参数共享双向补偿的风洞六分量天平静态标定OA

To address the issues of increased decoupling errors and mean drift of residuals for wind tunnel six-component balances under combined loading and positive-negative switching of main loads,a parameter-shared bidirectional compensation regression(PSBCR)calibration model is proposed to reduce both the maximum error and systematic bias simultaneously.This method decomposes the calibration model into a main calibration term shared in both positive and negative directions and a compensation term adjusted according to the sign of the main load in the second-order polynomial feature space,and jointly estimates the parameters using regularized regression.Here,single-component data is used to stabilize the main term parameters,while complex loading data is used to identify direction-related differences.Feature standardization and cross-validation are combined to determine the regularization coefficients.Experimental results show that this method reduces the drift in the residual mean from 0.030 0%FS to 0.004 8%FS,the maximum full-scale(FS)error from 0.088 8%FS to 0.067 5%FS,and the average root-mean-square error from 0.045 5%FS to 0.028 5%FS.The results verify that the model can improve decoupling accuracy and mitigate systematic shifts caused by direction switching under complex conditions,This approach is applicable to consistency calibration and data processing of positive and negative loads for wind tunnel six-component balances.