煤矿监控调度作业疲劳程度分级与判定研究

Research on classification and determination method of work fatigue level of coal mine dispatchers

为了能够准确地判断出煤矿监控调度员的疲劳程度,降低煤矿监控调度作业的失误率;运用眼动追踪技术进行煤矿监控调度模拟实验,采集作业者眼动数据与主客观疲劳判定值;利用K-means聚类算法划分疲劳等级数,训练神经网络搭建煤矿监控调度作业疲劳程度预测模型。结果表明:最佳疲劳等级数划分为3类,神经网络预测模型拟合度为90.58%。用预测模型对山西某煤矿监控作业模式进行测试,模型实地预测平均误差为6.26%,预测效果较好。

In order to accurately determine the level of fatigue of coal mine monitoring and dispatching operators and to reduce the error rate of coal mine monitoring and dispatching operations,a simulation experiment of coal mine monitoring and dispatching was carried out using eye tracking technology to collect the operator’s eye movement data and objective and subjective fatigue determination values.The K-means clustering algorithm was used to classify fatigue levels and train a neural network to build a fatigue prediction model for coal mine monitoring and dispatching operations.The results showed that the best fatigue level was classified into three categories and the fitting degree of the neural network prediction model was 90.58%.The prediction model was tested on a coal mine monitoring operation model in Shanxi Province,and the average error of the model field prediction was 6.26%,which was a good prediction effect.