Joint ELECTRA/SIRFN Workshop

Transcription

1 Joint ELECTRA/SIRFN Workshop October 24 th 2016, Niagara, Canada Web-of-Cells Concept and Control Scheme Helfried Brunner Technical Coordinator IRP ELECTRA This project has received funding from the European Union s Seventh Framework Programme for research, technological development and demonstration under grant agreement no

2 Electra Scope Real-time voltage and frequency control ( balancing ) for the future (2035+) power system Novel functional architecture incl. new concepts for network observability and robust controllers that act across different control boundaries 2

3 Trends and Assumptions Generation will shift from classical dispatchable units to intermittent renewables Generation will shift from few large units to many smaller units It will shift from central transmission system connected generation to decentralized distribution system connected generation Electricity consumption will increase significantly 3

4 Trends and Assumptions Large amounts of fast reacting distributed resources (can) offer reserves capacity Electrical storage will be a cost-effective solution for offering ancillary services Ubiquitous sensors will vastly increase the power system s observability Developments in Information and Communication Technologies will support the pathway towards more decentralized managed power systems 4

5 Trends and Assumptions From: Transmission grid connected dispatchable synchronous generators with downstream power distribution To: Large number of small intermittent RES generators that are distributed everywhere (all voltage levels) Increasing electrical loads (at medium/low voltage levels) and active control of flexible loads Reverse powerflows, local congestions, local voltage problems 5

6 The ELECTRA DoW Proposal TSO centric + improved TSO-DSO coordination Vertical Integration of horizontal distributed control schemes Early consultation with ETP SG (jul 2014) Do not underestimate the role of the DSO (think beyond TSO centric and TSD-DSO coordination ) Emphasize role of (grid-connected) microgrids and distributed storage Think out of the box ELECTRA De-centralized Web-of-Cells (WoC) concept 6

7 Web-of-Cells Concept Divide the power system (grid) in smaller entities (geographical areas) cells with local observability and control by a cell operator that is responsible for the real-time control of the cell Local problems are solved locally, in a secure manner, without system-wide communication, bottom-up aggregation and central decision making Cells are connected with each other via tie-lines (one or multiple, radial or meshed) Neighboring Cells can support each other in a autonomous distributed collaborative way (adjacent central aggregation) Neighboring cells can decide on local activation optimization (neighbor-to-neighbor central) 7

8 Web-of-Cells Concept Cells can contain/span multiple voltage levels Cells are dimensioned in relation to Computational complexity of Detection and Resolution (secure dispatching of reserves) Sufficient reserves providing resources Spatial correlation of weather forecasting for RES Cells do not need to be self-reliant for matching demand with supply They may depend on structural energy imports or exports (e.g. coming from large central RES power plants) as cleared in a system-wide optimized setpoint calculation They receive a setpoint (an import/export profile) as a reference for the real-time control (tie-line power exchanges) 8

9 Web-of-Cells Concept? 9

10 Voltage Control in WoC Voltage control: local by nature Detection is local Local reserves must be activated In each cell : Pilot nodes, AVR nodes with PVC (droop) control, and nodes with discontinuous voltage regulation (e.g OLTCs) Post-Primary Voltage Controller (PPVC): Optimal Powerflow Calculation (losses, security, robustness) to determine: Safebands for pilot nodes (security and robustness) Voltage setpoints for nodes (continuous and discontinuous) Droop settings for AVR nodes Regularly updating settings based on updated cell state forecast (proactive) or if safeband violation (corrective) 10

11 11

12 12

13 Objective : System balance restoration (load = generation) frequency is just an observable LFCA1 Frequency/Balance Control Current Control Scheme LFCA2 LFCA3 1. Frequency Containment Control FCC: Frequency Contain frequency deviation with (slow) inertia bearing generators Collaborative and global 2. Frequency Response Control FRC: Tie-line Powerflow and Frequency Restoring system balance and frequency Responsibilizing ( polluter pays ) Primary Trigger = system imbalance observed through frequency (aggregated deviations: imbalance netting!) local issues Secondary Trigger = LFCA imbalance (tie line powerflow versus plan/reference) 13

14 Frequency/Balance Control Current Control Scheme Frequency is/was a convenient observable (but inertia is declining, DC is coming, etc.) Many local imbalances causing congestions may exist at distribution grid level without a corresponding frequency deviation ( no FCC activation): imbalance netting How to ensure that reserves activations using distribution grid connected resources in response to a global observable (frequency) does not cause other, additional problems? How to make effective use of local resources to solve local problems locally? 14

15 Frequency/Balance Control Current Control Scheme Challenge 1: Central detection of the need for reserves activations Imbalance netting hides the many local problems (only considers the aggregated problem) Challenge 2: Secure and efficient activations of distribution grid connected reserves providing resources What, and how much, can be activated where, so that no new local voltage or congestion problems are caused by these activations Improve distribution grid observability/monitoring Improve TSO/DSO coordination But: Communication/aggregation complexity and delays (bottom-up and top-down) Central trade-off between security, efficiency and computational tractability 15

16 Balance Restoration Control In Web-of-Cells Divide-and-conquer (smaller cell large LFCA) : secure and efficient in computational tractable time Avoid communication/aggregation complexity and delays during real-time control Solve local problems locally based on local observables, acknowledging that: Cells have tie-line powerflow setpoints (schedules) Deviations are observed and corrective actions are taken using local (intra-cell) reserves detailed local information is needed and available to activate securely and effectively System balance is restored as the aggregated effect of restoring all cell balances No imbalance netting : security cost add Balance Steering Control Responsibilizing, the polluter pays But local collaboration possible 16

