面向高空温度观测的高精度温度传感器研究

Accurate measurement of the upper atmosphere temperature is of great significance to climate change research, but it is difficult to improve the observation accuracy to 0.05K order based on the structure,size,surface coating reflectivity and the solar radiation error correction method of the traditional sounding temperature sensor, and it can't meet the application requirements of climate change research. To tackle this problem, a sensor design of four leads and cross structure is proposed in this project. The sensitivity of the sensor observations for the radiosonde oscillation and rotation is reduced to 1%. On this basis, in order to realize the sensor probe size reduced to 0.4mmmagnitude and the surface reflectivity increased to 95%, the insulating film and silver film of micron order are prepared for the sensor surface by microelectronics coating technology. To further improve the accuracy of observation, a multi-physical field coupled heat transfer study is established by a computational fluid dynamics method. The radiation error correction equations are fitted by a neural network algorithm and simulation results. Therefore, a new method of correcting solar radiation error is proposed. Finally, the sun simulator and low wind tunnel platform are set up and the fly-out field experiments are carried out to verify the observation accuracy of the sensor. It is hopeful that the observation accuracy of the new temperature sensor is improved to 0.05K order by this project study, and it should promote the development of climate change research in the future.

高空气温的准确测量对气候变化研究具有重要意义，但传统探空温度传感器的结构、尺寸、表面涂层反射率和太阳辐射误差修正方法难以将观测精度提高到0.05K量级，无法满足气候变化研究的应用需求。为解决这个问题，本项目拟提出一种具有四引线十字型结构的传感器设计，尝试将传感器观测结果对探空仪摆动和旋转的敏感性降低到1%。在此基础上，利用微电子镀膜技术在传感器表面制备微米级绝缘膜和银膜，以实现传感器的探头尺寸缩小至0.4mm量级和表面反射率提高到98%。为进一步提高观测精度，采用计算流体力学方法对新研制的传感器进行多物理场耦合传热研究，获得的仿真结果通过神经网络算法拟合太阳辐射误差修正方程，从而提出一种新的太阳辐射误差修正方法。最后，搭建太阳模拟器、低气压风洞设备平台和进行外场放飞实验验证新型传感器的观测精度。希望通过上述研究，将新型传感器的高空温度观测精度提高到0.05K量级，推动气候变化研究的发展。

伴随着数值天气预报和气候变化研究精细化程度的不断提高，希望探空温度传感器观测精度提高到0.05K量级，但辐射误差的准确修正已成为制约观测精度提高的瓶颈问题。为解决这个难题，针对珠状热敏电阻提出一种在三维空间对称的四引线十字型结构设计，并运用计算流体动力学方法建立珠状热敏电阻和外围空气域的流固耦合模型，对珠状热敏电阻在不同太阳照射方向和不同引线数量条件下的温度分布进行数值模拟研究。结果表明：珠状热敏电阻的四引线十字型结构能够显著降低摆动对观测精度的影响，结合计算流体动力学方法有望将太阳辐射误差的观测精度提高到0.05K量级；两引线和六引线结构改善摆动影响的能力均弱于四引线结构，验证了四引线结构是目前最优的结构设计方案。为增加上述研究结果应用的普适性，通过BP神经网络算法可拟合多个变量的辐射误差修正方程，可实现影响因子连续变化时提供准确辐射误差修正数据的目标。为验证计算流体动力学方法和辐射误差修正方程的准确性，搭建了低气压风洞和太阳模拟器的实验平台，这个平台可提供不同太阳辐射强度和海拔高度的物理环境，从而可获得准确的辐射误差数据。