Font Size: a A A

Faults Classification Based On DAMADICS And Stiction Modeling And Compensation For Control Valve

Posted on:2014-02-01Degree:MasterType:Thesis
Country:ChinaCandidate:X W ZhouFull Text:PDF
GTID:2232330395977661Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
Process industries refer to those areas like petroleum, chemical plant, paper manufacturing and so on, which generally have a huge scale and complex procedure and can consume large amount of energy, playing an important role in economic development. Considering the scale and complexity, the feedback control loops are largely applied in the plants and the control performance is a key factor to guaranteeing the quality of the final products. In the control loop, pneumatic valve is a most common used actuator and whether it can work in good condition or not will definitely affects the control performance of the loop. This paper focuses on the pneumatic valve fault diagnosis, fault modeling and fault compensation so as to improve the control performance of control loop and guarantee product quality.Valve fault detection and isolation were performed based on a benchmark platform DAMADICS. More specifically, the residual method was applied to detect the fault of the simulating valve online via some variable data from the DAMADICS platform. If one fault detected, some features of the process would be extracted and data processing algorithm like PCA would be used to optimize the features. The SVM classifiers were chosen to isolate the faults of valve. A series of profound researches about valve stiction were done and the data-driven stiction model of valve was analyzed to study the characteristic of the stiction phenomenon. Based on the stiction model, this paper proposed a new method to compensate the factor of siction in the feedback loops. The compensator using the existed information of the loop, such as set point, output of controller and process variable, could generate new signals added to the input of stiction valve to eliminate the oscillation in the control loop.In order to verify these methods mentioned above, all of them were conducted on the MATLAB/SIMULINK environment. These simulating experiments proved the methods of valve faults classification, modeling and compensation feasible and reasonable. As to the stiction compensation, experiments were done based on the real experimental equipment, and the results also indicated the proposed method feasible.
Keywords/Search Tags:Process Industries, Control loop, Valve fault, stiction modeling, compensation
PDF Full Text Request
Related items