SYNAPS (Synchronous Analysis and Protection System)

Lead Research Organisation: University College London
Department Name: Statistical Science

Abstract

SYNAPS is an innovative project which brings together experts from the
power engineering, powerline communications, and statistical signal
processing communities to target the so-termed energy trilemma, namely
the challenge to improve energy security, reduce carbon emissions, and
reduce costs.
SYNAPS aims to develop a networked distribution automation platform for
low-voltage networks which will provide fault detection, classification
and location of faults, together with smart protection and
reconfiguration, at a significantly lower cost than has previously been
possible. In effect, this project will add a cost-efficient smart layer
across the national power grid which will not only solve long-standing,
industry-wide challenges but will also open up countless other
opportunities for stable, future-proofed growth as our cities and infrastructure become smarter and progress to the internet-of-things
future.
Since the low-voltage network was originally intended for one-way
distribution of energy, there has been little previous interest in
monitoring it. However, there is now a new imperative created by the
impact on network stability due to the growing deployment of consumer
operated renewable distributed generation equipment, electric
vehicles--- not to mention the 'exploding pavements' issue.
Currently, distributed generation amounts to only a small proportion of
the total network generating capacity, hence its impact on low-voltage
network performance is negligible. However, there is significant
industry concern about the effects of increased numbers of distributed
generation and electric vehicle installations, especially when these are
concentrated in co-located clusters.
The low-voltage electricity network needs to be able to support two way
electrical flow and real-time communication. About 9% of electricity is
lost in the distribution network, annually, and it has been reported
that 45% of Distribution Network Operator total network costs and 50% of
customer minutes lost are due to low-voltage cable faults.
Managing these new low carbon technologies present significant
challenges but early preparation and introduction of a Smart Grid should
make the transition easier and reduce overall costs. This project will
draw upon machine learning methodology to automatically monitor
low-voltage networks and detect and localise both known, and anomalous,
problem events. Furthermore, algorithms will also be progressed to
support software-based protection and reconfiguration of the network.
It is anticipated that such smart sensor networks will make a
significant contribution in network efficiency and future-proofing, and
have immense benefits for both consumers and EU/UK environmental and
energy policy targets.

Planned Impact

This project will deliver economic and commercial impact throughout the
energy sector and beyond. Through close collaboration with Powerline
Technologies Ltd, Techna International Ltd, Akya Ltd, and other
stakeholders, including Distribution Network Operators (DNOs), it will
facilitate rapid pull-through and knowledge transfer of advanced machine
learning and statistical signal processing to deliver enhanced network
performance, decrease downtime, and increase reliability. In turn, this
will support a range of EU/UK environmental and energy policy targets.
Anticipated impacts include:
. Distribution Network Operators are worried about the effect of new
renewable technologies on the quality of their network service. The
deployment of the Synaps platform will enable them to detect such
issues rapidly taking remedial action in a much shorter timescale.
Hence, better detection and mitigation of issues related to introduction
of photovoltaic and electric vehicles into the network, facilitating
wider deployment with resultant reduction in carbon emissions.
. Reduce the need for expensive site visits to investigate circuit
failures and enabling DNOs to not only fix feeder faults proactively
before they cause a power outage but also to plan their maintenance more
effectively by concentrating on those cables estimated most probable to
fail.
. Increased reliability of low-voltage network with reduced network
downtime, providing a better service to electricity consumers.
. Better low-voltage asset management potentially contributing to lower
energy costs. DNOs are under great pressure to extract more value from
their assets.
. provide are a necessary prerequisite to the evolution of DNOs into a
more active role, going beyond mere distribution to take on a role as a
manager of "microgrids" balancing renewable energy generation and
consumption, leading to a more resilient, cost effective and greener
industry.
In wider terms, it is anticipated that a successful project will
showcase the possibilities afforded by embedding advanced statistical
signal processing and machine learning methodology into a physical,
real-world system. By forming a genuinely interdisciplinary team from
academia and industry, this project will make a genuine leap forward
into the future and bring the long anticipated smart cities and
internet-of-things concepts one step closer to realisation.

Publications

10 25 50

publication icon
Nelson J (2015) Enhanced B-Wavelets via Mixed, Composite Packets in IEEE Transactions on Signal Processing

publication icon
Peng Chen (2017) Toward a Sparse Bayesian Markov Random Field Approach to Hyperspectral Unmixing and Classification. in IEEE transactions on image processing : a publication of the IEEE Signal Processing Society

 
Description Algorithms for fault detection have been developed in Three-Phase Low-Voltage Time-Series Data; subsequent exploration was explored with Powerline. Multiple publications arose as detailed.
Exploitation Route As described in the narrative impact this will be relevant to DNOs improving their operations. The algorithmic innovations may also be modified for related problems.
Sectors Energy

 
Description This functioning algorithm is being developed by power line technologies for exploitation, with the previously employed PDRA. If testing and validation is successful at PNDC (power network distribution centre) then DNOs (distribution network operators) will be able to exploit it. The PDRA is further exploiting the research by working with the funder.
First Year Of Impact 2018
Sector Energy
Impact Types Societal

 
Description Collaborating with Powerline, Techna, Akya and Intel 
Organisation Powerline Technologies Ltd
Country United Kingdom 
Sector Private 
PI Contribution This project is a joint endeavour between UCL and Bath, as well as several corporations, exploring how to target the so-termed energy trilemma, namely the challenge to improve energy security, reduce carbon emissions, and reduce costs. SYNAPS aims to develop a networked distribution automation platform for low-voltage networks which will provide fault detection, classification and location of faults, together with smart protection and reconfiguration, at a significantly lower cost than has previously been possible.
Collaborator Contribution Techna and power line are contributing with modelling expertise.
Impact This projects help to understand energy usage and how to detect faults.
Start Year 2015