· To assist Centrica with the design, testing and trial of the market platform known as the Cornwall local energy market.
Objectives
· To determine the data exchanges that are required to support the platform and the practicalities of purchasing and operating flexibility services via a market platform.
· To determine a means of optimising the selection of services from those available, which may include other factors than price, such as reliability.
· To investigate the impact of varying attributes such as market model, purchasing timing etc.
Learnings
Outcomes
Greater clarity was found over the functions that should be part of each element of the system and how these should work. The platforms was developed and tested successfully in each trial. Data was exchanged successfully using the web interfaces and APIs. Centrica will continue to develop this platform with a view to replicate it in other areas.
- Platform learning summary
Without the additional functionality, such as peer to peer trading, it is not possible to determine whether customers prefer platforms that support multiple services to those focussed on DSO requirements. Recruitment faced very similar issues to that by other projects with the financial rewards on offer relative to the time commitment required to participate being the main barrier.
The time commitment for sellers is non-trivial as it reflects the need to familiarise themselves with the platform mechanics, market concepts, review terms and conditions, have personnel ready to respond to events, collect metering data and monitor the platform continuously for bids. Therefore the strong local brand is not expected to translate into dramatically different levels of customer sign-up. However the less onerous metering requirements may make the platform attractive to a wider range of customers.
Exeter University have investigated the potential barriers to flexibility participation in greater depth. Their report is available using the following link. https://www.centrica.com/media/4671/future-prospects-for-local-energy-markets.pdf. This concludes that the regulatory and market barriers need to be addressed before local energy markets can flourish. As there is still a large degree of uncertainty as to how the market will develop, it is not yet possible to provide a reasonable estimate of likely platform operating costs.
- Phase 1 trial – Quote and Tender Market Model
The outcome of the Phase 1 trial is that we have demonstrated the market platform and data interfaces in the real world. A suitable open source tool for optimisation was found and an optimisation tool created. The trial generated significant learning which is summarised below.
Phase 1 trial learning summary
The Phase 1 trial showed that customers generally bid at prices that were almost identical to those quoted for Flexible Power (£300/MWh). Some events did not take place due to the inability to create a contract beforehand. This was suspected to reflect participants requiring notification of service requests that were relevant to them, a feature planned but not executed on the Phase 1 platform. Events that took place typically delivered 60% of the contracted capacity. The trial highlighted the issue of the network running in a non-standard configuration impacting service delivery. The practicality of optimisation solutions depends on the complexity of service definition. Standardised services, with consistent timings for providing availability data, arming and dispatch were preferred to individually tailored services.
- Phase 2 trial – Combined Spot Market Model
The main outcome of the Phase 2 trial was the demonstration of the market platform being used by multiple purchasers at the same time. This market mechanism was significantly different to the previous models that had been used and the platform performs sophisticated optimisation using complex parameters that also maintain grid security.
This generated significant learning which is summarised below.
Phase 2 trial learning summary
The Phase 2 platform successfully matched bids to offers for DSO and ESO services concurrently, while managing conflicts between services. The market mechanism did not affect prices or service delivery which were consistent with Phase 1. In terms of assessing service delivery analysis suggests that half hourly metering gives similar results to more granular metering for delivery proportions and does not result in consistently skewed payments. The baselining methodology and how any adjustments are applied were seen to have significant impact on measured delivery and this warrants further investigation. An issue was highlighted with domestic customers reporting negative service delivery, in contradiction to the local battery system monitoring suggesting that either the baselining is inappropriate or that service delivery can be negated by local changes in consumption.
Further details for all the outcomes and learning is available in the project closedown report.
Centrica have now completed publishing their reports on the Cornwall Local Energy Market which are available at this location. https://www.centrica.com/what-we-do/centrica-innovations/cornwall-local-energy-market-research-reports-and-papers/
Lessons Learnt
The project proved two viable alternative means of purchasing flexibility services using third party platforms. It showed that the pricing and reliability of the services under each method was similar and that it was possible to build in a degree of conflict avoidance within the spot market platform by introducing complex market clearance rules.
The trials also showed that changes in network configuration from the default arrangement could result in the impact of services being changed and that DSOs may have to share future network configurations to those buying locational services.
A potential issue around domestic customer benchmarking has been highlighted which suggests there may be a minimum number of customers or aggregated customer capacity that should be combined for benchmarking and service delivery purposes to benefit from the averaging effects.
The project proved two viable alternative means of purchasing flexibility services using third party platforms. It showed that the pricing and reliability of the services under each method was similar and that it was possible to build in a degree of conflict avoidance within the spot market platform by introducing complex market clearance rules.
The trials also showed that changes in network configuration from the default arrangement could result in the impact of services being changed and that DSOs may have to share future network configurations to those buying locational services.
A potential issue around domestic customer benchmarking has been highlighted which suggests there may be a minimum number of customers or aggregated customer capacity that should be combined for benchmarking and service delivery purposes to benefit from the averaging effects.
Another issue may be the relative scale of the power delivered by the battery. If battery power delivery is relatively small then concurrent increases in consumption in the home could negate the impact of the battery before it is registered at the meter. If service delivery by domestic customers is under-reported then the payments provided may be too low to encourage further development of the domestic market. Therefore further research is recommended to determine the impact of the number of customers, their aggregated capacity, the ratio of battery power to average consumption etc. on the reported delivery, but also on whether alternative benchmarking methodologies are more appropriate for domestic customers.
Terminology needs to be stated explicitly, for example both National Grid and WPD were used to using half hourly numbers to represent the 48 half hour periods in the day. However the National Grid convention was to start the day at 11pm as is the case for balancing services whereas the WPD convention was based on settlement where the day begins at midnight.
The project highlighted some of the difficulties of operating in an evolving market. The market rules were designed with a fluid flexibility services market with plenty of liquidity in mind, however the conditions during the trials were considerably different. With a limited number of customers participating in the trial, there were few instances where there were sufficient participants for bidding behaviour to reflect a mature market. Similarly, in a mature market the way in which service providers would price their flexibility could be expected to reflect knowledge gained from many months of previous activity, but this was not possible to replicate during the relatively short trials.
The development of market platforms for flexibility services continues and has recently been boosted by the BEIS FleX competition. Therefore the likelihood of third party platforms being deployed at scale is high, however it seems unlikely that all the platforms currently being developed will continue and it is more likely that there will be a period of consolidation.
