基于神经网络的高含硫气田L360钢腐蚀速率预测

高含硫气田地面集输系统广泛使用L360钢,由于腐蚀因素的多样性及协同效应,其腐蚀速率预测一直是个难题。文章介绍了不同腐蚀因素对L360钢腐蚀速率的影响。随着H2S和CO2压力的增高,腐蚀速率先降后升,在H2S和CO2压力为1.00和0.67 MPa时达到最小值;随Cl-质量浓度的升高,腐蚀速率增大,但当Cl-质量浓度高于40 g/L后,腐蚀速率反而降低;随着温度的升高,腐蚀速率增大,当温度超过70℃后,腐蚀速率反而降低。建立了三层结构BP神经网络模型,输入层有6个神经元,分别代表H2S,CO2分压、Cl-质量浓度、温度、流速和沉积硫6种腐蚀影响因素,隐层神经元数目为8个,输出层神经元数目为1个,代表腐蚀速率。结果表明,L360钢在试验水中的平均腐蚀速率的预测最大误差在15.9%以内,可以满足工程应用要求。

Prediction of L360 Steel Corrosion Rate in High-sulfur Gas Field Using Artificial Neural Network

L360 steel is widely used in gathering and transportation systems of the high-sulfur gas field.It is difficult to predict the corrosion rate due to many corrosion factors and their the synergistic effect.The impacts of different corrosion factors on the corrosion rate of L360 steel are studied.The results show that the corrosion rate of L360 steel falls initially,reaches the minimum at H2S/CO2=1.00/0.67(MPa),then increases gradually with the increase in the partial pressure of H2S/CO2.The corrosion rate increases with concentration of Cl-,and falls when the concentration of Cl-is over 40 g/L.The corrosion rate increases with the temperature and decreases when the temperature is over 70 ℃.A BP network of three layers has been developed.The input layer has six neurons,representing H2S partial pressure,CO2 partial pressure,Cl-concentration,temperature,chloride concentration,velocity and deposited sulfur respectively.The output layer has one neuron,representing corrosion rate.The neuron number of hidden layer is 8.The prediction results have shown that the maximum error is with 15.9%,which satisfies the requirements of the engineering