Optimising Thermal Comfort in Algerian Reference Hotel Across Eight Climate Zones: A Comparative Study of Simulation and Psychrometric Chart Results

Abstract

Since gaining independence in 1962, Algeria has significantly developed its tourism infrastructure, including notable projects by Fernand Pouillon. The thermal performance of hotel buildings, measured by discomfort hours and considering the design parameters for both PMV-PPD and adaptive comfort models, is a crucial study area across Algeria’s eight climate zones. This research focuses on the M’Zab Hotel in Ghardaïa, designed by Pouillon, utilising in situ measurements and dynamic simulations with EnergyPlus. After validating the simulation model, the performance of the optimised model, derived from sensitivity analysis parameters, is explored. A comparative study is conducted, analysing results obtained through simulation and psychrometric charts for both comfort models across Algeria’s climate zones. The findings indicate that the optimised design significantly reduces discomfort hours by 27.9% to 54.8% for the PMV-PPD model and 38.8% to 90.3% for the adaptive model, compared to the actual design performance. Strong correlations are observed between the simulation and psychrometric chart results for the PMV-PPD model, while the correlation for the adaptive model requires further investigation.

1. Introduction

In 2020, the average surface temperature underwent a significant increase of 1.02 °C compared to the period from 1951 to 1980, reflecting the pressing issue of global warming and climate change [1]. The urgency of energetically retrofitting both existing and new buildings has become apparent. Evaluating the thermal and energy performance of the building sector plays a crucial role in addressing this challenge. Prioritising thermal and energy building efficiency improvements is essential in the present and future [2].

Building modelling, software simulation, and climate analysis tools have become instrumental in tackling this challenge. These tools aid in estimating thermal comfort behaviour throughout the year using psychrometric charts. Prominent models for predicting human thermal comfort, such as the Predicted Mean Vote and Predicted Percentage of Dissatisfaction (PMV-PPD) and the Adaptive Comfort models, outlined in the ASHRAE Standard 55-2017 [3], are commonly used for mechanically conditioned (MC) and naturally ventilated (NV) buildings, respectively. The PMV-PPD model, developed by Fanger, is based on six primary parameters: metabolic rate, clothing insulation, air temperature, mean radiant temperature, air velocity, and relative humidity (RH). This model is typically applied in MC buildings, where indoor environmental conditions are more stable and controlled. In contrast, the adaptive comfort model is grounded in field studies and considers only two main parameters: operative indoor temperature and the prevailing mean outdoor air temperature. It is more suitable for NV buildings, where the occupants adapt to a wider range of indoor temperatures based on outdoor climatic conditions and behavioural adaptation. Nonetheless, real-world examples based on calibrated and validated models are crucial for establishing thermal and energy-efficient practices [4].

In Algeria, as in many parts of the world, the building sector accounts for approximately 40% of the total energy consumption and contributes 40% of energy-related carbon dioxide emissions (CO₂). This underscores the importance of addressing thermal and energy-related concerns within the tourism sector, particularly in the context of hotel buildings [5].

While several recent studies have focused on assessing and optimising thermal comfort through field measurements, surveys, and dynamic simulations, they often examined PMV and adaptive models independently. For instance, Koh et al. [6] assessed thermal comfort outcomes in retrofitted buildings in the Netherlands using material-based simulations, whereas Calama-González et al. [7] focused on predictive comfort modelling for Southern Spain’s existing housing stock based on validated simulation models. Jeong et al. [8] compared residential thermal comfort conditions across two Australian climates using adaptive comfort indices, highlighting regional variations. In the Algerian context, Amraoui et al. [9] investigated the thermal performance of neo-vernacular buildings in dry climates using envelope simulations and the PMV model, while Kaihoul et al. [10] explored passive design strategies in Southern Algeria based on psychrometric and qualitative analysis. Additionally, retrofit optimisation strategies that account for thermal comfort and cost considerations were studied by Calama-González et al. [11], showing the impact of simulation-driven decisions. However, there is a notable gap in the literature when it comes to comparative studies of a single case study employing both PMV-PPD and adaptive models. For instance, the study by Yao et al. [12] provided a comprehensive review of the evolution of adaptive thermal comfort models, comparing the heat balance approach (e.g., PMV), adaptive regression models, and hybrid adaptive heat balance models. It highlights the benefits and constraints of each and underscores the growing relevance of hybrid methods in climate-responsive design.

Similarly, research by Ter Mors et al. [13] on thermal comfort in primary school classrooms showed the limitations of PMV in predicting children’s thermal sensation and confirmed the importance of adaptive models in naturally ventilated spaces. Other studies, such as those by Orosa and Oliveira [14] and by Custódioet al. [15], further validated that adaptive models are more reliable in naturally ventilated or hybrid-mode buildings, while PMV remains accurate in fully air-conditioned environments. These findings are consistent with our analysis.

While previous studies have examined either PMV-PPD or adaptive models independently, often in limited climate contexts, our study bridges this gap by offering a dual-model evaluation of thermal comfort using both simulation-based and psychrometric chart methods across eight distinct Algerian climate zones. Through the optimisation of 13 climate-responsive design parameters applied to a real-world hotel case (M’Zab Hotel), this research provides a comprehensive comparison of NV and MC spaces, validated through regression analysis. By integrating simulation and graphical tools, the study not only confirms the reliability of passive design strategies but also lays the groundwork for region-specific thermal comfort zoning in arid and semi-arid regions—an area previously underexplored in the literature. This study aims to address this gap by first validating the simulation model and then examining the performance of the optimised model after applying parameters derived from a sensitivity analysis (SA). Subsequently, it conducts a comparative analysis of the results obtained through simulation and psychrometric chart predictions for both reference and optimised cases across Algeria’s climate zones. This research is motivated by a desire to contribute to the understanding of thermal behaviour in a large hotel building (M’Zab Hotel in Ghardaïa) and to offer insights into both MC and NV best practices. In doing so, it seeks to provide a thermal comfort assessment across Algeria’s eight climate zones, comparing the simulation results with those obtained from psychrometric chart predictions. The study aims to capture the attention of decision-makers and assist architects and engineers in enhancing the thermal efficiency of hotel buildings.

This study sets out to achieve the following objectives:

• Investigate the thermal behaviour of the M’Zab Hotel reference case across Algeria’s eight climate zones using psychrometric charts and dynamic simulations;
• Optimise the reference case by applying SA-derived parameters to both PMV-PPD comfort models in the eight climate zones of Algeria and compare the results with psychrometric chart predictions through linear regression;
• Identify best practice design parameters obtained through simulation and strategies derived from psychrometric charts in various Algerian climate zones.

The paper begins with a description of the method in Section 2; an examination of the theoretical foundation and a comprehensive literature review focusing on simulation-based optimisation studies are in Section 3. It is followed by an overview of the case study location and the extended regions included in the study, the description of the building, and the modelling and simulation in Section 4. Section 5 provides the optimisation analysis, and Section 6 provides the results achieved in detail. Finally, the discussion, implications and conclusions are derived.

2. Methods

The methodology employed in this study comprises the evaluation of psychrometric charts and dynamic simulations, along with a subsequent analysis of the results. The simulations were conducted after calibrating the models to investigate the thermal behaviour of the reference case with a climate-responsive design. Subsequently, an optimisation approach was applied to both comfort models across the eight climate zones of Algeria, guided by the best practice parameters identified. A comparative analysis was also carried out, employing linear regression to explore the results obtained through simulation and psychrometric chart predictions. Additionally, we explore the potential for design improvement through the application of a developed genetic algorithm (GA) and multi-objective optimisation methods, including occupant behaviour, economic feasibility, and environmental impact.

