Scenarios for Angola's Electricity Transition: Generation Capacity, Electricity Production, and Environmental Sustainability

1. Introduction

The electricity sector in Angola has been undergoing a period of rapid transformation, marked by rising demand for electricity, the need to diversify the generation mix, and the rehabilitation and expansion of infrastructure inherited from the period of conflict—particularly hydroelectric power plants and transmission and distribution networks [1]. Hydroelectric power plants remains the main source of electricity generation, supplemented by fossil-fuel-fired thermal power plants and a still-small share of new renewable energy sources, such as solar and wind power [1,2]. The sustainable expansion of Angola’s energy system is essential to balancing security of supply, universal access to energy, and the reduction of energy poverty with the need to limit CO₂ emissions and align with global climate change mitigation goals [1,3]. Electric power planning studies for specific regions of Angola, such as Namibe Province, show that combining increased hydroelectric capacity with the integration of solar, wind, and natural gas sources can reduce generation costs and greenhouse gas emissions, while increasing the reserve margin and system reliability [2].

The energy transition in emerging economies faces specific challenges: heavy reliance on fossil fuels, institutional and regulatory constraints, financial constraints, gaps in technical capacity, and the need to ensure a just and inclusive transition [4,5]. In this context, long-term energy planning plays a central role, using generation expansion planning (GEP) models to determine the optimal mix of technologies, minimizing total cost subject to technical, environmental, and security-of-supply constraints [6,7,8]. Recent reviews show that these models are rapidly evolving to incorporate high penetrations of renewables, energy storage, and interactions with carbon markets [6,7,8,9,10,11,12]. In the case of Angola, the literature has focused primarily on life-cycle analyses of the electric power system and on regional scenario studies using tools such as Long-range Energy Alternatives Planning System (LEAP) [1,2], highlighting the need to more explicitly integrate econometric demand modeling with long-term capacity expansion models. The main objective of this study is to develop a hybrid model that combines an econometric demand projection module with a capacity expansion model, analyzing the evolution of Angola’s electric power system under different development scenarios through 2050. The model assesses implications in terms of required additional capacity, generation mix composition, system costs, emission factors, and carbon intensity, thereby contributing to Angola’s sustainable energy planning. Electricity demand projection is fundamental to long-term energy planning, especially in developing countries [13,14]. In many models, demand is treated as an exogenous variable, derived from the growth trajectories of Gross Domestic Product (GDP), population, and energy prices [15,16], without adequately taking into account structural changes in the economy, the diffusion of technology, and improvements in energy efficiency [13,17]. The lack of long and reliable time series data limits the sophistication of demand models in transition economies; it is common to find planning studies in Africa that extrapolate demand based on historical trends, without rigorously calibrating the econometric parameters [17,18,19]. Most planning practices in low- and middle-income countries use relatively simple econometric models with limited parameter calibration [16,20], which reduces the ability to evaluate alternative scenarios and introduces uncertainty into long-term assessments [14,17]. Recent literature highlights the importance of hybrid approaches that link demand modeling to power expansion models [13,19], which have yet to be fully explored in the African context.

The challenge facing GEP—determining the type, size, location, and installation schedule of new power plants—has evolved from minimum-cost approaches to multi-objective models that incorporate uncertainty regarding resource availability, fuel prices, and demand [6,7,8,11]. The need to model system operations with temporal detail led to the development of models with high hourly resolution, capable of handling variable renewable generation and storage technologies to enable decarbonization scenarios [9,12,22]. In the specific case of Angola, existing studies focus on life-cycle analyses assessments of the electric power system [1] and in regional scenario analyses using LEAP [2]. However, there is still no comprehensive assessment of Angola’s electricity system that combines demand, optimal capacity, costs, and emissions through 2050, nor is there a robust analysis of macroeconomic and technological uncertainties. This study addresses these gaps through a hybrid model with sensitivity analysis applied to the 2024–2050 horizon.

2. Characterization of Energy Resources

2.1. Solar Resource

Figure 1a) and 1b) show the spatial distribution of global horizontal irradiation (GHI) and photovoltaic power potential in Angola. Both indicators increase from the northwestern coast toward the central and southern regions, with annual irradiation ranging from 1,607 to 2,433 kWh/m² and photovoltaic power potential from 1,241 to 1,972 kWh/kWp. The highest values are concentrated in Namibe, Cunene, Huíla, Cuando and Cubango, confirming their excellent suitability for utility-scale photovoltaic deployment, whereas the northwestern coastal region, particularly Cabinda, presents comparatively lower solar potential due to maritime influences.

