Surface Quality Prediction Model of Crane Boom Based on Deep Learning

Abstract

AbstractWith the development of urbanization, crane as an indispensable machine of construction, its workload is increasing. Crane has a certain load-bearing limit, but in the actual operation process, often because of its actual load-bearing, lifting frequently lead to cracks or even fracture of crane boom (CB). The CB is mainly welded structure truss, which has surface quality problems in the later period of use, that is, excessive use or improper use will cause cracks on the body surface, and the existence of cracks will greatly reduce the fatigue life of the boom. Therefore, based on the depth learning theory, the surface quality of CB is predicted, and the research and design of the prediction model of CB surface quality can provide reference for the use and maintenance of CB. In this paper, the characteristics of deep learning are expounded, and the reasons of using deep learning method in crane jib surface quality prediction (SQP) model are understood. In this paper, the necessity of predicting the surface quality of the CB is understood, the influencing factors of the surface quality of the CB are analyzed, the influencing factors of the cracks on the surface of the CB are described, and the quality prediction model of the CB is established. In this paper, five cranes are selected as the research object to study the actual value and the test value of the surface quality of the crane jib, so as to understand the feasibility of the prediction model of the surface quality of the crane jib. The experimental results show that the prediction model of the surface quality of the CB is still a little insufficient, and the error varies greatly due to the influence of different crane states. In the comparison results of crane jib SQP, the maximum error is 21.6% from crane No.4, and the minimum error is 0.3% from crane No.1.KeywordsCrane boomPrediction modelSurface quality predictionDeep learning