Recently, Zhang Kai's team has made significant progress in the field of intelligent wellbore. Their research results have been published in the journal Petroleum Science, in a paper titled "The real-time dynamic liquid level calculation method of the sucker rod well based on multi-view features fusion".
Innovation：We propose using a multi branch neural network for multi view fusion to monitor the dynamic liquid level of oil wells in real time. The core idea of the method is to design a special feature fusion layer through binomial distribution sampling. It can not only explore the correlation between different perspectives that affect the depth of oil well dynamic liquid level, but also avoid excessive interaction between different perspectives. As a more advanced regularization operation, it further prevents model overfitting. In addition, due to the particularity of this sampling method, when a sensor failure or incomplete recording in the oilfield results in missing data from a certain perspective, the missing perspective data will not participate in the dynamic liquid level calculation, nor will it affect the continued participation of other complete perspective data in the dynamic liquid level calculation. Therefore, this method can not only improve the real-time monitoring accuracy of dynamic liquid level, but also increase the utilization rate of oilfield data, reducing the risk of the real-time monitoring function of dynamic liquid level not being able to start when data is missing.
Abstract：In the production of the sucker rod well, the dynamic liquid level is important for the production efficiency and safety in the lifting process. It is influenced by multi-source data which need to be combined for the dynamic liquid level real-time calculation. In this paper, the multi-source data are regarded as the different views including the load of the sucker rod and liquid in the wellbore, the image of the dynamometer card and production dynamics parameters. These views can be fused by the multi-branch neural network with special fusion layer. With this method, the features of different views can be extracted by considering the difference of the modality and physical meaning between them. Then, the extraction results which are selected by multinomial sampling can be the input of the fusion layer. During the fusion process, the availability under different views determines whether the views are fused in the fusion layer or not. In this way, not only the correlation between the views can be considered, but also the missing data can be processed automatically. The results have shown that the load and production features fusion (the method proposed in this paper) performs best with the lowest mean absolute error (MAE) 39.63 m, followed by the features concatenation with MAE 42.47 m. They both performed better than only a single view and the lower MAE of the features fusion indicates that its generalization ability is stronger. In contrast, the image feature as a single view contributes little to the accuracy improvement after fused with other views with the highest MAE. When there is data missing in some view, compared with the features concatenation, the multi-view features fusion will not result in the unavailability of a large number of samples. When the missing rate is 10%, 30%, 50% and 80%, the method proposed in this paper can reduce MAE by 5.8, 7, 9.3 and 20.3 m respectively. In general, the multi-view features fusion method proposed in this paper can improve the accuracy obviously and process the missing data effectively, which helps provide technical support for real-time monitoring of the dynamic liquid level in oil fields.
Petroleum Science publishes high-level original research papers and review papers in the field of petroleum science both domestically and internationally. The impact factor for 2023 is 5.6. The JCR zone for 2023 will be Q1 zone, and the engineering technology category of the Chinese Academy of Sciences will be Zone 1 in 2023.
Paper link:
https://doi.org/10.1016/j.petsci.2024.05.005
Citation:
Chengzhe Yin, Kai Zhang*, Jiayuan Liu, Xinyan Wang, Min Li, Liming Zhang, Wensheng Zhou. The Real-Time Dynamic Liquid Level Calculation Method of the Sucker Rod Well Based on Multi-View Features Fusion[J]. Petroleum Science, 2024, 21(5): 3575-3586. https://doi.org/10.1016/j.petsci.2024.05.005
