基于光纤传感技术的油液污染度软测量研究

采用基于Mie理论的激光散射法测量油液污染度时,油液污染度与监测参数——入射光强、散射光强、出射光强以及颗粒参数之间存在着复杂的非线性关系,给油液颗粒污染度的准确测量带来困难。利用支持向量机优良的非线性映射和强大的泛化能力,建立了一个基于最小二乘支持向量机的油液污染度软测量模型,给出了相应的系统结构和算法;仿真和实际运行结果表明基于LS-SVM的油液污染度软测量模型具有较高的估算精度与泛化能力,为油液污染度的在线测量提供了一种简单、可靠的新方法。

Research on Soft-sensing of Contamination Degree of Hydraulic Oil Based on Optical Fiber Sensing Technology

When measuring the contamination degree of the hydraulic oil by the optical fiber scattering method hased on the Mie theory, there are nonlinear relationship between the oil contamination degree and the detecting parameters which are the incidence light density, scattering light density, transmission light density and the particle parameter. That makes very difficult to measure the contamination degree of the hydraulic oil particle. In this paper, using the better nolinear representation and strong generative capability of the least square support vector machine （LS-SVM） , a hydranlic oil contamination degree soft measuring model has been set up based on the LS-SVM. The relative system construetion and algorithm are put forward. The simulation and practicle operation result show that the soft measuring model based on the LS-SVM hy- draulic oil contamination degree has higher estimation accuracy and generative capacity, which provides a simple and new method for the on-line measurement of the hydraulic oil contamination degree.