2.2. Water Resources

Angola’s hydropower resources are estimated at approximately 18 GW, corresponding to an annual generation potential of about 72 TWh [24]. Around 86% of this potential is concentrated in the Kwanza, Queve, Catumbela, and Cunene, river basins, with nearly 100 identified sites, several exceeding 10 MW [24]. Figure 2 shows the spatial distribution of hydropower potential, confirming the dominance of these river basins, whereas the southern regions exhibit comparatively lower potential.

2.3. Biomass Energy

Biomass resources in Angola are mainly derived from forest residues and the sugar industry, with an estimated potential of up to 170 MW in some locations, particularly in the central and eastern regions [23]. The Ministry of Energy and Water identified 42 biomass projects totalling 3.7 GW, including 3.3 GW from forest biomass and more than 350 MW from sugarcane residues [23]. Figure 3 shows the spatial distribution of land cover, highlighting the concentration of forested areas in the northern and central regions, which confirms their greater biomass resource potential.

2.4. Wind Energy

Wind energy potential in Angola is concentrated along the Atlantic coast, particularly in the southwest, where average wind speeds at 80 m exceed 6 m/s [23]. The Ministry of Energy and Water identified 13 wind projects with a total potential capacity of 3.9 GW; several located near existing transmission infrastructure with up to 604 MW of grid connection capacity [23]. Figure 4 shows the spatial distribution of wind resources, confirming that Namibe has the highest wind potential, while most inland and northern regions remain below 5 m/s, making the southwestern coast the most suitable area for large-scale onshore wind development.

3. Materials and Methods

The proposed methodology projects the development of the Angolan electricity sector through the year 2050. The model assesses the expansion of installed capacity, technical electricity generation, carbon dioxide emissions, and changes in economic costs using life-cycle indicators. The simulations are based on historical data series from 2010 to 2024, as well as on the strategic energy planning targets established by the Government of Angola [24,26,27]. The overall methodological procedure is outlined in Figure 6. Numerical processing, econometric calibration, and the solution of dynamic algorithms were performed in the MATLAB technical computing environment, using a platform equipped with an Intel® Core™ i7-5500U processor (2.40 GHz), 16 GB of RAM, and the 64-bit Windows 11 Pro operating system.

3.1. Data Sources and Technical Parameter

The model is based on historical time series for installed capacity, electricity generation, GDP, and population for the period 2010–2024, presented in Table 1, which were used to calibrate and validate the dynamics of the Angolan energy system. The technical, economic, and environmental parameters by technology are presented in Table 2, including CAPEX, OPEX, useful life, capacity factors, and operational and life-cycle emission factors, compiled from scientific literature and international sources. The assumed capacity factors reflect typical operating conditions; for biomass, a range of 60–75% was considered, based on resource availability and subject to sensitivity analysis [23]. Monetary variables were harmonized in euros using an exchange rate of 0.92 EUR/USD (Bank of Portugal, May 20, 2026), ensuring consistency in the economic evaluation of the LCOE and Levelized Cost of Hydrogen (LCOH) indicators.

Figure 5. Methodology used.

Figure 5. Methodology used.

Table 3 summarizes the model’s overall parameters, including the discount rate, technological learning, fuel prices, and scenario assumptions for the 2024–2050 horizon. In the diesel–solar hybrid system, CAPEX was estimated by weighting the costs of the photovoltaic and diesel subsystems.

3.2. Historical Calibration and Demand Elasticity (2010–2024)

The model is calibrated using the observed time series from 2010 to 2024, employing a multiple linear regression based on natural logarithms to estimate the elasticity coefficients of Gross Domestic Product (GDP) and installed capacity. The econometric model adopted from Grimaldo et al. [50], and Omotola et al, [51], described by the Equation (1).

ln(G_hist,t) = β₁ + β₂ln(PIB_real,t) + β₃ln(Cap_GW,t) + β₄D_expansion,t

(1)

where: ln(G_hist,t), ln(PIB_real,t) e ln(Cap_GW,t) express the variables in natural logarithms; β₁ model intersection; β₂ Historical Elasticity of Demand Relative to GDP; β₃ historical elasticity of demand relative to installed capacity; β₄ Impact Coefficient of Structural Expansion; PIB_real,t real GDP at constant prices; Cap_GW,t total installed capacity. The dummy variable D_{expansion ,t} reflects the structural shift in electricity supply (2017–2021) associated with Laúca and Cambambe II. The statistical validation of the calibration is assessed through R² and MAPE. The R² measures the explained variance, according to the equation (2) [52]:

$$R^2=1-{\displaystyle \frac{{\mathit{SS}}_{\mathit{res}}}{{\mathit{SS}}_{\mathit{tot}}}}=1-{\displaystyle \frac{{\displaystyle {\sum }_{i=1}^N}{(y_i-f_i)}^2}{{\displaystyle {\sum }_{i=1}^N}{(y_i-\overline{y})}^2}}$$

