Research On A New Model Of Infrared Non-dispersed Gas Measurement

Single optical path dual wavelength infrared non-dispersed gas measurement technology is most widely used in the filed of infrared spectrum absorption measurement of gas concentration. But the technology is vulnerable to a variety of factors which can make the gas concentration is inaccurate. In order to improve the accuracy of gas concentration. The thesis research the principle, structure and software of single optical path dual wavelength infrared non-dispersed gas measurement system then conclude that it is easy to be affected by temperature, humidity, wind and other factors also has low sensitivity and slow response.According to the existing problems of single optical path dual wavelength infrared non-dispersed gas measurement system. A new calculation model of gas measurement is presented. The model can basically eliminate the influence of environment. This model firstly uses the k-means algorithm and the method of noise points to optimize the samples of RBF neural network. Secondly, the samples is used to train the RBF neural network that is optimized by gradient descent method then establish the model and forecast the environmental impact value. The real gas concentration is the value of removing the impact value. The real gas concentration can respond the change of gas concentration in real time.Through the test and analysis of the new gas measurement system, the results show that the new infrared non-dispersed gas measurement system can achieve the desired objectives. It has high accuracy and response speed and is suitable for gas concentration measurement.