Remuneration of Ancillary Services from Microgrids: A Cost Variation-Driven Methodology

Abstract

Microgrids (MGs) have emerged as pivotal players in the energy transition by enabling the efficient integration of distributed energy resources and the provision of ancillary services to the power system. Despite their technical capabilities, MGs still face economic and regulatory barriers that hinder their widespread deployment in electricity markets. This paper presents a structured methodological framework to assess the economic viability of MGs delivering services such as peak shaving, loss compensation, and voltage support, among others. The proposed approach considers three distinct scenarios: (1) MGs supplying energy to local loads, (2) hybrid MGs combining local supply with ancillary services, and (3) MGs exclusively dedicated to ancillary services. The framework incorporates adjusted levelized cost of electricity (LCOE), levelized avoided cost of electricity (LACE), and net value metrics, while accounting for tax incentives and market price signals. A case study based in Colombia (Cali and Camarones) validates the framework through simulations conducted in HOMER Pro V3.18.4 and MATLAB Online. The results indicate that remuneration schemes based on availability and service utilization significantly enhance the viability of MGs. The proposed methodology is applicable to emerging regulatory environments and offers guidance for designing public policies that promote the active participation of MGs in supporting grid operations.

1. Introduction

The acceleration of the energy transition has intensified the demand for more resilient, flexible, and sustainable power systems. In this context, MGs have emerged as strategic players by integrating distributed generation, energy storage, and intelligent control systems. Their potential extends beyond supplying local loads to providing ancillary services that enhance the operation and reliability of the main grid [1]. Despite proven capabilities in frequency regulation, voltage support, black starts, loss compensation, and demand-response programs (2), MG integration into electricity markets remains limited due to the absence of robust and tailored economic remuneration mechanisms.

The international literature has explored various economic assessment models to evaluate the financial viability of MGs under different participation frameworks [2,3,4,5]. The traditional levelized cost of electricity (LCOE) has proven insufficient to capture the added value of these services. In response, complementary metrics such as the levelized avoided cost of electricity (LACE) and net value have been developed to better reflect the marginal contribution of MGs to the power system [6,7,8,9].

This study introduces a novel methodological framework that integrates three complementary techno-economic pillars—LCOE, LACE, and net value—under specific regulatory and fiscal conditions relevant to the Colombian context. Unlike prior studies that often relied on purely technical or descriptive analyses, this approach incorporates a multi-scenario methodology that considers critical operational variables to assess the systemic contribution of microgrids. Additionally, the framework integrates fiscal incentives from Colombian legislation, particularly those established in Law 1715 of 2014 to promote non-conventional renewable energy sources through accelerated tax deductions of up to 50% of the investment. This convergence of economic, operational, and regulatory indicators represents a substantial methodological contribution to both national and international state-of-the-art practices.

In countries like Colombia, where the regulatory framework has yet to explicitly integrate MGs as active participants in the ancillary services market [10], it becomes essential to develop economic models that recognize their systemic value. Progress has been made through regulations such as CREG Resolution 101 019 of 2022 and 101 054 of 2024, which, although not directly promoting the use of MGs, enable their participation in demand-response initiatives. The methodological approach presented in this article could also be extended to other jurisdictions, as evidenced by markets such as CAISO and ERCOT, where storage and distributed resources compete in ancillary service auctions [11], or in NYISO and PJM, where new day-ahead uncertainty and ramping products are being developed to improve flexibility signals [12]. Likewise, international studies emphasize that microgrid-based schemes can be systematically incorporated into ancillary service remuneration mechanisms, reinforcing the broader applicability of the proposed framework [13].

The proposed methodology is structured based on three progressive scenarios: (1) MGs supplying energy to local loads and generating revenue solely from electricity sales; (2) hybrid MGs that both supply loads and deliver ancillary services; and (3) MGs exclusively designed to provide ancillary services. For each scenario, dedicated revenue and cost models were developed. The case studies conducted in Cali and Camarones focused on the delivery of peak shaving and load shifting services within a simulated urban environment, enabling validation of the practical applicability and sensitivity of the proposed models to variables such as availability price, activation price, surplus energy sales, and avoided cost.

This work contributes meaningfully to both the academic literature and emerging regulatory frameworks by:

• Developing an adaptable and replicable methodological framework to assess the economic viability of MGs in unregulated markets.
• Formulating the LACE model for the MG context, including scenario weighting criteria and avoided cost simulations.
• Proposing remuneration mechanisms based on availability and service utilization that enhance the ancillary services market delivered by MGs.
• Demonstrating the validity of one of the proposed models through a case study conducted in two Colombian regions with contrasting climatic and reliability conditions, supporting the formulation of policy and regulatory recommendations grounded in empirical evidence.

This article is organized as follows. Section 2 (Methodological Framework) describes the analytical, economic, and technical basis used to model the three MG scenarios. Section 3 (Case Study) presents the numerical results and economic indicators obtained from the evaluation of each scenario, with particular emphasis on the application of peak shaving and load shifting services. Section 4 (Conclusions) highlights the main findings, outlines the study’s limitations, and proposes directions for future research.

2. Methodological Framework

Assessing the financial viability of a microgrid that provides ancillary services requires expanding beyond traditional economic metrics. The LCOE is commonly used to compare generation technologies. This metric represents the total cost of building and operating a power plant divided by the total energy it produces over its lifetime [14]. However, the conventional LCOE presents key limitations. It fails to reflect temporal variations in value and does not account for added benefits such as ancillary service provision or system backup capabilities. For instance, LCOE overlooks the higher value of electricity during peak hours and the role of generation assets in enhancing grid stability. To address these gaps, complementary metrics have been proposed that incorporate both cost and system value contributions, including the LACE and net value.

To frame the evaluation of ancillary services, the following key assumptions are defined for the analysis horizon.

• Multi-year horizon: A planning horizon of N years is considered, corresponding to the expected lifetime of the microgrid or its main asset. This horizon includes technical parameters (e.g., capacity, efficiency) assumed to remain relatively constant, though degradation or equipment replacement may be incorporated if relevant. A discount rate $r$ is applied to bring future cash flows to present value, aligned with the project’s cost of capital.
• Hourly resolution: System operation is modeled on an hourly basis (8760 h per year or the necessary interval) to capture daily and seasonal fluctuations. For each hour of every year, the energy generation and the activation of each ancillary service are estimated.
• Project costs: The model includes the initial investment (capital costs of equipment, installation, etc.), fixed annual costs (e.g., fixed operation and maintenance not dependent on hourly output), and variable costs (e.g., fuel, wear and tear, variable O&M linked to energy generation or service provision). For simplicity, annual costs can be grouped into fixed annual CAPEX (including investment amortization if not paid upfront, fixed O&M, etc.) and hourly variable OPEX associated with operational activity in each hour (e.g., fuel cost/kWh generated in a given hour).
• Grid integration: The model assumes that the microgrid operates in coordination with the main grid, exchanging energy and services based on market prices or bilateral agreements. Network physical constraints are not explicitly considered. It is assumed that whenever the microgrid generates power or provides a service, it yields a quantifiable benefit to the system.

