1. Introduction
Reducing energy consumption in the hospitality sector has become a strategic priority in the context of the ongoing energy crisis, rising tourism demand, and increasingly strict climate commitments. Despite advancements in HVAC efficiency and the enforcement of sustainable building regulations [1,2], most energy management strategies in hotels remain generalized and fail to account for the specific thermal behavior of individual rooms [3,4]. This lack of granularity limits the potential for effective operational optimization, particularly in high-rise hotel buildings with significant vertical room distribution [5].
One promising approach is the energy-efficient allocation of guests, which has been explored mainly through dynamic pricing strategies or adaptive comfort models [6,7]. However, the scientific literature reveals a critical gap: few studies integrate physics-based building energy models (BEMs) into real-world guest allocation strategies. Most related works either focus on surface-level photovoltaic generation [8,9] or on aggregated HVAC performance simulations, without addressing room-level thermal behavior influenced by geometry, orientation, and floor height [10,11].
This study introduces a novel methodology based on a high-resolution simulation of the annual energy demand of each room in a real 17-story hotel tower. By leveraging EnergyPlus and Radiance, and using real climate data from Madrid [12], we simulate the impact of solar heat gains and building geometry on thermal loads. Unlike typical energy modeling approaches, our analysis remains intentionally relative, focusing on comparative energy demand between rooms rather than absolute consumption, thereby bypassing the need for detailed HVAC system calibration or sub-metered energy data.
The primary goal of this work is to theoretically quantify the potential energy savings enabled by guest allocation strategies informed by detailed BEM simulations. We demonstrate that even without retrofitting or technological upgrades, significant energy savings can be achieved simply by optimizing the room assignment process—especially under partial occupancy scenarios, where our results show the highest impact. This research provides a decision-support tool for hotel operators and paves the way for future real-world implementations.
Literature Review
The building sector is responsible for nearly 40% of global energy consumption and around 30% of CO2 emissions [1,2], making it a critical area for climate action. Within this context, the hospitality sector presents unique energy challenges due to its round-the-clock operation, seasonal variability, and heterogeneity in energy usage across spaces [3,4]. These conditions make hotels strong candidates for operational optimization through simulation-informed decision-making.
Building energy models (BEMs) have become essential tools in building performance analysis. They enable thermal simulations under different design and operational conditions, accounting for architectural features, materials, orientation, and usage schedules [10]. Recent advancements in geometric modeling and the incorporation of urban context—such as surrounding buildings, solar obstruction, and surface reflectivity—have increased the reliability of such simulations in dense environments [5,11]. Moreover, surrogate modeling and machine learning techniques have been introduced to enhance computational efficiency while maintaining accuracy, particularly in large and complex facilities such as hotels [13,14]. Surrogate models approximate high-fidelity simulations with simplified, data-driven alternatives, enabling rapid evaluations in tasks such as design optimization, sensitivity analysis, and control strategy development [15,16].
Many of these models have traditionally focused on photovoltaic (PV) performance analysis. Studies involving Building-Integrated Photovoltaics (BIPV) rely on 3D modeling and tools such as EnergyPlus and Radiance to simulate solar irradiance on façades and rooftops [8,9,17]. However, the emphasis is typically on estimating PV yield rather than assessing how solar irradiance translates into internal heat gains and HVAC loads. This gap has been addressed through recent advances in deep learning, reinforcement learning, and real-time predictive control systems that adapt HVAC operations based on thermal predictions and occupant presence [18,19,20].
This distinction is crucial in the context of hotels, where thermal heterogeneity between rooms—caused by floor height, façade orientation, and shading—directly affects HVAC energy demand. Yet, few studies focus on room-level solar heat gain as a driver of energy consumption, especially in operational strategies like guest allocation [12]. Emerging research, however, suggests that integrating occupancy modeling and predictive simulations can optimize energy efficiency by dynamically adjusting to guest behavior and thermal loads at the room level [21,22,23,24]. In contrast, our work estimates these gains as a function of space and time, enabling room assignment based on simulated thermal behavior rather than aggregated metrics.
Energy consumption in hotels is typically dominated by HVAC systems, especially during peak seasons [25]. In response, the literature has explored occupancy-based HVAC control, dynamic temperature setpoints, and zone-level management strategies to reduce energy use without compromising comfort [7,26,27]. Still, these strategies often assume uniform thermal conditions or ignore geometric variability between rooms.
Some studies have examined guest allocation models based on pricing strategies or booking optimization, but they rarely incorporate building physics or simulate energy behavior with spatial resolution [6], which limits their practical energy impact. In contrast, simulation-informed room allocation—taking into account solar heat gain, thermal mass, and room geometry–—remains largely absent from the literature.
Moreover, emerging work in indoor air quality (IAQ) monitoring and occupancy sensing, often using CO2 levels or PM concentration, provides additional tools to understand room usage patterns and population density in near real time. These techniques are increasingly linked to HVAC performance and could complement room-level energy simulations in future work.
In summary, there is a notable gap in the integration of BEM-based simulations with operational guest allocation strategies in hotels. The present study addresses this by developing a thermal simulation framework that enables energy-aware room assignment. It quantifies solar heat gains and resulting HVAC demand at room scale, offering a novel bridge between building physics and day-to-day hotel operation.
2. Methods
This study involves several key developments focused on simulating the thermal and energy performance of the selected building. It proposes an improved hotel guest room allocation strategy aimed at minimizing HVAC-related energy consumption. The methodological workflow–—summarized in Figure 1—includes detailed three-dimensional (3D) modeling of the building’s exterior, interior, and surrounding urban environment; the definition of HVAC operation assumptions and construction materials; and the characterization of local climate and solar trajectories. The results were validated phenomenologically in terms of physical correctness, evaluating in detail the expected solar gain curves, and were compared with the actual total hotel energy consumption. A realistic energy-saving potential has been calculated based on real hotel occupancy data. Each of these methodological aspects will be discussed individually in this section.
2.1. Case Study
The case-study hotel tower, the Eurobuilding operated by MINOR HOTELS EUROPE & AMERICAS, located in Madrid, Spain, was selected due to its relatively high installed thermal capacity and its potential for room allocation management interventions. It is a 17-floor building whose construction began in the 1960s and has since undergone numerous renovations to update its structure and installations, many aimed at improving energy efficiency.
The building has two floors at its base dedicated to event rooms and restaurants, with a total of 33 venues and 4 restaurants between the two floors. In the case of the guest rooms, there are a total of 412 single rooms in addition to some special double rooms and suites on the last floor. This case study has been limited to the modeling and efficient management of the 412 guest rooms spread over 15 floors.
2.2. 3D Modeling
2.2.1. Modeling of the Surrounding Built Environment
The 3D model of the hotel built environment was generated with the aim of capturing the shadows and reflections produced by the built environment on the hotel tower. Respective data were extracted from public sources of information such as public satellite images (Google Maps and Google Earth (Google LLC, Mountain View, CA, USA),
Satellite images were extracted from the renderings of Google Earth, from which a triangulated mesh was extracted using RenderDoc (RenderDoc (Baldur Karlsson, Open Source),
The triangulated mesh itself does not represent a solid object for 3D modeling software but a set of edges that describe the silhouette of the buildings observed by satellite imagery. Therefore, once the model was imported into SketchUp, the triangulated mesh served as the basis for generating the final 3D objects of the surrounding buildings. These objects, unlike the detailed model of the hotel building, were modeled as cubes with proportions equal to the building silhouettes but with flat surfaces, without detailed exterior or interior modeling as only the projections of shadows and reflections of these buildings were of interest. This can be seen in Figure 4, where the aforementioned cubes are shown superimposed on the original triangulated mesh.
2.2.2. Detailed 3D Modeling of the Hotel Tower
The hotel’s exterior (or facade) was modeled manually in OpenStudio v3.7.0, based on the architectural Computer Aided Design (CAD) plans of the tower as provided by the hotel operator (see Figure 5). OpenStudio allows for directly exporting the created models such that they can be read in by EnergyPlus for thermal modeling, see below.
In a first step, the facade of the hotel tower was modeled in high detail, including all balconies, etc., as shown in Figure 6. However, inclusion of corner balconies turned out to lead to incorrect energy calculations as OpenStudio was unable to generate properly enclosed spaces for the corner rooms in this case. Hence, corner balconies had to be removed from the final model.
