Learnings

Outcomes

At the start of the project, participation in FFR was extremely limited for dynamic portfolios of large assets.. The existing guidance and lack of clarity around process for large portfolios of smaller assets remained a barrier for large aggregate portfolios. Several processes were identified to be uneconomic for these providers in addition to frequency meters including: the manual onboarding process; testing; and the tender process. The baselining and settlement processes were also identified as risky for providers of dynamic portfolios. However, the barriers for residential asset aggregators to participate in the market were not identifiable yet to National Grid. The project identified these barriers and through research, development and demonstration proved the potential of both a new technology as well as operational practice for use on National Grid’s system.  

New Operational Practice 
The existing processes and technical requirements for service provision were clearly identified to reduce the confusion for new providers. In addition, National Grid ESO has become aware of the barriers for dynamic portfolios and have a set of key areas to focus on. They have a list of quick changes in the short run and considerations to include in updated guidance. In addition, solutions for longer term consideration have been highlighted which provide the most substantial benefits in the long run.  These solutions have been combined together to define a new operational practice with a cohesive method of combining on-boarding, tendering, delivery and performance monitoring and settlement (see Figure 5 - Progress Report 2020 Addendum ).  

With the conversations and findings of the project, the FFR service is already in the process of evolving by the end of the project. National Grid ESO has improved the testing process for new FFR units during the period of this project and recommendations have been provided to improve the process further. They are also trialling weekly auctions and alternative baselining methods, taking the project’s recommendations into consideration. In addition, such IT projects as ESO market portal and PAS are under development and offer promising potential to integrate the recommendations further.  

New Technology 
Two potential alternatives to frequency metering are under consideration by National Grid ESO and may significantly reduce the technical costs of participating in FFR series.   
The regional metering subtask developed a frequency meter located in a regional hub that can receive frequency signals from a number of devices in the local area (see Figure 6 -- Progress Report 2020 Addendum ). It determined that the communication signal can be dispatched and received with sufficient latency to allow the majority of assets to respond within the FFR requirements. However, reliability of the response is likely to be the challenge due to variability of communication over the home networks. This will require careful monitoring by the aggregator. 

The low cost local metering subtask considered using a lower-cost IC chip located on the device rather than a Synectic meter (see Figure 7- Progress Report 2020 Addendum). It established that sufficiently calibrated IC metering devices are able to achieve 99% accuracy, similar to that which can be seen from metering hardware currently widely used by FFR service providers. The subtask was able to build firmware for 3 individual IC evaluation kits to log frequency at 10Hz and stream to PC over serial port, and calibrate the local devices to achieve 98% savings in cost per device. The positive results of the firmware development costs as well as accuracy of the meters make this a potential opportunity to retrofit existing devices as well as include in new domestic devices especially in the EV and HVAC space.  

Review of Benefits 
Costs 
A key benefit of the project was identifying ways to reduce the costs for participation by residential scale assets. The Industry standard frequency metering costs £245 per asset. This cost was reduced to below £4 per asset with new local metering approaches developed in work package 3. The new testing methodology for the aggregate portfolio using type testing or automated evaluation would reduce testing costs to a negligible amount. Further automation of onboarding using an asset registry and ongoing performance monitoring using an API would reduce the cost of service provision to generate a net profit for residential scale assets. Once these changes are made, the consortium has agreed that it would be cost-effective to provide flexibility from their portfolios of EVs, batteries and other residential assets to National Grid ESO’s balancing services.  

As the number of these assets continues to grow and tender periods shorten, it will open up the market to allow new types of competitors. The competition for services will drive down the cost of balancing services. It will also contribute to lowering carbon emissions as they displace larger thermal plants that traditionally have been used to provide these services.  

Applicability to other projects 
This is the first project which recognizes the unique challenges with onboarding and management of large portfolios of residential-scale assets. The consortium workshops and calls enabled connections to be draw to other innovation work being carried out by ESO as well as network operators. These connections reduced duplication of work and ensured that the solutions were able to match future plans and also identified that the challenges identified are not unique to ESO.  

While the ESO has the most immediate need to facilitate large portfolios of residential flexibility assets in its balancing services, the Distribution Networks are also looking to these assets to provide flexibility. The learning from this project will be of direct use to all Distribution Network Operators (DNOs) as they start to develop their own markets for flexibility. The consortium is now engaging with DNO’s to share the knowledge on communication standards and flexibility and ensure that learnings can be directly applied.  In addition, Western Power Distribution (WPD) was able to provide feedback on results throughout the project as a strategic advisor on the project. The project will support WPD’s Future Flex as well as provide valuable information to other DNOS.   

The project also provides a good starting point to understand ESO’s energy service markets. These learning are already being fed into a number of projects which include ElectraLink’s and Electron’s asset registry as well as the ADE’s DSR working group on dynamic containment. 

