基于加速度分量测量的水下网衣破损定位方法OA北大核心CSTPCD

In order to obtain accurate identification and real-time monitoring of mesh damage location,a new modeling method for mesh damage location based on acceleration measurement was proposed by combining physical test and numerical simulation.The method increases the data difference between the monitoring points in the broken area and the non-broken area,so as to determine the location of the broken fishing net.The fishing net physical tests and numerical simulation calculations were combined to study the fishing net breakage localization.Firstly,we carry out the physical model test of mesh coat damage,validate the numerical model of mesh coat,and use the validated numerical model of mesh coat to deeply explore the influencing factors affecting the modeling of mesh coat damage location detection,which serves as the basis of the mesh coat damage location detection model proposed in this study.The damage localization model includes four monitoring areas,and eight damage conditions are listed to test the localization detection ability of the model.The model uses artificial neural networks to localize and identify the damages.The Y-direction acceleration data collected from the monitoring points in the four regions were processed as inputs,and the eight types of damage and one intact case were used as outputs.The training data are the data obtained from numerical simulation and the test data are the sensing data from the physical model.(1)The results show that under the effect of regular waves,when the center of the mesh line is selected as the monitoring point,the data of the monitoring points in the broken and non-broken areas almost overlap,and the difference is within 1%.When the mesh node of the mesh coat is used as the monitoring point,the data change trends of the monitoring points in the two monitoring regions are different,and the data peak and change range of the monitoring point in the broken region at 210-240 s are three times of that of the monitoring point in the non-broken region.(2)Y-direction acceleration component comparing the acceleration,in addition to this broken when the data difference changes the same,with the increase of the degree of broken,the acceleration data changes in the intact state in the range of 0.75-1.2 times,while the Y-direction acceleration changes in the range of 0.75-23 times,and comparing the other direction of the acceleration component,the intact region and the broken region of the data difference is greater.(3)The accuracy of the model in locating the breakage in the training set is 98.5%and the test accuracy is 97.9%,which can locate the breakage accurately.In this study,we proposed a modeling method for detecting breakage of mesh garments based on region division,proved that using mesh nodes as monitoring points can increase the difference between monitoring points that are in the intact and broken regions,and the selection of acceleration component as an indicator parameter can increase the ability of the monitoring points to perceive the breakage.The results of this study provide some guidance for the modeling of digital twin of net clothing,and at the same time provide theoretical basis for the sensor arrangement and type selection of fishing nets.