STEM newsletter

Applying STEM to the energy sector

31 July 2008

During peak hours, power for a country’s national grid is typically supplemented using flexible but high-cost producers of energy such as oil and gas (O&G) turbines.

Outside these periods, the lower level of demand on the grid (the base load) is met as far as possible using less expensive sources of energy. Base-load power plants are cheaper to run than O&G plants, but take a long time to start up and are relatively inefficient at less than full output. These plants – typically nuclear and coal-fired – run at all times throughout the year.

The long-term planning of the base-load power stations in the national grid must be done carefully in order to minimise the use of O&G turbines, without over-building. This article presents a basic STEM model that is able to forecast energy demand and the requirements for O&G turbines. (Of course hydro-electric plant are commonly used for peak production too, but our model focuses on the financial impact of the significantly more expensive O&G turbines.)

Planned base-load generation and incremental peak capacity

The demand for energy comes from three major types of users: residential, commercial and large businesses. These exhibit quite different patterns of demand, modelled using the following assumptions for a typical industrialised country:

  • Residential users mostly require energy during the morning and evening, every day of the year, and have a low energy profile of around 15kWh per day per household.
  • Commercial users require energy during working hours Monday to Friday, with a typical energy consumption of 160kWh per working day per site.
  • Large businesses (here assumed to be heavy consumers of energy) typically require 520GWh per day summed across the whole large business sector, and are able to smooth their consumption over the whole day, every day of the year.

The following figure presents the STEM implementation of such demand.

Energy demand

Coal, Nuclear and Hydro resources are used to handle base-load energy generation in this model (and other sources such as wind-powered generators could be added). The number of base-load generators is not driven directly by the total peak capacity required, but is instead determined by a long-term plan of the number of such generators. This centrally planned approach is modelled  by three ‘Installed generators’ transformations, one for each resource, as shown below. The peak requirement is passed on to the resources, to enable cost allocation (which is not treated in this article).

Base-load planning and peak requirements

Given the base-load capacity of the power grid as determined by the long-term planning of base-load resources, the additional capacity required at peak times can be calculated: this then drives the number of O&G turbines that must be installed, as shown below. Note that these turbines have a 15-year lifetime, so once installed they remain available even if base-load generation increases later.

Additional peak energy capacity

The supply of each type of generator is also modelled, and the associated capital and operational costs can be estimated. In simple terms, coal plants require supplies of coal and produce pollutants and greenhouse gases that entail further costs in line with national legislation and ‘Kyoto’ pollution permits. Nuclear plants require uranium and create radioactive waste that must be disposed of at some expense. Hydro-electric plants do not require fuel supplies (apart from rain), nor do they create by-products.

Additional resources required include high-voltage transmission networks, low-voltage distribution networks, and commercial and administrative overheads.

Optimising the mix of base-load and expensive peak generation

The model uses scenario planning to optimise the deployment of base-load generators and minimise costs (typically driven by the build-out of O&G generators, and the cost of oil and gas supplies), and also reduce the costs arising from pollution (here modelled by pollution permits). Our base case (over 25 years) assumes the following hypotheses, as illustrated in the figure below:

  • Gradual phase out of coal generators
  • Increase in amount of energy from nuclear sources over time, but not increasing as fast as the decline in output due to the phasing out of coal
  • Amount of energy from hydro-electric plants remains constant (assuming all suitable dam locations have been used).

Assumed base-load energy generation (coal, nuclear, hydro)

These assumptions have two consequences. Looking first at the peak energy requirement (below, left-hand graph), the peak demand modelled drives a significant consumption of the O&G resources, with the amount of O&G-generated energy required gradually decreasing over time.

However, looking at the annual energy supply the share taken by O&G is only 1%, since O&G turbines are primarily used to provide ad-hoc peak supply during the peak hour. The share of annual energy production by nuclear plants increases steadily to reach 50%, while the proportion of energy generated from coal decreases from 50% to around 35% in 25 years.

Peak and annual energy requirement and supply, by generator type

STEM can model energy supply and allocate costs for retail pricing

While the assumptions of this model can be modified for specific countries (or particular energy providers), this case study demonstrates the ability of STEM to model energy supply over long periods of time. Base-load planning can be optimised to minimise the use of expensive sources of energy during peak periods.

The model can also explore the consequences of certain choices regarding social benefits, such as pro-actively choosing low-emission energy sources (here nuclear or hydro-electric plants, or wind turbines – not modelled here), and to examine costs arising from pollution permits.

Distribution and overhead costs can also be included, allowing the model to calculate end-to-end costs (allocated to each type of service and energy user on a per-kWh basis). This cost allocation will become a useful tool if energy price regulation evolves from the current RPI-X method to a cost-based method.

© Implied Logic Limited