Energy Systems Operational Optimisation Emmanouil (Manolis) Loukarakis Pierluigi Mancarella Workshop on Mathematics of Energy Management University of Leeds, 14 June 2016
Overview 1: Perspective What’s this presentation about? 2: Electrical Distribution Networks Management 3: Other Problems 2/18 (Overview)
What’s the Problem? Towards increased Inflexible energy efficiency & Demand reduced emissions… Distribution medium/low voltage radial networks Large-Scale Storage Distributed Generation Flexible Transmission Large-Scale Demand Renewables high voltage meshed networks Distributed Storage Conventional Generation Heating CHP Resources Coordination Over Time Gas Electrical Distribution Network Networks Management Heat Network Hot Water / Other Processes Optimising Gas (or Optimising Heat other fuel) Usage Networks Operation 3/18 (Perspective)
Current State of Play Unit Commitment (every 24h to <1h) • significant uncertainty Forecasts • integer variables Transmission Large Distribution Is this the right time to optimise… Network System • typically coupled with Generators … devices at the end -user level? reliability requirements Not really! Detailed Simplified Aggregate … other energy vectors? model / bids models model/bids Probably not in detail! Optimise! Operating status Economic Dispatch (every 15min) • limited number of discrete controls • contingency considerations Transmission Large Distribution Network Generators System • limited look -ahead Aggregate Detailed Detailed measurements models models Is this the right time to optimise… … devices at the end -user level? Optimise! … other energy vectors? If not now…when?! Operating state / Control mode / Power set-points Local Device Controls (instant) real-time Transmission Large Network Generators 4/18 (Perspective)
Extending Dispatch 4 Area 1 Large-scale 2 401 413 419 431 444 450 456 generation IGs Area 2 Transmission (conventional 402 403 414 415 420 421 432 433 445 446 451 452 457 458 & renewable) 404 405 416 417 422 423 434 435 436 447 448 453 454 459 460 • Very large scale! 1 • Uncertainty! 406 407 418 424 425 426 437 438 439 449 455 461 462 463 Transmission system TSOs 3 4 • Peculiarities of individual (multiple areas) 408 409 410 427 428 440 441 464 465 devices. 5 Distribution 411 412 429 430 442 443 466 467 Area 3 6 Bus Users Aggregate Demand Large-scale generation (conventional Need for one more Transmission & renewable) optimisation step! Transmission system (multiple areas) Distribution system infeasible infeasible (high/medium voltage feeders) energy power Distribution Distribution system curtailment (medium/low Microgrid voltage feeders) curtailment Individual Users (inflexible & flexible 2 4 6 8 10 12 2 4 6 8 10 12 demand / small scale time-step time-step renewables) 5/18 (Perspective)
A Step Further Unit Commitment (every 24h to <1h) Forecasts Transmission Large Distribution Network Generators System Detailed Simplified Aggregate model / bids models model/bids Disaggregating the network operators schedule… Optimise! Operating status Economic Dispatch (every 15min) Forecasts Microgrid (Local) Dispatch (every 1min) Large Transmission Distribution Microgrids Generators Network System Microgrids Users Flexible & Aggregating Aggregate Detailed OPF Network inflexible energy function / OPF models model / bids constraints offers / requests Optimise! Optimise! Operating state / Operating state / Control mode / Control mode / Power set-points Power set-points Local Device Controls (instant) Transmission Large Users Network Generators 6/18 (Perspective)
Microgrid Dispatch … or in other words: close -to-real-time distribution network management Objectives IEEE- 123… the ● follow a given power output (market signal) good old days ● serve customers! ● alleviate constraints violations IEEE- 123… in a test case with lots of EVs if left uncontrolled Requirements ● solution time … up to a few minutes Controls ● many discrete: tap changers, capacitor banks, loads … ● some continuous: small- scale generation, storage, some EVs… 7/18 (Distribution Networks Management)
Modelling Considerations (part 1) Point 1 Return currents not of interest → Kron’s reduction! Point 2 Symmetrical components → no Point 3 advantage in 1p/2p Constant power models → not good loads enough → go ZIP + VI formulation Non-linear! Non-convex! 8/18 (Distribution Networks Management)
