【摘要】 基于BP神经网络技术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 ,对高含硫油井含水 2 .4 %～ 1 9.0 %的实际硫化物应力腐蚀 (SSCC)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未知输出信息的样本进行预测。证明BP神经网络技术能够正确地预测高含硫油井的SSCC ,且精度高于GM (1 ,1 )的预测结果 ,其预测结果用来指导油田的开发生产。更多还原

【Abstract】 With BP nerve network technique having the following advantages,stronger astringency,selfadaptability and organized faculty,good permissible errow,strong parrel processing ability and identification quickly and accurately,the technique is used to train the samples,the rate of sulfide stress corrosion (SSCC) in which the water cut is 2.4% 19.0% in high sulfur bearing oil wells.The original sample will be regressively simulated after the requirement value of accurcy is reached,then the sample,in which the input information is known,but not the output,will be predicted.The result identifies the BP nerve network can predict the SSCC in high surfur bearing oil wells,the accuracy is higher than the predicted result made by GM (1,1).更多还原