Intelligent diagnosis technology of oil well working condition based on deep residual network

In recent years, with the development of oilfield digitization and intelligence, the diagnosis technology for oil well conditions has been gradually studied in depth. At present, traditional methods are used to manually extract features, and with the help of experience and prior knowledge, it can no longer meet the screening, analysis, selection and judgment of a large amount of data information. For this reason, on the basis of deep learning, a method of offline training and online diagnosis of oil well conditions based on deep residual network is proposed. It can directly input the original load vector and displacement vector as the model input, through the pooling layer, the residual module, and the classification layers are connected to each other to further strengthen the learning ability of load and displacement vectors. Experiments have proved that the model can reach 99.05% of the oil well working condition fault diagnosis accuracy, realizes a good fault classification, and has a stable prediction effect and practical performance.