Intelligent diagnosis of petroleum equipment faults using a deep hybrid model

摘要

Performance assessment and timely failure detection of the electric submersible pump can reduce operation costs andmaintenance in the oil and gas field. Features of equipment malfunction are changes in vibration signals. Evaluationof vibrations based on accelerometer sensors can detect failures and allows assessment of system failures. This paperproposes a reliable deep learning-based method for electric submersible pump faults detection. The frequency, timeand spectral information of the vibrational signal are considered as input to the deep hybrid model. The spectral informationincludes the spectrogram obtained using the short-time Fourier transform and the scalogram as a result of thecontinuous wavelet transform and provides a detailed study of the vibration signal. The proposed approach is comparedwith k-nearest neighbors, support vector machines, logistic regression, and random forest. The experimental evaluationshows that the proposed deep hybrid model is superior to these machine learning methods, and can automaticallyand simultaneously detect failures of the electric submersible pump according to the vibration signal that is generatedduring system operation. The proposed approach gives good results and can help an expert in automatic diagnostics ofequipment and several complex technical systems.