Dynamic nonlinear PLS soft sensor modeling method based on Gaussian process regression

Abstract

The invention discloses a dynamic nonlinear PLS soft sensor modeling method based on Gaussian process regression, which can be used for industrial processes with strong time variability, coupling andnonlinearity. Firstly, the dynamic characteristics of industrial data are captured by augmented matrix. Then, Gaussian process regression is used to replace the linear relationship between the input and output score vectors in the partial least square modeling process, so that the partial least square method has the ability of nonlinear modeling. In order to verify the prediction ability of the model, the method is applied to the wastewater treatment simulation benchmark 1 model and the wastewater treatment process of a factory for soft sensor modeling. The experimental results show that the application of dynamic method and Gaussian process regression can significantly improve the prediction ability of partial least square regression model, which is more suitable for soft sensor modelingof complex industrial processes.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

AI Technical Summary

Benefits of technology

This patented technology describes a way to improve models for predictive purposes through utilizing advanced techniques such as machine learning or deep learning algorithms. It involves capturing both static and temporal aspects of complex systems with time series data from various sources like sensors or cameras. By doing things we can create more accurate predictions about these types of situations over longer periods than previously possible due to their high dimensionality compared to traditional methods.

Problems solved by technology

This patented technical problem addressed by this patents relates to improving the precision or efficiency of controller systems that use process measurements like temperature, pressure, flow rate etc., while also addressing issues related to nonlinearity caused by factors like noise and interference from external sources. These challenges require developing efficient and effective techniques based on mathematical relationships involving multiple variables rather than just one single factor.

Images