基于模糊聚类的阀门粘滞检测与量化

发布时间：2018-06-28 01:19

本文选题：控制回路 + 气动调节阀　；参考：《江南大学》2017年硕士论文

【摘要】：在流程工业中,气动调节阀是使用最广泛的一种执行器设备,其可靠性和安全性直接关系着控制系统的性能表现。由于调节阀通常安装在高温高压等较复杂的环境中,长时间与工艺介质直接接触,容易受到腐蚀,从而造成阀杆和密封部件之间的摩擦力增大,导致了粘滞非线性现象的产生。阀门粘滞非线性的存在影响控制器和调节阀达到最佳的运行状态,严重的粘滞故障还会引起控制回路发生振荡现象。因此,对阀门粘滞非线性的分析研究也是控制系统性能监控研究领域的一个重要组成部分。粘滞非线性的检测和量化研究是工程人员进行阀门维护以及改善控制回路性能的基础,有很强的工程应用价值,对实现流程工业的“安、稳、长、满、优”具有重大的意义。本文的主要研究内容如下:1、提出了一种改进的基于模糊聚类算法的阀门粘滞检测方法。在粘滞检测过程中,现有方法容易将具有非线性特征的外部扰动振荡误判为阀门粘滞非线性引起的。本文通过深入分析与研究,提出了改进方案,在其基础上增加了四边形凹凸性判断环节以及定义了一个新粘滞指标,使之能够有效的辨识阀门粘滞非线性和外部扰动所引起的振荡现象,并且对噪声信号有一定的抗干扰能力。通过实验验证了改进方法的有效性和准确性,适用于流量控制回路或带有智能阀的控制回路的阀门粘滞检测。2、提出了一种新型的模糊聚类下粘滞非线性量化方法。首先分析了阀门粘滞的量化问题和基本思路,并对基于椭圆拟合法的粘滞非线性量化算法展开了讨论与研究,分析了这类量化方法的有效性及其存在拟合椭圆形状对控制器参数敏感等不足之处。然后在所提出改进方法的基础上进一步考虑了对粘滞大小的量化,将其中两个聚类中心的有效距离作为粘滞的估计量。利用实际流程工业数据对两种量化方法进行了比较测试,验证了所提方法具有良好的效果以及推广使用的可能性。3、从实际应用角度出发,结合现有实验条件,设计和开发了阀门粘滞检测与量化系统软件,将本文所提方法推广至实际应用中去,主要实现理论研究的可视化与操作。介绍了软件的设计目标与设计说明,包括欢迎界面、数据导入、图像绘制、检测结果参数显示、工具箱等各模块的功能,并对软件的运行效果进行了操作展示。最后,总结了全文的研究工作,并对阀门粘滞非线性的检测与量化问题作了进一步的探讨与展望。
[Abstract]:Pneumatic control valve is the most widely used actuator in the process industry. Its reliability and safety are directly related to the performance of the control system. Because the regulating valve is usually installed in the complicated environment such as high temperature and high pressure, it is easy to be corroded because of the contact with the process medium for a long time, which results in the increase of friction between the valve stem and the sealing parts, which leads to the phenomenon of viscous nonlinearity. The nonlinearity of valve viscosity affects the optimal operation state of the controller and control valve, and the serious viscous fault will cause the oscillation of the control circuit. Therefore, the nonlinear analysis of valve viscosity is also an important part in the field of control system performance monitoring. The detection and quantification of viscous nonlinearity is the basis for engineering personnel to maintain valves and improve the performance of control circuits. It has strong engineering application value and has great significance for realizing "safety, stability, length, fullness and excellence" in process industry. The main contents of this paper are as follows: 1. An improved valve viscosity detection method based on fuzzy clustering algorithm is proposed. In the process of viscosity detection, the existing methods can easily misjudge the external disturbance oscillation with nonlinear characteristics as that caused by the valve viscosity nonlinearity. In this paper, through in-depth analysis and research, an improved scheme is put forward, based on which a quadrilateral concave convexity judgment link is added and a new viscosity index is defined. It can effectively identify the oscillation phenomenon caused by the viscous nonlinearity and external disturbance of the valve, and has certain anti-interference ability to the noise signal. The effectiveness and accuracy of the improved method are verified by experiments. It is suitable for the viscosity detection of valves in flow control loop or control loop with intelligent valve. A new nonlinear quantitative method for viscosity under fuzzy clustering is proposed. Firstly, the problem of valve viscosity quantization is analyzed, and the nonlinear quantization algorithm based on elliptic fitting method is discussed. The effectiveness of this method and its shortcomings such as the sensitivity of fitting ellipse shape to controller parameters are analyzed. Then the quantization of the viscosity is considered on the basis of the improved method, and the effective distance between the two clustering centers is taken as the estimator of the viscosity. This paper compares and tests the two quantitative methods by using the actual process industry data, and verifies that the proposed method has good effect and the possibility of popularizing and using. From the point of view of practical application, combined with the existing experimental conditions, The software of valve viscosity detection and quantification system is designed and developed. The method proposed in this paper is extended to practical application, and the visualization and operation of theoretical research are mainly realized. This paper introduces the design objectives and design instructions of the software, including the functions of various modules, such as welcome interface, data import, image drawing, parameter display of inspection results, toolbox and so on, and the operation effect of the software is demonstrated. Finally, the research work of this paper is summarized, and the detection and quantification of valve stickiness nonlinearity are discussed and prospected.
【学位授予单位】：江南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH138.52

【相似文献】