1. Introduction
Energy is essential for economic and social development and improved quality of life. However, much of the world’s energy is currently produced and used in ways that may not be sustainable in the long term. In order to assess progress towards a sustainable energy future (solar energy, wind energy, tidal energy and geothermal energy), energy indicators that can measure and monitor important changes are needed [
1].
Energy demand is a derived demand that arises to satisfy some needs which are met through the use of appliances. Hence, demand for energy then depends on the demand for energy services and the choice of energy using processes or devices. End-use service demand is affected by the cost of energy but also by other factors such as climatic conditions, affordability (or income of the decision-maker), preference for the end-use service, etc. Similarly, demand for end-use appliances depends on the relative prices of the appliances, relative cost of operation, availability of appliances, etc. [
2]. The expected future trends for these determining factors, which constitute ‘scenarios’, are exogenously introduced. The scenario is a potential image whose future could be revealed; a set of scenarios helps to understand the possible future evolutions of complex systems. For example, a scenario can consider in the long term a series of future socio-economic and technical data (population growth rate, GDP growth rate, structural changes in the economy, improvement of household size, improvement of efficiency) to examine energy demand [
3].
A model for peak-load demand should take into account the following factors or part of them, depending on the country in which this model is going to be implemented. There is no unique model that can be applied to utility companies.
These factors are
- ▪
The gross domestic product (GDP)
- ▪
The population (POP)
- ▪
The GDP per capita (GDP/CAP)
- ▪
The GDP per capita (GDP/CAP)
- ▪
The multiplication of electricity consumption by population (EP)
- ▪
The power system losses (LOSS)
- ▪
The load factor (LF)
- ▪
The cost of one kilowatt hour (the average rate per unit sale; R/S) (mill/kWh).
The first four factors depend on the behavior of the public; thus, they may vary from country to country, whereas the last three factors depend on the electric power system and the load itself, as well as the consumption of power generated [
4].
Electricity load forecasting is a multifaceted task that encompasses various sides of the economical and safety operations of power supply systems. The forecasting provides assistance for security analysis of supply systems, system control, maintenance, etc. It also provides feedback for planning. By considering these important facets of the forecasting, many methods have been applied to predict electricity load forecasting over years to get more accurate results for efficient planning, however, the prediction has still remained a challenging issue that contains many uncertain factors affecting electricity load utilization [
5].
There are mostly three categories of energy consumption studies in the literature according to the forecasting horizon. Long-term forecasting (5 to 20 years) is mostly applied for resource management and development investments. Mid-term forecasting (a month to five years) is mostly applied for planning the power production resources and tariffs, and short-term forecasting (an hour to a week) is mostly used for scheduling and analyses of the distribution network [
6].
Each one of these planning horizons is important; however, long-term forecasting can be considered the most critical horizon due to the consequences of the strategic and costly decisions, such as the expansion of utility plants. An overestimation of electricity demand, especially in long-term forecasting, will result in a significant increase in the construction of unnecessary electricity generation plants. In contrast, an underestimation of electricity demand forecasting will result in a shortage of electricity production and customer dissatisfaction [
7].
It is not surprising that factors affecting prediction horizon types may vary. In the short-term and mid-term load forecasting, the most important factor for load prediction is the complexity of an economy and a weather factor of the whole area, however, in long-term load forecasting, GDP, expected economic situations, changes in the demography, etc. are the most common factors which affect load usage [
8,
9].
The forecasting methods are classified into traditional (e.g., fuzzy logic, grey model, metaheuristic algorithms, regression models, simulation model, time series model) and intelligent (e.g., machine learning and neural network) models. Furthermore, the advantages and disadvantages of both traditional and intelligent forecasting methods as well as research limitations and future research have been determined [
10]. The historic time series data are used in many studies to predict energy consumption, the effect of variables such as GDP, consumer price index (CPI), humidity, temperature, population, energy price, daylight time, number of rainy days, etc. on energy consumption is analyzed in many others. Proposed models can be applied by the producers, suppliers and regulatory authorities who want to securely supply the electricity at a reasonable cost [
11].
Researchers use different models and methodologies to forecast electricity demand. For example, the work in [
12] used a hybrid load-forecasting method that combines classical time series formulations with cubic splines to model electricity load; only the hourly temperature is used in the proposed model and predictive power gains are achieved through the modeling of the 24-h load profiles across weekends and weekdays while also taking into consideration seasonal variations of such profiles. Long-term trends are accounted for by using population and economic variables. The authors of [
13] proposed a deep residual neural network model for day-ahead household electrical energy consumption forecasting and integrated multiple sources of information (weather, calendar and historical load). In [
14], the authors employed a fuzzy logic model for long-term load forecasting for one year based on the weather parameters (temperature and humidity) and historical load data. The work in [
7] presented long-term forecasting of electrical loads in Kuwait using Prophet and Holt-Winters models for ten years based on the real data of historical electric load peaks. The authors of [
15] also designed enhanced machine learning techniques for electricity load and price forecasting, and hourly data of one year are used for the forecasting process. In [
16], the author developed a scenario-based model for the African power system using the Schwartz methodology. The work in [
17] showed the superiority of triple seasonal methods of Holt-Winters and autoregressive moving average (ARMA) models for short-term electricity demand forecasting.
