Recent Trend of Magnetic Bearing System and its Application for Energy Storage Flywheel System

Transcription

1 CONTRIBUTION Recent Trend of Magnetic Bearing System and its Application for Energy Storage Flywheel System Professor Kenzo NONAMI Department of Electronics and Mechanical Engineering, Faculty of Engineering, Chiba University This paper is concerned with the recent trend of a research and development project of magnetic bearing control and its application for flywheel energy storage system. Some new results of magnetic bearing control system are described. Especially, a novel control method called zero power control is introduced. The zero power control means not only the zero bias current but also the zero control current for rotor near its equilibrium. The electromagnetic bearings are used for zero bias control as zero power nonlinear control. After the situation of the energy storage flywheel system is introduced, the application of the zero power control to the flywheel system for better efficiency of energy storage flywheel system is shown in this paper. Key Words: Magnetic bearing, Zero Power Control, Zero Bias Control, Flywheel, Energy Storage, Nonlinear Control, Experiments 1. Introduction Recently, energy storage flywheels with various capacities are becoming reality as a new application field of magnetic bearings. Energy storage flywheels are being energetically developed in Japan, U.S., and Europe, and some flywheel batteries are already marketed for uninterruptible power supply systems (UPS). In Japan, the research of the energy storage flywheel was started in the early 1990s by the feasibility project in NEDO. This project is based on the large vision of leveling the electricity loads in day and night by providing 10 MWh-class energy storage flywheels to about transformer station facilities in Japan so that nighttime residual electric power is used to rotate the flywheel in order to convert electric energy to kinetic energy for storage and the stored kinetic energy can be converted to electric energy when the daytime demand of electric power is at the peak. As a basic research for this, a research for flywheels with both small capacity and medium capacity has been performed which uses, for the purpose of improving the flywheel efficiency, not only super-conducting magnets but also magnetic bearings. In this project, a research for realizing a flywheel system or a magnetic bearing with high efficiency has been energetically performed. This paper describes recent trends of the control of a magnetic bearing and the application development of a magnetic bearing to an energy storage flywheel system. 2. Recent Trends of Magnetic Bearing Control The largest advantage obtained by the use of a magnetic bearing is that the magnetic bearing provides non-contact support which eliminates friction and abrasion. A magnetic bearing also has various advantages such as higher cleanliness, remarkably improved power transmission efficiency, or higher rotational speed. This allows a magnetic bearing to be frequently used by a turbo-molecular pump or a clean carrier vehicle in IC manufacturing equipment. Many researches on the leading-edge control of a magnetic bearing such as the sliding mode control 1), 2), 3), 4), the method based on a exact linearization 5), 6), 7), H control 8), and lcontrol 9) were reported in the past. On the other hand, a magnetic bearing has disadvantages such as generally high cost and the necessity of a control circuit for stabilizing an unstable system, except for types such an already-commercialized turbo-molecular pumps. Thus, a magnetic bearing has various problems in order to be used more widely. One of them is the consumption of much electric power because many magnetic bearings use electromagnets for their actuators. In a conventional magnetic bearing, a rotor passes through the changing magnetic flux and thus the rotor has therein eddy current. Although a normal magnetic bearing system causes no problem by the loss of the eddy current, this rotation loss requires to be reduced in order to realize an energy storage flywheel. In order to solve this problem, the authors suggested "zero power non-linear control magnetic bearing 10) ". The reason why the zero power non-linear control magnetic bearing uses a term "zero power" is that it uses only an 3