To take this simplification into account when calculating corner room energy consumption, the energy values of corner rooms with an upper balcony were corrected numerically by decreasing their cooling thermal energy by 4.4% and increasing the heating thermal energy by the same 4.4% in line with studies found in the literature [28]. The final 3D model of the hotel’s exterior, without modeling the corner balconies, is shown in Figure 7.
The modeling of the interior of the building and its floors was carried out on the basis of the architectural plans, respecting the actual dimensions (see Table 1) and orientation. The layout of the floor plan as shown was replicated on 14 floors, while the last floor 15—containing the significantly larger suites—was modeled accordingly.
The base floor, which houses event rooms, the entrance, and restaurants—located below floor 1—was simplified as a single thermal zone (see Figure 8) as it is not the primary focus of this study. On the other hand, an ”empty” floor was included above the top floor (floor 15), representing additional storage and technical spaces. Architectural details of this upper area were not available, and it was, therefore, modeled as a simplified unoccupied volume. This decision was made to ensure realistic boundary conditions for the top occupied floor. Without this buffer, the simulation of floor 15 would assume an exposed roof boundary, leading to exaggerated thermal losses and solar gains. By including an unconditioned, low-interaction volume above, the simulation preserves more realistic heat transfer dynamics. The influence of this empty floor on the results is minor as it remains unoccupied and thermally passive. In total, 439 individual thermal zones were modeled to represent the full hotel tower interior.
Table 1 gives an overview of the key geometric parameters of the modeled room types. Note that all suites are located on the 15th floor.
2.3. Building Thermal Behavior Modeling
2.3.1. Simulation Environment
The radiation and energy consumption simulations were carried out using EnergyPlus (EnergyPlus (U.S. Department of Energy—NREL, Golden, CO, USA),
2.3.2. Materials
The thermal simulations conducted in EnergyPlus require the specification of material properties for all building envelope components, including walls, floors, ceilings, and windows. In the absence of detailed information on the actual construction materials used in the hotel—such as layer composition, thicknesses, and thermal conductivities—we adopted predefined material templates available in OpenStudio. These templates reflect typical construction assemblies for commercial buildings and include standard configurations for external and internal walls, slabs, roofs, and glazing systems.
Table 2 summarizes the material properties assigned to the hotel model. While this approach does not reflect the exact construction of the case-study building, it is consistent with the overall modeling strategy of this study. Since the focus lies on the relative differences in thermal energy demand between rooms, rather than the absolute prediction of building-wide energy consumption, the use of standardized materials ensures that the comparative analysis remains valid. All rooms were modeled with identical material configurations and thermal assumptions, allowing us to isolate the effects of room orientation, geometry, and solar exposure on heating and cooling loads.
Although the materials listed in Table 2 are representative of typical hotel constructions, they were not tailored to the actual materials used in the studied building. This limitation introduces uncertainty in the absolute thermal behavior of the modeled rooms. However, as the study focuses on relative differences in HVAC-related energy demand across rooms within the same building–—assuming construction uniformity across floors–—the comparative validity of the results is preserved. Future work with access to as-built construction specifications could help refine the absolute energy estimates.
2.3.3. Thermal Control Modeling
Typically, thermal zones in building simulations encompass broad areas, integrating multiple rooms served by a common HVAC system. This simplifies the modeling of HVAC systems and, consequently, the simulation itself. However, for this study, individual results from each guest room are of particular interest, with the goal of comparing results based on different room configurations, such as orientation and floor level. Therefore, the resolution of the thermal zones was increased by assigning a thermal zone to each individual guest room and suite, resulting in a total of 439 thermal zones to simulate.
Regarding operational modes, two different modes were applied, replicating the control system of individual room HVAC units. Specifically, the influence of room occupancy was reflected in the simulations by assuming “comfort” mode for occupied rooms (using 22/23 degrees Celsius setpoints for winter and summer months, respectively) and “economy” mode (using 20/25 degrees Celsius setpoints for winter and summer months, respectively) for unoccupied rooms. The selected setpoints reflect the actual hotel’s control strategy.
2.3.4. HVAC Modeling
Accurately determining the real thermal behavior of the hotel building was not feasible in this study. Since the building is in continuous operation, all rooms and floors are actively air-conditioned year-round, making it impossible to observe the natural evolution of indoor temperatures under unconditioned conditions. Furthermore, no detailed technical documentation on the hotel’s HVAC system—such as control logic, efficiencies, or equipment specifications—was available. Due to these constraints, we adopted the “Ideal Loads Air System” (ILAS) from EnergyPlus to estimate the thermal energy required to maintain the comfort setpoints in each room.
The ILAS is a well-established simplification in early-stage or comparative energy modeling. Instead of simulating the full HVAC system–—including chillers, boilers, fans, ducts, and control delays–—it delivers the exact amount of heating or cooling needed to meet zone setpoints at every timestep, with perfect efficiency and no energy losses [29]. This approach eliminates mechanical system complexity and allows the model to compute only the theoretical thermal energy demand needed for comfort, independent of the specific HVAC installation.
In our context, this simplification was not only necessary but methodologically appropriate. The goal of the study is not to predict absolute energy consumption but to perform relative comparisons between rooms based on their location, geometry, and solar exposure. Since the ILAS model applies the same assumptions to all zones, it provides a consistent and neutral baseline. Differences in simulated energy demand between rooms can thus be attributed solely to physical and environmental factors—such as orientation, floor height, or solar gain–—and not to variations in HVAC performance or equipment.
This modeling strategy aligns with other studies that prioritize thermal zoning analysis or conceptual energy planning. While it omits real-world inefficiencies, it is particularly suitable when assessing spatial variation in thermal loads across a building with uniform construction and control assumptions. Future work may incorporate more detailed HVAC system modeling or calibration against sub-metered data to estimate absolute consumption. However, for the scope of this study–—quantifying potential energy savings from guest allocation strategies—the ILAS-based approach provides both robustness and clarity.
2.3.5. Weather Modeling
To be able to correctly estimate HVAC-related energy consumptions, it is necessary to replicate the meteorological conditions at the exact location of the hotel under study. In this way, key parameters such as external temperature and humidity, etc., can be taken into account. Some valuable information for the simulation was extracted from the hotel’s location, such as the relative position of the sun during each day of the year, while all the relevant meteorological variables were gathered in an EPW file, which is the format used by EnergyPlus for weather data. This file format can contain either monitored weather data or typical data based on the typical meteorological year (TMY) format. TMY files contain typical meteorological data derived from monitored data over a specific period, usually following the Sandia methodology developed by Hall et al. in 1978 [30].
In the present project, public data sources were used, specifically those provided by the Climate.OneBuilding organization, which developed their own format based on the Sandia method, called TMYx. These files contain typical data derived from hourly samples, in the case of this study, for the period 2009–2023 at station 082200 in Madrid.
Furthermore, the EnergyPlus Radiance module allows one to simulate the path of the sun at the exact location of the hotel (Figure 9). In this way—and taking into account the shadows generated by the surrounding buildings (see Figure 4)—it was possible to simulate within EnergyPlus how much solar radiation reaches the different parts of the hotel’s facade for each hour of a typical year. Finally, the resulting internal solar heat gains for each thermal zone (i.e., for each room) could be calculated, contributing to the heating of the room.
2.3.6. Additional Available Data
The hotel’s management kindly provided the following additional datasets to complement our study. On the one hand, they provided the hotel’s energy consumption data consisting of the hotel’s total (aggregated) electricity usage records for the period from 1 January 2020 to 7 May 2024 as provided by the hotel’s energy provider. The data were collected at 15-min intervals, allowing for an hourly, daily, or monthly analysis.
On the other hand, the hotel provided historical hotel occupancy data from 2020 to May 2024, recorded as daily number of sold rooms. Note that no information was provided regarding which exact guest rooms had been sold.
2.3.7. Occupancy-Dependent HVAC Energy Consumption Calculation
Using the 3D models of the hotel’s surrounding built environment, the hotel’s facade and interior, and applying the defined material parameters, ILAS air-conditioning with the defined “comfort” and “economy” settings and the weather data, simulations were carried out in EnergyPlus including the Radiance package, simulating an entire TMY in hourly time steps.
