Volume 11, Issue 3 (11-2021)
JEM 2021, 11(3): 2-11 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Shajiee M, Hosseini Sani S K, Shamaghdari S, Naghibi Sistani M B. Robust Dynamic Sliding Mode Torque Observer for Wind Turbine. JEM 2021; 11 (3) :2-11
URL: http://energy.kashanu.ac.ir/article-1-1445-en.html
URL: http://energy.kashanu.ac.ir/article-1-1445-en.html
Department of Electrical Engineering, Ferdowsi University of Mashhad
Abstract: (2030 Views)
Introduction: A wind turbine is an electromechanical system that converts wind kinetic energy into electrical energy. For a highly nonlinear model like a wind turbine, a structure like the Luenberger observer is not ideal. For this reason, based on the stability analysis of the error dynamics, a general framework for analysis and design of robust Dynamic Sliding Mode Observer (DSMO) is considered.
This paper investigates a new scheme in the design of a dynamic observer within the
scheme for generator torque estimation in the wind turbine and uses a dynamic sliding mode observer design that has never been used for a wind turbine. This analysis, in contrast to conventional design methods, considers the general form of the linear part as well as the Lipschitz constant of the nonlinear part.
Materials and methods: Section 2. describes the mathematical benchmark model of drive train in brief. It includes the significant wind turbine dynamic characteristics without becoming too complicated. In this paper, the partial load area is considered. The reason why the data for the wind speed from the cut-in and the rated wind speed range was considered was that this operational region presents a major opportunity to optimize the wind-turbine power-generation process. Design innovation is the use of dynamic gain in the design of the SMO and the guarantee of asymptotic stability condition by optimal control problem. The design methodology is the use of dynamic gain instead of static one in the sliding mode observer. The dynamic observer design offers an extra degree of freedom over the classical static gain observer. The additional degree of freedom proposed by this dynamic formula is used to cope with nonlinearity. Finally, using the real wind profile(with a realistic stochastic wind speed profile measured from a Binalood wind farm), the design procedure is implemented on a fully nonlinear model of the drive train 100 kW wind turbine of the Sun and Air Research Institute. Results and nonlinear model are also evaluated and validated with the wind turbine simulator software FAST.
Discussion and Conclusion: This paper presented the FAST validation of a drive train model, and a new elegant scheme to design a robust dynamic SMO with technique fortorque estimation was investigated. By the proposed dynamic sliding mode observer, the Lipschitz constant has been increased, which is equivalent to an increase in the operation range of a wind turbine and an increase in the observer's robustness to the nonlinear term. Fig.8 presents the comparison of system performance in four different cases. The Luenberger observer is with a static and dynamic gain, and a sliding mode observer is with static and dynamic gain. As it was illustrated, our DSMO provides a fast convergence rate that clarifies the satisfactory DSMO performance in comparison with all other observers, even the dynamic Luenberger [22] which is the advantage of the proposed design procedure.
Also, the second contribution of this paper is the superiority of the proposed algorithm over others is investigated. By adding process noise and measurement noise to the turbine, A modified mixed
strategy to deal with the noise is utilized in section 6. to attenuate the noise effect on the estimation error.
Applying the design method to the nonlinear faulty system and also using LMI solutions instead of the Riccati equation are suggested for future work.
This paper investigates a new scheme in the design of a dynamic observer within the
scheme for generator torque estimation in the wind turbine and uses a dynamic sliding mode observer design that has never been used for a wind turbine. This analysis, in contrast to conventional design methods, considers the general form of the linear part as well as the Lipschitz constant of the nonlinear part.Materials and methods: Section 2. describes the mathematical benchmark model of drive train in brief. It includes the significant wind turbine dynamic characteristics without becoming too complicated. In this paper, the partial load area is considered. The reason why the data for the wind speed from the cut-in and the rated wind speed range was considered was that this operational region presents a major opportunity to optimize the wind-turbine power-generation process. Design innovation is the use of dynamic gain in the design of the SMO and the guarantee of asymptotic stability condition by
optimal control problem. The design methodology is the use of dynamic gain instead of static one in the sliding mode observer. The dynamic observer design offers an extra degree of freedom over the classical static gain observer. The additional degree of freedom proposed by this dynamic formula is used to cope with nonlinearity. Finally, using the real wind profile(with a realistic stochastic wind speed profile measured from a Binalood wind farm), the design procedure is implemented on a fully nonlinear model of the drive train 100 kW wind turbine of the Sun and Air Research Institute. Results and nonlinear model are also evaluated and validated with the wind turbine simulator software FAST.Discussion and Conclusion: This paper presented the FAST validation of a drive train model, and a new elegant scheme to design a robust dynamic SMO with
technique fortorque estimation was investigated. By the proposed dynamic sliding mode observer, the Lipschitz constant has been increased, which is equivalent to an increase in the operation range of a wind turbine and an increase in the observer's robustness to the nonlinear term. Fig.8 presents the comparison of system performance in four different cases. The Luenberger observer is with a static and dynamic gain, and a sliding mode observer is with static and dynamic gain. As it was illustrated, our DSMO provides a fast convergence rate that clarifies the satisfactory DSMO performance in comparison with all other observers, even the dynamic Luenberger [22] which is the advantage of the proposed design procedure.Also, the second contribution of this paper is the superiority of the proposed algorithm over others is investigated. By adding process noise and measurement noise to the turbine, A modified mixed
strategy to deal with the noise is utilized in section 6. to attenuate the noise effect on the estimation error.Applying the design method to the nonlinear faulty system and also using LMI solutions instead of the Riccati equation are suggested for future work.
Type of Study: Applied Article |
Subject:
Electrical Engineering
Received: 2020/03/14 | Revised: 2022/01/2 | Accepted: 2020/08/4 | Published: 2021/11/1
Received: 2020/03/14 | Revised: 2022/01/2 | Accepted: 2020/08/4 | Published: 2021/11/1
Extended Abstract [PDF 80 KB] (76 Download)
| Rights and permissions | |
 |
This work is licensed under a This work is licensed under a Creative Common International License.. |