2 electromagnet without using a permanent magnet and the equilibrium status causes no flow of current (power). In other words, in the zero power non-linear control magnetic bearing, a bias current is not supplied and only control current is supplied to realize the stabilized levitation. Therefore, at the equilibrium point, the control current is zero and thus the equilibrium point theoretically provides completely zero power. When the magnetic bearing moves away from the equilibrium point, this also can be defined the zero bias control. Zero power control is also substantially realized because the rotor locates in the vicinity of the equilibrium point in high speed rotation range owing to rotor's self equilibrium action or gyroscopic effect. The magnetic bearing zero power control at the early stage had an objective of omitting bias current by the use of permanent magnet. The authors realized a zero power magnetic bearing system using both of permanent magnets and electromagnets by discrete time sliding mode control 11). This magnetic bearing uses an actuator with the combination of permanent magnets and electromagnets to substitute the suction power by the permanent magnets for the power generated by the bias current. Thus, this system does not need bias current and can reduce a significant amount of consumption of electric power as compared to the case of a normal magnetic bearing. However, the actuator with the combination of electromagnets and permanent magnets requires a complicated structure as compared to a magnetic bearing using only electromagnets, causing a problem of higher cost. Furthermore, this system is common to the conventional ones in that the rotor passes through the magnetic flux during the rotation. This leaves the problem of the eddy current unsolved and a problem of heat generation or the like. The non-linear magnetic bearing system which does not supply bias current has been researched concerning a sliding mode control by Charara et al. with regards to the magnetic bearing system having an input/output linearization 12) and the 13), 14), 15) output feedback control by the authors based on the backstepping procedure. However, the existence of unbalanced vibration or the like has caused a problem in which a levitation body does not reach the equilibrium point and control current is supplied only in accordance with the amount of inclination compensation of a rotor caused by the whirling vibration or the gravitational force. Thus, a completely zero power control has not been realized yet. 3. Energy Storage Flywheel There have been important tasks such as the reduction of exhaust gas CO 2 for preventing the global warming and the recycling of energy for protecting the environment. Thus, the commercial energy storage flywheel as one method for providing recycling renewable energy has been intensively researched and developed in Japan, U.S., Netherlands, Italy, Germany, and other European countries. Many papers have been reported concerning the energy storage flywheel, including: the ones by Netherlands for flywheel energy electric power storage project 16) for the windpower generation and the project regarding an energy storage flywheel to be mounted on an electrically-driven bus and a hybrid bus 17) ; and the one by U.S. NASA for developing the for-space-satellite energy storage flywheel 18) and the one for developing a flywheel uninterruptible power supply system (Flywheel UPS) 19). From the above viewpoint, also in Japan, researches on the energy storage flywheel have been performed at Shikoku Research Institute Inc., Ishikawajima- Harima Heavy Industries Co., Ltd., Koyo Seiko Co., Ltd., Chiba University, Tokyo Electric Power Company, Inc., Mitsubishi Electric Corporation and others in accordance with the NEDO project. Energy storage flywheel systems of various capacities have been researched ranging from a small capacity to an extremely large capacity. The one having an extremely large capacity has been researched for realizing a system with a single unit system having a capacity of 10 MWh as described in the above introduction. However, such a system is sometimes said to be difficult to be realized by the current science and technology level. Thus, a method is being researched in which a system with a predetermined capacity is intended to be realized not by a single system having a predetermined capacity but by the use of a number of systems having a small capacity. In the case of a single unit system, only the one unit needs to be controlled and thus the structure of the control system itself may be simpler but there have been actually many challenges regarding the system design, manufacture, transportation, maintenance, safety, reliability