基于BP神经网络的激光氧分压传感器温压补偿算法

Temperature and Pressure Compensation Algorithm of Laser Oxygen Partial Pressure Sensor Based on BP Neural Network

针对可调谐半导体激光吸收光谱(Tunable Diode Laser Absorption Spectroscopy,TDLAS)技术在密闭空间中测量氧分压时受温度和压力的影响,本文提出了基于BP神经网络的激光氧分压传感器温压补偿算法来消除温度和压力对氧分压测量的影响。采用模拟电路构建了激光氧分压传感器,并通过电连接器上传测量的二次谐波峰值、温度和压力值至上位机,上位机通过BP神经网络温压补偿算法进行温压补偿,获取准确的氧分压值。实验结果表明,该算法可以实现对测量的氧分压补偿,补偿后氧分压测量误差小于±1 kPa,能满足航天、航空等领域的要求,具有较好的应用前景。

Considering the influence of temperature and pressure on the measurement of the partial pressure of oxygen using TDLAS technology,a temperature and pressure compensation algorithm for laser oxygen partial pressure sensors is proposed for eliminating these effects.The laser oxygen partial pressure sensor was constructed using an analog circuit,and the measured second-harmonic peak,temperature,and pressure values were uploaded to the master computer through an electrical connector,which performed temperature and pressure compensation using the BP neural network temperature and a pressure compensation algorithm to obtain accurate oxygen partial pressure values.The experimental results show that the algorithm can achieve compensation for the measured partial pressure of oxygen,and that the error of the partial pressure of the oxygen measurement is less than±1kPa,which meets the requirements of aerospace,aviation and other fields,and has a desirable application prospect.