Fluid dynamics analysis and experimental study for solar radiation error correction of sounding humidity sensor.

Abstract

In order to correct the solar radiation error of relative humidity, the mainstream capacitive sounding humidity sensor HC103M2 is selected and investigated by simulation analysis and experimental verification. First, the basic theories for solar radiation error and sensor error itself are elaborated, and simulational and experimental platforms are introduced. The computational fluid dynamics (CFD) method is utilized to theoretically investigate the dry error of the humidity sensor caused by solar radiation heating, which is related to radiation intensity, altitude, and solar elevation angle as well as reflectivity, thickness, and shape of the shield. Then, in order to verify the accuracy of the simulation, an experimental platform including a humidity sensor and two temperature sensors to measure the solar radiation heating is built to analyze the relative error of humidity obtained by the CFD simulation and experiment. It is found that their maximum deviation is 3.30% and the average error is 1.94%, which indicates that the calculation using the CFD method is accurate and feasible. In order to easily and operationally predict the solar radiation heating of the humidity sensor, a back propagation (BP) neural network fusion algorithm based on three inputs of radiation intensity, air pressure, and solar elevation angle is proposed. Compared with the solar radiation heating obtained by CFD simulation, the maximum absolute error is about 0.2\xa0K, and the relative error of humidity is about ±1.30%. Finally, a case of vertical humidity profile correction considering the temperature-sensitive error of HC103M2 is analyzed. The response time of sensor measurement and the airflow into the shield are discussed as well. The corrected results after taking solar radiation heating into account are more similar to those measured by RS92 and cryogenic frost point hygrometer (CFH). This result shows that the prediction model is accurate, which may be applied to correct the dry error and further improve the accuracy of humidity measurement.