GPR modeling based on kernel slow feature analysis and time delay estimation

Abstract

The present invention discloses a GPR modeling method based on kernel slow feature analysis and time delay estimation, which is applied to chemical processes with time delay and non-linearity. The method is characterized in that: by using fuzzy curve analysis, the delay information in industrial data is fully tapped to obtain the optimal delay in the data, and the reconstruction of the modeling data is carried out; the kernel slow feature analysis method is further used to carry out nonlinear feature extraction on the reconstructed data; and finally, based on the extracted features, a Gaussian process regression model is established to realize accurate prediction of the key variables, so as to improve product quality and reduce production costs.

Description

technical field [0001] The invention relates to GPR modeling of kernel-slow feature analysis and time-lag estimation, and belongs to the fields of complex industrial process modeling and soft measurement. Background technique [0002] In real industrial processes, the measurement of some key variables is crucial to produce high-quality products. However, under the constraints of existing technical conditions and economic costs, it is very difficult to directly obtain key variables. [0003] Based on this background, soft sensing technology came into being. It infers and estimates difficult key variables by constructing the functional relationship between process measurable variables and key variables, so it has been widely used. Common soft sensor modeling methods such as partial least squares, neural network, least squares support vector machine, etc. can get good prediction results. Gaussian process regression (GPR) has been widely used in soft sensor modeling in recent...

AI Technical Summary

Problems solved by technology

Due to the complexity of the actual industrial process technology, it becomes impossible to estimate the time delay through the process technology

Zhang and Komulaine estimate the time delay by constructing the correlation coefficient between input variables and output variables, but because this method only considers the linear relationship between variables, it may get wrong conclusions for nonlinear cases

Ruan Hongmei et al. used the difference estimation (DE) algorithm to optimize the joint mutual information between process variables to determine the time delay, but the intelligent optimization algorithm is easy to fall into a local optimum, and the computational complexity of joint mutual information analysis correlation is high, and the running time longer

Images