Research On Indoor Formaldehydegas Detection Based On Deep Learning And NDIR Technology

Formaldehyde is a common and volatile harmful gas with a strong irritating odor.Long-term exposure to high concentrations of formaldehyde indoors can cause serious harm to human health,and may even induce cancer.Therefore,the detection of indoor formaldehyde has become particularly important.However,traditional gas detection methods have certain limitations,such as slow response speed and low detection accuracy.Therefore,based on non-dispersive infrared technology(NDIR),this study developed an indoor formaldehyde detection system and used an improved LSTM network for temperature and humidity compensation to achieve highprecision detection of indoor formaldehyde in complex environments.This paper analyzed mainstream formaldehyde gas detection methods and ultimately selected NDIR.In response to problems that may arise during the detection process,the technology used was optimized and improved.The main work completed includes the following three aspects:(1)NDIR technology has the advantages of long-term stability,fast response,high precision,and high selectivity.Based on the Lambert-Beer law,the nondispersive double-channel structure and differential principle are adopted to effectively suppress various external factors such as unstable infrared light sources,dust mixing,and other gases mixed in.The hardware and software design of the detection system have been completed,as well as the establishment and testing of the temperature and humidity compensation model.During the testing process,the experiment shows that the system has high precision and stability.(2)In response to the distortion of the target signal extracted due to the addition of noise when the signal-to-noise ratio of the output signal of the infrared detector is less than-60 d B,a Kalman filter algorithm was added to the end of the lock-in amplifier.After introducing the Kalman filter,the noise reduction ability of the lockin amplifier was enhanced,and the range of signal-to-noise ratio for identifying effective signals was further improved.The experiment showed that the use of the improved method significantly improved the measurement accuracy,and the fluctuation rate of effective signals also decreased.(3)In response to the impact mechanism of the error generated by the formaldehyde detection system under different temperature and humidity conditions,the ISSA-LSTM neural network model was established by adding the sparrow search algorithm to the LSTM network to automatically set hyperparameters and improve the temperature and humidity compensation effect.However,the sparrow search algorithm is prone to falling into local optima in the later stages of iteration.Therefore,chaos theory and other methods were used to optimize the sparrow search algorithm.The experimental results showed that the anti-interference ability of the detection system was significantly enhanced,providing new directions for temperature and humidity compensation.