137 JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY OF CHINA, VOL. 6, NO. 2, JUNE 2008 High Temperature Superconducting Magnetic Energy Storage and Its Power Control Technology Xiao-Yuan Chen, Jian-Xun Jin, Kai-Meng Ma, Ju Wen, Ying Xin, Wei-Zhi Gong, An-Lin Ren, and Jing-Yin Zhang Abstract⎯High temperature superconducting (HTS) power inductor and its control technology have been studied and analyzed in the paper. Based on the results of simulations and practical experiments, a controlled release scheme has been proposed and verified for developing a practical HTS SMES prototype. Index Terms⎯Electronic switch, SMES, high temperature superconducting (HTS) inductor, uninterruptible power system (UPS). 1. Introduction The power inductor energy storage technology has important applications in the modern scientific and technical field, i.e., high-energy physics, high-energy laser, electromagnetic propulsion, etc. Superconducting magnetic energy storage (SMES) devices can store the excessive electronic energy as electromagnetic energy in the superconducting inductor and release the stored energy if required. The advantages of SMES devices comparing with other energy storage devices include high energy storage density, high energy storage efficiency, long application life-time and few environmental pollution. With the development of applicable high temperature superconducting (HTS) materials, SMES technology has been progressed actively and is expected to apply in commercial applications[1]-[4]. In this paper, a HTS inductor design is introduced, with its geometry design and magnetic field analysis presented. Based on the results of simulations and experiments, a controlled release scheme is proposed to discharge the Manuscript received March 1, 2008; revised April 1, 2008. X.-Y. Chen, J.-X. Jin, K.-M. Ma, J. Wen are with the Center of Applied Superconductivity and Electrical Engineering, University of Electronic Science and Technology of China, Chengdu, 610054, China (e-mail: cxy_yjs@yahoo.com.cn, jxjin@uestc.edu.cn and hanqinmount@sina.com. cn 83201229). Y. Xin, W.-Z. Gong, A.-L. Ren, and J.-Y. Zhang are with Department of Development of Innopower Superconductor Cable Co. Ltd., Beijing, 100176, China (e-mail: yingxin@innopower.com, gong_weizhi@ innopower.com and ren_anLin@innopower.com). stored energy steadily and a simulation analysis on an uninterruptible power system (UPS) is also presented to verify the controlled release scheme for SMES applications. 2. HTS Inductor Design In this section, three HTS solenoid coils with different configurations are designed and analyzed. A common configuration of HTS solenoid coil is shown in Fig. 1. 2.1 Specifications of HTS Conductor Practically, Bi-2223 multifilament HTS tape conductor is chosen to deign a HTS solenoid coil. Its main specifications are: width a is 4.23 mm, thickness b is 0.23 mm, critical current Ic is 100 A (77 K, 0 T), and critical current density Je is 10 kA/cm2. 2.2 Inductance Calculation A formula for inductance calculation is given by [5] and can be expressed as follows L = 2πμ 0 N c2 R15 T ( p, q) (1) where μ0 = 4π × 10−7 , parameter Nc is given by Nc = N ( R2 − R1 ) D (2) where N is the number of coil turns, and T(p, q) is the function of size ratio p(=R2/R1), q(=D/R1). According to (1) and (2), there is 2 ⎡ ⎤ N L = 2πμ0 R T ( p, q) ⎢ ⎥ . ⎣ ( R2 − R1 ) D ⎦ 5 1 (3) Considering the filling factor K for practical design, there is Nab = ( R2 − R1 ) DK . (4) The total length of conductors S is given by R − R1 ⎞ ⎤ ⎡ ⎛ S = N ⎢ 2π ⎜ R1 + 2 2 ⎠⎟ ⎥⎦ ⎣ ⎝ (5) 2.3 Design of HTS Solenoid Coils For practical design of HTS solenoid coil, three 138 JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY OF CHINA, VOL. 6, NO. 2, JUNE 2008 different coils with same inner radius R1 and section area (R2−R1)D are analyzed in this section. The main geometries of the three HTS solenoid coils are shown in Table 1. Fig. 3. Magnetic field distribution of the HTS coil in Case 2. D C. Case 3 Referring to (3) and (4), the inductance L3 in Case 3 is 1.53 H and the length of HTS conductors S3 is 2224 m. The magnetic distribution is shown in Fig.4. R2 R1 0.772 0.647 0.523 0.398 0.273 0.148 0.024 Fig. 1. Scheme of HTS solenoid coils. 