Khodakaram Salimifard

Condition monitoring is one of the most important equipment health management techniques and maintenance and repairs (Maintenance) is based on conditions. In the cycle of health management methodology and defect prediction A somewhat more developed form of condition-based maintenance is condition evaluation as It is the most important part of this cycle. In this research, health management and fault prediction are defined for a number of sample equipments of National Gas Company with the help of oil analysis monitoring method with a predictive approach. Using this method can help in better management of assets and more efficient management of maintenance and repairs of the company on the one hand and availability of equipment in the company's operations on the other hand. In this research, using the data-driven forecasting approach and machine learning, with the help of predictive data mining and logistic regression, a model has been presented to predict the health status of turbines. Based on the proposed model, viscosity gauges at temperatures of 40 and 100 and chrome metal have shown the greatest impact on the health status of turbines. Therefore, the results of this research showed that soft computing methods can be reliably used as an alternative in diagnosing turbine health status based on oil analysis results.