The accurate characterisation of electricity consumption in HSPs is a fundamental stage in assessing the technical feasibility and economic viability of PV systems. This stage aims to identify the actual demand for electrical energy, define critical consumption patterns, and establish parameters for optimal system sizing. The process involves five structured sub-stages applicable across healthcare settings.

The sizing of the PV system is a crucial stage that defines its generation capacity in relation to the healthcare facility’s electricity demand, the local solar resource availability, and technical and regulatory constraints specific to HSPs. Proper sizing ensures that the PV system operates efficiently, meets energy reliability requirements, and optimises economic performance. This phase involves two key sub-stages: determining the required PV generation and selecting system components, including the storage subsystem. The sizing of the PV system in HSPs follows a structured methodology composed of four sequential stages.

Evaluating the economic feasibility of PV systems in HSPs under uncertainty requires a methodology that goes beyond deterministic cash flow analysis. The ROM addresses this challenge by incorporating flexibility into the investment decision process. Under this framework, the investment is treated as a real option ‒ analogous to a financial derivative ‒ where the investor holds the right, but not the obligation, to undertake the project depending on how future conditions evolve [42].

ROM allows for a dynamic assessment of the project’s value by considering not only the static net benefit but also the managerial flexibility to expand, defer, reduce, or abandon the project in response to changing variables such as electricity prices, solar resource availability, or regulatory signals. The ROM-based evaluation follows a structured process consisting of five stages.

Understanding the robustness of PV investments in HSPs requires evaluating how uncertainty in key variables affects financial outcomes and managerial flexibility. This stage integrates sensitivity analysis and scenario modelling within the ROM framework to assess how fluctuations in external conditions influence the project’s value and the optimal investment strategy. The analysis follows a two-pronged approach.

This final stage in the financial evaluation process for PV projects in HSPs involves selecting the most viable investment alternative. The selection is based on the results generated by the ROM, particularly the valuation of expansion, reduction, deferral, and abandonment options using both the Binomial and Black-Scholes models.

The ROM framework supports not only the quantification of investment value under uncertainty but also the strategic selection of the most advantageous course of action. To facilitate this decision-making process, a multi-criteria analysis (MCA) is used. This method integrates both financial metrics and strategic considerations, ensuring a balanced assessment of options under varying scenarios and risk profiles.

The proposed methodology was validated through its application in a real-world setting: a level 2 healthcare service provider (HSP) located in Barranquilla, Colombia. Validation included:

• Empirical implementation: The methodology was applied using actual energy consumption data and local economic parameters, ensuring contextual realism.

• Benchmarking against traditional methods: Financial results obtained through the ROM were compared with those from the NPV approach, showing improved flexibility and strategic value.

• Scenario and sensitivity analysis: The methodology was stress-tested against variations in electricity prices, solar irradiation, and investment costs. Elasticity values were calculated to evaluate model responsiveness.

• Multi-criteria evaluation: A structured decision-making framework was applied to rank investment options (expansion, reduction, deferral, abandonment), incorporating technical, financial, and strategic dimensions.

These steps confirmed the method’s robustness, adaptability to uncertainty, and practical applicability in healthcare settings. Although this validation is based on a single case, the structured nature of the framework facilitates replication in other institutions and sectors.

The methodology for the financial evaluation of photovoltaic projects using the ROM was applied in a hospital in Barranquilla, Colombia, classified as level 2 [58]. This Health Service Provider (HSP) offers inpatient, emergency, surgery, outpatient, diagnostic, and therapeutic support services, including intensive care, nursing, and medical specialties. Solar radiation in the area varies between 5.25 and 6.8 kWh/m² annually, with an average Peak Sun Hours (PSH) of 5.3 hours per day [45].

For the study, several areas of the HSP were selected, including the emergency department, intensive care, inpatient care, operating rooms, diagnostic imaging, clinical laboratories, outpatient clinics, and administration. These spaces use monitoring and life support equipment, emergency and general lighting, and air conditioning systems. They also have CT scanners, MRI scanners, X-ray equipment, and sterilisation systems. Additionally, they have computer and ventilation equipment, all of which require high energy consumption and service continuity. Based on the results of the ROM in the selected areas, a decision will be made as to whether the project to install photovoltaic systems at the HSP under study will be expanded, reduced, postponed, or abandoned.