To estimate differences in energy consumption between occupied and unoccupied scenarios for each hour of the year, two simulations were conducted: “comfort” mode was applied to all rooms and suites to simulate a fully occupied hotel, while in a second simulation, “economy” mode was applied to all rooms to simulate an unoccupied but still air-conditioned hotel. In this way, for each room, the required (thermal) energy was calculated to maintain the room either in “comfort” or in “economy” temperature conditions, depending on whether it was assumed to be occupied or not. We assumed no significant thermal energy exchange between “comfort” or “economy” air-conditioned rooms (due to their setpoints being rather similar). The total HVAC-related hotel energy consumption was, therefore, calculated as the sum of the individual rooms’ consumption.
In this way, a large range of occupancy scenarios and their respective HVAC-related energy consumption could be replicated without running individual simulations for each allocation configuration. For example, for a 70% occupancy rate, 70% of the rooms would be assigned to run in “comfort” mode, while the remaining 30% would run in “economy” mode. Which rooms would run in “economy” mode could be assigned randomly or following our strategic allocation recommendations based on the lowest estimated energy consumption for the specific day (for example, see sections below).
2.4. Simulation Results Validation
To validate the energy simulations, a physical validation approach was adopted in place of numerical validation. This choice was driven by two constraints: (1) the ILAS model provides only thermal energy demand, which does not directly reflect the electric consumption of HVAC systems (see Section 2.3.4), and (2) disaggregated electricity data for the HVAC systems serving individual hotel rooms was unavailable—only total building electricity consumption was accessible. Consequently, no reference data existed for direct numerical comparison.
The physical validation focused on differential factors across thermal zones (i.e., individual hotel rooms) to assess whether patterns of solar heat gains and thermal energy demand were physically plausible. This included analyzing variations by room orientation (E: east, NE: northeast, NW: northwest, SE: southeast, SW: southwest, and W: west) and floor level, considering both time of day and seasonal effects. Most rooms are oriented east or west, aligned with the building’s long facades, while corner rooms face diagonally; no rooms face due north or south.
Due to the hotel’s continuous operation during the study, detailed calibration of the energy model was not feasible. Additionally, information on material properties, HVAC system specifications, and disaggregated energy use was not available. As a result, the goal was not to achieve precise absolute, calibrated calculations of HVAC electricity use, but rather to analyze relative differences in thermal demand across rooms and allocation scenarios. Simulations assumed uniform construction and control logic for all rooms, attributing differences in HVAC-related energy demand solely to geometric factors such as room size, glazing area, orientation, and floor level.
2.4.1. Solar Heat Gain
The key objective of the study was to optimize room allocation, assuming all rooms share identical building materials, thermal installations, and control systems. Under these conditions, solar heat gains were considered the dominant factor influencing thermal behavior. This is because the simulation assumes uniform outdoor temperature effects across all rooms, regardless of orientation or floor level.
Therefore, simulation results were analyzed by classifying the rooms by orientation (E: east, NE: northeast, NW: northwest, SE: southeast, SW: southwest, and W: west), by floor of the building, by the time of day, and by the season of the year. The aim was to check that the simulation results are consistent with the real orientation of the building so that, for example, during morning hours the rooms on the eastern side receive higher solar gains than the rest of the rooms, while western rooms receive more solar gains in the afternoons.
2.4.2. External and Internal Air Convection
With regard to the convection of external air and its influence on the indoor temperature, it was expected that all rooms on the same floor are equally affected by external convection, regardless of their orientation. However, internal convection should affect the upper and lower floors of each floor, resulting in higher temperatures on the upper floors of the building.
2.4.3. HVAC Consumption
The function of the HVAC system in the building is to adapt the indoor environmental conditions to acceptable comfort conditions. In this sense, the consumption of these systems must be correlated with the thermal contribution from the outside, either by the convection of the outside air or by the solar gain mentioned above.
A visual validation was performed to analyze the relevant physical parameters using the Python (v3.12) libraries Matplotlib (v3.8.2) and Seaborn (v0.13.2). Note that we focused here on analyzing energy consumption ratios rather than absolute values, see Section 2.4. The ratio analysis focused on relative energy consumption, aiming to identify which thermal zones (i.e., hotel rooms) exhibit comparatively higher or lower HVAC energy use. These ratios were calculated and visualized by floor, orientation, room type, and by day or hour.
2.4.4. Simulated vs. Real HVAC Consumption
As discussed in Section 2.4, numerical validation was not feasible due to the lack of disaggregated HVAC energy consumption data and the limitations of the ILAS tool, which calculates thermal demand rather than actual electrical consumption. Only the hotel’s total electricity use was available, preventing direct numerical comparison.
In order to compare the simulated thermal energy demand with the real electricity consumption of the hotel, an estimation of the share of total electricity attributed to the HVAC system was required. Although disaggregated metered data were not available, a confidential energy audit provided by the hotel operator indicated that approximately 15% of the building’s annual electricity consumption could be attributed to HVAC operation. This figure was derived from internal assessments carried out by the hotel’s energy service provider based on aggregated billing data, building usage profiles, and HVAC system characteristics.
While this value cannot be independently verified due to the confidential nature of the source, it offers a reasonable and context-specific approximation consistent with typical HVAC shares reported in the literature for hotel buildings of similar size and operation patterns. Based on this estimate, the total annual electricity consumption curve was scaled by a factor of 0.15 to obtain an indicative baseline for HVAC-related electricity use.
In addition, our simulations yielded the daily energy consumption assuming that the hotel was fully occupied (100% occupancy, see Section 2.3.7). Therefore, the simulated (thermal) energy consumption was scaled by the daily occupancy rate to obtain an occupancy-adjusted daily consumption estimation. Finally, the simulated thermal energy consumption was divided by a Coefficient of Performance (COP) of 3.5 to convert it into estimated electrical energy. This value corresponds to a typical COP for air-source HVAC systems operating under moderate load conditions in commercial buildings [31]. Both consumption curves—actual vs. simulated—were analyzed visually and numerically.
2.5. Optimal Guest Allocation
The guest allocation strategy developed in this study builds on year-long, high-resolution energy simulations of the hotel under full occupancy conditions (see Section 2.3.7). For each of the 439 modeled rooms, daily estimates of heating and cooling thermal energy were calculated to maintain standard comfort conditions. These simulations form the foundation of an energy-based guest assignment system.
The simulation results were compiled into a structured lookup table that, for each day of the year, includes the heating and cooling energy demand for every individual room, as well as a daily ranking that orders rooms from lowest to highest HVAC-related energy consumption. This ranking captures the influence of multiple factors, including room orientation, floor level, shading from surrounding buildings, window size, and seasonal weather variability. As such, it enables the identification of the most and least energy-efficient rooms for any given day.
To quantify the potential energy savings of the proposed guest allocation strategy, we relied on the simulation results obtained under full hotel occupancy, which allowed us to compute, for each day of the year, a ranking of all rooms according to their estimated HVAC-related thermal energy demand. These daily rankings form the basis of the optimal allocation approach. Using them, we evaluated the saving potential in two ways. First, we simulated a range of hypothetical occupancy levels by varying the number of occupied rooms, N, from 1 to 411. For each occupancy level, we compared the total annual HVAC energy demand of two extreme scenarios: one where the N lowest-consuming rooms were assigned to guests (best case) and another where the N highest-consuming rooms were occupied (worst case). Second, to assess the real-world applicability of the strategy, we used actual hotel occupancy data from 2023 to assign guests, day by day, to the most energy-efficient rooms available according to the precomputed daily rankings. The resulting energy demand was then compared to the worst-case assignment for the same occupancy figures, allowing us to estimate the annual savings that could have been achieved by applying the strategy under real operational conditions.
Although the method relies on detailed simulation data, it can also inform simplified operational guidelines. For example, during summer periods, rooms located on lower floors with west-facing windows typically require less cooling, while in winter, rooms on higher floors facing southeast tend to be more efficient for heating. These trends, embedded in the ranking process, provide a practical basis for energy-conscious room assignment that balances technical insights with real-world applicability.
3. Results
3.1. Validation
Due to the absence of detailed sub-metered data for the HVAC system, a direct numerical validation of simulated energy demands against actual consumption was not possible. Instead, a phenomenological validation approach was adopted to assess the physical plausibility of the simulation results. This included analyzing whether the thermal behavior of the building followed expected patterns related to solar orientation, floor height, room type, and seasonal variation. We examined hourly and monthly demand curves aggregated by room, orientation, and floor and compared them to the overall electricity consumption profile of the hotel. This type of validation, although indirect, provides robust insight into whether the simulation responds correctly to key physical drivers of energy demand. Such approaches are commonly used in similar studies where empirical data are unavailable and help ensure that the model reflects realistic thermal dynamics.