This article presents a detailed viability assessment methodology for microgrids that provide ancillary services under three distinct scenarios, incorporating financial metrics.

• Scenario 1: A microgrid that exclusively supplies specific local loads and sells electricity at a price higher than its LCOE. The analysis assesses viability when the selling price exceeds the levelized cost of generating that energy.
• Scenario 2: A hybrid microgrid that simultaneously serves local loads, sells surplus electricity, and provides ancillary services to the grid. The methodology evaluates how to incorporate the additional costs and benefits of ancillary services using the net value approach based on both LCOE and LACE metrics.
• Scenario 3: A microgrid exclusively designed to deliver ancillary services to the grid, without supplying local loads or selling excess energy. The evaluation again applies the net value method to determine its economic feasibility.

Each scenario includes a sequence of steps in which mathematical models, financial frameworks, or remuneration schemes are developed. The analysis determines the economic feasibility of the proposed configurations in each case.

2.1. Scenario 1

This scenario serves as the baseline case. The MG focuses on supplying electricity to specific local loads, such as an isolated community or an industrial facility, and may inject surplus energy into the main grid. In this configuration, the explicit provision of ancillary services is not considered. The primary objective is to meet local demand in a cost-effective manner. The key assumption is that the microgrid sells electricity at a price higher than its LCOE, indicating potential profitability. The evaluation steps for this scenario center on comparing the LCOE with the actual selling price of the generated energy.

2.1.1. Calculation of the LCOE of the MG

This value reflects the cost/kWh of electricity generated by the microgrid, incorporating capital investment, operation and maintenance (O&M), fuel, and other related expenses [15]. These costs are expressed as a constant unit cost over the project’s lifetime, typically in USD/kWh or USD/MWh.

The LCOE can be calculated using the following equation:

$$LCOE={\displaystyle \frac{{\sum }_{t=1}^T\frac{C_{\mathrm{CAPEX},t}+C_{\mathrm{OPEX},t}}{{\left(1+r\right)}^t}}{{\sum }_{t=1}^T\frac{E_t}{{\left(1+r\right)}^t}}}$$

(1)

where:

• $C_{\mathrm{CAPEX},t}$ Capital expenditures in year t.
• $C_{\mathrm{OPEX},t}$ Operating and maintenance costs in year t.
• $E_t$ Energy generated in year t.
• r Discount rate.
• T Project lifetime

2.1.2. Determination of the Electricity Selling Price and Load Served

The selling price corresponds to the rate at which the microgrid can sell electricity to end users or the grid. In isolated systems, this price may reflect local consumer tariffs. In grid-connected schemes, it could represent the market selling price or a bilateral contract rate. Based on this scenario’s premise, the selling price exceeds the LCOE.

2.1.3. Compare Selling Price with LCOE

This represents the basic profitability test per unit of energy. If the selling price surpasses the LCOE, every kilowatt-hour sold yields a positive gross margin for the microgrid. This can be expressed as:

$${P_U}_{EN}=C_{EN}-LCOE$$

(2)

where:

• ${P_U}_{EN}$ Unit profit margin from energy sales “net value.”
• ${\mathrm{C}}_{EN}$ Energy selling price per kWh.
• $LCOE$ Levelized cost of electricity.

If ${P_U}_{EN}>0$, the MG generates profit for each unit of energy produced.

2.1.4. Consideration of Additional Savings or Avoided Costs (Tax Incentives)

In some countries, such as Colombia, national legislation has promoted the use of renewable energy sources (RESs). Law 1715 of 2014, for instance, established three key incentives that help accelerate RES deployment [16,17].

• Value-added tax (VAT) exemption: Equipment, machinery, components, and services associated with RES investments are exempt from VAT. This exemption applies to solar panels, wind turbines, batteries, efficient diesel generators, and other assets, effectively reducing the project’s upfront capital cost by the applicable VAT rate (19% in Colombia).
• Fifty percent investment deduction from income tax: Taxpayers may deduct up to 50% of the investment value in RES projects from their taxable income over a period not exceeding 15 years. In practice, this deduction is often applied in the early years to maximize its present value. A common application involves distributing the deduction over five years (i.e., 10% of the investment annually), allowing for partial capital recovery through tax savings.
• Accelerated depreciation over five years: Project assets may be depreciated over a shortened five-year period (e.g., 20% annually), instead of the typical 15–20 years. This approach enables significant tax savings during the early operational years by lowering taxable income.

A simplified way to include these tax benefits in the LCOE calculation involves applying a factor $\mathsf{\omega }$ that modifies the project’s CAPEX. This factor captures only the investment deduction and the accelerated depreciation incentives:

$$\mathsf{\omega }={\displaystyle \frac{1}{1-\mathsf{\tau }}}\left[1-\mathsf{\tau }\left(\sum _{j=1}^{T_1}{\displaystyle \frac{i}{{\left(1+r\right)}^j}}+\sum _{j=1}^{T_2}{\displaystyle \frac{d}{{\left(1+r\right)}^j}}\right)\right]$$

(3)

where:

• $\tau$ Effective income tax rate.
• $T_1$ Number of years for investment deduction.
• $T_2$ Number of years for accelerated depreciation.
• $i={\displaystyle \frac{0.5}{T_1}}$ Annual fraction for investment deduction.
• $d={\displaystyle \frac{1}{T_2}}$ Annual fraction for accelerated depreciation.
• $r$ Annual discount rate.

To incorporate the VAT exemption, simply deduct the VAT rate from the initial capital cost. The adjusted LCOE of the MG considering tax benefits is expressed as follows:

$$LCOE={\displaystyle \frac{{\sum }_{t=1}^T\frac{\omega \times \frac{C_{\mathrm{CAPEX},t}}{(1+IVA)}+C_{\mathrm{OPEX},t}}{{\left(1+r\right)}^t}}{{\sum }_{t=1}^T\frac{E_t}{{\left(1+r\right)}^t}}}$$

(4)

2.2. Scenario 2

The second scenario, based on [18], considers a microgrid that serves local loads while simultaneously delivering ancillary services (ASs) to the power system. This setup reflects a realistic use case for advanced grid-connected microgrids. For instance, a university campus may operate a microgrid to meet part of its internal demand while occasionally exporting surplus energy or providing voltage regulation to the utility. In this context, the microgrid plays a dual role: local generation and AS provider. The evaluation methodology builds upon Scenario 1 by analyzing the additional benefits that AS bring to a baseline case where the microgrid only supplies local loads.

2.2.1. Definition of the Baseline Case and the AS Case

Given the hybrid operation of the MG, two cases must be analyzed. The first, referred to as the “baseline case”, corresponds to Scenario 1 and follows the net value approach. The second, the “AS case”, incorporates the modifications necessary to enable AS delivery. Both cases should maintain the same infrastructure, except for essential and exclusive upgrades required to provide AS, such as additional energy storage capacity.

