Nigeria’s energy crisis, with a 24 GW demand gap against only 4–6 GW of grid supply, has accelerated interest in Decentralized Renewable Energy Systems (DRES). This study develops an integrated Socio-Technical System Dynamics model, calibrated with national survey data, to simulate household DRES adoption and emissions outcomes from 2024 to 2030. A Monte Carlo (MC) framework quantifies uncertainty, while Cost-Benefit Analysis (CBA) assesses economic viability. Within the modelled scenarios, Business-as-Usual (BAU) achieves only 4% DRES penetration and 11.2% emissions reduction by 2030, whereas Policy-Driven Transition (PDT) reaches 17% penetration and 19.0% reduction. The Behaviourally-Responsive Transition (BRT), driven by financial mechanisms and awareness interventions, achieves 10% penetration and 16.3% reduction. Global Sensitivity Analysis identifies financing as the dominant adoption driver (subsidy SRC = 0.62; affordability SRC = 0.60), while awareness growth governs emissions reduction (SRC = −0.98); grid improvement shows negligible influence (SRC = 0.001). The CBA confirms strong household-level returns, with BRT achieving an NPV of ₦6.76M (90% CI: ₦5.59M ₦7.96M) and a BCR of 20.28 (90% CI: 17.01–23.68). PDT delivers the highest returns (NPV ₦6.25M, BCR 20.84), while BAU achieves ₦5.07M and 13.65. Even under the most conservative scenario, BRT maintains NPV of ₦1.58M and a BCR of 5.77, confirming the economic viability of DRES adoption across all scenarios. These findings suggest that Nigeria’s clean energy transition is likely constrained more by affordability and behavioural barriers than grid capacity, highlighting the need for policies centred on inclusive finance, targeted subsidies, and integrated clean cooking interventions.
Combined heat and power (CHP)-centered microgrids are sensitive to electric and heating load forecast uncertainty in day-ahead scheduling, yet such uncertainty is often represented using fixed-percentage margins that do not reflect observed forecast-error behavior. This paper proposes a data-driven robust unit commitment and economic dispatch (UC/ED) framework that links forecast-performance-based uncertainty characterization with coupled electric–thermal scheduling. Seasonal uncertainty bounds are derived from actual-to-forecast ratios using empirical directional quantiles, providing separate downward and upward deviations for electric and heating demands. These asymmetric bounds are incorporated into a robust mixed-integer optimization model with independent electric and heating uncertainty budgets and are evaluated under both upward and downward deviations. The framework co-optimizes CHP units, battery energy storage, photovoltaic generation, and grid transactions over a 24-h scheduling horizon. A four-season case study using residential building data demonstrates substantial variation in forecast uncertainty, ranging from narrow autumn bounds of [−3.76%,+5.39%] for electric demand and [−4.55%,+4.40%] for heating demand to spring bounds of [−29.91%,+40.66%] and [−45.28%,+43.23%], respectively. Comparison with a conventional fixed ±15% benchmark shows that a uniform margin can overstate uncertainty in low-error seasons while understating its operational impact when empirical deviations are larger. The resulting operating-cost responses also vary with uncertainty direction and electric–thermal budget allocation, demonstrating the importance of preserving seasonal, directional, and load-specific forecast-error characteristics in robust microgrid scheduling.
