Method for reducing measurement errors of infrared methane sensor based on improved GWO-SVM

Abstract

The invention provides a method for reducing measurement errors of an infrared methane sensor based on a GWO-SVM. The method comprises the following steps of collecting voltage difference ratio signals output when the infrared methane sensor measures different methane concentrations; optimizing a penalty coefficient C and a kernel function coefficient g of a support vector machine by using an improved grey wolf algorithm to establish an SVM regression prediction model; S3, dividing the collected methane concentration and corresponding voltage difference ratio signals into a training set and atest set, and inputting the training set and the test set as characteristic parameters into the SVM regression prediction model established in the S2 for training and testing to obtain an improved GWO-SVM-based regression model; and inputting the actually measured voltage difference ratio signal data into the regression model based on the improved GWO-SVM to obtain the methane concentration corresponding to the actually measured data. Experimental results show that the support vector machine regression model established by adopting the grey wolf optimization algorithm is small in absolute error and relative error and high in precision.

Description

technical field [0001] The invention relates to the technical field of infrared measurement, in particular to a method for reducing the measurement error of an infrared methane sensor based on an improved GWO-SVM. Background technique [0002] Methane is a non-toxic, flammable and explosive gas that is widely used in industrial production and household life. However, excessive methane concentration will cause various hazards, so methane sensors are very important in places where methane is used. At present, methane sensors based on electrochemical principles, catalytic combustion principles, solid-state principles, and photoionization principles are commonly used in the market. Their detection accuracy is low and their service life is short, which often causes accidents and causes loss of life and property. . [0003] The infrared methane sensor has gradually become the mainstream methane sensor due to its advantages of low power consumption and high precision. The error o...

AI Technical Summary

Problems solved by technology

At present, methane sensors based on electrochemical principles, catalytic combustion principles, solid-state principles, and photoionization principles are commonly used in the market. Their detection accuracy is low and their service life is short, which often causes accidents and causes loss of life and property.

[0003] The infrared methane sensor has gradually become the mainstream methane sensor due to its advantages of low power consumption and high precision. The error of the combined result is high, which has a great impact on the detection accuracy of the infrared methane sensor

Images