2.2.2. Characterization of the Operation of the Hybrid Microgrid

This step involves defining the resources allocated to AS provision. Some shared resources may serve both local demand and AS, depending on technical or economic criteria defined by the microgrid operator.

The analysis identifies how the microgrid allocates capacity between local demand and ASs. For instance, it may operate a distributed generator below its maximum output to maintain headroom for frequency regulation or spinning reserve, avoiding violations of voltage limits [19]. Likewise, a diesel generator might run below its rated capacity to preserve inertial reserves for ASs, which increases the cost per kilowatt-hour due to reduced efficiency.

Mathematically, for any given hour h, if the MG has a maximum generation capacity $P_{Max}$ and a local demand $D_h$, it may allocate a generation level $P_{g,h}=D_h$ to supply local loads. Additional energy may be purchased from the grid if needed. However, the microgrid must retain a reserve $R_h$ such that $P_{g,h}+R_h\le P_{Max}$. If energy storage is available, it can support transient injections without disrupting $P_{g,h}$. In all cases, the microgrid dispatch must integrate AS requirements while ensuring the supply to critical loads remains uninterrupted.

2.2.3. Identification of Incremental Costs and Investments for ASs

Delivering ASs may require additional investment or operational expenditures. Based on Equation (1), the levelized cost of electricity for this scenario is:

$${LCOE}_{E2}={LCOE}_{Cb}+{LCOE}_{CAS}$$

(5)

where:

• ${LCOE}_{E2}$ Levelized cost of electricity for Scenario 2.
• ${LCOE}_{Cb}$ Levelized cost for the baseline case.
• ${LCOE}_{CAS}$ Levelized cost for the AS case.

2.2.4. Identification of Additional Benefits from ASs

This section quantifies the revenues or savings the MG can earn from providing ASs. These benefits may include direct payments from the grid operator or market administrator for offered reserve capacity, energy contributions that improve network parameters, or the avoidance of penalties.

The analysis focuses on estimating the economic value derived from AS provision. To this end, the services proposed in [20], are examined, classifying them into three categories:

• Conventional Ancillary Services

• Frequency control support
• Voltage support
• Black starts
• Reserve services: energy storage

• Pilot services

5.

Peak shaving and load shifting

6.

EV storage

• Under development

7.

Loss compensation

8.

Other services

• Frequency control support

Frequency control support is fundamentally an active power capacity service, requiring the availability of generation at the time the service is requested. The revenue a generating unit receives for providing reserve capacity for frequency control over one hour is calculated as follows:

$${\mathrm{R}}_{\mathrm{fc}}=C_{\mathrm{fc}}\times X_{\mathrm{fc}}$$

(6)

where:

• ${\mathrm{R}}_{\mathrm{fc}}$ Revenue from frequency control reserve.
• $C_{\mathrm{fc}}$ Price of frequency control reserve.
• $X_{\mathrm{fc}}$ Reserved capacity for frequency control support.

This service is bidirectional, meaning that an MG may be instructed either to increase or decrease its power output. When requested to increase active power output using the reserved capacity, the MG receives additional compensation for the energy delivered in response:

$$R_{\mathrm{EN},\mathrm{fc}}=C_{fc,t}\times E_{fc,t}$$

(7)

where:

• $R_{\mathrm{EN},\mathrm{fc}}$ Revenue from frequency control activation.
• $C_{fc,t}$ Activation price for frequency control.
• $E_{fc,t}$ Activated energy for frequency control.

2.

Voltage support

MGs are typically connected to a point of common coupling (PCC) through power electronic controllers capable of regulating reactive power. Voltage support involves the injection or absorption of reactive power (capacitive or inductive) to regulate voltage at the PCC. The available reactive power reserve depends on the device’s rated apparent power and real-time active power output:

$$Q_{\max }=\pm \sqrt{S_{\max }^2-P_{\mathrm{act}}^2}$$

(8)

where:

• $Q_{\max }$ Maximum available reactive power.
• $S_{\max }^2$ Maximum apparent power.
• $P_{\mathrm{act}}^2$ Actual active power.

This capacity enables voltage correction in both directions and serves as the basis for voltage support provisioning. For practical purposes, a microgrid’s reactive power capability is denoted $Y_{\mathrm{vc}}$, and the compensation for maintaining this reserve is defined as:

$${\mathrm{R}}_{vc}=C_{\mathrm{vc}}\times Y_{\mathrm{vc}}$$

(9)

where:

• ${\mathrm{R}}_{\mathrm{vc}}$ Revenue from voltage support reserve.
• $C_{\mathrm{vc}}$ Price of voltage support reserve.
• $Y_{\mathrm{vc}}$ Reserved capacity for voltage support.

However, simply reserving capacity does not ensure compliance with voltage regulation standards. It is necessary to maintain voltage levels within acceptable deviation limits (%V), typically defined by local regulation. Compensation for maintaining voltage within the required margin is calculated as:

$$R_{vc,t}=\left({\displaystyle \frac{T_{\mathrm{est}}}{T_{\mathrm{t}}}}\right)\cdot C_{vcmax}$$

(10)

where:

• ${\mathrm{R}}_{\mathrm{vc},\mathrm{t}}$ Revenue from voltage support performance.
• $C_{\mathrm{vc}\max }$ Maximum price for voltage support performance.
• $T_{\mathrm{est}}$ Time within compliance range.
• $T_{\mathrm{t}}$ Total billing time.

3.

Black Starts

A black start service is critical for restoring grid operation after a complete or partial blackout. It refers to a microgrid’s ability to initiate operation without external power supply. This service becomes especially valuable during emergencies and requires seamless transition from grid-connected to islanded mode. Remuneration methods for conventional power plants, such as the one described in [21], form the basis of the proposed model:

$$R_{bs}=\left(C_{bs}\times D_{bs}+{CT}_{+}\right)\times \left(1+F_m\right)$$

(11)

where:

• ${\mathrm{R}}_{bs}$ Revenue from black start availability.
• $C_{bs}$ Price for black start availability.
• $D_{bs}$ Required black start availability.
• ${CT}_{+}$ Additional test cost or compensation
• $F_m$ System recovery factor, evaluated by the system operator.

When the service is activated, the MG receives compensation based on the energy supplied, priced according to reconciliation mechanisms defined in the contract:

$$R_{\mathrm{EN},\mathrm{bs}}=C_{EN}\times E_{bs,t}$$

(12)

where:

• $R_{\mathrm{EN},\mathrm{bs}}$ Revenue from black start activation.
• $C_{EN}$ Energy selling price.
• $E_{bs,t}$ Energy delivered during black start.

To ensure availability, participating MGs must undergo periodic testing. If an MG fails these tests, it becomes ineligible for compensation until successful retesting confirms operational readiness.

In the event of a service failure during an actual contingency, the MG is suspended from receiving any compensation until retesting confirms proper function. Additionally, a financial penalty is imposed, requiring the MG to reimburse the cost of service provided by backup units during the outage.

