Partial least squares-based Gaussian regression soft measurement modeling method

Abstract

The invention discloses a partial least squares-based Gaussian regression soft measurement modeling method. The method can be applied to industrial processes with relatively strong time-varying characteristic, coupling, nonlinearity, hysteresis and other complex characteristics. The method comprises the following steps of: firstly, carrying out dimensionality reduction on multi-element input dataon the basis of a partial least squares method, and selecting proper score vectors as input of a Gaussian process regression model; secondly, selecting and combining covariance functions, and constructing different types of Gaussian process regression soft measurement models to predict output data; and finally, evaluating prediction ability of the models by using test set data. Modeling results ofpaper-making wastewater treatment process data prove that a partial least squares-based dimensionality reduction technology for measured variables can improve the prediction ability of the Gaussian process regression model; and the Gaussian process regression models constructed by different covariance functions provide multiple options for effluent indexes, so that the method is more suitable forcomplex and changeable paper-making wastewater treatment environment.

Description

technical field [0001] The invention relates to a soft-sensing method for effluent indicators in the papermaking wastewater treatment process, in particular to a Gaussian regression soft-sensing modeling method based on partial least squares. Background technique [0002] In the process of papermaking wastewater treatment, there are a large number of parameters that are difficult to measure or cannot be measured online, and these parameters closely affect the control of effluent indicators, such as chemical oxygen demand (Chemical Oxygen Demand, COD) and suspended solids (Suspended Solids ,SS), these parameters are called leading variables. It is particularly important to detect and control the leading variables timely and accurately. However, the high hardware cost has become the main limitation of online parameter measurement. The soft-sensing method can complete the prediction of the leading variable according to the selection and measurement of the measured variable. Th...

AI Technical Summary

Problems solved by technology

[0006] Now, in the field of papermaking wastewater treatment, the leading variables that affect the effluent indicators can be obtained by upgrading the measuring instruments, but such hardware instruments are either too expensive to be popularized, or have limited accuracy and affect the measurement results; generally speaking, the selection of measured variables Usually, based on process knowledge, the variable most closely related to the leading variable is initially selected, and then through correlation analysis combined with the knowledge of process experts, the measured variable is finally screened out in a relatively appropriate quantity. To analyze the correlation between the measured variable and the leading variable, it is difficult to obtain the appropriate measured variable and accurate prediction effect

Images