Towards Integrated Design Tools for Water–Energy Nexus Solutions: Simulation of Advanced AWG Systems at Building Scale

Abstract

This study investigated the integration of advanced Atmospheric Water Generators (AWGs) within the design process of building energy systems, focusing on the water–energy nexus in the context of a real-life hospital building. It is based on a simulation approach, recognised as a viable means to analyse and enhance AWG potentialities. However, the current state of research does not address the issue of AWG integration within building plant systems. This study contributes to fill such a research gap by building upon an authors’ previous work and proposing an enhanced methodology. The methodology describes how to incorporate a multipurpose AWG system into the energy simulation environment of DesignBuilder (DB), version 7.0.0116, through its coupling with AWGSim, version 1.20d, a simulation tool specifically developed for atmospheric water generators. The chosen case study is a wing of the Mondino Hospital in Pavia, Italy, selected for its complex geometry and HVAC requirements. By integrating AWG outputs—covering water production, heating, and cooling—into DB, this study compared two configurations: the existing HVAC system and an enhanced version that includes the AWG as plant support. The simulation results demonstrated a 16.3% reduction in primary energy consumption (from 231.3 MWh to 193.6 MWh), with the elimination of methane consumption and additional benefits in water production (257 m³). This water can be employed for photovoltaic panel cleaning, further reducing the primary energy consumption to 101.9 MWh (55.9% less than the existing plant), and for human consumption or other technical needs. Moreover, this study highlights the potential of using AWG technology to supply purified water, which can be a pivotal solution for hospitals located in areas affected by water crises. This research contributes to the atmospheric water field by addressing the important issue of simulating AWG systems within building energy design tools, enabling informed decisions regarding water–energy integration at the project stage and supporting a more resilient and sustainable approach to building infrastructure.

1. Introduction

There is a recognised nexus between energy and water, denoting an inseparable interconnection between these two topics [1]: energy production requires substantial water resources, while water management is energy-intensive [2]. Historically, development policies addressed these two themes separately; however, today, it is mandatory to adopt an integrated perspective [3] to enhance the effectiveness of solutions and reduce resource consumption. This integrated approach should be applied not only at a macroscopic scale—encompassing significant energy and water national infrastructures—but also at the urban level, particularly in the design and refurbishment of buildings. Simultaneously addressing the blue and green factors during the construction design phase represents a strategic means to optimise both resources and efficiency.

The building sector—residential and industrial alike—is one of the most energy-demanding, accounting for approximately 40% of total energy consumption, primarily due to heating and cooling systems [4]. Significant efforts are currently being made to reduce such consumption [5]. It is worth noting that, even when energy is supplied from renewable sources, an unlimited increase in its use may still contribute to global warming. This is because, according to the principles of exergy analysis, once the useful part of energy is fully utilized, it eventually degrades into low-grade heat, which accumulates in the environment, potentially contributing to a rise in ambient temperature [6]. Therefore, improving energy efficiency can be recognized as one of the most strategic and effective approaches to counteract global warming [7]. In this context, Heating Ventilation Air Conditioning (HVAC) systems play a key role. Solutions based on energy cascades—using multipurpose machines capable of simultaneously delivering multiple useful effects with the same energy input—can enhance the overall energy efficiency of the system while also reducing raw material consumption [8].

Additionally, in an urban environment, the provision of drinking water requires enhanced resilience [9], as the risk of adverse issues occurring in water supply systems is steadily rising [10]. This calls for strategies that go beyond better source management, promoting distributed solutions capable of mitigating periods of water stress [9]. Water-related issues are becoming increasingly pressing, not only due to worldwide development driving higher consumption and subsequent contamination of water sources, but also because of the growing impact of global warming. The latter phenomenon, which is exacerbated by the continuous increase in energy consumption, has various effects on the water cycle. The residence time of water vapour in the atmosphere, as well as its global content, are increasing in most regions [11]. The system global energy is higher, causing an increase in extreme events, such as droughts and floods [12]. Between 2001 and 2018, water-related disasters accounted for 74% of all natural disasters [13], while the United Nations estimates the percentage up to 90% [14]. During 2022, Europe faced a severe drought period, with 64% of its territory under alert/warning conditions due to dryness [15], while during 2023, droughts alternated with floods, the latter affecting 1.6 million people [16].