4.

Reserve Services: Energy Storage

Energy storage plays a critical role in MGs by enabling the optimal integration of distributed energy resources, especially renewable energy sources. A wide range of storage technologies exists, each with unique characteristics based on discharge time, energy content, and operational functionality. Additionally, storage systems act as enablers for multiple services, including those previously discussed and demand-response programs [22]. Various regulatory frameworks recognize storage as a support mechanism for transmission and distribution systems. For example, Colombia’s Resolution CREG 098 of 2019 [23] and Chile’s Decree 70 of 2023 [24] provide such guidance.

To provide greater clarity on the regulatory alignment between Chile and Colombia,

Table 1. Comparative overview of storage reserve remuneration frameworks in Chile and Colombia.

Feature	Chile’s Decree 70 of 2023	Resolution CREG 098 of 2019
Service	Capacity reserve via energy storage	Capacity reserve via energy storage
Remuneration basis	Availability (≥4 h discharge duration)	Availability (guaranteed through a regulated tender contract)
Payment structure	Power-duration adjusted factor	Annual expected revenue (IAE), defined by contract
Contract horizon	10 years of guaranteed recognition	15 years of guaranteed payments via UPME tender
Revenue source	Regulated capacity charges at the subsystem level	T&D tariffs, depending on project location
Penalty for non-compliance	Power reduction via availability factor; additional penalties may apply	Contractual reduction or legal penalties for breach
Incentive for high availability	Optional, not explicitly regulated	Not explicitly defined
Market interaction	Complementary: does not require energy market participation	Passive role: no direct participation in wholesale energy markets

compares how each jurisdiction approaches the remuneration of storage systems when used for reserve services. Both regulations base payments on availability; however, the Chilean mechanism relies on a power-duration factor and guarantees 10 years of recognition, while the Colombian approach secures payment through a 15-year expected revenue contract awarded by public tenders. In both cases, underperformance or low availability leads to penalties, and financing depends on regulated charges. This structured comparison underscores how regulatory frameworks can converge in their goals while differing in implementation details.

This section adopts Chile’s decree as a reference point, focusing exclusively on storage availability and remuneration. Firstly, the MG-offered power capacity for storage services must be quantified using:

$$P_{ESS}=\sum _{t=1}^{360}{P}_n\left(\mathrm{t}\right)\times {\mathrm{F}}_{ESS}\left(t\right)$$

(13)

where:

• $P_{ESS}$ Energy storage capacity.
• $P_n$ Nominal storage power.
• ${\mathrm{F}}_{ESS}$ Duration factor $=min\left({\displaystyle \frac{Duration}{Contrat}},1\right)$.

The duration factor ${\mathrm{F}}_{ESS}$ should reflect the technical characteristics of the MG. It equals 1 (100%) if the full installed capacity remains available for the minimum contracted duration each day. Accordingly, the revenue from storage availability is calculated as:

$$R_{ESS}=P_{ESS}\times C_{ESS}$$

(14)

where:

• ${\mathrm{R}}_{\mathrm{ESS}}$ Revenue from energy storage availability.
• $C_{ESS}$ Price for energy storage availability.

5.

Peak Shaving and Load Shifting

Peak shaving and load shifting can be addressed within a unified remuneration mechanism in MGs. These services offer strategic functionality by reducing peak demand through the coordinated use of storage systems, local generation, and demand-side management. They not only contribute to grid stability but also enable active participation of consumers as distributed energy resources.

Regulatory frameworks have increasingly recognized their relevance as ancillary services. In the Colombian context, although Resolution CREG 101 019 of 2022 [25] and 101 054 of 2024 [26] do not specifically target MGs, they may serve as enablers for their integration. This section outlines a remuneration mechanism based on those regulations and the model developed by [27].

First, the availability revenue is determined by:

$$R_{DR}=X_{DR}\times C_{DR}$$

(15)

where:

• $R_{DR}$ Revenue from demand-response availability.
• $X_{DR}$ Firm available capacity.
• $C_{DR}$ Payment for available capacity.

In addition, activation revenue must be considered. This is based on the actual reduction in energy consumption relative to the baseline agreed upon in the contract. The reduced energy is calculated as:

$$E_{DR}={\int }_{T1}^{T2}{(E}_{bl}-E_l)dt$$

(16)

where:

• $E_{DR}$ Reduced energy.
• $T1,T2$ Start and end times of service activation.
• $E_{bl}$ Baseline consumption.
• $E_l$ Actual consumption.

Finally, the revenue from service activation is calculated as:

$$R_{DR,t}=E_{DR}\times C_{DR,t}$$

(17)

where:

• $R_{DR,t}$ Revenue from demand-response activation.
• $C_{DR,t}$ Activation payment.

6.

EV storage

The temporary energy storage service through electric vehicles, known as vehicle-to-grid (V2G), represents an emerging capability for optimizing MG operations. This service enables connected EVs to inject energy into the grid during peak hours, thus contributing to demand reduction during critical periods. V2G constitutes a form of active demand-side management that leverages the flexibility of mobile resources without requiring additional stationary storage infrastructure. This section proposes a remuneration mechanism for EV users participating in V2G schemes based on an energy arbitrage and battery degradation model as presented in [28].

For a user to find energy injection economically viable, the hourly selling price must exceed the combined cost of energy acquisition and battery degradation, adjusted by system efficiency:

$$LBMP\left(t\right)>{\displaystyle \frac{C_{grid}+C_{dg}}{n^2}}$$

(18)

where:

• $LBMP$ Location-based marginal price.
• $C_{grid}$ Grid energy import cost.
• $C_{dg}$ Battery degradation cost.
• $n$ Charge efficiency.

Battery degradation is modeled based on depth of discharge, total storage system cost, and expected battery cycle life:

$$C_{dg}={\displaystyle \frac{E_{S,max}\times c_b}{2\times DoD\times L_C\times E_{S,max}}}$$

(19)

where:

• $c_b$ Storage system cost per unit of energy.
• $DoD$ Depth of discharge.
• $L_C$ Battery cycle life.
• $E_{S,max}$ Maximum storage capacity.

If the economic condition is satisfied, the revenue from energy injection is calculated as:

$$R_{EVS}=\left(E_S\left(t-1\right)-E_S\left(t\right)\right)\times \mathsf{\eta }\times LBMP$$

(20)

where:

• $E_S\left(t\right)$ Stored energy at time t.
• $\mathsf{\eta }$ Unidirectional efficiency.

7.

Loss compensation

Loss compensation refers to the ability of a microgrid to inject energy at strategic points in the distribution or sub-transmission network to reduce technical losses, mainly caused by joule heating in conductors and transformers. These losses depend primarily on the square of the current, conductor resistance, and line length. Technical guidelines such as those by EPM [29] offer analytical methods for quantifying losses in transmission and distribution systems:

$$P=NCP\times 3\times {\left({\displaystyle \frac{I}{NCP}}\right)}^2\times \cdot R\times l\times F_p$$

(21)

where:

• $I$ Phase current.
• $NPC$ Number of conductors per phase.
• $R$ Conductor resistance per km.
• $l$ Line length.
• $F_p$ Loss factor.

