Strategic Planning for Power System Decarbonization Using Mixed-Integer Linear Programming and the William Newman Model

Abstract

This paper proposes a comprehensive framework for strategic power system decarbonization planning that integrates the William Newman method (diagnosis–options–forecast–decision) with a multi-objective Mixed-Integer Linear Programming (MILP) model. The approach simultaneously minimizes (i) generation cost, (ii) expected cost of energy not supplied (Value of Lost Load, VoLL), (iii) demand response cost, and (iv) CO₂ emissions, subject to power balance, technical limits, and binary unit commitment decisions. The methodology is validated on the IEEE RTS 24-bus system with increasing demand profiles and representative cost and emission parameters by technology. Three transition pathways are analyzed: baseline scenario (no environmental restrictions), gradual transition (−50% target in 20 years), and accelerated transition (−75% target in 10 years). In the baseline case, the oil- and coal-dominated mix concentrates emissions (≈14 ktCO₂ and ≈12 ktCO₂, respectively). Under gradual transition, progressive substitution with wind and hydro reduces emissions by 15.38%, falling short of the target, showing that renewable expansion alone is insufficient without storage and demand-side management. In the accelerated transition, the model achieves −75% by year 10 while maintaining supply, with a cost–emissions trade-off highly sensitive to the carbon price. Results demonstrate that decarbonization is technically feasible and economically manageable when three enablers are combined: higher renewable penetration, storage capacity, and policy instruments that both accelerate fossil phase-out and valorize demand-side flexibility. The proposed framework is replicable and valuable for outlining realistic, verifiable transition pathways in power system planning.

1. Introduction

The global energy landscape faces significant challenges due to heavy reliance on fossil fuels, which account for 73% of greenhouse gas emissions. In this context, decarbonizing the power sector is essential to meeting the targets set by the Paris Agreement. This agreement seeks to limit global temperature rise to 1.5 °C and reduce emissions by 45% by 2030 [1,2].

Achieving these goals requires an energy transition centered on renewable sources, which supplied 29% of global electricity in 2020. With proper implementation, renewables could reach 50% by 2035, driven by advances in solar, wind, and energy storage technologies [3,4].

Currently, power systems generate 40% of global CO₂ emissions, underscoring the urgency of sustainable strategies. At the same time, climate change impacts result in annual economic losses estimated at USD 250 billion, according to the World Bank [5,6].

In response, Mixed-Integer Linear Programming (MILP) enables the integration of renewable energy and storage into power systems, reducing investment costs by 23% and energy losses by 44% [7,8]. Furthermore, decentralized systems such as microgrids can lower local emissions by up to 80%, as reported by the International Energy Agency [9,10].

Clean energy investment reached USD 1.1 trillion in 2022, surpassing fossil fuels for the first time, according to BloombergNEF [11,12]. In this sense, policies and technologies based on MILP are projected to reduce operating costs by 10–15% by 2030 and enhance grid reliability against extreme climate events. These strategies position power systems as key drivers of global sustainable development [13].

The primary objective of strategic decarbonization planning is to eliminate carbon emissions from electricity generation. To this end, the process aims to maximize renewable integration and reduce emissions associated with the energy system. Advanced models such as MILP support this process by incorporating renewable resources, energy storage, and control technologies, enabling efficient and sustainable system design and operation. These models also address the inherent variability of renewables, ensuring reliable and cost-effective supply [14,15].

Hybrid systems that combine renewable generation with hydrogen production and storage are also essential for decarbonization. These technologies enhance efficiency, reduce operational costs, and improve grid stability [16,17]. Advanced bi-level collaborative models optimize capacity planning and load distribution in wind–solar–hydrogen systems [18]. Likewise, hydrogen systems and batteries maximize the use of intermittent renewable resources. As a result, such strategies can reduce CO₂ emissions by 50% while increasing renewable penetration [18,19].

Integrated planning must consider both generation and transmission, accounting for physical system constraints and variations in supply and demand. This approach reduces computation time by 82% and keeps costs within 1.5% of the optimal solution [20,21].

Robust optimization based on MILP mitigates uncertainties in generation and demand, ensuring reliable and economically viable operations. Similarly, advanced methods enhance system flexibility and reduce risks associated with outages or fluctuations [22,23].

These strategies are thus fundamental to achieving global emission reduction targets. Transitioning to resilient and sustainable energy systems ensures the capacity to meet future demand for clean and efficient electricity.

The current energy transition is gaining momentum due to rising environmental concerns, with the power sector and other industries as major contributors to pollution. Reducing CO₂ emissions from the power sector has become a top priority for many countries, with favorable results driven by the feasibility of renewable energy systems that enable large-scale decarbonization of electricity [24,25]. The central aim is to reduce dependence on fossil-based fuels in system operations.

In parallel with renewable deployment, the power sector has undergone major transformations linked to technology and digitalization [26]. The development and implementation of renewable generation systems foster more flexible and efficient energy management across nations, either as business ventures or production initiatives [27].

A long-term vision is needed to enable a broader portfolio of renewable technologies, including solar, wind, and hydro. Clean energy generation through photovoltaic systems is increasingly common, driven by rapid PV development that offers innovative and cost-competitive alternatives. For the power sector, the growing use of solar PV and utility-scale generation systems is proving highly productive [28,29].

In Ecuador, energy sector reports published by regulatory entities describe gross annual generation of 32,206.9 GWh, a demand of 4.21 GW, and an installed capacity of 8794.4 MW [30]. Historically, part of the sector has relied on thermal plants, burning fossil fuels such as coal and natural gas to produce electricity. These operations have persisted for decades, making them a significant source of greenhouse gas emissions and highlighting the need for long-term projects aimed at their progressive replacement with clean energy sources [30].

Strategic planning thus becomes a fundamental process for both power utilities and enterprises. Based on past, present, and future conditions, it supports effective responses to unforeseen events and anticipates emerging challenges, ensuring proactive and efficient management [31,32]. The shift toward renewable technologies calls for a new strategic planning model in the power sector [33].

For calculations, the strategic model is combined with Mixed-Integer Linear Programming, supported by mathematical optimizers to achieve final outcomes such as minimizing CO₂ emissions in connected systems. Additionally, optimization seeks to maximize factors such as power consumption growth, generation adequacy, and energy efficiency within a long-term planning horizon [34,35,36].

Finally, building on William Newman’s framework, this work integrates strategic planning principles with a Mixed-Integer Linear Programming (MILP) formulation to address power system decarbonization. The originality of this research lies in bridging a qualitative strategic management model with a quantitative optimization approach, enabling the evaluation of realistic decarbonization pathways under growing demand conditions. Unlike most previous studies, which focus separately on unit commitment, capacity expansion, or emission reduction policies, the proposed approach combines four key objectives: minimizing generation costs, minimizing the expected cost of unserved energy (VoLL), reducing demand response costs, and constraining CO₂ emissions. This integration provides a structured and replicable decision-making framework for utilities and policymakers.

The main contributions of this paper are threefold. First, it establishes a novel methodological link between strategic management (the William Newman model) and numerical optimization (MILP), offering a dual strategic–operational perspective. Second, it applies this integrated framework to a representative case study adapted to Ecuador’s power system context, which still relies significantly on fossil-based generation despite its renewable potential. Third, it delivers a scenario-based evaluation (baseline, gradual transition, and accelerated transition) complemented by carbon pricing sensitivity analysis, quantifying both the economic and environmental trade-offs of decarbonization pathways.

The remainder of this paper is organized as follows. Section 2 describes the methodological framework, detailing how the William Newman model is embedded into the MILP formulation. Section 3 introduces the case study, including generation, cost, and emission data. Section 4 presents the results under different decarbonization scenarios, while Section 5 concludes with key findings and outlines avenues for future research.

Related Work and Systematic Literature Review

MILP for power system planning and operation. Mixed-Integer Linear Programming (MILP) has become a standard tool for generation dispatch, unit commitment, and capacity expansion under technical constraints (ramping, minimum up/down times, reserve margins) and policy targets. Prior studies apply MILP to co-optimize short-term scheduling and long-term investment while capturing transmission limits and security criteria [7,8,20,21,22,23,33]. These models typically minimize fuel and operating costs subject to nodal balance and generator limits.

Renewables, storage, and hybrid systems. A large body of work analyzes wind/solar integration and the role of utility-scale and distributed storage to mitigate variability and curtailment. Recent contributions include hybrid architectures (PV–wind–H₂ production and storage) and battery–hydrogen co-optimization that enhance system flexibility and reduce emissions [16,17,18,19]. Results consistently show that storage and flexible demand are necessary complements to renewable expansion.

Demand response and flexibility markets. Demand response (DR) has been modeled as price-based or incentive-based programs (e.g., TOU, critical peak, direct load control), enabling load shifting and peak shaving in both transmission and distribution contexts [37,38,39,40,41]. MILP formulations commonly represent DR via sectoral elasticities and interruption costs (VoLL-based), ensuring explicit trade-offs between system cost and consumer disutility.

Carbon policies and decarbonization pathways. Carbon pricing, technology-specific emission factors, and retirement schedules have been incorporated to evaluate pathways consistent with Paris-aligned targets. Evidence shows that renewable additions alone are insufficient; carbon price signals, storage build-out, and DR materially accelerate decarbonization while maintaining adequacy [1,2,3,4,5,6]. Sensitivity to carbon cost is often decisive for optimal portfolios.

