Breakout Session 1.5 Innovation in Electricity Network Design LCNI - - PowerPoint PPT Presentation

1

Breakout Session 1.5 Innovation in Electricity Network Design

LCNI Conference Wednesday 6 December 2017

2

The ATLAS project (Architecture of Tools for Load Scenarios)

Dr Rita Shaw Model Development Lead

3

Future demand is uncertain ... and it may fall Load may rise...

4

Objectives of our work

Credible demand and generation scenarios, reflecting uncertainty Tailored to our region, assets and data Enabling good decisions about solutions to capacity problems, and informed dialogue with National Grid and other stakeholders Support well-justified strategic planning of network capacity

5

This presentation Next steps Overview of the ATLAS project New approach to MW (P) forecasting New approach to MVAr (Q) forecasting

6

Two NIA projects on load scenarios

Winter / summer peak load Heat pumps & air con The Real Options CBA model April 2015 - October 2016 Half-hourly (hh) through year Demand & generation Seasonal peak and min P (MW) & Q (MVAr) Nov 2015 – December 2017 Demand Scenarios with Electric Heat and Commercial Capacity Options ATLAS (Architecture of Tools for Load Scenarios)

7

ATLAS scope

Full half-hourly view of true MW demand MVAr scenarios learning from REACT NIA, for whole DNO network MW scenarios learning from the Demand Scenarios NIA, with more customer detail Prototype tools for GSP, BSP and Primary scenarios

8

ATLAS – demand definitions

True demand Latent demand

Loads DG units

Measured demand

9

ATLAS – true demand

True demand Measured demand Monitored DG exports Effects of DG

• n reducing

customer demand

Monitored component

• f true demand

Non- monitored DG

Latent demand

10

Data processing - monitored component

Data corrections (half-hourly & daily analyses) Identification of data problems

See detailed methodology at www.enwl.co.uk/atlas

11

Aggregated MW demand across GSPs

500 1000 1500 2000 2500 3000 3500 4000 4500 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 P (MW) time (hr)

( )

Generation Measured Demand

Peak true demand (23/11/2016)

500 1000 1500 2000 2500 3000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 P (MW) time (hr) Generation Measured Demand

Min true demand (05/07/2016)

Latent demand varies over time

12

Substation-specific weather correction

Correlate daily weekday demand over five years, with temperature and daylight hours Scale half-hourly demand to the historic temperature range of that month

13

MW forecast model per G&P substation

Integrated scenarios approach for all GSPs, BSPs and primary substations Scenarios presenting peak/average/ min diurnal profiles

• f demand and

generation Working with Element Energy, extending their work with UKPN and NPG Baseline of processed half hourly (hh) true demand + database of installed DG Model on FY17 baseline used for 2017 scenarios

14

MW forecast approach

Demand Technologies Generation Technologies Energy Storage Technologies Electric vehicles Solar PV Domestic storage (with solar PV) Heat pumps (domestic and I&C) Wind I&C storage behind the meter Air conditioning (domestic and I&C) Micro and larger CHP Frequency response Flexible generation Other generation Underlying demand based on 35 customer archetypes matched to substations Efficiency, demographics, economic activity

15

What does ATLAS add?

All prototype development in 2017 – transfer to BAU in 2018

1

Full views of true demand and latent demand, linked to measured demand

2

Not just peaks - 48hh per day

3

New weather- correction approach

4

New long-term MW forecast approach

5

Add connections activity

6

New time-series MVAr forecast approach with network modelling

7

Combine MW and MVAr to meet all reporting and planning needs

16

4,000 4,500 5,000 5,500 6,000 6,500 FY17 FY18 FY19 FY20 FY21 FY22 FY23 FY24 FY25 FY26 FY27 FY28 FY29 FY30 FY31 MVA Peak Demand Scenarios Green Ambition Active Economy Central Outlook Focus on Efficiency Slow Economy 2017 2024 2031

2017 peak true demand scenarios

Using the ATLAS prototype approach Long-term scenario adjusted for known major demand projects

17

Use scenarios to make decisions

Site demand scenarios Choose timescale etc. Repeat analysis for Strategy A and Strategy B Cost and risk distributions Calculations Summary metrics Inputs Define strategies with up to 3 interventions, including post- fault DSR

18

Why forecast reactive power?