The loss factor accounts for hourly load variability and is defined as:

$$F_p=0.7\cdot F{c}^2+0.3\cdot Fc$$

(22)

where:

• $Fc$ Load factor, i.e., the ratio between average and peak power consumption.
• $Fc={\displaystyle \frac{PotenciaPromediokVA}{PotenciaMáximakVA}}$.

Compensated energy is estimated as:

$$E_{lc}=P\times T_{prom}$$

(23)

where:

• $E_{lc}$ Compensated energy.
• $T_{prom}$ Average activation time.

The corresponding revenue from service activation is then given by:

$$R_{lc,t}=E_{lc}\times C_{lc,t}$$

(24)

where:

• $R_{lc,t}$ Revenue from loss compensation activation.
• $C_{lc,t}$ Activation payment for loss compensation.

8.

Other Services

This section presents a remuneration mechanism for multiple services that support the distribution or transmission system operator. The compensation structure is tied to the ability of these services to defer capital investments. The services considered here include congestion management, phase balancing, and oscillation damping.

The first step involves calculating the present cost of the investment being deferred. This requires knowledge of the amount and timing of the investment. The net value (NV) of the investment without MG intervention is computed following the methodology outlined in [18]:

$${\mathrm{N}\mathrm{V}}_{YBC}={\displaystyle \frac{C_{Inv}}{{\left(1+r\right)}^{YBC}}}$$

(25)

where:

• ${\mathrm{N}\mathrm{V}}_{YBC}$ Net value of the investment without MG intervention.
• $YBC$ Base-case year of the investment without MG support.
• $C_{Inv}$ Investment cost in year $YBC$.
• $r$ Interest rate.

If an MG is integrated and provides services that allow the deferral of this investment, the resulting economic benefit can be captured by recalculating the NV considering the postponed investment year:

$${\mathrm{N}\mathrm{V}}_{YMG}={\displaystyle \frac{C_{Inv}}{{\left(1+r\right)}^{YMG}}}$$

(26)

where:

• ${\mathrm{N}\mathrm{V}}_{YMG}$ Net value of the investment with MG intervention.
• $YMG$ Postponed investment year due to MG support.
• $C_{Inv}$ Investment cost in year $YMG$.
• $r$ Interest rate.

The savings from deferring the investment, ${NV}_{Inv}$, are calculated as the difference between the two NVs:

$${NV}_{Inv}={\mathrm{N}\mathrm{V}}_{YBC}-{\mathrm{N}\mathrm{V}}_{YMG}$$

(27)

The microgrid is entitled to a portion of the savings as compensation for its contribution to investment deferral:

$$R_{Inv}={f\times NV}_{Inv}$$

(28)

where:

• $R_{Inv}$ Revenue earned by the MG for deferring the investment.
• $f$ Agreed percentage of the total savings.

This revenue may be paid as a lump sum or distributed evenly across the project’s time horizon, assuming annual settlement of the investment deferral benefits.

As a summary,

Table 2. Summary of remuneration models—Scenario 2.

Service	Revenue from Availability	Revenue from Activation
Frequency control support	${\mathrm{R}}_{\mathrm{fc}}=C_{\mathrm{fc}}\times X_{\mathrm{fc}}$	$R_{\mathrm{EN},\mathrm{fc}}=C_{EN}\times E_{fc,t}$
Voltage support	${\mathrm{R}}_{vc}=C_{\mathrm{vc}}\times Y_{\mathrm{vc}}$	$R_{vc,t}=\left({\displaystyle \frac{T_{\mathrm{est}}}{T_{\mathrm{t}}}}\right)\times C_{vcmax}$
Black Start	$R_{bs}=\left(C_{bs}\times D_{bs}+CT\right)\times \left(1+F_m\right)$	$R_{\mathrm{EN},\mathrm{bs}}=C_{EN}\times E_{bs,t}$
Reserve services “energy storage”	$R_{ESS}=P_{ESS}\times C_{ESS}$	N/A
Peak shaving	$R_{DR}=X_{DR}\times C_{DR}$	$R_{DR,t}=E_{DR}\times C_{DR,t}$
Load shifting
EV storage	Not applicable	$R_{EVS}=\left(E_S\left(t-1\right)-E_S\left(t\right)\right)\times \mathsf{\eta }\times LBMP$
Loss compensation	Not applicable	$R_{lc,t}=E_{lc}\times C_{lc,t}$
Congestion management	$R_{Inv}={f\times NV}_{Inv}$	Not applicable
Phase balancing
Oscillation damping

presents the revenue from availability, revenue from service use, and total revenue for each evaluated service.

2.2.5. Calculation of the LACE of Service

The levelized avoided cost of electricity (LACE) represents the benefits that an MG contributes to the power system. As defined in [6,30], it refers to the average revenue per unit of energy that a new generation project or MG provides to the grid. According to [8], the LACE also captures the potential income for the project owner from energy and capacity sales. Lastly, in [31] it is described as the monetized value per unit of electricity delivered, and is calculated as:

$$LAC{E}_p={\displaystyle \frac{{\sum }_{t=1}^T\frac{{\mathrm{R}}_D+{\mathrm{R}}_U+R_{EN}}{{\left(1+r\right)}^t}}{{\sum }_{t=1}^T\frac{E_s^p+E_t^p}{{\left(1+r\right)}^t}}}$$

(29)

where:

• $LAC{E}_p$ Levelized avoided cost for service p.
• ${\mathrm{R}}_D$ Revenue from availability of service p in year t.
• ${\mathrm{R}}_U$ Revenue from activation of service p in year t.
• ${\mathrm{R}}_{EN}$ Revenue from energy sales in year t.
• $E_s^p$ Energy used for providing service p in year t.
• $E_t^p$ Energy used for selling electricity in year t.
• r Discount rate.
• T Project lifetime.

The availability and activation revenues are computed using the models in Table 2, while energy sales revenues correspond to energy used for satisfying local demand and selling surplus (if applicable) to the grid at the contractual rate.

The LACE expresses the monetized value of energy, linking the revenues generated to the amount of energy required to produce them.

2.2.6. Calculation of the Net Value of the Microgrid

The net value (NV) of the scenario quantifies the difference between the benefits and the costs of the MG, serving as an indicator of the project’s profitability relative to its expenditures. Both the LACE and LCOE must be computed using the same time horizon, discount rate, and energy basis to ensure consistency. The net value is calculated as:

$$\mathrm{N}\mathrm{e}\mathrm{t}\mathrm{V}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e}=LAC{E}_p-LCOE$$

(30)

If net value < 0, the MG is financially unviable. If net value = 0, the project breaks even. A positive net value implies economic viability.

Table 3. Net value reconciliation example.

