Wavelet-fused richer convolution feature for multiphase flow visual monitoring of oil and gas production processes

Abstract

Abstract Multiphase flow monitoring of the oil and gas production process is of great significance to the safety of oil and gas exploitation and production. Electrical capacitance tomography (ECT) is one of the most attractive technologies in the field of multiphase flow measurement due to the advantages of its non-radioactive and non-contact properties, good visualization, and low cost. We propose a reconstruction algorithm based on wavelet and richer convolution feature (W-RCF) for solving the problem of artifacts and edge blurring in ECT imaging. In the algorithm, the two-channel source images of Tikhonov regularization and Landweber are simultaneously decomposed by a three-level wavelet. On this basis, an image fusion rule combining Bayesian decision and maximum entropy threshold is established to optimize the wavelet coefficients at each scale. The rule can reduce image artifacts and compensate for the defects in the source images. Afterward, the fused images are input to the RCF network for training and testing, and ECT reconstructed images with higher quality are obtained. Based on the simulation and experimental results, it can be seen that the image reconstruction quality of W-RCF is significantly better than that of the linear back projection, Tikhonov regularization, Landweber, and convolutional neural network algorithms. Therefore, the W-RCF algorithm has higher accuracy and stronger adaptability for multiphase flow under different flow patterns, which provides an effective method of ECT image reconstruction and is more suitable for visual monitoring of multiphase flow in the oil and gas production process.