Application of BP Neural Network Based on Genetic Algorithm in Quantitative Analysis of Mixed GAS

Aiming at the problem of mixed gas detection in neural network and analysis on the principle of gas detection. Combining BP algorithm of genetic algorithm with hybrid gas sensors, a kind of quantitative analysis system of mixed gas is designed. The local minimum of network learning is the main reason which affects the precision of gas analysis. On the basis of the network study to improve the learning algorithms, the analyses and tests for CO, CO₂ and HC compounds were tested. The results showed that the above measures effectively improve and enhance the accuracy of the neural network for gas analysis.