Scenario Description	LCOE USD/kWh	LACE USD/kWh	Net Value USD/kWh
Base Case	0.180	0.235	0.055
High Diesel Penetration	0.245	0.210	−0.035
Optimized Hybrid (Service)	0.165	0.270	0.105

provides a one-row reconciliation example for clarity:

These values demonstrate how design and operational decisions significantly influence the financial feasibility of the microgrid. The “high diesel penetration” scenario results in a negative net value, while the optimized configuration reduces costs and enhances benefits, ensuring economic viability.

2.3. Scenario 3

In this scenario, the MG is designed and implemented solely to provide ancillary services to the power system without serving any local load. A typical example may involve an industrial microgrid equipped with generation and battery assets dedicated to providing frequency regulation or reserve services to the system operator. Since the primary purpose is not electricity sales to end users, but rather grid support, the evaluation methodology must quantify the cost of service delivery and compare it to the corresponding compensation. The procedure follows a similar structure to Scenario 2.

2.3.1. Definition of the Ancillary Services and Required Resources

The first step is to identify the ancillary services the MG will deliver (e.g., frequency regulation, voltage control, black start) and size the system accordingly. This involves determining the generation or storage capacity needed and the operating strategy (e.g., operating generators below rated capacity to provide reserves, or using fast-responding storage for inertial response). It is assumed that the MG will operate in accordance with the system operator’s requirements.

2.3.2. Calculation of the LCOE of the Microgrid

In this context, the LCOE is computed using Equation (1) with the distinction that only the energy effectively delivered for service provision is included. Internal energy flows (e.g., energy cycling in storage systems) are excluded:

$$LCOE={\displaystyle \frac{{\sum }_{t=1}^T\frac{C_{\mathrm{CAPEX},t}+C_{\mathrm{OPEX},t}}{{\left(1+r\right)}^t}}{{\sum }_{t=1}^T\frac{E_t^{AS}}{{\left(1+r\right)}^t}}}$$

(31)

2.3.3. Calculation of the LACE of the Microgrid

As in the previous scenario, the LACE is computed, except that no energy is sold to local loads or exported to the grid:

$$LAC{E}_p={\displaystyle \frac{{\sum }_{t=1}^T\frac{{\mathrm{R}}_D+{\mathrm{R}}_U}{{\left(1+r\right)}^t}}{{\sum }_{t=1}^T\frac{E_t^{AS}}{{\left(1+r\right)}^t}}}$$

(32)

2.3.4. Calculation of the Net Value of the Microgrid

The same evaluation criteria from Scenario 2 apply here. However, both the LCOE and LACE are computed considering only the ancillary services delivered.

Figure 1 presents the methodological framework, which compares three scenarios (energy supply only, hybrid supply plus ancillary services, and ancillary services only) through common analytical steps of data collection, technical-economic modeling, and regulatory integration, leading to the calculation of LCOE, LACE, and net value for assessing microgrid viability.

To better illustrate the relationship between the three scenarios, Figure 2 shows the common elements and how to obtain economic benefits.

3. Case Study

This section presents the development of a case study for the peak shaving and load shifting services in two locations in Colombia: the rural settlement of Camarones in La Guajira and the city of Cali in Valle del Cauca. A simulation is conducted using HOMER Pro software to obtain dispatch data, enabling an economic assessment of the viability of these services at both sites. Data processing was carried out in MATLAB.

The first step involves sizing the MG under evaluation. HOMER Pro software was used for this purpose, leveraging geographic, technical, and economic data to size a grid-connected or islanded MG. In this case, a grid-connected MG is designed with a power constraint to emulate the peak shaving service. The procedure followed for sizing the MG is illustrated in Figure 3.

In addition to technical specifications, configuring the simulation in the software also requires financial input, including both capital expenditures (CAPEX) and operational expenditures (OPEX). For both case studies, financial data were sourced from technical reports such as [22,32,33]. The input parameters used in HOMER Pro are summarized in

Table 4. Input data for simulation in HOMER Pro.

Component	Input Parameters	Value	MG
Diesel generator	Initial capital	USD 12,500	Cali/Camarones
	Replacement	USD 12,500	Cali/Camarones
	O&M	USD 0.750	Cali/Camarones
	Fuel price	USD 2	Cali/Camarones
	Power	25 kW	Cali/Camarones
	Fuel curve intercept	0.825 L/h	Cali/Camarones
	Fuel curve slope	0.273 L/h/kW	Cali/Camarones
PV system	Initial capital	USD 1200/kW	Cali/Camarones
	Replacement	USD 800/kW	Cali/Camarones
	O&M	USD 30/year	Cali/Camarones
	Time	25 years	Cali/Camarones
	Rated capacity	164 kW	Cali/Camarones
Battery	Initial capital	USD 300/kW	Cali/Camarones
	Replacement	USD 300/kW	Cali/Camarones
	O&M	USD 30/year	Cali/Camarones
	Time	10 years	Cali/Camarones
	Throughput	800 kWh	Cali/Camarones
	Nominal voltage	12 V	Cali/Camarones
	Nominal capacity	1 kWh	Cali/Camarones
	Maximum capacity	83.4 Ah	Cali/Camarones
	Capacity ratio	0.403	Cali/Camarones
	Rate constant	0.827/h	Cali/Camarones
	Roundtrip efficiency	80%	Cali/Camarones
	Maximum charge current	16.7 A	Cali/Camarones
	Maximum discharge current	24.3 A	Cali/Camarones
	Maximum charge rate	1 A/Ah	Cali/Camarones
Sinexcel 150 kW	Initial capital	USD 45,000	Cali/Camarones
	Replacement	USD 25,000	Cali/Camarones
	O&M	USD 1000/year	Cali/Camarones
Load	Average	976.96 kWh/day	Cali/Camarones
		40.71 kW	Cali/Camarones
	Peak	73.19 kW	Cali/Camarones
	Peak month	August	Cali/Camarones
	Load factor	0.56	Cali/Camarones
Grid	Annual purchase capacity	40 kW	Cali/Camarones
	Mean outage frequency	7.7/year	Cali
	Mean repair time	12.2 h	Cali
	Mean outage frequency	34/year	Camarones
	Mean repair time	51 h	Camarones

.

The mean outage frequency and mean repair time are key reliability indicators used to compare the two distinct locations. In Cali, both metrics fall below the national average, whereas Camarones, a remote locality, ranks among the areas with the poorest reliability indices in the country. These parameters play a crucial role in sizing the MG, as it must not only provide peak shaving and load shifting services but also ensure the continuity of supply to the entire load demand. These reliability values were configured in the “Reliability” section of the “Advanced Grid” module in HOMER Pro. Figure 4 displays the corresponding inputs for the Cali scenario.

