OZ-ENERGY-ANALYSIS.ORG   -   open science for the new millennium

THE STORIES   |   DATA   |   ANALYSIS   |   MODELS   |   LITERATURE   |   DISCUSSIONS

status: [24th Jan 2012] Round Three (R3)

Introduction

Electricity use (demand) data for the National Electricity Market (NEM) states, 1999 to 2011 inclusive.

The data presented here (and as discussed below) is the AEMO "price_demand" data, repackaged from 700+ csv files into a single demand flatfile, and alternatively as basic yearly CSV files.

The demand data covers the NEM (National Electricity Market - Australia), consisting of SA, QLD, NSW, VIC, TAS and a region SNOWY (now disbanded). Data for mainland states starts in Dec 1998; Tasmania starts in May 2005; the SNOWY region was disbanded into NSW and VIC from July 2008, and in any case the values involved are small compared with these states.

This is Round Three data (first presented here 19th January 2012); it includes data to end 2011. This data page is one of three closely connected pages: (this) demand data page, the Electricity Price Data page and the Price, Demand and Value - Electricity Market Analysis page.

The multiple data sources

There are at least two versions of demand data provided publicly by AEMO; that given here (dug out as indicated on the price data page), and that found in: http://www.nemweb.com.au/REPORTS/CURRENT/HistDemand/ and http://www.nemweb.com.au/REPORTS/ARCHIVE/HistDemand/ These two versions are not identical, but are highly similar. Unless and until some reason to revert to the above HistDemand data arises, we will now use the "price_demand" data.

The Data

Most data runs Jan 1999 - Dec 2011 inclusive. The 30 min data is the 'raw' data, with daily and weekly averages derivative from this. In constructing the daily average data we allowed up to three null ('NaN') data points (and the average that day is calculated from the remaining 47 or 46 or 45 data values); this suffices to patch all the missing data, of which there is very little in any case. The weekly averages are built as the average of all seven daily values; the weekly data drops the last day or two of each year in order to give a clean 52 weeks of data for each year. Units are MW average over the time interval.

• 30 min - OzEA_R3_nem_demand_30min.flat.gz (3.2 MB gziped flat / text file)
• Daily - OzEA_R3_nem_demand_daily.txt (380 KB text file)
• Weekly - OzEA_R3_nem_demand_weekly.txt (55 KB text file)

Also, here are some basic CSV files containing the thirty minute data for the NEM states for each of the years: 2009, 2010, and 2011 (can provide other years in this format if requested).

Demand Data - Descriptive Statistics

First we look at an overall picture in terms of the weekly average use. No smoothing (beyond the averaging).
A wide browser window will give the plots in pairs.
In all cases the 'y-axis' range represents a doubling.
Some comments at the end.

It is interesting to note the seasonal patterns in the various states, and how these combine to give the overall NEM demand. The strong summer peaks in SA are thought to correspond to paticularly hot periods.

Demand Extremes

It is of particular interest to examine the high demand extremes. The following plots show the demand extremes in an overall view, year-by-year, for each of the NEM states (based on the 30 min data). The 10⁰ point represents the 1% level; that is the demand level that is exceeded 1% of the time. Note also that individual plots link through to PDF versions that show an expanded view down to 0.01% (i.e. 10^-2).

When these plots were first done (see comment 6 below), looking at the years up to and including 2009, growth in the demand extremes was almost stepwise (especially apparent for SA and VIC). With the addition of the 2010 and 2011 data, we see some moderation in the demand extremes. It would be great to understand what underlies this.

Here are the cut-off values (in MW) for the demand extremes, as in the plots above (for 2009, 2010 & 2011).

It is seen that, for SA and VIC especially, the last 10% of the demand requires the system (generation and distribution) to be half as big again as that required for the other 90% of the time.

Comments

This is a data page, concerned mostly with presenting the raw data, having a place to discuss any data-centric issues, and including here straight data characterisation works.

There are other aspects of the demand data that can usefully be characterised, including the daily cycle; however, this is it for the 'head post' for now [Jan 2012]. Any further work and commentary can proceed in the open discussion below.

6 OzEA_DED0006

francis Subject: Demand extremes - plots Date: 2010-09-03 (at 00:40:38)

[fc: 24th Jan 2012]
Am keeping these visible to show how it looked previously; i.e. with the peak demand apparently growing strongly year on year.

Here are plots showing the demand extremes, year-by-year, for each of the NEM states. The 10⁰ point represents the 1% level; that is the demand level that is exceeded 1% of the time.

The growth in demand extremes (as opposed to overall usage) is clearly observed for SA and VIC in particular. This is expected to be the consequence of Air Conditioner use, something that we'll look at when we get to examining the timing of the extreme demand events.

[Show Full Lists]

 

  TOP   |   REFRESH PAGE   |   DATA

fc - Jan. 2012