Pageflex Server [document: D-Aalto-565839ED_00001]

Aalto University, P.O. Box 11000, FI-00076 Aalto www.aalto.fi Author Timo Manninen Name of the doctoral dissertation Fault Simulator and Detection for a Process Control Valve Publisher School of Engineering Unit Department of Engineering Design and Production Series Aalto University publication series DOCTORAL DISSERTATIONS 146/2012 Field of research Machine technology Manuscript submitted 12 June 2012 Date of the defence 30 November 2012 Permission to publish granted (date) 15 October 2012 Language English Monograph Article dissertation (summary + original articles) Abstract In this study a novel fault simulator for a quarter-turn pneumatic control valve is presented. Another contribution of this study is a novel fault detection and diagnosis method for a control valve. It has been possible to analytically model control valve dynamics, despite their inherent nonlinearities for the fault simulator. These nonlinearities of the system have been identified and estimated through selected parameters. The models that were derived have been verified with measurements and the modeling error is found to be acceptable for the fault simulations. Some typical control valve faults have been simulated and the impacts on the internal variables of the flow control loop and control performance analysed. The fault simulator presented here can be used for fault detection and diagnosis, as well as robust control research. On the basis of simulations and test bench test runs it is possible to detect and diagnose typical control valve faults before they have a severe impact on flow control loop performance. This can be done with the online method introduced in this study, which requires low computing power. This means that the method is implementable in an intelligent valve controller and diagnosis can be performed without disturbing the process. The method that is introduced is based on the observation that the internal variable closest to the fault compensates and reacts first to the fault when feedback control is utilised. That leads to an operation point shift for all the internal variables before the fault in the chain of the internal variables in the system. When these operation point changes are being detected, the faults can be detected and diagnosed. This principle can be utilised in all feedback-controlled mechatronic systems. The fault detection and diagnosis method introduced here was verified with the fault simulator and test bench runs and found to be applicable to the detection and diagnosis of all the faults that were tested.In this study a novel fault simulator for a quarter-turn pneumatic control valve is presented. Another contribution of this study is a novel fault detection and diagnosis method for a control valve. It has been possible to analytically model control valve dynamics, despite their inherent nonlinearities for the fault simulator. These nonlinearities of the system have been identified and estimated through selected parameters. The models that were derived have been verified with measurements and the modeling error is found to be acceptable for the fault simulations. Some typical control valve faults have been simulated and the impacts on the internal variables of the flow control loop and control performance analysed. The fault simulator presented here can be used for fault detection and diagnosis, as well as robust control research. On the basis of simulations and test bench test runs it is possible to detect and diagnose typical control valve faults before they have a severe impact on flow control loop performance. This can be done with the online method introduced in this study, which requires low computing power. This means that the method is implementable in an intelligent valve controller and diagnosis can be performed without disturbing the process. The method that is introduced is based on the observation that the internal variable closest to the fault compensates and reacts first to the fault when feedback control is utilised. That leads to an operation point shift for all the internal variables before the fault in the chain of the internal variables in the system. When these operation point changes are being detected, the faults can be detected and diagnosed. This principle can be utilised in all feedback-controlled mechatronic systems. The fault detection and diagnosis method introduced here was verified with the fault simulator and test bench runs and found to be applicable to the detection and diagnosis of all the faults that were tested.