Gap addressed by this work. Compared with the above literature, this paper contributes a strategic planning framework that explicitly couples the William Newman management model (diagnosis–options–forecast–decision) with a multi-objective MILP. Our formulation jointly minimizes the following: (i) generation cost (EGSC), (ii) expected cost of energy not supplied (EECC via VoLL), (iii) DR program cost (ELRC), and (iv) carbon emissions cost (ECEC), under power balance, unit commitment binaries, renewable and storage requirements, and DR bounds. Unlike prior studies focused solely on expansion or single-objective dispatch, we evaluate transition trajectories (baseline, gradual –50%/20 years, accelerated –75%/10 years) on the IEEE 24-bus system with realistic demand growth, technology-specific costs, and emission factors [14,15,18,19,33,42]. The framework is transparent and replicable, providing decision support that links managerial strategy with numerical optimization.

2. Theoretical Framework

2.1. Analysis of the Current Power System

The first step requires collecting, compiling, and analyzing information on the present state of the power system.

Ecuador’s power sector is highly diverse. By 2021, its nominal installed capacity reached 8734.40 MW, while the effective capacity was 8100.70 MW. The generation mix is classified into renewable and non-renewable sources, with hydropower dominating nominal capacity, followed by biomass, photovoltaic, and other plants. Hydropower accounts for nearly 70% of total national electricity production, with notable facilities such as the Coca Codo Sinclair project.

Nevertheless, Ecuador also relies on non-renewable sources. Thermal power plants, which operate using fossil fuels, hold the second largest share of the mix. These facilities play a critical role in electricity supply, particularly in the insular region and parts of the Amazon. In addition, new trends have emerged, such as hybrid plants, including an expansion project currently operating in the Galápagos. Overall, the generation system has been designed to meet the needs of both large and small consumers. Despite challenges, Ecuador’s power sector has experienced significant growth in the past decade, achieving a high electrification rate that extends to rural areas [43].

Types of Power Plants

A power plant is defined as a facility specialized in electricity generation, serving either national or regional demand, and typically connected to the grid. Plants can be categorized as conventional or non-conventional depending on their primary energy sources. This includes thermal, hydropower, solar, and wind facilities [44,45].

Power plants must meet the energy requirements of various sectors, including industrial, residential, and others. They are also expected to ensure high reliability, efficiency, and sustainability of supply through diversified and well-maintained generation assets [46].

Table 1. Effective and nominal capacity of power plants in Ecuador.

Plants	Energy Sources	Nominal Capacity [MW]	Effective Capacity [MW]
Hydropower	Renewable	5106.85	5072.26
Photovoltaic	Renewable	27.70	26.80
Wind	Renewable	21.20	21.20
Biogas	Renewable	8.40	7.25
Biomass	Renewable	144.40	136.45
Thermal	Non-renewable	3426.15	2836.90

shows both nominal and effective capacities. Hydropower plants, as renewable sources, present the highest values, including both reservoir-based and run-of-river facilities. Non-renewable thermal plants include internal combustion units, which are the most widely used, followed by steam and gas turbine plants.

2.2. Description of Thermal Power Plants

Ecuador’s energy sector includes a variety of operational plants. Hydropower produces approximately 6877 GWh annually, while thermal plants contribute around 346,355 GWh. Thermal plants operate using gas, internal combustion engines, gas turbines, and steam technologies [47].

2.3. Technical Characteristics of Thermal Plants

Each thermal plant has distinct operating characteristics, depending on the type of fuel injected during production. These features are evaluated annually to assess global energy performance. Some are described in the following subsections [48].

2.3.1. Internal Combustion Engine (ICE) Plants

Thermal power plants with internal combustion engines (ICE) use diesel, bunker, or natural gas and are typically deployed for peak support due to their fast response and modularity [49,50]. In this study, only aggregated techno-economic parameters are employed in the MILP optimization model. The detailed datasheet of the Guangopolo II power plant (units, ratings, efficiencies, and nameplate data) has been moved to Appendix B,

Table A7. Guangopolo II thermal power plant (ICE).

GUANGOPOLO II THERMAL POWER PLANT
SUPERVISOR:	CELEC-EP/TERMOPICHINCHA
TECHNOLOGY TYPE:	Internal Combustion Engine (ICE)
GENERAL DATA
Country	Ecuador
Province	Pichincha
City	Quito
TECHNICAL DATA
Installed Capacity:	52.38 MW
Type:	ICE
Number of Units:	7 Units
Capacity per Unit:	8.73 MW
Fuel Type:	Diesel–Bunker
Plant Factor:	3.82%
Average Energy:	582.58 GWh/year
Engines
Type:	MITSUBISHI-MAN
Rating:	5,200 kW
Power Factor:	0.80
Type:	WÄRTSILÄ DIESEL
Model:	8SW28
Rating:	1,980 kW
Frequency:	60 Hz

.

2.3.2. Gas Turbine Plants

Gas turbines (GT) offer good part-load efficiency and reduced start-up times, with natural gas as the prevalent fuel and lower emission factors than ICE [51,52]. To avoid overloading the main text, the technical datasheet of the Machala II plant (GE TM2500 units, capacity, and operational parameters) is included in Appendix B,

Table A5. Machala II thermal power plant (Gas Turbine).

MACHALA II THERMAL POWER PLANT
SUPERVISOR:	CELEC-EP/TERMOMACHALA
TECHNOLOGY TYPE:	Gas Turbine
GENERAL DATA
Country	Ecuador
Province	El Oro
City	Machala
TECHNICAL DATA
Installed Capacity:	252 MW
Type:	Gas Turbine (GT)
Number of Units:	8 Units
Capacity per Unit:	6 of 20 MW and 2 of 66 MW
Fuel Type:	Natural Gas
Plant Factor:	37.50%
Average Energy:	406.70 GWh/year
Gas Turbine
Type:	General Electric
Model:	TM2500
Frequency:	60 Hz

.

2.3.3. Steam Turbine Plants

Steam turbine (ST) plants convert thermal energy into electricity using boilers operated with fossil fuels [53]. In the MILP formulation, only aggregated techno-economic parameters are considered. The detailed datasheet of the Trinitaria thermal plant is provided in Appendix B,

Table A6. Trinitaria thermal power plant (Steam Turbine).

TRINITARIA THERMAL POWER PLANT
SUPERVISOR:	CELEC-EP/ELECTROGUAYAS
TECHNOLOGY TYPE:	Steam Turbine Thermal Plant
GENERAL DATA
Country	Ecuador
Province	Guayas
City	Guayaquil
TECHNICAL DATA
Installed Capacity:	133 MW
Type:	Steam Turbine (ST)
Number of Units:	1 Unit
Capacity per Unit:	133 MW
Fuel Type:	Fuel Oil #4
Efficiency:	16%
Plant Factor:	54.10%
Average Energy:	629.50 GWh/year
Turbine
Type:	DKY2-INDRI
Manufacturer:	ASEA BROWN BOVERI
Rated Power:	133 MW
Frequency:	60 Hz
Temperature:	583 °C
Pressure:	140
Phases:	3
Poles:	2
Speed (RPM):	3600

.

2.4. Identification of Opportunities and Challenges

Based on the analysis of the current state, opportunities for decarbonization within thermal power plants can be identified strategically. Their feasibility can thus be evaluated as follows:

Advantages:

• According to CELEC EP, thermal plants require relatively less maintenance.
• Thermal plants play a crucial role by replacing hydropower during climate events such as droughts.
• Their relatively simple deployment makes thermal plants widely used worldwide, offering lower costs compared to other generation technologies.
• Thermal generation is not weather-dependent, allowing adaptation to any environment.
• Thermal efficiency is key, as it enables higher energy conversion with minimal losses.

Disadvantages:

• The use of fossil fuels causes significant environmental damage, and fuel prices may fluctuate, directly affecting electricity production.
• Thermal plants have a considerable environmental impact, producing high levels of CO₂ emissions regardless of plant type.
• High emissions not only harm the environment but also contribute to respiratory illnesses, affecting both operators and local populations.
• Maintenance—whether corrective or scheduled—can require long periods, sometimes months, to replace major components.

2.5. Principles and Foundations of the William Newman Model for Power System Decarbonization

The William Newman model provides a structured methodology for strategic planning that can be adapted to the decarbonization of power systems, particularly those relying on conventional thermal generation plants. Its strength lies in offering a clear sequence of phases that support decision-making, ensuring that complex transitions are approached systematically and transparently.

Figure 1 illustrates the William Newman model and its adaptation to the power sector. The diagnosis phase establishes the baseline conditions, the options phase defines renewable integration and demand response strategies, the forecast phase evaluates their long-term impact using MILP simulations, and the decision phase identifies the optimal decarbonization pathway. This integration highlights the usefulness of the Newman model as both a conceptual framework and a practical decision-support tool for planning sustainable power systems [42,54].

Application of the William Newman Model to Power System Decarbonization

The Newman framework is used here as a practical scaffold—rather than as a generic description—to organize the transition analysis on the IEEE 24-bus system. Each phase maps to concrete modeling tasks:

Diagnosis. We define a data-driven baseline with the original generator portfolio and transmission limits, hourly demand from

Table A3. Load profile.

Hour	System Demand [MW]	Hour	System Demand [MW]
1	1775.835	13	2517.975
2	1669.815	14	2517.975
3	1590.3	15	2464.965
4	1563.795	16	2464.965
5	1563.795	17	2623.995
6	1590.3	18	2650.5
7	1961.37	19	2650.5
8	2279.43	20	2544.48
9	2517.975	21	2411.995
10	2544.48	22	2199.915
11	2544.48	23	1934.865
12	2517.975	24	1669.815

, costs in Table 5 and Table 6, and emission factors in Table 7.

Options. Two policy pathways are constructed: Gradual (meta de $-50\%$ en 20 años) y Acelerada (meta de $-75\%$ en 10 años). Las opciones incluyen límites de emisiones (25), cuotas mínimas renovables (26), y disponibilidad de almacenamiento/demanda (27) and (28).

