Data Analytics for Future Energy Systems Dr Stuart Galloway & - - PowerPoint PPT Presentation

▶
data analytics for future energy systems
SMART_READER_LITE
LIVE PREVIEW

Data Analytics for Future Energy Systems Dr Stuart Galloway & - - PowerPoint PPT Presentation

Dec 18, 2023 533 likes •731 views

Data Analytics for Future Energy Systems Dr Stuart Galloway & Dr Bruce Stephen Advanced Electrical Systems Group Institute for Energy and Environment Department of Electronic and Electrical Engineering University of Strathclyde Glasgow G1

slide-1
SLIDE 1

Data Analytics for Future Energy Systems

Dr Stuart Galloway & Dr Bruce Stephen Advanced Electrical Systems Group Institute for Energy and Environment Department of Electronic and Electrical Engineering University of Strathclyde Glasgow G1 1XW United Kingdom {stuart.galloway|bruce.stephen}@strath.ac.uk

slide-2
SLIDE 2

Introduction

  • Smart Grid is essentially about data
  • More informed power system: sensing at an

extent and rate previously unknown

  • Q: How to get business value from this

investment in monitoring?

– New services? – Augment existing services?

  • A: Understand current operation & constraints…
slide-3
SLIDE 3

WHAT DRIVES THE NETWORK?

Characteristics of Demand

slide-4
SLIDE 4

Demand

  • Smart Meters/BEMS

– 30 minute reads

  • Understand demand to manage it

– DSM

  • When used/when not
  • Realign peak demand with renewable generation
  • All data driven!
slide-5
SLIDE 5
  • Loads on the LV network were always

assumed to have a high degree of variability to them due to the nature of domestic routine

  • AMI & BEMS deployment allows much

greater insight into the behaviour of these loads

  • Patterns are difficult to capture given the

complex nature of the data

  • High dimensional – how do they all

relate?

  • Non-stationary (many sub-behaviours
  • bserved e.g. weekends, holidays)
  • Tools have been developed for our past

projects that:

  • Formulate classes of energy use and

automatically categorise residents

  • Quantify behavioural consistency in

terms of movers and stayers

  • Forecast aggregated residential load
  • Model appliance usage and its

variability

  • Wet appliances
  • Heating (space and water)
  • EV Charging?

Patterns repeat in load profiles both across days and customers…

Load characteristics can be attributed to appliance usage patterns, which can also be quantified…

slide-6
SLIDE 6

HOW DOES THE NETWORK BEHAVE?

Power System Operation and Control

slide-7
SLIDE 7

Network Operation

  • Load flows

– Power flow directions in the network – Voltage excursions along feeders

  • State estimation

– Infer unmeasured quantities from measured ones

  • Data model drives physical models

– Real loads are not homogenous

  • Homogenous assumption may result in different outcomes…
slide-8
SLIDE 8

Feeder/Microgrid Model

Use to evaluate power system health metrics (load flows, voltages etc.) and inform. Load models plug into point loads and resulting aggregate can be evaluated (minus losses) at a desired point e.g. grid infeed.

1 2 3 4 5 6 7 8 9 1 1 1 1 2 1 3 1 4 1 5 1 6 1 7 1 8 1 9 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

11 kV sub- station

20 36 38 39 16 25 49 30 42 32 14 23 8 31

Double laid Single laid Triple laid

2 81 43 50 81 15 7 21 12 18 41 9 33 19 48 46 44 26 28 11 27 29 3 40 13 24 22 5 17 47 37 4 35 6 10 45 34 1 11kV

11kV 0.4 kV

Topology is OK but the real interest comes from what and where (how far) loads are attached from the feeder end. Solution is to superimpose the network model on an existing geographic housing layout… …but what is realistic? Realism is important from the DNO perspective as the following examples illustrate…

slide-9
SLIDE 9

Rural/Suburban

Heterogeneous properties of a similar vintage (1973), generous but uneven spacing, increasing PV penetration… …one infeed transformer – possibly specified for fewer houses than were eventually built.

Stephen, B., Mutanen, A., Galloway, S., Burt, G. & Jarventausta, P. (2013) Enhanced load profiling for residential network customers. IEEE Transactions on Power Delivery. ISSN 0885-8977 (In Press) Stephen, B., Isleifsson, F., Galloway, S., Burt, G. & Bindner, H. (2013) Online AMR domestic load profile characteristic change monitor to support ancillary demand services. IEEE Transactions on Smart Grid. ISSN 1949-3053 (In Press)

slide-10
SLIDE 10

HOW DOES THE NETWORK AGE?

Condition Monitoring

slide-11
SLIDE 11

Network Monitoring

  • Condition monitoring

– Could be SCADA, could be more…

  • Whats ‘right’, whats ‘wrong’?

– Huge volumes of data – What does fault(s) look like?

  • How to predict fault onset?
slide-12
SLIDE 12

Fault Detection/Diagnosis

  • Offshore wind Generation

– Look at the way the power curve behaves – Plant wear and degradation

  • How do changes manifest?
  • Transmission Power Transformers

– Partial Discharge detection – Different frequency compositions

  • What types of faults do they represent?
  • Distribution Switchgear

– Trip coil testing – Mechanism and control system faults

  • What does the shape of the test record indicate?

Stephen, B., Galloway, S.J., McMillan, D., Hill, D.C. & Infield, D.G. (2011) A copula model of wind turbine

  • performance. IEEE Transactions on Power Systems, 26 (2). pp. 965-966. ISSN 0885-8950

Baker, P., Stephen, B. & Judd, M. (2013) Compositional modelling of partial discharge pulse spectral

  • characteristics. IEEE Transactions on Instrumentation and Measurement, 62 (7). 1909 - 1916. ISSN 0018-9456

Stephen, B., Strachan, S.M., McArthur, S.D.J., McDonald, J.R. & Hamilton, K. (2007) Design of trip current monitoring system for circuit breaker condition assessment. IET Generation Transmission and Distribution, 1 (1). pp. 89-95. ISSN 1751-8687

slide-13
SLIDE 13

HOW CAN THE BUSINESS MAKE INFORMED DECISIONS?

Asset Management and Replacement

slide-14
SLIDE 14

Asset Management

  • Track fault occurrences/onsets on individual

plant

  • Look across whole fleet

– Inter and intra plant faults

  • Manage/model lifecycle of fleet

– Spares inventory management

  • Identify subsystems failures

– Repair prioritisation

slide-15
SLIDE 15

Intra-Plant Condition

Who is most similar who will fail next?

slide-16
SLIDE 16
  • Markov chain type

models

  • How do assets

move through their lifecycles

  • ver time?
  • Fault tree type

models

  • How do systems
  • f components

fail?

Wilson, Graeme and McMillan, David (2013) Modeling the effects of seasonal weather and site conditions on wind turbine failure modes.In: ESREL 2013, 2013-09-30, Amsterdam.

slide-17
SLIDE 17

How to go forward?

  • Power Systems can benefit from Big Data
  • Hence Smart Grid - Operational robustness & business value through informed operation
  • Good physical models and domain knowledge exist
  • ‘Black box’ type models may fail to capitalise on this
  • Other Smart Grid stakeholders could see benefit from these
  • utcomes. How can we all work together?
  • IT Service Providers
  • Community Energy Groups
  • Distribution Network Operators
  • Work with Power Systems engineers to develop models that leverage big

data insight with well understood domain knowledge

slide-18
SLIDE 18