To mitigate this scenario, various actions should be carried out to achieve a mosaic solution comprising better management, wastewater treatment recycle and reuse, aqueduct revamping, and unconventional source exploitation [17,18]. Among the non-traditional sources, atmospheric water has lately been rediscovered as one possibility to address water scarcity [17,19]. In particular, in urbanised areas, Atmospheric Water Generators (AWGs) can enhance resilience, offering a means to locally produce water [20]. The most mature technology for atmospheric water harvesting is based on Vapour Compressor Refrigeration Cycles (VCRCs). AWGs encompassing such a principle are recognised as the most compact and scalable solution, able to provide from tens to thousands of litres a day [21]. Moreover, they are proved to produce the cleanest water [22], given that they are equipped with air filtration [23] and built with materials that do not release pollution [24]. Nevertheless, in comparison to other techniques, they have the highest energy consumption [22], which is one of the main drawbacks slowing their diffusion. To overcome this issue, there are successful efforts to combine them with renewable energy sources, and to enhance their design to create multipurpose, integrated machines [22]. The latter solution is particularly interesting because it encompasses the water–energy integrated approach, offering a technology able to provide simultaneously, and with the same energy input, water harvesting, heating and cooling [8]. They are based on the same principle as VCRCs machines, which cool a flux of environmental air below its dew point, causing the condensation of part of its vapour content. The key difference lies in their smart design, which allows users to benefit from the side effects of the atmospheric water harvesting through the VCRC process, namely a flux of dehumidified and cooled air and a flux of heating energy (extracted from the air to condense the vapour).

When such a multipurpose machine is smartly inserted into building design it can yield significant savings. For example, in a real-life installation in a worker village in Dubai [25], one of these AWGs was used to upgrade the existing HVAC plant. In addition to producing drinking water, it also provided heating and cooling—using the same energy input—thereby increasing the overall efficiency of the system. An economic evaluation of the potential energy and water savings indicated a payback period of just two years, including maintenance costs. These advanced AWGs are therefore recognised as promising solutions for sustainable applications [22,26] as the energy used for water harvesting also generates useful thermal effects, improving the efficiency of existing plants, thus reducing overall energy consumption. To maximise these benefits, it is crucial to ensure proper integration of such systems within the building design. It is clear that, in order to incorporate AWG technologies—along with other HVAC solutions—into a building, designers must be equipped with appropriate tools that allow their inclusion among the available plant system options.

The authors started to address such a topic in a previous research work [27], studying how to include an advanced AWG in a widespread commercial software, DesignBuilder (DB) [28], which is a dynamic energy simulation graphical tool based upon Energy Plus [29]. That research presented a methodology that aimed to combine the results coming from another dynamic simulation tool, AWGSim [30], which is a tailored software describing the AWG behaviour, with the commercial software. The case study focused on the integration between a building—housing inverters and a transformer dedicated to a 2 MW photovoltaic field—and a multipurpose AWG. This device provided water for panel cleaning and hydrogen production, as well as cooling power for the building itself, to dissipate heat generated by the electronic components.

The results were encouraging, as they demonstrated the feasibility of modelling the AWG cooling behaviour within the energy simulation tool. This made it possible to compare two different configurations of the building cooling system: a traditional setup and one incorporating the multipurpose machine in order to determine which was more effective under identical environmental conditions. The most interesting outcome—beyond energy savings provided by the AWG solution—was the possibility of conducting assessments during the design phase of the structure.

The current research builds on that work by broadening the scope of the analysis and addressing some of its key limitations—namely, the simple geometry and plant system of the building model and the sole consideration of the AWG cooling capability. The aim of this paper was to address the open questions raised in the previous work, particularly regarding the extension of the analysis to more complex scenarios. It also seeks to contribute to bridging the research gap related to the lack of design tools that enable practitioners to incorporate, at the design stage, solutions that address the water–energy nexus at the building scale.

To this end, this paper begins with a brief overview of advanced multipurpose AWG technologies, followed by a literature review that highlights the main research gaps. The proposed methodology is then presented and applied to a real-world case study involving a hospital building. The simulation results are subsequently discussed in detail, along with the limitations of the research. The paper concludes with insights into potential future developments and a summary of the main findings.