Forecast. Cada opción se evalúa con un MILP multiobjetivo (costos de generación, costo esperado de energía no servida, costo de respuesta de la demanda y costo del carbono), sujeto a balance (19), límites de generación (20) y activación binaria (29).

Decision. The comparison is carried out using Pareto fronts and aggregated metrics (total cost and tCO₂), reported in the Figure 11 and Figure 15, and with the direct comparison of emissions between scenarios (Figure 9).

2.6. Implementation of Mixed-Integer Linear Programming for Power System Decarbonization: Considerations and Constraints

Mixed-Integer Linear Programming (MILP) is an effective tool for optimizing complex systems such as power networks, especially in the context of decarbonization. However, applying MILP in these cases requires several considerations and constraints to be addressed [33].

Considerations

Data and Information Considerations:

The quality of MILP results depends heavily on the accuracy of the input data. Comprehensive and precise datasets are required on existing generation capacity, production costs, and the emission factors of different technologies.

Technical Factors Considerations:

Several technical factors must be taken into account when implementing MILP, including generation capacity limitations of different technologies, the availability of resources (for example, solar and wind depend on weather conditions), and constraints in the transmission network.

Regulatory and Policy Considerations:

Government policies and regulations, along with decarbonization targets and electricity tariffs, must also be incorporated into the model. These considerations depend on the geographical region in which the power system is deployed.

2.7. Distributed Resources

The integration of non-conventional renewable generators into the Power System (PS) poses significant technical challenges, since their operation depends directly on variable meteorological conditions. This inherent uncertainty prevents the guarantee of a continuous supply, potentially leading to operational imbalances or even partial grid collapses. Nevertheless, despite these limitations, this type of generation represents a sustainable alternative to conventional methods, which, based on fossil fuels, perpetuate pollutant emissions and the depletion of non-renewable resources [55,56].

2.7.1. Photovoltaic Generation

Photovoltaic (PV) systems convert plane-of-array irradiance into electric power through semiconductor modules arranged in series–parallel strings. The following relations describe cell temperature, instantaneous power, period energy, and annualized cost; explicit symbol definitions are given immediately below each equation to ensure clarity [57,58,59].

$$T_{cell}=T_{amb}+{\displaystyle \frac{(NOCT-20)}{800}}G$$

(1)

where $T_{cell}$ = cell operating temperature [°C]; $T_{amb}$ = ambient temperature [°C]; G = plane-of-array solar irradiance [W/m²]; $NOCT$ = Nominal Operating Cell Temperature. The factor $(NOCT-20)/800$ converts irradiance into a temperature rise above ambient per the manufacturer’s rating.

$$P_{ph}=P_{stc}{\displaystyle \frac{G}{1000}}\left[1+\alpha (T_{cell}-25)\right]$$

(2)

where $P_{ph}$ = PV output power at operating conditions [W]; $P_{stc}$ = module/string power at STC (1000 W/m², 25 °C) [W]; $\alpha$ = power temperature coefficient [1/°C]; $T_{cell}$ from (1) [°C]; G [W/m²]. The term $G/1000$ scales output with irradiance relative to STC.

$$E_t=3.24M_{pv}\left[1-0.0041(T_t-8)\right]S_t$$

(3)

where $E_t$ = energy produced in period t [kWh]; $M_{pv}$ = installed PV capacity (sum of nameplate ratings) [kW]; $T_t$ = average ambient temperature in period t [°C]; $S_t$ = site-specific insolation factor for period t [kWh/kW]. Constant $3.24$ reflects the adopted daily resource normalization; the temperature modifier $0.0041$ represents the empirical derating per °C deviation from 8 °C.

$$F\left(P_s\right)=\left(a{I}^p{P}_s\right)+\left(G^E{P}_s\right)$$

(4)

where $F\left(P_s\right)$ = annualized PV cost attributed in the planning model [$]; $P_s$ = PV capacity (or equivalently dispatched power proxy) [MW_Megawatt]; $I^p$ = specific investment cost [$/MW]; $G^E$ = fixed O&M per unit capacity [$/MW year]; a = capital recovery factor in (5) [1/year].

$$a={\displaystyle \frac{r}{1-{(1+r)}^{-N}}}$$

(5)

where r = real discount rate [–]; N = project lifetime [years]. This converts upfront investment into an equivalent uniform annual cost.

2.7.2. Wind Power Generation

This technology is based on the conversion of wind kinetic energy into electricity using adaptive wind turbines. These systems dynamically adjust their operation to compensate for forecast discrepancies and maximize efficiency.

The performance of wind turbines depends on key climatic parameters such as average wind speed, diurnal fluctuations, seasonal variations, and the geographical characteristics of the site. These factors determine the optimal technical sizing of the turbine.

Energy conversion is quantified through Equation (6) [60], which models generated power as a function of atmospheric variables and turbine physical properties.

$$P={\displaystyle \frac{1}{2}}\left(\rho ·A·u^3\right)$$

(6)

In this expression, $\rho$ represents air density, P denotes the generated power, u is wind speed, and A is the swept area of the turbine rotor.

The economic valuation of wind power is described by Equations (7) and (8) [59].

$$F\left(P_e\right)=\left(a·I^P·P_e\right)+\left(G^E·P_e\right)$$

(7)

$$a={\displaystyle \frac{r}{[1-{(1+r)}^{-N}]}}$$

(8)

Here, $P_e$ is the wind power output expressed in MW, a is the annuity factor, $I^p$ represents the investment cost per MW, and $G^E$ covers operation and maintenance expenses. N is the estimated operational lifespan in years, and r is the interest rate used in the project’s financial assessment.

2.7.3. Diesel-Based Power Supply

Diesel generators, widely used in large-scale systems, provide electricity in isolated locations and support critical loads. Their high reliability and rapid load response make them a common choice in these contexts [61,62].

However, their dependence on fossil fuels limits efficiency and causes a significant environmental impact. Additionally, operating costs are high and subject to fluctuations in fuel prices [63].

The economic cost of diesel generation is determined using Equation (9), as established in [59].

$$C_i\left(P_{Gdi}\right)=\sum _{i=1}^{NG}[a_i+b_i·P_{Gdi}+C_i·P_{Gdi}^2]$$

(9)

In this expression, $a_i$ is the startup cost of each generator, $b_i$ represents the fuel cost per unit of generated power, $P_{Gdi}$ is the output power of each diesel unit, and $NG$ is the total number of generators in the system.

2.7.4. Battery-Based Backup Systems

Batteries store energy in direct current and are used to maintain supply during faults or interruptions in the main grid.

Despite their advantages, they also present challenges such as high cost, periodic maintenance requirements, and limited lifespan. Cell temperature must be controlled and kept below 25 °C to preserve performance, as this directly affects both efficiency and durability.

Battery performance depends on several factors, including materials, construction quality, operating temperature, and charge/discharge conditions [64]. The economic valuation considers technical parameters such as expected lifetime, operating temperature range, and current profiles. For lead–acid battery banks, the cost is calculated using Equation (10) [59].

$$C_b=a·I_{IB}+C_{O\&M}+C_R+{\displaystyle \frac{C_c}{n_b}}$$

(10)

Here, a is the annuity factor, $I_{IB}$ is the initial installation cost, and $C_{O\&M}$ accounts for operation and maintenance expenses. Similarly, $C_R$ corresponds to replacement costs during the battery lifetime, $C_c$ represents charging-related costs, and $n_b$ denotes the efficiency of the battery bank.

2.7.5. Charge and Discharge Index

Battery storage systems are commonly represented using two technical formulations. The first is based on tracking the State of Charge (SOC) to estimate available energy. The second employs an Electromotive Force (EMF) equation that dynamically characterizes charging and discharging processes as a function of internal system voltage.

In Hybrid Renewable Energy Systems (HRES), surplus energy produced by primary sources is transferred to storage. When renewable generation is insufficient, the stored energy provides backup. Battery efficiency improves under these conditions, optimizing overall energy management.

The sizing of the battery bank is determined by technical parameters such as maximum Depth of Discharge (DOD), thermal dissipation during operation, and system lifetime. In HRES, the storage bank plays a critical role in stabilizing the energy balance between available generation and instantaneous demand.

Because charging and discharging alternate, input power may take positive or negative values. The State of Charge is determined by productivity and consumption time according to Equation (11).

$$P_{pv}^T+P_{WIND}^T+P_{BIO}^T=P_{DEMAND}^T$$

(11)

In this equation, $P_{pv}^T$ is the energy generated by photovoltaic installations, $P_{WIND}^T$ is the energy from wind turbines, and $P_{BIO}^T$ is the power supplied by biomass plants. $P_{DEMAND}^T$ is the total system demand.

To maintain stable battery capacity and avoid storage fluctuations, condition (12) applies.

$$P_{pv}^T+P_{WIND}^T+P_{BIO}^T>P_{DEMAND}^T$$

(12)

When hybrid generation exceeds demand, the battery charges. The stored energy at time T is determined by Equation (13).

$$E_{battery}^T=E_{battery}^{T-1}·(1-\tau )+{\displaystyle \frac{[(P_{pv}^T+P_{WIND}^T+P_{BIO}^T)-\frac{P_1^T}{n_{inverter}}]·n_{bc}}{1}}$$

(13)

Here, $E_{battery}^T$ is the stored energy at time T, and $E_{battery}^{T-1}$ is the available charge from the previous period. $P_1^T$ is the demand at a specific hour, while $P_{pv}^T$, $P_{WIND}^T$, and $P_{BIO}^T$ are contributions from photovoltaic, wind, and biomass systems. $n_{bc}$ represents charging efficiency, $n_{inverter}$ is inverter efficiency, and $\tau$ is the self-discharge rate per hour.