3.1.1. Solar Radiation
Figure 10 shows a series of graphs representing the solar gain throughout the day (on the horizontal axis) in different seasons (winter, spring, summer, and autumn) and on different floors of the hotel (from 1 to 15 with a step of 2), with different orientations (E, NE, NW, SE, SW, and W).
Each column of the figure corresponds to a season of the year, from left to right: winter, spring, summer, and autumn. Each row of the figure corresponds to a floor of the hotel. As for the X-axis of each graph, it represents the time of day, while the Y-axis represents the solar heat gain in Watts, the line being the mean of the distribution and the band representing the deviation. Finally, the following color code has been used to represent the orientations: blue for east, orange for northeast, green for northwest, red for southeast, purple for southwest, and brown for west.
The simulated solar gain was generally higher during spring and summer and lower in winter and autumn. As one moves up the floor, the solar gain increases, which was most noticeable on the top floor. The highest peak solar gain occurred in the northeast and southeast orientations in the early hours of the day, while the northwest and southwest orientations reached a lower peak but later in the day. Note that the position of the hotel is slightly tilted towards the east; hence, rooms in the northeastern corner receive a lot of sunlight in the mornings.
In addition, east and west orientations show significantly less solar gain than the corners as the east and west room sizes are smaller than the corner rooms and have less window area. Figure 11 shows eight graphs representing the accumulated solar heat gain per month on different floors of the hotel (1, 3, 5, 7, 9, 11, 13, and 15) and with different orientations (E, NE, NW, SE, SW, and W). The first row of the figure shows floors 1, 3, 5, and 7, while the second row shows floors 9, 11, 13, and 15. The X-axis represents the month of the year from January to December, while the Y-axis indicates the accumulated solar heat gain of the month in Watts. Finally, the colors show the different orientations of the hotel.
The cumulative solar heat gain per month was higher in the middle months of the year (May, June, July, and August) than the rest (January, February, November, and December). Further, the cumulative solar heat gain per month increased with the floor level. In the case of intermediate floors, the increase was moderate, while the jump occurring at the top floor was much larger. The cumulative solar heat gain per month was generally highest in the northeast and southeast orientations, followed closely by the northeast and southwest orientations, and lastly by the eastern and western-facing rooms.
The lowest solar heat gain throughout the year is seen in the west orientation, probably due to the eastward tilt of the hotel and the generation of shadows from the surroundings; as seen in Figure 10, rooms in the west tend to receive relatively low solar radiation and only at the end of the day. On the lower floors (1, 3, and 5), it can be seen how the southeast and northeast orientations have similar cumulative solar gain values, probably due to the effect of other buildings projecting shadows onto these lower floors.
Figure 12 shows the distribution of the accumulated solar gain per month with the difference that in this case different orientations are shown, having in the first row the northeastern, eastern, and southeastern orientation. In the second row, the northwestern, western, and southwestern orientations are shown, while the colors are used to represent the floors. The observations in the figure are identical to those in the previous figure, except that the impact of the floor height on the accumulated solar gain can be seen more clearly.
From these results, we concluded that the solar path simulation performed in EnergyPlus yielded physically meaningful results that take into account the building orientation and shadows generated by surrounding buildings.
3.1.2. Cooling Energy Demand
Figure 13 shows the hourly distribution of cooling thermal energy as a function of orientation and time of year. Each graph represents each orientation of the hotel rooms, with the first row having northeastern, eastern, and southeastern orientations, while the second row has northwestern, western, and southwestern orientations. The X-axis represents the time of day, while the Y-axis indicates the simulated cooling thermal energy demand in Watts. Finally, the color shows the season of the year: blue for winter, orange for spring, green for summer, and red for autumn.
The cooling thermal energy required was higher in summer and spring seasons, while it remained low in autumn and winter. Values peaked in the early morning hours in the northeast, east, and southeast-facing rooms, while for the northwest, west, and southwest-facing rooms, the peak occurred later in the day. This coincides with the solar heat gain.
It can be seen how the thermal energy demand was higher in the corner rooms (northeast, southeast, northwest, and southwest), especially in the eastern and western corners due to their larger size, larger window area, and receiving more solar exposure time and, thus, solar heat gain.
Figure 14 represents the distribution of the accumulated thermal energy of the month to maintain the temperature of the rooms within the “comfort” profile and shows the impact of orientation and floor height. The first row of Figure 14 shows northeastern, eastern, and southeastern orientations, while the second row shows the northwest, west, and southwest. The X-axis represents the time of day, while the Y-axis indicates the accumulated cooling thermal energy demand in Watts. Finally, the colors represent the floor of the room.
In the central months of the year, the cooling thermal energy demand was higher than in the remaining months of the year. As in the previous case, the corner-related orientations had higher energy needs than the eastern and western orientations because of their larger size and window area and because they receive more solar heat gain. As the floor height increases, the cooling thermal energy demand also increases, especially on the top floor where the suites are located. The variation in thermal energy demand with varying floor heights (removing the 15th floor) was much smaller in the northwestern and northeastern orientations than in the other orientations.
Figure 15 shows similar results with the difference that each graph represents different floors, with floors 1, 3, 5, and 7 as the first row and floors 9, 11, 13, and 15 as the second row, while the color changes to represent the orientation of the room. The observations are identical to the previous case, but the figure serves to see more clearly the effect of the orientation on the thermal energy required for monthly cooling.
3.1.3. Heating Energy Demand
Figure 16 shows the distribution of the cumulative heating energy per month required to maintain the room temperature in the “comfort” profile as well as the effect of floor height and orientation. The first row shows results for northeastern, eastern, and southeastern orientations, while the second row represents northwest, west, and southwest. The X-axis represents the month of the year from January to December, while the Y-axis indicates the accumulated thermal energy required for heating of the month in Watts. Finally, the floor height is represented by the color.
In the central months of the year, the heating thermal energy was zero, while the peaks were mainly in the months of January and December (winter). In general, the northeastern, eastern, and southeastern orientations required less thermal heating energy than the northwest, west, and southwest, and, similarly, the northeastern and northwestern orientations required more energy than the southeast and southwest, which coincides with the solar heat gain.
As for the effect of floor height, the higher the floor, the less heating thermal energy was required, although the effect was very small, except for the top floor, which has the largest energy needs (probably due to the suites being located there).
The results for cooling and heating energy demand coincided with those for the solar heat gain and make sense from a physical point of view, assuming that the summer months require more cooling, as do the (larger) corner rooms and suites. Typically, cooling demand exceeded heating demand, which was confirmed by the hotel management.
3.1.4. Influence of Weather Conditions
Based on the simulated energy data and the weather-related EPW input file used for the simulations, the Pearson correlation between all variables was calculated: total thermal energy (cooling and heating), solar heat gain, air temperature, dew point temperature, relative humidity, atmospheric pressure, wind speed, etc. The resulting triangular correlation matrix is presented in Figure 17, with the first column displaying, both numerically and via color scale, the correlation of each variable with total thermal energy demand.
There was a strong positive correlation between the total thermal energy, solar heat gain, and outside air temperature, reaching values of 0.68 and 0.63 Pearson correlation, respectively. The positive correlations are physically consistent with heat transfer mechanisms, particularly solar radiation and convective exchange with the outdoor air. While negative correlations are generally less significant in this context, relative humidity and cloud cover exhibited notable negative correlations with total thermal energy demand, with coefficients of –0.56 and –0.52, respectively.
3.1.5. Simulated vs. Actual Energy Consumption
Figure 18 shows the curves of real and simulated electricity consumption, plotted at hourly, daily, and monthly resolutions. These curves are based on several assumptions regarding the hotel’s HVAC electricity use and the conversion of simulated thermal loads into estimated electrical consumption in kilowatt-hours (kWh); see Section 2.4.4 for details.
Figure 19 presents the real hotel occupancy rate during 2023. In general, the real and simulated consumption curves follow similar temporal patterns, primarily influenced by occupancy. For example, both curves show a marked decrease in consumption during August, when the hotel operates at minimal occupancy. In winter months, the real consumption curve exhibits higher peaks as it includes the total electricity use of the hotel—covering HVAC, lighting, appliances, elevators, and other systems. In contrast, the simulated curve represents only the thermal loads estimated by EnergyPlus, converted to electricity using a fixed Coefficient of Performance (COP = 3.5).
