论文部分内容阅读
基于ＢＰ神经网络技术具有较强的收敛性及自适应、自组织学习能力、较好的容错性，并行处理强、识别预测迅速准确、稳健性好的特点，以高含硫油井在含水２．４％～１９．０％之间的实际硫化物应力腐蚀（ＳＳＣ）速率作为训练样本，应用ＢＰ网络进行训练，达到精度要求后，对原样本进行回判模拟，再对只知输入信息而输出信息未知的样本进行预测。证明ＢＰ神经网络技术能够正确地预测高含硫油井的ＳＳＣ，且精度高于ＧＭ（１，１）预测结果。其预测结果可用来指导油田的开发生产。
BP neural network based on BP neural network technology has strong convergence and self-adaptability, self-organizing learning ability, better fault tolerance, strong parallel processing, rapid and accurate identification and prediction, and good robustness. Actual SSC rate between 4% and 19.0% is used as a training sample, which is trained by BP network. After reaching the precision requirement, the original sample is subjected to backtracking simulation and then output to the input information only Samples with unknown information are predicted. It is proved that BP neural network can correctly predict the SSC of high sulfur oil wells and the accuracy is higher than that predicted by GM (1,1). The forecast results can be used to guide the development and production of oilfields.