "fail to safety", with plant tripping or being shed by automatic systems to avoid palnt damage. But I don't mean 'fail to safety' in the same strict sense as a railway or chemical engineer would.
N+1 and 2N can be viewed as very similar and when combined they give a very resiliant system. Computer halls are very very fragile systems to the likes of a power break or dip. This can cause untold down time and such as they bring systems back on line after the outage so best response to this is to factor out the possibilities.
2N being as was pointed out two systems running in parallel to support the load. At any one time the load can be supported by one system. The load sees no break. N+1 being an alternative system running side by side. Failure of one system will result in a transition to the other system through static switchers or other menas so that the supported load sees no break.
In other words look at a process that requires absolute reliance on security of supply. You would have a dual string capacity - each string being able to support the load in its entirety (N+1). Now break this down even further to include 2N on each string. Two no-break UPS units tied in sync together. Each one of those able to support the load on its own. You now have a system that approaches cast iron stability. You have accounted for failure of supply and also for failure of a unit.
Main thing to to imagine what can go wrong and try to work around it so that you can cover for any eventuality on the system. Generally systems grow in direct proportion to the amount of money that would be lost if it fell over and the fear that the suits at the top have of this happening.
In think we should define N-1 a little better. Usually the '1' is the largest item that can fail, for argument sake if you have a 2MW load and a 2MW and 1MW genset then you do not have N-1 redundancy. It is not about the number, but the size!
Regarding a comment in a previous post: "You have accounted for failure of supply and also for failure of a unit."
Achieving true 2N is not as easy as it sounds. Be careful that you don't introduce any single points of failure downstream of your expensive 2N or N+1 system. It would be a shame for a failed system controller, tie breaker or static transfer switch downstream of a 2N UPS system to shut down your loads. The ideal 2N scenario is to carry your two completely separate systems all the way to the indiviual loads if they can accept two power supplies.
This was the system that I was trying to describe maybe Im not that clear but the end result is that it goes down to two power leads going into the back of the server, both powered up and able to run that item even if the other goes down. Thus resiliance of the system is maintained.
