Energy services may be described as energy supplied in a form appropriate for the accomplishment of some social or economic task. Most energy services are provided in the form of motive power, visible light and heat at various temperatures. Some services, such as those requiring information technologies, can only use electricity with current technologies. Electricity services are those energy services using electricity. Some of these services, such as information processing and lighting, are such that electricity is the only fuel that can be realistically used. Others, especially heating, can readily use other fuels.

Objective: Minimise service cost of systems incorporating demands with storage, and switchable and intermittent sources.

To meet this objective, demand and supply technologies will evolve with each other.

For example: one reason off-peak electric heaters were sold was to fill the night-time trough because it’s so expensive to turn big generators on and off. Therefore it is important that the electricity model takes some account of the load management potential with stores.

1.1.1 Environment

Environmental concerns increasingly constrain the electricity industry. A burgeoning body of international law and accords is setting limits to the emission of various pollutants, and to other environmental impacts.

The emission of sulphur dioxide, nitrogen oxides and particulate matter from large combustion plant is limited by EC law under the Large Combustion Plant Directive of 1988. This law applies to the total emission from all large plant in each country (a bubble limit), and to specific emissions from individual new or refurbished plant.

At present there is no legislation pertaining to the total or specific emissions of most greenhouse gases from countries or technologies. However there are some verbal commitments to limits, for example stabilising the emission of carbon dioxide. Also there is discussion of economic levers which will encourage energy efficiency. These include the possible carbon tax for the EC, and the imposition of VAT on domestic fuels in the UK.

Like other countries, the UK is increasingly subject to international law and negotiations which limit the future impact of energy processes generally, and the emission of certain pollutants in particular. EU law, reviewed in 1994, placed mandatory limits on the emission of acid gases (SO2 and NOx) and dust from large combustion plants. Figure 2 shows the current profile of allowed sulphur emission from UK power stations.

Figure 1 : Power station sulphur limits

1.1.2 Carbon dioxide

Currently there is no law limiting the emission of carbon dioxide. However international limits have been suggested and some governments have made commitments to national limits. The UK government is presently committed to limiting xxcarbon emission in 2010 to 12.5% less than 1990 levels under the Kyoto agreement. Other governments have policies which aim at reductions.

1.2 Predictability

The better the predictability of demand and intermittent supply, the better the control of the system and use of stores and switchable generators. What factors are important about prediction: GW total, rate of change of GW, hours ahead?

Wind forecasting is addressed in some detail in the SDC report and the accuracy of forecasting one-hour ahead is well illustrated by Figure 9 (page 24). The weather patterns that give our winds typically take 12 to 48 hours to cross the UK, so wind speeds can be accurately forecast for a few hours ahead of time; which means that one does NOT need power stations running at no load in case the wind suddenly stops. Over a land area of just 100km by 100km (which is only 4% of the total UK land area) the mass of air - up to a height of say 100 metres – is about a billion tonnes. When this is in motion it isn’t going to suddenly stop.

Demand. Storage (e.g. heat stores) and interruptible services can be used for some demand-supply matching. Such user stores, and interruptible demands, can be controlled more accurately with intelligent systems and telecommunication. The heat load, and its potential use for balancing, varies across four seasons. Perhaps 15% of electricity demand is for space and water heating, and most of this has storage. Other demands, such as refrigeration and cooling (about 8%), also include some passive or active heat (cool) storage. In addition there are interruptible demands, predominantly in the industrial and services sectors. Therefore, perhaps more than 20% of annual demand could be used for temporal load modification over periods of minutes to hours.

Direct resistance heating may diminish in energy efficient scenarios, but electric heat pumps may increase in number. Perhaps electric heating would be preferentially installed near wind farms?

[Graphs output from Mark Barrett electricity model. Historic, hourly simulation for a day from each season.]

Supply. Heat up costs and maximum heat up rate of large generators means they are used less than expected from steady state marginal costs. Hence mid-high marginal cost generators eat into the trough as shown below. This has an especially large effect on marginal fuel burn and emissions.

The following pages illustrate the use of EleServe as applied to the UK electricity service system. It is intended as a demonstration rather than a closely argued scenario for the UK. First a context for why demand side management (DSM) is likely to gain favour in the UK is provided. Then the use of EleServe to investigate demand, DSM, supply and emissions is illustrated with a sequence of graphs and accompanying text.