Many future electricity demand projection studies in the literature research have employed a bottom-up approach using the Model for Analysis of Energy Demand—Energy Demand module (MAED_D), which adopts socio-economic and technological development indicators to make long-term projections in the case of Syria, Turkey, Nigeria, Nepal, Tanzania, Bangladesh and Lesotho.
The final energy demand and the final electricity demand of Syria will grow annually. Both final energy and electricity demand growth rates are lower than the corresponding GDP growth rates, which indicate positive development trends in the elasticity evolution. However, the growth of electricity is continuously higher than that of final energy with increasing tendency over the whole study period. This is a direct result of more automation in industry and more electrical equipment in households and services [
18]. The MAED model was executed for the above three scenarios and Turkey’s annual electric energy demand values were predicted from the study period. The result shows that the future electricity demand will be growing rapidly in Turkey [
19]. The forecast of the energy demand of Tanzania for all economic sectors was analyzed by using the Model for Analysis of Energy Demand (MAED) for the study period. In the study, three scenarios were formulated to simulate possible future long-term energy demand based on socio-economic and technological development with the base year. The study results show that the highest growth rate of electricity demand is in the industry sector, followed closely by the service and household sectors [
20].
The second module of MAED, MAED_EL (EL = electricity) was used to determine the annual peak load of future Syrian electric power demand. The results indicate that the current residential behavior of the Syrian power system will shift in the reference scenario more and more to the typical industry behavior characterized by higher load factors. The future peak load will grow annually [
21].
The major contribution of the work in this paper is to improve estimates of future electrical energy needs of Taoussa using the Model for Analysis of Energy Demand (MAED) on the long-term evolution of electricity demand for the sustainable development of the regions of northern Mali. National economic and energy statistics provide the input to the calibration of the model to accurately reflect the current electricity system as well as its interaction with the principal drivers of electricity demand, such as demographics, economic development, technological change, environmental policy, etc. Energy planners can then compare these scenarios (different futures) with their ability to support national development objectives and goals. Critical policy and investment aspects of different energy strategies can be defined, undesirable consequences can be identified, and the most cost-effective approach to meeting future electricity needs can be determined.
The rest of this article is organized as follows.
Section 2 presents general information on the case study.
Section 3 talks about the IAEA’s tools and models.
Section 4 describes scenarios and background information employed in the MAED model.
Section 5 predicts the electricity demand under three scenarios and indicates future research directions, and finally, conclusions are elaborated in
Section 6.
3. Energy System Models of IAEA
3.1. International Atomic Energy Agency’s Tools and Models
With specific progress, the IAEA has developed a set of energy and environmental impact modeling tools appropriate for the execution of this program, as well as several mechanisms for transferring the tools and the necessary expertise to developing countries.
The IAEA transfers these models to the interested member states, along with training and technical help, to make them available for national applications. The models provide a convenient framework for conducting energy studies covering: the analysis and assessment of energy demand; energy technology assessment including technical, economic and environmental aspects; the evaluation of alternative supply strategies for energy and electricity; financial analysis; and the assessment of environmental impacts and other external costs.
Presently, over 150 countries and 21 international organizations are using the IAEA’s energy models. The IAEA’s energy planning tools (models) include:
- ▪
EBS (Energy Balance Studio)—to facilitate collection and organization of energy data;
- ▪
ESST (Energy Scenarios Simulation Tool)—to explore energy system development, offering scenarios in terms of capacity expansion, investment and GHG emissions;
- ▪
MESSAGE (Model for Energy Supply System Alternatives and their General Environmental Impacts)—to analyse energy supply strategies;
- ▪
MAED—to study future energy demand;
- ▪
WASP (Wien Automatic System Planning Package)—to plan power sector expansion;
- ▪
FINPLAN (Financial Analysis of Electric Sector Expansion Plans)—to assess financial implications of a power project;
- ▪
SIMPACTS (Simplified Approach for Estimating Impacts of Electricity Generation)—to analyse impacts to human health and agriculture of a power project;
- ▪
ISED (Indicators for Sustainable Energy Development)—to analyse and monitor sustainable energy development strategies;
- ▪
CLEW (Climate, Land use, Energy, and Water)—to analyse interactions among key resource systems [
26].
