摘要
利用人工神经网络技术对Q235钢在中原油田注水系统中的腐蚀进行了研究.所用网络结构为14-15-1的形式,以试验用水的温度和水质成分等14种环境因素作为网络输入,以Q235钢在试验用水中的平均腐蚀速度作为网络输出.学习算法为反向传播算法.结果表明,利用训练好的网络对Q235钢在试验用水中的腐蚀进行预测,误差较小,方法可行.同时,对影响Q235钢的腐蚀因素进行了研究,找出了影响腐蚀的几种主要因素.
The corrosion of Q235 steel in the oilfield injeted water system and its influencing factors were studied with artificial neural network technique. In the study, a neural network with 14-15-1 structure was used, regarding 14 physic-chemical factors of the water as network input, average corroding rate of Q235 steel as network output. The learning algorithm is BP (Back Propagation) algorithm.The research results show that error of corrosion prediction for Q235 steel in the water by artificial neural network is smaller and this method is suitable for the study of oilfield injected water system corrosion. Furthermore, influencing factors on corrosion of Q235 steel in the oilfield water were discussed. The results indicate that Cl-、HCO3-、O2 and TGB are main factors influencing corrosion of Q235 steel in oilfield water.
出处
《腐蚀与防护》
CAS
1999年第4期151-153,共3页
Corrosion & Protection
关键词
神经网络
Q235钢
水腐蚀
油田
注水系统
Artificial neural network Q235 steel Water corrosion Oilfield injected water system
