A study on processing method for outliers obtained from CNC gear measuring center

In order to decrease the effects of outliers on measuring accuracy, grey system theory is introduced into outliers processing of CNC gear measuring center. The reason for the effect of the choice of initial condition on the precision of GM(1,1) model is analyzed, and an algorithm in the choice of initial condition which can obtain the global optimum predictive value is proposed. Through an analysis of the properties of the GM(1,1) model, it is found that the GM(1,1) model is suitable to simulate concave sequence, and there is no requirements for monotonicity of the sequence. Then the GM(1,1) model is improved upon the choice of initial condition and concavity of the sequence, after which the improved model is simulated and the simulation result shows that the precision of the improved model increases. The improved GM(1,1) model is applied to outlier processing of CNC gear measuring center, taking measuring method of involute master as an example. Experimental results indicate that outlier identification and revision are realized automatically and the amplitude of outlier is reduced to 25% of the previous. Therefore, the measuring accuracy of CNC gear measuring center is improved.