Static Calibration Experiment Research Based On Parallet Six-axis Force Sensor With Heavy-load Capacity

This article is written for the demand of the task of sensor in high precision measurement,static calibration of multi-axis force sensor is studied based on the parallel six-axis force sensor with heavy-load capacity.The research work includes calibration algorithm,optimum of algorithm,model reconstruction and loading method,static calibration for six dimensional force sensor it provides a certain reference value for the study of the static calibration of six-axis force sensor.Firstly,through the static calibration experiment with load in a wide range of the parallel six-axis force sensor with heavy-load capacity,the calibration data is obtained.The calibration data is calculated by three different kinds of algorithms,through the comparision between the non-linear errors obtained from calibration algorithms,the conclusion is got that BP neural network algorithm is better than the other two algorithms,at the same time the correctness of the algorithms is proved by the data test.Secondly,the artificial fish swarm algorithm is introduced detailedly,a BP neural network algorithm based on the artificial fish swarm algorithm is put forward,the initial weights and biases are optimized by the optimizer based on the optimization principle of the artificial fish swarm algorithm,the calibration data is calculated by the optimized BP neural network algorithm and it is shown that the result of the BP neural network algorithm based on the artificial fish swarm algorithm is relatively stable and it is not easy to fall into local extremum.Then,the static calibration experiment with load in a small range of the parallel six-axis force sensor with heavy-load capacity is conducted,the case of the load applied on the force plate fall into three categories: vertical force,parallel force and slant force.According to the reconfiguration theory the optimal measurement models of the sensor are selected through the calculation of the least square method and BP neural network algorithm,and the optimal measurement models of the sensor in the case of three-dimensional force loading of the two static calibration experiments are selected.Finally,unit load,orthogonal multiple load and mixed multiple load are introduced and the corresponding load tables are designed.The 3D model of the sensor is loaded with reference to the load tables through Workbench software according to the virtual calibration method.The data obtained from unit load and mixed multiple load is calculated by the least square method and BP neural network algorithm,and the calibration results of the two kinds of loading methods are tested through the data obtained from orthogonal multiple load.