« Previous | Contents | Next »
Listen
5 Energy Delivery Scenarios
A scenario based methodology was developed for the study as a means of concisely specifying various mixes of renewable energy sources. The balance between generated and consumed electrical energy is expressed both at a national level and in disaggregated form at area level by a number of key figures. This section of the report describes the analysis methodology in more detail and defines the key figures derived for each scenario.
5.1 Scenario Selection
Scenarios in this study are based on time series of renewable generation and demand. The temporal resolution of these was one hour. The total period of time was limited to three years, mainly because input data for offshore-wind and wave modelling was only available for 4.5 years of which the three full years, 2001 through 2003, were chosen for the time series analysis. Results were derived for each year individually and for the whole period. Due to the importance of seasonal variations in generation and demand, the winter and summer periods were examined separately. In this study the months of December, January and February represent winter, while June, July and August represent summer.
The creation of time series for renewable generation has been described in Section 3 for the four technologies considered. Time series for demand were derived as described in Section 4. These time series were used to compare, on an hour-by-hour basis or over periods of time, the electricity generated from a technology mix with the corresponding load demands. This process is illustrated in Figure 5.1.
Figure 5.1 Model of scenario development.
Results of this study are presented as a series of scenarios. A scenario is specified by a particular mix of renewable energy sources. The results corresponding to each scenario follow from comparisons that are made between the time-series of renewable energy generation and of load demand. They are presented in the form of graphs and tables in Section 6.
Within the scenarios, each technology type was implemented by using incremental placements of capacity. The placements were ordered by relative cost rather than by location. However, as clustering as well as resource quality was taken into account in the calculation of relative cost, geographically adjacent groups of machines were likely to stay together in the same placement groups. The total plant capacity was increased between scenarios by a factor of two. The starting scenario contained 750 MW of plant with the subsequent ones containing 1.5 GW, 3.0 GW and 6.0 GW. The placements were cost ranked, so that in each case, the smaller placements were composed of the generators offering the cheapest energy and were sub-sets of the larger blocks.
Time series for a particular scenario were formed by summing the output of all cells which had been selected on the basis of relative lifetime economics. To examine different aspects of the expansion of renewable electricity generation, two types of scenarios were chosen. These technology scenarios and area scenarios are described in detail below.
5.1.1 Technology Scenarios
A series of scenarios was chosen to identify the effects and consequences of different mixes of generation technologies. The first group of scenarios was made up of increasing capacities of a particular technology, either onshore-wind, offshore-wind, wave, or tidal-current. The second group of scenarios comprised various mixes of the four technologies.
Onshore-wind The projected lifetime economics of all feasible onshore-wind developments in Scotland were identified. Detailed results in tabular form are given in Section 6 for the best blocks of 750 MW, 1.5 GW, 3.0 GW and 6.0 GW. To draw graphs indicating trends beyond these capacity limits, a number of further onshore-wind scenarios were made, ranging from 75 MW through to 9.0 GW.
Offshore-wind The capacity and projected lifetime economics of all offshore-wind sites were similarly identified and placed as for onshore-wind. The placements had capacities of 750 MW, 1.5 GW and 3.0 GW. The range of suitable shallow water sites in Scotland suggested that an installation of 6.0 GW would only be feasible with deep-water technology. Therefore no results were calculated at this level. For the presentation of the results in graphical form, some further runs with capacities down to 75 MW were made.
Waves An initial cost analysis determined the estimated lifetime production costs at suitable locations. As wave energy converters need to be arranged in lines facing the wave front, the final placement was done manually. The ranking between cells was then based on the production figures of all the cells with energy converters. Scenarios with 750 MW, 1.5 GW and 3.0 GW of plant were developed. The wave resource is much greater than these numbers would indicate, but due to potential conflicts with navigation it is unknown how much more of the resource can be exploited in areas not fully reported as shipping lanes. The resources may also be greater in real terms than estimated here if the average wind speeds return to customarily higher values than those in the years used to prepare the forecasts. Graphs in Section 6 show an expansion of up to 6 GW.
