Frequency Response Provisions Graham Stein 3 March 2015 Agenda Introductions Background Discussion Topics Options for Progression 2 Purpose of the day PP14/59 was presented to the Panel in November Panel members recommended splitting the issues for the purpose of progression Workshop was recommended to help scope the work 3 Purpose of the day Four Grid Code “defects” were outlined in the paper Are there other things to consider? 4 Purpose of the day Workshop Conclusions Applicable Objectives Stop: record the issue as closed with reasons why or Outline Terms of Reference to form a workgroup Recommendation Grid Code Review Panel Agreed Next Steps or Outline scope to develop proposals for Consultation 5 Purpose of the day What we can do today Recommend undertaking further work which may result in a Grid Code change What we can’t do today Recommend a new market mechanism for frequency response Develop requirements for parties not subject to the Grid Code (eg demand side not connected to the transmission network) However, points raised will be recorded 6 Our output could look like… Issue Stop Further Assessment B D Develop Consultation A C Form a Workgroup …. 7 Background GC0022 Frequency Response Workgroup and Technical Subgroup GC0022 Frequency Response Workgroup GC0022 arose from long running industry discussions in the Frequency Response Workgroup Joint Grid Code/CUSC Workgroup Main purpose of Workgroup was to examine market arrangements Grid Code proposals were packaged up in GC0022 Proposals did not proceed past the draft consultation stage Incorporated the findings of the Frequency Response Technical Subgroup including an assessment of synthetic inertia an exploration of Rapid Response Why Inertia is Important for Frequency 1) Initial Rate of change of frequency limited by stored energy of the rotating mass (ie the size of the inertia) Frequency (Hz) 2) Primary Response acts within 10 seconds and sustained for a further 20 to contain frequency deviation 49.2 Hz -4s 0s ~8s Time (s) Power (MW) PTemp (5-10% PNom) Additional power delivered by synchronous machines – area under the curve is the kinetic energy released by the rotating mass Power output of decoupled Wind Generation – Ideally such plant should behave like that of a synchronous plant through controlled action PNom Recovery in Kinetic Energy following restoration of System Frequency Primary response delivered by de-loaded machines PDeload -4s 0s 2s 12s Time (s) 10 Frequency 16 October 2014 (1000 MW loss) 49.75 49.75 49.65 49.65 49.55 49.55 09:10… 49.85 09:09… 49.85 09:08… 49.95 09:07… 49.95 09:06… 50.05 09:05… 50.05 09:04… Frequency (Hz) Recent Example of a Secured Event 11 Synthetic Inertia Potential Advantages Direct replacement for inertia from synchronous plant Available without curtailment Issues Widely discussed but poorly defined concept Active power recovery: Likely to have value in most power systems meaning manufacturers might want to develop it as a standard capability recovery period for wind turbines operating just below rated wind speed (after the synthetic inertial injection) can result in significant reduction in active power output causing a double dip some 10-15s after the initial loss Control: Df/dt controllers are noise amplifying and can fail to operate properly especially when small time constants are involved Response needs to be proportional to avoid instability 12 Synthetic Inertia: Active Power Recovery Wind Speed Power Curve p.u. 1.2 Power (p.u.) 1 0.8 0.6 0.4 0.2 0 0 20 40 60 80 100 Wind Speed (%) Worst Case frequency drop 50.5 Frequency (Hz) 50 49.5 49 48.5 48 47.5 0 10 20 30 40 50 60 Time (s) 13 Rapid Frequency Response Potential Advantages Standard control philosophy Extension of existing Balancing Service No major technical hurdles identified Wide range of plant already capable Issues Potential to disrupt new technologies Extension of mandatory requirement is unpopular Perceived potential to restrict competition Curtailment required 14 Outstanding Issues Frequency response of generators at low load Current provisions potentially limit generators’ ability to operate at low loads by requiring full response capability whilst operating at or above DMOL. This feature ensures that response is available when generators are running but potentially restricts operation at lower loads unnecessarily. Frequency Response from other on-site sources Current provisions limit the use of alternative solutions (eg onsite batteries, storage or standby units) due to all units being required to meet the Grid Code rather than the overall output of the site. 15 Where are we now? The world has changed since GC0022 was packaged up Generation and Demand outlook is different Commercial Services have been developed Role of demand expanding Grid Code governance under review System Operability Framework developed to provide a better view of how the behaviour of the electricity system might evolve 16 The Issues 17 Rapid Frequency Response a Issue Rapid response for nonsynchronous generators Defect The current mandatory