Scenario-Based Simulation of Balinese Vernacular Passive Cooling for Energy-Efficient Hotels in Hot-Humid Climates

2.1 Research design

This study adopts an evidence-informed, scenario-based building performance simulation design. It is not presented as a mixed-methods study: no new interview, survey, or occupant data are collected or reported here. The environmental principles that define the design variables are drawn from the literature on tropical passive cooling and from the authors’ previously published field investigation [20], which is cited as prior work rather than reproduced as new evidence. The measurable impact of those variables on thermal performance and energy use is then evaluated through simulation [22]. The field observations presented in this section summarize the observational evidence previously reported by the authors [20] and are included only to justify the selection of simulation variables rather than to introduce new empirical findings.

The study proceeds in three stages. The first translates vernacular environmental principles into explicit model inputs: operable opening configuration, roof height and ventilation, envelope thermal properties, orientation, shading, and spatial openness. The second isolates each variable in a separate simulation run and then combines them into an integrated package. The third interprets the energy and thermal outputs and synthesizes them into a performance-based design framework.

2.2 Derivation of design variables

This stage identifies and classifies the vernacular architectural elements that contribute to passive cooling performance, and establishes the cultural and physical rationale for selecting them as simulation variables.

1. Literature review

A comprehensive literature review was conducted to examine the principles of Balinese vernacular architecture, passive cooling strategies in tropical climates, and previous studies on sustainable and green building design. This process enables the identification of key architectural elements such as cross-ventilation, roof configuration, and material properties, while also positioning the research within existing scientific discourse [16, 17]. Furthermore, the literature review serves to establish the theoretical framework and clarify the research gap related to the lack of integration between vernacular knowledge and building performance analysis [13, 18].

2. Field observations

Field observations were conducted on both traditional Balinese compounds and contemporary hotel buildings to document the physical characteristics and environmental responses of each typology. The observations focused on spatial configurations, building orientation, ventilation systems, and material usage. Through this process, the study captures empirical evidence of how traditional architectural principles are implemented and identifies the extent to which these principles are adapted or neglected in modern buildings. This comparative observation provides a critical basis for understanding the practical relevance of vernacular strategies in current architectural contexts [8, 15].

3. Prior field evidence

Interviews with architects, building practitioners, and local cultural experts were conducted and reported in the authors’ earlier publication [20]. Those findings inform the selection of design variables in the present study but are not re-analyzed or presented here as new results. Readers are referred to study [20] for the interview procedure, participant characteristics, and qualitative findings.

2.3 Simulation phase

This stage evaluates the thermal and energy performance of the selected variables through simulation. It provides measurable evidence of how each vernacular design element influences building performance and is the stage at which the study departs from the authors’ earlier field investigation [20].

Table 1. Building model and simulation inputs

1. Simulation tools

Two widely recognized building performance simulation tools were employed, namely EnergyPlus [23] and DesignBuilder [24]. EnergyPlus was utilized for detailed energy modeling and thermal analysis, while DesignBuilder functioned as an interface for model development, visualization, and scenario testing. These tools enable accurate simulation of building performance under varying environmental conditions. The model was built in DesignBuilder version 7.3.2 using the EnergyPlus version 23.2 calculation engine, run over a full calendar year at six timesteps per hour using the IDN_JW_Denpasar-Ngurah.Rai.Intl.AP.972300_TMYx.2011-2025.epw weather file (Denpasar / I Gusti Ngurah Rai International Airport, WMO 972300; climate.onebuilding.org). Natural ventilation was represented using scheduled natural ventilation, with opening control governed by occupancy, outdoor air temperature 22–30 ℃, indoor temperature >24 ℃, and HVAC OFF during the ventilation period (see Table 1). The model was NOT calibrated against metered energy data, because no measured monthly consumption series was available for the case buildings. No claim of compliance with ASHRAE Guideline 14 is therefore made [25, 26]. The results should be read as relative differences between scenarios under identical assumptions, not as absolute predictions of operational energy use. The absolute baseline is compared only against published energy-use ranges for tropical hotels [1-3]. This limitation is stated again in Section 4.5. The full set of model inputs required to reproduce the study is given in Table 1.

The climatic boundary conditions applied in all six runs are those embedded in the weather file identified in Table 1. Table 2 reports the monthly statistics extracted directly from that file, so that the thermal environment driving the simulation is documented rather than described in general terms.