2. Model

EleServe provides an integrated framework for assessing technical options and their associated costs and environmental implications. EleServe can be used to explore these options and so help construct policies aimed at meeting given economic and environmental objectives.

There has been no discussion here of the policy context for strategy. There is an international policy context that has an ever increasing effect on UK policy: commitments to carbon emission limits and EU policies, regulation and law may be mentioned here. Within this international context there are many facets of UK economic and energy policy that bear on electricity service policy. These range from the regulation of the electricity supply industry (e.g. the distribution price control formula and cost plus basis of the RECs, non-fossil obligations); to appliance and building standards; to consumer investment financing; to allocation of emission bubble limits by the DETR; and to strategic fuel supply policy concerning domestic and imported fuels.

EleServe is an integrated model of electricity demand and supply. It has been developed to analyse a wide range of options which may be exercised in the provision of electricity services. It can assist on a range of issues: from appraising the market potential for energy efficient technologies, to demand and marketing analysis for electricity supply companies, to the formulation of strategies for generators or policymakers. It is particularly suited to assessing the effect of environmental concerns on strategy.

Overview of EleServe

EleServe is basically divided into a demand and a supply module. Modules may be used together or separately.

The demand module makes forecasts for the demand for electricity in terms of energy (TWh) and power (GW) or load shape. The demand for energy is calculated from a number of determinants which include numbers of households, ownership levels, industrial and service sector production and efficiency. The domestic element of the model includes a detailed stock model. The effect of demand side management programmes on demand and programme costs can be assessed. Programmes can be driven by normative or regulatory assumptions, or by market signals.

The supply module includes an hourly simulation model operated on a station by station or set by set basis. Station specifications include thermal efficiency, availability patterns and pollution control equipment. Up to three different fuels may be specified for each station with the cheapest being used. It calculates generation and fuel burn for each station and marginal supply costs. Total supply costs including capital, fuel and O&M costs are calculated. The emission of acid gases and carbon dioxide are calculated.

The power or load management module sits between demand and supply. It iteratively calculates the economically optimum levels of load management given the hourly marginal supply costs calculated by the supply module.

The calculations of EleServe are reported in a number of graphical and tabular formats. These may be printed, or pasted into reports. Results data may be output for further use in spreadsheets or other programmes. Data input for demand and supply can be supplied as ordinary text files, or stored in spreadsheets and databases.

3. Technology

3.1 Components

Demand

Sources

Heat engines

Thermal power stations

Energy to heat them to operating temperature

Energy to sustain them at operating temperature

Energy for generation

Flow

Light

Transmission: V=IR

3.2 System

Components in space and time.

Demand

Time

Flows and stores.

4. Demand

A starting point for analysis is a breakdown of electricity demand by sector, cost, type of end use and time of use. The degree of disaggregation depends on the data available, and the purpose of the analysis. Figures 3 and 4 show the structure of demand for the UK in 1991.

Figure 2 : Electricity consumption by end use

Domestic sector

What happens to the SEC of appliances when household owns more than one appliance? Only important for TVs?

Non-domestic sectors

No detailed stock data. How to simulate penetration of efficiency?

Price response model

Decision of consumers to purchase efficiency based on economic and other factors.

Economic model: NPV analysis ('implicit,inferred!) of

Regulatory constraints

5. Demand side management

DSM can be divided into the management of Power and of Energy.

The pressure for more DSM arises from desires to improve economic efficiency and to mitigate environmental impacts. International and national policies, regulation and law increasingly reflect these desires.

DSM is likely to gain favour in the UK because it is more efficient economically, and because it helps meet environmental constraints.

5.1.1 Economic efficiency

The electricity service systems can be divided into supply (the generation and distribution of electricity to the consumers' premises) and demand. The long run marginal cost (LRMC) of electricity supply, and the LRMC of saving electricity can be estimated. They are illustrated in Figure 1. Maximum economic efficiency occurs when the total cost of electricity services is minimised. This point is where the LRMC of conservation equals that of supply. The optimum (in this Figure) occurs at a level of demand 50% less than without conservation. The annual cost saving is equal to the area shown.