To enhance understanding of the employed methodology, Figure 1 presents an illustrative framework for this study.

3. Theoretical Foundations for Research

3.1. ‘Simulation-Based Optimisation’: A Literature Review

Ali-Toudert and Weidhaus [16] highlight the scarcity of best practice models for energy-efficient buildings in Algeria. They affirm that energy savings can be achieved by optimising parameters such as insulation levels, building orientation, window-to-wall ratio (WWR), and the use of passive design measures for natural ventilation and shading compared to locally inappropriate building types. This optimisation has the potential to eliminate the need for heating and significantly reduce cooling demands.

For subsequent research endeavours, it is recommended to focus on parameters that exhibit a discernible impact on thermal behaviour and energy performance. Utilising these parameters for multi-objective optimisation through GA is particularly pertinent in both the current and anticipated scenarios of climate change [17,18]. Given the computational demands and time consumption associated with using numerous parameters in optimisation, it becomes imperative to assess the sensitivity of these parameters in each climate. Categorising them based on their influence allows for the selection of the most impactful ones during the optimisation process [19]. Moreover, an increase in optimisation sampling and time has been observed to yield more optimal solutions [20]. The effectiveness of the optimisation algorithm, in terms of multi-objective optimisation, is evident across various algorithm types, including the bare-bones particle optimisation (PSO), which simplifies control parameters like inertia weight and acceleration coefficients, GAs that simulate the process of natural selection, and differential evolution (DE), which uses population-based optimisation methods. Each of these algorithms demonstrates unique strengths in balancing energy consumption and economic or environmental considerations in building energy performance [21].

3.2. Thermal Comfort and Energy Performance in Hotel Buildings

The growing demand for tourism activities necessitates a substantial number of hotels, making it imperative to measure and evaluate thermal behaviour and energy consumption from both the perspectives of NV and MC environments. Establishing benchmarks in this sector is crucial. Nguyen and Rockwood [22] contribute to this discourse by exploring the energy benchmark of hotels in Vietnam through surveys. Subsequently, they develop a reference model to assess other hotels and pave the way for additional studies focused on labelling hotels in terms of energy consumption intensity.

In a recent study, Koh et al. [6] examined the performance of rehabilitated cavity walls in the Netherlands to optimise thermal comfort and energy consumption. The study findings indicate that the use of an aerogel composite can lead to a substantial reduction in both cooling and heating demand, ranging from 51% to 72%. Furthermore, improvements in thermal comfort are notable, showing a decrease from 51% to 18%.

Calama-González et al. [7] investigated the thermal performance and sensitivity of a typical building within Spain’s common housing stock. The study reveals a notable 68% discomfort hours over the year and identifies the parameters that are most influential on thermal comfort.

In another study, Calama-González et al. [11] focused on optimising a typical building within the housing stock after conducting an SA of design parameters. Employing a multi-objective optimisation approach using a GA, the study aimed to simultaneously reduce thermal discomfort and cost across four climate zones in the southern region of Spain. The retrofit strategies included improvements such as enhanced insulation, upgraded glazing, and the implementation of shading devices. The results of the study demonstrate noteworthy improvements, with investment costs from approximately EUR 20–200/m² leading to significant reductions in annual overheating and undercooling hours across different climatic areas.

Jeong et al. [8] delved into the perception of thermal comfort using the adaptive model in two distinct climates in Australia, employing field data collection. The study’s results reveal a broader temperature range for thermal comfort, estimating an acceptable range of around 11 K, which is 4 K wider than the range prescribed by ASHRAE’s adaptive model. Additionally, the study indicates that the 80% acceptability range can shift according to the climate, with a 1.5 K difference observed between the two study regions.

Rawal et al. [23] introduced an adaptive thermal comfort model in India based on field surveys conducted across the country’s five climate zones. The study’s findings indicate that the proposed model predicts a comfort range of 16.3–35 °C, a broader span compared to both the PMV model of ASHRAE and the India Model for Adaptive Comfort for Commercial Buildings.

Du et al. [24] conducted a study using the Chinese Thermal Comfort Database to analyse the performance of the PMV-PPD model in both MC and NV buildings. The findings of the study indicate that the PMV model demonstrates greater accuracy in MC buildings compared to NV ones.

4. The Case Study

Algeria, situated in the northern part of Africa, has a landscape where approximately 84% of its total surface is covered by desert, resulting in a predominantly hot–dry climate. The renowned M’Zab Hotel, designed by Fernand Pouillon, is located in Ghardaïa, about 600 km south of Algiers. Algeria has four main climate zones based on the Köppen–Geiger classification: Mediterranean (Csa), semi-arid steppe (BSk), arid (BWh/BSh), and hyper-arid (BWh). The selected cities, namely Algiers, Saïda, Guelma, Tiaret, Ghardaïa, Ouargla, Béchar, and Illizi, represent these varied zones, thus offering a comprehensive basis for analysing thermal comfort strategies. Detailed information on various regions, climate classifications across Algeria, and specifics about the case study can be found in the study [4], where it was defined according to the Algerian Technical Regulation Document (DTR C3-T), which establishes the national thermal zoning system used for building energy performance standards and climate-responsive architectural design [25]. The Algerian climate transitions from temperate–humid along the northern coast to temperate–dry in high plateaus, ultimately reaching a hot–dry climate in the Algerian desert. This shift is accompanied by significant seasonal and daily temperature variations.

In Ghardaïa, where the M’Zab Hotel is situated, summers are characterised by high temperatures ranging from 15 to 45 °C, while winters can reach colder temperatures ranging from 0–25 °C. The RH in the region typically ranges from 16 to 34%, and the wind velocity varies between 3.4 and 5.1 m/s, with minimal rainfall throughout the year.

4.1. Ghardaïa: M’Zab Hotel, by Fernand Pouillon

Constructed between 1967 and 1971, the M’Zab Hotel, designed by Fernand Pouillon, has a capacity of 600 beds. The study was conducted during a renovation period in 2019. The M’Zab Hotel, located in the Saharan city of Ghardaïa, is a representative example of mid-20th-century modernist architecture in Algeria. It features passive design elements such as thick walls, internal courtyards, and limited glazing, which reflect local vernacular strategies for thermal comfort. In this study, a representative thermal zone within the hotel was modelled and simulated using dynamic software. The design was evaluated using both PMV-PPD and adaptive models across eight Algerian climate zones to explore the thermal performance of passive strategies and propose optimised retrofitting solutions. Additional details about the hotel and the study can be found in the previous work by [4] (refer to Figure 2a–d).