Lessons Learnt

The project provided many interesting learnings to both National Grid ESO as well as the service provider. This project has changed the way National Grid ESO understands the barriers to portfolio-based approaches to provision of flexibility with low voltage assets. This is the first step to ensuring equitable market access, and has the potential to significantly expand the supply base for ancillary services, increasing market liquidity and competition, and reducing costs. 

Project Objectives and Outcomes 
The project had three main objectives relating to the delivery of the three work packages. It was identified at the beginning of the project that National Grid ESO would not implement the solutions during the duration of the project, however they would be proposed with clear guidance for implementation as business-as-usual in the future. In addition, it was highlighted that for the trials in work package 3, finding both positive and negative results on the trials testing approaches to frequency metering would successfully meet the objective.  
Objective 1: Recommend updates to the approvals and commissioning process for onboarding new assets, ensuring these are fit for purpose for large portfolios.  
Objective 1 has been met and the recommendations are presented in the final work package 1 report: Proposal For Approvals / On-boarding Process For Large Portfolios of Flexible Assets.  
In this report, the existing testing, approvals, and onboarding processes for large portfolios has been clarified and laid out in Section 1 Current Process overview and Gap analysis. This includes step-by-step description of the process, the main barriers, and a comprehensive list of issues (see Figure 1 - Progress Report 2020 Addendum ).  

Section 2 includes the proposals for how the future process could look in both the short and long term to address these barriers (see Figure 2 - Progress Report 2020 Addendum).  Alternative solutions of improved guidance, type testing or automation of testing verification and an asset registry have been proposed as solutions. Section 3 considers further suggestions for National Grid ESO to consider to ensure that the solutions implemented are fit-for-purpose. This includes approaches to producing guidance documents and integration with the functionality of the ESO Market Portal and a potential asset registry such as RecorDER. The solutions link to work being developed currently by National Grid ESO and Elexon as well as consideration of future services.  

National Grid ESO has taken the recommendations on board for consideration and active engagement has begun with DNO’s to incorporate considerations into their future services. 
 
Objective 2: Determine and agree the procedures for data capture and requirements for portfolio assets to back up performance-based assessment of portfolios managed close to real-time.  
Objective 2 has been met and the results are presented in the final work package 2 report: Establish Dynamic Management Processes for Residential Aggregated Assets Portfolio Participation in FFR. 
Section 1 of the report identifies the current tender and service delivery processes. The step-by-step guide portrays the current processes includes determining the available capacity of assets, tendering, allocating assets to a portfolio, managing and monitoring the portfolio during delivery, and the payments and evidence provision at the end of delivery.  
Section 2 looks at the key gaps and issues with the current process based on a gap analysis and case studies.  Four key themes were identified with the portfolio management processes: 
Manual interactions across the whole process add unnecessary costs and minimise flexibility for dynamic portfolio management within the balancing services markets. A large part of this relates to processes relying on faxing forms and countersigning for confirming portfolio changes; 
The systems used by providers and NG ESO throughout these processes are disjointed increasing the complexity of ongoing participation in the market for providers with more dynamic portfolio of assets (e.g. Coupa for tendering, faxing forms for availability changes and emailing excel template for performance monitoring); in addition, these processes are not connected to the manual on-boarding processes;                                                                                                                              
Current processes are not explicit enough for providers wanting to participate in the market with a portfolio of dynamic assets (especially where the assets have a different primary use). Current processes are only well-defined for large-scale single asset participants with static availability. Lack of detail around processes for changing portfolio assets, no definition behind how performance data should be captured on a portfolio level; no baselining definitions different types of assets, are making participation in the market difficult; 
Auction time frames are uneconomic for providers with dynamically changing portfolios with the current portfolio management processes and baselining methodology.  Current timeframes for tendering, limitation on number of portfolio changes, and inaccurate baselining methodology lead to increased costs and unnecessary uncertainty particularly to dynamic portfolios of small assets that could technically participate in the markets. 
Section 3 proposes an alternative portfolio management process for National Grid ESO to implement. The solutions resolve the identified barriers by determining methods for dynamic real-time portfolio management and data capture. A methodology was proposed for implementing these solutions for baselining, weekly auctions, and ongoing performance monitoring that utilises the IT solutions National grid is currently developing. A summary of these solutions includes:    

• Move the balancing services markets to weekly auction timeframes and introduce day-ahead auction through trials;  
• Clarify and formally define asset allocation processes required within dynamic portfolios. In the long term these processes should be further linked to asset onboarding processes and tied together via asset registry platform;
• Test and formalise an additional baselining methodology fit for portfolios of residential assets. This should include the ability for service providers to update the baseline closer to real time;  
• Move to formalised ongoing operational performance monitoring for improved settlement processes and formalise process for aggregating performance data from large asset portfolios;
• Establishing transparent and automated data exchange processes across the full range of NG ESO balancing services markets to cover portfolio management, availability, and settlement;  
• Develop integrated IT infrastructure in the longer term to enable cohesive operation end-to-end (from processes pre-service delivery to settlement post-service delivery) across the range of balancing services.

