A Nonlinear Integrated Modeling Method of Extended Kalman Filter Based on Adaboost Algorithm

Abstract In the zinc hydrometallurgical purification process, the concentration ratio of zinc ion to trace nickel ion is as high as 10(5), so that the nickel spectral signal is completely covered by high concentration zinc signal, resulting in low sensitivity and nonlinear characteristics of nickel spectral signal. Aiming at the problem that it is difficult to detect nickel in zinc sulfate solution, this paper proposes a nonlinear integrated modeling method of extended Kalman filter based on Adaboost algorithm. First, a non-linear nickel model is established based on nickel standard solution. Second, an extended Kalman filter wavelength optimization method based on correlation coefficient is proposed to select wavelength variables with high signal sensitivity, large amount of information and strong nonlinear correlation. Finally, a nonlinear integrated modeling method based on Adaboost algorithm is proposed, which uses extended Kalman filter as a basic submodel, and realizes the stable detection of trace nickel through the weighted combination of multiple basic models. The results show that the average relative error of this method for detecting nickel is 4.56%, which achieves accurate detection of trace nickel in zinc sulfate solution.