.gif)
SYSTEM RELIABILITY / WHITEPAPERS
Reliability Calculation
Harmonics and Harmonic Distortion
Power Problems
Power Quality
Power Factor
Solutions to Power Problems
Introduction to Motor Generators
Hybrid Rotary UPS
RELIABILITY FOR THE LAYMAN
DEMONSTRATED RELIABILITY IS TRUE RELIABILITY
Protection for your critical load is the reason that you are going to purchase an Uninterruptible Power Supply (UPS). It is only natural that you wish to purchase the most protection you can for a given investment. However how do you sort through all of the claims of ultra high reliability, MTBF of over 1,000,000 hours and other statements made by UPS manufacturers.
In this article we will try to simplify some of the technical jargon and look at the problem through the eyes of a layman.
- First some basic facts:
- Every single item ever made has some probability of failure.
- Different items of different complexity have different probabilities of failure.
For example...
A short piece of copper wire has a very low probability of failure. A manufacturing defect, electrical current high enough to melt the copper, or physical damage to the wire are about the only events that would cause the wire to fail.
An Automatic Voltage Regulator (AVR) on the other hand is a very complex item. An over simplified description of an AVR's function is to measure the existing voltage, compare it to a voltage set point and to send a correction signal if actual measured voltage and set point voltage are not the same. AVRs typically do the above function hundreds of times a second. The AVR is also made up of hundreds of individual parts, all of which have to function correctly or the AVR will fail.
From the above two examples you should agree that the probability of failure of a short piece of copper wire is much, much less than the probability of failure of an AVR.
Now assume that the wire is connected to the AVR and that only these two items make up a sample system. Both items are equally important to the system but they have vastly different probabilities of failures.
By the analogy that "A chain is only as strong as its weakest link," the probability of failure of our two item sample system, a wire and an AVR, can not be better that the probability of failure of the weakest item. In our sample system this is the AVR. This same analogy holds true for very complex systems with thousands of parts.
As you might imagine trying to calculate the probability of failure of a very large complex system can be a frightening task. Many years ago the U.S. Government came to the rescue.
In attempting to define the reliability of some Defense Systems the term "MTBF" (Mean Time Before Failure) was defined. (See MILHDBK-217D). The theory was that each item was assigned a probability of failure and all the various probabilities of failure were mathematically manipulated until a final number of hours was calculated. However please note that the final number of hours calculated was not the number of hours before a failure occurred but rather the number of hours at which there was a 63.2% probability of failure! Let me restate this point: A system having a mathematically calculated MTBF of 1000 hours has a 63.2% probability of failure by the 1000 hour mark. This is in fact the basic idea behind mathematically calculated MTBF's.
Power Systems & Controls on the other hand uses an experience calculated MTBF. We calculate the total number of systems running, multiply by the number of hours being run on each system. Add these numbers and divide by the number of failures that have occurred. This gives you a true picture of the actual length of time you can expect your system to run without any failures.
Power Systems & Controls is also very conservative in the definition of a failure. If any item in our system fails that is counted as a failure even if the critical load is transferred to bypass and not affected by the failure. Other competitive systems only count interruption to the critical load as a failure.
When you see very different MTBF values, ask yourself the following questions:
- Is it a mathematically calculated MTBF?
- Is it an experience calculated MTBF?
- What is classified as a failure?
Before we complete this section there is one more point to make. Remember from our very basic item sample, a wire and an AVR, the loss of either item would be a failure of the system. Therefore the MTBF of the system can not possibly be greater than the MTBF of the lowest rated item. In our sample system this is the AVR.
PS&C's major rotary competitor quotes an MTBF of more than 600,000 hours for a UPS module. This value does not count a bypass transfer as a failure and it may be a mathematically calculated value. However there is only one AVR on that module and without a properly functioning AVR the module will not operate. If you recall from above, an AVR is made up of hundreds of parts making hundreds of calculations each second. It just does not seem possible nor have our engineers been able to find an AVR that will run 68 years (600,000 hours) without a failure. Please examine this claim very closely.
Now assume that we have a sample system each with two pieces of wire and two AVR's. I think it is obvious that the MTBF of this system is one half of our original sample system. It has twice as many equally rated parts. Therefore there is twice the probability of failure. Therefore the MTBF is approximately one half. (This statement can be proven with the Laws of Probability and statistics but that would defeat our purpose of writing this article for the layman.)
Power Systems & Controls makes use of the above theory in several ways in order to provide the most reliable UPS on the market. First we design our systems to have the fewest possible parts. The fewer the number of parts the fewer the number of failures
We also manufacture many different size ratings. This seemingly very simple fact offers huge advantages in reliability. If one module will do the job of two modules the expected reliability will be double.
REDUNDANT SYSTEMS
Because of the value of the critical load that UPS's protect, most systems are redundant. That is, there is one more module than is required to supply the load so that in event of trouble on one module there is sufficient capacity available to protect the critical load. Let us examine this situation in detail.
One module of a multi-module redundant system is down for repairs. The critical load is being protected by the remaining modules. If there is a failure of one of the remaining modules before the first module is repaired and back on line, then the critical load will be unprotected. The seemingly simple option of larger size ratings now begins to pay dividends.
As an example: assume that the critical load is 2000 kVA and the system must be redundant.
- System one consists of four 750 kVA modules
- System two consists of three 1000 kVA modules.
From our above discussions we can assume that the probability of failure of system one is greater that the probability of failure of system two. Again fewer parts, fewer failures. However one failure is not catastrophic because our example system is redundant. Another module failure before the first module is repaired is catastrophic.
During the repair of the first module, system one has three non-redundant modules running while system two has only two non-redundant systems running. System one has a greater probability of failure than system two by a factor of 3 to 2.
MEAN TIME TO REPAIR (MTTR)
An equally important factor of system reliability is MTTR (Mean Time To Repair). If you consider our example above, the critical load is exposed to a catastrophic failure during the time of repair to the first module, The longer the first module is not repaired, the greater the exposure to catastrophic failure. Power Systems & Controls has designed our products to have the shortest possible MTTR.
All of the components are easily accessible and no part needs to be removed to gain access to another part. This is an extremely important feature and can only be really appreciated by actually looking at the UPS.
CONCLUSION
This article was titled to be for the layman. I hope it is useful to you. I have tried to show how and why Power Systems & Controls can claim to provide the most reliable UPS on the market. Throughout the product development history of Power Systems & Controls, one thought has been paramount: Provide the Most Reliable UPS. Every single design decision is made to increase reliability. The result of this single minded devotion to reliability can be summed up in the following points:
- We use a single well proven control concept.
- We use the fewest number of parts in the industry.
- We supply the widest selection of module sizes in our marketplace.
- We design for the shortest possible MTTR.
After all, you are going to buy a UPS for RELIABLE protection of your critical load. You should know the facts. The facts confirm PS&C has the most reliable UPS available to the market today.