(2)

where: R² coefficient of determination; SS_res represents the sum of the squares of the residuals and SS_tot defines the sum of the squares around the sample mean $\overline{y}$. In addition, the absolute magnitude of the residual error was quantified using the MAPE, calculated according to the equation (3) [52]:

$$\mathit{MAPE}={\displaystyle \frac{100\mathrm{\%}}{N}}{\displaystyle \sum _{j=1}^N}\left|{\displaystyle \frac{y_i-f_i}{y_i}}\right|$$

(3)

where: MAPE mean absolute percentage error, and in Equations (1.2) and (1.3), y_i actual historical value of demand; f_i value estimated by the model and N = 15 total number of observations. Figure 7 shows the results of the historical calibration and statistical validation of the model.

Figure 6. Historical calibration and validation of the electrical model for Angola (2010–2024): a) comparison between actual generation [G] and model generation [m]; b) statistical validation (R²) and (MAPE).

Figure 6. Historical calibration and validation of the electrical model for Angola (2010–2024): a) comparison between actual generation [G] and model generation [m]; b) statistical validation (R²) and (MAPE).

The model exhibits high fit to the historical data, with an R-squared value of 0.9659 and a MAPE of 5.53%. These results confirm the model’s predictive robustness and the validity of the calibration for the BS, GTS, and RDS scenarios, falling below the 10% threshold commonly accepted in the energy planning literature.

3.3. Demographic Trends and Projected Macroeconomic Elasticity (2024–2050)

For the 2024–2050 period, the population and GDP grow annually at the exogenous rates specified for each scenario. To project these variables, we adopt the approach proposed by Kitessa et al. [53], as expressed in Equations (4) and (5):

Pop_t = Pop_t−1 + (GR_Pop,t−1× Pop_t−1)

(4)

GDPP_t =GDPP_t−1 + (GR_PIB,t−1× GDPP_t−1)

(5)

where: Pop_t is the population for the year t; Pop_t−1 the previous population; GR_Pop,t−1 population growth rate; GDPP_t is the GDP for the year t ; GR_PIB,t−1 the rate of economic growth and GDP_t−1 the initial value. The elasticity of demand follows a dynamic, nonlinear trajectory linked to GDP per capita (GDP_pc,t = GDPP_t/Pop_t), adopting the generalized logistic function of Khalili et al. [54], as given by Equation (6):

$${\epsilon }_{\mathit{din},t}={\epsilon }_{\mathit{min}}+{\displaystyle \frac{{\epsilon }_{\mathit{max}}-{\epsilon }_{\mathit{min}}}{1+e^{\left(\frac{{\mathit{GDP}}_{\mathit{pc},t}-{\mathrm{GDP}}_{\mathit{ref}}}{{\alpha }_{\mathit{pc}}}\right)}}}$$

(6)

where: ε_din,t dynamic elasticity over the time period t; ε_min = 0.85 e ε_max = 1.45 are the lower and upper operational limits of elasticity; GDP_pc,t It is GDP per capita in dollars (USD/capita) that drives the energy transition; GDP_ref = 8695.65 USD corresponds to the reference income level at which elasticity reaches half of its transition; e α_pc = 2173.91 USD is the model's dimensional scale constant.

The value of GDP_ref it was set above the all-time high recorded in Angola (6,378.59 USD per capita in 2014), ensuring a gradual transition in elasticity between emerging and mature economies. Thus, total electricity demand is dynamically projected by coupling effective rates and energy efficiency, based on the methodological approaches of Liddle et al. [55] and Khalifa et al. [56], as shown in Equation (7):

$$G_t=G_{t-1}\times \left(1+{\epsilon }_{\mathit{din},t}\times {\mathit{GRPIB}}_{t-1}\right)\times \left(1+{\displaystyle \frac{{\mathit{Pop}}_t-{\mathit{Pop}}_{t-1}}{{\mathit{Pop}}_{t-1}}}\right)\times (1-{\eta }_{\mathit{effective}})$$

(7)

where: G_t is the estimated electricity demand for year t (TWh); G_t−1 demand observed in the previous year (TWh); GRGDP_t−1 GDP growth rate for each scenario and η_effective = 0.02 the energy efficiency factor.

3.4. Definition of Scenarios

Three distinct energy scenarios were defined to assess different development trajectories:

Baseline Scenario (BS) – continuation of current trends, with moderate expansion of renewables and a significant share of conventional thermal sources; Green Transition Scenario (GTS) – high penetration of renewables (solar, wind, and green hydrogen), aligned with decarbonization. Includes 2.8% hydrogen by 2050, contingent on the pace of investment and technology transfer; and Restricted Development Scenario (RDS) – a conservative scenario with limited expansion of installed capacity due to economic and institutional constraints.

