基于RBF的起重作业岗位人因可靠性预测

为提高起重作业可靠性,防止人因失误酿成事故,针对人因失误的随机性、模糊性和不确定性特点,提出运用具有非线性映射能力和容错能力的径向基函数(RBF)神经网络,分析人因失误非线性动力学过程。以起重机操作岗位作为人因可靠性分析(HRA)实例,首先,建立基于"作业人员、交流界面、作业环境、作业特性、作业组织"的人因可靠性预测指标体系,并对指标进行量化;其次,根据人因可靠性原理,统计出人因失误次数,给出人因失误率;最后,通过对"人的疲劳和情绪、交流通道、作业复杂程度和时间裕度、照明环境和风力影响、工作强度和安全监管"等因素的分析,构建基于RBF的起重机操作岗位人因可靠性预测分析神经网络模型。分析结果表明,RBF预测分析同时包含人的操作可靠性与认知可靠性,预测结果同现场实际观测结果的符合度达到92.0%。

Human Reliability Prediction of Lifting Operation Based on RBF Neural Network

In order to improve the reliability of lifting operation,and prevent accident caused by human errors,bearing randomness,fuzziness and uncertainty of human error in mind,an RBF neural network-based method for analyzing the human error's nonlinear dynamics process was put forwarded.Taking the lifting operation as example,firstly,an indexes system about the human reliability prediction was constructed,which included the factors of the operator,the communion interface,the operating circumstance,the operating characteristics and the operating organization.Then the indexes were quantified.Secondly,according to human reliability analysis(HRA) theory and the scene record,the human error data were calculated out,and the human error rates were given.Finally,basing on the analysis of the operator's tiredness and emotion,information channels,operation complexity and time margin,lighting and wind power conditions,working pressure and safety supervision,an RBF neural network-based model for lifting operation human reliability was built.The results indicate that the RBF prediction includes the operation reliability as well as the cognitive reliability,and that the predictions results conform with the actual observed values up to 92.0%.