Artificial neural network based fast pattern recognition of ESP ammeter card

Ammeter card diagnosis is one typical means to diagnose the working conditions of electrical submersible pump (ESP). The tradition pattern recognition of ammeter card needs artificial operation and has technical barriers, so subjective error can be introduced. As a kind of machine learning algorithm, artificial neural network can make up for the error of artificial recognition. In order to realize fast, accurate and objective pattern recognition, this paper established the artificial neural network model based on the corresponding relationship between the actual working condition and the current data characteristic value obtained after the data pretreatment of the collected ammeter cards. Then, this artificial neural network model was applied to working condition diagnosis. It is indicated that the artificial neural network model is absolutely more advantageous than the traditional artificial recognition of ammeter card. An artificial neural network model was established by means of above mentioned method and applied to pattern recognition. The working condition diagnosis model was verified by extracting the ammeter card data of untrained wells, which indicates higher accuracy. The research results indicate the feasibility and reliability of artificial neural network to fast pattern recognition of ammeter card and working condition diagnosis.