To configure the tariffs, this study employed standard hourly electricity rates used by Colombian energy retailers For this case, the analysis used ENERTOTAL’s May tariff data [34]. Additionally, the month of August was set as the peak consumption period. The available capacity was set at 1.5 times the peak demand (75 kW) to ensure coverage during contingencies and allow for future system expansions. A capacity availability price of USD 100/kW was defined and distributed across the entire project life cycle. The time-of-use structure divides the day into five bands (F1 to F5) according to demand profiles. Off-peak periods correspond to F1 and F2: F1 spans from 00:00 to 04:59 with a price of USD 0.2175/kWh and a sellback rate of USD 0.1724/kWh; F2 covers 05:00 to 09:59 and 13:00 to 18:59, with a price of USD 0.2214/kWh and a sellback rate of USD 0.1756/kWh. Peak periods correspond to F3, F4, and F5: F3 runs from 10:00 to 12:59, F4 from 19:00 to 21:59, and F5 from 22:00 to 23:59. The prices for these bands are USD 0.2229/kWh (F3), USD 0.2244/kWh (F4), and USD 0.2232/kWh (F5), respectively, while the sellback rates are USD 0.1767/kWh (F3), USD 0.1781/kWh (F4), and USD 0.1769/kWh (F5). These values were configured in the “Rate Definition” section of the “Advanced Grid” module in HOMER Pro [35], as illustrated in Figure 5 and Figure 6.

Figure 7 depicts the demand-response baseline constructed from the average hourly consumption profile of regulated customers. The curve exhibits three distinct phases: a low-consumption period between 00:00 and 05:00 h, with values close to 20 kW; a moderate increase during the morning (06:00–12:00 h), reaching approximately 30 kW; and a pronounced afternoon–evening peak (16:00–20:00 h), with maximum values above 50 kW. This profile was adopted as the counterfactual reference against which demand reductions are measured during activation events. The baseline reflects the typical daily load behavior of medium-sized consumers and provides a consistent foundation for quantifying demand-response impacts in the proposed methodology.

We conducted four simulations grouped into two categories: “with energy sales” and “without energy sales”. Each category included two locations (Camarones and Cali). Additionally, a sensitivity analysis was performed for grid import limits of 30 kW, 40 kW, and 50 kW to identify the most cost-effective threshold for grid purchase, as illustrated in Figure 8a. These simulations yielded four optimized configurations of distributed energy resources. For each case, the configuration with the lowest LCOE was selected, aiming to minimize generation costs and maximize net value, which ultimately reflects the economic benefit.

Table 5. Simulation outcomes in HOMER Pro.

Resource	Location (MG)	Selected Value		Other Simulations with Sales
		with Sales	Without Sales	PV and Storage	Generator and Storage
Solar PV	Cali	228 kW	131 kW	236 kW	-
Generator		25 kW	25 kW	-	25 kW
Battery		557 kWh	556 kWh	809 kWh	534 kW
Purchase		50 kW	40 kW	50 kW	50 kW
LCOE		USD 0.149/kWh	USD 0.247/kWh	USD 0.171/kWh	USD 0.329/kWh
Solar PV	Camarones	423 kW	383 kW	474 kW	Not applicable
Generator		25 kW	25 kW	-	Not applicable
Battery		1420 kWh	1340 kWh	1967 kWh	Not applicable
Purchase		40 kW	40 kW	50 kW	Not applicable
LCOE		USD 0.300/kWh	USD 0.399/kWh	USD 0.316/kWh	Not applicable

presents the selected results.

Once all modules were configured, a single-line diagram of the microgrid was generated, as shown in Figure 8b. The legend for interpreting the graphs is provided in Figure 9, and the dispatch results for Cali appear in Figure 10 and Figure 11.

In both figures, the light-blue line represents the grid energy purchase. During demand peaks (black line), this constraint forces the MG to rely on its internal resources to meet the remaining load. The orange line represents the inverter power output, highlighting the microgrid’s contribution during periods of high consumption. The teal line indicates battery charging, while the green line shows battery discharging, which plays a critical role in mitigating demand peaks that exceed the grid purchase limit. During low-demand periods, the MG charges the batteries, and during high-demand events, it discharges to reduce grid dependence, effectively delivering the intended peak shaving service. Surplus energy exports, visible only in Figure 10, occur when internal generation exceeds load requirements. Together, these graphs demonstrate the coordinated operation of all components to flatten the grid purchase curve and increase the value delivered by the MG.

The Camarones MG displays similar dispatch behavior. However, its higher reliance on diesel backup (purple line) results from reduced grid reliability in the region. Figure 12 illustrates a time interval when the grid fails and the MG operates in islanded mode, supplying the full load independently.

The data exported from HOMER Pro were processed using a MATLAB script that implemented Equations (2), (4), (15)–(17), (29), and (30) to assess the financial viability of each scenario for the peak shaving service.

Table 6. LACE and LCOE results calculated in MATLAB.

MG	with Energy Sales		Without Energy Sales		Lifetime	r
	LCOE	LACE	LCOE	LACE
Cali	USD 0.1584/kWh	USD 0.2019/kWh			25 years	0.0808
			USD 0.2368/kWh	USD 0.2269/kWh
Camarones	USD 0.1909/kWh	USD 0.2002/kWh
			USD 0.1996/kWh	USD 0.2169/kWh

summarizes the simulation results.

Among the four configurations analyzed, the MG in Cali under the “with energy sales” scenario achieved the lowest cost-to-output ratio. This performance results from the high energy production relative to its associated costs. In both locations, when energy export is allowed, the LACE exceeds the LCOE, indicating that the service not only covers generation costs but also delivers added value to the MG. When energy sales are not permitted, this trend reverses in Cali, diminishing its economic appeal. Conversely, Camarones still exhibits a slight margin where the LACE surpasses the LCOE, suggesting that geographic location and energy export capabilities significantly influence project viability. Figure 13 and Figure 14 illustrate the LACE-versus-LCOE comparison across all scenarios.

These results enabled the estimation of the profit margin in each case. For Scenario 1 (Section 2.1) (E1), Equation (2), was used, considering a selling price $C_{EN}$ of 0.17 USD/kWh. For Scenarios 2 (Section 2.2) (E2) and 3 (Section 2.3) (E3), the net value equation (30) was applied using each scenario’s respective LCOE.

Table 7. Profit margin evaluation.

MG	Profit Margin
	with Energy Sales		Without Energy Sales
	E1	E2	E1	E3
Cali	USD 0.0116/kWh	USD 0.0435/kWh
			USD −0.0654/kWh	USD −0.0100/kWh
Camarones	USD −0.0209/kWh	USD 0.0093/kWh
			USD −0.0296/kWh	USD 0.0173/kWh

presents the profit margins for both MGs under both energy sales conditions.

The analysis of profit margins highlights the direct impact of the MG’s remuneration and operational schemes on its profitability. In E2, both locations show positive margins, confirming that peak shaving with energy export schemes yields favorable economic outcomes. Cali achieved the highest margin in this scenario. In E1, Cali maintains a positive margin, while Camarones shows a negative one, indicating that the local sale price does not sufficiently cover its costs. In the “without energy sales” scenarios, Cali’s margins remain negative for both E1 and E3, suggesting that these configurations are financially unviable. However, in Camarones, while E1 continues to underperform, E3 yields a positive margin, revealing that local conditions and revenue structure can ensure project viability even in the absence of energy exports.