It is widely accepted that certain elements of DSM can reduce the overall cost of electricity services. Figure 1 illustrates the marginal costs of supply and conservation for the UK. As efficiency levels are driven up the marginal cost of electricity conservation increases and eventually equals that of supply (generation and distribution). This point is where the total cost of electricity service is minimised and is economicall optimum.

Therefore, insofar as Government policy aims to improve economic efficiency the prospects for cost-effective DSM will be enhanced.

Figure 3 : Conservation vs supply

5.2 Energy management

The demand for electrical energy is determined by the level of economic acitivity in the various sectors; and by the market share and energy efficiency of electricity use.

The question then is: how will the average stock index (I) evolve under different circumstances? This will depend on INew for each year, and on the rate of stock turnover.

Assuming strongest programme

General questions

What if new efficient technologies have different lives from those they replace (e.g. fluorescent bulbs)?

How will efficiency costs fall with mass production and technical development?

Second hand market important?

Premature replacement. Account for scrap value.

What about lights - different hours

Domestic sector

What happens to the SEC of appliances when household owns more than one appliance? Only important for TVs?

Non-domestic sectors

No detailed stock data. How to simulate penetration of efficiency?

Price response model

Decision of consumers to purchase efficiency based on economic and other factors.

Economic model: NPV analysis ('implicit,inferred!) of

Regulatory constraints

5.2.1 Market share

Electricity competes with other fuels in certain end use markets - mainly heat and, to a degree, motive power. About 20% of UK electricity supplied is used for space and water heating. For technical reasons the unit cost of supplying electrical energy and the environmental impact of so doing is greater than for most other fuels for most consumers. Therefore, as the pressures for economic efficiency and environmental improvement increase, so it will be difficult to sustain electricity's share in competitive markets.

5.2.2 Energy efficiency

Improving energy efficiency means providing a given level of service for less energy. Two examples: an efficient refrigerator prvides the same cool storage volume for 30% less energy than the current average: a fluorescent bulb provides the same amount of visible light as an incandescent bulb but consumes 70% less electricity.

Improving energy efficiency will reduce energy consumption for a given level of socioeconomic activity. Only if efficiency gains are greater than activity level increases will delivered fuel requirement be diminished.

In general, altering energy efficiencies will change the load shape of demand diurnally and seasonally. The weather sensitivity of total demand may also be changed. Increasing the efficiency of space heating and lighting will reduce the weather dependency of demand.

Market models

We have SECs expressed in terms of indices which are proportions of current average stock SEC. Starting points are the average SECs for the whole stock (IStock); for equipment actually purchased in a particular year (INew); and the most efficient equipment available in any year (IMin). market data and other analysis should provide values for these in the base year. An assumption must be made about what the lowest index could ultimately be, i.e. the maximum efficiency of that equipment (IMinUlt).

The question then is: how will the average stock index (I) evolve under different circumstances? This will depend on INew for each year, and on the rate of stock turnover.

Description	Name	Sample values
base average stock	IStock	100% in base year
base average newly purchased	INew	90% in base year
available lowest	IMin
available lowest base year	IMinBase	50% in base year
available lowest ultimately	IMinUlt	20% in future year

Assuming strongest programme

General questions

What if new efficient technologies have different lives from those they replace (e.g. fluorescent bulbs)?

How will efficiency costs fall with mass production and technical development?

Second hand market important?

Premature replacement. Account for scrap value.

What about lights - different hours

5.2.3 Demand forecast

Forecasts of the demand for electrical energy depend on numerous factors: from fundamental drivers such as number of households and economic output, to ownership levels and technical change including efficiency.

Figure 5 illustrates a Reference (REF) forecast based on projections of drivers. It is assumed that modest efficiency improvements (10% of maximum) occur without specific DSM programmes.

Figure 4 : REF sectoral electricity consumption (TWh)

The following Figures illustrate the REF scenario in which 20% of cost effective DSM savings are introduced across all sectors. The DSM programme does not include a reduction in market share for electricity in the competitive heat markets, although this could be argued for on environmental and economic grounds. The DSM programme commences in 1995.

Figure 5 and 6 show how consumption is disaggregated by sector, sub sector and end use for the domestic and non-domestic sectors.