or the like. For the realization of a system having a predetermined capacity using a number of systems with small capacity, although the existence of many systems to be controlled makes it difficult to control the systems to be connected, it is considered easy for an individual system to be designed, manufactured, transported, and maintained. From the past researches, it has been considered that energy storage for few hours or more is impossible due to the loss by the rotation. Judging from the strength of the flywheel material, it has been also said that the high-speed rotation is impossible. As a result, it has been considered that a system having a large capacity is very difficult to be designed 20), 21). Thus, a flywheel system having a small capacity has been researched and reported with regards to the vibration analysis 22), the calculation of physical parameters 23), the development of a medium size flywheel system 24), the estimation of physical parameters and energy for a flywheel system having a large capacity 25), the concept design and the structure design of a flywheel system having large capacity 26), and the detailed research for a flywheel system having small capacity 27). The present situation is that experimental researches on the small capacity system have solved many problems and the next step of the development of a medium capacity system is performed. Solving the problem for the medium capacity may lead to the research and development of a large capacity system. It is especially required to research on the problem of elemental technology for realizing such a system, e.g. the 4

3 specification and the structure of a flywheel body, the production engineering, the levitation (magnetic bearing) system, the motor generator, the highly-efficient energy conversion technique, and the protection for emergency 28). On the other hand, the authors have been intensely working on the research on the dynamics of an energy storage flywheel system 29), H control 30), sliding mode control 31), distributed control 32), lcontrol 33) or the like. 4. Zero Power Non-linear Control-type Energy Storage Flywheel As conventional magnetic bearing causes bias current resulting in energy consumption even when the rotor is in the equilibrium condition, application of such a conventional bearing to an energy storage flywheel results in undesirable overall energy storage efficiency. A conventional magnetic bearing allows the rotor to pass through the changing magnetic flux and thus causes the rotor to have therein eddy current. A normal magnetic bearing system has no problem by the loss due to the eddy current but, to realize the energy storage flywheel, this rotation loss requires to be reduced. Due to this reason, the zero power control is applied to the energy storage flywheel system to design the zero power-type energy storage flywheel system. The authors have used ANSYS to design a vibration analysis model having a gyroscopic effect with regards to the 10 MWh-class high temperature superconducting magnetic levitation flywheel system and performed a vibration analysis having the gyroscopic effect. Next, the authors have developed a one-dimensional finite element method model based on the result of the vibration analysis to subject the model to Cholesky factorization for mode separation so that a zero power compensator is designed. Then, simulation was performed to check from the result of the simulation whether the control input is zero in order to consider the feasibility of the zero power control 34). However, there are still many problems for the realization of the 10 MWh-class high temperature superconducting magnetic levitation flywheel system. Thus, based on the design procedure of the 10 MWh-class high temperature superconducting magnetic levitation flywheel system, the zero power control system having a small capacity magnetic bearing flywheel system was designed. By a simple calculation based on the weight of the rotor and the number of rotations, the system is defined as a 0.5 KWh-class magnetic levitation flywheel system. The rotor section of the 0.5 KWhclass magnetic levitation flywheel system used in this research is shown in Fig. 1. Figure 1 (a) shows the rotor section attached to the 0.5 KWh-class flywheel. The black component in Fig. 1 (a) is a flywheel with a CFRP-made rotor having the diameter of 400mm and the height of 40mm. Figure 1 (b) shows the appearance of an experiment device having the height of 0.65 m and the diameter of 0.6 m. In