A3.1 INTRODUCTION
If rural energy demand is to be modelled for all sectors, all major fuels and nine separate regions within Botswana, as well as on a national basis, a large database will have to be set up as input for the model algorithms.
In addition, the number of calculations involved in such an operation is such that any kind of sensitivity analysis necessitates the use of dedicated computer programs.
By developing an appropriate suite of programs, the following major advantages are gained:
- the setting up of a large database, defining many rural energy demand scenarios is facilitated;- the model can be run for several different scenarios with little effort and consistent mathematical accuracy;
- the model results and output can easily be summarised in tabulated form.
A3.2 APPROACH
The approach adopted for the development of this computer model is one that attempts to maximise the advantages outlined above. The following criteria were adopted when designing the software structure:
- ease of use for the casual user;
- flexibility and sensitivity for more in-depth analysis;
- simplicity of scenario definition;
- clearly tabulated results.
To enable these criteria to be satisfied, several features were specifically built in to the software:
- a menu-driven system of options by which the user accesses the various functions of the model;- the facility to define a base-case scenario which incorporates many assumptions which are unlikely to change. This base-case scenario can then be used as the skeleton for other scenarios, so saving the user the effort of redefining many model parameters every time a new scenario is to be defined;
- total editing power to change every parameter used by the model, should this be required.
A3.3 SYSTEM SOFTWARE
The choice of system software was strongly influenced by the requirement that ease of database management should have a higher priority than run-time speed of the model. Considering also, that the model may need to be updated in the future (either to change base parameters or to build in new options/sectors/regions/fuels), it was decided that the database-management software package, dBASE II, should be utilised for software development.
This system software allows easy development and updating of database files to hold both scenario input and output data. In addition, it provides a structured programming language by which to manipulate the data held in the files.
A3.4 MODEL STRUCTURE AND DATA REQUIREMENTS
Each sub-model has its own individual structure and assumptions associated with it. However, in general, the model requires information in the following categories:
- 1981 for regional population data, 1984 for per capita fuel consumption rates;- initial (1985), (1) values of a particular parameter, e.g. per-capita woodfuel consumption, regional population or number of small scale irrigation schemes; information on how a parameter changes over time, which may be in three forms:
i. % annual growth rates over a given number of growth periods;ii. the value of a parameter at the end of each of a given number of periods (model linearly interpolates);
iii. the increase (decrease) in the value of a parameter over a given period (i.e. the value to be added/subtracted over that period).
(1) 1981 for regional population data, 1984 for per capita fuel consumption rates.
In later sections the parameters relating to each sub-model are defined, together with their inter-relationships and the ways in which the Editor allows them to be changed.
To illustrate the typical manner in which a set of data could be edited, the first sub-model, for Domestic Woodfuel (Section A2.6), is explained in rather more detail.
A3.5 USING THE MODEL
The model options are accessed via a full-screen formatted menu system.
Each model run uses input data defining a future rural energy-demand scenario. The input data assigns values to such parameters as:
- number of population growth periods for region X, together with associated annual % growth rates;- number of forecast dates for the scenario;
- growth (positive or negative) of annual per-capita consumption of a given fuel over certain periods;
- change in efficiency of government boilers over defined period;
- growth in % of land cultivated by tractor.
All the data necessary to run the model already exist on the user-disc database under the scenario name "BASE", made for modelling purposes in the ERL report "A Study of Energy Utilisation and Requirements in the Rural Sector of Botswana".
The basis of the approach adopted is that the user may base his scenario on any scenario previously defined (Including the BASE case). The data defining this existing scenario is effectively copied, and may then be edited to match the requirements of the scenario being created.
A3.6 DOMESTIC SECTOR
The domestic sector sub-model draws heavily on the results of ERL's village survey in terms of per capita fuel demands.
A.3.6.1 Woodfuel
From the survey, it was possible to derive mean woodfuel per capita demand figures for both the summer and winter seasons. These per capita figures were sorted according to the wealth class 1/ and size of the households surveyed. Thus, by weighting the summer and winter figures, on the assumption of an 8-month summer period, a mean per-capita demand (kg/person/day) matrix was obtained. (See Table A3.6 (a))
1/ As defined in ERL's report.
Table A3.6 (a) - Mean Per Capita Demand (kg/person/day)
|
|
Household Size | ||
|
Wealth class of household |
1-4 |
5-8 |
9+ |
|
1 |
3.3 |
1.47 |
0.89 |
|
2 |
2.73 |
1.49 |
0.92 |
|
3 |
2.51 |
1.34 |
0.92 |
The summer and winter mean annual matrices are given in Table A3.6 (b).
Table A3.6 (b) - Annual Per Capita Woodfuel Demand (kg)
|
|
Winter |
|
Summer |
||||
|
Household Size |
Household Size |
||||||
|
1-4 |
5-8 |
9+ |
1-4 |
5-8 |
9+ |
||
|
Wealth Category |
|
Wealth Category |
|
||||
|
1 |
1,672 |
777 |
438 |
1 |
986 |
420 |
266 |
|
2 |
1,259 |
785 |
496 |
2 |
865 |
427 |
256 |
|
3 |
1,069 |
719 |
467 |
3 |
840 |
376 |
270 |
The other information needed, on a yearly basis, to model woodfuel demand in this way is:
- regional population;
- mean size of each household-size category, for each region;
- regional household distribution, according to wealth class and household size.