Conversely, when renewable generation is insufficient to cover demand, condition (14) applies.

$$P_{pv}^T+P_{WIND}^T+P_{BIO}^T+P_{dis}^T=P_{DEMAND}^T+P_{ch}^T$$

(14)

In this case, the battery discharges, limited by its nominal capacity. The discharge model is expressed in Equation (15).

$$E_{battery}^T=E_{battery}^{T-1}·(1-\tau )+{\displaystyle \frac{[\frac{P_1^T}{n_{inverter}}-(P_{pv}^T+P_{WIND}^T+P_{BIO}^T)]·n_{bf}}{1}}$$

(15)

Here, $n_{bf}$ denotes battery discharge efficiency.

2.7.6. DOD

The charge extracted from a battery during the period $Q_d$ is expressed as a percentage of the total available charge. This relationship is defined by Equation (16).

$$DoD={\displaystyle \frac{Q_d}{C}}·100\%$$

(16)

In this formulation, $Q_d$ is the discharged energy during the analysis period, and C is the total usable capacity of the storage system, according to the limit defined in [65].

2.7.7. DR

Demand response (DR) is based on voluntary adjustments made by end-users to their electricity consumption patterns in response to market signals. The mechanism aims to reduce peak loads, minimize energy-related costs, and increase system reliability, while enabling consumers to actively participate in the dynamic balance of energy demand, as outlined in [37,38].

In residential and commercial contexts, different devices can be integrated into DR schemes, such as HVAC systems for climate control, programmable appliances such as dryers, dishwashers, and freezers, and energy storage systems such as electric vehicle batteries. Heat pumps, refrigeration units, and certain industrial processes, such as roller presses, can also be incorporated into load-control strategies [39].

• DR Programs

DR programs can be understood as structured mechanisms operating under incentive-based schemes, divided into two main modalities. The first corresponds to traditional DR, which includes direct load control and interruptible tariffs. The second relies on market-based dynamics, including programs designed for energy contingencies, reserve and capacity markets, consumption-dependent auctions, and ancillary service platforms [39].

Implicit DR, or price-based DR programs, involve voluntary schemes in which consumers face variable electricity tariffs depending on time. An example is time-differentiated pricing, such as day–night rates [40]. These programs are structured according to the costs of electricity generation during different time periods and adapt to usage conditions and regional limitations. In northern Europe, for example, consumers actively participate in flexible pricing models such as Time of Use (TOU) tariffs, critical peak pricing, and real-time price adjustments [39].

The operational environment of DR involves multiple stakeholders across the power sector. These include generators, infrastructure operators, commercial networks, specialized intermediaries, and policy-making agencies. Transmission System Operators (TSO) and Distribution System Operators (DSO) manage the transmission and distribution networks, respectively. Balance Responsible Parties (BRP) are responsible for energy balancing, while end-users participate as residents or property owners. New actors, such as aggregators, provide advanced services including the aggregation and management of retail customer consumption. This diversifies and strengthens the structure of the energy market. The relationships among these actors are presented in

Table 2. Main actors in the demand response market.

Actors	Offers	Users
BRP	Energy loss payments; Market access; DR incentives	Consumer
Aggregator	Ancillary services; Tariffs; Grid balancing services	TSO; DSO
Supplier/Retailer	Incentive packages and contracts for implicit DR programs; DR incentives	Consumers
Regulator	DR regulations; Knowledge for DR management	All actors
Consumer	Demand profile; Direct control; Large consumers can provide flexibility directly	Aggregator; Supplier/Retailer; DR market

.

Controlled power under DR strategies must remain within the permitted values, as shown in Equation (17).

$$0\le P_{DR}\left(t\right)\le P{\left(t\right)}_{max}^{DR}$$

(17)

In this equation, $P_{DR}\left(t\right)$ represents the power managed through demand-side mechanisms, and $P{\left(t\right)}_{max}^{DR}$ indicates the maximum permissible value. This upper limit is determined as a fraction of the system’s total capacity, considering both conventional and non-conventional generation technologies [41].

3. Problem Formulation

The increasing energy demand and the urgency to reduce environmental impacts have driven decarbonization strategies in power systems. Globally, thermal power plants remain essential for ensuring grid stability. However, their high carbon emissions demand an accelerated transition toward renewable energy sources.

One of the main challenges of this transition is maintaining system reliability when incorporating intermittent sources such as wind and solar power. The inherent variability of these technologies can compromise electricity supply and increase operating costs if appropriate management strategies are not implemented. Therefore, it is essential to develop tools that balance renewable integration with system stability.

To address this issue, the present study proposes an optimization model based on Mixed-Integer Linear Programming (MILP) and the William Newman approach. This methodology enables strategic planning of decarbonization by minimizing both costs and emissions without compromising grid stability. The following section presents the development of the William Newman method.

3.1. Decarbonization Process Using the William Newman Approach

The transition to a sustainable power system requires clear and well-structured strategies that reduce carbon emissions without compromising supply reliability. To achieve this, an approach adopted is based on the William Newman method, which provides an organized framework for strategic planning. This method integrates decarbonization strategies into specific phases, ensuring coherent and well-grounded decision-making.

The process is divided into four main phases: problem diagnosis, strategy definition, implementation and execution, and evaluation and control. Each phase addresses critical aspects of the energy transition, from identifying challenges to measuring results. The strategies proposed in each phase focus on resource optimization and environmental impact mitigation.

In the diagnostic phase, the main issues of the current power system are analyzed. Subsequently, in the strategy definition stage, specific actions are established to overcome these challenges. Implementation and execution ensure effective application of the measures, while evaluation and control allow for strategy adjustments based on observed results.

3.2. Demand Variability

For each scenario, a specific demand profile is used, considering that both demand and generation grow over time.

Figure 2 presents hourly electricity demand over a 24 h horizon. Between 6:00 and 9:00, demand rises sharply from 1000 MW to more than 5000 MW. From 10:00 to 17:00, demand remains high, ranging between 5000 and 6000 MW. From 18:00 onward, demand progressively decreases until reaching approximately 2000 MW by midnight. The shaded area between the 10th and 90th percentiles highlights significant hourly variability, especially during peak hours.

Figure 3 shows the demand evolution over a 10-year horizon. The average curve reveals steady growth from approximately 4400 MW to nearly 5400 MW by the end of the period. The shaded area between the P10 and P90 percentiles remains relatively stable, reflecting consistent variability over time. By year 10, the P90 percentile reaches about 6100 MW, while the P10 percentile remains above 5200 MW.

Figure 4 extends the time horizon to 20 years, showing an increase in average demand from approximately 4000 MW in the first year to almost 6000 MW by year 20. During this period, the range between the P10 and P90 percentiles slightly expands, reflecting growth both in average demand and in its dispersion. By the end of the analysis, the P90 percentile reaches nearly 6500 MW, while the P10 percentile stands close to 5500 MW.

3.2.1. Phase 1: Problem Diagnosis

The first stage of the William Newman method focuses on diagnosing the power system. This detailed analysis precedes the application of decarbonization strategies. At this stage, a baseline scenario is defined without additional restrictions on the energy mix, which allows for the assessment of conventional generation performance in its current state.

The main objective is to quantify operating costs, evaluate supply reliability, and analyze the environmental impact in terms of carbon emissions. To achieve this, mathematical models are formulated, including objective functions and constraints, as presented below.

• Cost of Unserved Energy

In the baseline scenario, the absence of investment in emerging technologies and demand management strategies can affect supply security. To address this issue, an objective function is formulated to minimize the costs associated with unserved energy, as expressed in Equation (18).

$$\begin{array}{cc}\hfill min{f}_1\left(t\right)& =min\left[EEC{C}_{L_b}\left(t\right)\right]\hfill \\ \hfill & =min\left[\sum _{L_b=1}^L{P}_{L_b}^{loss}·VoLL\right]\hfill \end{array}$$

(18)

In this expression, $EEC{C}_{L_b}\left(t\right)$ represents the expected cost associated with unserved energy at load node $L_b$ ($). $P_{L_b}^{loss}$ is the unserved energy at node $L_b$ (MWh). $VoLL$ defines the value of lost load, representing the economic cost of unsupplied electricity ($/MWh).

• Constraints in the Baseline Scenario

In a power system, it is essential to ensure that available energy meets demand without interruptions. Therefore, constraints are established to guarantee stability and reliability, verifying both sufficiency of generation and production capacity limits.

• Power Balance

Power balance is essential for proper system operation, as it ensures that generated energy matches demand at every instant. If this balance is not maintained, outages or overloads may occur, compromising system stability. This principle is expressed in Equation (19).

$$\sum _{g=1}^{G}P{G}_g\left(t\right)=\sum _{L_b=1}^{L}P{D}_{L_b}\left(t\right),\forall t$$

(19)

Here, $P{G}_g\left(t\right)$ is the power generated by unit g at time t (MW), G is the set of all generating units, $P{D}_{L_b}\left(t\right)$ is the demand at load node $L_b$ at time t (MW), L is the set of all load nodes, and t represents the time interval of system analysis.

• Generation Limits

Each power plant operates within a specific generation range, determined by technical characteristics and economic constraints. To ensure safe and efficient operation, the minimum and maximum generation limits are defined as shown in Equation (20).

$$P{G}_g^{min}\le P{G}_g\left(t\right)\le P{G}_g^{max},\forall g\in G,\forall t$$

(20)

Here, $P{G}_g\left(t\right)$ represents the power generated by unit g at time t (MW), $P{G}_g^{min}$ is the minimum power unit g can generate (MW), and $P{G}_g^{max}$ is the maximum power unit g can generate (MW). G is the set of all generating units in the system, and t represents the analysis time interval.

