Research On Fatigue Life Monitoring Of "Sensing-with-Computing" For Mechanical Equipment

With the continuous progress of social productivity,mechanical equipment is developing in the direction of large-scale,systematized,and high-speed.The complexity of its structure and working environment is also increasing.Factors such as heavy load and strong vibration make mechanical equipment work Fatigue failure problems of key weak parts are emerging one after another,which seriously threatens the safe operation of machinery and equipment and the lives and health of workers.In view of the fatigue and aging of mechanical equipment in service,in order to prevent large-scale safety accidents,the fatigue damage monitoring and early warning of in-service structures needs to be solved urgently.However,the relatively complete fatigue life monitoring technology involves the intersection of multiple disciplines,and there are problems such as difficult to monitor load signals,difficult to process online data,and difficult to assess damage status.To this end,the thesis starts with the fatigue life analysis theory of the in-service structure and data-driven load spectrum prediction technology.It focuses on the actual working scene of the equipment,combined with digital technologies such as advanced sensing and big data analysis,and explores real-time load perception,online damage monitoring,and lifespan.Platform-based,integrated,and scenario-based fatigue life monitoring and early warning technology for in-service equipment for precise assessment and dynamic risk early warning.The main work of the thesis is as follows:(1)Fatigue load spectrum advanced adaptive prediction model: In view of the complexity and time-variability of the load height during equipment service,and the sparseness of online monitoring data,the strain time history is obtained through methods such as noise reduction and rain flow counting to obtain the cyclic load spectrum,Using interval division,KS test and nuclear density estimation to transform the prediction of continuous load spectrum into time series prediction of discrete points,and predict the future load spectrum of the structure through NAR neural network combined with particle swarm algorithm and Monte Carlo method.(2)Fatigue life prediction model under mixed loads of high and low cycles: Aiming at the mixed stress of high and low cycles during equipment service,this paper is based on the Manson-Halford two-stage loading model and introduces correction factors to establish the fatigue life prediction model of the structure under multi-stage loads.The construction of the intelligent monitoring system and the development of in-situ monitoring sensors provide a theoretical basis for the following.(3)The design of the intelligent monitoring and early warning system for the fatigue life of the service structure: In response to the real-time damage monitoring requirements,a four-channel strain acquisition instrument is designed,and based on the fatigue load spectrum prediction model and the fatigue life prediction model,the real-time strain acquisition and material model parameters are comprehensively considered Set up,load statistical analysis,remaining life prediction and risk dynamic early warning,etc.,using Labview and Matlab to jointly build a set of human-computer interaction in-service structural fatigue life intelligent monitoring and early warning system.(4)"Integration of sensing,calculation and storage" fatigue life in-situ monitoring and early warning sensor design: In order to reduce the signal dependence of monitoring and reduce the deployment cost,it integrates multiple sub-units such as control transmission,strain acquisition,signal processing and life evaluation,based on embedded Develop in-situ monitoring equipment for structurally weak parts of the system to achieve in-situ monitoring and early warning of structural fatigue life.(5)Verification of intelligent sensing system and theoretical model: In order to verify the feasibility of long-term stable monitoring of intelligent sensing system,this paper uses multi-level block spectrum test of standard samples to monitor the functions of online system and in-situ sensors and long-term The stability of the monitoring was verified separately;in addition,in order to verify the load spectrum advanced adaptive prediction algorithm,the actual load data of a certain engineering TBM cutter-head was compared with the model prediction data to verify the effectiveness and necessity of the algorithm in this paper.