Downloaded from https://www.everyspec.com AMCP 706-19 V CHAPTER 10 RELIABILITY GROWTH L, U *subscripts which Imply Upper and ing involves changes In desig~n details, pro- Lower uncertainty Ouction methods, inspections, etc. Mr) *mean number of failures In the Not all fixes cure the ilis they were sup- interval (0 TI posed to cure, and sometimes fixes introduce -drin tj par. new illp that weren't there before. 1iis Is "I number of failures drigt(p.
one reason that the Army Is properly skep- 10-*2-2) tical about claims that product ills have been cured, when there is no physical testing to nr i number of failures In the interval support that assertion., It is also reason to (0O r) bu skeptical of reliability growth miodels that t utimecan't decrease. Thes repeated process ot experience-fIx is *time, at which failure I occurred, called Reliability Growth, Often If con'tinues test time of uldt I (par. 10-2.2) throughout the life-cycle of a product.
Since relinbility growth ought to be present In all T = time at which prediction is made programs, It Is desirable to have a method at' predicting wh&ut the railiAbility will eventually 40) - measure of uncertainty In lIt WO grow to. The prediction method uses a mathematical model, i~e., a set of equattions u constant in the equation for mn0) whose parameters can be estimated from historical di-ta on this product or similar X~) -failure' occurrencc rate, peril rate; ones. There are many such models, only one2 intensity of Poisson process of which is treated lit detail In this chapter.
X * failure rate of unit 1 10.2 DUANE MODEL 10.1 INTRODUCTION The Duane model (Refs, I and 2) and modificatio'ns thereof are based on the idea Contrary to conventional wisdom, a prod- that the more resoutces which are devoted uct when It comes from design engineering, to finding and fixing what is w~i ,)ng, the does not have an "inherent" reliability which bettor the reliability will be. The model is is then degraded by the manufacturing do- usually expressed as "The 'log cumulative partment. Rather, the product contains MTBP' Is prc'nortional to the 'log cumulative weaknesses which are эemoved by an experi- test-time', if there has been no major re- ence-fix acquence.
Experience Includes design." This statement is rneither clear, nor design-development testing, production compnlete, In several ways: screening, field use, etc. .- anyway that ex- perlence with the~ product Is obtained. Fix- 1. The concept of MTBF being used