capability is not efficient in addressing future frequency response requirements on its own. Non-synchronous equipment does not contribute to system inertia in the same way as synchronous plant does. Faster frequency response will help to manage the resulting higher rates of change of system frequency and reduce the risk of excursions outside limits. It will reduce the total volume and cost of Frequency Response and allow wind turbines, interconnectors and photo-voltaics to provide more value to the consumer. Pros Reduced Balancing Services costs and therefore reduced costs to consumers. Enhanced system security Cons Increased costs for generators which would be expected to be reflected in prices to consumers. Will require further assessment in the future as the volume of non synchronous plant increases. There is a possibility that synthetic inertia may be required in the future Rapid Frequency Response What is it? Any plant which is insensitive to changes in system frequency and does not contribute to system inertia to be capable of providing fast frequency response ie applicable to non-synchronous plant (eg wind, interconnectors etc) Primary Response to be delivered in the period 0-5s after an incident (allowing for a 1s delay) Low Frequency response to be sustained for 25s High Frequency response to be sustained at not less than the defined initial rate as long as required All other aspects of Primary Response unchanged eg minimum required volume of 10% of Registered Capacity (RC) 19 Rapid Frequency Response 20 Rapid Frequency Response Why is it valuable? Start of frequency deviation (eg circuit breaker opens) Period of highest rate of change of frequency Response (MW) Frequency (Hz) 50.0 49.2 0 1 2 5 Time (seconds) 10 System Frequency Primary Response Rapid Primary Response 21 Scheduling Frequency Response 3 Generation Scenarios at Low Demand Scenario Description A Generation matches demand within the available regulating range B In an example where more frequency response is needed, more generators are required to deliver response. Eventually you run out of regulating room. C In the absence of other options, you need to curtail either wind or other “inflexible” generation to solve the problem 22 How do you decide whether to progress? Ramp Rates and Delay b Issue Clearer delay and ramp-rate requirements Defect Current provisions leave some uncertainty over the performance requirements for generators delivering frequency response. There is clear system sensitivity to the ramping capability of responsive generation and how quickly response is initiated. How quickly a generator meets its primary response requirement in 10 seconds can be critical. Pros Clearer ramping requirements and initiation times would reduce uncertainty in calculations of response requirements and in generator compliance assessment. The ENTSO-E European Network Codes also ask for these to be defined. Cons Risk of requirement being too onerous, especially for certain plant types 24 Ramp Rates and Delay National Grid makes assumptions about how quickly frequency response is delivered when working out how much is needed Our experience has been that response times vary Issues have arisen during compliance testing due to varying interpretations of the Code 25 Low Load Operation c Issue Low Load Operation Defect Current provisions potentially limit generators’ ability to operate at low loads by requiring full response capability whilst operating at or above DMOL. This feature ensures that response is available when generators are running but potentially restricts operation at lower loads unnecessarily. Pros Facilitates flexible operation from generators and contributing to system inertia and voltage control Cons Requirements may be complex to define and agree. 26 Low Load Operation 27 Alternative on-site sources d Issue Defect Alternative on- Current provisions limit the use of site sources alternative solutions (eg on-site batteries, storage or standby units) due to all units being required to meet the Grid Code rather than the overall output of the site meeting Grid Code. Pros Alternative ways of providing response. This may however be addressed through the ENTSO-E Requirements for Generators Code Cons Potentially limited application 28 Alternative on site sources The Grid Code is not clear on the use of alternative sources to provide additional frequency response, for example storage Could help in cases where technology struggles to provide frequency response The Grid Code Connection Conditions place obligations on the “Generating Unit” which makes it difficult to have separate onsite sources for the provision of additional Ancillary Services RfG could simplify this process as the obligations for Generating Units will in future be on the basis of a Synchronous Power Generating Module not the Unit The current version of RfG (dated 21 January) only covers Pumped Storage. Storage is however covered within DCC and is considered as part of a Demand Unit. Is there a demand for a change? 29 Discussion Options to Progress