4.2. Psychrometric Chart

The contribution of passive strategies to achieving comfort is outlined in the bioclimatic psychrometric chart, adhering to ASHRAE Standard 55-2017 [3]. The primary goal is to explore discomfort hours throughout the year and effectively reduce them through passive strategies. The simplified graphical approach, utilising the psychrometric chart to represent the thermal comfort zone for an identified occupant, was employed as per the PMV-PPD model. The Climate Consultant (V. 6.0) tool [26] was utilised by inputting Ghardaïa’s climate data (Figure 3). Specifically, the software estimates annual discomfort hours by mapping hourly climate data onto a bioclimatic psychrometric chart, following ASHRAE Standard 55-2017. The tool identifies whether each hour falls inside or outside the thermal comfort zone, which is defined by user-specified parameters such as metabolic rate, clothing insulation, and operative temperature. Hours outside this zone are considered “discomfort hours”. In our study, Ghardaïa’s climate data was used as the input weather file, and the default values for summer and winter activity levels and clothing were adjusted to reflect realistic use conditions. For the summer scenario, the following values were applied: clothing insulation (Clo) = 0.5, metabolic rate (Met) = 1.1, and comfort range = 23 °C to 27 °C. For the winter scenario, the values were clothing insulation (Clo) = 1.0, metabolic rate (Met) = 1.1, and comfort range = 20 °C to 24 °C. These values were selected based on commonly accepted indoor activity levels and seasonal attire, which is consistent with ASHRAE Standard 55-2017 guidelines. The psychrometric chart suggests various passive strategies adopted in the case study to achieve comfortable conditions. The selection criteria for passive strategies in Climate Consultant (V. 6.0) were based on the percentage of annual hours that each strategy contributed to maintaining indoor conditions within the thermal comfort zone, as per ASHRAE 55-2017. Strategies showing a high potential impact, such as sun shading of windows, high thermal mass with night ventilation, natural ventilation cooling, passive solar direct gain (low mass), and internal heat gain, were prioritised. In the manuscript, we now explain how these strategies were interpreted in the building design. For example, sun shading was translated into optimised overhangs and vertical fins; thermal mass with night flushing was represented by the use of thick masonry walls and operable windows to enable night ventilation; and natural ventilation was integrated through cross-ventilation openings.

The results indicate that these strategies contribute to a total of 5169 comfortable hours out of 8760 h, representing 59.0%. Without applying any strategies, the comfort hours in the specific region are 1622 h (18.5%). Specifically, the software allows users to overlay annual hourly climate data on the psychrometric chart and identifies which hours fall inside or outside the comfort zone defined by the selected comfort model (PMV-PPD in our case). To determine the seasonal discomfort hours, we divided the year into summer months (June–August) and winter months (December–February) and manually counted the number of hourly points that fell outside the comfort zone during each season and applied each of the strategies. The psychrometric chart further illustrates the potential to achieve comfort conditions passively up to 79.0%, totalling 6919 h annually. The use of the Psychrometric Chart aims to reduce discomfort hours to 3591 h (41.0%), minimising summer discomfort to 22.9% and winter discomfort to 18.1% of the total yearly hours.

It is crucial to note that the adaptive model, though tested to evaluate the comfort zone, does not have a clear impact in adapting to the harsh, hot–dry climate of Ghardaïa. This model establishes a comfort zone based on the prevailing mean outdoor air temperature, suitable for naturally ventilated buildings. The results were obtained by configuring Climate Consultant (V. 6.0) with adaptive model parameters and importing the typical meteorological year (TMY) climate data for each city. The software then generated a psychrometric chart that visualises how many hours throughout the year fall inside or outside the adaptive comfort zone. The discomfort hours over the year are limited to 58.0% when applying the existing case study strategies, compared to 59.0% achievable with the simplified PMV-PPD model. Furthermore, a limit of 74.5% is verified by applying all passive strategies, while the simplified PMV-PPD shows 79.0%. This discrepancy arises because the adaptive model is applicable within a specific range of mean outdoor air temperatures (10 °C to 33.5 °C), which is exceeded in the hot summer months in these regions.

4.3. Simulation Model

In August 2019, in situ measurements of the case study were conducted using the Testo^® 480 instrument (Forbach, France), a tool that provides detailed information on the indoor thermal environment and aligns with the International Organisation for Standardisation (ISO) standard 7726-2002 [27]. Detailed information on the instruments is shown in

Table 1. Detailed information on instruments.

Description	Parameter Measured	Range	Accuracy
Digital temperature and humidity recorder (Ø 12 mm)	Air temperature	−20 to +70 °C	±0.2 °C (+15 to +30 °C) ±0.5 °C (Remaining Range)
	Relative humidity	0 to 100%RH	±(1.0%RH + 0.7% of mv) (0 to 90%RH) ±(1.4%RH + 0.7% of mv) (90 to 100%RH)
Surface probe (Ø 4 mm)	Surface temperature of wall and roof	−60 to +1000 °C	±1.5 °C (−40 to +375 °C) ±0.4% of reading (+375 to +1000 °C)
Vane/temperature probe (Ø 16 mm)	Air velocity	+0.6 to +50 m/s	±(0.2 m/s + 1% of mv) (0.6 to +40 m/s) ±(0.2 m/s + 2% of mv) (40.1 to +50 m/s)

mv = measurement value. Source: Testo^® (Forbach, France) official website [28].

. These measurements followed the guidelines recommended by ISO 7726 [27] and ASHRAE 55 [3]. The case study, representing a multi-zonal design, was modelled in EnergyPlus/DesignBuilder (Version 4.8.0.068), a tool for evaluating thermal and energy requirements. The model was calibrated for further investigations, focusing on the eastern wing due to the hotel’s extensive size. Each floor, including rooms and corridors, has a total surface area of 3333 m², with an additional 3481 m² for the patio. The model comprises multiple zones on each level (11 zones in level one, 8 in level two, 5 in level three, 2 in level four, and 1 in level five) (refer to Figure 4a,b). The in situ measurement campaign, including the selection of rooms and instrumentation protocol, was based on a previously validated study of the same hotel. Given the vast scale of the hotel, each wing, such as the eastern one analysed here, with a volume of approximately 44,120.5 m³, is treated as an autonomous thermal zone. This approach is consistent with the architectural and operational separation between wings, as documented in a prior validated study [4] (refer to Figure 5a,b).

For the simulation, an (.epw) file containing detailed hourly weather data for the reference year in Gardhaïa was imported. Night flushing ventilation is implemented by opening windows at night if the indoor temperature exceeds 22 °C. Lighting is applied automatically as per the minimum software requirements. The thermal characteristics of envelope materials and standard internal gains of occupants (excluding appliances) are considered. In this study, air changes per hour (ACH) values were not assumed to be constant but were dynamically calculated based on the Algerian thermal regulation DTR C3-T, which considers climate, building envelope characteristics, and occupancy-related requirements. Infiltration ACH values were computed separately for summer and winter using wind-driven and permeability-based equations. Natural ventilation (air renewal) rates were determined according to occupancy, with the extract flow rates defined per room type. These calculations allowed for seasonal and occupancy-sensitive ventilation inputs in the simulation model. For a full description of the equations and calculation methods, please refer to the DTR C3-T regulation [25], serving various simulation steps. The occupancy schedule is determined through walkthrough visits and interviews regarding the occupants’ behaviour (

Table 2. Typical occupancy schedule for the M’Zab Hotel simulation.

Space Type	Time Period	Occupancy Ratio (Fraction)	Notes
Guest Rooms	00:00–07:00	1.0	Full occupancy during night
	07:00–09:00	0.5	Morning activity (pre-checkout, grooming, breakfast)
	09:00–17:00	0.2	Most guests out, minimal presence
	17:00–23:00	0.8	Evening use, dinner, rest
Circulation Spaces	08:00–12:00	0.3	Regular guest movement
	12:00–14:00	0.5	Peak check-in/check-out and lunch movement
	14:00–18:00	0.3	Afternoon circulation
	18:00–23:00	0.4	Evening movement
Patio and Common Areas	08:00–23:00	0.3	Intermittent and informal use by guests and staff

) [29].

Table 3. Fundamental characteristics of the simulation process.