National Grid ESO has accepted these recommendations and has provided the report to the teams working on the development of the related services, e.g. the PAS team.  
 
Objective 3: Evaluate, develop, and test options for measuring power and frequency that meet NGESO requirements for verifying service provision 
Objective 3 was met and the results can be found in the work package 3 final report: Alternative Approaches to Frequency Metering for Large Portfolios of Residential Assets. 
The first section of the report confirms the technical requirements to participate in FFR as published by National Grid in the testing guidance. The team determined that the testing guidance contains power metering requirements that are able to be met by the low voltage assets.  Frequency metering requires a 10 Hz recording, and accuracy comparable to industry standard Synectic meters.  

The technical requirements were used to develop 3 alternative options that could be scaled cost-effectively. These three options included remote metering, aggregate metering and low cost device level metering. The team was able to successfully test all three approaches in a range both lab and field trials.  The process of successfully developing new publishing designs, firmware and software is outlined in the report for future replication.  

All three metering arrangements were able to provide frequency readings in ideal settings in the lab. The metering arrangements were improved through calibration of the meters and refinement of the publishing design. A series of tests were run across different locations, times, and numbers of meters to test the robustness of the results at a larger scale. The final results show two of the three approaches were able to provide accuracy and latency within the technical boundaries: remote metering and low cost device metering using IC chips.  
The regional frequency metering results of all the testing provided encouraging initial results. The values found for data transmission are typically between 200-800ms with a mean of around 500ms (see Figure 3 - Progress Report 2020 Addendum). Additionally the associated latency of a site was stable across several days, meaning an aggregator could monitor and manage aggregate response. Device inverter response times were not tested and further questions around cyber security requirements remain. However, the results positively suggest that regional metering could be a practical solution of DER participating in frequency services.  
The low-cost device level metering using IC chips with updated firmware also provides a viable approach. Based on the trials the meters show high levels of accuracy with more than 99% of all measurements being within 0.01 Hz tolerance threshold (see Figure 4 - Progress Report 2020 Addendum). The consortium recommends that NG ESO look to sign off the device level metering with IC chips as being suitable for services requiring 1Hz, and potentially 10Hz. A number of further actions are outlined further in the report to verify some additional testing that are desirable to confirm the robustness of the solution.  
The aim, methodology, results, and conclusions of all three trials are presented in the report along with recommendations for further work. National Grid has accepted the initial conclusions of the trials and noted the areas for further testing.  
 
Lessons Learnt 
One key finding of the project was the benefit of communication between service providers and the decision makers at National Grid ESO to allow national grid to buy-in to the gaps that need to be addressed.  

Communication 
The project has opened up communication between portfolio aggregators and decisions makers at National Grid ESO that otherwise wouldn’t have existed. Monthly calls were put in place to provide a more open dialogue between National Grid ESO and the delivery team. As a result, it became clear how communication with National Grid ESO could enable new providers to both understand the existing processes quicker, but also identify where bespoke or alternative solutions where possible. Further opportunities for open communication between service providers and National Grid ESO may support the development of these services.     

Much of the initial desk-based research on the technical requirements and existing processes for work package 1 and 2 proved difficult due to a lack of clear guidance available as well as limited applicability to low-voltage assets. The project plan initially proposed a workshop at the end of each work package to discuss findings with National Grid ESO. After the initial work package 1 workshop, many perceived risks or issues of participating were resolved without needing to change the actual process or requirements.  For example, clarification to the new testing guidance that the aggregate portfolio could be tested rather than individual assets. These learnings strengthen the usefulness of the reports and may save time and money for service providers entering the market.  

The project also provided key learnings into how National Grid ESO’s current systems and processes are structured and the rationale behind them. By gaining a better understanding of how ESO processes work in terms of their capabilities, real requirements, and flexibility, it was possible to prioritise solutions in work package 1 & 2 for both short term quick-fixes and longer term that were acceptable to ESO.   

Systematic Change and IT systems 
Another key finding was that the existing processes in place are largely manual. This hasn’t been a problem when providing services with a few assets large assets, but will be when trying to scale these processes to large, dynamic portfolios.  Systematic change will be required to be able to dynamically test, tender, deliver and monitor services. Even after the processes are updated to consider low voltage assets, participation in the market will not become scalable without streamlined IT capabilities.   

The project has gained an understanding how small changes in IT are big undertakings for ESO; however many of the solutions require all the existing processes in place to be completely re-done to include IT components such as an asset registry and performance monitoring. While the solutions cannot be implemented straight away, an integrated roadmap to changing the whole process and the integration with the developing IT systems will be critical.  Work package 2 aimed to integrate the proposed solutions for testing and managing portfolios.