0.732 0.586 0.440 0.294 0.149 0.074 0.024 Table 1. Main geometries of the three HTS coils R1 (mm) R2 (mm) D (mm) Case 1 81 162 81 Case 2 81 81 121.5 162 243 40.5 Case 3 Fig. 4. Magnetic field distribution of the coil in Case 3. A. Case 1 From Table 2, the size ratios (p, q) in Case 1, 2, and 3 are (2, 1), (1.5, 2), (3, 0.5), respectively. According to [6], T(p, q) in Case 1, 2, and 3 are 0.3290, 0.2046, and 0.5028, respectively. A solenoid coil having the size ratio of (2, 1) is a so-called Brooks coil, which can give a maximum inductance for a given length of conductor. In the design of Case 1, the HTS coil should be wound with a certain insulating layer. The total width and thickness of Bi-2223 HTS conductor with insulating layer are 6 mm and 0.6 mm, respectively. Therefore, a reasonable filling factor is 32.2%. Referring to (3) and (4), the inductance L1 in Case 1 is 1 H and the number of turns N1 is 2186. Referring to (5), the total length of HTS wires S1 is 1668 m. To realize some magnetic specifications of the designed coil, QuickField software is introduced. Assume that the current through the designed coil I1 is 100 A, the magnetic field distribution is generated as shown in Fig. 2. Line a 0.866 0.656 0.445 0.234 0.142 0.054 0.020 Fig. 2. Magnetic field distribution of the coil in Case 1. B. Case 2 The filling factor K1, the length of HTS wires S1, and the current through the designed coil I1 in Case 1 are also used to deign the coils in Case 2 and Case 3, with an aim to optimize practical inductor design. Referring to (3) and (4), the inductance L2 in Case 2 is 0.62 H and the length of HTS conductors S2 is 1390 m. The magnetic field distribution is shown in Fig. 3. From the designs above, the main specifications of the three deigned coils are summarized as shown in Table 3. Table 3: Main geometries of the three HTS coils Turns number Volume (m3) Inductance (H) Wire length (m) Case 1 2186 0.005 1 1668 Case 2 2186 0.004 0.62 1390 Case 3 2186 0.007 1.53 2224 From Table 3, a Brooks coil is chosen to obtain maximum inductance for HTS conductors with a given length. The required volume and HTS conductor length of the coil in Case 2 are minimum values in three coils, however the inductance in Case 2 is lowest. The inductance of the coil in Case 3 can obtain maximum inductance within the same section area of the three coils, but the required volume and wire length are maximum values, which may lead to largest placement volume and highest cost to fabricate a practical coil. 2.4 Magnetic Field Analysis The maximum flux density of the solenoid coil is normally located at the cross circle of the central plane of the solenoid coil and its inner cylindrical surface. Assume that the center coordinates of magnetic distributions in Fig. 2, Fig. 3 and Fig. 4 are (0, 0) and the coil was symmetrically placed around (0, 0), a line a from (−80, −100) to (−80, 100) is added to analyze the flux density distributions, as shown in Fig. 2. Parallel flux density Bx1 in Case 1, parallel flux density Bx2 in Case 2 and parallel flux density Bx3 in Case 3 are shown in Fig. 5. Bx (//c) decreases the HTS conductor Ic more severely than the B and By (⊥c). Quench occurs when the current value through the HTS coil reach or exceed the Ic value. It is essential to reduce Bx so that the operating current of a practical HTS coil can reach a larger value. In Fig. 5, Bx3 is B B B B B B B 139 CHEN et al.: High Temperature Superconducting Magnetic Energy Storage and Its Power Control Technology lowest and therefore the operating current can reach largest value in practical application. Bx1 Bx2 Bx3 10 5 0 −0.25 -0.25 00 Y (mm) 50 50 30.4 t (s) (b) 30 30.8 B i (A) 100 5 0 50 0 30.8 30.4 t (s) (a) 30 30 30.4 t (s) (b) 30.8 Fig. 8. Discharging current waveforms with different charging current I: (a) I=10 A and (b) I=100 A. 10 i (A) Based on the above analysis, conclusions can be made as follows: 1) The Brooks coil in Case 1 can obtain maximum inductance for HTS conductors with a given length, but its energy storage may not be maximum value because of its high Bx; 2) The HTS coil in Case 2 is better for cooling and requires shortest conductor length and lowest placement volume, but its inductance is a minimum value and not ideal for obtaining high energy storage; 3) Though the required volume and conductor length of the HTS coil in Case 3 are maximum values, its inductance and operating current can reach maximum values, which are more suitable to obtain maximum energy storage. 