The MAED software is used for simulating the annual electricity demand and peak load forecast for the electrification in Taoussa of Mali under different scenarios from 2020 to 2035.
3.2. Description of MAED
The MAED model software is the simulation model (not optimization) and it evaluates future energy demand for the medium and long term (not short term) [
27].
The general structure of MAED is shown in
Figure 2, which also illustrates the flow of information from the various modules and associated data files [
28].
The number assigned to each module indicates the proper sequence of their execution [
28]:
Module 1 (Energy Demand Calculations) processes information describing the social, economic and technological scenario of development and calculates the total energy demand for the desired years. The breakdown of this demand by energy form and by economic sector considered is also provided as part of the results of the analysis. This module creates a file (FT12: “LOADSCEN”) to be used later by Module 2. A scratch file (FTll: “TEMP”) is also used as a temporary working file by Module 1.
Module 2 (Hourly Electric Power Demand) uses the total annual demand of electricity for each sector (contained in file “LOADSCEN”) to determine the total electric power demand for each hour of the year or, in other words, the hourly electric load which is imposed on the power system under consideration. During its execution, this module creates a file (FT13: “LOADYEAR”) to be used by Module 3.
Module 3 (Electric Load Duration Curve) uses the hourly loads (contained in file “LOADYEAR”) to produce the load duration curve of the power system as required for the execution of a WASP study. Two output files are created by this module during execution. The first one, file (FT14: “LOADWASP”), contains the WASP input data, and the second (FT15: “LOADPLOT”) contains the same information but is presented in a different format as required for plotting the load duration curve. (At IAEA, plotting of curves is done by means of a software product, i.e., TELL-A-GRAF, marketed by Integrated Software Systems Corporation, U.S.A.)
Module 4 (Load Modulation Coefficients) is an auxiliary module of MAED-1 which may be used to analyze the past evolution of the coefficients, describing the variation of the hourly electric loads based on load curves determined from statistical data. An additional input file, i.e., “LOADCHRO”, with the chronological electric power demand hour-by-hour for past years of statistics is required for executing this module.
5. Results and Discussions
The main objective of the project was to contribute to the sustainable development of the regions of northern Mali, as well as poverty reduction, environmental management and specifically for the long term, to produce 332.57 GWh/year, 472.61 GWh/year and 647.02 GWh/year electricity in the low, reference and high scenarios by 2035.
5.1. Final Electricity Demand by Sector
In
Table 10,
Table 11 and
Table 12, electricity demand in 2020 was 15.41 MW (106.84 GWh) for all scenarios; the average annual growth rate of electricity demand is slightly higher during the 2021–2035 period, 8.13% in average from 17.88 MW (117.28 GWh) to 49.40 MW (332.57 GWh) for low LS, 10.31% in average from 18.62 MW (129.97 GWh) to 66.46 MW (472.61 GWh) for RS and 12.56% in average from 19.53 MW (136.65 GWh) to 89.47.01 MW (647.02 GWh) for HS. The share by sector is 56.77% in industry, 25.60% in household and 17.63% in service for LS, 56.13% in industry, 28.55% in household and 15.31% in service for RS and 61.16% in industry, 25.67% in household and 13.17% in service for RS, respectively. We construe that electricity demand for the industry sector will be equal to household and service demand.
5.2. Demography and Comparative Evolution of GDP by Scenario
The projection for the total population and GDP of the Taoussa area are presented in
Table 11. It has been found that the demography will grow annually at an average rate of 3.97% from 0.754 million people in 2020 to 1.351 million people in 2035. GDP in 2020 was USD 1.30 billio for all scenarios and the annual average GDP growth rates will be 2.33% from USD 1.33 billion in 2021 to USD 1.84 billion in 2035, 4.20% from USD 1.35 billion in 2021 to USD 2.41 billion in 2035 and 6.33% from USD 1.39 billion in 2021 to USD 3.27 billion in 2035 for LS, RS and HS, respectively.
The forecast annual average growth rate of GDP/Cap and electricity demand/Cap is shown in
Table 11. During 2020–2035, the GDP/Cap will be 4% from 0.08 kWh/USD to 0.19 kWh/USD for the LS, 6.3% from 0.08 kWh/USD to 0.20 kWh/USD for the RS and 8.5% from 0.08 kWh/USD to 0.19 kWh/USD for the HS, whereas the electricity demand/Cap is projected to account for 5.6% from 0.14 MWh/cap to 0.25 MWh/cap, 6.1% from 0.14 MWh/cap to 0.35 MWh/cap and 6% from 0.14 MWh/cap to 0.47 MWh/cap in the LS, RS and HS, respectively.
