Στατιστικές τεχνικές αποδοχής εξαρτημάτων μέσω σύνθετων τεστ εκκίνησης
Statistical methods for units testing through composite strat-up demonstration tests
View/ Open
Subject
Markov processes ; Bernoulli shifts ; Στατιστική ανάλυσηAbstract
In a start-up demonstration test we wish to decide if an equipment is qualitatively acceptable; we test a randomly selected sample of similar components, which are subjected to continuous start-ups and based on a specified criterion we determine whether the sample is acceptable or not. The start-ups can be modeled either by independent and identically distributed Bernoulli trials or by dependent trials (for example, assuming Markovian dependence between trials). The criterion for the decision of the acceptance or rejection of the sample of equipment which is used each time is the feature that differentiates the start-up demonstration test. The aim of each test is to arrive at the correct decision (acceptance or rejection) of the equipment by carrying out as few tests as possible, so as to save time and cost. Applications of start-up demonstration tests include the testing of electric generators, lawn mowers, car batteries, pumps, outboard motors, chain saws, snow blowers and other power operation equipment, computers, electrical appliances (e.g. TV, VCR) etc. The aim of the present dissertation is the presentation and study of the basic processes that have been proposed in the literature for the acceptance or rejection of equipment through start-up demonstration tests. The methods of inference of the parameters in case that we have statistical data from a start-up procedure will also be reviewed in some detail. Special attention will be drawn to the presentation of complicate start-up demonstration tests in which the acceptance or rejection are used to multiple rules based on the number or on consecutive successful or/and failed tests.