Criteria	Tool	Value
In situ measurements	Testo 480 (Forbach, France)	Aligns with ISO 7726-2002 and ASHRAE 55
Weather data	(.epw)	8769 h for the reference year
Simulation model	EnergyPlus/DesignBuilder (Version 4.8.0.068)	Surface area of 3333 m2 +3481 m2 for the patio 27 zones in total
Night flushing	EnergyPlus/DesignBuilder (Version 4.8.0.068)	Opening windows at night (T > 22 °C)
Lighting	EnergyPlus/DesignBuilder (Version 4.8.0.068)	Minimum software requirements
Envelope materials	Algerian Technical Regulation Document (DTR C3-T).	Material	Transmittance (W/m2K)
		First balcony wall composition	2.426
		Second load-bearing wall composition	1.341
		Roof composition	1.964
		Floor composition	1.611
		Single clear glazing	5.871
Internal gains from occupant	EnergyPlus/DesignBuilder (Version 4.8.0.068)	Standard internal gains of each space according to the occupancy (excluding appliances)
Air changes per hour (ACH)	Algerian Technical Regulation Document (DTR C3-T).		Summer	Winter
		Infiltration	0.1177	0.055
		Air renewal	0.1	0.6
Occupancy schedule	EnergyPlus/DesignBuilder (Version 4.8.0.068)	Walkthrough visits and interviews

presents the fundamental characteristics of the simulation model. Additional details about the simulation process are available in [4].

4.4. Model’s Validation

To validate the model for accurately describing thermal behaviour over extended periods, it is essential to address the uncertainty values associated with both the simulated and measured data. Two key equations are used for this purpose: the normalised mean bias error (NMBE) (Equation (1)) and the coefficient of variation of the root mean square error (CVRMSE) (Equation (2)). These equations are crucial tools for evaluating and quantifying the accuracy of the model’s predictions compared to the actual measured values. The NMBE provides insights into the overall bias between the simulated and measured data, while the CVRMSE assesses the precision of the model in capturing the variability within the dataset. To ensure model accuracy, the validation process employed NMBE and CVRMSE metrics in accordance with ASHRAE Guideline 14 and the IPMVP framework, both of which are internationally recognised standards for building energy simulation validation [30].

$$NMBE={\displaystyle \frac{\sum _{i=1}^n\left(y_i-{\widehat{y}}_i\right)}{\left(n-p\right)·\overline{y}}}·100$$

(1)

$$CVRMSE={\displaystyle \frac{\sqrt{\frac{\sum _{i=1}^n{\left(y_i-{\widehat{y}}_i\right)}^2}{\left(n-p\right)}}}{\overline{y}}}·100$$

(2)

where $y_i$ is the actual measurement for the period $i$, ${\widehat{y}}_i$ is the simulated value for period $i$, $\overline{y}$ is the average measurement, $n$ is the number of measured periods, and $p$ = 1.

EnergyPlus/DesignBuilder (Version 4.8.0.068) was used with measurements taken in August 2019 using the Testo^® 480 instrument (Forbach, France). Calibration was performed for both temperature (outside and indoor) and RH. The results, with an NMBE of 9.04% and −1.09% and a CVRMSE of 13.95% and 5.2% for outdoor and indoor temperatures, respectively, fall within ASHRAE’s acceptable limits. For RH, an NMBE of −4.57% and CVRMSE of 23.39% indicate reasonable accuracy, despite higher indoor variability.

Simulating the entire year using the ASHRAE 55 PMV-PPD model revealed discomfort hours, where the indoor environment did not achieve thermal comfort conditions, accounting for about 3741 h (42.7% of simulated hours). The summer discomfort hours were approximately 2154 h (24.6%), and the winter discomfort hours were around 1587 h (18.1%) of the whole year’s hours.

Comparing the psychrometric chart (ASHRAE 55 PMV-PPD model) with the EnergyPlus simulation results, both estimated discomfort hours percentages were at 41.0% and 42.7%, respectively. The psychrometric chart suggests that the maximum discomfort hours with passive strategies are around 21% annually. Additionally, the psychrometric chart indicates that the discomfort hours can be passively reduced from approximately 41.0% to about 21% with the incorporation of ‘two-stage evaporative cooling’ and ‘passive solar direct gain high mass’.

Table 4. Comfort and discomfort hours: psychrometric vs. simulation (PMV-PPD model).

Estimation Tool (Through a Single Year—8760 h)	Comfort (h)			Discomfort (h)
	Summer	Winter	Total	Summer	Winter	Total
Psychrometric chart *	2387 (27.2%)	2782 (31.8%)	5169 (59%)	2005 (22.9%)	1586 (18.1%)	3591 (41%)
Simulation **	2226 (25.4%)	2793 (31.9%)	5019 (57.3%)	2154 (24.6%)	1587 (18.1%)	3741 (42.7%)

* By Climate Consultant tool V 6.0 (build 13). ** By DesignBuilder Version 4.8.0.068.

and

Table 5. Comfort and discomfort hours: psychrometric vs. simulation (adaptive model).

Estimation Tool (Through a Single Year—8760 h)	Comfort (h)			Discomfort (h)
	Summer	Winter	Total	Summer	Winter	Total
Psychrometric chart *	2295 (26.2%)	2783 (31.8%)	5078 (58%)	2097 (24%)	1585 (18%)	3682 (42%)
Simulation **	-	-	6492 (74.1%)	-	-	2268 (25.9%)

* By Climate Consultant tool V 6.0 (build 13). ** By DesignBuilder Version 4.8.0.068.

provide details on the estimated comfort and discomfort hours throughout the year that were obtained from the psychrometric chart and the software model’s simulation of both the PMV-PPD and adaptive models. A regression analysis indicates a strong correlation between the two tools, with a Pearson correlation coefficient (R) of 0.9947 and a regression coefficient (R²) of 0.9895 (refer to Figure 6). These results confirm that the simulation aligns with the psychrometric chart estimation of the passively achievable comfort and discomfort hours.

5. The Optimisation Process

The selection of the 13 design parameters and their respective ranges was based on validated findings from a previous study on the same case, which included an in-depth literature review and sensitivity analysis. This prior work established a robust foundation for the optimisation process adopted in the present research. The sensitivity analysis and parametric simulations were conducted using DesignBuilder^® software (Version 4.8.0.068), which utilises EnergyPlus for dynamic simulation. The One-Variable-at-a-Time (OAT) method was applied to assess the individual influence of each design parameter on discomfort hours Building upon the exploration of the sensitivity of 13 design parameters on the thermal behaviour of both comfort models in a prior paper [4], which includes considerations such as azimuth orientation, max equipment power, type of external walls, type of glazing, rate of infiltration, type of internal walls, max general lighting, type of roofs, type of shading, max space density, scheme of natural ventilation (schedule and rate of ACH), and WWR, this study aims to optimise the reference case model across the eight climate zones of Algeria.

The baseline scenario, or reference case, aligns with the actual settings of the case study without any intervention. The baseline scenario in this study refers to the original architectural and thermal conditions of the M’Zab Hotel, as designed by Fernand Pouillon and built between 1967 and 1971. This scenario represents the existing state of the building prior to any optimisation. It includes:

-

The original materials and envelope construction (see Table 2), including double-layered hollow concrete block walls with air gaps, traditional plaster, and granite-tiled floors and roofs;

-

The geometry and zoning of the hotel’s eastern wing, which was selected due to the hotel’s large size. This includes multi-story rows of rooms surrounding a patio and detailed zoning based on floor plans;

-

The climatic data used for Ghardaïa over a reference year (2004–2018), with an 8760-h simulation in EnergyPlus/DesignBuilder (Version 4.8.0.068);

-

The internal conditions, such as the standard internal gains, ventilation assumptions (based on DTR C3-T regulations), and night ventilation control, based on indoor temperature thresholds (e.g., 22 °C);

-

The absence of mechanical HVAC operation, due to the building being unoccupied during the renovation period, ensures that the baseline represents purely passive performance conditions.

This baseline model served as the reference case for dynamic simulations. In contrast, the optimised case reflects modified settings derived from the sensitivity study. This optimisation process seeks to enhance the thermal performance and comfort conditions, taking into account the diverse climatic conditions present across Algeria’s different climate zones.