This study contributes to the field by extending the analysis to more complex building scenarios, refining the methodology for simulating the behaviour of multipurpose AWG within a widely adopted design tool and supporting the integration of water–energy nexus solutions at the building scale during the design stage.

2. Short Description of Advanced Integrated Multipurpose AWGs

There are many techniques for collecting atmospheric water from the air, which can be classified in different ways. For example, Tu et al. [31] distinguished among integrated systems (e.g., machines collecting water and producing cooling, or desiccant based heat pumps), direct harvesting methods (e.g., fog or dew water capturing), and vapour concentration (e.g., systems embedding selective membrane or vapour absorption–desorption cycles). Nevertheless, one of the most common classifications is based on energy consumption [32]. Accordingly, AWGs are referred to as “active” when they require high-grade energy to operate and “passive” when natural thermal gradients and/or solar thermal energy are sufficient to drive the process. There is a steady stream of notable review papers [21,22,32], providing comprehensive overviews of the possible harvesting methods, including also the ancient ones [33]. The aim of this work is not to provide yet another review of existing techniques but to contribute to the field by proposing a methodology for integrating advanced multipurpose AWGs into a design tool for building-scale energy systems, thereby addressing the energy–water nexus. Accordingly, the purpose of this section is to offer a brief description of these AWG systems in order to support a better understanding of their operating principles and potential applications. These machines are recognised as effective means to enhance atmospheric water sustainability [34] and are particularly well-suited to address the energy–water nexus, as previously mentioned.

Advanced multipurpose AWGs can be defined as integrated active systems and are developed from the VCRC-based approach with the aim of overcoming the energy-related limitations of that technology. Specifically, VCRC-based AWGs harvest atmospheric water by cooling air below its dew point through a refrigerant evolving in a compressor-driven reverse cycle. One of the earliest examples of this approach dates back to the late 1960s [35]. The evolution of compression-based cooling techniques eventually enabled engineers to design the commercially viable AWGs of the current era [36]. The working principle of VCRC-based AWGs can be summarised as follows. An environmental air flux is moved by fans into a heat exchanger (called the evaporator), where it comes into thermal contact with a refrigerant. This fluid evaporates at a pressure corresponding to a temperature below that of the air, and in particular, well below its dew point.

The refrigerant, in order to change phase, absorbs part of the heat from the air, causing the air temperature to decrease below its dew point. In this way, the water vapour content in the air can no longer remain in the gaseous phase and part of it is forced to condense. Then, the air is released into the external environment at a certain distance from the inlet section to avoid air short-circuits.

Once entirely in the gaseous phase, the refrigerant is compressed to a higher pressure corresponding to a temperature above that of the environment. The next step consists of the condensation of the fluid by means of a second heat exchanger, where external air, moved by other fans, extracts heat from the refrigerant until it rejects the same amount of heat absorbed during the evaporation stage and added by the compression work. Finally, an expansion valve reduces the fluid pressure to the evaporation isobar. The process requires energy not only to operate the compressor but also the fans.

It is worth remembering that, in order to produce meaningful quantities of water—ranging from tens to thousands of litres per day—impressive masses of air must be moved. This is because the water vapour content of air is on the order of grams per kilogram [8]. Moreover, air filtration is one of the recognised methods to prevent severe pollution of the condensate, and this stage cannot be applied to systems that do not employ forced ventilation [23].

It can be said that atmospheric water harvesting, in meaningful quantities, in a comparatively small space, and while avoiding pollution issues, often requires external energy [37]. Nevertheless, it is possible to minimise the energy footprint, not only by employing renewable energy sources and/or desiccant materials [38], but also by applying the VCRC approach in a smart way, through the use of multipurpose integrated AWG machines [26]. This kind of system represents an evolution of the VCRC approach, not merely pursuing condensation energy minimisation by itself, but rather adopting a holistic strategy aimed at maximising the overall efficiency of the entire process. The energy input is thoroughly exploited to obtain the maximum useful effects from the reverse cycle.

The working scheme, represented in Figure 1 [25], of this kind of AWG is similar to that described for VCRC-based systems but it presents two substantial differences:

• The machine is designed to allow users to channel the dried and cooled air exiting the evaporator so that it can be used as primary air or as a support for air conditioning systems.