The first step in the financial evaluation was to characterise the energy consumption of the selected areas. This stage allows for identifying electricity demand, analysing consumption patterns, assessing power quality, and determining the energy a solar generation system can supply.

Energy data for 2023 was collected from billing records, on-site measurements with power grid analysers, peak demand analysis, and an electrical equipment inventory. Figure 2 shows the hourly load profile of the HSP, including the disaggregated consumption by department and the total daily demand.

Figure 2.

Load profile in the HSP under study: (a) hourly electricity consumption distribution by functional area; (b) total consumption

Figure 2a shows the hourly electricity demand disaggregated by hospital departments. The largest contributions during peak hours come from the Surgical Suite, Outpatient Clinics, and Inpatient Care, reflecting the concentration of clinical services between 08 h and 14 h. This detailed breakdown helps to identify consumption hotspots and underlines the importance of matching PV generation with daytime load.

Figure 2b depicts the load profile of the monitored hospital areas, averaging 91.8 kWh/ day (33 522 kWh/ year) with a peak demand of 12.5 kW. The early-morning valley (00:00–06:00 h) ranges between 0.7 kW and 1.7 kW, consistent with the operation of only emergency lighting, patient-monitoring devices, and life-support equipment. From 06:00 h the load rises steadily, reaching 4.6 kW at 09:00 h as clinical and administrative activity ramps up, and peaks at 12.5 kW around midday due to the simultaneous operation of diagnostic equipment and HVAC. Demand then declines to 7.8 kW at 14:00 h and 5.9 kW at 15:00 h, before falling below 2 kW after 18:00 h and stabilising near 0.7 kW overnight.

Approximately 60% of daily energy is consumed between 08:00 and 16:00 h, coinciding with the 5.3-hour local peak-sun window. Although Colombia applies a flat electricity tariff, self-consuming solar generation during these hours still reduces the total kilowatt-hours purchased from the grid and thereby improves the project’s NPV. The marked intra-day swing in demand (about 35% volatility between valley and peak) also feeds directly into the ROM. By quantifying management’s ability to expand photovoltaic capacity if future consumption grows or defer upgrades if clinical schedules change, the ROM assigns a flexibility premium that the static NPV cannot capture. Consequently, the load profile not only supports the positive deterministic NPV but also underpins the strategic options valued through the ROM, strengthening the investment case under uncertain operational and regulatory conditions.

Figure 3 shows the percentage distribution of energy consumption across the functional areas, critical areas, air conditioning and lighting systems, and administrative equipment.

Figure 3.

Distribution of energy consumption by functional areas

According to Figure 3, the largest percentage of consumption is attributed to air conditioning systems (30%), which are essential for maintaining thermal comfort for patients and ensuring the proper operation of temperature-sensitive medical equipment. Critical areas, including intensive care units, operating rooms, emergency rooms, and clinical laboratories, account for 25% of total consumption, reflecting their high energy demand. Lighting consumes 20% of the energy, due to its extensive use throughout the hospital, while specialised medical equipment absorbs 15%, primarily due to the operation of monitoring and treatment devices.

Figure 3 indicates that air-conditioning systems account for 30% of the hospital’s electricity consumption (they preserve thermal comfort for patients and safeguard temperature-sensitive equipment). Critical medical areas (intensive-care units, operating theatres, emergency rooms and clinical laboratories) contribute an additional 25%, reflecting their continuous high-power demand and strict reliability requirements. Lighting represents 20%, driven by round-the-clock operation in corridors, wards and diagnostic suites, whereas the remaining 15% corresponds to specialised medical devices (mainly monitoring and treatment equipment that operate intermittently at elevated power levels).

This breakdown has two direct implications for the photovoltaic project. First, because HVAC and lighting loads are concentrated during daylight hours, most of the solar generation can be self-consumed, which supports the positive NPV calculated for a system sized to cover about half of the annual demand. Second, the inflexible demand of critical areas establishes the minimum level of security supply and therefore shapes the downside scenarios considered in the ROM (for instance, the decision to defer expansion if additional backup capacity is required). By distinguishing flexibility from non-flexible loads, Figure 3 thus provides the operational rationale for the strategic options valued in the ROM analysis.