6. Results

The study aimed to compare the impact of various climate-responsive design strategies. The data were derived from the ASHRAE 55 standard values, where the metabolic rate is 1.00 for men, 0.85 for women, and 0.75 for children, or an average value if there is a mix of sizes. For clothing insulation, a 0.5 Clo level was used in summer, and a 1.0 Clo level was used in winter. Thirteen parameters were evaluated, and multiple scenarios were analysed using the local sensitivity analysis (LSA) with the OAT technique. These parameters included azimuth orientation, equipment power density, external wall types, glazing types, infiltration rates, internal wall types, general lighting power density, roof types, shading devices, space density, natural ventilation schemes (schedule and ACH), and WWR. The selection of these parameters was based on their known influence on building thermal behaviour, and the ranges of their variations were determined using the existing literature and preliminary studies.

The baseline scenario, reflecting the reference case’s actual conditions without any modifications, served as the reference for assessing the model’s behaviour across different climatic conditions. The outcomes of these optimised configurations compare the performance metrics across different scenarios, highlighting the effectiveness of each strategy in terms of energy savings and thermal comfort improvements. This comprehensive analysis provides insight into how each parameter was adjusted to achieve optimal building performance. By detailing the qualitative aspects of these strategies, the comparison between the baseline and optimised scenarios becomes clearer, demonstrating the tangible benefits of implementing these design modifications [4].

Table 6. Optimised design parameters for each climate zone.

N°	Climate Zone	Optimised Design Choices
		PMV-PPD Model		Adaptive Model
x1 (Azimuth)	Algiers/A	135° (northern west–southern east)		135° (northern west–southern east)
	Saïda/B
	Guelma/B1 and B2
	Tiaret/C	315° (northern west–southern east)
	Béchar/D	135° (northern west–southern east)
	Ghardaïa (El-Goléa)/E
	Illizi (Djanet)/E1
	Ouargla (Hassi Messaoud)/F
x2 (Equipment power)	Algiers/A	20 W/m2		20 W/m2
	Saïda/B			0 W/m2
	Guelma/B1 and B2
	Tiaret/C
	Béchar/D			16 W/m2
	Ghardaïa (El-Goléa)/E			12 W/m2
	Illizi (Djanet)/E1	0 W/m2		0 W/m2
	Ouargla (Hassi Messaoud)/F	20 W/m2		16 W/m2
x3 (Ext. walls)	Algiers/A	(P+I+W+P), W [0.4 m], I [0.1 m] (W) Adobe brick wall		(P+W+P), W [0.5 m] (W) Lime–sandstone wall
	Saïda/B	(P+W+I+P), W [0.4 m], I [0.1 m] (W) Hollow brick wall		(P+W+P), W [0.5 m] (W) Hollow concrete block wall
	Guelma/B1 and B2	(P+I+W+P), W [0.2 m], I [0.05 m] (W) Lime–sandstone wall		(P+W+P), W [0.5 m] (W) Hollow concrete block wall
	Tiaret/C	(P+W+I+P), W [0.2 m], I [0.05 m] (W) Adobe brick wall		P+W+P), W [0.5 m] (W) Lime–sandstone wall
	Béchar/D	(P+W+I+P), W [0.4 m], I [0.1 m] (W) Hollow brick wall		P+W+P), W [0.5 m] (W) Lime–sandstone wall
	Ghardaïa (El-Goléa)/E	(P+W+I+P), W [0.2 m], I [0.05 m] (W) Hollow brick wall		(P+W+P), W [0.3 m] (W) Hollow concrete block wall
	Illizi (Djanet)/E1	(P+I+W+P), W [0.4 m], I [0.1 m] (W) Adobe brick wall		(P+W+P), W [0.1 m] (W) Hollow concrete block wall
	Ouargla (Hassi Messaoud)/F	(P+W+I+P), W [0.4 m], I [0.1 m] (W) Lime–sandstone wall		P+W+P), W [0.5 m] (W) Lime–sandstone wall
x4 (Glazing)	Algiers/A	Double glazing + Krypton gas (SageGlass Climaplus Blue No Tint) U = 1.267		Double glazing + Krypton (SageGlass Climaplus Blue No Tint) U = 1.267
	Saïda/B
	Guelma/B1 and B2
	Tiaret/C
	Béchar/D
	Ghardaïa (El-Goléa)/E
	Illizi (Djanet)/E1
	Ouargla (Hassi Messaoud)/F
x5 (Infiltration)	Algiers/A	1 ac/h		0.3 ac/h
	Saïda/B			1.4 ac/h
	Guelma/B1 and B2	0.9 ac/h		0.7 ac/h
	Tiaret/C	1.4 ac/h		0.4 ac/h
	Béchar/D	0.7 ac/h		0.7 ac/h
	Ghardaïa (El-Goléa)/E			1 ac/h
	Illizi (Djanet)/E1	1.4 ac/h		1.4 ac/h
	Ouargla (Hassi Messaoud)/F	1 ac/h		0 ac/h
x6 (Int. walls)	Algiers/A	(P+W+P), W [0.1 m] (W) Hollow concrete block wall		(P+W+P), W [0.3 m] (W) Lime–sandstone wall
	Saïda/B
	Guelma/B1 and B2
	Tiaret/C	P+W+P), W [0.3 m] (W) Hollow brick wall
	Béchar/D	P+W+P), W [0.1 m] (W) Hollow brick wall		(P+W+P), W [0.3 m] (W) Hollow concrete block wall
	Ghardaïa (El-Goléa)/E	(P+W+P), W [0.1 m] (W) Hollow concrete block wall		(P+W+P), W [0.1 m] (W) Hollow concrete block wall
	Illizi (Djanet)/E1	P+W+P), W [0.1 m] (W) Hollow brick wall		(P+W+P), W [0.3 m] (W) Hollow concrete block wall
	Ouargla (Hassi Messaoud)/F	(P+W+P), W [0.1 m] (W) Hollow concrete block—case study wall		(P+W+P), W [0.1 m] (W) Hollow concrete block wall
x7 (Lighting)	Algiers/A	15 W/m2		0 W/m2
	Saïda/B
	Guelma/B1 and B2	12 W/m2
	Tiaret/C
	Béchar/D
	Ghardaïa (El-Goléa)/E
	Illizi (Djanet)/E1	6 W/m2
	Ouargla (Hassi Messaoud)/F	12 W/m2
x8 (Roofs)	Algiers/A	(G+C+I+R+P), R [0.3 cm], I [0.15 m] (R) Hollow polystyrene block slab		(G+C+I+R+P), R [0.1 cm], I [0.05 m] (R) Hollow concrete block slab—case study
	Saïda/B
	Guelma/B1 and B2	(G+C+R+P), R [0.3 m] (R) Hollow polystyrene block slab		(G+C+I+R+P), R [0.1 cm], I [0.05 m] (R) Hollow brick block slab
	Tiaret/C	(G+C+I+R+P), R [0.3 cm], I [0.15 m] (R) Hollow polystyrene block slab		(G+C+I+R+P), R [0.3 cm], I [0.15 m] (R) Hollow concrete block slab—case study
	Béchar/D			(G+C+R+P), R [0.3 m] (R) Hollow brick block with earth material slab
	Ghardaïa (El-Goléa)/E	(G+C+I+R+P), R [0.3 cm], I [0.15 m] (R) Hollow brick block with earth material slab
	Illizi (Djanet)/E1	(G+C+R+I+P), R [0.1 m], I [0.05 m] (R) Hollow concrete block slab—case study		(G+C+R+I+P), R [0.3 m], I [0.15 m] (R) Hollow polystyrene block slab
	Ouargla (Hassi Messaoud)/F	(G+C+I+R+P), R [0.2 cm], I [0.1 m] (R) Hollow polystyrene block slab		(G+C+R+P), R [0.3 m] (R) Hollow brick block with earth material slab
x9 (Shading)	Algiers/A	No shading		Overhangs + sidefines [1 m]
	Saïda/B	Horizontal overhangs [0.5 m]		Overhangs + sidefines [1.5 m]
	Guelma/B1 and B2	No shading
	Tiaret/C	Horizontal overhangs [1.5 m]
	Béchar/D	No shading		Horizontal overhangs [1.5 m]
	Ghardaïa (El-Goléa)/E	Horizontal overhangs [0.5 m]		Overhangs + sidefines [1.5 m]
	Illizi (Djanet)/E1	Overhangs + sidefines [1.5 m]
	Ouargla (Hassi Messaoud)/F	No shading
x10 (Space density)	Algiers/A	0 People/m2		0.4 People/m2
	Saïda/B			0.2 People/m2
	Guelma/B1 and B2			0.3 People/m2
	Tiaret/C			0.4 People/m2
	Béchar/D			0.2 People/m2
	Ghardaïa (El-Goléa)/E
	Illizi (Djanet)/E1			0.1 People/m2
	Ouargla (Hassi Messaoud)/F			0.2 People/m2
x11 (Ventilation scheme)	Algiers/A	On 7/24		On 7/24
	Saïda/B
	Guelma/B1 and B2
	Tiaret/C
	Béchar/D
	Ghardaïa (El-Goléa)/E
	Illizi (Djanet)/E1
	Ouargla (Hassi Messaoud)/F
x12 (Ventilation rate)	Algiers/A	12 ac/h		12 ac/h
	Saïda/B
	Guelma/B1 and B2
	Tiaret/C
	Béchar/D
	Ghardaïa (El-Goléa)/E
	Illizi (Djanet)/E1
	Ouargla (Hassi Messaoud)/F
x13 (WWR)	Algiers/A	50%		10%
	Saïda/B	30%
	Guelma/B1 and B2
	Tiaret/C	10%
	Béchar/D	30%
	Ghardaïa (El-Goléa)/E
	Illizi (Djanet)/E1	10%
	Ouargla (Hassi Messaoud)/F	30%
‘Shaded cells’ present design choices that are different between PMV-PPD model and adaptive model
Wall: (P) Plaster * (W) Wall’s main materials: - Hollow brick wall - Adobe brick wall - Lime–sandstone wall - Hollow concrete block—case study wall - Hollow concrete block wall (I) Insulation (expanded polystyrene) (A) Air gap			Roof: (G) Granite tile ** (C) Cement layer *** (R) Roof’s main materials: - Hollow polystyrene block slab - Hollow brick block with earth material slab - Hollow brick block slab - Hollow concrete block slab—case study (P) Plaster * (I) Insulation (expanded polystyrene)