The file describes a persistently warm and humid maritime tropical climate with a narrow annual temperature amplitude and high year-round humidity. Cooling demand is therefore continuous rather than seasonal, and the temperature window governing the natural-ventilation control in Table 1 (22–30 ℃) is satisfied for a substantial share of annual hours. This is the condition against which the marginal contribution of each vernacular variable is evaluated.

Table 2. Monthly climate statistics extracted from the Denpasar-Ngurah Rai International Airport, WMO 972300 file

2. Simulation parameters

The simulation analysis focuses on two primary categories of performance indicators. The first is building energy performance, evaluated using annual Building Energy Use Intensity (EUI), annual cooling-energy use, total annual electricity consumption, cooling share of total electricity consumption, and peak cooling demand. The second is indoor thermal performance, evaluated using occupied-zone mean air temperature under representative warm operating conditions [27, 28].

Adaptive thermal-comfort compliance is not formally assessed in this study because the required variables, including operative temperature, prevailing mean outdoor temperature, air velocity, humidity, clothing insulation, metabolic rate, and occupied-hours compliance, were not evaluated as a complete comfort dataset. Similarly, room-scale air-velocity distribution is not reported because the annual EnergyPlus model does not provide spatially resolved airflow velocity fields without a separately documented computational fluid dynamics (CFD) analysis [29]. The present study therefore limits its thermal interpretation to indicators directly generated and supported by the building-performance simulation.

The six simulation scenarios were developed by systematically modifying the architectural variables identified in the preceding stage, first individually and then in combination, allowing their conditional marginal effects to be compared with the performance of the integrated package.

2.4 Study area

The study was conducted in four major tourism regions in Bali, each representing distinct microclimatic and morphological characteristics. Sanur, as a coastal area, is characterized by high humidity levels and the influence of sea breezes. Kuta represents an urban tourism zone with dense development and significant urban heat island effects. Ubud, located in a highland cultural area, features a cooler climate and a strong presence of traditional architectural forms. Nusa Dua, as a planned tourism zone, reflects controlled development with standardized building regulations [12, 30].

The selection of these locations enables a comparative analysis across diverse environmental conditions, thereby enhancing the robustness and generalizability of the proposed design model in different tropical contexts [31].

2.5 Research variables

This study defines variables based on the integration of vernacular architectural elements and building performance metrics to establish a clear relationship between design strategies and energy efficiency outcomes.

1. Independent variables (design variables)

The independent variables represent key elements derived from traditional Balinese architecture, including cross-ventilation configuration, roof height and geometry, building orientation, and the use of local materials. These variables are selected based on their relevance to passive cooling performance and are systematically manipulated within simulation scenarios in six simulation runs to evaluate their individual and combined effects on building performance [32, 33]. Run P0 is the conventional baseline. Runs P1–P4 each introduce one predefined design-strategy package relative to P0, while retaining the other packages at their baseline settings. Run P5 applies the complete integrated package. Because interaction between variables means the combined effect need not equal the sum of the individual effects, the single-variable results are interpreted as conditional sensitivities rather than universal causal contributions. The scenario definitions are given in Table 3.

Table 3. Definition of the six simulation runs

The six runs were not selected arbitrarily. P1-P4 correspond to the four vernacular principles most consistently reported across the field precedents documented in Section 3.4 and the authors' earlier field study [20]: cross-ventilation through opposing openings, elevated ventilated roof volumes, envelope material properties, and courtyard-based spatial openness. Each was isolated in its own run so that its marginal contribution could be separated from the others, and P5 combines all four to test the integrated package actually promoted in vernacular practice. Other combinations (for example, pairwise interactions such as ventilation with roof design alone) were not simulated because a full factorial design across four two-level variables would require sixteen runs; the one-factor-at-a-time matrix was chosen as a tractable first step, and this trade-off is acknowledged as a limitation in Section 4.5. The P0–P5 values in Table 3 are drawn from the observed range across the four case regions in Section 3.3 (Sanur, Kuta, Ubud, Nusa Dua) rather than from a single building, so the prototype is representative of the range of conventional and vernacular practice observed across those regions rather than of any one specific hotel.

2. Dependent variables (performance indicators)

The dependent variables comprise measurable building-energy and indoor thermal-performance outcomes. Energy indicators include annual Building EUI, annual cooling-energy consumption, total annual electricity consumption, cooling share of total electricity use, and peak cooling demand. Indoor thermal performance is represented by occupied-zone mean air temperature under representative warm operating conditions. Formal adaptive-comfort compliance and room-scale air velocity are not treated as dependent variables because these require additional environmental and occupant-related inputs or CFD analysis beyond the scope of the present model.

