Research on Valve Life Prediction Based on PCA-PSO-LSSVM

The valve is a key control component in the oil and gas transportation system, which, due to environment, transmission medium, other factors, susceptible internal leakage, resulting failure. Conventional testing methods cannot judge service life of valves. Therefore, it important carry out prediction research for safety. In this work, method based on PCA-PSO-LSSVM algorithm proposed. main factors affecting are obtained by principal analysis (PCA), least squares support vector machine (LSSVM) used predict life, parameters optimized using particle swarm optimization (PSO), model established. results show that predicted closer actual value, with an average relative error (MRE) 16.57% root mean square (RMSE) 1.2636. Valve accuracy improved, which provides scientific technical maintenance replacement