Tidal-current The number and capacity of tidal-current sites are limited by minimum spring tide velocities of 2 m/s and by the 30 to 50 metre range of water-depths. The fraction of the energy flux that can be exploited without significantly altering the tidal regime depends on the flow characteristics of the channel or sea-area and on the relative 'blockage' to flow caused by turbines. This aspect is likely to be very site specific. In the study, the main placement amounted to 750 MW and this was located manually in the most accessible regions. More tidal-current capacity will become available with later generations of devices that are able to operate in deeper waters. For graphical trend analysis, the capacity was varied from 75 MW through to 1 GW.
Technology mixes Two sets of mixed portfolios of generation technologies were also created. The first set comprised the four technologies always in the same relative proportions but with increasing total capacity. As onshore-wind is likely to dominate the renewable energy mix in the near future, it was assigned 75% of the total capacity. By 2020 the amount of installed offshore-wind and wave power plant may still be comparatively small and they were each assigned 10% of the total capacity of the mixes. As the total exploitable tidal-current resource is likely to be smaller than the wind or wave resources, a 5% contribution was assigned to it. Having assigned the relative contribution of each technology to the mixes, the total capacity was then increased from 750 MW to 6.0 GW in three capacity-doubling steps.
In the second set of mixed portfolio scenarios, the total capacity was kept fixed at 6 GW whilst the proportion assigned to onshore-wind was varied. Thus the contribution of onshore-wind was increased from 0 to 6.0 GW with the balance coming from a 2:2:1 mix of offshore-wind, wave and tidal-current. As with single-technology scenarios, further calculations were made to obtain data for a graphical representation.
Demand For comparison of renewably generated electricity with the load demand, the technology scenarios treated Scotland as a single area. Therefore only one demand time-series was needed for the calculations, the aggregated demand for Scotland. This meant that the SP and SSE three-year time series were added and scaled up to 2020 levels.
Figure 4.2a shows the maximum capacities for each of the four technologies placed: 6.0 GW of onshore-wind, 3.0 GW of offshore-wind, 3.0 GW of wave and 750 MW of tidal-current plant. Each 1 km2 cell is highlighted to make it more visible. Figure 5.2b shows the corresponding map for area scenarios which are discussed in the next section. More details can be obtained from Maps 10 and 11 in the appendix.
Figure 5.2 Energy converter placement for simulation scenarios.
(a) Technology scenarios; (b) Area scenarios.
5.1.2 Area Scenarios
The second series of scenarios was chosen to assess the regional characteristics of the resources. For the purpose of the study, Scotland was divided into ten areas: Shetland, Orkney, Western Isles, Highlands North, Highlands South, North East, North Central, Argyll and Bute, Central and South. Some of the boundaries between areas were based on actual boundaries between planning authorities. Where available, Ordnance Survey information was used to determine the sea boundaries between such areas. Another boundary corresponds to the division between the Scottish and Southern Energy area in the north and the Scottish Power area in the south of Scotland.
In principle the generator placements from the technology scenarios could have been used in the calculation of area scenarios. However, some areas might then have been represented by a small number of machines that were known to be amongst the most economic from a national perspective. This could then have led to anomalously high plant capacity-factors being calculated for the local area. To avoid this problem, and as all the areas are similar in size, equal generating capacities were placed in them.
The total capacities of the area scenarios were reduced compared to the technology scenarios. 3.0 GW of onshore-wind, 1.5 GW of offshore-wind, 1.5 GW of waves and 375 MW of tidal-current plant were placed. All areas can accommodate some onshore-wind power plant. Offshore-wind power plants are more likely to appear in the east and south of Scotland. Nevertheless it would technically be possible to develop projects in each area. Therefore 150 MW of offshore wind was placed into each area. In contrast, wave and tidal-current project developments are unlikely to be evenly distributed across Scotland. Wave power developments will most probably be confined to the northern and western areas including Shetland, Orkney, Western Isles and possibly Argyll and Bute. A quarter of the 1.5 GW total wave capacity was identified in each of these four areas. Tidal-current developments are promising in Shetland, Orkney, Highlands North, Argyll & Bute and South. 75 MW of plant was placed in each area, potentially slightly exceeding the exploitable resource in Shetland.
For the final area scenario, the performance of a mixed portfolio of technologies with a 75-10-10-5% split was examined: 4.5 GW of onshore-wind, 600 MW of offshore-wind, 600 MW of wave and 300 MW of tidal-current capacity.
