煤矿作业人员人因可靠性的径向基神经网络评价模型研究

依据人因可靠性原理、事故致因理论，结合煤矿生产系统的特点，提出了观测可靠度的概念，确立了一系列便于统计和赋值的人因可靠性评价指标。针对人因失误事件过程的动态性、复杂性以及数据的不完整性，利用径向基（RBF）神经网络学习速度快、抗干扰能力强的特点，建立了基于RBF神经网络的煤矿作业人员人因可靠性评价模型。对于岗位工龄短或有效记录不足的煤矿作业人员，利用训练好的RBF神经网络模型进行人因可靠性评价，经过实例验算证实了模型的可信性。

Study on evaluation model for human reliability of coalmine workers based on RBF neutral network

Based on the human reliability principle and accident causation theory, combining with the characteristics of coalmine production systems, the concept of observed reliability was put forward, and a series of evaluation indices for human reliability being facilitated to statistics and assign were established. An evaluation model for human reliability of coalmine workers based on RBF neural network was proposed and constructed in light of the dynamic and complicated accident process and data imperfection of human errors, considering about the fast learning speed and strong anti-interference characteristics of RBF neural network. The optimized RBF neural network model was applied to evaluate the human reliability of workers with short job seniority or insufficient job record, and the model was verified to be credible by checking up actual examples.