神经网络用于油田地面集输管道结垢预测

THE SCALE PREDICTION OF SURFACE GATHERING PIPELINE IN OIL FIELD BASED ON ARTIFICIAL NEURAL NETWORK

利用典型的误差反传神经网络理论，对油田地面集输管道结垢进行预测和评判，避开了各种因素对其结垢影响规律的难题，准确地预测和评判地面集输管道的结垢情况。应用人工神经网络分析某油田地面集输系统管道的结垢情况后表明，人工神经网络无需建立数学模型，学习过程通过自动调节神经元之间的连接权值完成，在选取有代表性的训练样本情况下，人工神经网络能够成功地预测和评判地面集输管道的结垢情况。

According to the theory of artificial neural network, prediction and judgement the scale of the gathering pipeline in oil field can avoid the problem of various factors on scaling and the scale of the gathering pipeline can be predicted and judged correctly and easily. After application of this theory in analysis of a pipeline scale statue in one oil field, it is shown that the artificial neural network needn't to establish mathematical model, instead, it adjusts the connected weight values automatically. Under the condition to get the representative training sample, the artificial neural network can successfully predict and judge the scale statue of the surface gathering pipeline in oil field.