The research framework runs from prior published field evidence [20], through the identification of vernacular environmental principles, to their translation into measurable simulation variables, the six-run simulation matrix (P0–P5), and finally the synthesis of a performance-based design framework. The findings are then synthesized into a comprehensive design model that integrates vernacular architecture with passive cooling strategies. The complete methodological sequence is illustrated in Figure 1.

1.png

Figure 1. Research design: from prior published field evidence to the six-run simulation matrix and the resulting performance-based design framework

The findings indicate that principles drawn from Balinese vernacular architecture, when expressed as explicit model inputs, improve simulated building energy performance in a hot-humid climate [16, 18]. This section discusses the implications of these findings in relation to the research questions, existing literature, and the development of a simulation-based integrative design framework.

4.1 Vernacular architecture as a passive cooling system

The first research question concerns the identification of vernacular architectural elements that function as passive cooling systems. The results confirm that elements such as cross ventilation, elevated roof structures, spatial openness, and local material utilization are not only culturally embedded but also perform effectively as climate-responsive design strategies [7, 8].

It is fair to ask whether these four variables are distinctively Balinese or simply generic passive-design strategies that happen to appear in Bali. The individual mechanisms - cross-ventilation, elevated roofs, low-conductance envelopes, and courtyard openness-are indeed used in passive design worldwide and are not unique to Bali in isolation. What is specific to the Balinese case, and is carried into the scenario-specific simulation inputs in Table 3, is (i) the particular combination and proportioning of these elements as codified in vernacular built form - for example the ratio of open pavilion (bale) to enclosed space, the roof-ridge height relative to occupied-zone height that produces the stack effect described in Section 3.3, and the natah (central courtyard) as an organizing spatial principle rather than an incidental gap between buildings-and (ii) the cultural rationale of Tri Hita Karana that determines why these forms were combined in this way rather than another. The contribution claimed here is therefore the translation of that specific combination and its underlying rationale into testable simulation variables, not the discovery of passive-cooling mechanisms that are novel in a physical sense.

Unlike modern buildings that rely heavily on mechanical air conditioning systems, traditional Balinese architecture demonstrates a deep understanding of environmental adaptation. The use of open pavilions (bale), central courtyards (natah), and permeable building envelopes allows for continuous airflow and heat dissipation. This indicates that vernacular architecture operates as an integrated environmental system rather than a collection of isolated design features [13, 35].

4.2 Quantified contribution of individual design variables

RQ2 concerns the marginal contribution of each design variable to the overall reduction in annual building EUI. The single-variable simulation runs yielded reductions of 8.3% for cross ventilation, 10.9% for elevated roof design, 14.9% for envelope material improvement, and 12.0% for spatial openness, compared with an overall reduction of 38.0% achieved by the fully integrated vernacular design package [16, 20]. The arithmetic sum of the individual reductions (46.1%) exceeds the integrated reduction by 8.1 percentage points, indicating that the passive strategies interact rather than contribute independently. This interaction arises because multiple design interventions simultaneously reduce the same sources of heat gain and cooling demand, resulting in diminishing incremental benefits when combined.

This finding is consistent with previous studies highlighting the effectiveness of passive cooling strategies in reducing building energy demand; however, the present study extends the existing literature by quantifying the individual contribution of each vernacular design element within a unified simulation framework. Unlike earlier investigations that primarily reported overall performance improvements or relied on qualitative interpretation, this study decomposes the contribution of cross ventilation, roof configuration, envelope materials, and spatial openness before evaluating their combined performance within a normalized Balinese hotel prototype. This methodological approach provides a clearer understanding of how individual passive strategies interact and offers a more reproducible basis for climate-responsive hotel design in tropical regions [18, 21, 36].

Table 9. Contextual comparison with selected previous studies

Table 9 provides contextual benchmarks rather than directly comparable estimates of energy savings. The meta-analysis in reference [16] pools residential buildings across multiple hot climates rather than hotels specifically, and reports a mean saving averaged across heterogeneous simulation tools, climates, and baseline definitions, so its ~29% figure is a broad benchmark rather than a like-for-like comparator. Reference [18] evaluates a single vernacular element (the rowshan screen) in a hot-dry residential context in Jeddah using coupled building-performance simulation and CFD at room scale, reporting indoor air temperature reductions of 15–22% rather than a whole-building EUI reduction; it is included as evidence that vernacular elements produce measurable thermal benefit in other climates, not as a numerically comparable result. Reference [20] is the authors' own prior field-metered study of the same building type and climate, and is the most directly comparable entry, though it reports an occupied-building metered saving (32.4%) rather than a simulated one. None of the three comparators shares building type, climate, simulation tool, and baseline definition simultaneously with the present study, so Table 9 should be read as contextual positioning rather than as a controlled benchmark.

