油品含水量智能检测仪的开发与信息融合技术的研究

【作者】 罗桂娥；

【导师】 杨欣荣；

【作者基本信息】 中南大学 ， 控制理论与控制工程， 2002， 硕士

【摘要】 本文介绍了一种油品含水量智能检测系统，该测量系统具有响应快、精度高和使用方便等优点。因水的介电常数远大于油的介电常数，使两者呈现不同的射频阻抗，这是油品含水量测量仪的理论依据。但在油品检测过程中发现传感器的输出值不只取决含水量一个参量，环境温度发生变化时系统的输出值也会发生变化，即传感器存在对温度的交叉灵敏度。存在交叉灵敏度的传感器，其性能不稳定，测量精度较低。因此在油品含水量智能检测系统中，通过对温度和水分两个参量进行监测，并采用多传感器信息融合技术对传感器输入信息温度和水分、输出信息温度电压和水分电压进行融合处理以提高目标参量的测量精度和系统的温度稳定性。二传感器信息融合的方法有多种，作者主要探讨了曲面拟合法和神经网络法。对传感器用机油进行实验的结果表明：传感器的零位温度系数、灵敏温度系数、温度的稳定性及测量精度与未进行信息融合前相比均有较大的提高。 将多传感器信息融合技术应用于油品含水量测量系统中，解决了传感器的非线性和温度对水分的交叉灵敏度问题，能提高系统对目标参量的辨识能力及快速有效获得高精度的测量结果。同时将虚拟仪器技术引入该检测系统，采用图形化编程软件LabVIEW进行程序设计，使得界面设计灵活、简单，测量直观，操作简单易行。更多还原

【Abstract】 In this article, a kind of intelligent measuring system used to measure the water content in oil products is introduced. This system has many advantages: quick response , high accuracy , and convenience to use. Based on the theory of radio-frequency impedance, the dielectric constant of water is much greater than that of oil, so they show different properties in the respect of radio frequency impedance. Taking advantage of this characteristic, we designed Radio frequency capacitancesensor to detect the water-content in oil products.In the process of meterage, we found that the output of the sensor is not determined by the parameter of water content and it is also changed when the temperature of environment is changed. That is to say, the system has certain cross sensitivity about the environment which also affects measurement precision. In order to improve the measurement precision and maintain the temperature stability of the system, we put forward poly information fusion technique to deal with this problem. In bi-sensors information fusion techniques, there are a lot of methods, in which we applied two of them: curved surface fitting and neural network. In sample experiments by using lubricant oil, it is showed that the sensor’s zero temperature coefficient, sensitivity temperature coefficients, temperature stability and measurement precision compared with results before information-fusion processing were greatly improved.The results of experiment are proven that it is an effective way. At the same time, we applied Virtual Instrument technique to develop detector, which saved valuable development time and resources.The main characteristics of the virtual detector are high precision, flexibility and convenience to operate.更多还原