Identifying Significant Health Measurement of Equipment Affecting the Quality of a Continuous Product (Case Study: Unit 2, Parand Gas Turbine Power Plant)

Majorproducers consider quality as a major criterion in decision making.Quality characteristics are affected by maintenance and repair decisions. In this study, a model is developed to determine significant measurements of production equipment affecting the quality of a continuous product to identify which measurements are more critical in terms of quality. Diversity of parameters affecting the quality and the delay until effects on quality come into view, are the main aspects of the issue. Genetic algorithm with a fitness function including prediction accuracy, convergence rate, and number of measurements is developed to obtain optimum set of measurements. Artificial neural networks are also used to evaluate the reliability and validity of the solutions. The proposed model was applied and evaluated by a case study in unit 2, Parand Gas Turbine Power Plant. The results demonstrated the optimum set of measurements which are significantly related to quality characteristic. In addition, the available data demonstrating that the terminal equipment in production process has more significant effects on quality. The proposed model enjoys the capability of identifying the most important health measurements affecting the output quality of a continuous product in some limited steps of optimization algorithm by processing the history data from Condition Monitoring Process. With these significant measurements available, the decision makings in maintenance and repair can happen on the grounds of quality.