As described in Section 2.4, direct numerical validation was not feasible due to the lack of disaggregated HVAC electricity consumption data. To enable comparison, the hotel’s total recorded electricity consumption was scaled by 0.15 based on an internal energy audit provided by the hotel operator. This audit, although confidential and not independently verifiable, indicated that approximately 15% of the total annual electricity use could be attributed to HVAC systems—an estimate consistent with values reported in the literature for similar hotel buildings.
Furthermore, the simulation assumed full occupancy throughout the year. To reflect the actual operational context, the simulated thermal energy was scaled by the recorded daily occupancy rate and converted to estimated electricity using the assumed COP of 3.5. The resulting simulated curve thus represents an adjusted estimation of HVAC-related electricity use under real occupancy conditions.
Following this approach, both curves were analyzed for temporal consistency. Although the simulated values are approximately eight times lower in absolute terms, the temporal correlation between both curves is high. Pearson correlation coefficients of 0.605 (hourly) and 0.710 (daily), both statistically significant (p < 0.05), confirm that the simulated model captures the main temporal dynamics of energy demand. This supports its validity for assessing relative variations and for evaluating allocation strategies aimed at reducing HVAC-related consumption.
3.1.6. Summary of Validation Results
In Table 3, a summary of the validation results of the simulations is given, organized by variables used for the analysis:
3.2. Analysis of Consumption Patterns and Energy-Saving Potentials
Based on the simulated data, this section focuses on assessing the overall impact of the floor height, room orientation, room type, and individual room selection on the yearly HVAC-related energy consumption. Furthermore, we use the real occupancy data over the year to obtain a more realistic estimate of the energy-saving potential.
3.2.1. Thermal Energy Demand Related to Floor Height
In Figure 20, the cumulative cooling (left) and heating (right) thermal energy consumption per floor over the year is shown. The X-axis represents the hotel floors, while the Y-axis shows the ratio of each floor’s cumulative annual thermal energy consumption to that of the most energy-efficient floor (i.e., the one with the lowest annual consumption). By definition, the most-efficient floor has a Y-axis value of 1, and all other floors have values greater than 1. Additionally, the bold number above each bar indicates the percentage by which a floor’s annual consumption exceeds the minimum. Note that all types of rooms and orientations are aggregated here per floor.
The ground floor had the lowest annual thermal cooling demand, with energy requirements increasing progressively up to the eighth floor, where they stabilized. This indicates that during summer, lower floors—particularly up to the eighth—are more energy-efficient for cooling. The difference between the most and least-efficient floors was substantial, with the top floor consuming 106% more cooling energy than the ground floor—slightly more than double.
In contrast, heating demand showed the opposite trend. It gradually decreased with height up to the eighth floor and then increased again on the top floor. This suggests that in winter, the upper floors—from the eighth floor upwards, excluding the top floor hosting the suites—are preferable for minimizing heating energy consumption. The disparity in heating demand was also significant, with the least-efficient floor consuming 204% more than the most-efficient one—approximately three times as much.
3.2.2. Thermal Energy Demand Related to Orientation
Figure 21 is analogous to the previous one with the difference that instead of showing the cumulative annual thermal energy ratio with respect to the best option per floor, it shows ratios related to the energetically best orientation (with the most-efficient orientation having value 1). Note that all floors and types of rooms are aggregated here per orientation.
For cooling, east-facing orientations—particularly southeast and northeast—exhibited higher annual thermal energy demand compared to their west-facing counterparts (southwest and northwest), with differences reaching approximately 40%. This is largely due to greater solar exposure during the morning hours. Additionally, rooms facing north, northeast, and northwest tended to require slightly less cooling than those facing south, southeast, or southwest.
Rooms oriented directly east or west show the lowest cooling demand overall, primarily due to their smaller size and reduced solar gain. Among these, west-facing rooms generally have the lowest energy requirements.
For heating, the highest thermal energy demand was observed in northwest-facing rooms, followed by southwest and northeast orientations, which show similar levels of consumption. Southeast rooms follow, while east- and west-facing rooms again show the lowest heating demands.
3.2.3. Thermal Energy Demand Related to Room Type
Similarly, Figure 22 shows the annual cumulative energy ratio for cooling (left) and heating (right) according to the type of room: normal, corner, or suite. The most-efficient room type has a ratio of 1. Note that all floors and orientations are aggregated here per type of room.
From lowest to highest cumulative annual cooling consumption, the corner rooms and suites consumed between 106% and 256% more energy than normal rooms. In the case of heating consumption, the corner rooms and suites consumed between 194% and 719% more than normal rooms.
3.2.4. Heat Maps Showing Relative Energy Consumption for Individual Rooms
Figure 23 shows the distribution of the annual cumulative cooling thermal energy demand per room for the eastern facade (top row) and the western facade (bottom row) to generate a “heat map” of energy consumption. Again, all values were normalized with respect to the room consuming least amount of energy (assigned value 1 and shown in blue colors).
Each row contains three figures (from left to right): the first shows the entire hotel facade, the second excludes the top floor (where suites are located), and the third excludes both the top floor and corner rooms to provide a more detailed view of the energy demand distribution per room across the facade.
The top-floor suites showed the highest energy consumption, with values up to eight times greater than those of the most energy-efficient room. Similarly, individual corner rooms consumed significantly more energy than centrally located rooms—up to four to five times more. The lowest cooling demands were concentrated in rooms on the lower floors of both facades. When the top floor and corner rooms are excluded, the impact of shading from surrounding buildings on the second floor became more apparent.
Figure 24 is analogous to the previous one with the difference that it shows the cumulative annual heating energy distribution per individual room instead of the cooling energy distribution.
As in the previous case, the highest cumulative annual energy consumption in terms of heating was concentrated in the suites, reaching up to 20 times the consumption of the most energy-efficient room. Corner rooms also showed significantly higher heating demand, likely due to their larger size and increased window area. For heating in particular, the highest consumption was observed in the lower floors of the hotel, likely due to reduced exposure to solar radiation (see Section 3.1.1).
3.2.5. Energy-Saving Potential Estimation
In order to analyze the impact of hotel occupancy on possible energy savings, the generated look-up table (see Section 2.5) comprising the simulation results of a fully occupied hotel (all rooms in “comfort” mode) and the results of the non-occupied hotel (all rooms in “economy” mode) was used in the following way.
For each day of the year, all rooms (excluding suites) were ranked according to their thermal energy demand, from low to high. Using the N first rooms (from lowest to higher energy consumption) would represent the “best case” scenario of room allocation. Using the N last rooms (from highest to lower energy consumptions) would represent the “worst case” scenario of room allocation.
Then, we selected a specific number of occupied rooms N and calculated the “best case” cumulative annual energy consumption, assuming that every day of the year N rooms would be occupied. Hence, every day of the year, the N rooms would be allocated in the energetically best way. This process was repeated to calculate the “worst case” cumulative annual energy consumption assuming a constant number of N occupied rooms throughout the year. Finally, the ratio between the “best case” and “worst case” cumulative annual energy consumptions assuming N rooms to be occupied was calculated. The entire process was performed varying N from 1 up to 411 rooms (suites excluded) to cover the range from lowest to highest possible hotel occupation. This approach was aimed at saving potentials associated with the hotel’s occupation.
The results are presented in Figure 25. For both cooling and heating, the energy-saving potential increases with occupancy up to a certain point—peaking at approximately 50% occupancy—before declining again. This pattern is explained by the system’s degrees of freedom: when occupancy is either very low or very high, there is limited flexibility in room allocation, reducing the potential for optimization. Maximum annual savings are observed at medium occupancy levels, around 160–180 occupied rooms. At this range, cooling energy savings can reach up to 6%, while heating energy savings are significantly higher, reaching up to 26%.
To obtain more realistic estimates of the energy-saving potential, actual hotel occupancy data from 2023 were used in a next step. This allowed the previously described process to be repeated using the real number of occupied rooms, N_real, for each day of the year. The results are presented in Figure A1 of the Appendix A.
The X-axis shows the months of the year. The primary Y-axis represents the cumulative annual thermal energy consumption in Watts: the blue bars indicate the most energy-efficient monthly consumption based on the “best case” room allocation strategy, while the green bars show the energy saved compared to the “worst case” allocation. The percentage savings achieved each month are displayed above the bars. Additionally, the secondary Y-axis (purple line) shows the actual monthly occupancy ratio.
