Fault Diagnosis Of Pumping System Under The Background Of Big Data

In the context of the era of big data,the intelligentization of digital oilfield fault diagnosis is of great significance to the processing of massive data detected on-site in the oilfield.Based on the background of big data,this paper optimizes the network parameters in the RBF neural network to improve the accuracy of fault recognition and speed up the recognition speed;and use this as the core of the algorithm to develop an intelligent pumping unit fault diagnosis system.First,this article takes the collected indicator diagram data as the research object,and preprocesses the grayscale correction and image segmentation on the indicator diagram before feature extraction.The gray-scale matrix-based feature extraction method is used to calculate the gray-scale matrix using the breadth search algorithm and the depth search algorithm,and the matrix statistical features are used as the image feature data.The extraction method has a small amount of calculation,and has a good performance on the invariance of image rotation,scale transformation,and translation.Second,use neural networks for big data processing.The PSO algorithm is used to optimize the connection weight and width in the RBF neural network,and considering the central value problem in the RBF neural network,the FCM algorithm is used to optimize it.The acceleration factor ? of the adopted PSO algorithm uses the dynamic descent method,and the FCM algorithm optimizes the initial value sensitive problem by adding a penalty function.The optimized RBF neural network is compared with the traditional RBF neural network and PSORBF neural network.The experimental results show that the error rate of the optimized network is lower than that of the two networks,and the accuracy rate is higher than 19% and 5%,respectively.Finally,the optimized RBF neural network is used as the core of the algorithm to build a fault diagnosis system for the pumping unit.Through the example verification,the system can quickly and accurately identify the failure of the pumping unit,and give corresponding treatment suggestions.This paper enriches the extraction methods of pumping unit indicator diagrams,and applies the improved FCM algorithm and the improved PSO algorithm to the RBF neural network in the context of big data,which enriches the existing diagnosis methods.