火炮抽气装置压力测试中温度影响的实验研究

【作者】 桂宝佳；

【导师】 袁人枢；

【作者基本信息】 南京理工大学 ， 机械电子工程， 2007， 硕士

【摘要】 压力传感器的输出值通常不仅决定于目标参量(压力),还会受到非目标参量的影响,如在火炮抽气装置压力测试中,火药气体的温度对压力传感器的影响。论文首先概述了压力传感器温度补偿的一般方法和压阻式压力传感器结构、原理以及温度补偿的必要性。利用压阻式压力传感器,组建基于PXI总线仪器的实验系统,获取温度补偿所需要的实验数据:并在Labview开发环境下编写实验数据采集、温度误差分析、灵敏度温度系数分析等程序,编写基于双线性、双三次、双三次样条二维插值法的压阻式压力传感器的温度补偿程序:论文利用Matlab神经网络工具箱设计了BP神经网络模型,然后根据实验数据对BP神经网络进行训练,得到压力传感器受温度影响的模型,并由Labview通过Matlab script节点调用该模型来开发压力传感器的温度BP神经网络补偿系统;论文对二维插值函数法和BP神经网络法进行了比较,结果表明二维插值函数法和BP神经网络法对压阻式压力传感器温度补偿效果非常明显。更多还原

【Abstract】 Usually, the output of pressure sensor not only depends on target parameter (pressure), but also is affected by non-target objection parameter. For example, the pressure sensor is affected temperature of powder gas in pressure measurement of gun bore evacuator. First of all, the paper describes theory and method of temprature error compensation of piezoresistive sensor theory, then, the paper gives the theory, concept and character of piezoresistive pressure sensor concept and character of piezoresistive pressure sensor. Because piezoresistive pressure sensor’s sensitivity and measurement precision is influenced by temperature. To take piezoresistive pressure sensor for example, the paper constructs experiment system based on PXI bus instruments, to acquire experiment data for compensating software to compensatate the temperature error of piezoresistive sensor. Then, temperature error and sensitivity parameter of temperature were analyzed, and a method for compensating temprature error and non-linearity of piezoresistive pressure sensor is designed using 2-dimensional interpolation of bilinear, bicubic and bicubic spline function. The application programs of data acquisition, error analysis, and temprature error compensation of piezoresistive sensor have been completed on development environment of Labview. Finally, the paper presents the realization of pressure sensor temperture error compensation system in Labview and the application of the BP neural network trained by experimental samples in Matlab neural network toolbox as the model for Labview Matlab script node call. The paper compare 2-dimensional interpolation functions, all of experimental results show that the effect of temperture error compensation of 2-dimensional interpolation and BP neural network is excellent.更多还原