This study compares the conditional financial viability of a standardized 5 MW/10 MWh battery energy storage system (BESS) across Japan, Great Britain, PJM, Germany, and Vietnam. Wholesale-arbitrage and ancillary-service opportunities are reconstructed from observed market prices in four markets, while Vietnam is assessed using explicitly assumed revenues. Japan’s reference ancillary data cover April–December 2024 without annualization, and 288 missing five-minute PJM Regulation observations are excluded. A common unlevered, pre-tax 15-year model assumes USD 1.92 million CAPEX, USD 51,900 annual fixed operations and maintenance, 1% annual revenue degradation, and 10% real discount rate. First-year revenue comprises the larger standalone stream plus 50% of the smaller, with a 50% ancillary award rate assumed for the four observed-price markets and no capacity-market payments.Under these assumptions, Japan, PJM, and Germany yield positive NPVs of USD 2.625, 1.841, and 1.228 million, with IRRs of 31.38%, 25.41%, and 20.57%, respectively. Great Britain yields first-year revenue of USD 192,884, an IRR of 0.01%, and an NPV of −USD 0.922 million. Vietnam’s USD 90,000 revenue scenario yields an IRR of −14.37% and an NPV of −USD 1.665 million; its required first-year revenue for NPV = 0 is USD 320,640 at reference CAPEX. A matched fiscal-year reconstruction increases Japan’s IRR to 42.35%. Reducing PJM’s assumed ancillary award rate from 50% to 25% lowers its NPV to −USD 0.073 million, reversing the positive reference result. These findings describe modelled revenue opportunities, not realized project returns or market-wide bankability.Electrical capacity and energy are expressed using SI-based power units and conventional electricity-energy units. Market settlement prices retain the units of their respective source datasets, including JPY/ΔkW per 30-minute interval, GBP/MW/h, and EUR/MW per four-hour product. Financial results are reported in constant 2024 USD unless otherwise stated.
With increasing concerns over climate change, the utilization of renewable energy has gained significant attention. In Japan, the construction of woody biomass power plants continues to expand, highlighting the need for economically efficient facility location models. However, existing studies remain limited, as most rely on aggregated data and partial cost components. This study proposes an integrated optimization framework combining stand-level biomass estimation with a full supply chain cost assessment. Biomass availability is estimated at the forest sub-compartment level and aggregated to obtain total supply, enabling a high-resolution representation. For each stand, the costs of biomass collection, chipping, and transportation are evaluated based on terrain conditions. This study contributes by (i) establishing a stand-level biomass estimation framework that quantifies available biomass and provides a detailed procurement cost for each forest stand; (ii) using terrain- and biomass-condition-dependent cost models for individual supply chain components, including collection, chipping, and transportation; and (iii) integrating Particle Swarm Optimization (PSO), road networks, terrain, and stand-level GIS data into an automated optimization system that achieves high-precision identification of the optimal plant location. Due to the computational complexity of evaluating more than 65,000 forest stands and continuous candidate locations, the PSO algorithm is used to efficiently identify the optimal plant location based on total supply and land acquisition costs. A case study in the Fuji region of Shizuoka Prefecture shows that procurement costs can be reduced to 8,496 JPY/t, demonstrating improved economic feasibility.
Agrivoltaic (AV) system design requires a balance between photovoltaic (PV) electricity generation and the light that reaches crops at ground level. This study presents a multi-objective layout screening framework that jointly evaluates PV irradiance capture, the PV energy yield ratio, ground level light reduction, a crop specific light yield ratio based on photosynthetically active radiation (PAR), and the light productivity factor (LPF). Five structural configurations are compared under a representative Malaysian climate: fixed north–south, fixed east–west, standard tracking, reverse tracking and customised tracking. Each is assessed with monofacial and bifacial modules at approximately half and full array density for Andrographis paniculata, lettuce, turnip and corn, which have contrasting light requirements. Ground level irradiance is simulated through geometric shadow mapping, and four optimisation cases progressively introduce the PAR yield ratio as an objective and the PV energy yield ratio as a constraint. In the multi-objective case, the annual irradiance on the modules was 1853.9–1935.1 kWh m−2, compared with 1819 kWh m−2 for the benchmark layout. The energy output was 67.95–70.61 kWh m−2 for monofacial and 77.91–83.53 kWh m−2 for bifacial modules, 24–53% above the benchmark of 54.69 kWh m−2, while the ground level light reduction was 35.20–42.65%. The LPF ranged from 0.91 to 2.33, and the highest value was obtained for Andrographis paniculata under full density bifacial standard tracking. The framework supports early-stage screening of food and energy trade-offs and identifies candidate layouts for field validation. It does not predict harvested crop yield.