For each area a particular time-series of load demand was applied based on the appropriate daily and annual patterns, from either Scottish and Southern Energy or from Scottish Power.
5.2 Terminology and Graphs
The presentation of results from the scenarios requires the use of specialised terminology and graphics. The following explanations and definitions may be helpful.
5.2.1 Generation and Demand Curves
As an example, Figure 5.3 shows six hours of output from a 40 MW wind farm connected to the network in the vicinity of a town. The wind farm and town are considered to form an area.
During the first hour there is perfect matching of turbine output and town demand. During the following four hours there is a local shortage of renewable energy. Plant connected elsewhere to the network provides the balancing energy. In the last hour production exceeds demand and the network allows electricity to be exported to another area. However, the generated energy is now higher than the network limit and the distribution network operator ( DNO) curtails the wind farm output to a value below the rated power of the turbine. Note that for graphical distinction, in Figure 5.3 the average demands during each hour are represented by points which are joined by lines.
Figure 5.3 Simplified generation and demand curves for a fictitious area.
From the six hours of data some figures can be calculated which help to illustrate the definition of terms which follow:
- Total energy demand = (20 + 25 + 20 + 30 + 15 + 20) MWh = 130 MWh;
- Total renewable energy supply = (20 + 15 + 5 + 0 + 10 + 35) MWh = 85 MWh;
- Local renewable energy supply = (20 + 15 + 5 + 0 + 10 + 20) MWh = 70 MWh;
- Renewable energy export = 15 MWh;
- Renewable energy not supplied (100% target) = (10 + 15 + 30 + 5) MWh = 60 MWh;
- Renewable energy not supplied (40% target) = (3 + 12) MWh = 15 MWh.
5.2.2 Special Terms
Plant capacity factor The plant capacity factor of a generating unit is calculated as the total of the energy generated divided by the nominal or nameplate rating and the period of time:
. (5.1)
A plant capacity factor of 100% would indicate that the plant was working continuously at its full rating. During the six hours of the example shown in Figure 5.3 a value of 85 / (40 ยท 6) = 35% is reached. It is important to note that the plant capacity factor of a machine depends on its design and that the most economic plant does not necessarily have the highest plant capacity factor. A plant that is over-rated for the chosen site will have a low plant capacity factor, and a plant that is under-rated will spend longer operating at a capacity limited to the nameplate rating. For each of the four technologies used in the study a particular machine was used throughout, so that its performance at different locations could be compared. However, caution is required when plant capacity factors of different technologies are compared.
Long-term gross matching The ratio of generated and supplied renewable energy to demand gives long-term gross matching. For the example of Figure 5.3 a value of 85 / 130 = 65% is reached. Long-term gross matching figures tend to be favourable, as excess of production at a certain hour balances out shortfall at another. The figures are expressed as percentage values and can exceed 100%. A matching of 50% suggests that the total renewable energy produced over the time considered was half of the electrical energy supplied in that region. The stricter approach of using long-term local matching, as described below, was used in the study.
Long-term local matching The long-term local matching of an aggregated group of generators was calculated as the total of the renewably-generated energy that is used to satisfy local demand divided by the total energy demanded in that area:
(5.2)
For the example of Figure 5.3, the long-term local matching reduces to 70 / 130 = 54%. This definition emphasises the shortfall in energy required locally to meet demand and therefore was used throughout the study.
Energy export When renewable electricity production exceeds local demand, then the excess must be exported to other areas or the output of the machines must be reduced. In the example of Figure 5.3 there is one such hour requiring export. Relative to the total energy demand, the amount of exported energy is 15 / 130 = 12%. Note that this figure is the same as the difference between gross and local matching (with the discrepancy of 1% arising from the presentation of the results in integer format).
Output limitation With increasing penetration of renewable generators it is possible that at times the network operator may have to limit the plant output. In the present study, the network was assumed to be ideal and output limitation was not implemented.
Energy shortfall Shortfall is expressed as the percentage of total energy demand in any locality that is not supplied by the corresponding renewable-energy generators:
. (5.3)
When the target demand level to be supplied by renewable generators is 100%, then the sum of energy shortfall and long-term local matching add up to 100%. In the example of Figure 5.3, the energy shortfall (with 100% demand target) is 60 / 130 = 46%. This amount of energy must be supplied by other generators connected to the network. Had no renewable energy been produced at all during the six hours, then the energy shortfall would be 100%.
