This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government. SAND2018-9697C Energy Storage Design Considerations for an MVDC Power System L. J. Rashkinl, J. C. Neely", D. G. Wilson', S. F. Glover', N. Doerry 2 , S. Markle 2 , T. J. McCoy 3 Sandia National Labs, Albuquerque, NM, USA NAVSEA, PMS 320, Washington, D.C., USA 2 McCoy Consulting, Box Elder, ND, USA 3 Sandia National Laboratories VAL SEA S COMMAND Iw mArqkv,9 U.S. DEPARTMENT OF ENERGY V A k4 Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International, Mc., for the U.S. Department of Energys National Nuclear Security Administration under contract DE- NA0003525. The views expressed in the article do not necessarily represent the views of the U.S. Department of Energy or the United States Government.
Sandia Outline National Laboratories ■ Introduction ■ Model Development ■ System Layout ■ Control Algorithm ■ Profiles Load ■ Simulation Results ■ Energy Storage Requirements ■ Power and Energy Requirements ■ Frequency Analysis ■ Conclusions 2
Sandia Introduction National Laboratories ■ Controls are recognized as a primary challenge to fielding a medium voltage DC (MVDC) power system for future Navy ships ■ The service power demands of these future naval warships may include advanced mission systems which need large amounts of power in short pulses ■ Energy storage is a key component of shipboard MVDC architecture ■ Minimum sizing ■ Trade offs between performance and size 3
Scaled notional model of ship power system Sandia National Laboratories was developed • An electric ship is emulated using the Secure Scalable PM Gene 1.7 kW En Microgrid Testbed (SSMTB) M rid hardware at Sandia National Pulsed d DC Bus Vdc (SNL) d Laboratories E Vdc Microgrid 3 3 networked microgricls DC Bus rd 200 Vdc ridl 2 • Power electronics interfaces DC Bus • Agent based control PM Gene 23 • Repeatable experiment profiles • Simulink model library 4
Sandia Optimization control determines set-points National Laboratories based on system status • Based on guidance control algorithm Guidance Control • Reads terminal voltages, Load Estimator IDC-DC Power Command pp )1 Compute : output currents, and state of yor,o, s 2 I 4 rLiaikProf 2V; Pload ITES • Filter iload (V; pf0,- charge (SOC) for all sources ¡Li = • rl i V b • Estimates load A A . 0 * .* A 1 ,1 L1 • Determines power demands Diesel Generator 1 Programmable • Sets operational set points .* Loads Vb,2 ,1 b,2 , Vs2 11 2 ,1 L' Diesel for sources Generator 2 v cnr , Pulsed vb 5 3 5 lb , 5 +,71-1 1 3 Load .,-.Esl : ir • Filter used to determine Energy Storage 1 power balance between *** energy storage and generator resources 5
Load vignettes were obtained for different Sandia National Laboratories types of loads Mission Load #1 Mission Load #2 ,80 80 e• -- "••• 2 2 40 a) o 0 Propulsion Load 11111111 11111111 11 0 - 0 0 0 200 400 600 80 o 200 400 600 , Time (sec) Time (sec)+ 2 40 a) 80 o oo 200 400 600 2 4 0 Service Load a) 80 40 200 400 600 0 a_ Time (sec) 200 400 600 Total Load Time (sec) 6 A.M. Cramer, X. Liu, Y. Zhang, J. D. Stevens and E. L. Zivi, "Early-stage shipboard power system simulation of operational vignettes 201 for dependability assessment " 2015 IEEE Electric Ship Technologies Symposium (ESTS), Alexandria, VA, S_ pp. 382-387_
Potential shipboard load profiles are Sandia National Laboratories identified and positioned on power system ■ Propulsion load — 60 MW variable load ■ Split between Microgrid 1 and 2 ■ Service load — 20 MW variable load ■ Microgrid 3 ■ Mission Load 1 — 10 MW pulsed load ■ Microgrid 3 ■ Mission Load 2 — 700 kW pulsed load ■ Microgrid 3
Vignettes were scaled to hardware Sandia National Laboratories DC capabilities Total Power for 200 V bus 4.5 4 ,.,-,..,-,.., -, 3.5 • Original system S 3 , I ,_ 2.5 • 82 MW a) 2 • 20 kV 1.5 1 • SSMTB system 0.5 • 5 kW 100 200 300 400 500 600 • 200 V Time (sec) • Four load cases considered • Five filter constants considered • No pulsed loads • Mission load 1 only • 0.1356 sec • Mission load 2 only • 0.5299 sec • All Mission loads • 2.0049 sec • 7.5117 sec • 28.5643 sec 8
