论文部分内容阅读
通过测量大庆地区区域土壤的理化性质以及碳钢的短期腐蚀数据,分析土壤传质过程的逻辑关系,构建了碳钢短期土壤腐蚀预测模型.通过用该模型在BP人工神经网络中进行学习、训练及模拟,并与现场碳钢埋片腐蚀实验结果对比,进一步验证了腐蚀模型的合理性.结果表明:含水量、空气容量、pH、Cl-含量、SO24-含量和可溶盐总量六种土壤环境参数为影响区域土壤中碳钢腐蚀的主要因素;运用基于Matlab平台的人工神经网络,通过不断地积累土壤腐蚀信息,多次训练后可以建立起稳定性好、泛化能力强的土壤腐蚀预测模型,能较好地预测了大庆地区碳钢在土壤中的腐蚀速率.
By measuring the physical and chemical properties of soil in the Daqing region and the short-term corrosion data of carbon steel, the logical relationship between soil mass transfer process and the short-term soil corrosion prediction model of carbon steel was analyzed. By using this model to study and train in BP artificial neural network And simulations, and compared with the corrosion test results of on-site carbon steel buried pieces, the rationality of the corrosion model is further verified.The results show that: the water content, air capacity, pH, Cl- content, SO24- content and the total amount of soluble salts The soil environmental parameters are the main factors affecting the corrosion of carbon steel in the soil. By using artificial neural network based on Matlab platform, soil erosion can be accumulated continuously after many trainings, and soil erosion with good stability and extensive generalization ability can be established The prediction model can well predict the corrosion rate of carbon steel in the soil in Daqing area.