The size of the photovoltaic system was determined to cover approximately 50% of the energy consumption in the selected areas of the HSP. This system would supply 16,761 kWh/year, thereby reducing grid energy consumption and lowering the cost per unit of energy, ensuring a favourable return on investment. Figure 4 shows the comparison between the HSP's load profile and the estimated generation from the proposed photovoltaic system.

Studies, such as those presented in [58] and [59], recommend sizing photovoltaic systems in HSPs to cover between 50% and 60% of total demand. This recommendation considers several factors. Firstly, the variability of solar irradiation can reduce energy production on cloudy days or during periods of low radiation. Secondly, although rooftop installation is the preferred strategy, the available surface is typically limited due to the presence of technical infrastructure (e.g., HVAC systems, ventilation units, water tanks), safety clearances, and access routes. These constraints reduce the effective area for PV deployment, making it difficult to fully meet energy demand with solar power alone. Finally, the investment costs of a fully self-sufficient system would be high and could exceed the budget of many healthcare institutions.

Figure 4.

Load profile and projected photovoltaic system for the HSP section under study

Figure 4 confirms that daytime photovoltaic output overlaps well with the hospital load curve, covering a significant share of consumption between 09 h and 13 h. After midday, the load begins to exceed solar generation by as much as 6.5 kW, and during the night production is zero, so the grid remains indispensable. Deterministically, this profile justifies sizing the system at 50% of annual demand because the kilowatt-hours displaced coincide with the hours of highest irradiance (even though the Colombian tariff is flat). Strategically, the midday gap represents an embedded option to add modules or storage later if electricity prices rise, clinical services extend into the evening or carbon-reduction targets tighten; this upside is invisible to the NPV but captured in the ROM.

Maintaining power quality is critical in healthcare applications, therefore a 15 kW medical-grade inverter (pure sine wave, 98% efficiency) was selected. Its modest headroom over the 12.4 kW of installed modules prevents saturation on clear days (and preserves the possibility of connecting an extra string without replacing power electronics, an expansion path valued in the ROM). The array comprises twenty-three monocrystalline panels rated at 540 W each (21.3% efficiency, 25-year warranty). Although these mid-to-high-end modules increase the upfront cost, they limit degradation and reduce downside risk, which enhances the option value. The complete bill of materials (Table 9) totals USD 13,740 and includes a 5 kWh lithium-ion battery reserved for critical circuits; while this storage has limited influence on the baseline NPV, it provides resilience benefits explicitly monetised in the ROM, reinforcing the hospital’s ability to manage grid disturbances and future regulatory changes [60].

Table 10 presents the parameters used in the financial evaluation of the photovoltaic project at the HSP under study. The evaluation considers installation, operation, and maintenance costs, as well as the uncertainty associated with key variables such as cash flow volatility and electricity prices.

The values in Table 10 establish the framework for applying the Binomial and Black-Scholes models in project valuation. Discount rates are typically considered to range from 8% to 12% in energy infrastructure in emerging markets [61]. For photovoltaic projects, cash flow volatility in solar investments generally varies between 30% and 40%, depending on electricity price fluctuations and weather conditions [62]. Additionally, the risk-free rate in energy investments is linked to the yield on long-term government bonds (10 years or more), which typically range from 4% to 6% in developing countries [63].

To calculate the NPV, the benefit of the investment in period t, eq. (4) was applied:

The cost incurred in period t is calculated using eq. (5):

The NPV over the 10 years of the project's life was calculated by applying eq. (3):

This result indicates that the present value of the expected benefits exceeds the initial investment, justifying the project's financial viability under traditional methods.

For the application of the Binomial model, the growth and decrease factors of the value of the underlying asset were determined by applying eqs. (8), (9), (10) and (11).

These values indicate that the project has a 50.1% probability of generating positive returns under uncertain scenarios, reinforcing the investment's strategic flexibility.