Modelling Considerations (part 2) Point 4 If V in polar coordinates the energy balance (right part) is non- linear → use rectangular coordinates! Point 5 ● Voltage constraints → non-convex Still non-linear! 9/18 (Distribution Networks Management)
Modelling Considerations (part 3) non-linear exact curve real{ I } (p.u. c P ) 1 linear feasibility approximation region 0.8 0.6 0.8 0.9 1 1.1 1.2 voltage (p.u.) Approximation 2 outer 1 ● Approximate P-part, as approximation imag{ I } (p.u. I max ) 0.5 a ZI-part 0 -0.5 -1 -2 -1 0 1 2 real{ I } (p.u. I max ) Approximation 3 ● Imbalance / capacity bounds → linearize! Linear (assuming Z part is fixed)! 10/18 (Distribution Networks Management)
Modelling Considerations (part 4) Formulation Multi-time-step? stochastic? Formulation Single-time-step? Approximation 4 deterministic? 650 ● Modified utility function to prioritise demand 646 645 632 633 634 ● Follow the market 632A 632B power reference 632C 632D 632E 611 684 671 675 652 680 Point 6 ● Do we really need tight voltage bounds? 11/18 (Distribution Networks Management)
Does It Work? 650 646 645 632 633 634 632A Collect info Approximate problem at a 632B 632C 632D from smart given voltage reference 632E 611 684 671 675 meters frame 848 652 680 822 846 820 844 YES 818 864 842 Algorithm 1 Needs adjustment? 800 802 806 808 812 814 850 816 824 826 858 834 860 836 840 832 888 890 862 NO 810 838 852 828 830 854 856 Send energy schedules to devices Time (sec) IEEE-13 0.0176 0.0161 0.9945 0.17 250 32 30 29 300 0.0031 0.0024 0.9999 (0.25) 51 111 112 113 110 114 33 31 28 50 49 109 107 25 47 IEEE-34 48 46 108 26 0.0056 0.0168 0.9929 0.24 104 45 27 106 64 44 103 23 43 65 105 450 0.0003 0.0006 1.0000 (0.50) 24 102 63 100 42 41 66 101 21 40 99 IEEE-37 0.0002 0.0005 0.9998 0.23 62 71 98 22 39 70 35 97 38 18 69 36 0.0001 0.0001 1.0000 (0.43) 19 68 20 37 75 67 74 60 IEEE-123 0.0079 0.0099 0.9964 0.33 73 57 58 14 85 610 72 11 59 61 0.0012 0.0013 1.0000 (1.80) 79 10 9 56 77 78 55 54 2 53 52 76 80 13 94 8 7 84 149 1 96 34 88 12 90 81 92 17 86 87 15 89 91 93 95 3 5 82 83 6 16 4 12/18 (Distribution Networks Management)
Tap-Changers Approximation 5 848 ● Taps are continuous 822 846 820 844 818 864 842 800 802 806 808 812 814 850 816 824 826 858 834 860 836 840 832 888 890 862 810 838 Collect info Approximate demand & from smart taps at a given voltage 852 meters reference frame 828 830 854 856 Update Solve for state and trust-region taps (approximate) Iterations Solution Max. tap Power time (sec) rounding errors change YES Algorithm 2 IEEE-13 0.32 5 0.0029 -3.78% Needs adjustment? IEEE-34 0.89 4 0.0038 -5.16% NO 0.49 4 0.0030 -8.03% IEEE-37 IEEE-123 2.62 16 0.0031 -4.42% Send energy schedules to devices 13/18 (Distribution Networks Management)
Discrete Controls Approximation 5 ● Solve continuous relaxation → restricting deviations from nearest Mixed integer integral solution programming… Collect info Approximate demand & from smart taps at a given voltage Tap Added Solution meters reference frame EVs number time (sec) controls number Solve for state and IEEE-13 3 592 7.7 Update taps (approx. IEEE-34 9 316 4.3 trust-region continuous relaxation) IEEE-37 3 409 4.9 YES IEEE-123 9 623 14.9 Adjust Needs adjustment? penalty NO An feasible integral solution NO Algorithm 3 Is integral? Due to high number of was recovered… small controls… no YES significant difference Send energy schedules to between the continuous devices relaxation objective value… 14/18 (Distribution Networks Management)
Summing Up months/ minutes / min. sec. real-time years ahead hours ahead ahead ahead Model Model detail Model detail Model detail detail Uncertainty Uncertainty Not Uncertainty now! There are more problems out there! How should we solve them? Important! • Problem characteristics! • Solver characteristics! 15/18 (Distribution Networks Management)
Another Problem : energy district management (1) The problem… …optimising over time subject to network constraints from gas from / to electricity and detailed device supply network supply network and building models OTHER BUILDINGS / INSTALLATIONS HEAT / POWER GENERATION gas INSTALLATION network heat boiler network pump CHP heat exchanger Computational difficulties ● thermal network storage electrical capacity BUILDING network other gas ● thermal network dynamics demand building heating hot water renewables other electrical demand storage 16/18 (Other Challenges)
Recommend
More recommend