* Plaster thickness is fixed for all cases (0.025 m). ** Granite tile thickness is fixed for all cases (0.025 m). *** Cement layer thickness is fixed for all cases (0.025 m).

provides an overview of the optimised parameters in the eight cities for both the PMV-PPD and adaptive models.

The subsequent sections present the results of the optimisation for both models. The discomfort hours data is assessed based on the humidity ratio and operative temperature conformity with the regions specified in ASHRAE Standard 55-2017 for the PMV-PPD model and within the regions shown in ASHRAE Standard 55-2017 for the adaptive model. The eight representative Algerian climate regions used in the study are as follows:

-

Zone A—Algiers (Coastal Mediterranean, mild climate);

-

Zone B—Saïda (continental semi-arid);

-

Zone B1 and B2—Guelma (semi-arid with more seasonal variation);

-

Zone C—Tiaret (high plateau, dry);

-

Zone D—Béchar (arid steppe, near desert);

-

Zone E—Ghardaïa (El-Goléa) (hot–dry, Saharan fringe);

-

Zone E1—Illizi (Djanet) (hot–dry, deep Sahara);

-

Zone F—Ouargla (Hassi Messaoud) (extreme hot–dry desert);

The following segment delves into a comparative study between the optimised and base cases, focusing on the discomfort hours for both the PMV-PPD and adaptive models. Additionally, the simulation results are compared to those obtained from the psychrometric chart to establish a correlation between the two tools. The results for each climate zone are presented separately for the PMV-PPD and adaptive models in the following sections.

6.1. Results of the PMV-PPD Model

Table 7. Discomfort hours: simulation vs. psychrometric chart (PMV-PPD model).

	Model	Algiers/A	Saïda/B	Guelma/B1 and B2	Tiaret/C	Béchar/D	Ghardaïa (El-Goléa)/E	Illzi (Djanet)/E1	Ouargla (Hassi Messaoud)/F
Climate
Sim.	Base (h)	3945	3836	3825	3895	3790	3741	3717	3784
	Optim. (h)	1801	1734	1744	2445	2015	2277	2679	2390
Chart	Base (h)	4257	3406	3833	3763	3560	3591	2763	3629
	Optim. (h)	2330	2479	2575	3197	1769	1840	674	2006

Base (actual design) discomfort hours per year in (h). Optim. (optimised design) discomfort hours per year in (h) (in shaded cells).

details the estimated discomfort hours for the eight cities in both the base and optimised cases obtained by the psychrometric chart and the software model’s simulation.

After simulating the actual design in the eight cities, the base case discomfort hours are as follows: 3945 h, 3836 h, 3825 h, 3895 h, 3790 h, 3741 h, 3717 h, and 3784 h per year. This represents 45%, 43.8%, 43.7%, 44.5%, 43.3%, 42.7%, 42.4%, and 43.2% of the total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively. The psychrometric chart of the adapted passive strategies predicts the discomfort hours in the eight cities as 4257 h, 3406 h, 3833 h, 3763 h, 3560 h, 3591 h, 2763 h, and 3629 h per year. This corresponds to 48.6%, 38.9%, 43.8%, 43%, 40.6%, 41%, 31.5%, and 22.9% of the total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively.

The results show close alignment between the simulation and the psychrometric chart in Guelma, Béchar, Ghardaïa, Ouargla, and Tiaret. However, the differences slightly increase in Algiers, Saïda, and Illizi, indicating that the design does not precisely meet the psychrometric chart predictions in these locations. Notably, only Algiers and Guelma have fewer values in the simulation compared to the psychrometric chart predictions.

For the optimised cases in each zone, the discomfort hours obtained by the simulation are as follows: 1801 h, 1734 h, 1744 h, 2445 h, 2015 h, 2277 h, 2679 h, and 2390 h per year. This represents 20.6%, 19.8%, 19.9%, 27.9%, 23%, 26%, 30.6%, and 27.3% of total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively. The psychrometric chart for the adapted passive strategies predicts the discomfort hours in the eight cities as 2330 h, 2479 h, 2575 h, 3197 h, 1769 h, 1840 h, 674 h, and 2006 h. This corresponds to 26.6%, 28.3%, 29.4%, 36.5%, 20.2%, 21%, 7.7%, and 22.9% of the total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively.

The results for the optimised case show proximity between the simulation and the psychrometric chart in Béchar, Ghardaïa, and Ouargla. However, the differences increase in Algiers, Saïda, Guelma, Tiaret, and Illizi. The simulation results indicate that only Algiers, Saïda, Guelma, and Tiaret have fewer values than the psychrometric chart predicts (refer to Figure 7).