Figure 5 : REF domestic electricity consumption (TWh)

Figure 6 : Non-domestic electricity consumption (TWh)

EleServe includes a detailed stock turnover model to simulate the scrapping of existing stocks and the introduction of new stocks with different efficiencies and costs. Figure 8 shows the change in dishwasher stocks as the ownership level rises and reaches saturation, growth then slows to match the increase in numbers of households.

Figure 7 : Domestic dishwasher stocks (M)

Extra efficiency, beyond some minimum free gain, adds to the capital cost of most equipment. Figure 9 show the extra capital investment required in domestic equipment. The unit cost of savings (p/kWh) increases with proportion saved. As the efficiency levels of the equipment and ownership levels reach maxima the annual investment levels off. Figure 9 shows the investment in efficiency in domestic appliances.

Figure 8 : Domestic efficiency capital (M£)

Figure 9 : BAU service cost (M£)

The extra cost of efficiency may be added to the cost of electricity delivered to arrive at the total cost of service to the consumer. Strictly speaking the total cost of the equipment (basic plus extra for capital) should be given. However this is not very meaningful for all end uses. Also, the extra costs for marketing efficiency are not included. Figure 10 shows the total cost of electricity services for all consumers. Even though efficiency improvements are assumed in the BAU case, in this instance they do not incur extra costs. The total annual cost of service rises from £15 billion to £23 billion.

In the DSM case (Figure 11) the cost of service rises from £15 billion to £18 billion. The financial savings to consumers of DSM thus amount to £5 billion per year by the end of the scenario. Investment in efficiency rises to £6 billion per year which is on third of the total cost of service, or half the cost of delivered electricity.

Figure 10 : DSM service cost (M£)

5.2.4 Power profiles

The demand for electricity varies through the day, week and year. it varies according to sector and end use. The demand for some end uses (space and water heat, lighting) depends on short and long term variations in weather. The changing composition of demand will alter the variations in total demand. The temporal pattern of demand, and particularly peak demand, determines the capacity and mix of electric power sources required. EleServe calculates the hourly power for each component of demand as it varies through the day, week and year. The following Figures show demand and power management on a winter's weekday for the year 2000. Figure 12 shows domestic demand, a similar disaggregation is available for the other sectors..

Figure 11 : Domestic demand (GW)

Demand components result in aggregated into overall load profiles as shown in Figure 13. Note the load management due to off-peak electric heating between midnight and 6 am — without this, the peak to trough ratio of diurnal demand would be much larger.

Figure 12 : Total demand (GW)

These demands may aggregated by useful energy type. This aids an analysis of the potential for load management.

Figure 13 : Diurnal demand by useful energy type (GW)

5.3 Load (power) management

Temporal

There are two basic means of load management: time shifting and disconnection. A consumer may want to load manage if the benefits exceed the costs.

5.3.1 Time shifting

This is where the demand for electricity is shifted from the time of service demand to some other time. This is accomplished by energy storage (e.g. off peak electric heating) or by changing time of service (e.g. running dishwasher at night).

Storage

The storage method of time shifting means that the timing and quality of service is not changed substantially. However, storage generally results in energy losses incurred during conversion when charging or discharging the store (e.g. electrochemical battery) or as losses (heat losses from a hot water tank); so that the total amount of electrical energy supplied is increased. For some technologies, where heat is not involved, losses may not increase subatnitally with time. such as batteries,

For heat storage the losses depend on the duration of energy storage and can be considerable; perhaps 20-40% of the heat from off peak electric storage heaters is wasted.

Energy storage incurs capital and running costs. These latter include maintenance and the costs of lost energy.

Presently the control of most storage is unsophisticated shifting is a simple time clock. More sophisticated controls are used in other countries and may be developed. These include controls which respond to time varying signals of parameters such as price per kWh. Control could be complex: for example; inputs to a store might be arranged over several hours or a day to xxabsorb a forecast input of renewable electricity.

: if the supply price is greater than some level the load is reduced or shifted.

Changing the time of service

Changing the time of service can engender capital costs such for timeswitches, although they may not be major. In general there is a cost to the consumer in that service shifting will bring an inconvenience. This may be small or even negative (running a dishwasher at night) or quite high (showering at 5 am rather than at 8 am).

5.3.2 Power/Load loss interuption

This is where a consumer's electricity supplies are reduced or cut off for a period of time. The consumer does not recoup this lost energy supply at some other time. Therefore there is a loss of service and total energy demand is reduced. Generally (in the UK) the probability of this occurring is small.