3.5. Logistics Trajectory of Total Installed Capacity

The expansion of the Angolan power system’s total installed capacity follows a logistic saturation curve, limited by a structural ceiling, according to the model proposed by Nakicenovic et al. [57] and applied by George et al. [58] and Akaev et al. [59]. The parameterization by Harijan et al. [60] is adopted, as shown in Equation (8):

$$P_{\mathit{total},t}={\displaystyle \frac{K}{1+e^{[-r_{\mathit{log}}\times (t-t_0\left)\right]}}}$$

(8)

where: P_total,t is total installed capacity in year t (GW); K the maximum capacity of the scenario (44.40, or 38 GW); r_log = 0,15 the intrinsic logistic growth rate; t simulation year; t₀ = 2038 the turning point in the expansion of power.

3.6. Learning Dynamics and CAPEX Trends for Emerging Technologies

To estimate the reduction in capital costs for emerging technologies (solar photovoltaics, onshore wind, and green hydrogen) as a function of cumulative installed capacity, we adopted a framework that combines Wright’s Law [61], with the minimum cost threshold from the IMACLIM-R model [62], as shown in Equation (9):

$${\mathit{CAPEX}}_{\mathit{din},j,t}=\mathit{max}\left({\mathit{CAPEX}}_{\mathit{min},j},{\mathit{CAPEX}}_{\mathit{initial},j}\times {\left({\displaystyle \frac{P_{\mathit{acum},t}}{P_{\mathit{initial},j}}}\right)}^{-b_j}\right)$$

(9)

where: CAPEX_din,j,t the dynamic investment cost of technology j in year t; CAPEX_initial,j e P_{initial, j} define, respectively, the initial cost and the initial cumulative power; P_acum,t corresponds to the cumulative installed capacity; CAPEX_{min, j} asymptotic minimum cost limit; b_j it is the index of technological learning elasticity.

This framework combines Wright’s Law [61] with a minimum cost threshold inspired by the IMACLIM-R model [62], ensuring progressive reductions in CAPEX without converging to economically unrealistic values. The learning exponent (b_j) is determined endogenously from the learning rate ((LR_j), as shown in Equation (10) [63]:

$$b_j=-{\displaystyle \frac{\log (1-{\mathit{LR}}_j)}{\log \left(2\right)}}$$

(10)

3.7. Leveled Cost Structure for Electricity and Hydrogen

3.7.1. Levelized Cost of Electricity (LCOE) Excluding Fuel

The Levelized Cost of Electricity (LCOE) for renewable technologies without direct fuel consumption (solar photovoltaic, onshore wind, and biomass) is calculated using Equation (11). The equation is based on the general LCOE model proposed in Ref. [64] and has been adapted to represent only capital and operating costs, under the assumption of tax neutrality (unit tax factor, Tax Factor = 1):

$${\mathit{LCOE}}_{j,t}={\displaystyle \frac{{\mathit{CAPEX}}_{\mathit{din},j,t}\times {\mathit{CRF}}_{j,t}+{\mathit{OPEX}}_{j,t}}{E_{\mathit{annual},j}}}$$

(11)

where: CAPEX_j,t represents the cost of invested capital; CRF_j,t constitutes the Capital Recovery Factor used to annualize the investment; OPEX_j,t It is the cost of operation and maintenance e E_annual,j annual energy production generated by technology j. The Capital Recovery Factor it is calculated based on the financial discount rate r, which is equal to 0.08, and the asset's useful life n_j, according to Equation (12) [62,63]:

$${\mathit{CRF}}_j={\displaystyle \frac{{r\times (1+r)}^{n_j}}{{(1+r)}^{n_j}-1}}$$

(12)

Annual operating and maintenance costs are indexed to the invested capital using a fixed percentage (OPEX_%,j) calculated as a function of CAPEX, according to Equation (13) [63]:

OPEX_j,t = OPEX_% × CAPEX_din,j,t

(13)

3.7.2. Levelized Cost of Electricity (LCOE) by Fuel Type

For systems based on the combustion of fossil fuels (gas-fired CCGTs and pure or hybrid diesel generators), the Levelized Cost of Electricity (LCOE) explicitly includes the annualized fuel cost component. This formulation is based on the general LCOE model with indexed operating costs set forth in Ref. [67], according to Equation (14) [67]:

$${\mathit{LCOE}}_{j,t}={\displaystyle \frac{{\mathit{CAPEX}}_{\mathit{din},j,t}\times {\mathit{CRF}}_{j,t}+{\mathit{OPEX}}_{j,t}+C_{,j,t}}{E_{\mathit{annual},j,t}}}$$

