Concentration soft measurement method for isolation wall rectifying tower based on nonlinear filtering

TLDR: In this article, a concentration soft measurement method for an isolation wall rectifying tower based on nonlinear filtering is proposed. But the method is not suitable for the case of noisy data.

Abstract: The invention relates to a concentration soft measurement method for an isolation wall rectifying tower based on nonlinear filtering. Nonlinear unscented Kalman filtering is adopted to perform error correction on a concentration soft measurement method integrated with a data driving model. When a concentration detector detects once every two hours or more, an error corrector performs error correction by adopting nonlinear unscented Kalman filtering; the error corrector can also adopt improved non-linear unscented Kalman filtering to perform error correction, wherein the improvement lies in that when error correction is carried out on nonlinear unscented Kalman filtering, in the case that a covariance P matrix of a corrected concentration point k-1 at the last moment is not a positive definite matrix, a nearest symmetric positive definite matrix P~ on an F-norm of the covariance P matrix is used for replacing P , and Cholesky decomposition is carried out. Accordingto the invention, the problem that oscillation exists due to the fact that the concentration detection interval is increased when extended Kalman filtering is adopted is solved, and rapid error-freecontrol over the isolation wall rectifying tower is achieved.