17 In each cell : Balance Restoration Control In Web-of-Cells Balance Restoration Controller (BRC): monitor and restore cell tieline powerflow profiles (net import/export) to centrally cleared secure values (representing system balance) Leveraging many fast acting resources (flex loads and storage instead of synchronous generators): very high ramping rates Cell imbalances = deviations from planned import/export profile caused by: Intra-cell incidents or forecast errors Intra-cell reserves activations for FCC or PVC PVC: unavoidable FCC: avoid, or smart/adaptive (based on cell state), or only in selected cells Deviations in neighbouring cells (physical connections) allows for local collaborative balance restoration effort based on powerflows ( based on global observable like frequency) 17

18 Still needing FCC? BRC can be primary (and secondary combined)? Leverage opportunities of distributed storage and flexible loads Can act much faster than the inertia bearing generators that were used before Can restore faster than current FCC can contain? Very high ramping rates + ICT : fast enough? No longer need for separate primary and secondary control Or Adaptive FCC Balance Restoration Control in Web-of-Cells Avoid frequency deviation triggered activations in Cells that are in balance Cells that are not causing the deviation: responsibilisation Local collaboration full responsibilisation secure activation (not causing new/additional problems like voltage problems or congestions) Introduce locality and proportionality by adding smarter controller paradigms 18

19 Balance Steering Control in Web-of-Cells BRC loses the benefit of imbalance netting: excess amount of reserves activations Balance Steering Control (BSC) introduces distributed local (neighbour to neighbour) imbalance netting where neighbours agree on a modified but still secure setpoint Power BRC possible ramping rate : fast acting resources without inertia (but with communication and calculation delay) Needed ramping rate to contain freq deviations (can be influenced by IRPC) Fast BRC without BSC DP, D Fast BRC with corrective BSC (undoing activations) Needed amount of activations if no imbalance netting (BRC without BSC) DP, D Inertia Steering Control/ Inertia Response Power Control (IRPC) provides a stable amount of (virtual) inertia irrespective of actual energy mix Slow BRC with proactive BSC (preventing activations) Needed amount of activations if local imbalance netting (BRC with BSC) Theoretical needed amount of activations if system-wide imbalance netting Time 19

20 20

21 21

22 22

23 23

24 24

25 WoC vs. Microgrids A cell is by design not a microgrid In ELECTRA, microgrids are defined as being able to operate in grid-connected as well as island -mode Being able to operate in island mode is not a requirement of a cell. but of course microgrids can easily be a cell once gridconnected, independent from the size However, cells will more and more have attributes of microgrids. This will lead to future integrated grids that are a combination of Cells that to a great extent, can operate in island-mode as well meeting the needs of the identified essential loads with available cell resources 25

26 Summary Web-of-Cells Concept ELECTRA Decentralized Web-of-Cells concept Load Frequency Control Areas smaller cells that are responsible both detecting the need for reserves activations as well as for the reserves activation itself Local control and collaborations between cells based on local observables (tieline powerflows) instead of global collaborative control based on global observable (frequency) delegate responsibility for local balance/frequency and voltage control to local cell operators Solve local problems locally : less complexity, less communication, more efficiency (less losses), more secure (less reverse power flows) Divide-and-conquer : more optimal and more secure within computational tractability limits 26

27 Summary Web-of-Cells Control Scheme VOLTAGE CONTROL CURRENT GRID Primary voltage control (PVC) Secondary voltage control (SVC) Tertiary voltage control (TVC) FUTURE GRID Primary voltage control (PVC) Post-primary voltage control (PPVC) CURRENT GRID FUTURE GRID - Inertia steering control BALANCE CONTROL Frequency containment control (FCC) Frequency restoration control (afrc) Frequency replacement control (mfrc) Adaptive Frequency containment control (FCC) Balance restoration control (BRC) Balance steering control (BSC) 27

28 Lab-scale Validation Experimentally implement Web of Cell (WoC) based distributed real-time control in a number of respected European laboratories Demonstrate the effectiveness of distributed controls in relation to a number of grid scenarios Prove the role of the Smart Grid Architecture Model (SGAM) in the setting of experimental plans and cooperation of multiple partners Investigate the local coordination of numbers of devices when subject to uncertainty in system operation while maximizing the effective utilization of flexibility Compare performance demonstrated across multiple laboratories Understand on the basis of experiments the implications of controller conflict(s) and the relative merits of different controls 28

29 Methodological approach Methodology Involved partners and laboratories 29

30 Discussion Point: How to validate with impact? 1. What are the major challenges for the validation of distributed control approaches (like the WoC real-time control approach) in the domain of smart grids? 2. What is the main advantage of a laboratory experiment (hardware, software, Hardware-in-the- Loop (HIL)) over a pure software simulation? 3. What is the value in experimental teams using SGAM? 4. What key features should be included in scenarios that stimulate genuine interest? 5. What aspects of smart grid systems evaluation are more important to represent in hardware rather than emulated in software models? 6. How do you assess if a Technology Readiness Level (e.g. TRL6: validated in operational environment) has actually been achieved? 30

31 CONTACT INFORMATION Helfried Brunner Chris Caerts ELECTRA IRP website link: 31