• Load Shedding

When a power system lacks new generation plants or advanced demand-side management mechanisms, periods of insufficient energy may occur. This situation causes load shedding, leading to temporary supply interruptions in certain areas. To avoid these issues, it is essential to maintain a balance between supply and demand.

To control this phenomenon, Equation (21) is applied, ensuring that unserved energy does not exceed total demand.

$$0\le P_{L_b}^{loss}\left(t\right)\le P{D}_{L_b}\left(t\right),\forall L_b\in L,\forall t$$

(21)

In this expression, $P_{L_b}^{loss}\left(t\right)$ represents the unserved energy at load node $L_b$ in period t (MWh). $P{D}_{L_b}\left(t\right)$ is the total demand at load node $L_b$ in period t (MWh). L is the set of load nodes in the system, and t represents the system analysis period (years or time intervals).

3.2.2. Phase 2: Strategy Definition

In the second phase of the model, strategies are designed to modify energy use and reduce dependence on traditional sources. The objective is to decrease carbon emissions while ensuring that electricity remains reliable and affordable. To this end, rules are established to regulate generation technologies and manage associated costs.

Two main transition plans toward cleaner energy are proposed, as follows:

Gradual Transition: This approach seeks to progressively reduce emissions over a 20-year period. It allows for the incremental integration of renewable energy sources, ensuring a balance between sustainability and system stability.

Accelerated Transition: In this case, emissions are reduced over a shorter period of 10 years. To achieve this, renewable energy deployment is intensified, and energy storage systems are implemented.

• Carbon Emission Minimization

The proposed model aims to minimize carbon emissions in the power system through a mathematical function that evaluates the environmental impact of each generation technology. Equation (22) establishes the relationship between carbon emissions and energy generated from fossil fuels.

$$\begin{array}{cc}\hfill min{f}_4\left(t\right)& =min\left[ECE{C}_{G_b}\left(t\right)\right]\hfill \\ \hfill & =min\sum _{b=1}^G\left(\begin{array}{c}EP{G}_{oil}^{G_b}\left(t\right)·(C{E}_{oil}^{ct}+C{E}_{oil}^{st})\hfill \\ +EP{G}_{cl}^{G_b}\left(t\right)·C{E}_{cl}\hfill \end{array}\right)\times EC\hfill \end{array}$$

(22)

Here, $EP{G}_{oil}^{G_b}\left(t\right)$ represents the energy generated by oil-fired units at node $G_b$ during period t (MWh). $C{E}_{oil}^{ct}$ and $C{E}_{oil}^{st}$ are the CO₂ emission factors for combustion and steam turbines operating with oil (tons/MWh). $EP{G}_{cl}^{G_b}\left(t\right)$ denotes the energy generated by coal-fired units at node $G_b$ during period t (MWh). $C{E}_{cl}$ is the CO₂ emission factor for coal units (tons/MWh). $EC$ reflects the cost associated with carbon emissions ($/ton).

• Minimization of Total Generation Cost

Ensuring the efficiency and sustainability of the power system requires optimizing generation costs. This process involves reducing both fixed and variable costs, depending on the technology used to produce energy, as expressed in Equation (23).

$$\begin{array}{cc}\hfill min{f}_2\left(t\right)& =min\left[EGS{C}_{G_b}\left(t\right)\right]=\hfill \\ \hfill & \sum _{b=1}^G\left(\begin{array}{c}{C}_{oil}^{G_b}(1-{\alpha }_{oil}^{G_b})Fi{x}_{oil}+EP{G}_{oil}^{G_b}Va{r}_{oil}\hfill \\ C_{cl}^{G_b}(1-{\alpha }_{cl}^{G_b})Fi{x}_{cl}+EP{G}_{cl}^{G_b}Va{r}_{cl}\hfill \\ C_{ncl}^{G_b}(1-{\alpha }_{ncl}^{G_b})Fi{x}_{ncl}+EP{G}_{ncl}^{G_b}Va{r}_{ncl}\hfill \\ C_{hy}^{G_b}(1-{\alpha }_{hy}^{G_b})Fi{x}_{hy}+EP{G}_{hy}^{G_b}Va{r}_{hy}\hfill \\ C_{wnd}^{G_b}Fi{x}_{wnd}+EP{G}_{wnd}^{G_b}Va{r}_{wnd}\hfill \end{array}\right)\hfill \end{array}$$

(23)

In this expression, $C_i^{G_b}$ represents the installed capacity of technology i at node $G_b$, measured in MW. ${\alpha }_i^{G_b}$ corresponds to the fraction of conventional generation replaced by renewable sources. $Fi{x}_i$ defines the fixed operation and maintenance cost of technology i, expressed in $/MW. $EP{G}_i^{G_b}$ denotes the energy generated by technology i at node $G_b$, measured in MWh. $Va{r}_i$ reflects the variable cost per unit of energy generated by technology i, expressed in $/MWh.

• Minimization of Demand Response Cost

Guaranteeing system stability requires strategies that reduce dependence on conventional generation. Efficient demand-side management becomes a key mechanism to lower operating costs and optimize energy use, as expressed in Equation (24).

$$\begin{array}{cc}\hfill min{f}_3\left(t\right)& =min\left[ELR{C}_{L_b}\left(t\right)\right]\hfill \\ \hfill & =min\sum _{L_b=1}^L\left(\begin{array}{c}ED{R}_{res}^{L_b}\left({\beta }_{res}^{L_b}\right)\times I{C}_{res}\hfill \\ ED{R}_{ind}^{L_b}\left({\beta }_{ind}^{L_b}\right)\times I{C}_{ind}\hfill \\ ED{R}_{com}^{L_b}\left({\beta }_{com}^{L_b}\right)\times I{C}_{com}\hfill \\ ED{R}_{lrg}^{L_b}\left({\beta }_{lrg}^{L_b}\right)\times I{C}_{lrg}\hfill \\ ED{R}_{agr}^{L_b}\left({\beta }_{agr}^{L_b}\right)\times I{C}_{agr}\hfill \\ ED{R}_{gov}^{L_b}\left({\beta }_{gov}^{L_b}\right)\times I{C}_{gov}\hfill \\ ED{R}_{off}^{L_b}\left({\beta }_{off}^{L_b}\right)\times I{C}_{off}\hfill \end{array}\right)\hfill \end{array}$$

(24)

In this equation, $ED{R}_i^{L_b}$ is the energy shifted at node $L_b$ by sector i (MWh). ${\beta }_i^{L_b}$ indicates the demand response elasticity factor for sector i at node $L_b$. $I{C}_i$ represents the interruption cost of sector i ($/MWh).

• Constraints of the Second Phase

The proposed model integrates a set of constraints designed to ensure an efficient and controlled energy transition. These conditions regulate the gradual reduction of emissions, establish a minimum renewable generation threshold, and optimize both energy storage and demand management, as presented below.

• Emission Reduction Constraint

To meet energy transition objectives, a limit is set on total system emissions for each period, as defined in Equation (25). This mechanism enables progressive and controlled reduction of greenhouse gases, ensuring a balance between energy generation and environmental sustainability.

$$\begin{array}{cc}\hfill min{f}_4\left(t\right)& =min\left[ECE{C}_{G_b}\left(t\right)\right]\hfill \\ \hfill & =min\sum _{b=1}^G\left(\begin{array}{c}EP{G}_{oil}^{G_b}\left(t\right)·(C{E}_{oil}^{ct}+C{E}_{oil}^{st})\hfill \\ +EP{G}_{cl}^{G_b}\left(t\right)·C{E}_{cl}\hfill \end{array}\right)\times EC\hfill \\ \hfill & \le x\times E_{base}\hfill \end{array}$$

(25)

Here, x is the emission reduction factor, which takes values between 0.25 and 0.5 in phase 2. $E_{base}$ represents the total emissions in the baseline scenario (tons of CO₂).

• Renewable Energy Integration Constraint

To promote decarbonization of the power system, a minimum renewable generation threshold is established, as defined in Equation (26). This measure ensures significant participation of clean technologies in the energy mix, driving the transition toward a more sustainable and environmentally friendly model.

$$C_{wind}^{G_b}\left(t\right)+C_{sol}^{G_b}\left(t\right)\ge y\times C_{total}\left(t\right),\forall G_b,\forall t$$

(26)

In this expression, $C_{wind}^{G_b}\left(t\right)$ denotes the installed wind generation capacity at node $G_b$ during period t (MW). $C_{sol}^{G_b}\left(t\right)$ represents the installed solar generation capacity at node $G_b$ during period t (MW). $C_{total}\left(t\right)$ indicates the total installed capacity in the power system (MW). y expresses the minimum percentage of renewable generation required in the energy transition.

• Energy Storage Capacity Constraint

The variable nature of renewable energy sources requires reliable backup through storage systems, modeled by Equation (27). These systems are essential to compensate for fluctuations in solar and wind generation, ensuring a stable and continuous electricity supply.

$$S_{bat}^{G_b}\ge 0.2\times P{D}_{L_b}\left(t\right),\forall G_b,\forall t$$

(27)

Here, $S_{bat}^{G_b}$ corresponds to the storage capacity used at node $G_b$ during time t (MWh). $P{D}_{L_b}\left(t\right)$ represents the total demand at load node $L_b$ during period t (MW).

• Demand-Side Management Constraint

Demand response plays a key role in reducing reliance on conventional generation and preventing grid overloads. Through this strategy, expressed in Equation (28), energy resources are used more efficiently, contributing to system stability.

$$\sum _{i\in S}ED{R}_{L_b}^i\left(t\right)\le \delta \times P_b^{DL}\left(t\right),\forall L_b,\forall t$$

(28)

In this expression, S indicates the set of sectors participating in demand response, including residential, industrial, and commercial sectors.

