Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly...
Transcript of Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly...
![Page 1: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/1.jpg)
Shaving the peaks through statistical learning, smart use of solar energy and storage solutions A. Lind, J. Selj, J. A. Tsanakas, P. Arnestad and A. Severinsen
![Page 2: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/2.jpg)
v
v
Motivation
2
• Theavailabilityofhourlyenergymetering,inpar6cularforelectricityconsump6on,havegreatpoten6altocontributetomoreefficientenergyuseinbuildings
• Toreachfullpoten6alweneedØ ImproveddatamanagementØ Improvedanalysistovisualizeeconomicpossibili6es
• Wehaveperformedacasestudytodemonstratethepoten6alelectricalenergycostsreduc6onsincommercialbuildings.
• Suchcostsavingsareachievablethrough:Ø peakdemandshavingØ u6lizinglocalPVenergyproduc6onØ storage
![Page 3: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/3.jpg)
v
v
Method
3
• Weanalyzedtheloadprofileof600differentstorestoiden6fythosewiththegreatestcostreduc6onpoten6al.
• Theenergyconsump6onisanalysedbyusingthecoefficientofvaria6on(CV)Ø TheCVvaluewasusedtoevaluate600differentstores
• PVsystwasusedtodeterminePVproduc6ononanhourlybasis.Ø Basedonavailableroofareaandlocalirradia6onandclima6ccondi6ons
• TRNSYShasbeenusedtolookatdetailedpowerflowandsimula6onofstorage Ø Basedonthefollowinginforma6on:
Ø ResultsfromPVsystØ HourlyloadprofilesØ Effect-tariffsØ Spotprices
![Page 4: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/4.jpg)
v
v
Grid tariffs
4
Commercialcustomer:peak-loadtariffs Price Unit
Fixedchargeperinstalla;on 340 NOK/month
Peakloadcharge(Jan-Feb,Dec) 150 NOK/kW/month
Peakloadcharge(MarandNov) 76 NOK/kW/month
Peakloadcharge(Apr-Oct) 11 NOK/kW/month
Energycharge(Jan-Mar,Nov-Dec) 0.052 NOK/kWh
Energycharge(Apr-Oct) 0.03 NOK/kWh
![Page 5: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/5.jpg)
v
v
Spotpriser 2016 (NO1)
5
0
100
200
300
400
500
600
700
800
900
1000
[NOK/MWh]
![Page 6: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/6.jpg)
v
v
• Energy consumption data from 600 Norwegian retail stores, with a total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving peaks in their
energy consumption profile, we propose using the coefficient of variation (CV). • The CV is a standardized measure of dispersion, defined as the ratio
of the standard deviation σ to the mean µ (Everitt, 1998):
𝐶𝑉= 𝜎/𝜇
2017-06-05 6
Selecting the stores with the highest cost saving potential
![Page 7: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/7.jpg)
v
v
7
Statistics for three selected stores with high and intermediate CV for the winter months
Store ID Average KWh Min KWh Max KWh CV Sum KWh Area (m2)
2448 103 60 160 18,7 227 208 1780 2487 76 26 180 44,7 154 870 2200 2703 151 78 212 17,8 307 298 2237
2448 2487 2703
kWh
![Page 8: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/8.jpg)
v
v
Analysis • For each individual store, three different cases were analysed: • Existing system (i.e. only power from grid) • Battery storage system combined with power from the grid • Battery storage system with power production from PV and electricity
purchased from the grid
8
![Page 9: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/9.jpg)
v
v
0
20
40
60
80
100
120
140
160
180
200
220
240
260E O
ut.Inv.[GWh]
9
Cumulative PV-production
2448
2487
2703
![Page 10: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/10.jpg)
v
v
10
• Hourly production values • Based on local irradiation and
climatic conditions • Modelled on the actual roof
orientation and inclinations
![Page 11: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/11.jpg)
v
v
11
Battery specifications
CurrentbaNerycost
[US$/kWh]
Op;mis;cbaNerycost
[US$//kWh]
BaNeryefficiency
[-]
Stateofcharge(low)
[-]
776 388 0.9 0.2
