为了将多元热流体的各因素对注热管柱腐蚀的影响程度定量化，提出了基于"层次分析法（AHP）-BP神经网络法"组合赋权的评价指标权重确定模型。模型首先采用主成分分析法提取腐蚀评价指标，然后根据组合赋权法计算评价指标权重。验证试验表明：组合赋权法与理论研究结果相辅相成。组合赋权法将基于主观判断的层次分析法和基于客观数据的BP神经网络法相结合，综合利用专家自身的知识经验和室内试验数据，提高了腐蚀评价指标权重的准确性和可靠性。

In order to quantify the influence of each element of multi-element heat fluid on the corrosion of tubing, a weight determination model of evaluation index based on "AHP-BP neural network" combination weighting was proposed. In the model, principal component analysis (PCA) was used to extract the corrosion evaluation index, and then the weight of the evaluation index was calculated according to the combination weighting method. The verification test showed that the combination weighting method and the theoretical research result supplement each other. Combination weighting method, combining AHP based on the subjective judgment with BP neural network based on the objective data, can comprehensively use expert's knowledge and experiment data,and eventually improve the accuracy and reliability of the corrosion evaluation index weight.