-

In order to reduce the costs and the down-time of industrial machinery, a vital step for the prevention of malfunction is the maintenance inspection. On the basis of the statistical calculations
and matching the rate of failure ,down-time and inspection time with the function of the exponential distribution, action is taken to minimize the aggregates of down-time periods through
frequency of inspection (n).


(o) shows the rate of failure without having conducted any type of maintenance inspection.. Regarding the importance and sensitivity of some industries, it is essential to determine
frequencies of inspection with a certain reliability secured through the relation in this formula on the basis of statistical calculation.
The alternative methods, on the basis of the artificial neural
networks, are: 1) pattern recognition, and 2) optimization.
The usage of MLP capability signifies the learning of the
system behavior and generalization of the past function towards
future in the course of machinery normal life and timings of


maintenance inspection based on such function. The object function of this network is in which is the number of actual or inspections by a statistical method in time unit and
(n) is the number of calculable inspections through neural network whereby J 1 could be minimized.
Attempt has been made in this article to use the optimization capability of the neural network so as to minimize the aggregates of machinery down-time aided by the object function. In the final analysis, minimization of the total amount of down-time in conjunction with the maximization of the reliability in respect of minimizing the object function through the neural network is of great significance.
The acquisition of tangible and overall results in the form of optimum integration of down-time (which has an opposite relation with the number of inspections) and reliability (which has a direct relation with the number of inspections), have manifested the relative cutting edge of the new method.