Top Banner
Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei
20

Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

Mar 29, 2015

Download

Documents

Stephon Trott
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

Investigating Intrinsic Energy Consumption

Seasonality

Adedamola Adepetu, Elnaz Rezaei

Page 2: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

2

outline

• Introduction• Problem Definition• Data Sets• Our Approach• Challenges

Page 3: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

introduction

Seasonality is a repetitive behavior observed in time series

Intrinsic inherent repetitions in load

Days, Weeks, Months, Calendar seasons

3

Page 4: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

daily season

4

Time of day

Page 5: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

weekly season

5

Page 6: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

annual season

6

SummerWinter Winter

Page 7: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

problem definition

• Does the intrinsic load seasonality correspond to predefined ‘seasons’?

• Why does this seasonality exist (if any)?

• How is this affected by exogenous factors: temperature, prices, hours of day etc.?

7

Page 8: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

expected results

Establish a procedure for determining actual load seasonality, resulting in:

• Improved electricity pricing structure• More informed load prediction

process• Storage sizing for peak load

reduction

8

Page 9: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

datasets: ieso• Description: Aggregate load for the whole province

of Ontario

• Time Period: 2002 till date

• Resolution: 1 hour

• Weather data available from Environment Canada

• Electricity prices also available

• This is at the macro level9

Page 10: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

datasets: essex• Description: Load profiles for 6800 homes in

Windsor, Ontario

• Time Period: 1.5 years

• Resolution: 1 hour

• Weather data available from Environment Canada

• This is the micro level

• Also contains outages10

Page 11: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

datasets: irish• Description: Load profiles for 4560

homes, somewhere in Ireland

• Time Period: 1.5 years

• Average weather data for Ireland?

• Resolution: 30 minutes

• Electricity & gas11

Page 12: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

our approach

• Observe (without preconception) seasonality in all datasets

• Determine impact of exogenous features

• Match macro & micro levels, i.e., Ontario & Essex patterns

• Validation

12

Page 13: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

observing seasonality

• Autocorrelation– Includes Partial Autocorrelation

13

Page 14: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

observing seasonality

• Fourier series analysis– Spectral analysis

14

Time (months)

Page 15: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

observing seasonality

• Additive modely = trend + seasonal_component +

error

• Multiplicative modely = trend *seasonal_component + error

Aggregation of home load? Yes & No.

15

Page 16: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

impact of exogenous features

• Temperature

• Hours of day & night

• Prices (any feedback?)

• Weekends & Holidays (not exogenous per se but still important)

16

Page 17: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

regression

• Seasonal AutoRegressive Moving Average (SARMA)

• Seasonal ARIMA (verify nonstationarity of data)

• ARMAX? …X for ‘exogenous’

• Use regressive model from macro level as a feature at micro level

17

Page 18: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

validation

• Bootstrapping“…statistical inference based on building a sampling distribution for a statistic by resampling from the data at hand.” – Fox, 2002

Used to establish confidence intervals

18

Page 19: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

challenges

• What if we find nothing?

• Longer time periods required for micro-level datasets

19

Page 20: Investigating Intrinsic Energy Consumption Seasonality Adedamola Adepetu, Elnaz Rezaei.

references• Bisgaard, Soren and Kulahci, Murat. John Wiley &Sons, Inc., 2011. ISBN 9781118056943. doi:

10.1002/9781118056943.ch9.

• Espinoza, M., Joye, C., Belmans, R., and DeMoor,B. Short-term load forecasting, prole identication, and customer segmentation: A methodologybased on periodic time series. Power Systems, IEEETransactions on, 20(3):1622 { 1630, aug. 2005. ISSN0885-8950. doi: 10.1109/TPWRS.2005.852123.

• (IESO), Independent Electricity System Operator.Market Data. Accessed on 18 October 2012,at http://www.theimo.com/imoweb/marketdata/marketdata.asp .

• (OEB), Ontario Energy Board. Electricityprices. Accessed on 18 October 2012, at http://www.ontarioenergyboard.ca/OEB/Consumers/Electricity/Electricity+Prices .

• Singh, R. P., Gao, P.X., and Lizotte, D. J. On Hourly Home Peak Load Prediction. In IEEE SmartGridComm 2012, 2012.

• StatSoft. How To Identify Patterns in Time Series Data: Time Series Analysis. Accessed on 18 October 2012, at https://www.statsoft.com/textbook/time-series-analysis/ .

• Fox, J. (2002). Bootstrapping Regression Models. Annals of Statistics 9, 1-14. Available at: http://www.jstor.org/stable/10.2307/2240411.

• http://www.itl.nist.gov/div898/handbook/pmc/section4/pmc4463.htm

20