PROBABILISTIC DERIVATION OF OPTIMAL MAINTENANCE SCHEDULES FOR CYCLICAL EQUIPMENT

• Main Author: Kraut, Willi K
• Format: Report
• Language: English
• Subjects: CATAPULTS; CURVE FITTING; FAILURE(MECHANICS); MAINTENANCE; MARKOV CHAINS; MONTE CARLO METHOD; Operations Research; OPTIMIZATION; PROBABILITY; RELIABILITY; SCHEDULING; STATISTICAL DISTRIBUTIONS; STATISTICAL FUNCTIONS

A theoretical study concerned with the mathematical treatment of maintenance data acquired from the records of the TC13 catapult at NATF(SI) is presented. A mathematical approach is presented for determining equipment maintenance schedules to help reduce equipment malfunctions and resultant downtimes. Methods are derived for determining the optimum number of satisfactory equipment operating cycles after which preventive maintenance is to be performed. The TC13 catapult was chosen to illustrate derived methods. Wartime operations of a one-catapult aircraft carrier were simulated through use of a 1604 Control Data computer. Techniques used included a finite Markov chain to describe reliability, Monte Carlo simulation to estimate transition probabilities of the Markov chain, curve fitting to approximate log normal downtime and launching time probability distribution functions, and simulation of three independent probability distribution functions for 10,000 days of aircraft-carrier operations. Some of the statistical and computer methodologies should be of benefit to those analysts solving large-scale operations-research type problems that involve probability distribution functions of a discrete nature as applied to cyclical-type equipment.