order to control the radial direction, a radial magnetic bearing unit is provided. A gap sensor provided in the radial direction is used to detect the displacement in the radial direction. In addition to the gap sensor and the electromagnet, a protection bearing is also provided. The gap between the magnetic bearing and the rotor has the length of 250 lm and the distance from the protection bearing is 100 lm. This system is a test device for the basic research by NEDO International Collaborative Research (Koyo Seiko, Co., Ltd., Chiba University, and CCM). This project plans to realize the zero power magnetic bearing-type flywheel within this year in cooperation with CCM company of Netherlands. This project also plans to mount the zero power magnetic bearing-type flywheel to a bus in the future. Table 1 shows the values of various parameters of the system of Fig. 1. (a) Rotor (b) Appearance Fig KWh class flywheel system Table 1 System parameters and values Item Value Mass of the rotor, kg 4.85 Mass of the flywheel, kg Constant of upper magnetic attractive force, Nm 2 /A 2 Constant of Lower magnetic attractive force, Nm 2 /A 2 Nominal air gap of x direction, m Nominal air gap of y direction, m Moment of inertia about x and y axis, kgm 2 Moment of inertia about z axis, kgm 2 Distance from the center of gravity (upper), m Distance from the center of gravity (lower), m

4 This system was subjected to a vibration analysis for the case of a rigid mode and an elasticity mode, the case in which the gyroscopic effect is ignored, and the case in which the gyroscopic effect is considered. The result of the vibration analysis was used to develop a control model so that the sliding mode control and the H control are performed to consider the stability of the closed-loop. Concerning the design of the zero power compensator of this system and the results of the simulation and the experiment, see the papers 35), 36). It is known that the conventional PID control requiring a bias current always supplies the current of about 1.5 A. Thus, the above system provides the same performance with about 25% of the current of the conventional PID control. Specifically, the above system results in the reduction to 6.25% of power. This research also plans to further increase the rotational speed so that a few suggested methods are examined and considered. 5. Future Problems and Perspective The authors have already succeeded in the experiment of the non-linear control with zero power and zero loss by only the normal conductive magnetic bearing. This is particularly innovative in that the control performance enabling the reduction to 6% of the amount of energy consumption of the conventional type magnetic bearing. Thus, this can be considered as an innovative technique. This success allows the high-efficiency energy storage flywheel to be more likely. It has already demonstrated that this technology can be applied to a system or elastic rotor having a strong gyroscopic effect. It is expected to realize further improvement of the performance by an ultra high-speed rotation and a completely zero power control. Methods for designing the zero power non-linear control system include the researches of "frequency shaped non-linear robust control", "LMI base gain schedule type H control 37) ", and "sliding mode control combined with this 38) ". The system having the gyroscopic action is a time modulation system in which the natural frequency is the function of rotational speed. Thus, the control by one fixed compensator cannot provide superior characteristics for all regions of rotational speed. On the other hand, the gain schedule type compensator always has an optimal performance for all regions. Thus, it is considered that a gain schedule type H control based on Linear Matrix Inequality (LMI) or a sliding mode compensator with this hyperplane is designed to realize a zero power control having higher performance. Furthermore, it would be also effective to allow a system having a large unbalanced vibration to be used with "unbalanced vibration adaptive suppression control by unknown disturbance frequency estimation type adaptive algorithm for eliminating model 39), 40) ". Today, it is expected that the energy storage flywheel battery is used to an energy storage flywheel type uninterruptible power supply device, electric or hybrid cars, train compartment or the like. It is also increasingly expected that a lead battery having a shorter service life and a higher cost for the disposal is substituted by a safe and pollution-free battery that provides an electric power storage with a large capacity and that can be used semi-permanently. 