基于BP神经网络的镍基合金腐蚀性能的预测分析

The prediction for Nimonic 80A corrosion properties based on BP neural networks

利用三层BP神经网络,根据已有的镍基合金(Nimonic 80A合金镍基沉淀硬化)在不同重油杂质环境参数下的腐蚀速度数据,建立了燃气轮机的涡轮部件在重油杂质环境中腐蚀失重的人工神经网络模型,并进一步预测了腐蚀失重.预测结果表明:在重油杂质条件下,温度和有机硫化物含量越高,腐蚀损失越大.预测结果能正确地反映出环境参数对涡轮部件腐蚀失重的影响.

A three layers BP neural network model for the corrosion weight - loss of the gas turbine wheel part in heavy oil impurity environments was set up by using Nimonic 80A corrosion weight -loss data in the different heavy oil impurity environments, and the influence law of the gas turbine wheel part corrosion weight -loss was predicted. The results show that, under the heavy oil impurity condition, the higher the tempera- ture and the organic sulfide content are, the larger the corrosion loses is. These prediction results can correctly reflect the influence of environment parameter on the turbine wheel part corrosion weightlessness.