10 5 0 30.05 t (s) (a) 30 30.1 5 0 30 30.4 t (s) (b) 30.8 Fig. 9. Discharge current waveforms with different Rs: (a) Rs=10 Ω and (b) Rs=1 Ω. 3.1 Theoretical Analysis and Simulation The electromagnetic energy stored in the inductor E(J) can be described by (6) E = 1 / 2(Ls I 2 ) where Ls(H)is the inductance of the inductor, I(A) is the charging current. The current i (t ) in the discharging circuit can be described by t (7) where t is the discharging time, τ (=Ls/Rs) is the time constant. The discharging time becomes longer while τ increases. The principle of charging and discharging circuit is shown in Fig. 6. Ls is assumed to be the ideal inductor and Rs is the resistance of practical inductor. The inductor is charged while the switch S is on, and discharged through the diode D while S is off. Some simulation results are shown in Fig. 7, Fig. 8 and Fig. 9. The simulation results conclude that the discharging time becomes longer while Ls increases or Rs decreases, however it is independent of the charging current. 3.2 Experimental Test and Analysis Practically a circuit test platform is designed and built to verify the simulation results. A 0.2 H Cu inductor is chosen to analyze charging and discharging tests. The resistance of the inductor Rs decreases while the operating temperature decreases. Its value is 5.6 Ω and 1.6 Ω at room temperature and liquid nitrogen (LN2) temperature, respectively. The Cu inductor is charged by a 70 V power supply and discharged through the diode D while the switch K is on and off, respectively. The discharging waveform is shown in Fig. 10. The discharging current decreases exponentially and can not be used effectively, which matches with the theoretical analysis using (7). The operating temperature increases along with LN2 evaporation. The above simulation results are well verified by the relational curves among different parameters, as shown in Fig. 11. 40 i (A) 3. Inductive Energy Release τ 0 30.8 10 B − 30.4 t (s) (a) 5 Fig. 7. Discharging current waveforms with different Ls: (a) Ls= 0.2 H and (b) Ls=0.02 H. 100 100 Fig. 5. Bx distribution of the designed three HTS coils. i (t ) = Ie 30 i (A) −50 -50 10 i (A) 00 −0.5 -0.5 −100 -100 Ls D Fig. 6. The principle of a charging and discharging circuit. i (A) Bx (T) 0.25 0.25 Rs i (A) 0.5 0.5 S DC 20 0 0.032 0.5 t (s) 1.0 Fig. 10. Discharging current waveform (Rs=1.6 Ω). 140 JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY OF CHINA, VOL. 6, NO. 2, JUNE 2008 filter, R is the load. 50 im(A) t (ms) t(ms) 500 300 DC 10 100 1 3 Rs (Ω) 5 1 3 Rs (Ω) (b) (a) S3 Lf R Cf D2 D1 Fig. 12. A controlled release circuit. Based on the above simulation and experimental analysis, conclusions can be made as follows: 1) The charging current i(t) is directly related to the. charging current I, but the discharging time t is not affected by I; 2) The discharging time t becomes longer while Ls increases or Rs decreases, and it is expected to enlarge the inductance or reduce the resistance for practical applications. The circuit in Fig. 6 can not satisfy the requirements for practical applications, the power release control strategy should be developed to improve its performance. 4. A Controlled Release Scheme It has been verified that energy release without any control is useless and it is expected to develop a controlled release scheme for practical applications. Assume that the inductor can be discharged with a constant power P0(W) within a time period ts(s), the stored energy E(t) at t(<ts) can be described by E (t ) = E − P0 t . (8) When t=ts, the current through the inductor Is is P (9) Is = 0 v where v(V) is the voltage across the inductor. If the discharging current drops below Is, the inductor will be discharged with a reduced power, which depends on the discharging depth coefficient λ. λ can be described by P0 t s E S2 S1 5 Fig. 11. Relational curves among different parameters: (a) relational curve between t and R and (b) relational curve between the peak value of discharging current im and Rs. λ= Ls S 30 (10) Referring to (6) and (9), the discharging current i (t ) at S t S1 t S2 t S3 t (a) (c) (b) Notes: (a) Energy-charging state, (b) Energy-storing state, (c) Energy-discharging state. Fig. 13. The control voltage waveforms on S, S1, S2 and S3. From Fig. 13, the steady-energy release control circuit has three operating states: 1) Energy-charging state; 2) Energy-storing state; 3) Energy-discharging state. In order to achieve steady-energy release control, the discharging current IR (or discharging voltage VR) is compared with a reference current Iref (or reference voltage Vref). If IR > Iref, (or VR > Vref), the steady-energy release control