5.3. Total Annual Electric Energy Demand and Peak Load
Simulation of total hourly electricity demand forecast for the period from 2020–2035 must meet the peak load time, a short period of critical time during which electricity consumption is highest within a year and could have a strong influence on the reliability of electricity supply.
In
Table 12, the times ahead of system peak load illustrate that the average annual growth rate of the system peak between 2020 and 2035 represents 7.92% in the Low scenario; in this case, the system peak demand will increase from 20.8 MW in 2020 to 64.88 MW in 2035. The system peak electricity demand time of day in the Reference scenario is forecasted to increase to 10.53% per year from beginning to end of the study period and from 25.35 MW in 2021 to 92.2 MW in 2035. On the other hand, the system peak demand in the High scenario is forecasted to increase at approximately 12.91% annually, from 26.65 MW in 2021 to 126.22 MW in 2035.
The information on total electric energy by month in Gwh for all the scenarios is presented in (
Appendix A). This allows for visualizing the proportion of months in which consumption is higher than a certain level of power; the lowest consumption is in the cool season in January from 4.89 GWh in 2020 for all scenarios to 15.18 GWh for LS, 21.57 GWh for RS and 29.53 GWh for HS in 2035, respectively, while the month of high consumption is in the hot season in May from 12.40 GWh in 2020 for all scenario to 38.37 GWh for LS, 54.52 GWh for RS and 74.64 GWh in 2035, respectively.
5.4. Comparative Evolution between Annual Electricity Demand and Peak Load by Scenario
The difference between the annual electricity demand and peak load forecasts was relatively big for all scenarios. This difference will increase at an average annual growth rate of about 7.49%, 10.31% and 13.87% in the Low scenario from 5.39 MW in 2020 to 15.48 MW by 2035, in the Reference scenario from 6.73 MW in 2021 to 25.74 MW by 2035, and in the High scenario from 7.12 MW in 2020 to 36.75 MW by 2035, respectively.
The ways to reduce the electricity demand during peak load include replacing energy-intensive appliances with more energy-efficient ones, shifting the operation of electricity-consuming equipment during a peak period to periods of inactivity (off-peak hours), organizing the implementation of equipment according to a plan, raising awareness of the population towards energy saving and relieving certain groups of necessary equipment.
5.5. Annual Load Duration Curves by Scenario
From
Figure 5,
Figure 6 and
Figure 7 show the annual load duration curves (LDC) permit to view the proportion of the time during which the consumption is peak load or minimum load. It is shown that the peak loads during 2020 to 2035 are in the hot dry season between 17–23 May at 9:00 p.m., and the minimum load is between 15–21 January at 7:00 a.m.
The electricity demand system load is divided usually into four seasons;
Figure A1 shows the average electricity demand for each day by season at the same time. In the Taoussa area, there are three types of days (workdays, Fridays and weekends); the workdays include four days (Monday to Thursday) and the weekend includes two days (Saturday and Sunday). Taoussa’s electricity demand is virtually the same for all-day types in the dry cool 1 season and dry cool 2 season. Generally, for all-day types, peak hours are from 8:00 p.m. to 11:00 p.m., and off-peak hours are from 3:00 a.m. to 9:00 a.m.
Future work will be based on the hybrid renewable energies (solar power, hydroelectric power, wind power) optimization technique using the MESSAGE model to satisfy this electricity demand.
6. Conclusions
During the study period 2020–2035, MAED analysis has shown that the GDP, electric capacity and electricity demand will increase to USD 1.84 billion and 49.40 MW (332.57 GWh) for the Low scenario (LS), USD 2.41 billion and 66.46 MW (472.61 GWh) for the Reference scenario (RS), and USD 3.27 billion, 89.47 MW (635 GWh) for the High scenario (HS), respectively. The total electricity demand in the Taoussa area increased at an average rate of 8.13% in the LS, 10.31% in the RS and 12.56% in the HS in all sectors. The demographic was identical for all scenarios and will increase to 1.351 million people.
The industry sector (including manufacturing, construction, mining and agriculture) will be the biggest electricity consumer of the Taoussa area. Under the calculated growth, the system peak load in the planning horizon and the electricity peak demand is expected to grow at about 7.92%, 10.53% and 12.91% corresponding respectively to the three scenarios (LS, RS and HS). The days of peak load are in the hot dry season between 17–23 in May. The growth rate has gradually increased year by year. Based on the calculation results, the following conclusions can be drawn. There is a need for energy conservation and energy efficiency measures in all electricity sectors, which can reduce the future electricity demand, decrease the greenhouse gas emissions and consequently fight against climate change and bring down the public investment in the electricity production in the Taoussa area.