(14)

where: CAPEX_din,j,t it is dynamic investment capital; CRF_j,t constitutes the Capital Recovery Factor used to annualize the investment; OPEX_j,t fixed and variable operating and maintenance costs (excluding fuel); C_j,t it is the total annualized cost of the fuel consumed e E_annual,j,t corresponds to annual electricity production. The estimate of the annual dynamic fuel cost incorporates changes in fuel prices, modeled using an annual escalation rate, as described by Short et al. [68]. Fuel energy conversion is modeled based on thermal efficiency and lower heating value, as proposed by Duffie and Beckman [69] and Kaabeche and Ibtiouen [70]. The annual fuel cost results from integrating these parameters with annual electricity production, as shown in Equation (15):

$$C_{j,t}=\left({\displaystyle \frac{P_{\mathit{base}}{(1+{\alpha }_t)}^{t-\mathit{base}}}{{\eta }_j\mathit{*PCI}}}\right)\times E_{\mathit{annual},j,t}\times D_{j,t}$$

(15)

where: C_j,t annual fuel cost (EUR/year); P_base fuel price in the base year; α_t annual price escalation rate; η_j thermal efficiency of the technology; PCI lower heating value and the D_j,t annual dispatch factor.

3.7.3. Levelized Cost of Green Hydrogen (LCOH)

The estimate of the levelized cost of hydrogen (LCOH_t), in €/kg, includes the annualized investment and operating costs, as well as the electrolysis conversion efficiency [71,72]. Capital costs are annualized using the Capital Recovery Factor (CRF), while the conversion of electricity to hydrogen is modeled by the ratio of hydrogen's lower heating value (${\mathit{LHV}}_{H_2}=33.33\mathit{kWh}/\mathit{kg}$) and the efficiency of the electrolyzer (η_eletr) [72,73]. Integrating these parameters with annual hydrogen production yields Equation (16):

$${\mathit{LCOH}}_t={\displaystyle \frac{{\mathit{CAPEX}}_{H_{2}V,t}\times {\mathit{CRF}}_{H_{2}V,t}+{\mathit{OPEX}}_{H_{2}V,t}}{E_{\mathit{annual},H_{2}V,t}}}\times \left({\displaystyle \frac{{\mathit{LHV}}_{H_2}}{{\eta }_{\mathit{eletr}}}}\right)$$

(16)

3.8. Carbon Footprint and CO₂ Emissions Inventory

Environmental accounting distinguishes between operational emissions and life-cycle emissions [74]. The former refer to the direct combustion of fossil fuels, while the latter encompass the entire life cycle of energy technologies [74]. Both inventories are calculated as the product of the electricity generated (E_annual,j) and the corresponding emission factors, as shown in Equations (17) and (18)) [74]:

$$E_{\mathit{op},t}={\displaystyle \sum _{j=1}^M}{E}_{\mathit{annual},j}\times {\mathit{EF}}_{\mathit{op},j}$$

(17)

$$E_{\mathit{LCA},t}={\displaystyle \sum _{j=1}^M}{E}_{\mathit{annual},j}\times {\mathit{EF}}_{\mathit{LCA},j}$$

(18)

where: E_op,t represents operational emissions e E_LCA,t life-cycle emissions; EF_op,j corresponds to the operational emission factor for technology j (gCO₂/kWh) and EF_LCA,j to the corresponding life-cycle emission factor j.

3.9. Sensitivity Analysis

A univariate sensitivity analysis was performed (Table 4), in which the main parameters were varied individually to identify the primary sources of uncertainty. A Monte Carlo analysis was not applied at this stage due to the lack of robust empirical distributions; it will be reserved for future work.

4. Results

4.1. Trends in Macroeconomic Demand and Demographic Transition (2024–2050)

The validation of the Angolan electricity sector’s trajectories was based on the baseline (BS), green transition (GTS), and restricted development (RDS) scenarios, revealing structural divergences after 2024 compared to the stability observed between 2010 and 2024, as illustrated in Figure 8. Electricity demand shows a sharp acceleration, reaching 283 TWh in 2050 under the green transition scenario (GTS), 194 TWh under the baseline scenario, and 101 TWh under the restricted development scenario (RDS), reflecting different dynamics of economic growth and electrification rates, as shown in Figure 7a). Per capita consumption ranges from less than 500 kWh/capita to 3,350 kWh/capita (GTS), 2,250 kWh/capita (BS), and 1,430 kWh/capita (RDS). At the same time, the total population is projected to range between 84 and 86 million in the GTS and BS scenarios, and 70 million in the RDS scenario, directly influencing the pressure on the global energy system, as illustrated in Figure 7b) and 7c).

Figure 7. Long-term projections of electricity demand and demographic trends in Angola (2010–2050) under the BS, GTS, and RDS scenarios: a) total electricity demand (TWh); b) per capita electricity consumption (kWh/capita); c) population trends (millions of inhabitants). Historical data (2010–2024) are indicated by discrete markers.