circuit is operated in the energy-storing state and IR (or VR) decreases gradually; If IR < Iref (or VR < Vref), the steady-energy release control circuit is operated in the energy-discharging state and IR (or VR) increases gradually. Thus the stored energy is discharged steadily and a steady current output (steady voltage output) can be applied in practical applications. A steady-current release control simulation module is developed based on Matlab/Simulink. Assume that Ls=0.2 H, Lf=21 mH, Cf=47 μF, R=0.05 Ω , Iref=2A. From Fig. 14, the steady-current release time increases form 50 ms to 140 ms while the charging current increases from 10 A to 100 A. So the steady-current output is achieved and the discharging time increases in direct proportion to the stored energy in the HTS coil. Based on the above analysis, a controlled release prototype is developed in Fig. 15. t ts (11) Referring to the (8) and (10), the stored energy at any given time t can be described by Pt Eλ t E (t ) = E − t = 0 s (1 − λ ) (12) ts ts λ Based on the above theoretical analysis, a controlled release scheme is proposed to achieve steady release, as shown in Fig. 12. Ls is a HTS inductor, switches S1, S2, S2, S3 and diode D1, D2 are used to form the charging and discharging circuits, inductor Lf and capacitor Cf form a 2 1 0 i (A) v 1− λ 1− λ 0.4 0.5 t (s) 0.6 0.1 0.05 0 0.4 0.5 t (s) (a) 0.6 2 1 0 0.7 v (V) P0 v (V) i (t ) = i (A) any given time t can be described by 0.7 0.4 0.5 0.6 t (s) 0.7 0.4 0.5 t (s) (b) 0.7 0.1 0.05 0 0.6 Fig. 14. Steady-current discharging waveforms: (a) the charging current is 10 A and (b) the charging current is 100 A. 141 CHEN et al.: High Temperature Superconducting Magnetic Energy Storage and Its Power Control Technology because its control technique is rather mature. A chopper can play a key role in the charging and discharging process of HTS coil, and provide a stability voltage for the load. Chopper control algorithms can be operated in three operation states: energy-charging state, energy-discharging state and energy-storing state, as shown in Fig. 17. Each operation state can be achieved by turning on or off the two control switches in chopper. Dewar HTS coils T4 Current detector T1 T2 C A B LS T3 R Control board Load Cf Current detector Fig. 15. A controlled release circuit prototype. SMES 5. SMES Application in UPS 5.1 Application Scheme SMES is expected to be a promising application in UPS. A SMES-UPS scheme is shown in Fig. 16. AC power SMES (a) (b) (c) Fig. 17. The three operation states of chopper algorithms: (a) energy-charging state, (b) energy-discharging state, (c) energy-storing state. 5.2 Simulation and Analysis A SMES simulation module is established by using Matlab/Simulink, as shown in Fig. 18. Its main parameters are as follows: energy storage inductance L=5 H; the equivalent resistance of energy storage circuit R1 is 0.1 Ω; steady charging current I0 is 400 A; a three-phase 380 V output with its effective power of 50 kW loads in the three-phase parallel connection RLC circuit. L is discharged after charging for 0.1 s and its simulation waveforms are shown in Fig. 19. Load SMES Fig. 16. A SMES-UPS scheme. A voltage-type converter is chosen to design the circuit ter R2 Mosfet1 L Uref Pulses Gain2 Mosfet2 Saturation PI Vc1 C ter2 Discrete PI controller Discrete PWM generator Voltage regulator C Set current value 1 Vref(pu) Vabc(pu) Vd-ref (pu) Vabc-inv m Uref Pulses v Vab-inv A B C 50 kW 380 V rms 50 Hz Fig. 18. A SMES-UPS simulation module. 1/z ter3 Vab-load v Scope1 v Vdc Diode1 Diode 0 L1 OR Mosfet3 PWM generator1 ter1 NOT R1 Step Vdc1 SMES Discrete Ts=2e006s Vabc A a B b C c Measure A B C A B C LC filter IGBT inverter g A B C v 1000 0 −1000 Vab-load(V) 0 500 0 −500 0 1 1 1 t(s) t(s) t(s) (a) 2 2 V1(V) 1000 500 0 0 3 1000 500 0 0.1 Vab-inv(V) JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY OF CHINA, VOL. 6, NO. 2, JUNE 2008 3 1000 500 0 0.1 Vab-load(V) Vab-inv(V) V1(V) 142 2 1000 500 0 3 0.1 0.2 t(s) 0.3 0.2 t(s) 0.3 0.2 t(s) (b) 0.3 Fig. 19. Simulation waveforms in SMES module: (a) the waveforms of Vc1, Vab-inv and Vab-load, (b) enlargement of (a). 