【摘要】 针对注水管道中腐蚀速率和腐蚀影响因素之间复杂的映射关系 ,提出了利用灰色理论对腐蚀速率进行有效预测 ,同时为提高预测精度 ,对标准的GM (1,1)模型进行了合理改进 ,提出了 4种改进方法 :改进背景值、考虑初始点影响、灰色理论和BP神经网络相结合 (简称灰色神经网络 )以及灰色理论和遗传算法 (简称灰色遗传算法 )相结合等。通过示例表明 ,经过改进后的 4种方法预测精度都有所提高 ,特别是灰色理论和神经网络结合、灰色理论和遗传算法相结合预测得到的腐蚀速率和实测值能较好吻合 ,预测精度最高 ;因此可以运用这两种改进方法较准确地预测腐蚀速率随着时间的变化趋势更多还原

【Abstract】 According to the complicated corrosion rate reflection relation with influence factors, gray model is put forward to predict corrosion rate. At the same time, in order to increase prediction precision, four reasonable modification methods are brought out, namely, modifying back value, modifying initialization values, gray neural network and gray genetic algorithm. Examples are used to verify that modified four methods can increase prediction precision. Gray neural network and gray genetic algorithm methods are in good agreement with field data and can be used to predict corrosion rate variation with time in injecting pipeline.更多还原