To: Nick Lappos
Quote: “I was right, he couldn't keep it below 23 lines. I count well over 50”!
Well at least we know that you can count. Your response is typical of those you made in the past where you tried to vindicate yourself in the eyes of the other contributors to this forum by saying you were right all along but in the process you avoided responding to the technical content of the post.
To: tecpilot
Please understand I am not being critical of the design or the application of the fly-by-light control system, as I believe it is the wave of the future for both aircraft and helicopters. What I was critical of was the statement of having a catastrophic failure rate of 1 10 9. It is the means of determining that it is possible to meet the requirement. Reliability and its’ companion System Safety are totally dependent upon the manipulation of numbers in a database. For any given component you may have as many as 50 different entries showing the failure rates. Notice I indicated rates as opposed to rate as they are all different. In order to calculate the reliability of a system the analyst will pick a number that will allow him/her to show that the system meets the spec requirements. If the analyst can’t find the number he or she is looking for they can pick a number from a component that is used in a totally different environment and multiply it by certain K factors dealing with operational environment. This provides the required number but it does not truly represent the same item that is in the system.
Once the system reliability has been established (on paper) the end failure rates are plugged into a mathematical calculation that is representative of a fault tree. The fault tree consists of And gates and Or gates. On an Or gate any failure that occurs will pass onto the next level. With an And gate all of the failures that feed into it must be present for the failure to pass on to the next level. The failure that passes through any gate is calculated by either adding the entering failure rates or multiplying those failure rates. This process continues upward until the catastrophic failure rate for the system is calculated. Using the failure rates compiled by the reliability engineers who in most cases are not truly representative of the real failure rate the safety engineer runs his calculations. The end result is that at the system level the analyst can show that the system has met or exceeded the requirement for a catastrophic event established by the certifying authorities (As indicated in Nick Lappos’ post above). On one program I worked on the calculations for the systems catastrophic failure rate was as high as 1 10 18 (I can’t even count that high. Maybe Nick can).
There is one major problem in this method. Only system catastrophic failure rates are determined. There is no calculation at the aircraft level. If you can visualize the fault tree with its’ various combinations of And or Or gates then take it one step further and create an Or gate that is representative of the aircraft. Any system failure that can down the aircraft feeds into this Or gate so that any one will cause the loss of the aircraft. If you use the same Boolean Algebra used at the system level then it can be shown that the aircraft does not meet the certification requirements of 1 10 9 catastrophic failure rate. This is the point I was trying to make to Nick Lappos. The 1 10 9 rate is for the systems not the aircraft. I have worked to FAA and JAR requirement as well as Def Stan requirements and the methodology of calculating failure rates at the aircraft level are all the same. To put 1 10 9 in perspective Jesus Christ was born 17,542,776 (approximately) hours ago. The ratio of his birth to 1 10 9 is 1:57. Granted, we are not working with a system in isolation. The 1 10 9 represents accumulated flight hours for the fleet. Lets address a typical long haul jet that flies 3500 hours per year and there are 100 in the fleet. It would take 2,857 years to accumulate one billion hours. Can you visualize any mechanical or electrical system lasting that long or longer prior to catastrophic failure? If you placed an anvil in an airless chamber it would not last that long. Neither will the airless chamber.
However if you employed the approved data bases and used the approved calculation methods the certification authorities will approve the design. The only document related to the whole process that is turned over to the certification authorities is the safety hazards analysis. They never see the FMECAs or the reliability analysis unless they come into the offices of the firms Assurance Engineering Department. It all boils down to Garbage in-Garbage out.
Nick to save you the time in calculating the verbiage in my post here are the stats:
Pages 2 (when prepared in Word)
Characters (no spaces) 3923
Characters (with spaces) 4800
Paragraphs 9
Lines 65