In 2023, the months with the lowest occupancy were January, July, and August. The winter months—when the building requires heating—showed higher percentage savings (approximately 20%) than the summer months—when cooling is needed—where savings were around 5%. However, in absolute terms, a greater saving potential was observed during the summer due to much higher overall HVAC energy consumption, resulting in savings between 4% and 5% of the cumulative monthly thermal energy.
The average annual energy savings achieved through the optimized guest allocation strategy amounted to approximately 9.4%, with a standard deviation of ±7.0% across the months. This standard deviation reflects the strong seasonal variability of the saving potential, rather than statistical uncertainty, and captures the dispersion between peak heating and cooling periods.
3.2.6. Summary of Consumption Analysis and Saving Potentials
A summary of the above saving potential analysis is presented hereafter in Table 4, organized by simulation variables and geometrical aspects of the building.
While the energy savings estimates presented in this study are based on deterministic simulations, they should be interpreted as indicative values rather than absolute predictions. Several sources of uncertainty may affect the exact magnitude of savings in real-world scenarios, including climatic variability beyond the typical meteorological year used, temporal changes in occupancy patterns, and operational deviations from the assumed HVAC control strategy. Although standard deviations were not computed for the aggregated savings figures, these contextual uncertainties were qualitatively considered throughout the analysis.
4. Discussion
This study presents a novel, simulation-based approach to reduce HVAC energy consumption in hotels by strategically allocating guests to rooms with more favorable thermal profiles. In particular, we calculate here internal solar heat gains generated by solar irradiation and estimate the resulting HVAC energy demands to maintain occupied rooms at comfort levels. Unlike prior literature focused on optimal architectural design, system retrofitting, or benchmarking, this work emphasizes operational optimization using software tools and without requiring physical alterations to the building. This conceptual distinction positions the proposed approach as innovative and implementable, especially in existing hotel infrastructures.
The building energy model (BEM) simulation results revealed strong spatial and temporal variability in HVAC-related energy needs across the guest rooms of the large hotel tower under study located in Madrid, Spain. The simulations revealed shadowing effects of the surrounding built environment up to floor number eight, resulting in considerable less cooling demands in lower floors during summer months and less heating demands in upper floors during winter. Furthermore, the orientation of the rooms strongly influenced seasonal performance: annual cooling demands peaked in southeast- and northeast-facing rooms, while heating demands were concentrated in northwest- and southwest-facing rooms. Suites and corner rooms located on the upper floors were found to consume substantially more energy—up to eight times more in some cases–—compared to the normal guest rooms facing east or west. These differences were primarily attributed to larger room surface areas, increased glazing surface and higher solar exposure (less impact of shadows from surrounding buildings). The results are consistent with known physical behavior of solar incidence and thermal transfer, providing thus a form of phenomenological or “physical” validation for the BEM used.
A second layer of validation was obtained through comparison between simulated and actual hotel electrical energy consumption data. Although the simulations employed the Ideal Loads Air System (ILAS)—a simplified method assuming perfect HVAC performance and instantaneous temperature control—the seasonal and occupancy-based trends matched those observed in the real hotel consumption. The observed divergence in absolute values is consistent with known limitations of ideal simulations and reflects missing components such as system inefficiencies, infiltration losses, and control dynamics. Nevertheless, the alignment in behavioral trends reinforces the model’s ability for relative energy analyses as carried out here.
The potential for energy savings was not uniform throughout the year. The developed “best case” guest allocation strategy, assigning lowest consuming rooms first, proved most effective at intermediate hotel occupancy levels, particularly between 30% and 70%, where room selection flexibility allows for optimal distribution based on thermal performance. At near-full or near-empty occupancy, the impact of allocation was significantly diminished. These results underscore the importance of dynamically adapting the strategy to actual occupancy data in order to maximize energy savings. It is also worth noting that guest behavior profiles—such as differences between business and leisure travelers—may influence room usage patterns and corresponding HVAC loads. While not addressed in this study, these factors could be considered in future work to further refine energy-aware allocation strategies.
When situated within the broader literature, this study fills a critical gap. Prior works such as those by Nguyen (2019) [3] and Luo (2021) [26] emphasized pattern recognition and energy performance benchmarks, yet do not translate insights into actual operational strategies. Similarly, approaches by Yu (2024) [25] focus on physical upgrades or architectural redesign, while Chen (2022) [32] explores zonal control without implementing fine-grained spatial optimization. In contrast, the present work demonstrates how advanced, geometrically detailed simulations (including 439 internal thermal zones to model individual hotel rooms) can directly inform operational decisions at the room level, offering a pathway for data-driven energy management that does not rely on new construction or invasive retrofitting.
It is worth noting that the proposed method shows higher theoretical potential in climates with strong seasonal variations, such as continental or Mediterranean climates. In more thermally stable regions, such as tropical climates, the reduced variability in solar exposure may limit the relative effectiveness of the strategy. This highlights the importance of climatic context when evaluating the applicability and scalability of the proposed method.
5. Conclusions
This study has demonstrated that simulation-informed guest allocation strategies, based on solar heat gain and room geometry, can contribute to reducing HVAC energy demand in hotels. Using a high-resolution building energy model (BEM) of a real 17-floor hotel tower in Madrid, with detailed geometrical modeling of both the building envelope and internal room distribution, we identified strong spatial and seasonal variability in room-level energy requirements. All rooms were modeled as individual thermal zones (439 in total), while HVAC behavior was kept uniform across zones to ensure comparability. This allowed us to generate daily rankings of thermal demand under consistent system assumptions.
By combining these rankings with real hotel occupancy data from 2023, we estimated that applying the proposed allocation strategy could lead to HVAC energy savings of approximately 5% during summer and up to 20% during winter, with an annual average around 9.4%, with a standard deviation of ±7.0%. These savings are achieved entirely through software-based decision-making, without requiring changes to physical infrastructure or sacrificing guest comfort. The method shows particular effectiveness under intermediate occupancy levels, where allocation flexibility is highest.
This approach exemplifies how existing building infrastructure can be optimized through operational intelligence. Rather than focusing on architectural retrofits or hardware upgrades, it leverages simulation to support energy-aware decisions in day-to-day hotel management. The findings contribute to the broader field of smart building operation, offering a replicable framework that could be extended to other typologies such as hospitals, dormitories, or residential buildings.
This study presents a theoretical framework based on idealized simulations, and several limitations should be acknowledged. The HVAC energy demand was modeled using the Ideal Loads Air System (ILAS), which does not account for real-world inefficiencies, though it allows consistent comparison between rooms. Construction materials were based on standard templates as detailed information from the hotel was unavailable. Validation was carried out at an aggregate level due to the lack of sub-metered or room-specific empirical data. Additionally, the allocation strategy does not yet differentiate between guest types or usage patterns. Finally, the method has not been tested in a real operational setting. These limitations define the scope of the current work and highlight areas for future refinement and validation.
Future research should focus on incorporating real HVAC system specifications, material properties, and occupancy behavior profiles into the model. Pilot testing in an operational hotel environment could help evaluate the practical feasibility and performance of the strategy under real-world conditions.
Conceptualization, I.L.d.B. and J.L.B.; Methodology, I.L.d.B., M.F.I. and J.L.B.; Software, M.F.I. and J.O.G.; Validation, I.L.d.B., M.F.I., J.O.G. and J.L.B.; Formal analysis, I.L.d.B., J.O.G. and J.L.B.; Investigation, I.L.d.B. and J.L.B.; Resources, J.L.B.; Data curation, M.F.I. and J.O.G.; Writing—original draft, I.L.d.B. and J.L.B.; Writing—review & editing, I.L.d.B. and J.L.B.; Visualization, M.F.I. and J.O.G.; Supervision, V.F. and J.L.B.; Project administration, V.F. and J.L.B.; Funding acquisition, V.F. and J.L.B. All authors have read and agreed to the published version of the manuscript.
The original contributions presented in this study are included in the article. Further inquiries can be directed to the corresponding author.
We gratefully acknowledge the support of MINOR HOTELS EUROPE & AMERICAS and in particular the management team of the Eurobuilding hotel tower in Madrid for granting access to the hotel consumption, occupancy, and floor plan data.
Víctor Fabregat was employed by the company Regenera Energy. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Figure 1 Workflow of the simulation-based guest allocation method.
Figure 2 Workflow for obtaining a detailed 3D model of the surrounding built environment.