The condensation stage can be carried out using a plate heat exchanger, which is specifically designed to directly heat domestic water or another thermal vector, thus effectively exploiting the condensation heat instead of simply disposing of it into the environment.

The integrated AWG system used for the current research is similar to that described in [25] and in [27]; therefore, it comprises the following:

• Finned-tube evaporation coil and a heat recovery system, in food-grade material;
• Heat recovery system, which pre-cools the inlet air flux and recovers the cooling power from the same air flux exiting the evaporation coil;
• Evaporation fans, capable of conveying an average of 10 thousand cubic meters of air per hour across the heat exchanger, operated by the machine control system;
• Screw compressor with a declared cooling capacity at standard conditions of 30 °C and 70% relative humidity of 100 kW;
• Plate-coil condenser, for domestic water heating;
• Finned-tube condenser for condensation excess removal, when exceeding the required load;
• Condensate collection basin and water treatment unit in food-grade material.

A more articulate description can be found in [25].

This kind of system is particularly suitable for integration into buildings, whether residential, or industrial, where, in addition to water, both cooling and heating are simultaneously required. Regarding quality, which is one of the recognised important aspects of atmospheric water [39], levels comparable to bottled water for human consumption were achieved [40], while for industrial uses, purities similar to those required for electrolysis and photovoltaic panel cleaning were obtained [27].

Integration with other HVAC systems is of utmost importance in order to achieve the aforementioned advantages; thus, a building design tool that allows designers to study possible plant configurations including such systems is fundamental if the energy–water nexus is to be addressed at the building scale.

In the next section, the research gap concerning the above topic, based on a literature review, is presented.

3. Literature Review and Research Gap Identification

Research interest in the field of atmospheric water has been continuously rising starting from the first decade of the 2000s, as demonstrated by the increasing number of published articles [20], lately culminating in the foundation of the International Atmospheric Water Harvesting Association [41]. Predicting the behaviour of AWGs, by means of models, is considered beneficial for the atmospheric water field [42] and numerous researchers developed their own models and incorporated them into their studies. For example, Milani et al. presented a TRNSYS model representing a solar assisted AWG employing a desiccant wheel. The model was validated with experimental data and used for subsequent analysis in different climates [43]. Kabeel et al. [44] used a three-dimensional computational fluid dynamics (CFD) model to simulate the fluid flow region of their solar-based thermoelectric AWG. The model was used to predict behaviour in three different climatic conditions during summer. Karimi et al. [45] proposed a novel absorption-based AWG employing a two-stage modelling approach: first, the system was thermodynamically simulated in Engineering Equation Solver (EES) to assess the effects of seven design parameters; next, the EES outputs were mapped into MATLAB via artificial neural networks and subjected to a tri-objective genetic algorithm optimisation. Finally, correlations between weather data and system performance were derived from the optimised model. Barletta et al. [42] developed three machine-learning regression models—Support Vector Regression, Gradient Boosting Regression, and a Multilayer Perceptron—trained on an experimental dataset to predict water yield. The Multilayer Perceptron proved the best performer (89% accuracy) and was used for predicting the water harvesting potential in a particular climate. Raveesh et al. [46] conducted a parametric analysis of a VCRC AWG, adopting a physically based modelling approach employing MATLAB Simscape. The model was validated by means of a real-life AWG and allowed the researchers to find important relations with operation parameters like the evaporation temperature and the air flux. Ayodha Ajiwiguna et al. [47] used the open-source software DWSIM to reproduce the behaviour of a small VCRC AWG and found that, with a cooling capacity of 0.3 kW, it was possible to obtain maximum 0.285 L/h when environmental conditions were 30 °C and 80% of relative humidity, r.h.

Pokorny et al. [48] built a hybrid AWG prototype, encompassing a desiccant wheel and a VCRC, and tested it in a climatic chamber. The authors used experimental data to develop a non-linear regression model for estimating the AWG performance in arid climates, taking into account temperatures and r.h. Ansari et al. [49] developed a neural network model to represent the behaviour of a commercial VCRC-based AWG. Although not physically based, this data-driven approach could become increasingly valuable as more experimental performance data become available. Ferrari et al. [50] developed an Energy Management System (EMS) for a smart grid, including an AWG whose behaviour was implemented using the MATLAB-Simulink environment. The EMS-based approach demonstrated both economic and environmental advantages over standard management, highlighting the benefits of integrated layouts—especially with energy storage—for enhancing renewable energy exploitation and supporting the energy transition through flexible AWG systems.

