Page 29 - Gi flipbook - November 2018

Amazon and Netflix have been in a fashion that ensures continued
successfully harnessing machine public confidence in the system.
As regulation looks to keep pace
learning for some time... that insouciance with digitalisation in the energy sector
and more generally (most notably via
may not equally apply to our energy use the Network and Information Systems
Regulations 2018 (NIS) and the
implementation of the EU General
learning-based predictive algorithms necessarily conservative standard Data Protection Regulation (GDPR)),
can offer significant advantages to operating parameters. and energy suppliers face maintaining
traders, given their sophisticated power Systems incorporating very high public confidence in the security of
to draw out relevant patterns in volumes of data, that need to be their data, the skills required within
market drivers and behaviour and their analysed and acted upon in near supply-side companies are changing.
ability to handle larger datasets, real-time, are ideal for machine Alongside the already complex energy
enabling a broader range of parameters learning-based decision-making. Deep specific regulation faced by the sector,
to be considered. It is also clear that analysis, using probabilistic decision there is a growing body of data and
government, Ofgem and National Grid algorithms, allows optimised decisions cybersecurity regulation, which energy
are keen not to stymie this opportunity. to be made at speeds not humanly companies and their advisors must be
The government and Ofgem’s 2017 possible. Digitalisation therefore resourced to navigate and to apply.
Smart Systems Plan was accompanied allows operators to make better use On smart metering specifically, the
2
by a series of balancing services of existing infrastructure, avoiding provisions of GDPR, coupled with the
reforms from National Grid, most billions in capital upgrades. Despite security requirements imposed by NIS
recently in the form of confirmed plans these advantages, handing control of and the Smart Energy Code (SEC),
to further open up access to the the nation’s power to AI would represent a subset of regulation quite
balancing mechanism to ‘virtual power undoubtedly make some feel uneasy. distinct in nature from the transactional
plants’, made-up of distributed In the event of a failure, those and regulatory understanding required
generation and demand assets . machine-based decisions may be hard to ensure the physical infrastructure is
3
Digitalisation is also meeting to reconstruct or explain to those installed appropriately.
challenges in energy infrastructure responsible for oversight. There is no denying that
operation and maintenance. The ability At the consumer level, the (at times cybersecurity is challenging, but the
to deploy increasingly cheap sensors, controversial) implementation of smart banking industry has demonstrated
tied to the ability to collect and metering represents a tangible example that with the right regulation and
analyse the resulting datasets, allows of the reach of digitalisation, with the industry approach, consumer
the construction of ‘digital twins’ for potential for a truly two-way, real-time confidence can be gained. A reminder
major assets. Digital twins are a highly relationship between supplier and of the benefits of digitalisation and
accurate digital representation of a customer. The benefits for the the profound changes underway in
complex physical asset that uses consumer have been heavily promoted, our energy system means that hitting
continual machine learning to model but inevitably there is a certain amount the stop button simply isn’t an option.
the performance of the asset of public disquiet about what happens If the energy industry can harness the
throughout its lifetime. Groups of to the data they produce and the fact power of digitalisation, it will bring
digital twins can be modelled that suppliers have been tasked with enormous efficiency savings for the
collectively to predict and optimise leading on roll-out. nation and for individuals, not to
complete sections of energy systems. The reality is that the mind-boggling mention greatly facilitating the
Hence, machine learning can facilitate a amount of data (35,000 data points journey to decarbonisation. ■
transition from planned or condition- per year, for each household) is simply
based maintenance practices to too huge to be analysed effectively ■ Compton Energy Associates is a
predictive maintenance regimes. The – by humans. But with AI, deep-seated sustainable energy consultancy
prediction of failures allows the patterns of consumption behaviour specialising in renewables and whole
avoidance of unscheduled maintenance. can be teased out of huge datasets, energy system issues. For more
When integrated with market almost in real time, through the use of information, email andy.compton@
predicting algorithms previously analytics, and this is where the comptonenergyassociates.co.uk.
described, corrective maintenance can valuable information starts to appear. LexisNexis is a leading global provider
be scheduled to avoid high-revenue Amazon and Netflix have been of legal, regulatory and business
opportunity periods. Further, it can successfully harnessing machine information and analytics. In October
optimise operating regimes to ensure learning for some time to recommend 2017 LexisNexis launched LexisPSL
the asset reaches the intended product and film choices and most Energy, a dedicated energy sector
maintenance period without failure. consumers seem relatively relaxed platform. For more information visit
Digital twins allow the opportunity to about this. However, that insouciance www.lexisnexis.co.uk
simulate performance under extreme may not equally apply to our energy
scenarios without risk to the physical use. Legitimate concerns around
asset. Better knowledge of the asset’s cybersecurity and data privacy will REFERENCES
1. www.iea.org/digital
performance could allow operators to require robust safeguarding measures. 2. www.gov.uk/government/publications/upgrading-our-
release additional performance, Similarly, the question of what we allow energy-system-smart-systems-and-flexibility-plan
3. www.nationalgrid.com/sites/default/files/documents/
previously considered risky under AI to control will need to be addressed Wider%20BM%20Access%20Roadmap_FINAL.pdf

29

EnergyMarketSmartenUp.indd 2 18/10/2018 13:37