Dynamic analysis of nonlinear features for high-precision fault diagnosis of large pumps and compressors in oil and gas fields

The nonlinear and dynamic complexity of the oil and gas engineering system itself leads to an increase in the chances of accidents in the system, and the financial and material losses caused by system failures are incalculable. The study proposes a new fault diagnosis method based on evidence theory and the KPCA algorithm, and analyzes the fault characteristics of large pumps, compressors, and other high-precision equipment under high-speed operation. A set of fault feature vectors reflecting the change of nonlinear characteristics of the system is extracted, and the similarity between the patterns is used to obtain the mass function of each evidence. The simulation experiment results show that the fault diagnosis accuracy, leakage rate and F1 score of the DS-KPCA algorithm in the TE process dataset are 97.56%, 1.98% and 98.99%, respectively, and the method is significantly better than the traditional method in fault detection. Practical application shows that in the production process of oil and gas fields, this rapid fault processing method is characterized by clarity, speed, and accuracy, which significantly improves the efficiency and accuracy of fault processing.