• Binary Constraint for Generator Unit Activation

The model incorporates binary variables to represent discrete decisions, such as the activation of generating units, as expressed in Equation (29). These constraints ensure that a unit only produces power when it is operating, which is a fundamental requirement to formulate the problem as a Mixed-Integer Linear Programming model.

$$P_{G_b,t}^{min}·u_{G_b,t}\le P_{G_b,t}\le P_{G_b,t}^{max}·u_{G_b,t}$$

(29)

Here, $P_{G_b,t}$ represents the generated power, $u_{G_b,t}\in \{0,1\}$ indicates the operational status of the unit, and $P_{G_b,t}^{min}$ and $P_{G_b,t}^{max}$ are the technical generation limits.

Finally, Algorithm 1 presents the decarbonization process described above. Additionally,

Table 3. Variables used in Algorithm 1.

Symbol	Description	Unit
N	Population size	-
$\alpha =[{\alpha }_1,\dots ,{\alpha }_G]$	Wind penetration coefficients for each generator	-
$\beta =[{\beta }_1,\dots ,{\beta }_L]$	Percentage of load to be shifted in each load sector	-
$V_w$	Wind speed	m/s
$V_{ci},V_r,V_{co}$	Cut-in, rated, and cut-out wind speeds of the turbine	m/s
$P_r$	Rated power of the wind turbine	MW
$A,B,C$	Coefficients of the wind power function	-
$L\left(t\right)$	Load demand at time t	MW
$P_k$	Power shifted by demand response	MW
$\Omega$	Set of simulation periods	-
$P_{W_b}$	Power generated by the wind farm	MW
$P_{los{s}_l}$	Unserved energy losses	MW
$VoLL$	Value of Lost Load	$/MW
$ED{R}_{L_b}^{res}\left({\beta }_b^{res}\right)$	Energy reduced by demand response	MWh
$I{C}_{res}$	Incentive cost for load reduction	$
$C_{G_b}^{oil}$	Oil-fired generation cost	$/MWh
$Fi{x}_{oil}$	Fixed cost of oil-fired generation	$
$EP{G}_{G_b}^{oil}$	Energy generated with oil	MWh
$C{E}_{oil}^{ct},C{E}_{oil}^{st}$	Oil generation emission factors	ton CO2/MWh
$P_{G_g}$	Power generated by unit g	MW
$P_{D{L}_l}$	Power demanded in sector l	MW
$P_{Gmin},P_{Gmax}$	Minimum and maximum generation limits	MW

describes the variables involved.

Algorithm 1 Pseudocode of the optimization model
Power System Decarbonization Process
Step 1	Initialization Define main parameters: - Population size N - Maximum number of iterations or convergence criterion - Coefficient of variation of EECC < 5% - Parameters of genetic operators Initialize random solution population: $\alpha =[{\alpha }_1,{\alpha }_2,\dots ,{\alpha }_G]$ $\beta =[{\beta }_1,{\beta }_2,\dots ,{\beta }_L]$ Define binary variables: $x_{G_b}^{wind}\in \{0,1\}$ (installation of wind farm at node $G_b$) $u_{G_b,t}\in \{0,1\}$ (ON/OFF status of generator $G_b$ at t) $y_{L_b}^{DR,t}\in \{0,1\}$ (activation of DR at node $L_b$)
Step 2	Evaluation of each solution in the population For each individual x in the population: Wind simulation and wind power generation: Use ARMA model for $V_w$ Calculate wind power output: $P_{W_b}=\left\{\begin{array}{cc}0,\hfill & 0\le V_w<V_{ci}\hfill \\ (A+B{V}_w+C{V}_w^2)P_r,\hfill & V_{ci}\le V_w<V_r\hfill \\ P_r,\hfill & V_r\le V_w<V_{co}\hfill \\ 0,\hfill & V_w\ge V_{co}\hfill \end{array}\right.$ Load modeling and demand response: Modify load curve: $L\left(t\right)=L\left(t\right)+{\displaystyle \frac{{\sum }_{t\in \Omega }(L\left(t\right)-P_k)}{N}}$
Step 3	Calculation of optimization objectives Cost of Unserved Energy (EECC): $f_1\left(x\right)={\sum }_t{P}_{loss,l}\times VoLL$ Generation Cost (EGSC): $f_2\left(x\right)={\sum }_G(C_{G_b}^{oil}\times (1-{\alpha }_{G_b}^{oil})\times Fi{x}_{oil}+\dots )$ Demand Response Cost (ELRC): $f_3\left(x\right)={\sum }_l(ED{R}_{L_b}^{res}\left({\beta }_b^{res}\right)\times I{C}_{res}+\dots )$ Carbon Emissions (ECEC): $f_4\left(x\right)={\sum }_G(EP{G}_{G_b}^{oil}\times (C{E}_{oil}^{ct}+C{E}_{oil}^{st}))$
Step 4	Verification of constraints Generation–demand balance: ${\sum }_g{P}_{G_g}={\sum }_l(P_{D{L}_l}+P_{loss,L_b}\left(t\right))$ Technical loss limit: $0\le P_{loss,L_b}\left(t\right)\le \gamma P_{D{L}_b}\left(t\right)$ Generation limits: $P_{Gmin}\le P_{G_g}\le P_{Gmax}$ Load shedding: $0\le P_{loss,l}\le P_{D{L}_l}$ Generator activation: $EP{G}_{G_b}\left(t\right)\le P{G}_{G_b}^{max}\times x_{G_b}$ Demand response activation: $ED{R}_s^{L_b}\left(t\right)\le ED{R}_s^{max}\times y_{L_b,t}^{DR}$
Step 5	Solution selection Identify non-dominated solutions in the population
Step 6	Fuzzy selection of the best solution For each individual x in the solution pool: For each objective $i=1,\dots ,4$: Normalize objective values Select the solution with the smallest error as optimal
Step 7	Genetic Operators Apply selection, crossover, and mutation on the solution set
Step 8	Convergence criterion If verification coefficient < 5% or iterations ≥ 100,000: Terminate algorithm Else: Update population and repeat from Step 2
Step 9	Final result Return the solution set and the best solution
Step 10	End

Methodological Novelty and Contribution

Unlike prior applications that treat strategic planning and numerical optimization separately, our work tightly couples William Newman’s management cycle (diagnosis–options–forecast–decision) with a multi-objective MILP that operates over explicit transition trajectories. Concretely:

• Diagnosis is encoded as a calibrated baseline on the IEEE 24-bus system (demand growth, technology costs, and emission factors) that fixes the reference emission level $E_{base}$ and the unconstrained cost benchmark.
• Options are translated into policy/technology levers that enter the MILP as constraints and variables: emission caps (25) via factor x, minimum renewable share (26) via y, storage adequacy (27), and sectoral demand response bounds (28). This mapping makes qualitative strategy selectable and testable in a single optimization run.
• Forecast is implemented as a multi-period formulation that evaluates long-horizon pathways (baseline, 20-year gradual, 10-year accelerated), jointly enforcing power balance (19), unit limits (20), commitment binaries (29), and VoLL-based unserved energy (18).
• Decision is carried out through a multi-objective trade-off among (i) generation cost (EGSC, (23)), (ii) expected cost of energy not supplied (EECC, (18)), (iii) demand response program cost (ELRC, (24)), and (iv) carbon cost (ECEC, (22)). We use the fuzzy-ranking step in Algorithm 1 to select a single actionable plan from the non-dominated set.

This integrated design is not a direct reuse of MILP or Newman in isolation: the strategic phases become first-class constraints and objectives inside the optimizer, enabling auditable, scenario-consistent decisions and making the managerial framework operational for system planning under decarbonization targets.

3.2.3. Case Study

This study employs the IEEE 24-bus reliability test network to analyze the energy transition. The original system, detailed in

Table 4. Generation system data.

Fuel	Technology	Node	Capacity [MW]
Oil	Combustion turbine	1	40
		2	40
	Steam turbine	7	300
		13	591
		15	60
Coal	Steam turbine	15	155
		16	155
		23	660
Nuclear	Nuclear steam	18	400
		21	400
Water	Hydraulic turbine	22	300

, consists of 10 generation buses with conventional and hydroelectric technologies. Detailed data on generators, lines, and demand are provided in the annexes.

From the economic perspective, Table 5 specifies fixed costs by technology, while Table 6 details operational, fuel, and transmission variables.

Table 5. Generation system costs.

Fuel	Technology	Cost	FOM
		[$/MWh]	[$/MW]
Oil	Combustion turbine	10.22	409
	Steam turbine	-	-
Coal	Steam turbine	24.52	1154
Nuclear	Nuclear	54.84	2117
Hydro	Hydraulic turbine	0.92	1535
Wind	Wind turbine	60	1477

Table 5. Generation system costs.

Fuel	Technology	Cost	FOM
		[$/MWh]	[$/MW]
Oil	Combustion turbine	10.22	409
	Steam turbine	-	-
Coal	Steam turbine	24.52	1154
Nuclear	Nuclear	54.84	2117
Hydro	Hydraulic turbine	0.92	1535
Wind	Wind turbine	60	1477

Table 6. Variable generation system costs.

Fuel	Technology	Cost	VOM	Fuel	ElecT
		[$/MWh]	[$/MWh]	[$/MWh]	[$/MWh]
Oil	Combustion turbine	4.09	14.8	16.06	1.3
	Steam turbine	-	-	-	-
Coal	Steam turbine	3.07	40	-	0.8
Nuclear	Nuclear	0.43	0.4	-	-
Hydro	Hydraulic turbine	0.003	-	-	-
Wind	Wind turbine	26.67	-	-	-

Table 6. Variable generation system costs.

