Compu&ng  Electricity  Consump&on  Profiles  from  Household  Smart  Meter  Data  

O.  Ardakanian,  N.  Koochakzadeh,  R.  P.  Singh,  L.  Golab,  S.  Keshav  University  of  Waterloo  

 3rd  Workshop  on  Energy  Data  Management  

March  2014  Real-Time Distributed Control for Electric Vehicle Chargers

Omid Ardakanian, S. Keshav, and Catherine Rosenberg Cheriton School of Computer Science, University of Waterloo

Abstract

Problem Definition

Solution Approaches

spatial & temporal uncertainties

EV charging load is significant

~

impacts on the grid

At high penetrations, uncontrolled electric vehicle (EV) charging has the potential to cause line and transformer congestion in the distribution network. Instead of upgrading components to higher nameplate ratings, we investigate the use of real-time control to limit EV load to the available capacity in the network. Inspired by rate control algorithms in computer networks such as TCP, we design a measurement-based, real-time, distributed, stable, efficient, and fair charging algorithm using the dual-decomposition approach. We show through extensive numerical simulations on a test distribution network that this algorithm operates successfully in both static and dynamic settings, despite changes in home loads and the number of connected EVs. We find that our algorithm rapidly converges from large disturbances to a stable operating point. Given an acceptable level of overload, we show in a dynamic setting that only 30 EVs could be fully charged without control, whereas up to around 300 EVs can be fully charged with our control algorithm, which compares well with the ideal maximum of 383 EVs.

Control

Pre-dispatch Real-time

Centralized

Utility-oriented

User-oriented

Distributed

Utility-oriented

User-oriented

Control Architecture

Single Snapshot Optimization Problem

Results

References

max log 𝑟𝑎𝑡𝑒∈𝒮  

subject to 0 ≤ 𝑟𝑎𝑡𝑒 ≤ 𝑚𝑎𝑥𝑟𝑎𝑡𝑒 ∀𝑠 ∈ 𝒮 𝐸𝑉  𝑙𝑜𝑎𝑑 + ℎ𝑜𝑚𝑒  𝑙𝑜𝑎𝑑 ≤ 𝑠𝑒𝑡𝑝𝑜𝑖𝑛𝑡 ∀𝑙 ∈ ℒ

MCC nodes

Smart chargers

Every line/transformer is associated with a nameplate rating and a setpoint

line/

trans

form

er lo

adin

g

1 2 3 4

(measured by an MCC node)

(setpoint)

available capacity

Certain distribution branches may be subject to significant overloads, while the whole system has sufficient capacity

this provides proportional fairness

O. Ardakanian, S. Keshav,  C.  Rosenberg,  “Real-time  Distributed  Control  of  Electric  Vehicle  Charging”,  submitted  to  IEEE  Transactions  on  Smart  Grid. O. Ardakanian, C. Rosenberg, S. Keshav,  “Distributed  Control  of  Electric  Vehicle  Charging”,  Proc.  ACM  e-Energy, May 2013. O. Ardakanian, C. Rosenberg, S. Keshav,  “Fast  Distributed  Congestion  Control  for  Electrical  Vehicle  Charging”,  ACM  SIGMETRICS  Performance  Evaluation  Review,  December 2012.

Your  Smart  Meter  is  Watching!  

From:  hSp://www.thestar.com/opinion/2009/11/17/your_smart_meter_is_watching.html  2  

Smart  Meters  are  Ubiquitous  

Electricity     Gas   Water  

Energy  Retail  Associa&on  Projects  

Trials  

3  

Mo&va&on  for  Smart  Metering  

4  

Electricity  Consump&on  Profiles  

5  

The  Need  for  Electricity  Consump&on  Profiles  

6  

Prior  Work  on  Electricity  Consump&on  Profile  Genera&on  

•  Rely  on  data  that  is  not  easily  available  

•  Use  a  black  box  method  which  is  not  interpretable  

•  Are  not  robust  to  noise  

•  Do  not  remove  the  effect  of  temperature  and  ac&vity  –  cannot  be  extended  to  other  regions  and  ac&vity  paSerns  

7  

Takeaways  

•  Electricity  consump&on  profile  genera&on  has  several  applica&ons  

•  A  profiling  framework  must  be  simple,  interpretable,  yet  prac&cal  

•  Time  series  analy&cs  can  be  used  to  generate  such  consump&on  profiles  

8  

Key  Observa&ons  

9  

Residen&al  Load  Varies  with  Temperature  

10  

Residen&al  Load  Varies  with  Ac&vity  

11  

Residen&al  Load  Varies  with  Ac&vity  

Time  of  Day  

0                  1                  2                  3                  4                    5                  6                  7                  8                9                10                    11            12  

Power  (W

)  

12  

Ac&vity  level  must  be    inferred  from  data  

Our  Methodology  

13  

PARX  Model  

recent  history   temperature-­‐sensi6ve  load  

outliers   intercept  and  noise  terms  

season  index  

14  

PARX  Model  –  cont’d  

Cooling  

Hea&ng  

Overhea&ng  

15  

Handling  Outliers  

10%  of  Observa6ons  

10%  of  Observa6ons  

16  

Compu&ng  Consump&on  Profiles  

•  Parameter  Es&ma&on  – Number  of  seasons  – Coefficients  

•  Subtrac&ng  the  effect  of  exogenous  variables  

17  

Weekday  and  Weekend  Profiles  

18  

Comparison  –  Predic&ve  Power  •  Data  set  –  Residen&al  hourly  electricity  consump&on  data  of  1000  homes  from  March  2011  to  October  2012  

–  Hourly  air  temperature  data  of  that  region  

•  Prior  work  –  3-­‐Line  Method  

•  Fits  a  tree-­‐piece  linear  regression  aker  removing  outliers  –  Hourly  Mean  –  Convergent  Vector  

•  The  same  as  ours  but  does  not  remove  the  effect  of  exogenous  variables  

19  

Results  Av

g.  RMSE  

20  

Conclusions  •  Electrical  consump&on  profile  genera&on  is  important  and  has  many  applica&ons  – water  and  gas  consump&on    

•  Time  series  auto-­‐regression  framework  enables  us  to  remove  the  effects  of  temperature  and  ac&vity  

•  We  demonstrated  a  simple,  interpretable,  and  prac&cal  profiling  model  with  high  predic&ve  power  

21