The first two pieces of information have been extracted from the 1981 Census data for small settlements. The household distribution matrix was derived by combining these data with wealth and household size data from the consultants 5-village rural survey, and by assuming a correlation between village wealth and the distribution of households, by wealth class, within the village.
From the survey results it was possible to build up such a matrix for each village surveyed. The wealth assignment was made up using a criteria-related points system which revolved around household possessions considered to be valid wealth indicators, e.g.
- cattle;
- pit latrine;
- permanent roofing and/or walls/floors.
Using the Census data and a wealth assignment system based on similar criteria, a 'village wealth' index was assigned to every village in each region (>500, <5,000 population). These included the five villages surveyed, and their wealth indices covered most of the wealth index range which occurred elsewhere, i.e. from Lecheng, at the lower end of the scale through to Good Hope, at the upper end.
Then, by assuming the household wealth distribution within each village surveyed was representative of most other villages of a similar village wealth index, a household distribution was built up for each village according to its basic wealth index and its household sizes information from the Census data.
By aggregating these data on a regional basis, the mean household distribution matrices for each region were estimated, after including the households in villages of <500 population (all assumed to be of the poorest wealth level).
The Regional Household Distribution Matrices so derived are shown in Table A3.6 (b) - the matrix values represent the fraction of all households in the region which fall into each size/wealth category.
Therefore the data relating to Domestic Woodfuel Demand which can be changed for the purposes of scenario modelling are:
- per-capita annual demand matrix;
- regional household distribution matrix;
- regional population.
It is assumed that the Mean Household size of each size category is constant.
Per capita demand matrix
The following options are available:
- to alter the values of the 1983 matrix which the model requires as initial input;
- to keep this matrix constant over time;
- to allow the matrix values to change over time, e.g. to allow for substitution by other fuels.
Table A3.6 (c)
|
|
Household Size |
||||
|
1-4 |
5-8 |
9+ |
All |
||
|
Region 1 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.02 |
0.03 |
0.01 |
0.06 |
|
|
2. |
0.06 |
0.07 |
0.03 |
0.15 |
|
|
3. |
0.31 |
0.33 |
0.14 |
0.79 |
|
|
Total |
0.40 |
0.42 |
0.18 |
|
|
Region 2 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.01 |
0.00 |
0.00 |
0.01 |
|
|
2. |
0.02 |
0.01 |
0.01 |
0.04 |
|
|
3. |
0.36 |
0.40 |
0.19 |
0.95 |
|
|
Total |
0.38 |
0.42 |
0.20 |
|
|
Region 3 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.02 |
0.02 |
0.01 |
0.05 |
|
|
2. |
0.05 |
0.05 |
0.02 |
0.11 |
|
|
3. |
0.33 |
0.34 |
0.17 |
0.84 |
|
|
Total |
0.39 |
0.41 |
0.20 |
|
|
Region 4 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.02 |
0.02 |
0.01 |
0.04 |
|
|
2. |
0.04 |
0.05 |
0.03 |
0.12 |
|
|
3. |
0.32 |
0.33 |
0.19 |
0.84 |
|
|
Total |
0.38 |
0.40 |
0.23 |
|
|
Region 5 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.04 |
0.04 |
0.02 |
0.11 |
|
|
2. |
0.07 |
0.07 |
0.03 |
0.16 |
|
|
3. |
0.29 |
0.30 |
0.14 |
0.74 |
|
|
Total |
0.40 |
0.41 |
0.19 |
|
|
Region 6 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.02 |
0.02 |
0.01 |
0.06 |
|
|
2. |
0.03 |
0.04 |
0.02 |
0.09 |
|
|
3. |
0.31 |
0.36 |
0.19 |
0.85 |
|
|
Total |
0.36 |
0.42 |
0.22 |
|
|
Region 7 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.04 |
0.04 |
0.01 |
0.09 |
|
|
2. |
0.05 |
0.05 |
0.02 |
0.12 |
|
|
3. |
0.47 |
0.25 |
0.09 |
0.80 |
|
|
Total |
0.55 |
0.33 |
0.12 |
|
|
Region 8 |
|||||
|
Household Wealth |
|||||
|
|
1. |
0.06 |
0.06 |
0.03 |
0.15 |
|
|
2. |
0.08 |
0.08 |
0.04 |
0.20 |
|
|
3. |
0.21 |
0.27 |
0.17 |
0.65 |
|
|
Total |
0.35 |
0.41 |
0.25 |
|
Regional household distribution matrix
The same options which are available for the per capita demand matrix are also available for this matrix, only that each region has its own distribution matrix (to allow for household size and wealth trends across the country) - see Table A3.6 (c).