The Binomial tree is constructed from these values for the project's 10-year life, as shown in Figure 5. The tree is divided into three periods, starting from year 0. The first period evaluates the project at year 3.33, the second period at year 6.66, and the third and final period at year 10.

Figure 5 displays the three-stage binomial lattice used to project the photovoltaic system’s value over the ten-year horizon. Starting from the base NPV of USD 10,316 at t = 0, the up factor (u = 1.919) and down factor (d = 0.521) generate a dispersion that ranges from USD 72,901 in the best case to USD 1,459 in the worst. The risk-neutral probability of an up move is 0.501, so the distribution of terminal values is slightly skewed toward favourable outcomes. This spread illustrates how cash-flow volatility of 35.7% magnifies both upside and downside potential, information that a single deterministic NPV cannot convey.

The Black-Scholes model was applied to quantify the real option value of the photovoltaic project.

The parameters d₁ and d₂ were calculated using eqs. (14) and (15) as follows:

Figure 5.

Binomial tree corresponding to the 10 years of the useful life of the project

The value of the “call” option, according to the Black-Scholes model, was calculated using eq. (12) as follows:

For a put option, which grants the right to sell an underlying asset at a strike price, the valuation is calculated using eq. (13) as follows:

The values of N(d₁) and N(d₂) were obtained using statistical tables of the standard normal distribution [64]. The value of d₁ = 0.8421 indicates a high probability that the project will generate benefits when adjusted for market risk. In financial terms, this suggests that the project's future net cash flow is likely to exceed the initial investment, with a cumulative probability of approximately 80%, reinforcing the expectation of profitability. Conversely, d₂ = −0.2869 represents the risk that the investment may not achieve the expected returns. Although negative, its absolute value is three times smaller than d₁, indicating a moderate and controlled risk of loss.

Applying backward induction, the ROM assigns an option value of USD 5,335 for expansion (call) and USD 2,560 for abandonment or scale-back (put). Adding the call premium to the base NPV raises the project’s risk-adjusted worth to USD 15,651, a 52% improvement. In practical terms, management can commit to the current 12.4 kW array while retaining the right to install additional panels if electricity prices increase or if daytime demand grows beyond the present load curve (for example, through the addition of a new diagnostic wing). Conversely, the put option quantifies the benefit of being able to postpone or cancel further investment should regulatory changes or technology prices evolve unfavourably, thereby limiting downside exposure to roughly one quarter of the initial outlay.

Comparing both metrics clarifies the strategic landscape. The positive NPV confirms economic viability under expected conditions, but the ROM shows that ignoring flexibility undervalues the project by omitting real managerial choices. For hospital administrators who operate under flat tariffs and strict service-continuity constraints, the call option legitimises a phased investment plan (initial installation plus pre-sized inverter, with space for extra modules). Simultaneously, the put option supports a defensive stance if future maintenance costs rise or if new efficiency regulations reduce the price of grid electricity. Therefore, when evaluated through the ROM, the photovoltaic project aligns not only with present cost savings but also with long-term adaptability goals, providing a more robust foundation for capital-budget approval than the NPV alone.

A sensitivity analysis was conducted on four critical variables, each subjected to a 20% increase. In each case, the NPV was recalculated using eqs. (3) and (16), with the results presented in Table 11.

Table 11 confirms that electricity price and solar irradiance are the two dominant drivers of profitability. A 20% rise in either variable lifts the NPV by USD 5,151 (from USD 10,316 to USD 15,467). The corresponding elasticity of +2.50 means that every 1% increase in tariff or insolation raises project value 2.5%, emphasising how strongly revenue streams depend on external market and climatic conditions. From a ROM standpoint, these variables also enlarge the upside captured by the expansion option, since higher prices or better irradiance improve the payoff of adding extra module strings.

In contrast, a 20% escalation in the initial investment cuts the NPV to USD 7,920 (a loss of USD 2,396). Although the elasticity of −1.16 denotes moderate sensitivity, the indicator remains positive, suggesting that the project can tolerate reasonable cost overruns. Nevertheless, the put option valued in Stage 3 becomes more attractive under this circumstance, as managers could defer or scale back the installation if capital costs rise unexpectedly.

