露天矿运输的PSO-BP神经网络风险评估法

Transportation Risk Assessment of Open-pit Mine based on PSO-BP Neural Network

为对露天矿运输路线进行风险评估,提出一种基于PSO-BP模型的露天矿运输风险评估方法。以某露天矿内15条运输路线为例,选取17个主要因素作为风险指标体系的二级指标并进行定量化处理,构建了BP风险评估模型,利用PSO算法优化BP模型的权值和阈值,建立了PSO-BP风险评估模型。以模糊综合评判得出的风险分值为依据,分别对PSO-BP模型与BP模型的训练结果和测试结果进行对比分析。结果表明:PSO-BP模型的评估结果与现场实际情况基本相符,评估准确率明显高于BP模型。

In order to assess open-pit mine transportation risk,a new assessment method based on the model of PSO-BP neural net was proposed. Taking the 15 transportation routes of an open-pit mine for example,seventeen major factors were selected as the second-level indexes of risk index system to make quantitative processing. The BP neural net model was built,and its weights and the threshold values were optimized with PSO algorithm. Depended on the risk scores of fuzzy comprehensive evaluation,a comparison of training results and test results between PSO-BP model and BP model was made respectively. Results show that the assessment results of PSO-BP model is greatly consistent with the actual situation and has higher assessment accuracy than that of BP model. The proposed PSO-BP evaluation model.