The various authors addressed the theme of AWG behaviour prediction, employing different approaches but always achieving encouraging results. The valuable insights gained from previous research suggest the need for reliable simulation tools to improve the effectiveness of AWGs by enabling a better understanding of their behaviour across different climates and the optimization of their operating conditions. This need is particularly relevant in scenarios where AWGs operate in synergy with existing HVAC systems, supporting their cooling and heating functionalities. In this case, the ability to simulate their performance within building design tools becomes essential for supporting informed design decisions and system integration, thereby addressing the water–energy nexus in a tangible way. Nevertheless, such integration represents a current research gap, as no previous efforts to include AWGs in building design tools had been reported prior to an initial study recently presented by the authors [29], as described in the Introduction. The current work represents a further advancement in this field, aiming to overcome some of the limitations of the previous research and to contribute to bridging the identified gap.

In particular, the objectives of this study are:

• To develop the methodology presented in [29] in order to apply it to a more complex case study—a wing of a real-life hospital characterized by elaborate geometry and an HVAC system, which includes a chiller and a multipurpose methane boiler;
• To enhance the validation of the building dynamic model by employing real-life energy consumption data over an extended period;
• To improve the integration between the two software tools by including the AWG heating capabilities and by validating the results through output comparison.

The next chapter provides a brief summary of the methodology presented in [29], with particular emphasis on the enhancements required to address the complexities introduced by the case study.

4. Methodology

In their previous research [29], the authors developed a methodology to integrate AWGSim results into DB as a first step towards incorporating AWG evaluation in the early stages of building design. The need for this approach arises from the fact that, despite extensive DB libraries—including construction materials, HVAC configurations and components, occupancy schedules, ancillary equipment, and more—it does not include anything similar to an integrated AWG system.

It is worth noting that both tools are dynamic simulation software. AWGSim is a physically based simulation tool specifically designed to model the integrated behaviour of AWG systems, using outdoor air total pressure, temperature, and relative humidity as input forcing variables. The tool applies the principles of energy and mass conservation, and allows users to analyse both the system transient response and steady-state operation. Each main component of the AWG can be included in the model based on its operational and heat exchange characteristics—for example, compressor performance under different conditions (according to standards), the geometry and exchange coefficients of heat exchangers, fan behaviour in terms of airflow and energy use, control logic, and more. It is important to note that the simulation intentionally excludes peripheral energy users, such as electronic control boards, power conversion devices, display systems, and structural thermal losses. This choice is consistent with the objective of isolating the thermodynamic core for high-fidelity modelling. Nevertheless, based on real-life tests, these auxiliary energy consumptions have been found to be negligible for AWG units of the size considered in this study (cooling capacity: 100 kW). If needed, such contributions can be easily incorporated into the simulation by introducing dedicated ancillary components.

A brief description of the main features of AWGSim can be found in Appendix A of [25].

DB, version 7.0.0116, is a widely used graphical tool that integrates EnergyPlus (version 24.2.0), which is a simulation engine for modelling building energy performance, and it allows designers to create a building model, taking into account the following:

• The thermal properties of building structures (e.g., transmittance, linear thermal bridges, thermal conductivity, specific heat, and density);
• Hourly profiles of indoor heat gains from occupants and equipment, including zone-specific activity templates and schedules;
• Hourly outdoor environmental data (temperature, relative humidity, solar irradiance, wind direction, and wind speed);
• The dynamic effects of indoor thermal inertia;
• Detailed HVAC component specifications and their time-dependent behaviour and energy performance;
• Hourly patterns of natural ventilation;
• Time-dependent shading effects.

In the current paper, the aforementioned methodology is summarised and enhanced to handle more complex cases, making greater use of DB plant configuration capabilities to simulate both the heating and cooling capacities of the AWG system.

Case study choice. This methodology step concerns the selection of appropriate applications that can benefit from the integrated water–energy approach. In particular, the choice concerns buildings that require high-quality water and at least one of the other useful effects of advanced AWG and are located in zones where traditional sources can be affected by water stress.