In terms of percentage reduction, the discomfort hours of the optimised design are 54.8%, 54.4%, 54.3%, 46.8%, 39.1%, 37.2%, 36.8%, and 27.9% less than the actual design performance of the PMV-PPD model for Saïda, Guelma, Algiers, Béchar, Ghardaïa, Tiaret, Ouargla, and Illizi, respectively. For the psychrometric chart, the results show that the discomfort hours of the optimised design are 75.6%, 50.3%, 48.8%, 45.3%, 44.7%, 32.8%, 27.2%, and 15% less than the actual design performance of the PMV-PPD model for Illizi, Béchar, Ghardaïa, Algiers, Ouargla, Guelma, Saïda, and Tiaret, respectively.

The regression analysis between the two tools for the base case indicates a strong correlation with a Pearson correlation coefficient (R) of 0.7968 and a regression coefficient (R²) of 0.5741 (refer to Figure 8a). For the optimised case, the regression analysis shows a moderate correlation, with a Pearson correlation coefficient (R) of −0.4641 and a regression coefficient (R²) of 0.2154 (refer to Figure 8b). The regression analysis between the two tools for both the base and optimised cases indicates a strong correlation with a Pearson correlation coefficient (R) of 0.7127 and a regression coefficient (R²) of 0.5080 (refer to Figure 8c). We emphasise that, although some differences occur—particularly in the optimised case—the overall positive correlation confirms the general alignment between early-stage design insights from the psychrometric method and the performance outcomes from the dynamic simulation.

6.2. Results of the Adaptive Model

Table 8. Discomfort hours: simulation vs. psychrometric chart (adaptive model).

	Model	Algiers/A	Saïda/B	Guelma/B1 and B2	Tiaret/C	Béchar/D	Ghardaïa (El-Goléa)/E	Illzi (Djanet)/E1	Ouargla (Hassi Messaoud)/F
Climate
Sim.	Base (h)	2810	2345	2711	2050	2620	2268	3398	2312
	Optim. (h)	528	1434	765	198	995	1333	2050	774
Chart	Base (h)	3421	3376	3496	3767	3616	3682	2716	3657
	Optim. (h)	2592	3171	3136	3836	2146	2234	963	2356

Base (actual design) discomfort hours per year in (h). Optim. (optimised design) discomfort hours per year in (h).

details the estimated discomfort hours for the eight cities in both the base and optimised cases obtained by the psychrometric chart and the software model’s simulation.

After simulating the actual design in the eight cities, the base case discomfort hours are as follows: 2810 h, 2345 h, 2711 h, 2050 h, 2620 h, 2268 h, 3398 h, and 2312 h per year. This represents 32%, 26.8%, 31%, 23.4%, 29.9%, 25.9%, 38.8%, and 26.4% of total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively. The psychrometric chart for the adapted passive strategies predicts the discomfort hours in the eight cities as 3421 h, 3376 h, 3496 h, 3767 h, 3616 h, 3682 h, 2716 h, and 3657 h per year. This corresponds to 39%, 38.5%, 39.9%, 43%, 41.3%, 42%, 31%, and 41.7% of the total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively.

The results show differences between the base case simulation and the psychrometric chart predictions in Algiers, Guelma, and Illizi. The disparity increases in the other cities, indicating that the design in the adaptive model does not precisely meet the psychrometric chart predictions. Additionally, only Illizi has more values in the simulation than what the psychrometric chart predicts.

For the optimised case, the discomfort hours obtained by simulation are as follows: 528 h, 1434 h, 765 h, 198 h, 995 h, 1333 h, 2050 h, and 774 h per year. This represents 6%, 16.4%, 8.7%, 2.3%, 11.4%, 15.2%, 23.4%, and 8.8% of the total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively. The psychrometric chart for the adapted passive strategies predicts the discomfort hours in the eight cities as 2592 h, 3171 h, 3136 h, 3836 h, 2146 h, 2234 h, 963 h, and 2356 h. This corresponds to 29.6%, 36.2%, 35.8%, 43.8%, 24.5%, 25.5%, 11%, and 26.9% of the total year’s hours for Algiers, Saïda, Guelma, Tiaret, Béchar, Ghardaïa, Illizi, and Ouargla, respectively.

The results for the optimised case show a complete contrast between the simulation and the psychrometric chart in all cities. Additionally, the prediction of the simulation reveals that only Illizi has more values than what the psychrometric chart predicts (refer to Figure 9).

In terms of percentage reduction, the discomfort hours of the optimised design are 90.3%, 81.2%, 71.8%, 66.5%, 62%, 41.2%, 39.7%, and 38.8% less than the actual design performance of the adaptive model for Tiaret, Algiers, Guelma, Ouargla, Béchar, Ghardaïa, Illizi, and Saïda, respectively. For the psychrometric chart, the results show that the discomfort hours of the optimised design are 64.5%, 40.7%, 39.3%, 35.6%, 24.2%, 10.3%, and 6.1% less than the actual design performance of the adaptive model for Illizi, Béchar, Ghardaïa, Ouargla, Algiers, Guelma, and Saïda, respectively. However, in the case of Tiaret, the results of the optimised case are 1.8% more than the actual design; this contradiction of strategies between seasons should be investigated further in the psychrometric chart prediction.

The regression analysis between the two tools for the base case indicates a strong correlation with a Pearson correlation coefficient (R) of −0.8898 and a regression coefficient (R²) of 0.7918 (refer to Figure 10a). For the optimised case, the regression analysis shows a strong correlation with a Pearson correlation coefficient (R) of −0.7527 and a regression coefficient (R²) of 0.5666 (refer to Figure 10b). The regression analysis between the two tools for both THE base and optimised cases indicates a weak correlation with a Pearson correlation coefficient (R) of 0.1939 and a regression coefficient (R²) of 0.0376 (refer to Figure 10c). The significant differences between the simulation model and the psychrometric chart in the optimised adaptive model are mainly due to the simplified assumptions of the chart, which does not account for indoor thermal dynamics or the cumulative effects of passive strategies. The dynamic simulation, however, reflects the real-time interactions of building systems, leading to more accurate and often lower discomfort hour predictions. Illizi’s case, where the simulation results exceed the chart estimations, highlights limitations in passive cooling effectiveness in extremely hot climates and underscores the need for climate-specific design refinements.

The PMV-PPD model is a heat balance model grounded in laboratory studies. It assumes that occupants are in a steady-state condition with limited ability to adapt, and it requires fixed parameters, such as metabolic rate, clothing insulation, air speed, and humidity. It is typically more sensitive to deviations from the thermal neutrality defined by those fixed parameters and, thus, tends to report higher discomfort hours—especially in hot and arid climates where adaptive behaviours are common but not captured by the model.

In contrast, the Adaptive Model is derived from field studies and reflects the dynamic behaviour of occupants who actively adjust to their environment (e.g., changing clothing, opening windows, and using fans). It is inherently more tolerant of variations in indoor conditions, especially in naturally ventilated spaces, and accommodates psychological and behavioural adaptations that reduce discomfort.

This conceptual divergence explains why the Adaptive Model often yields significantly fewer discomfort hours in both the base and optimised cases, especially in warm climates. Additionally, the simulation results using the Adaptive Model showed a stronger linear correlation with the psychrometric chart than the PMV-PPD model in the base case (R = −0.8898 vs. R = 0.7968), highlighting the closer alignment between the psychrometric design logic and the adaptive comfort assumptions.

The differences are particularly visible in cities like Illizi and Tiaret, where climatic extremities challenge the assumptions of the PMV-PPD model. For example, in Illizi, the Adaptive Model simulation predicted 2050 discomfort hours in the optimised case, while the PMV-PPD model predicted 2679 h under similar conditions—reflecting a greater sensitivity of the PMV-PPD model to high temperatures and limited indoor adaptability.