Figure 7. Long-term projections of electricity demand and demographic trends in Angola (2010–2050) under the BS, GTS, and RDS scenarios: a) total electricity demand (TWh); b) per capita electricity consumption (kWh/capita); c) population trends (millions of inhabitants). Historical data (2010–2024) are indicated by discrete markers.

4.2. Optimal Power Expansion and Evolution of the Technology Mix

The installed capacity mix reflects a profound structural transition between 2030 and 2050, with marked differences among the scenarios analyzed. In the BS scenario, total installed capacity expands from 10.2 GW to 37.8 GW; the share of hydropower declines from 56% to 43%, while solar power rises to 21% and onshore wind power reaches 9%.

Natural gas (CCGT) stabilizes at 10%, diesel drops to 4%, biomass remains at a negligible level, and hybrid systems settle at 4% in 2050 (Figure 8a)). In the GTS scenario, total installed capacity reaches 34.3 GW. Reliance on hydropower declines from 55% to 37%, offset by strong penetration of solar (33%) and wind (14%). Natural gas combined-cycle power plants (CCGT) and diesel generator sets are limited to 5% in aggregate, green hydrogen reaches a 5% share, biomass remains marginal, and hybrid technology stands at 3% (Figure 8b)).

Finally, in the RDS scenario, total installed capacity amounts to 32.6 GW. Large-scale hydropower plants remains dominant in the sector (though its share declines slightly from 58% to 49%), solar photovoltaic reaches 12%, and wind power stands at 5%. Natural gas-fired thermal power expands to 17% and diesel to 10%, with biomass accounting for a negligible share, virtually no development of hydrogen, and a 3% share for hybrid systems (Figure 8c)).

Figure 8. Optimal power expansion and evolution of the energy mix in Angola (2030–2050) under the following scenarios: a) BS; b) GTS; c) RDS, broken down by generation technology.

Figure 8. Optimal power expansion and evolution of the energy mix in Angola (2030–2050) under the following scenarios: a) BS; b) GTS; c) RDS, broken down by generation technology.

4.3. Electricity Generation Mix and Decarbonization Pathway on the Supply Side

Electricity generation reflects the capacity expansion guidelines, showing a significant divergence in trends across the scenarios, as illustrated in Figure 9. In the BS scenario, annual production grows from 30 TWh to 194 TWh. The contribution from hydropower decreases from 60% to 47%, while solar and onshore wind account for 11% and 9%, respectively. Natural gas (CCGT) and green hydrogen account for 12% and 6% of the mix, respectively, while diesel declines to 3%. In the GTS scenario, total generation amounts to 283 TWh. Hydropower reduces its share to 44%, while solar photovoltaic accounts for 19% and wind power for 14%. Green hydrogen accounts for 9% of total generation and biomass for 5%, limiting natural gas to 7% and diesel to 2%. In the RDS scenario, total production is capped at 101 TWh. Hydropower maintains a clear economic and operational dominance (61%–51%), solar reaches 6%, and onshore wind stands at 5%. Natural gas expands its share to 20% and diesel to 5%, reflecting a delay in the transition and greater structural dependence on fossil fuels.

4.4. Economic Performance and Levelized Cost of Electricity (LCOE)

4.4.1. Assessment of the Levelized Cost of Electricity (LCOE) by Technology (2050)

The LCOE analysis for the 2050 horizon confirms the superiority and economic competitiveness of renewable technologies compared to conventional thermal alternatives, in line with the data in Figure 10. In the three simulated scenarios, large-scale solar photovoltaics (19.2–25.5 EUR/MWh) and onshore wind (28.5–30.5 EUR/MWh) have the lowest marginal generation costs. In contrast, pure diesel systems are the least efficient and most expensive option (202.3–238.0 EUR/MWh), Natural gas combined-cycle plants operate in an intermediate range between 130.5 and 142.3 EUR/MWh, while large-scale hydropower continues to face high structural costs (179.0–193.4 EUR/MWh), due to the significant CAPEX associated with large civil infrastructure projects and the length of the national transmission lines.

Diesel-solar hybrid systems demonstrate a robust capacity for cost-effective mitigation, reducing decentralized generation costs to levels between 86.3 and 93.9 EUR/MWh. In addition, green hydrogen has competitive production costs, ranging from 2.4 to 3.0 EUR/kg H₂, establishing itself as a viable vector for energy storage and industrial decarbonization in the Angolan system.