Figure 3 Triangulated mesh of the hotel’s surroundings exported from Google Earth.
Figure 4 Simplified built environment of the hotel where surrounding buildings were modeled as cubes (red). The hotel (prior to detailed modeling) is shown in the center in gray.
Figure 5 Detailed 3D model of the entire hotel under study.
Figure 6 Initial detailed model of the facade, including corner balconies.
Figure 7 Final model without corner balconies.
Figure 8 A 3D model of the entire hotel tower in cross-section view resulting in a total of 439 thermal zones (left) and detail on corner rooms (right). The base has been excluded from the analysis.
Figure 9 Simulated solar trajectory in June (left) and December (right).
Figure 10 Simulated hourly mean solar heat gain, organized per floor (from bottom to top) and per season (left to right), with deviations (color band). Note that “normal” rooms are oriented towards east (E) and west (W). All other orientations represent corners.
Figure 11 Accumulated solar heat gain per month, per floor (left to right); per orientation (color code).
Figure 12 Accumulated solar heat gain per month by orientation; per floor (color code).
Figure 13 Average simulated thermal energy demand for cooling per hour, including suites. Organized by orientation and season (color code).
Figure 14 Thermal energy demand for cooling accumulated per month, including suites; per orientation and per floor (color code).
Figure 15 Thermal energy for cooling demand accumulated per month, including suites; per floor and per orientation (color code).
Figure 16 Thermal energy for heating accumulated per month, including suites; per orientation; per floor (color code).
Figure 17 Correlations between simulated thermal energy demand and weather data (red: large positive correlations; blue: large negative correlations).
Figure 18 Ratios of real and simulated electricity consumption, plotted at hourly and daily resolutions. (top: daily; bottom: hourly).
Figure 19 Real hotel occupancy rate over the year 2023.
Figure 20 Ratios of cumulative annual thermal energy demand per floor (top: cooling; bottom: heating), including corner rooms and suites.
Figure 21 Ratios of cumulative annual thermal energy demand by orientation (top: cooling; bottom: heating).
Figure 22 Ratios of cumulative annual thermal energy demand by room type (left: cooling; right: heating).
Figure 23 Heat maps of the ratios of the cumulative annual cooling thermal energy demand per individual room; the least consuming room has a value of 1.0 (left: all rooms included, middle: without top floor, and right: without top floor and without corner rooms).
Figure 24 Heat maps of the cumulative annual heating thermal energy per individual room; the least consuming room has a value of 1.0 (left: all rooms included, middle: without top floor, and right: without top floor and without corner rooms).
Figure 25 Ratio of “best case” to “worst case” cumulative annual energy consumption as a function of hotel occupancy level (number of occupied rooms, X-axis) for both cooling (blue curve) and heating (orange curve). Energy-saving potential increases with occupancy, peaking around 160–180 occupied rooms, and then decreases due to reduced allocation flexibility at extreme occupancy levels. Maximum potential savings reach ∼6% for cooling and ∼26% for heating.
Geometric parameters of the modeled hotel rooms.
| Room Type | Surface [m2] | Volume [m3] | Window Surface Area [m2] |
|---|---|---|---|
| Standard | 39 | 107 | 4.5 |
| Corner | 44 | 119 | 11 |
| Suite1 | 74 | 204 | 17 |
| Suite2 | 74 | 204 | 17 |
| Suite3 | 77 | 210 | 19 |
| Suite4 | 122 | 335 | 17 |
Materials used for the simulations and their thermal properties in OpenStudio.
| Type | Boundary | Pre-Set | Materials | Conductivity [W/m·K] |
|---|---|---|---|---|
| Roof | External | Insulated roof | Roof tiles 20 mm, | 0.84 |
| glass fiber quilt 100 mm, | 0.04 | |||
| plywood 25 mm | 0.15 | |||
| Wall | Zone | Party Wall 1 | Plaster board 20 mm, | 0.7264 |
| standard brick 100 mm, | 0.8 | |||
| plaster board 20 mm | 0.7264 | |||
| Floor | External | Ground Floor 1 | Common earth 200 mm, | 1.28 |
| gravel 200 mm, | 1.28 | |||
| heavy mix concrete 100 mm, | 1.4 | |||
| horizontal air 20 mm, | ||||
| chipboard 25 mm | 0.15 | |||
| Window | External | DG—low-E—Krypton | Clear 3mm Soft LoE, | |
| Krypton 14 mm, | ||||
| clear 3 mm | ||||
| Roof | Zone | Ceiling 1 | Chipboard 25 mm, | 0.15 |
| EPS 100 mm, | 0.035 | |||
| plaster board 20 mm | 0.7264 | |||
| Wall | External | External wall 1 | Standard brick 100 mm, | 0.8 |
| Thermawall TW50 200 mm, | 0.022 | |||
| inner concrete block 100 mm | 0.51 | |||
| Floor | Zone | Internal floor 1 | Plaster board 20 mm, | 0.7264 |
| EPS 100 mm, | 0.035 | |||
| chipboard 25 mm | 0.15 |
Summary of simulation validation results.
| Variable | Category | Observations |
|---|---|---|
| Solar Gain | Seasonality and Height | Highest in spring and summer, decreases in autumn and winter Solar gain increases with floor height, especially on the top floor |
| Orientation | NE and SE have high morning peaks NW and SW peak later with lower intensity E and W have lower gains due to smaller rooms and windows | |
| Monthly Distribution | Peak cumulative gain from May to August Gain increases with floor height, notably on the top floor | |
| Cooling Energy Demand | Seasonality | Highest in spring and summer Lowest in autumn and winter |
| Orientation | NE, E, SE demand peaks in early hours NW, W, SW demand peaks later, aligning with solar gain | |
| Height Influence | Demand increases with height Highest on top floor (suites) Corner rooms consume more due to size and solar gain | |
| Heating Energy Demand | Seasonality | Required mainly in January and December Negligible in summer months |
| Orientation | NE, E, SE require less heating NW, W, SW require more heating | |
| Height Influence | Heating demand slightly decreases with height Top floor shows a more noticeable decrease | |
| Meteorological Conditions | Correlations | Strong positive correlation with solar gain (0.68) Positive correlation with outdoor air temperature (0.63) |
| Actual vs. Simulated Consumption | Trend Comparison and Magnitude | Simulated consumption trends match actual patterns (driven by occupancy) Simulated values are lower due to simplifications and omitted losses Strong and statistically significant (p < 0.05) correlation confirms similar behavioral patterns |
Summary of simulated energy consumption patterns and saving potential.
| Aspect | Condition/Grouping | Key Observations |
|---|---|---|
| Floor Height | Cooling Demand | Lowest on ground floor, increases progressively to 8th floor Top floor consumes 106% more than ground floor |
| Heating Demand | Decreases up to 8th floor and then rises on top floor (suites) Maximum difference reaches 204% between most and least-efficient floors | |
| Room Orientation | Cooling Demand | SE and NE have up to 40% higher demand than SW and NW E and W facades (smaller rooms/windows) show lowest demand |
| Heating Demand | NW shows highest demand, followed by SW and NE E and W again show lowest heating consumption | |
| Room Type | Cooling Demand | Corner rooms and suites consume 106–256% more than normal rooms Top-floor suites consume up to 8× more than most-efficient room |
| Heating Demand | Corner rooms and suites consume 194–719% more than normal rooms | |
| Room-Level Variation | Individual Rooms | Top-floor suites and corner rooms have highest consumption Centrally located, lower-floor rooms show lowest cooling demand, and higher rooms show less heating demand Shading effects from nearby buildings visible on lower floors |
| Saving Potential (Simulated) | Ideal Allocation (Constant Annual Occupancy) | Max savings at 50% occupancy (160–180 rooms) Up to 6% savings for cooling and up to 26% for heating Very low and very high occupancies reduce optimization flexibility |
| Actual 2023 Occupancy | Lowest occupancy in January, July, August Winter: higher relative savings Summer: higher absolute savings Average annual savings: 9.4% ± 7.0% (std dev) |
Appendix A
Figure A1 Simulated monthly thermal energy consumption and saving potential in 2023 based on actual hotel occupancy. Blue bars represent optimal energy consumption using the “best case” room allocation; green bars indicate energy savings compared to the “worst case” allocation. Percentage savings are shown above each bar. The purple line (secondary Y-axis) depicts the actual monthly occupancy ratio.
