Research On Temperature And Humidity Compensation Method Of Gas Sensor Based On Neural Network

The detection of toxic and harmful gases is widely used in the fields of environmental protection,petrochemical industry,coal mine gas,earthquake and other natural disasters.These toxic and harmful gases not only affect air quality,but also seriously endanger human life and health.Therefore,the detection of toxic gases has become very important.However,due to the limitation of the sensitivity mechanism of various gas sensors,the change of temperature and humidity in the measured working condition environment has a serious impact on the detection accuracy and accuracy of toxic and harmful gas concentration.In this paper,an improved neural network data fusion algorithm is designed and studied to realize the temperature and humidity compensation of toxic and harmful gas concentration,and realize the high-precision and accurate detection of toxic and harmful gas concentration under complex working conditions.Based on the comprehensive investigation of the research status and development trend of gas sensor detection mechanism and gas sensor temperature and humidity compensation methods at home and abroad,this paper analyzes in detail the structure,working principle and the influence mechanism of ambient temperature and humidity on the measurement error of three typical toxic gas sensors（electrochemical sensor,infrared gas sensor and PID sensor）.Finally,three typical gas sensors are selected to form the gas detection system.On this basis,the test and verification platform of temperature and humidity compensation method is constructed.Aiming at the influence mechanism of gas sensor measurement error in different temperature and humidity environment,this paper uses the improved neural network algorithm as a way to solve the problem to realize temperature and humidity compensation.Neural network algorithm can compensate temperature and humidity of gas sensor,but the algorithm is easy to fall into local extreme value and convergence speed is slow.To solve this problem,particle swarm optimization algorithm is introduced to optimize the parameters of neural network,but premature convergence is easy to occur.Finally,genetic algorithm and particle swarm optimization algorithm are introduced to optimize the neural network,which overcomes the slow speed of genetic algorithm and makes up for the premature convergence of particle swarm optimization algorithm,and provides an effective method for temperature and humidity compensation of gas sensor.On the test and verification platform of toxic gas temperature and humidity compensation method,the whole system was built and temperature and humidity experiment was carried out first,which provided data support for the test and verification of compensation method.Then BP neural network,particle swarm optimization neural network（PSO）,genetic algorithm and particle swarm optimization neural network（GA-PSO-BP）were used to compensate the temperature and humidity of four kinds of gas CO,SO₂,CH₄ and volatile organic compounds（VOC）,and calculate the error between the compensation value and the actual measured concentration.The experimental results show that the optimized neural network（GA-PSO-BP）combined with genetic algorithm and particle swarm optimization has the best temperature and humidity compensation effect on the gas sensor,and the average error of the gas sensor detection system is 1.64%,which verifies the advanced performance of the compensation method.