Declining minimum Q (MVAr) demand from distribution

Source: NG SOF 2016

 High voltage problem on transmission network Develop ATLAS method to put scale on future Q exports to transmission

19

Simplified view of MVAr (Q) flows

Qprimaries

Empirical Rule: QGSP = Qprimaries + QEHV-absorbed - QEHV-gains

= QEHV-absorbed - QEHV-gains

I2X

V2C’ℓω

20

Historical Q/P-ratio at primaries (linear fitting of seasonal trends per GSP)

EHV Network Component Future measured Q demand at primary substations Future measured Q demand at GSPs and BSPs

Primary true P demand (scenario results) Primary latent P demand (scenario results) EHV generation (P and Q of existing DG & scenario results for P) EHV demand of large customers (P and Q demand of existing load & scenario results for P)

Primary Substations GSP & BSP substations

ATLAS Q Forecasting method

Empirical or modelled approach?

Historical Q/P-ratio at primaries (linear fitting of seasonal trends per GSP)

Future measured Q demand at primary substations

Primary true P demand (scenario results) Primary latent P demand (scenario results)

Primary Substations

21 1000 2000 3000 4000 5000 6000 7000 8000

time(hr)

• 50
• 25

25 50 75 100

Q(MVAr)

EHV absorbed EHV gains Primaries

Q forecasting – empirical rule

Q absorption → reduced for more lightly loaded EHV, but not for reverse flows Q at primaries → more capacitive primaries (declining Q/P trends) Q gains → increased when more cables or higher voltage targets are used

22

50 100 150 200 250 300 350

time(hr)

100 200 300 400 500

P(MW) Kearsley 132 GSP

simulation NG data CLAVA

50 100 150 200 250 300 350

time(hr)

• 150
• 100
• 50

50 100

Q(MVAr)

Validation using historical network and half-hourly monitoring data

Q forecasting – network modelling

Network Modelling Time-series analyses (i.e. daily simulation using

• perational

aspects) REACT approach... but with enhanced inputs P and Q profiles at primaries (and BSPs for large customers)

23

Central Outlook scenario, avg DG output , minimum Q demand = max Q exports

5 10 15 20 25 30 35

year (starting from FY17)

• 1500
• 1000
• 500

Q(MVAr) sum of min Q at GSPs

min Q Q at min P

5 10 15 20 25 30 35

year (starting from FY17)

1 1.5 2

Q/Q

2 0 1 7 (pu exports)

min Q (max Q exports)

Q exports in this scenario: +5% in 5 years +11% in 10 years +83% in 35 years But... in reality max Q exports could be even higher in different scenario and with different generation

• utput

24

Future application of the ATLAS methods

So next year we will: Use 2018 scenarios to estimate max Q exports at GSPs Request NG’s expected Q export limits at GSPs / compare to Q export scenarios Scope interventions to alter max Q in ED2 By 2020: NG as SO will use powers under RfG / DCC to set Q export limits at GSPs, via expanded NOA process Could add significant costs on DNOs in ED2 period And in FY20 we will: Use 2019 scenarios to estimate max Q exports at GSPs Compare max Q exports in our scenarios to limits per GSP Create high-level intervention programme for ED2 WJBP

25

Final months of the project Available capacity for generation Thermal and fault level Transition G&P approach to BAU, but keep under review Scope approach for secondary networks, build

• n improved

baseline data in new NMS

26

For more information

Please contact us if you have any questions or would like to arrange a one-to-one briefing about our innovation projects www.enwl.co.uk/innovation innovation@enwl.co.uk 0800 195 4141 @ElecNW_News linkedin.com/company/electricity-north-west facebook.com/ElectricityNorthWest youtube.com/ElectricityNorthWest

e