4.3 Integration of culture and technology

The third research question focuses on how to formulate an effective integrative design model. The findings demonstrate that vernacular architectural principles can be transformed into a simulation-evaluated design framework that bridges the gap between cultural heritage and modern sustainability requirements [11, 14, 18].

This integration represents a paradigm shift in architectural design, where local wisdom is no longer treated as a symbolic or aesthetic element but as a functional and measurable performance strategy. The study shows that cultural architecture can be operationalized through simulation tools, enabling its adaptation in contemporary building design [15, 25].

This model illustrates how traditional architectural knowledge is translated into measurable performance outcomes and integrated into modern building design. It serves as a guideline for developing sustainable buildings that are both environmentally efficient and culturally responsive.

The relationships summarized in Table 10 are set out as an operational sequence in Figure 3. The framework traces the path from vernacular environmental principle, through its expression as a measurable model input, to the simulated performance outcome, and finally to the design decision it supports. Read in this direction, the framework is not a claim that vernacular form is inherently efficient; it is a procedure by which such a claim can be tested for any given element, climate file, and building program.

Table 10. Integration framework

3.png

4.4 Implication for sustainable architecture in tropical regions

The framework may be relevant beyond Bali, but the results are not transferable as they stand. Applying it elsewhere requires local weather files, envelope properties, occupancy patterns, and operating schedules, followed by re-simulation; the percentages reported here should not be carried over to other tropical settings [31, 37].

From a practical perspective, this approach provides architects and developers with a context-sensitive design strategy that is both environmentally and culturally sustainable. From a theoretical perspective, the study contributes to the development of a new framework that combines vernacular knowledge, environmental performance, and technological validation [38, 39].

4.5 Limitations

This study has six limitations that bound the interpretation of its results. First, the model was not calibrated against metered energy data; no measured monthly consumption series was available, so no ASHRAE Guideline 14 compliance is claimed and the absolute values should not be read as predictions of operational energy use [25, 26]. Second, the analysis is confined to a single climate context in southern Bali and one hotel typology; the reported percentages are not transferable to other tropical regions without re-simulation. Third, deterministic occupancy, equipment, and opening schedules do not represent the variability of real hotel operation, and simulated savings therefore tend to exceed those achieved in occupied buildings. Fourth, the one-factor-at-a-time design estimates conditional sensitivities under a fixed baseline; it does not capture the full interaction structure that a global sensitivity analysis would reveal. Fifth, room-scale airflow velocity was not modelled, so claims about perceived air movement are outside the scope of these results. Sixth, embodied carbon, life-cycle cost, moisture risk, acoustic performance, fire safety, and long-term durability were not assessed, and a design that performs well on operational energy may not perform well on those criteria.

Future work should therefore prioritize: monthly metering of the case buildings to permit formal calibration; global sensitivity analysis across the full parameter space; CFD for occupied-zone airflow; and life-cycle assessment and cost analysis to test whether operational savings survive a whole-life accounting.

The end-use and peak-demand figures in Table 5 are model outputs, not measurements. Because the model was not calibrated against metered consumption, the disaggregation should be read as an internally consistent attribution of the simulated reduction across end uses, not as a validated account of how energy is actually used in the case buildings.

The authors would like to express their sincere gratitude to Universitas Udayana for providing academic support, research facilities, and an enabling environment for the completion of this study. The authors also extend their appreciation to the Doctoral Program in Engineering, Faculty of Engineering, for their continuous guidance and institutional support throughout the research process.

Special thanks are addressed to the architects, building practitioners, and local cultural experts whose insights informed the identification of vernacular design principles and the selection of variables in the present work.

The authors further acknowledge the contribution of all parties involved in data collection, field observations, and simulation processes, including technical assistants and research collaborators. Their support has been instrumental in ensuring the accuracy and reliability of the research findings.

Finally, the authors would like to express their appreciation to colleagues and academic peers who provided constructive feedback and critical discussions during the development of this manuscript.