| Crane is an indispensable part of modern industrial production,by the end of 2020,the number of crane in China is as high as 2,538,400.While crane has greatly improved labor efficiency,safety accidents caused by electrical faults have occurred from time to time,often causing heavy economic losses and even casualties.Social attention on the safety of crane continues to increase,but compared with foreign countries,the research in the field of functional safety in China is still lagging behind.This thesis studies the intelligent evaluation method for the functional safety of crane electrical systems,and the related software is designed.The main work contents are as follows.Several methods for determining safety integrity level are proposed in IEC61508 and this thesis analyzes their shortcomings,and proposes a more comprehensive and intelligent functional safety evaluation scheme.The functional safety assessment of the electrical system of crane is divided into hardware functional safety,system functional safety and risk profile.This method is more detailed and accurate.A quantitative method for determining crane’s risk level based on negative function and dynamic weight is designed.In this thesis,the failure modes and consequences of the crane electrical system are analyzed,and the possible risk sources of the system are judged.The output parameters of the crane are collected by using black box test method,and the score of each parameter is calculated by using negative function.The static weight of each parameter is calculated by AHP,and when a parameter is in an abnormal state,its weight will be adjusted appropriately.According to the score and the weight of each parameter,the result of the whole machine is calculated comprehensively to determine the risk level of the crane.In this thesis,the reliability block diagram method proposed in IEC61508 is decomposed and refined,and the aging compensation algorithm of element is added to make the results more realistic.The electrical system of crane is divided into three subsystems,the redundant voting group structure and series or parallel composition form of each subsystem are analyzed,the PFD of each subsystem is calculated.Considering the different failure rates of cranes at various time,a modified algorithm for failure rate based on Weibull function is designed.The parameters in the Weibull function are estimated with the help of Excel,and the failure rate of crane in the wear stage is modified,which makes the calculation result more realistic.Aiming at the software failure,electrical interference and improper management of crane,the neural network is used to evaluate the system function safety based on the previous data.The prediction accuracy,convergence performance and convenience of PNN neural network,BP neural network and Hopfield neural network are analyzed and compared.Finally,Hopfield neural network is used to realize intelligent evaluation.The software for intelligent assessment has been developed,the main program is written by C++,graphical interface is realized through QT.The software has login module and data management module,it uses SQLite database to store user’s information,and uploads the data through Excel.The software is applied to the actual crane,and the usability and correctness of the software are verified by testing. |