4.4.2. Comparative Analysis of LCOE: Pure Diesel Systems vs. Hybrid Systems

Figure 11 details the temporal evolution of the LCOE for the decentralized Pure Diesel and Hybrid (Diesel–Solar) systems over the period from 2030 to 2050. The Pure Diesel system shows a continuous increase in linear costs, rising from 150 EUR/MWh (2030) to 208 EUR/MWh (2050) in the BS scenario, and peaking at 238 EUR/MWh in the GTS scenario, driven by the volatility and rising prices of fossil fuels, In contrast, the hybrid configuration introduces tariff stability, keeping the LCOE between 70 and less than 95 EUR/MWh. This translates into economic efficiency and cost savings of over 60% compared to the standalone fossil fuel mix in 2050, demonstrating the mitigation of the risk associated with the international fuel market.

4.5. Carbon Intensity Assessment: Operational Emissions vs. Life-Cycle Emissions

Environmental accounting based on carbon intensities reveals severe disparities between thermal power and renewable energy systems when direct operational emissions are compared with the integrated life cycle assessment (LCA) footprint, Figure 12. Diesel-fired generation alone results in the system having the most burdensome indicators, ranging from 0.78 tCO₂e/MWh (operational, EF_op) and 0.90 tCO₂e/MWh (LCA, EF_LCA), Natural gas CCGTs come in right behind, with emissions ranging from 0.49 to 0.64 tCO₂e/MWh. The introduction of solar hybridization mitigates the environmental impact of diesel, reducing emissions to 0.43 tCO₂e/MWh (operational) and 0.45 tCO₂e/MWh (LCA). 100% renewable technologies eliminate operational emissions, leaving only residual life-cycle footprints associated with manufacturing and construction: solar photovoltaic (0.04 tCO₂e/MWh), hydropower (0.01 tCO₂e/MWh), wind power (0.01 tCO₂e/MWh), and biomass (0.23 tCO₂e/MWh). These indicators demonstrate the climate benefits of hybridization and the central role of renewables in meeting Angola’s decarbonization goals.

4.6. Multivariate Sensitivity Analysis and Systematic Uncertainty Analysis

The sensitivity and uncertainty analysis demonstrates the model’s stability and robustness in the face of extreme fluctuations in the input parameters, as shown in the diagrams in Figure 13. In the total electricity demand vector, yield elasticity (epsilon) is assumed to be the dominant variable, generating deviations in the range of -60% to -31%, followed by variations in GDP (-59% to +19%), while fluctuations in CAPEX are negligible or residual. With regard to the system’s present value, the discount rate (r) is the parameter with the highest sensitivity and financial risk, causing variations ranging from -37.5% to +50.0% and outweighing the influence of elasticity and GDP, while CAPEX has only marginal effects limited to < plus or minus 4.1%. To aggregate CO₂ emissions, uncertainty is driven almost exclusively by the magnitude of final demand, with an additional impact of only approximately 8% associated with the specific emission factor for natural gas. The robustness tests confirm that the trajectories of the Angolan electricity system respond with greater volatility to macroeconomic and demographic factors than to technological parameters.

5. Discussion

5.1. Bridging the Gap Between Official Energy Planning and Electric System Optimization in Angola

The results point to a profound structural transformation of Angola’s electricity system by 2050, which is only partially consistent with the guidelines of the Angola 2025 Long-Term Strategy [81], the Angola Energy Plan 2025 [82], the Angola 2050 Long-Term Strategy [83], and the 2040 Electricity Sector Master Plan [84]. There is a critical mismatch between current official planning and the simulated long-term dynamics.

Electricity demand reaches 283 TWh in the GTS scenario, far exceeding the projections of the Ministry of Energy and Water (MEW), which estimate approximately 65 TWh and 11.2 GW of peak load by 2040 [84]. The Master Plan itself acknowledges significant methodological limitations in its demand data and acknowledges the existence of a weak historical correlation between GDP and electricity consumption [84], which reinforces the need to adopt forward-looking and structural approaches such as the one developed in this model. The results also indicate high demographic pressure (84–86 million inhabitants) and a macroeconomic trajectory aligned with the Angola 2050 Long-Term Strategy [83] and the targets of SDG 7 of the UN 2030 Agenda. The energy mix is evolving from a historically hydro-centric system to a diversified matrix, characterized by strong growth in solar photovoltaics (33%), onshore wind (14%), and green hydrogen (9%) in the GTS scenario. This configuration significantly exceeds the modest targets of the 2040 Master Plan, which limits expansion to levels less than or equal to 100 MW for solar and 652 MW for wind, while completely omitting the role of hydrogen [79,81].