Fuel	Technology	Cost	VOM	Fuel	ElecT
		[$/MWh]	[$/MWh]	[$/MWh]	[$/MWh]
Oil	Combustion turbine	4.09	14.8	16.06	1.3
	Steam turbine	-	-	-	-
Coal	Steam turbine	3.07	40	-	0.8
Nuclear	Nuclear	0.43	0.4	-	-
Hydro	Hydraulic turbine	0.003	-	-	-
Wind	Wind turbine	26.67	-	-	-

Complementarily, Table 7 quantifies carbon emissions and their associated costs, which define parameters for emission reduction by replacing fossil fuels with renewables.

Table 7. Carbon emission rates and costs by technology.

Fuel	Technology	CE	EC
		[tonCO2/MWh]	[$/tonCO2]
Oil	Combustion turbine	0.618	35
	Steam turbine	-	-
Coal	Steam turbine	0.743	35
Nuclear	Nuclear steam	0.835	-
Water	Hydraulic turbine	-	-
Wind	Wind turbine	-	-

Table 7. Carbon emission rates and costs by technology.

Fuel	Technology	CE	EC
		[tonCO2/MWh]	[$/tonCO2]
Oil	Combustion turbine	0.618	35
	Steam turbine	-	-
Coal	Steam turbine	0.743	35
Nuclear	Nuclear steam	0.835	-
Water	Hydraulic turbine	-	-
Wind	Wind turbine	-	-

To reflect realistic operating conditions, interruption costs from demand-side management are included (

Table 8. Load interruption cost by customer segment due to rescheduling.

Load Type	IC
	[$/MWh]
Residential	150
Industrial	13,930
Commercial	12,870
Large consumers	13,930
Agriculture	650
Government	3460
Office	3460

) [41].

To illustrate the proposed methodology, the IEEE 24-bus test system is employed as a case study, as shown in Figure 5.

Rationale for the Test System and Real-World Applicability

The IEEE 24-bus Reliability Test System was selected as the reference network for the present study. This benchmark has been widely adopted in the literature because it ensures reproducibility, is computationally tractable for large-scale Mixed-Integer Linear Programming (MILP), and integrates diverse technologies such as thermal, hydro, and nuclear units together with realistic transmission constraints. These characteristics make it particularly suitable for evaluating strategic decarbonization approaches.

To ensure external validity, the parameters of the IEEE 24-bus system were calibrated against empirical ranges reported for real power systems, specifically those of Ecuador.

Table 9. Parameter mapping from real-world ranges to IEEE 24-bus model inputs.

Dimension	Real-World Range (Reference)	Model Treatment	Input Used
Demand growth	0.8–1.6%/year	Scaled hourly profile with compound growth	1.2%/year baseline
Carbon price ($/tCO2)	0–100+	Scenario parameter	{0, 25, 50, 100}
Hydro share	30–70% of energy	Hydro units modeled with annual energy budget	45% baseline
Oil/coal heat rates	Technology-specific	Reflected in fuel + VOM costs	As in Table 5 and Table 6
Wind capacity factor	28–40%	Hourly availability profile	33% mean, 0.15 std
Storage round-trip efficiency	0.75–0.90	Charging/discharging efficiencies	0.85 (base), 0.75–0.90 (sensitivity)

summarizes the parameter mapping that links real-world conditions to the IEEE 24-bus test system. This mapping preserves merit-order relations, fuel cost signals, and congestion patterns, while maintaining a transparent structure that facilitates replication.

This calibration ensures that the optimization results obtained with the IEEE 24-bus system are representative of realistic operating conditions, thereby strengthening the applicability of the findings to actual power system decarbonization strategies.

3.3. Sensitivity Analysis

The robustness of the optimization results was examined through a systematic sensitivity analysis. The assessment considered four key parameters: (i) carbon price ($EC$), (ii) maximum allowable wind capacity ($C_{\mathrm{wind}}^{max}$), (iii) annual demand growth rate (g), and (iv) round-trip storage efficiency (${\eta }_{\mathrm{rt}}$). For each parameter, the MILP was re-optimized, and the resulting system costs and emissions were compared against the baseline scenario.

The system cost in each sensitivity scenario s was calculated as

$$\mathrm{SystemCost}\left(s\right)=EGSC\left(s\right)+ELRC\left(s\right)+EC·\mathrm{Emissions}\left(s\right),$$

(30)

where total emissions were given by

$$\mathrm{Emissions}\left(s\right)=\sum _t\sum _{i\in \left\{\mathrm{oil},\mathrm{coal}\right\}}EP{G}_i\left(t\right)·C{E}_i.$$

(31)

Table 10. Sensitivity analysis results: variation in system costs and emissions relative to baseline.

Scenario	$\Delta$ Total Cost	$\Delta$ Emissions	Main Effect
$EC=25$/t	–6% to –12%	–18% to –25%	Acceleration of oil-to-hydro/wind shift
$EC=50$/t	–10% to –20%	–30% to –45%	Coal curtailed; storage and DR gain importance
$EC=100$/t	–20% to –35%	–55% to –70%	Fossil almost eliminated except peaking
$C_{\mathrm{wind}}^{max}=+25\%$	–4% to –8%	–12% to –18%	Integration feasible with moderate curtailment
$C_{\mathrm{wind}}^{max}=+50\%$	–7% to –14%	–20% to –30%	Transmission congestion becomes binding
$g=0.8\%$/y	–3% to –5%	–4% to –6%	Lower demand reduces peaking requirements
$g=1.6\%$/y	+4% to +7%	+5% to +9%	Higher demand increases reliance on DR
${\eta }_{\mathrm{rt}}=0.75$	+2% to +4%	+1% to +3%	Reduced arbitrage efficiency of storage
${\eta }_{\mathrm{rt}}=0.90$	–2% to –3%	–2% to –4%	Enhanced renewable absorption

summarizes the observed changes in system cost and emissions under the different scenarios. The analysis reveals that moderate carbon prices (≥50 $/tCO₂) induce significant emission reductions with limited cost impacts. Wind capacity expansion shows clear benefits when complemented by storage and transmission reinforcement. Higher demand growth increases system costs and highlights the importance of demand response programs. Improvements in storage efficiency further enhance the system’s capacity to integrate renewable energy.

4. Results Analysis

To evaluate the model, Phase 3 of William Newman’s method is implemented, applying the constraints and objective functions established in Phase 2. The analysis examines the impact of replacing fossil generation with renewable energy, demand-side management, and carbon emission reduction on system operating costs and emissions.

The study develops three scenarios. The first, the base scenario, represents the current situation. The second, gradual transition, targets a 50% emission reduction over 20 years. Finally, the third, accelerated transition, targets a 75% reduction over 10 years.

4.1. Current Situation Analysis

In the base scenario, the system operates without emission reduction constraints or wind integration. Only conventional generation technologies are considered: coal, oil, and nuclear. The analysis evaluates the distribution of generation, carbon emissions, and associated costs.

Figure 6 shows the hourly evolution of generation by fuel type. Oil-based generation (red line) remains relatively constant, while nuclear (green line) exhibits greater variability. In contrast, coal-fired generation (blue line) experiences significant fluctuations.

Figure 7 presents the aggregated generation values by technology.

Oil (GenOil) shows the largest share, followed by nuclear (GenLW) and coal (GenCoal). Wind generation (GenW) is absent in this scenario.

Carbon dioxide (CO₂) emissions by energy source are shown in Figure 8.

The comparative analysis of total CO₂ emissions across the three evaluated scenarios provides a clear visualization of the decarbonization potential of each pathway. While the base case represents the current fossil-dependent configuration, the gradual transition incorporates renewable expansion with moderate reductions, and the accelerated transition emphasizes a more ambitious substitution of fossil technologies. Figure 9 summarizes the cumulative effect of these strategies, highlighting the emission reduction trajectories and their alignment with long-term sustainability targets.

Figure 9. Comparison of total emissions by scenario. The gradual pathway reduces tCO₂ by 15.38% compared to the baseline case, while the accelerated pathway achieves 75% in 10 years.

Figure 9. Comparison of total emissions by scenario. The gradual pathway reduces tCO₂ by 15.38% compared to the baseline case, while the accelerated pathway achieves 75% in 10 years.

Oil and coal are the main emitters, with values exceeding 12,000 tons of CO₂, reaching 12,000 tons for coal and 14,000 tons for oil. In contrast, nuclear and hydroelectric technologies do not produce direct emissions.

Finally, Figure 10 shows the relationship between the total generation cost and the cost of emissions.

The graph highlights an inverse correlation between energy generation costs and emission-related expenses. The quantitative analysis reveals a decreasing linear trend: as the emission cost rises from 0 to 600$ annually, generation expenditure decreases from 90,000 to 32,000$ in the same period.

This relationship suggests that reducing fossil-based generation and increasing renewable integration lowers system operating costs. However, it simultaneously increases expenses from emissions, creating a critical balance between environmental sustainability and economic feasibility.

4.2. Gradual Emission Reduction Analysis

The emission reduction analysis is conducted over a 20-year horizon, considering a gradual increase in energy demand with a constant 20% expansion factor. Within this context, a structured transition process is implemented, in which conventional technologies progressively reduce their share, while renewable sources increase proportionally. These initial conditions form the operational basis for interpreting the results on generation, emissions, and costs presented below.

