Research On Big Data Mining Method For Crane Safety Evaluation

During the long-term service of the crane,various faults or damages often occur due to its internal or external factors,resulting in huge casualties and property losses.Effectively analyze and evaluate the health state of the crane,providing an important support for timely identifying the abnormities and ensuring the safe,reliable and economic operation of the crane.Big data is an important part of the new generation of information technology and has great potential in the field of fault identification and safety evaluation.In this paper,the big data mining method in crane safety evaluation is studied to mine the potential value of the data and realize the intelligent and rapid warning of the crane.The main work of the paper is as follows:(1)The source and classification of big data for crane safety evaluation are introduced.And the big data system of crane is established by analyzing the characteristics and status of the data.Based on the big data of crane safety evaluation,the data preprocessing technologies such as repeated data detection,missing data filling,noise and abnormal data processing are discussed and realized,laying the foundation for subsequent association rule mining and intelligent warning analysis.(2)An improved multi-support association rule mining method based on MIS-Tree is proposed,and the superiority of the algorithm is verified.Using actual case data,the association rule mining of the health monitoring data in a quayside bridge is carried out,and the correlation between vibration and stress data is studied to obtain some interesting rules.The rules are applied to the prediction of crane monitoring data,and the prediction error is decreased to 0.11%-8.81%,which verified the validity and accuracy of the prediction based on the association rule model.(3)A feature extraction method based on AR model's autoregressive parameter is proposed,and Principal Component Analysis(PCA)is used to reduce the dimension of extracted initial feature samples.Aiming at the deficiency of Random Forest(RF),an integrated classification method based on GA-RF is proposed by using the Genetic Algorithm(GA)to optimize the parameters and select the characteristics.And GA-RF model and Random Forest are respectively used to identify the safety state of a crane gearbox bearing,and four indicators including accuracy,precision,recall andF-measure are used for evaluation.The experimental results show that the GA-RF model has better predictive recognition effect and stronger generalization ability.(4)Detailed overall scheme and the demand of functional module are designed,and the functional architecture of each module is determined.According to the basic requirements of the system,the specific data structure and mutual relations are designed and the database of security and early warning system is constructed.An intelligent and early warning system of crane is designed and developed.The research results of this paper provide a new way for crane safety evaluation and intelligent warning,and lay a solid foundation for the application of big data technology in the crane industry.