Figure 13b illustrates the comparison between the LCOE and LACE for Cali under the no-export E3. In this case, the LACE value falls below the LCOE, highlighting a clear limit of economic viability. This outcome shows that the revenues derived from ancillary services are insufficient to offset the levelized generation costs, reducing the project’s financial appeal. As a result, the microgrid cannot recover its costs with the available remuneration, confirming that under this operational scheme, economic sustainability in Cali is not feasible.

Figure 15 shows the one-way sensitivity analysis for scenario E3 in Cali, assessing the response of LCOE, LACE, and net value to variations in CAPEX, OPEX, and diesel price. The results demonstrate that battery CAPEX (Bat) is the most critical variable: increases beyond 100% drive the LCOE above USD 0.30/kWh while reducing the net value to strongly negative levels. Photovoltaic CAPEX (FV) also has a relevant impact, although to a lesser extent, whereas variations in OPEX and fuel price produce only marginal changes. This behavior confirms that high investment costs, particularly in storage, are the main driver of infeasibility in the “without energy sales” scenario.

Figure 16 summarizes these findings through a tornado diagram that ranks the parameters according to their effect on the net value. In this case, battery CAPEX stands out as the dominant variable, capable of shifting the project between feasible (orange) and infeasible (blue) conditions. Photovoltaic CAPEX ranks second in importance, while OPEX and diesel price have a negligible influence on the project’s economic outcome. This prioritization clearly identifies the critical parameters that must be addressed to improve profitability.

Taken together, Figure 15 and Figure 16 demonstrate that the infeasibility of scenario E3 does not stem from operating costs or fuel expenditures, but rather from high upfront investment, particularly in storage. Therefore, reducing CAPEX or implementing remuneration schemes that properly account for the value of ancillary services emerges as a necessary condition to ensure the economic sustainability of the microgrid.

After identifying that investment costs are the main driver of infeasibility in scenario E3 (Figure 15 and Figure 16), a bivariate analysis was carried out to assess the combined effect of photovoltaic CAPEX and battery CAPEX. Figure 17 presents the results: the left panel shows that only under simultaneous reductions of both investment costs (yellow areas) are positive margins in the net value (NV) achieved. The feasible region is restricted to values below 90% for both variables, confirming that high upfront investment remains the main barrier to project viability. The central panel illustrates the evolution of the LCOE, which steadily increases as investment costs rise, reaching values above USD 0.30/kWh in pessimistic scenarios. The white dot in each panel represents the base case at 100% for both variables. Finally, the right panel indicates that the LACE remains constant, reflecting that remuneration in this scenario relies exclusively on ancillary services rather than energy sales. This outcome reinforces the conclusion that financial feasibility is strongly constrained by CAPEX, particularly that of storage.

To explore alternatives that could offset high investment costs, Figure 18 presents the bivariate analysis of available capacity for demand response ($X_{DR}$) and the corresponding payment ($C_{DR}$). The left panel highlights that simultaneous increases in these variables significantly improve the net value (NV), reaching feasible areas (yellow zones) that surpass the profitability threshold. The central panel confirms that the LCOE remains constant, consistent with its dependence on generation costs rather than remuneration. In contrast, the right panel reveals that the LACE grows with both capacity availability and payment, showing that ancillary services can become the main lever to enhance economic sustainability when energy exports are not allowed. Consequently, designing tariff schemes that properly reward demand response emerges as a key strategy to ensure the viability of scenario E3.

Taken together, the sensitivity analysis results confirm that the infeasibility of scenario E3 is not driven by minor operational aspects, but by the underlying cost and remuneration structure.

The sensitivity analysis shows that project feasibility strongly depends on the CAPEX and OPEX of solar PV and storage. For policymakers, this highlights the need for incentives or tariffs that reduce the risks of capital-intensive technologies. For investors, it points to prioritizing projects where storage costs are declining or where remuneration rewards availability and performance. The impact of diesel price volatility further stresses the importance of regulatory frameworks that promote renewable integration. Overall, the results provide guidance for aligning investment strategies and policy design toward more cost-effective microgrid deployment.

4. Conclusions

The findings from this study provide a robust methodological foundation for assessing the economic viability of MGs that deliver ancillary services to the power system. The scenario comparison reveals that remuneration schemes based on availability and utilization significantly improve profit margins, particularly when surplus energy sales are permitted. In the case of Cali, the hybrid scenario (E2) yields the highest economic return (USD 0.0435/kWh), whereas the absence of energy sales undermines project viability. In Camarones, despite more stringent reliability conditions, the ancillary services-only scenario (E3) achieves positive margins, demonstrating that MGs can remain profitable even without energy sales, provided that an appropriate remuneration framework is in place.

Additionally, the sensitivity analysis performed on key parameters (CAPEX, OPEX, and demand-response variables) strengthens the validity of the findings. The results demonstrate that the infeasibility of scenario E3 in Cali primarily stems from high investment costs, particularly those associated with storage. However, they also reveal that remuneration schemes based on availability and capacity payments for demand response offer an effective pathway to restore financial viability. This highlights the importance of designing regulatory mechanisms that explicitly recognize the value of ancillary services in contexts where surplus energy sales are not allowed.

Beyond peak shaving, the models developed for frequency support, voltage regulation, black starts, loss compensation, and storage management effectively quantify the system-level benefits of these services. The proposed formulations integrate technical and economic criteria, making them particularly relevant for market agents and grid operators interested in leveraging MGs for ancillary service provision. The introduction of the LACE metric as a valuation tool, along with net value analysis, offers a comprehensive approach to recognizing the systemic contributions of MGs, surpassing the limitations of traditional LCOE assessments.

This research holds particular relevance for the Colombian power sector, where current regulations do not fully address the integration of MGs into ancillary service markets. The proposed models may guide the development of public policies, regulatory frameworks, and tariff instruments that reflect the real value of these services, thereby promoting the deployment of more sustainable and decentralized MG-driven technologies in the country.

Future research could expand upon this work by developing dynamic and optimized models that account for seasonal variations in availability and activation pricing, as well as uncertainty in consumption and renewable generation patterns. Additionally, exploring interactions between multiple MGs within the same node could inform the design of aggregation mechanisms or local ancillary service markets. Finally, validating the remaining models through simulations or pilot projects in Colombia would provide the empirical grounding needed to refine the remuneration mechanisms based on actual system behavior.

Future research should also emphasize the role of pilot projects as a means of empirical validation. While the proposed framework demonstrates technical and economic consistency, field implementations are essential to capture operational uncertainties, regulatory constraints, and behavioral factors that cannot be fully modeled. Pilot experiences would not only refine the remuneration mechanisms but also provide evidence for regulators and investors, strengthening confidence in the scalability of microgrid participation in ancillary service markets.