基于BP神经网络的矿山安全文化评价体系

Mine Safety Culture Evaluation System Based on BP Neural Network

为了完善和加强矿山安全文化的建设,根据矿山自身的特点,本文提出了一种对矿山安全文化进行合理、科学、全面评价的方法。根据国际SMART准则,从安全环境、安全管理、安全培训、班组长素质、员工行为、安全事务6个方面确立了21项安全文化评价指标,提出了安全文化等级划分标准,并在MATLAB平台上建立了基于BP神经网络的安全文化等级评价体系。以6个矿山的样本数据作为训练样本,经过训练,结果显示当迭代次数为355时,网络的全局误差达到要求,并且输出值与期望输出值之间呈基本吻合,误差比较小。最后,以该神经网络对恒祥煤矿的安全文化等级进行预测,输出结果表明该矿山的安全文化等级为4级,与该矿山的安全文化实际状况相符。因此,BP神经网络具有很好的学习功能,通过样本训练后能够准确地对矿山安全文化进行评价,为提高矿山安全文化水平提供参考和依据。

In order to improve and strengthen the construction of mine safety culture, according to the characteristics of mine, this paper puts forward a reasonable, scientific and comprehensive evaluation method for mine safety culture. According to the international SMART standards, from the 6 aspects： security environment, safety management, safety training, class leader quality, employee behavior, security affairs, establishing 21 index of safety culture evaluation, the grade division standard of safety culture is put forward, and safety culture assessment system was established based on BP neural network in the MATLAB platform. With six mines of sample data as the training sample, after training results show that when the number of iterations is 355 which meet the requirements of the network of the global error, and the output value and the expected output value was basically consistent error is relatively small. Finally, with the neural network predicting Heng Xiang mine safety culture level , the output indicates the mine safety culture level is level 4 and consistent with the actual situation of the mine＇ s safety culture. Therefore, BP neural network has a good learning function. Through the sample training, the mine safety culture can be evaluated accurately which can provide reference and basis for improving the safety culture level of mine.