1. Zhou, T. Percentage-based thermal zoning approach for enhanced stock-level building energy performance modelling. Energy Build.; 2025; 329, 115231. [DOI: https://dx.doi.org/10.1016/j.enbuild.2024.115231]
2. Borowski, T. Hotel Adapted to the Requirements of an nZEB Building—Thermal Energy Performance and Assessment of Energy Retrofit Plan. Energies; 2022; 15, 6332. [DOI: https://dx.doi.org/10.3390/en15176332]
3. Nguyen, T. Developing an energy benchmarking system for hotel buildings using simulation. J. Green Build.; 2019; 14, pp. 1-22. [DOI: https://dx.doi.org/10.3992/1943-4618.14.3.1]
4. Permana, T. Energy Consumption Analysis Using Weighted Energy Index and Energy Modeling for a Hotel Building. Buildings; 2023; 13, 1022. [DOI: https://dx.doi.org/10.3390/buildings13041022]
5. Giuzio, T. HVAC operational efficiency using EnergyPlus. Energy Rep.; 2022; 8, pp. 3069-3081.
6. Cunha, T. Benchmarking for realistic nZEB hotel buildings. J. Build. Eng.; 2020; 30, 101298. [DOI: https://dx.doi.org/10.1016/j.jobe.2020.101298]
7. Sha, T. Development of a key-variable-based parallel HVAC energy predictive model. Build. Simul.; 2022; 15, pp. 1193-1208. [DOI: https://dx.doi.org/10.1007/s12273-021-0885-0]
8. Tharushi Imalka, S.; Yang, R.J.; Zhao, Y. Machine learning driven building integrated photovoltaic (BIPV) envelope design optimization. Energy Build.; 2024; 324, 114882. [DOI: https://dx.doi.org/10.1016/j.enbuild.2024.114882]
9. Boccalatte, A.; Fossa, M.; Ménézo, C. Best arrangement of BIPV surfaces for future NZEB districts while considering urban heat island effects and the reduction of reflected radiation from solar façades. Renew. Energy; 2020; 160, pp. 686-697. [DOI: https://dx.doi.org/10.1016/j.renene.2020.07.057]
10. Shen, T. Green Design Studio: A Modular-Based Approach for High-Performance Building Design. Doctoral Dissertation; Syracuse University: Syracuse, NY, USA, 2023.
11. Zhou, T. Multi-parameter building simulation for hotel optimization. Energy Procedia; 2023; 158, pp. 4041-4046.
12. Sezer, N.; Yoonus, H.; Zhan, D.; Wang, L.L.; Hassan, I.G.; Rahman, M.A. Urban microclimate and building energy models: A review of the latest progress in coupling strategies. Renew. Sustain. Energy Rev.; 2023; 184, 113577. [DOI: https://dx.doi.org/10.1016/j.rser.2023.113577]
13. Zhang, J.; Yuan, C.; Yang, J.; Zhao, L. Research on Energy Consumption Prediction Models for High-Rise Hotels in Guangzhou, Based on Different Machine Learning Algorithms. Buildings; 2024; 14, 356. [DOI: https://dx.doi.org/10.3390/buildings14020356]
14. Gao, G.; Li, J.; Wen, Y. Surrogate modeling for long-term and high-resolution prediction of building energy demand. Energy Built Environ.; 2022; 3, pp. 175-184. [DOI: https://dx.doi.org/10.1016/j.enbenv.2021.11.004]
15. Landa del Barrio, I.; Fernández-Vigil Iglesias, M.; Peppas, A.; Bruse, J.L.; Echeverría Trueba, J.B.; Fernández Bandera, C. Ensemble surrogate modeling of a real-time HVAC set-point optimization framework for PV self-consumption maximization. Energy Build.; 2025; 336, 115478. [DOI: https://dx.doi.org/10.1016/j.enbuild.2025.115478]
16. da Costa Paulo, B.; Aginako, N.; Ugartemendia, J.; Landa del Barrio, I.; Quartulli, M.; Camblong, H. Surrogate model of a HVAC system for PV self-consumption maximisation. Energy Convers. Manag. X; 2023; 19, 100396. [DOI: https://dx.doi.org/10.1016/j.ecmx.2023.100396]
17. Romero Rodríguez, L.; Duminil, E.; Sánchez Ramos, J.; Eicker, U. Assessment of the photovoltaic potential at urban level based on 3D city models: A case study and new methodological approach. Sol. Energy; 2017; 146, pp. 264-275. [DOI: https://dx.doi.org/10.1016/j.solener.2017.02.043]
18. Aftab, M.; Chen, C.; Chau, C.K.; Rahwan, T. Automatic HVAC Control with Real-time Occupancy Recognition and Simulation-guided Model Predictive Control in Low-cost Embedded System. arXiv; 2017; arXiv: 1708.05208[DOI: https://dx.doi.org/10.1016/j.enbuild.2017.07.077]
19. Gao, G.; Li, J.; Wen, Y. Energy-Efficient Thermal Comfort Control in Smart Buildings via Deep Reinforcement Learning. arXiv; 2019; arXiv: 1901.04693
20. Merabet, G.H.; Essaaidi, M.; Ben Haddou, M.; Qolomany, B.; Qadir, J.; Anan, M.; Al-Fuqaha, A.; Abid, M.R.; Benhaddou, D. Intelligent Building Control Systems for Thermal Comfort and Energy-Efficiency: A Systematic Review of Artificial Intelligence-Assisted Techniques. arXiv; 2021; arXiv: 2104.02214
21. Kim, J. LSTM-based Space Occupancy Prediction towards Efficient Building Energy Management. arXiv; 2020; arXiv: 2012.08114
22. Hoque, K.A.; Cauchi, N.; Abate, A. Analyzing Occupancy-Driven Thermal Dynamics in Smart Buildings. arXiv; 2019; arXiv: 1903.06123
23. Gao, G.; Li, J.; Wen, Y. Long-term analysis of cooling demand and heat recovery in hotel buildings using EnergyPlus. Energy Build.; 2024; 287, 112983. [DOI: https://dx.doi.org/10.1016/j.enbuild.2023.112983]
24. Dobbs, J.R.; Hencey, B.M. Model Predictive HVAC Control with Online Occupancy Model. arXiv; 2014; arXiv: 1403.4662[DOI: https://dx.doi.org/10.1016/j.enbuild.2014.07.051]
25. Yu, T. Long-term analysis of cooling demand and heat recovery in hotel HVAC systems. Energy Build.; 2024; 321, 114657.
26. Luo, T. Energy consumption patterns in hotel HVAC systems. Renew. Energy; 2021; 172, pp. 1001-1012.
27. Ji, T. Sensitivity analysis of HVAC control variables. Energy Build.; 2022; 263, 111990.
28. Alhuwayil, W.; Mujeebu, M.A.; Algarny, A. Impact of External Shading Strategy on Energy Performance of Multi-Story Hotel Building in Hot-humid Climate. Energy; 2018; 169, pp. 1166-1174. [DOI: https://dx.doi.org/10.1016/j.energy.2018.12.069]
29. U.S. Department of Energy. EnergyPlus Engineering Reference: The Reference to EnergyPlus Calculations; U.S. Department of Energy: Washington, DC, USA, 2023; Chapter: Ideal Loads Air System
30. Hall, I.J.; Prairie, R.R.; Anderson, H.E.; Boes, E.C. Generation of Typical Meteorological Years for 26 SOLMET Stations; Technical Report SAND78-1601 Sandia National Laboratories: Albuquerque, NM, USA, 1978.
31. ASHRAE. Air-Source Heat Pump Systems in Commercial Applications. ASHRAE J.; 2023; 65, pp. 35-40.
32. Chen, T. Guest comfort vs energy savings in hotel zoning. J. Build. Perform.; 2022; 13, pp. 87-96.
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.
Details
; Flores Iglesias Markel 2 ; Odriozola González Juan 2 ; Fabregat Víctor 3
; Bruse, Jan L 2
1 Fundación Vicomtech, Basque Research and Technology Alliance (BRTA), Mikeletegi 57, 20009 Donostia-San Sebastián, Spain, School of Architecture, University of Navarra, 31009 Pamplona, Spain
2 Fundación Vicomtech, Basque Research and Technology Alliance (BRTA), Mikeletegi 57, 20009 Donostia-San Sebastián, Spain
3 Regenera Energy, C. Molina de Segura 8, 30007 Murcia, Spain