O&M expenses show the smallest influence. A 20% increase trims the NPV by only USD 310, bringing it to USD 10,006. The elasticity of −0.16 demonstrates that routine cost fluctuations have little bearing on viability, largely because O&M outlays account for a small share of total cash flow. This mild response underpins the project’s operational resilience and reduces the likelihood that the abandonment option would be exercised purely on the basis of higher maintenance costs.

The Tornado diagram in Figure 6 visualises these effects: electricity price and solar irradiance each add 50% to the NPV for a 20% rise, whereas initial investment subtracts 23% and O&M cost only 3%. The direction and length of the bars confirm that favourable shifts in market tariffs or solar resource enhance profitability, while higher capex and operating costs erode it. These findings support the strategic advantage of low operating costs, which not only enhance project resilience but also open a pathway for potential system expansion. Although the present design targets a 50% coverage due to spatial and budgetary constraints, the ROM captures the value of scaling up generation capacity in future scenarios with favourable economic or regulatory conditions.

Figure 6.

Tornado diagram

To translate single-factor insights into composite outcomes, three scenarios were constructed with the results presented in Table 12.

In the optimistic case (+20% on price and irradiance, −20% on capex and O&M) the NPV reaches USD 15,467, strengthening the rationale for exercising the expansion option and increasing system capacity. The baseline case maintains an NPV of USD 10,316, supporting the current design without further adjustments. The pessimistic scenario (−20% on price and irradiance, +20% on capex and O&M) lowers the NPV to USD 5,168, a level that remains positive but narrows the safety margin and underscores the value of the deferral option. Consequently, the combined sensitivity and scenario analysis corroborates the earlier ROM results: upside potential is governed primarily by electricity market conditions and solar availability, while downside risk is cushioned by managerial flexibility rather than small variations in operating costs.

Although the pessimistic scenario still yields a positive NPV, more severe downside conditions, such as a simultaneous and prolonged drop in both electricity prices and solar irradiation, could indeed push the project into a negative NPV territory. For instance, a drop of over 30% in both parameters would significantly reduce revenues while fixed costs remain constant, challenging financial viability. In such a case, mitigation measures become crucial. These may include deploying performance-based O&M contracts to reduce recurring costs, leveraging governmental incentives or carbon credits, incorporating adaptive energy management strategies, or scaling the system incrementally to limit upfront investment. Additionally, the ROM supports managerial flexibility by valuing the deferral or expansion options, which become even more relevant under uncertain or unfavourable conditions.

This stage integrates the results obtained in the previous phases to support strategic decision-making. The ROM allows comparing different investment strategies under uncertainty, considering not only their financial performance but also their flexibility and resilience. To this end, a multicriteria analysis was performed to determine the most suitable alternative for the hospital.

Several key factors were analysed to compare traditional financial valuation methods with ROM, including the consideration of uncertainty, flexibility in decision-making, impact on valuation, sensitivity analysis, resilience to market changes, and the possibility of multi-criteria evaluations. Table 14 presents a qualitative and quantitative comparison of both approaches, highlighting the differences in how each method addresses these criteria in the HSP under study.

This confirms that the ROM approach enhances decision-making by incorporating market uncertainties and strategic flexibility. It provides a dynamic assessment that considers potential future adjustments, ensuring a more resilient investment strategy for the photovoltaic project.

The proposed methodology presents several strengths. First, it integrates real options analysis with operational and regulatory constraints specific to the healthcare sector, offering a novel and replicable framework for evaluating energy projects in critical infrastructure. Second, it provides a probabilistic assessment under uncertainty, incorporating scenario analysis, elasticity estimation, and strategic options such as expansion and abandonment. This allows decision-makers to better capture the value of managerial flexibility and respond to market and environmental variability. Third, the application to a real-world hospital strengthens the practical relevance and transferability of the approach.

However, the study also has limitations. The analysis is based on a single case study, which may limit generalisability to other contexts without proper adaptation. In addition, the simulation assumes constant technical performance of the photovoltaic system over time, without modelling panel degradation or long-term maintenance effects. Finally, while the methodology is compatible with hybrid decision-support systems, this study does not yet integrate artificial intelligence or advanced predictive tools, which could enhance forecasting accuracy and responsiveness in future applications.