Table 9. Linear regression: simulation vs. psychrometric chart (PMV-PPD and adaptive model).

Linear Regression Criteria	PMV-PPD			Adaptive
	Base	Optim.	Base and Optim.	Base	Optim.	Base and Optim.
Pearson correlation coefficient (R)	0.79	−0.46	0.71	−0.88	−0.75	0.19
Regression coefficient (R2)	0.57	0.21	0.50	0.79	0.56	0.03

Base (actual design) discomfort hours obtained by simulation and psychrometric chart. Optim. (optimised design). Discomfort hours obtained by simulation and psychrometric chart.

illustrates the outcomes of the linear regression analyses that compare the simulation results with those derived from the psychrometric chart for both the base and optimised cases in both the PMV-PPD and adaptive models.

7. Discussion and Implications

This study focused on optimising the thermal performance of the M’Zab Hotel model across eight climate zones in Algeria using both the PMV–PPD–Percentage of People Dissatisfied and adaptive comfort models. The key conclusion is that integrating passive solutions into retrofitting or new building designs is crucial for enhancing thermal behaviour and reducing energy loads. The study found that both NV and MC spaces could be optimised effectively, yielding promising results.

In the base case, the adaptation of passive strategies in the M’Zab Hotel demonstrated effectiveness in terms of discomfort hours simulation and achieving psychrometric chart predictions of the PMV-PPD model. However, the results were contrary to the adaptive model, where the simulation predictions were fewer than what the psychrometric chart estimated. The optimised design exhibited effective behaviour throughout the year, showcasing the studied passive strategies’ impact on both the PMV-PPD and adaptive models. The simulation predictions were lower than the psychrometric chart estimates for the PMV-PPD model in terms of discomfort hours (except for the southern cities, which are 50% of the case studies). This is due to the limitation of the adopted method through one-variable-at-a-time optimisation. For the adaptive model, the simulation values were significantly lower than the psychrometric chart estimates (except for Illizi).

The significant reduction in discomfort hours in the optimised design is primarily attributed to the integration of tailored passive design strategies that respond to the local climate characteristics of each city. These include improved building envelope insulation, optimised window-to-wall ratios, enhanced shading devices, and appropriate thermal mass use—each selected based on bioclimatic needs (e.g., solar radiation control in hot–arid climates or night ventilation in high diurnal range zones). From a qualitative standpoint, the optimisation improved the building’s ability to reduce indoor overheating and maintain thermal stability throughout the year. The dynamic simulation results demonstrate how these strategies collectively minimise thermal swings, delay heat gains, and enhance natural ventilation effectiveness—factors that are not fully accounted for in the base case or the static psychrometric evaluations. These modifications enable buildings to remain within the comfort zone for more extended periods, significantly lowering the predicted discomfort hours in both the PMV-PPD and Adaptive Models.

However, the thermal comfort issue of the indoor environment has several related aspects that could affect it, such as energy performance, indoor air quality, and daylighting. The need for further studies corresponding with this theme using developed methods is inevitable. For example, a recent study by Fangli Hou et al. [31] has optimised thermal comfort, in addition to indoor air quality and energy performance, through an extreme learning machine model optimised by an algorithm. Another study [32] shows the effectiveness of a multi-objective optimisation algorithm in improving the thermal comfort and energy consumption of a building’s air conditioning and mechanical ventilation. Also, a study [33] calculates the optimal temperature to ensure indoor thermal comfort using Building Information Technology and the Internet of Things. However, more studies should be conducted to develop the thermal comfort of the indoor environment and to set up a thermal comfort map for both NV and MC buildings over the climate zones of Algeria, which is the aim of this study.

While this study focused on thermal comfort, it is essential to acknowledge that indoor environmental quality involves multiple interconnected aspects, such as energy performance, indoor air quality, and daylighting. Future studies should explore these dimensions using advanced methods, for instance by using machine learning models and algorithms to assess thermal comfort alongside indoor air quality and energy performance [31]. Multi-objective optimisation algorithms have proven effective in enhancing thermal comfort and reducing energy consumption in buildings [32]. Additionally, studies exploring optimal temperatures for indoor comfort using Building Information Technology and the Internet of Things contribute to the comprehensive understanding of indoor thermal comfort [33].

Further research is necessary to develop thermal comfort maps for both NV and MC buildings across the diverse climate zones of Algeria. The methodology presented in this research offers valuable insights for architects aiming to optimise the thermal design of buildings in similar climate zones globally.

While this study primarily focused on the thermal comfort performance of passive design strategies, it is important to acknowledge their economic implications. Many of the implemented optimisations, such as enhanced insulation, strategic shading, and improved natural ventilation, are low-cost measures that can be integrated with minimal disruption, particularly in new constructions or planned retrofits. These strategies generally involve modest capital investment while offering substantial reductions in operational cooling demand. Preliminary evidence from similar contexts suggests that these interventions often present favourable payback periods and long-term energy cost savings. Future research will aim to quantify these economic indicators through life cycle cost analysis (LCCA) tailored to each climatic zone studied.

This study adopts static assumptions for internal loads and usage schedules to ensure consistent comparison across regions. However, hotels experience dynamic fluctuations in occupancy, equipment use, and ventilation demand, which can substantially impact thermal comfort outcomes. Future work should incorporate variable occupancy profiles and stochastic modelling to more accurately reflect real operational conditions.

However, while the OAT method facilitated the isolation of individual parameter effects across multiple climates, future studies should adopt global sensitivity analysis techniques, such as the Sobol method, to better capture the interaction effects among passive design variables and provide more comprehensive optimisation outcomes.

On the other hand, Future research could benefit from incorporating computational fluid dynamics (CFD) or numerical modelling of hydrodynamic air fields, based on the Navier–Stokes equations, to evaluate airflow distribution and natural ventilation performance alongside thermal comfort outcomes.

Although the study provides valuable insights into the climatic adaptability of passive strategies, it does not encompass local construction techniques, cost-related constraints, or cultural design practices, which are critical for practical implementation. Future studies should incorporate field-based surveys to evaluate these factors and ensure that proposed optimisations align with regional construction traditions and socio-economic contexts.

It is noteworthy that the methodologies and findings of this study contribute to the broader knowledge on sustainable building design and the importance of considering passive solutions for thermal comfort optimisation. Architects and researchers worldwide can benefit from the study’s insights, particularly in regions with similar climatic conditions.

8. Conclusions

The study focused on the M’Zab Hotel as a case study, employing a calibrated model in dynamic simulation following an SA of 13 design parameters related to thermal performance. The subsequent optimisation of the reference case was conducted and compared with the results obtained through psychrometric chart prediction. The findings demonstrated a correspondence between psychrometric chart predictions and simulation outputs for both the PMV–PPD–Percentage of People Dissatisfied and adaptive comfort models, particularly concerning discomfort hours.

The optimised passive design was shown to be effective, with the discomfort hours in the optimised design ranging from 27.9% to 54.8% less than the actual design performance of the PMV-PPD model. Similarly, the adaptive model displayed 38.8% to 90.3% fewer discomfort hours than the actual design performance. These results highlight the potential improvement in thermal performance across various climates by applying multiple design parameters.

This study presents a methodological optimisation procedure for a well-known hotel in Algeria, addressing both the PMV-PPD and adaptive comfort models. The comparative analysis of the results obtained through simulation and psychrometric chart predictions adds depth to the climate analysis tool. The research underscores the need for further investigations and surveys to enhance the thermal and energy efficiency of buildings, emphasising the significance of passive and active solutions in diverse climatic conditions.