Fault Diagnosis of Manufacturing Processes via Genetic Algorithm Approach

Instantaneous detection and diagnosis of various faults and break-downs in industrial processes is required to reduce production losses and damage to equipments. A solved knowledge-based system and developed techniques within artificial intelligence principles are showing its capability in assisting process operators to detect and diagnose faults and anomalies in industrial processes, especially in chosen machine tools environment. In this contribution, there is investigated a self-learning of diagnostic rules through genetic algorithms approach. A success of a genetic algorithm is crucially connected with the adaptive plan which controls „mutation“ and „crossover“ actions, and hence the production of the new solutions. The maximum entropy principle with minimum crossentropy updating, provides a way of making assumptions about the missing specification that minimises the additional information assumed. Solved genetic algorithm is also a stochastic computational model that seeks the optimal solution to an objective function.