Skip to content.
Sections
Home About Us Consumer Industry Contact Vacancies
Document Actions

Draft Grid Planning Assumptions (GPAs) released in May 2007 - demand forecasts

The information on this page was developed, released for consultation in September 2006, and published in May 2007.

National electrical energy demand forecast

The Commission carried out a substantive review of the demand forecasting methodology used for the 2005 Statement of Opportunities (SOO) as part of the preparation of the May 2007 draft Grid Planning Assumptions (GPAs).

The 2007 forecast is substantially lower than the 2005 forecast. The reasons relate to forecasting methodology. The Commission uses econometric models to forecast demand, where the historical relationship between demand and key drivers such as gross domestic product (GDP) is used to project future demand. The data used in the earlier forecast included a period in the 1980's when there was either a transition from oil and coal, to gas and electricity, or a structural change in industry make-up. The New Zealand Institute of Economic Research (NZIER) investigated the possible fuel transition for the Commission (NZIER Electricity vs total energy demand pdf [427 KB])

Following consultation with interested parties, the Commission decided to truncate the fit of historical data to start after the possible fuel switching or structural change was finished - about 1986. The impact was to reduce predicted future electricity demand.

The lower energy forecast is accompanied by a move to more sophisticated modelling of prudent peak demand (described below) which leads to higher prudent peak forecasts. Accordingly, this new forecast need not necessarily lead to any delay in the need for transmission investment.

Regional energy demand forecast

Regional forecasting is difficult because there is insufficient data to develop good regional econometric forecasts. The regional forecasting approach used by the Commission is actually an allocation of the national forecast to regions, based on population and GDP in each region, and population growth in each old Electric Power Board region.

Short-term changes in energy intensity, such as the intensification of farming, may result in high energy consumption growth relative to changes in GDP and population in those areas. The Commission’s approach incorporates the ‘inertia’ in demand growth associated with such changes by using a simple regression on the recent regional energy growth rate, which over five years or so, 'morphs' into the econometric allocation to the region. This means that a region that has been having very high energy growth rate, but little population growth, will continue to have that high growth rate in the initial years of the forecast, but will gradually come back to the econometrically based long run expectation.

A description of the regional forecasting approach is included in the National Demand Forecast Review (June 2006) pdf [522 KB]

document noted above.

Expected and prudent peak demand forecasts (annual)

The Commission has prepared forecasts of expected and prudent national, island and regional annual peak demand, as part of developing the May 2007 draft GPAs.

The primary uses of the expected and prudent peak forecasts within the GPAs were to produce the “after diversity maximum demand” (ADMD) peak forecasts and the generation scenarios.

The methodology used for calculating these peak demand forecasts is substantially different from that used in the Initial 2005 SOO. The goals driving the changes are:

  • to make sure that if a region has had a rapid increase in peak demand over the last few years, then the peak forecast continues to increase at a similar rate for the next few years; and
  • to provide a 'prudent' forecast, allowing for various influences which may result in higher peak demand than expected.

The expected forecast is a projection of historical peak demand data, with future growth rates driven in the long term by regional energy growth projections, and in the short term by extending recent peak demand growth.

The prudent forecast is based on a 10% probability of exceedence (POE) criterion; it is calculated as the 10th percentile of the probability distribution of future peak demand, allowing for various influences on demand growth.

For some purposes, the instantaneous peak demand is more relevant than the half-hourly peak.  The Commission has not produced an instantaneous peak forecast at this stage and plan to derive one when relevant data become available. 

The Commission note that the forecast presented is a 'business as usual' forecast, not explicitly allowing for any changes in consumer behaviour or technology. It also makes no explicit allowance for the possibility of increased future demand-side response with the intention of reducing peak loads.  The Commission prefers to treat increased demand-side response as a potential tool for meeting future peaks, rather than as a reduction applied to the peak forecast.

Load probability curves

The Commission has developed a methodology for producing forecasts of regional and national load probability curves. These forecasts are intended for use in reliability analysis, as a part of Transpower’s application of the GIT (Grid Investment Test) for major projects.  They would be used to determine the timing of major investments under the GRS and to calculate expected unserved energy.

The load probability curve (LPC) forecasts are consistent with the other regional forecats (energy, prudent and expected peak). They are expressed on the same scale (half-hourly demand at GXP level, net of embedded generation) and cover the same geographical regions. Regional, half-island, island and national forecasts are included, covering the period from 2007 to 2036. Both seasonal and annual forecasts are included.

The LPC forecast for a given year indicates the probability distribution of load in a randomly chosen trading period.  The spread of this distribution is a combination of the variation of load levels within a reference year and the uncertainty in peak and energy growth forecasts.  For example, the annual LPC forecast for the upper North Island is shown in the plots below.

The Commission has decided to include these LPC forecasts in the next draft of the Grid Planning Assumptions, along with examples as to how the Commission might expect to see them used in the application of the GIT.

As part of preparing the seasonal LPC forecasts, the Commission prepared seasonal prudent and expected peak forecasts (as opposed to the annual prudent and expected peak forecasts described in the previous section).  These seasonal forecasts will probably not be included in the next SOO, as they are not required for use in the GIT.  However, they are used as inputs to the LPC forecasts described above, and hence are included here for completeness.

After Diversity Maximum Demand peak forecasts

The ADMD peak forecasts are the peak forecasts that form part of the GPAs to be used for the application of the GIT under the Rules. They are intended to incorporate an allowance for possible high peak growth in the short to medium term, while presenting an expected growth path in the long term.  

The ADMD forecasts are prepared at a GXP level. They incorporate the annual regional prudent and expected peak forecasts, with a number of additional adjustments.

Individual grid exit point peaks, diversity factors and power factors have been calculated using historical meter data. Each grid exit point peak is forecast out to 2040 using the regional energy demand forecasts, and then scaled based on the regional peak forecast in each year. Additional adjustments are made to account for the impact of embedded generation.

Demand side scenarios

While the Commission has carried out a number of econometric demand forecasts, there is a need to explore more widely the range of possible outcomes which are the result of social or technical change rather than just economic drivers.  This is the role of demand scenarios. The preparation of demand scenarios is very much work-in-progress. 

As part of the consultation package circulated during the preparation of the May 2007 draft GPAs, the Commission published a report prepared for it by Dr Jonathan Lermit which described four demand scenarios looking at fuel substitution (electric cars) and energy efficiency.

Demographic change, change in energy consumption behaviour (e.g. a move to warmer temperatures in homes), and change in appliance use (e.g. substitution of heat pumps for other forms of heating) have also been noted as possible drivers of demand scenarios.

See also Modelling Group work in progress on the impact of electric vehicles on demand.
Last update on 12 August 2008 12:27 PM