Design Of A New Surface Temperature Sensor And Correction Of Radiation Error

In meteorological observations,the surface atmospheric temperature is an important parameter to reflect the surface conditions,and the overall rate of change of atmospheric temperature is 0.1 ℃ per 10 years.In the actual measurement,due to the influence of solar radiation traditional radiation shield will produce about 1 ℃ order of magnitude of radiation error,radiation error is directly related to the accuracy of weather prediction and atmospheric science research.In order to reduce the radiation error and improve the detection accuracy,this paper proposes a new type of surface temperature sensor.The design accelerates the air flow around the sensor to achieve forced ventilation and promote radiant heat exchange by means of the piezoelectric vibrator bending and vibrating to drive the temperature measurement probe to swing.The computational fluid dynamics method is used to quantify the radiation error of the sensor in a multi-physical field environment,including different environmental parameters such as radiation intensity,altitude,subsurface reflectivity,airflow velocity and solar altitude angle.The simulation results show that the radiation error is positively correlated with the radiation intensity,altitude and lower cushion reflectivity,and the radiation error is negatively correlated with the airflow velocity,and the solar altitude angle has less influence on the radiation error,which verifies the effectiveness of the structure design.Then to ensure that the sensor achieves effective measurement,the hardware temperature acquisition circuit is designed.To optimize the radiation error in different environments,three radiation error correction algorithms,BP neural network,particle swarm optimization BP neural network and Gaussian process regression,are proposed for optimization,and the optimal prediction model is determined by comparison.The results show that the root mean square error between the corrected and true values of the particle swarm optimization BP neural network algorithm is 0.0054 ℃,and the average absolute error is 0.0035 ℃,with the best radiation error correction effect.Then,in order to verify the actual measurement effect of the sensor,an external observation platform of radiation error is built,and the temperature value of 076 B forced ventilation observation instrument is used as the benchmark to carry out the observation experiment.The experimental results show that the mean absolute error between the measured value of the new surface temperature sensor and 076 B after error correction is 0.041 ℃,and the root mean square error is 0.055 ℃.It is shown that the new surface temperature sensor has the advantages of high measurement accuracy and better stability to meet the demand for high-precision temperature data in meteorological detection.In order to meet the current trend of intelligent informationization in the field of meteorological observation,finally,the Java language is used to design the radiation error query interface so that the data can be displayed more intuitively to the users and help them to save and analyze the information of temperature and other related meteorological parameters.