油水井计量仪表诊断分析技术与平台系统设计

Oil Well Measuring Instrument Diagnostic Analysis Technology and Platform System Design

摘要: 由于目前国内大数据分析技术正在起步阶段，特别是在数据分析法筛选失准(或故障)油水井计量仪表领域还没有实质性研究成果。为此，开展了数据预处理模型、数据挖掘和失准(或故障)定位模型的研究，并根据研究成果完成油水井计量仪表诊断分析平台的建立。通过研究，建立了基于深度自编码学习的数据预处理模型；建立基于神经网络数据分析的油水井计量仪表失准分析模型；开发了基于神经网络分析的油水井计量仪表诊断分析平台。通过数据分析平台对胜利油田油水井计量数据的分析，找出失准或故障传感器，大大减少了传感器校准的工作量，表明应用基于神经网络分析的油水井计量仪表失准分析模型是可行的，具有科学性；据此开发的油水井计量仪表诊断分析平台具有可用性。

Abstract: At present, the domestic big data analysis technology is in its infancy; especially in the field of misa-lignment (or failure) of data analysis method screening oil and water Wells, there are no substan-tive research results. To this end, the research of data preprocessing model, data mining and misa-lignment (or fault) location model is carried out, according to the research results, to complete the establishment of oil and water well metering instrument diagnosis and analysis platform. Through the research, a data preprocessing model based on deep self-coding learning is established. Based on neural network data analysis, the misalignment analysis model of oil and well measuring in-strument is established. A diagnostic and analysis platform for oil and well measuring instruments based on neural network analysis is developed. By analyzing the measurement data of oil and water Wells in Shengli Oilfield on the data analysis platform, the misalignment or fault sensors are found out, which greatly reduces the workload of sensor calibration. It shows that the application of the misalignment analysis model of oil and water Wells measuring instruments based on neural net-work analysis is feasible and scientific. Based on this, the diagnostic analysis platform of oil and well measuring instrument is available.