From: Irena 2015
![Page 12: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/12.jpg)
v
v
Oseana
PV price assumptions
![Page 13: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/13.jpg)
v
v
13
Yearly grid supply and production
0
200
400
600
800
1000
1200
1400
Ba_ery Ba_ery+PV Ba_ery Ba_ery+PV Ba_ery Ba_ery+PV
2448 2487 2703
[MWh/year]
PV
Grid
![Page 14: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/14.jpg)
v
v
14
Store 2487: Reductions in grid effect
0
5
10
15
20
25
30
35
40
0
10
20
30
40
50
60
[%]
[kW] Ba_+PV
Ba_
Ba_+PV
Ba_
![Page 15: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/15.jpg)
v
v
020406080100120140160180
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
[kW]
PV Grid Charge Discharge Load
15
Store 2487: Powerflow
March 3rd
Battery-strategy Use: (Load – PV) >= 130 kW Charge: Load <= 110 kW (20 kW or PV_surpluss)
020406080100120140160180200
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
[kW]
May 19th
![Page 16: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/16.jpg)
v
v
16
Store 2487: Powerflow
0102030405060708090100
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
[kW]
PV Grid Charge Discharge Load
May 10th
Battery-strategy Jan, Feb, Dec: Use: (Load – PV) >= 130 kW
Charge: Load <= 110 kW (20 kW eller PV_surpluss)
March: Use: (Load – PV) >= 100 kW Charge: Load <= 80 kW (20 kW eller PV_surpluss)
April - Sep: Use: (Load – PV) >= 70 kW Charge: Load <= 50 kW (20 kW eller PV_surpluss)
![Page 17: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/17.jpg)
v
v
17
Store 2487: Powerflow
Battery-strategy Jan, Feb, Dec: Use: (Load – PV) >= 130 kW
Charge: Load <= 110 kW (20 kW eller PV_surpluss)
March: Use: (Load – PV) >= 100 kW Charge: Load <= 80 kW (20 kW eller PV_surpluss)
April - Sep: Use: (Load – PV) >= 50 kW Charge: If surpluss from PV (PV_surpluss)
0
10
20
30
40
50
60
70
80
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
[kW]
PV Grid Charge Discharge Load
May 16th
![Page 18: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/18.jpg)
v
v
18
Payback time
0
4
8
12
16
20
776US$/kWh 388US$/kWh776US$/kWh+Øktne_leie
388US$/kWh+Øktne_leie
Payba
ck;me[years]
2448
66kWh
33kWh
0
4
8
12
16
20
776US$/kWh 388US$/kWh776US$/kWh+Øktne_leie
388US$/kWh+Øktne_leie
Payba
ck;me[years]
2487
66kWh
33kWh
≈ 66 kWh battery 0
4
8
12
16
20
776US$/kWh 388US$/kWh776US$/kWh+Øktne_leie
388US$/kWh+Øktne_leie
Payba
ck;me[years]
66kWh
33kWh
2703
![Page 19: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/19.jpg)
v
v
Summary
20
• Thecoefficientofvaria6on(CV)isausefulsta6s6calmethodtoanalysethedispersionofenergyconsump6onincommercialbuildingsØ (thereisaneedtosortoutbuildingsthathasrecentlycarriedthroughenergy
efficiencymeasures!)• Highpeakloadtariffswillgivehighsavingpoten6alforbuildingswithhighCV
values,thoughpeakloadshaving• ThreebuildingsofintermediatetohighCVvalueswereanalysedindetail:
Ø Threescenarios:a)grid,b)gridandba_ery,ogc)grid,ba_eryandPVØ CommercialbuildingswithhighCVvaluesmayprofitfromop6mizedba_ery
systemsalreadytoday.Ø Expectedincreasesinpeakloadtariffsandreduc6onsinba_eryandPVpriceswill
increasetherelevanceofsuchsystems.Ø Moreadvancedba_erystrategyisiden6fiedaskeytoachievethemost
costop6malsolu6onsØ Furtherworkincludesimplemen6ngmachinelearningstrategiesforload
predic6onsandba_erystrategies,inaddi6ontoPVforecas6ng.
![Page 20: Shaving the peaks through statistical learning, smart use ...€¦ · total of 5.3 million hourly observations for 2016. • In order to identify stores with high potential of shaving](https://reader035.fdocuments.in/reader035/viewer/2022070921/5fb9d602d3efd35218623e53/html5/thumbnails/20.jpg)
v
v
21
Thank you for the attention!
Josefine Selj, Phd Research scientist, IFE E-mail: [email protected] Arne Lind, PhD Senior research scientist Institute for Energy Technology e-mail: [email protected]