Research On Observation Data Analysis And Correction Algorithm Of AWS Temperature And Humidity Sensor

The temperature and humidity sensors on the application of Automatic Weather Stations (AWS), which is in the observations of the actual business, due to the influence of temperature, humidity, pressure and other environmental parameters, may introduce additional measurement uncertainty. Therefore, considering the temperature and humidity sensor under the multi-param-eter measurement uncertainty can have scientific, comprehensive grasp of the quality of the observational data.In this thesis, according to the environmental impact parameters of the temperature and humidity sensor in the actual business observing the process, firstly, using theoretical analysis to identify the main influencing factors, which make a detailed experimental program, through analysis of the experimental data, we conclude that temperature have a big impact on humidity sensors, at the same time, humidity have a greater impact on the temperature sensors. Then correct the influence factor using the improved genetic algorithm (Genetic Algorithm GA) combined with back-propagation neural network (Back Propagation BP) algorithm, least squares and radial basis function (Radical Basis function RBF) neural network fusion algorithm and the Fourier-based neural network algorithm. Finally, re-evaluate the uncertainty of the temperature and humidity sensors, which will provide a scientific basis for the assessment of the quality of the observational data, and improve the ground detection technology development plan. The main contents include the following:Analysis and study the influence factors of temperature and humidity sensors. Two angles from the theoretical analysis and experimental studies demonstrate the impact of environmental parameters, qualitative drawing mainly affect factor, thus further quantitative experimental study. Analysis of the experimental data, we conclude that some experimental results, one is the presence impact of temperature on the humidity sensor, the other is the presence influence of humidity on the temperature sensor.Correction algorithm model for the influencing factor of humidity sensor. Starting from the experimental data of temperature and humidity sensors, respectively, using the improved GA-BP neural network algorithm and least squares fusion algorithm, combined with the RBF neural network to establish an appropriate mathematical model, study the impact factor temperature compensation. The results show that the two methods can be effectively compensated for the impact of temperature and humidity sensors, which improve the reliability of the observations of the humidity sensor.Correction algorithm model for the impact factor of the temperature sensor. Fourier-based neural network algorithm, using this method to establish the mathematical model of the temperature sensor impact factor correction, combined with experimental data measured compensation humidity and other environmental parameters of sensor measurements, effectively guarantee the quality of data on temperature sensor.Uncertainty assessment of the measurement results for temperature and humidity sensor. Through the analysis of the sources of uncertainty in the actual observations of temperature and humidity sensor business, establish a corresponding mathematical model, the calculated uncertainty evaluation method in accordance with the evaluation method of the Class A and Class B, which is more scientific and accurate judgment sensor performance.