Research On Modeling And Compensating For Temperature Error Of Pendulous Accelerometer

Accelerometer is one of the important inertial sensors in the inertial navigation system and its precision will directly influence the precision of inertial navigation system. Error in measurement may occur due to the working environment and condition of the accelerometer, especially the influence of temperature shall not be neglected. So modeling and compensating for temperature error is an important approach to achieve higher accuracy. The paper mainly focuses on the characteristics of static temperature of pendulous accelerometer.Firstly, according to the structural characteristics of pendulous accelerometer, the error model in the conditions of line motion and angle motion is analyzed theoretically in this paper.Secondly, the error model of I/F conversion circuit is researched in this thesis. The principle of I/F conversion circuit and the effect of I/F conversion circuit factors are analyzed, and the temperature model of I/F conversion circuit is described through polynomial fitting method. After compensating the temperature error by using the temperature model, the data acquisition precision of I/F conversion circuit has been improved.Thirdly, in the temperature-control turntable, the static temperature experiment of accelerometer is tested. The relation of static model coefficients and temperature is analyzed, and the temperature models of accelerometer bias and scale factor are established between -50℃and +68℃. After compensating the temperature error through the temperature model, the post-fit residuals of the accelerometer output is lowed about an order of magnitude, and the trend term of accelerometer changing with temperature vanished basically. The metrical precision of the accelerometer has been improved observably.Finally, the temperature models of accelerometer and I/F conversion circuit are tested in the experiment of inertial measurement unit.The results show that the temperature models of pendulous accelerometer are correct and effective. Compensating these models is an efficient way to enhance the accuracy of accelerometer.