Statistical modelling of district-level residential electricity use in NSW, Australia

Electricity network investment and asset management require accurate estimation of future demand in energy consumption within specified service areas. For this purpose, simple models are typically developed to predict future trends in electricity consumption using various methods and assumptions. This paper presents a statistical model to predict electricity consumption in the residential sector at the Census Collection District (CCD) level over the state of New South Wales, Australia, based on spatial building and household characteristics. Residential household demographic and building data from the Australian Bureau of Statistics (ABS) and actual electricity consumption data from electricity companies are merged for 74 % of the 12,000 CCDs in the state. Eighty percent of the merged dataset is randomly set aside to establish the model using regression analysis, and the remaining 20 % is used to independently test the accuracy of model prediction against actual consumption. In 90 % of the cases, the predicted consumption is shown to be within 5 kWh per dwelling per day from actual values, with an overall state accuracy of −1.15 %. Given a future scenario with a shift in climate zone and a growth in population, the model is used to identify the geographical or service areas that are most likely to have increased electricity consumption. Such geographical representation can be of great benefit when assessing alternatives to the centralised generation of energy; having such a model gives a quantifiable method to selecting the ‘most’ appropriate system when a review or upgrade of the network infrastructure is required.

[1]  Dejan Mumovic,et al.  A review of bottom-up building stock models for energy consumption in the residential sector , 2010 .

[2]  V. Ismet Ugursal,et al.  Modeling of end-use energy consumption in the residential sector: A review of modeling techniques , 2009 .

[3]  V. I. Ugursal,et al.  A residential end‐use energy consumption model for Canada , 1998 .

[4]  Stuart James,et al.  Intelligent Grid - A Value Proposition for Distributed Energy in Australia , 2009 .

[5]  Marcos Pereira Estellita Lins,et al.  Regional Variations in Energy Consumption of Appliances: Conditional Demand Analysis Applied to Brazilian Households , 2002, Ann. Oper. Res..

[6]  Zhengen Ren,et al.  AusZEH Design: software for low-emission and zero-emission house design in Australia , 2011 .

[7]  R. O’Brien,et al.  A Caution Regarding Rules of Thumb for Variance Inflation Factors , 2007 .

[8]  Michael H. Kutner.  Applied Linear Statistical Models , 1974 .

[9]  Cheng Hsiao,et al.  A Bayesian Integration of End-Use Metering and Conditional-Demand Analysis , 1995 .

[10]  Robert Gould,et al.  A Modern Approach to Regression with R , 2010 .

[11]  Alan S. Fung,et al.  Modelling of residential energy consumption at the national level , 2003 .

[12]  V. Barnett,et al.  Applied Linear Statistical Models , 1975 .

[13]  P. Jones,et al.  MODELLING BUILDING ENERGY USE AT URBAN SCALE , 2001 .

[14]  Xiaoming Wang,et al.  Assessment of climate change impact on residential building heating and cooling energy requirement in Australia , 2010 .

[15]  C. Riedy,et al.  Study of factors influencing electricity used in Newington , 2006 .

[16]  Vijay Modi,et al.  Spatial distribution of urban building energy consumption by end use , 2012 .

[17]  Y. Shimoda,et al.  Residential end-use energy simulation at city scale , 2004 .

[18]  V. I. Ugursal,et al.  IMPLEMENTATION OF A CANADIAN RESIDENTIAL ENERGY END-USE MODEL FOR ASSESSING NEW TECHNOLOGY IMPACTS , 2009 .

[19]  H. Hens,et al.  Impact of energy efficiency measures on the CO2 emissions in the residential sector, a large scale analysis , 2001 .