Research on Intelligent Diagnosis of Oil Pumping Well Based on Optimized BP Neural Network

It achieve remarkable results that classify and diagnose the dynamometer card of pumping well with the help of artificial neural network. Aiming at the training defects of the traditional BP neural network, it adds momentum term, changes learning parameters, optimizes the parameters of neural network, speeds up learning and training, avoids falling into local minimum, shortens diagnostic training time, improves calculation accuracy, without changing the overall structure. Calculate separately the five pumping typical fault at the end of this paper. It is basically in agreement that identification results and actual test results of fault type.