Research Of Temperature Compensation Method For Typical Coal Mine Gas Infrared Sensors

Infrared spectrum is an essential technical method for quantitative analysis of coal mine gases.However,currently widely used portable infrared sensors are highly affected by the change of ambient temperature in the field testing process.In order to improve the analysis accuracy and applicability of infrared sensors,this paper carried out theoretical analysis and experimental research about the necessity of constant temperature heating,reasonable heating temperature range,test data error optimization model and other scientific issues of infrared sensors.Firstly,the test platform was built,which included the temperature compensation before,constant temperature heating test,temperature compensation after and different test environment temperature conditions.The accuracy of CO₂、CH₄、CO infrared sensors without heating were tested under the standard gas concentration of CO₂（0.5%,1%,2.43%）,CH₄（0.1%,1%,4%）,and CO(10×10^-6,100×10^-6,500×10^-6).And the test errors were 1.4%-2.0%,（3000-5000）×10^-6 and（50-160）×10^-6,respectively,which did not meet the error range stipulated by the coal industry standard.Secondly,the thermostatic heating test of 3 gas sensors was carried out at 80%of full scale and ambient temperature from 25℃to 55℃,and the most suitable heating temperature of infrared sensors were determined to be 45℃in terms of gas spectral absorption intensity and absolute error convergence,etc.The accuracy test of CO₂,CH₄and CO standard gas was further carried out under this temperature condition,and it was found that the test errors of the 3 kinds of gases were only 0-0.045%,0-900×10^-6,0-9×10^-6,which is better than the error range specified in the coal industry standard.Thirdly,the variable temperature environment of-20℃～45℃was set in the high-low temperature test chamber to simulate the application of coal mine scenarios,and the temperature compensation verification was carried out on the 3 kinds of gas sensors after heating.The test found that the lower the ambient temperature and the higher the concentration of the gases to be tested,the greater the detection error of the sensors.The maximum test errors of standard gas at full scale concentration were 2.7%,1.9%and 110×10^-6,respectively.Finally,the error analysis optimized models were established based on error back propagation（BP）neural network and genetic algorithms-error back propagation（GA-BP）neural network,the optimized fusion of ambient temperature,test concentration and standard gas concentration with the optimized model showed that the mean absolute error（MAE）values optimized by GA-BP neural network model were reduced by 0.14951%、0.05315%and 5.7214×10^-6,respectively.And the mean square error（MSE）and root mean square error（RMSE）calculation results also verified the stability and feasibility of GA-BP model in gas analysis.This paper verified the influence of temperature on the accurate quantitative analysis of infrared sensors,analyzed the necessity of constant temperature heating of sensors,obtained the appropriate heating range through experiments,respectively carried out the temperature compensation before,temperature compensation after and experimental verification test under variable temperature conditions,and established the coal mine gas infrared sensors temperature compensation model under different ambient temperature conditions.It layed the technical foundation for constructing the in-situ on-line monitoring system of major gas and trace gas in coal mine.There are 49 figures,28 tables and 114 references in this dissertation.