Within this study, a supply target of 40% of demand was used and figures are calculated relative to this:
. (5.4)
With the values from Figure 5.3, the energy shortfall (with 40% demand target) is now 15 / 130 = 12%. Despite the high long-term local matching value of 54% there is still this amount of energy which could not be supplied by the renewable generators to meet the 40% target on an hour-by-hour basis. Had no renewable energy been produced at all during the six hours, then the corresponding energy shortfall would be 40%.
Exceedance hours When energy shortfall is calculated on an hour-by-hour basis then the absolute hourly energy amounts can be represented in an exceedance curve asrcentageortfall over fractionsts can then be represented in an exceedance curve which energy exceedance or shortfall over frac energy exceedance or shortfall against percentage of time.
In some cases it may be useful to know the number of hours when production exceeds a certain level of demand. In the example of Figure 5.3, 100% of demand is reached or exceeded for 33% of the time, 40% of the demand is exceeded for 67% of the time. For the scenarios in Scotland, the number of hours when the 40% demand level is exceeded was calculated and expressed as a percentage of time.
Coincident hours A long-term local matching value of 50% achieved with renewable energy sources suggests that the total renewable energy produced over a particular season or year is equal to half of the electrical energy consumed within the area. During any shorter time period, the generated energy could be well above or well below the local demand. The use of a single numerical value for matching masks this important information. It is therefore useful to present the statistics of matching in the form of a coincident-hours histogram which is an extension of the exceedance hours concept.
As the time series of six hours in the example in Figure 5.3 above is too short to produce a histogram, data from actual scenarios (with 26,280 time steps) is used to illustrate the concept. Figure 5.4a shows an idealised coincident-hours histogram for an imaginary situation where generated and demanded energy are always in balance. Each column in this bivariate histogram represents the relative amount of time during which a particular combination of generation as a percentage of total capacity and demand as percentage of peak demand is true. The combined heights of all of the columns represent 100%. On the diagonal of green columns capacity-factors and load-factors are matched (to within a 10% tolerance band). The blue area represents times when demand percentage is less than generation percentage and the red area represents times when it is higher. Note that if demand and generation axes were scaled to the same maximum value in MW, then the diagram would look different.
A different situation is illustrated by the coincident-hours histogram of Figure 5.4b for a 6 GW renewable mix across Scotland. Note that the height scale of the columns in both diagrams is different. It can be seen that demand never falls below approximately 30 % of its peak value. Generation, in contrast, varies between zero and total nameplate rating. The number of columns makes it somewhat difficult to compare scenarios, but the extremes are of greater concern. For example, the hours when generation is less than 10% of capacity and demand exceeds 90% of its peak value. This value is represented by the small red column at the very right of Figure 5.4b. Due to its importance, this worst-case coincident hours figure is listed in the spreadsheets of Section 6.
Figure 5.4 Coincident-hours histograms.
(a) Imaginary situation where generation and demand are always matched;
(b) Scenario with 6 GW broad-mix of technologies.
Power output variability Other statistical information that can be extracted from the data includes the rate of change of power output. When renewable output power reduces, for instance because the wind calms down, then balancing plant needs to be brought online. Likewise, the output of balancing plant can be reduced or the plant can be taken offline when enough renewable generation is available. Forecasting and system monitoring allow dispatching centres to act on these events, provided that spare balancing plant is available with output that can be changed at the required rate. The probability of a certain change of renewable output power (as a percentage of installed capacity) can be displayed in form of a histogram. For instance, a high-wind cut-out of a wind turbine represents a -100% change while a subsequent cut-in at strong winds can create a +100% change. An important parameter of the power output variability is the time difference between the two observations. Two different calculations were made in this study, one with a one hour time lag and another with a six hour time lag. The one hour results need to be interpreted with caution since the offshore wind and wave data originated from three-hourly records with interpolations in between. However, detailed time series analysis of wave power records suggests that there is little change from one hour to the next. Example histograms are shown in Section 6.3.
« Previous | Contents | Next »