Load management algorithm

Objective to reduce total costs of service.

Demand side. Value to consumer of service at a particular time, and costs of time shifting that service.

Supply side. Avoided capacity costs (generation, transmisison, distribution) and energy costs (fuel, O&M).

Power loss. Gross cost saving should reflect avoided cost of not taking power. This includes some reflection of The cost to the consumer is that the value placed on electricity at that time. This might reflect lost production etc. etc.

Power shift.

The gross cost savings include those for power loss. In addition..??

Power or load management can be applied change the profile of demand. If the costs of shifting demand to another time, or of disconnecting load, are less than the cost savings in delivered electricity then power management is worthwhile. Figure 14 illustrates this. The reduction in peak load is 3.8 GW.

Figure 14 : Before and after power management (GW)

6. Supply

The supply module of EleServe is used to calculate the generation costs and emissions. Assumptions are made about future capacity, fuel supply and emission control.

6.1 Generating mix

Figure 15 shows a possible future capacity mix. The installed capacity of private generation and CHP is included. The capacity scenario is for demonstration. No claim is made that it necessarily represents the optimum from an economic, environmental or other point of view.

Figure 15 :Installed capacity (GW)

The generation by each station is calculated for each hour of each sampled day. Figure 16 shows generation on a winter's day aggregated by plant fuel type.

Figure 16 : Diurnal supply (GW)

The model can output snapshot results for any hour of the year (see Fig 17). These may may be used for purposes such as inputs to the European electricity trade model.

Figure 17 : Stations energy costs and carbon emission

The diurnal generation results are summed over the year. Generation, fuel burn, costs and pollution emission calculated. The remaining Figures illustrated some of these.

Figure 18 shows the mix of generation over the period. From this, and the following Figures the long term problem of generation mix may be realised. If the long term supply of gas is problematic, then it will decline. If nuclear remains unacceptable, then it will also decline. This leaves abundant supplies of 'dirty' coal, less abundant and more insecure oil, and renewables. These latter are by definition inexhaustible, but what is their potential within the foregoing context?

Figure 18 : Generation by source type (TWh)

Sulphur emissions (Figure 19) decline steeply because of the use of gas and some retrofitting of FGD. In the longer term they increase from a low base because of increased coal burn albeit in plant which remove most sulphur.

Figure 19 : Sulphur emission (MtS)

Figure 20 shows carbon emission. In the first decade it falls principally because of the increased use of gas with a lower carbon content and of renewables. However, in the longer term carbon emission starts to rise again with demand and the increasing proportion of fossil generated electricity. Carbon emission can only be further reduced by a combination of very high levels of efficiency and a very large renewable component. An unceasing increase in the basic demand for energy services will eventually overwhelm any technical developments. Therefore the possibility and nature of long term economic growth requires critical discussion.

Figure 20 : Carbon emission by fuel (MtC)

Figure 21 :Installed capacity (GW)

The model can output snapshot results for any hour of the year (see Fig 17). These may may be used for purposes such as inputs to the European electricity trade model.

Figure 22 : Stations energy costs and carbon emission

The diurnal generation results are summed over the year. Generation, fuel burn, costs and pollution emission calculated. The remaining Figures illustrated some of these.

Figure 23 : Generation by source type (TWh)

Figure 24 : Sulphur emission (MtS)

Figure 25 : Carbon emission by fuel (MtC)

7. Remarks

The illustrative scenario shown demonstrates some of the capabilities of EleServe and how it can be used to explore options for the provision of electricity services. The scenario illustrates the difficult problems that will be faced in the longer term provision of electricity services, and the consequent need to devise a coherent approach and strategy to best face these problems.

EleServe runs on an IBM compatible PC with Windows 3.0 or later. At present it is suitable for experienced users only.

Market share

7.1.1 Energy efficiency

Market models

The question then is: how will the average stock index (I) evolve under different circumstances? This will depend on INew for each year, and on the rate of stock turnover.

Description	Name	Sample values
base average stock	IStock	100% in base year
base average newly purchased	INew	90% in base year
available lowest	IMin
available lowest base year	IMinBase	50% in base year
available lowest ultimately	IMinUlt	20% in future year