Figure 11 summarizes the system’s energy, environmental, and economic performance during the planning horizon. The annual generation chart shows a gradual reduction in oil and coal use, progressively replaced by hydro, wind, and, to a lesser extent, nuclear sources. This reconfiguration of the energy matrix results in sustained CO₂ emission reductions, achieving a 50% decrease between years 9 and 20, as illustrated in the annual emission evolution chart. Economically, annual operating costs rise moderately due to the integration of renewable technologies. Additionally, annual emission costs steadily decline in terms of monetized environmental impact, demonstrating that the transition strategy is both technically and economically viable.

Figure 11. Energy, environmental, and economic performance of the system during the planning horizon—gradual transition.

Figure 11. Energy, environmental, and economic performance of the system during the planning horizon—gradual transition.

During the planning horizon, shown in Figure 12, electricity demand is continuously met despite the gradual phase-out of fossil sources. This balance is achieved through the progressive redesign of the energy mix, with a sustained increase in wind and hydro generation compensating for the reduction of coal and the eventual phase-out of oil in the final period. Nuclear generation maintains a steady and moderate contribution, serving as a backup technology in the transition.

The cumulative generation by source, described in Figure 13, shows that hydro accounts for 36% of the total, followed by oil at 28.6%, despite its progressive elimination in the final stages. Wind generation represents 16%, reflecting its growing incorporation into the energy mix. Coal and nuclear contribute 11.7% and 7.8%, respectively. This distribution highlights a system still partially dependent on conventional technologies, yet with notable progress toward renewable sources in terms of cumulative contribution.

The cumulative emissions balance, represented in Figure 14, shows that oil was responsible for 8238 tCO₂, marking the largest environmental impact, followed by coal with 4563 tCO₂, together totaling 12,802 tCO₂. This concentration of emissions in two technologies fully justifies their progressive reduction and eventual retirement within the optimization model. The evidence confirms that the adopted strategy successfully guided the system toward a cleaner configuration without compromising coverage or technical operability.

4.3. Accelerated Emission Reduction Analysis

The accelerated transition scenario covers a 20-year period, considering a constant 20% increase in electricity demand. Under this scheme, an intensive substitution strategy is applied, in which conventional technologies rapidly decrease their share, favoring an accelerated penetration of renewable sources. This operational configuration provides the foundation to analyze system outcomes—focused on generation, emissions, and costs—with a 75% reduction achieved within the first 10 years.

Figure 15 illustrates an accelerated reconfiguration process of the system, where renewable sources, especially hydro and wind, gain early participation, while coal and oil use are drastically reduced starting in year 7. This technological substitution enables a 75% reduction in annual CO₂ emissions by the end of year 10, as shown in the emission chart. Although generation costs show a sustained increase, emission-related costs decline significantly, evidencing an environmentally efficient transition with a moderate economic penalty.

Figure 15. Energy, environmental, and economic performance of the system during the planning horizon—accelerated transition.

Figure 15. Energy, environmental, and economic performance of the system during the planning horizon—accelerated transition.

During the accelerated transition, shown in Figure 16, electricity demand is continuously met despite the early retirement of conventional technologies. The gradual reduction of coal and the complete elimination of oil in the final year are offset by the steady growth of hydro generation, reinforced nuclear output, and greater wind participation.

This dynamic shows that, even under an accelerated substitution scenario, the system preserves operational stability and response capacity without compromising performance.

Meanwhile, the cumulative generation profile, detailed in Figure 17, indicates that, despite the short planning horizon, coal remains the dominant source with 35.9%, followed by nuclear at 20.2% and hydro at 19.1%. Oil generation falls significantly to 14.6%, while wind power reaches 10.2%, reflecting its gradual incorporation into the energy mix. These results highlight that, although the transition is accelerated, the operational inertia of conventional technologies continues to dominate cumulative contributions.

Cumulative emissions reach 6171 tCO₂, with coal accounting for 77% (4742 tCO₂) and oil the remaining 23% (1430 tCO₂), as shown in Figure 18. Compared to the base scenario, this represents a 75% reduction by the end of year 10, aligning with the model’s target. Therefore, the accelerated substitution strategy proves effective in reducing environmental impact over short periods without affecting system operation.

To provide a clear quantitative comparison across scenarios,

Table 11. Comparative summary of scenarios: emissions and costs.

Scenario	CO2 Emissions [tCO2]	Reduction [%]	Annual Cost [$]
Base case	26,000	0.0	90,000
Gradual transition (20 years)	22,000	15.38	40,000 + 600 (emissions)
Accelerated transition (10 years)	6,171	75.0	48,000 + 150 (emissions)

summarizes the main indicators, including total CO₂ emissions, percentage reduction relative to the base case, and annual generation costs. This comparison highlights the progressive effectiveness of the gradual and accelerated transition pathways relative to the baseline scenario.

4.4. Sensitivity Analysis to Carbon Pricing

An additional analysis was conducted to evaluate the sensitivity of system costs to carbon pricing. This assessment is essential because the economic performance of decarbonization pathways depends not only on technology integration but also on regulatory instruments, such as carbon pricing.

Figure 19 shows the total system generation cost under different carbon price scenarios, ranging from 0 to 100$/tCO₂. The results reveal a clear upward trend: as carbon prices increase, generation costs rise due to higher penalties on fossil fuel technologies. This behavior demonstrates that, under stricter carbon pricing regimes, the optimal system configuration shifts more rapidly toward renewable sources and storage, thereby reinforcing the role of clean technologies in long-term cost minimization.

4.4.1. Impact of Renewable Penetration Threshold

Figure 20 shows the effect of increasing the minimum renewable penetration threshold y from 20% to 60%. As renewable requirements increase, fossil fuel generation decreases significantly, leading to emission reductions. Nonetheless, higher renewable shares also require additional investments in storage and demand response, slightly increasing operational costs.

4.4.2. Discussion

The sensitivity results confirm the robustness of the model. While both carbon pricing and renewable mandates effectively drive decarbonization, the combination of these policies yields the most significant emission reductions at manageable costs. This validates the proposed framework as a flexible decision-support tool adaptable to diverse policy and economic contexts.

4.5. Phase 4: Evaluation and Validation of the Energy Transition

In Phase 4 of William Newman’s method, the results of the energy transition are evaluated and validated. This stage verifies the effectiveness of the strategies applied in Phase 3 by comparing proposed objectives with the outcomes obtained.

The analysis considers key metrics such as emission reduction, operating costs, and the share of renewable energy. These indicators make it possible to identify potential adjustments or improvements in the optimization model.

4.5.1. Emission Reduction Evaluation

The model set a target of reducing emissions by 50% through a gradual 20-year transition. However, the results show a reduction of only 15.38%, from 26,000 to 22,000 tons of CO₂.

This outcome indicates that, although the applied strategy had a positive impact, it did not achieve the expected level of decarbonization. Therefore, additional strategies are needed to achieve a more effective reduction.

4.5.2. Cost Validation

The base scenario represents a system without emission restrictions and with predominant fossil generation. In this context, the annual generation cost reaches $90,000, while emission costs remain negligible at around $0. This reflects dependence on coal and oil, which are high operating-cost technologies without environmental penalties.

In contrast, the gradual transition scenario reduces fossil participation and increases renewable share. Here, generation costs decrease to $40,000 annually, but emissions costs reach $600 annually due to penalties. This shows that replacing polluting sources with renewables optimizes operating costs, although it introduces temporary financial impacts from environmental regulations.

The downward trend confirms that decarbonization entails initial costs but reduces expenses in the long term. Nevertheless, to ensure economic viability, complementary strategies are required—such as energy storage systems to guarantee operational stability and fiscal incentives to facilitate the progressive retirement of fossil assets.

5. Conclusions

The comprehensive analysis of the current power system revealed that coal- and oil-based thermoelectric generation remain the main contributors to greenhouse gas emissions. In the base scenario, these technologies account for 12,000 and 14,000 tons of CO₂, respectively, underscoring the urgent need for an accelerated and structured energy transition.

The implementation of Mixed-Integer Linear Programming as a core optimization tool allowed for the evaluation of strategic pathways for decarbonization. Under the gradual transition scenario, the integration of wind and hydro power reduced oil-related emissions to 10,000 tons, while coal remained steady at 12,000 tons, achieving an overall reduction of 15.38%. Although this demonstrates tangible progress, the results fell short of the targeted 50% reduction, highlighting that incremental renewable integration alone is insufficient to reach ambitious climate goals.

The accelerated transition scenario, by contrast, confirmed the technical feasibility of achieving a 75% emission reduction within 10 years. This was made possible through the rapid phase-out of fossil-based generation, combined with significant renewable penetration and the deployment of energy storage systems. However, the results also revealed higher operational costs during the early years, emphasizing the economic trade-offs that policymakers and system operators must address.

From an economic perspective, the study showed that decarbonization not only reduces long-term operational costs but also enhances system resilience against volatility in fossil fuel markets. Nevertheless, short-term financial penalties associated with carbon pricing and renewable integration costs require complementary measures. These include fiscal incentives, regulatory frameworks that accelerate the retirement of carbon-intensive assets, and investments in grid flexibility and storage technologies.

Technically, the research highlights the importance of balancing renewable penetration with reliability requirements. Demand response programs, Hybrid Renewable Energy Systems, and binary operation of generation units proved essential to maintaining supply security and system stability throughout the transition.

In summary, the proposed framework validates that sustainable and economically viable decarbonization of the power sector is achievable under specific conditions. Success requires a combination of four pillars: (i) accelerated renewable integration, (ii) large-scale deployment of storage systems, (iii) active demand-side participation, and (iv) strong regulatory and fiscal policies. These elements together ensure that emission reduction targets can be realistically met without compromising energy security or economic competitiveness. Beyond its immediate findings, the framework developed is replicable and adaptable, offering a practical roadmap for other power systems facing similar decarbonization challenges.