Load Responses were simulated for each Sandia National Laboratories controller time constant • Faster time constants result in more pulse delivery from generators • Slower time constants result in more pulse delivery from storage Controller time constant of 0.1356 sec Controller time constant of 28.56 sec 15 15 10 10 5 5 11111111- 0 0 11111111 -5 -5 - 10 - 10 — Pulsed Load - Pulsed Load — Bow — Bow Storage Storage — Starboard Storage — Starboard Storage - 15 - 15 — — — Port Storage — — Port Storage — Starboard Generator 260 220 — Starboard Generator — — — Port Generator 220 260 t — — Port Generator (sec) , - 20 - 20 t (sec), 0 200 400 600 0 200 400 600 Time (sec) Time (sec) 9
Load Responses were simulated for each Sandia National Laboratories controller time constant • Inertia and rate of power extraction govern speed variation • Faster time constant results in more spikes in speed Comparison of Generator Speeds 1000 Reference Speed (RPM) 980 Generator Speed Experiment 1 Generator Speed Experiment 5 960 Speed 940 920 900 0 100 200 300 400 500 600 t (sec) 10
Control effort is considered by a set of cost Sandia National Laboratories functions functions: Cost is determined by the If NG„, I (i b ,(r)—i b , dr j ) J 1 = i ) to N ES t f ( , 2 f I(iEsi(r)—iESI) dr J 2 = lo A 1 ) where i bi (t) are the currents delivered to the respective busses by the starboard and port generator converters as a function of time, G ens is the number of generators, and i Esi (t) are the bus N currents from the energy storage systems. N E s A 1 t r i b , = i b ,(T)ch t—T ia A 1ESi = 1 .i. i (r)dt ESi T fa t—T f where T fa is the period of the fast average. 11
System behavior is dependent on load and Sandia National Laboratories control filter 120 o No Mission Loads o Mission 1 Only Mission 2 Only 0 All Mission Loads Greater Filtering of Generator Power 0 Cornmand 8 c c 0 0 50 100 150 200 J 1 (Generator Control Effort Fast Average) 12
Sandia Energy storage power and energy National Laboratories requirements are determined from simulations 10 No Mission Loads Mission Load 2 —Mission Load 1 - - -All Mission Loads 10 4 10 5 Energy (Wh) 13
Energy Storage technologies vary in specific Sandia National Laboratories power / specific energy and frequency response Energy storage strategies vary in the technology used; each technology has different size/weight and performance capabilities, examples include: • Flywheel energy storage • Electrochemical Cells/Batteries (i.e. Lithium lon) • Super Capacitor Capabilities are usually identified over a range of values based on demonstrated systems [1] Energy Power Specific Specific Approx. Technology Density Density Energy Power Bandwidth (Wh/L) (W/L) (Wh/kg) (W/kg) (Hz) Flywheel 20-90 1000-5000 5-100 400-1500 20 Lithium-lon 150-500 1500-10000 75-200 150-2000 80 Super Cap 10-30 >100000 2.5-15 500-10000 80 [1] Xing Luo, Jihong Wang, Mark Dooner, Jonathan Clarke, Overview of current development in electrical energy storage technologies and 14 the application potential in power system operation, Applied Energy, Vol 137, 2015, pgs 511-536,
and system size are Storage technology Sandia National Laboratories a Ragone plot determined from 10 6 Fneumat,c bA 101.g , Flywheco Elastic Element 100 kg 1nternal Combust on kg * 1000 P-4 10,000 kg 10 2 Batte Fuel Cell 100,000 kg rz1-1 10 0 10 4 10 0 10 2 Specific Energy (Wh/kg) 15
Energy storage frequency requirements Sandia National Laboratories determined from chirp response of the system • Applied a log-sine chirp to the load on microgrid 3 • A sin (2n - f 0 (! ci i - /ti t) • Where f o is the initial frequency in Hz and f i is the frequency at time t1 in Hz • System input is load power on microgrid 3 • System output is the output power of the generators and energy storage systems • Frequency domain behavior of inputs and outputs are found using a fast Fourier transform (MATLAB fft function) 16
Recommend
More recommend