6. Conclusions A HTS inductor design and the principle of the charging and discharging of the inductor are studied and analyzed in the paper. The common power inductor can not be used in practical applications owning to the exist of resistance and the stored energy can hardly be utilized if energy release without any control, a steady-energy release control method has been proposed and verified by a UPS simulation. More analysis and SMES prototype design will be presented in near future. References [1] J. X. Jin, “High Tc superconducting materials for strong current applications: Approach at the first stage,” Journal of Electronic Science and Technology of China, vol. 5, no. 1, pp. 38-43, 2007. [2] J. X. Jin and L .H. Zheng, “Development and applications of high temperature superconducting material,” Journal of University of Electronic Science and Technology of China, vol. 35, no. 4, pp. 612-627, Aug. 2006 (in Chinese). [3] J. X. Jin, Z. G. Wang, J. Wen, Y. G. Guo, and J. G. Zhu, “High temperature superconducting energy storage techniques,” Journal of the Japan Society of Applied Electromagnetics and Mechanics, vol. 15, Supplement, pp. 108-111, Sep. 2007. [4] J. X. Jin, C. Grantham, Y. C. Guo, J. N. Li, R. Bhasale, H. K. Liu, and S. X. Dou, “Magnetic field properties of Bi-2223/Ag HTS coil at 77 K,” Physica C, vol. 278, pp. 85-93, Sep. 1997. [5] J. X. Jin, C. Grantham, S. X. Dou, H. K. Liu, Z. J. Zeng, Z. Y. Liu et al., “Electrical application of high Tc superconducting saturable magnetic core fault current limiter,” IEEE Transactions on Applied Superconductivity, vol. 7, no. 2, pp. 1009-1012, 1997. [6] J. X. Jin, S. X. Dou, C. Grantham, and H. K. Liu, “Preparation of high Tc superconducting coils for consideration of their use in a prototype fault current limiter,” IEEE Transactions on Applied Superconductivity, vol. 5, no. 2, pp. 1051-1054, 1995. [7] J. X. Jin, C. M. Zhang, and Z. M. Li, “A Power Inductor Energy Control Technique,” 2008 IEEE International Conference on Industrial Technology (ICIT 2008), WA2-B1 (CD-Rom), Sichuan University, Chengdu, China, 21-24 April 2008. [8] S. W. Wu, “Inductance tables of air-cored cylindrical coil,” Journal of Zhengzhou University (Engineering Science), vol. 24, no. 3, pp. 106-112, 2003. Xiao-Yuan Chen was born in Jiangxi Province, China, in 1986. In 2007, he received the B.S. degree from the Chengdu University of Technology. He is currently pursuing the M.S. degree with UESTC. His research interest is in high temperature superconductivity. Jian-Xun Jin was born in Beijing, in 1962. He received B.S. degree from Beijing University of Science and Technology in 1985, M.S. degree from University of New South Wales, Australia in 1994, and Ph.D. degree from University of Wollongong, Australia in 1997. He was a research fellow and Australian ARC project chief investigator and senior research fellow with Australian University of Wollongong from 1997 to 2003. He is currently a professor and the Director of the Center of Applied Superconductivity and Electrical Engineering, UESTC. His research interests include applied high temperature superconductivity, measurement, control and energy efficiency technology. Kai-Meng Ma was born in Sichuan Province, China, in 1977. In 2001, he received the B.S. degree from Tianjin Polytechnic University. He is currently pursuing the M.S. degree with University of Electronic Science and Technology of China (UESTC). His research interest is in superconducting magnetic energy storage technology. Ju Wen was born in Hunan Province, China, in 1982. In 2005, he received the B.S. degree from Henan University of Science and Technology. He is currently pursuing the M.S. degree with UESTC. His research interests include control theory and control engineer. Ying Xin was born in Heilongjiang Province, China, in 1953. He received the B.S. degree from Tianjing University, Tianjin, in 1991, and the Ph.D degree from University of Arkansas, in 1991. He is now working in Innopower Superconducting Cable Co., Ltd. Dr. Xin is active in high temperature superconductivity. Wei-Zhi Gong was born in Province, China. He received the B.S. degree from the University, His research interest is the electrical application of the superconductor. An-Lin Ren was born in Hebei Province, China, in 1969. He received the B.S. degree from the Tianjin University, Tianjin, in 1991. He is now working in Innopower Superconducting Cable Co., Ltd. His research interests include the electrical application of the superconducting. Jing-Yin Zhang was born in Hubei Province, China, in 1979. He received the B.S. degree in 2001 in mechanical engineering and M.S. degree in 2006 in electrical engineering, both from the Beihang University (BUAA), Beijing. His research interests are superconductor applications in power grid.