Fault diagnosis method for oil pumping unit based on adaptive unscented Kalman filtering and RBF (Radial Basis Function) neural network

Abstract

The invention provides a fault diagnosis method for an oil pumping unit based on adaptive unscented Kalman filtering and an RBF (Radial Basis Function) neural network. The method comprises the steps of first, modeling a decision parameter by use of the RBF neural network; then, updating a weight, a center and a width of a hidden layer of a neural network model in real time by use of an unscented Kalman filtering algorithm m to obtain an optimal neural network parameter; and finally, improving the stability of the model by use of an adaptive filtering algorithm, and establishing the fault diagnosis method for the oil pumping unit based on the adaptive unscented Kalman filtering in combination with the RBF neural network. The fault diagnosis method for the oil pumping unit based on the adaptive unscented Kalman filtering and the RBF neural network has the remarkable effects that because the unscented Kalman filtering has real-time updating performance, the nonlinear dynamic modeling of the RBF neural network is implemented; the adaptive filtering algorithm can improve the stability of the model and satisfy the requirement on the model precision under complex environments. With the method, the precision rate for fault diagnosis is improved, and the purpose that a running status of the oil pumping unit is detected in real time is really achieved.

Description

technical field [0001] The invention relates to a pumping unit fault diagnosis technology, in particular to a pumping unit fault diagnosis method based on an adaptive unscented Kalman filter and an RBF neural network. Background technique [0002] The fault diagnosis of pumping units requires scientific and reasonable methods. At present, people mainly judge artificially based on the dynamometer diagram, and can only make qualitative analysis. The diagnosis results are affected by expert experience, technology, etc., and the diagnosis has a certain lag , can not achieve real-time accurate diagnosis. The operation process of the pumping unit has the characteristics of nonlinearity and strong coupling, which brings great difficulties to the fault diagnosis. The RBF neural network has a strong nonlinear mapping ability, which is suitable for solving nonlinear system modeling problems, and provides a new idea for the process modeling of the scheme. The present invention adopts...

AI Technical Summary

Problems solved by technology

[0003] The present application provides a pumping unit fault diagnosis method based on adaptive unscented Kalman filter and RBF neural network to solve the problems caused by failure to detect the pumping unit failure in the prior art when the pumping unit is in operation. Technical problems that miss the best maintenance period

Images