A.3.6.2 Paraffin and Gas
The parameters for paraffin and gas are similar to woodfuel except that the annual per capita demand matrix contains only three values, one for each wealth category.
A3.6.3 Electricity
Because national grid-electricity demand is modelled as a category in its own right, domestic and commercial demand for this fuel is dealt with under that sub-model section.
A3.7 COMMERCIAL SECTOR
Parameters:
- per capita demand figures (wood, gas, paraffin, diesel) in villages where commercial fuel consumption is considered to be significant (villages of population >800 were considered in ERL's report, following its rural survey); and- % of the regional population living in these so-called "commercially-active" villages.
The per capita consumption figures apply to all regions, whereas the population fractions can be set separately for each region. The consultants' base case used the following percentage fractions, representing, for each region, the population living in villages of >800 people (1983).
Table A3.7 (a) - Percentage of regional population living in villages of> 800 people
|
Region |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
|
Pop'n % |
40% |
15% |
31% |
29% |
29% |
18% |
100% |
20% |
24% |
This fraction can be changed over time (one period of change only) by entering:
- start date;
- end date;
- % fraction at end date.
A3.8 AGRICULTURE SECTOR
The agriculture sub-model consists of three further sub-models:
- borehole pumping demand;
- irrigation pumping;
- tractors/machinery;
A3.8.1 Borehole Pumping
Because a similar approach is used, pumps for large and small village water supply are treated here together with those for cattleposts and lands areas.
Large villages
Parameters:
- number of pumps in 1983;
- mean annual diesel consumption (tonnes/pump/year);
- number of growth periods (max 3);
- end date of each period;
- annual % growth rate for each period.
Small villages
Parameters:
- small villages connected to supply in 1983;
- average number of pumps per village (x.x);
- mean annual diesel consumption/per pump (tonnes);
- number of growth periods (max 4);
- end date of each period;
- number of extra villages connected during each period.
Cattleposts/lands
Parameters:
- number of operating boreholes in 1983;
- average annual diesel consumption/pump (tonnes);
- number of growth periods (max 2);
- end date of each period;
- number of pumps added during each period.
Electric/RET pump substitution
To calculate the effect on diesel and grid-electricity demand, substitution is allowed for.
Parameters are the same for each type of pump:
- number of growth periods (max 5);
- end date of each period;
- total number of pumps introduced/substituted by the end of the period;
- mean annual diesel saving of each pump (tonnes/pump/year).
If the first real growth period does not begin in 1983, a dummy growth period must be included where no pumps are introduced. The end date of this period will be the start date of the first real growth period. This applies to all other sub-models where no start date is specified.
A3.8.2 Irrigation Schemes
Again three further sub-models for:
- existing small schemes;
- existing large schemes;
- projected schemes.
Existing small and large schemes
Parameters the same for large and small:
- number of schemes in 1983;
- mean annual consumption rate/pump (tonnes);
- number of growth periods (max 3, starting 1983);
- annual % growth rate for each period.
Projected schemes
Data can be entered on any number of schemes, as follows:
- scheme name (15 characters max);
- number of growth periods (max 6);
- area (ha) under irrigation by end of each period;
- diesel consumption (tonnes/ha/yr).
A3.8.3 Tractors/Machinery
Parameters:
- area of farmland under cultivation in 1983 (ha);- % cultivated by tractor in 1983;
- tonnes paraffin used for agricultural purposes in 1983;
- tractor fuel consumption (litres of diesel/paraffin mix) per hectare of land ploughed, sowed and harrowed - 1983 value;
- number of growth periods (max 4, starting 1983);
- end date of each period;
- total area under cultivation by each end date;
- % cultivated by tractor by each end date;
- paraffin usage by each end date;
- tractor fuel consumption rate by each end date.
A3.9 GOVERNMENT SECTOR
This sub-model considers demand for the following fuels for electricity generation in government buildings:
- diesel;
- grid electricity;
- coal;
- renewables (e.g. photo-voltaics).
Parameters:
- diesel generated power in 1983 (GWh);
- BPC grid power in 1983 (MWh);
- renewables power in 1983 (MWh);
- coal generated power in 1983 (GWh);
- efficiency of government boilers in 1983 (xx%);
- number of growth periods (max 3) for total generated demand;
- end date of each period;
- annual % growth rate of total demand during each period;
- BPC grid power supplied to rural areas by the end of each period;
- renewables power by the end of each period;
- coal power by the end of each period;
- efficiency of government boilers by the end of each period.
A3.10 GRID ELECTRICITY SECTOR
Demand from waterpumps and government installations are modelled in those sub-models, respectively. Parameters are also required, however, for domestic and commercial demand, and since national figures are required, the data are dealt with outside of the main, regionalised Commercial and Domestic sectors.
Parameters:
- domestic demand in 1983 (MWh);
- commercial demand in 1983 (MWh);
- 'other' (miscellaneous) demand in 1983 (MWh);
- number of growth periods (max 5, starting 1983);
- end date of each period;
- domestic annual % growth rate for each period;
- commercial annual % growth rate for each period;
- 'other' annual % growth rate for each period.