Despite the conceptual alignment with the sustainability goals of the Angola Energy 2025 Plan [82], the current pace of implementation is insufficient to meet long-term security of supply needs. From an economic standpoint, renewable sources offer an unequivocal advantage in terms of LCOE (19 - 30 EUR/MWh) when compared to the costs associated with diesel-powered generators (150 - 238 EUR/MWh) or large-scale hydropower, which is burdened by high CAPEX for grid infrastructure; this reinforces the economic rationale for energy decentralization. In off-grid systems, diesel-solar hybridization reduces generation costs by more than 50%, providing technical validation for PEGC and MEW policies on rural electrification. From an environmental perspective, the transition scenarios significantly reduce the sector’s carbon intensity, in close alignment with Angola’s Nationally Determined Contributions (NDCs) [86]. Conversely, the RDS scenario increases the risk of climate non-compliance due to its persistent reliance on fossil fuels

Finally, the sensitivity analysis confirms the predominance of macroeconomic variables over technological constraints, indicating that the success of Angola’s energy transition depends fundamentally on financial and regulatory stability. It is concluded that only the GTS scenario ensures simultaneous coherence among energy security, socioeconomic development, and deep decarbonization by 2050.

5.2. Analysis of Policy Gaps: Model vs. National Energy Plan

The comparative analysis presented in Table 5 reveals a systematic bias toward underestimation in official planning when compared with the GTS (2050) scenario. There is a discrepancy of more than four times in total demand and a shortfall of more than 23 GW in peak installed capacity, which is inextricably linked to an underestimation of official demographic projections (35 million according to official figures versus the projected 84–86 million). On the technological front, there is a significant underestimation of solar and wind power, as well as a structural lack of pathways for green hydrogen.

This transition involves shifting from a hydro dominated energy mix (> 60%) to a balanced and resilient structure (37 - 51%). From an economic perspective, power generation based strictly on diesel incurs costs up to 10 times higher than those of solar photovoltaics (150 - 238 EUR/MWh versus 19 - 26 EUR/MWh), exhibiting high price volatility.

Decentralized hybrid systems, on the other hand, enable reductions of more than 50% in both operating costs and greenhouse gas emissions. Overall, the results show that Angola’s current energy planning simultaneously underestimates long-term demand, capacity expansion needs, and the depth of the necessary technological transition, thereby compromising the robustness of strategic decisions and alignment with international decarbonization commitments.

6. Conclusions

This paper presented a prospective and structural analysis of Angola’s energy system for the time horizon from 2024 to 2050, evaluating different transition pathways based on dynamic multi-scenario modeling. The results reveal a profound structural shift from the historical pattern observed during the 2010–2024 period, driven by strong demographic pressure—with the population projected to reach between 84 and 86 million—and by a sharp increase in electricity demand. This demand could reach 283 TWh in the Green Transition (GTS) scenario, in stark contrast to the 194 TWh estimated in the Reference Scenario (BS) and the 101 TWh in the Restricted Development Scenario (RDS). Per capita consumption rises significantly to 3,350 kWh/capita in the GTS scenario, underscoring the urgent need for an accelerated expansion of national installed capacity.

The electricity system is gradually evolving toward a more balanced and diversified generation mix. Although large-scale hydropower remains the backbone and structural foundation of the supply (accounting for a share of 37% to 51%, depending on the scenario), there is strong penetration of unconventional renewable sources. Large-scale solar photovoltaics account for up to 33% of total installed capacity, while onshore wind and green hydrogen account for up to 14% and 9% of total generation, respectively, drastically reducing the country’s structural dependence on fossil fuels.

From an economic perspective, renewable technologies demonstrate a clear competitive advantage in terms of lower-cost planning, with an optimized LCOE ranging from 19.2 to 30.5 EUR/MWh. These indicators contrast with the high generation costs associated with the expansion of regulated hydropower infrastructure (179.0-193.4 EUR/MWh) and, most importantly, with the volatile costs of pure diesel generation (271.4-307.1 EUR/MWh), confirming the direct economic benefit of decarbonizing the national grid. In the field of decentralized and rural electrification, diesel-solar hybrid systems are emerging as a highly mature solution, reducing generation costs to levels ranging from 86.3 to 121.5 EUR/MWh, representing savings of more than 50%. At the same time, they reduce the subsector’s carbon intensity from 0.78 - 0.90 tCO₂e/MWh to just 0.43 tCO₂e/MWh, validating the decentralized regulatory strategies coordinated by PEGC and MEW.

Finally, the sensitivity and systemic uncertainty analysis strongly indicates that macroeconomic variables—namely revenue elasticity, GDP growth, and the discount rate — exert greater statistical dominance than the technological parameters of CAPEX and OPEX, demonstrating that the viability of Angola’s energy transition is intrinsically dependent on factors of institutional, exchange rate, and macrofinancial stability. In summary, it is concluded that decentralized solar hybridization and the large-scale expansion of interconnected renewable energy parks constitute technically viable and economically imperative solutions to safeguard security of supply, ensure climate sustainability, and enable sustained cost reductions in Angola’s electricity system through 2050.