6. Conclusion This paper described the new trend of the magnetic bearing control and the application to the energy storage flywheel. "Research for power storage flywheel" is an age-old problem and the combination of a magnetic bearing and a flywheel has been tried over and over again. However, the immature elemental technology has prevented the power storage flywheel from success. However, today's environment is totally different; various elemental technologies have been remarkably improved and a great proposition of the protection of global environment has been highly publicized. A historical ambition of "magnetic bearing-type energy storage flywheel" has been already realized and it is considered that conditions for the development and commercialization of the magnetic bearing-type energy storage flywheel also have been almost satisfied. It is expected that the magnetic bearing-type energy storage flywheel will be further researched and developed for providing a new big market for a magnetic bearing. References 1) H. Tian, K. Nonami, Tans. Japan Society of Mechanical Engineers, C, 60, 580 (1994) ) A. E. Rundell, S. V. Drkunov and R. A. Decarlo, IEEE Trans. on Control Systems Technology, 4, 5 (1996) Tans Japan Society of Mechanical Engineers, C, 58, 545 (1992) ) K. Nonami, H. Yamaguchi, Tans. Japan Society of Mechanical Engineers, C, 58, 545 (1992) ) H. Tian, K. Nonami, Tans. Japan Society of Mechanical Engineers, C, 60, 569 (1994) ) R. D. Smith and W. F. Weldom, IEEE Trans. on Magnetics, 31, 2 (1995) ) D. L. Trumper, S. M. Olson and P. K. Subrahmanyan, IEEE Trans. on Control Systems Technology, 5, 4 (1997) ) F. Matsumura, T. Namerikawa, A. Murata, Trans. IEE of Japan, 118, 4 (1997) ) W. Cui, K. Nonami, Tans. Japan Society of Mechanical Engineers, C, 58, 553 (1992) ) K. Nonami, Tans. Japan Society of Mechanical Engineers, C, 68, 666 (2002) ) K. Nonami, Y. Ariga Proceedings of the 11th Electromagetic Force Symposium (1999) ) K. Nonami, K. Nishina, M. Saito, Tans. Japan Society of Mechanical Engineers, C, (1996) ) A. Charara, J. D. Miras, B. Caron, IEEE Trans. on Control System Technology, 4, 5 (1996) ) Y. Ariga, K. Nonami, K. Sakai, Tans. Japan Society of Mechanical Engineers, C, 67, 654 (2002)

5 14) S. Sivrioglu, K. Nonami, A. Kobo, R. Takahata, 8th International Symposium on Magnetic Bearing, August (2002) Mito, Japan, ) Y. Ariga, K. Nonami, K. Sakai, Tans. Japan Society of Mechanical Engineers, C, 67, 654 (2002) ) 17) 18) D. A. Christopher, R. Beach, IEEE (1997) ) 20) Y. Miyamgawa, H. Kameno, R. Takahata and H. Ueyama, Preprint of ASC '98 in Palm Springs, CA (September 17, 1998) 21) K. Nakagawa, Proceedings of the R&D Project on High Temparature Superconducting Flywheel Energy Storage 5 (1998) ) R. Takahata, H. Ueyama, Y. Miyagawa, H. Kameno, H. Higasa, H. Ishikawa, Fifth International Symposium on Magnetic Bearings Kanazawa, Japan (1996) ) M. Ahrens, A. Traxler, P. Burg, G. Schweitzer, Fourth International Symposium on Magnetic Bearings, ETH Zurich (1994) ) R. Shimada, Bull. Res. Nucl. Reaction, 22 (1998) ) H. Higasa, T. IEE Japan, 113-B, 7 (1993) ) H. Borneman, IEEE Trans. On Applied Superconducting, 7-2 (1997) ) Takahata, H. Higasa, H. Ueyama, Y. Miyasawa, H. Kameno And F. lshikawa, T. IEE, 117-D, 9 (1997) , (in Japanese) 28) R. E. Horner, /96 (1996) IEEE, ) Edo, Kanki, Kawanishi, Proceedings of the JSME, (1998.3), ) Y. Zhang, K. Nonami, H. Higasa, Tans. Japan Society of Mechanical Engineers, C, 67, 660 (2001) ) Y. Zhang, K. Nonami, H. Higasa, Tans. Japan Society of Mechanical Engineers, C, 67, 662 (2001) ) J. Cao, L. Yu, Y. Xie, Seventh International on Magnetic Bearings (August 23-25, 2000) ETH Zurich, ) U. Schonhoff, J. Luo, G. Lu, E.Hiton, R. Nordmann, P. Allaire, Seventh International on Magnetic Bearings (August 23-25, 2000) ETH Zurich, ) Y. Zhang, K. Nonami, H. Higasa, Japan AEM, 10, 2 (2002) ) S. Sivrioglu, M. Saigo, K. Nonami, Proceedings of the 15th Electromagnetic Forces and Dynamics (2003) ) Y. Zhang, K. Nonami, Proceedings of the 15th Electromagnetic Forces and Dynamics (2003) ) S. Sivrioglu, K. Nonami: Tans. Japan Society of Mechanical Engineers, C, 63 (610) (1997) ) K. Nonami, S. Sivrioglu: Sliding Mode and Nonlinear Control, Springer-Verlag UK (1999) ) Z. Liu, K. Nonami: Tans. Japan Society of Mechanical Engineers, C, 66 (649) (2000) ) Z. Liu, K. Nonami: Tans. Japan Society of Mechanical Engineers, C, 65 (638) (1999)