In this study, to increase the water–energy analysis, a building needing not only cooling but also heating was chosen. Hence, the relevant portion of an important hospital located in Pavia, Italy (Mondino Hospital) was selected. The analyses concern the inpatient ward sector, called “Degenze Nord-Est” (DNE), as it is occupied by patients and hospital staff throughout the entire day and requires a fresh air supply and Domestic Water Heating (DWH). It is worth mentioning that, in a hospital, high-quality water is needed not only for drinking purposes, but also for medical activities.

Actually, in Pavia, water stress has not, up to now, affected the hospital water supply; however, given the current trend of extreme events, it is not such a remote possibility that some issues could occur in the near future. As for the current case study, part of the produced water was used for photovoltaic panel cleaning—an activity more negatively influenced by the seasonal water crisis currently affecting the area.

Data collection and adaptation. This step concerns the data collection of the building characteristics, HVAC system and ancillary devices of the study case. Such types of data are essential for the construction of a reliable DB model and should be sourced from official design documentation. In the context of existing buildings—for example, when retrofitting or revamping projects are involved—it is advisable to include real-life energy consumption data, when available, to enable the validation of the model against empirical evidence. This empirical grounding enhances the robustness and credibility of the simulation process, ensuring that the model reflects not only theoretical specifications but also actual operational performance.

For the current study, it was possible to collect real-life data concerning the main components of the building and the HVAC system. In particular, building information was gathered directly from the following:

• The “Relazione di Legge 10” (Law 10 Report), a document containing comprehensive data on HVAC systems, including model, size, operating cycles, and the thermal characteristics of the materials used in the building construction. Such a document is required at the start of the design process, outlining the measures taken to comply with Italian Law no. 10/1991 on energy consumption in buildings;
• Original drawings, such as plans, elevations, and sections, which provide all the geometrical data related to the building.

For the HVAC system, the following were available:

• Plant blueprints;
• A list of the main components with their characteristics;
• Real consumption data for the year 2023.

Moreover, an onsite survey was conducted to verify and confirm the key characteristics of both the building and the HVAC system.

For the photovoltaic field, intended for the produced water use, data about its installed power, orientation and size were collected directly from the technical documents of the real installation. Also in this case, an onsite survey was carried out.

Weather data and simulation parameters. Weather data are the primary driving variables for both the DB and AWGSim simulations. Specifically, AWGSim requires only the total atmospheric pressure; dry-bulb temperature, t; and relative humidity, r.h.; whereas DB additionally requires wind speed and solar radiation data. Given the availability of real energy consumption data for the year 2023, weather data from the same year were used as input for the simulations. In this case, the purpose of using weather data was not to provide a general climatic representation of the region, but rather to verify whether the model response to the specific climatic forcing was comparable to the actual behaviour of the building. This alignment ensured consistency between external conditions and real building performance, allowing the authors to validate the hospital model and establish a reliable benchmark for subsequent simulation scenarios.

Furthermore, the weather data for 2023 were not based on statistical averages but were directly sourced from the weather station database of the “Agenzia Regionale Protezione Ambientale” (ARPA) [51] located in Pavia. From this database, an hourly dataset for the year 2023, comprising dry bulb temperature, relative humidity, solar irradiation and wind speed was arranged as input for both DB and AWGSim.

AWG Sizing, Modelling and Validation in AWGSim. Based on the water demand and the HVAC system characteristics determined in the previous steps, it is possible to define both the size of the multipurpose AWG and the required functionalities. At this stage of the methodology, it is crucial to establish the primary objective of the project. In the present case, water production was identified as the main priority. Consequently, the same machine as that employed in [27] was selected.

Once the AWG unit was defined, the AWGSim tool was used to simulate its behaviour. Each main component of the AWG is incorporated into the model according to its operational behaviour and heat exchange characteristics. This includes, for example, the compressor performance under varying operating conditions (as defined by relevant standards), the geometry and thermal exchange coefficients of the heat exchangers, the fan performance in terms of airflow and energy consumption, and the associated control logic.

Following the simulation, the model was validated—according to the approach described in [27]—against real-world data obtained from the operation of the same AWG unit. As previously mentioned, the advanced AWG used in the present study (described in Section 3) is identical to that analysed in the earlier work [27], where a detailed validation of the AWGSim model was carried out. Given the excellent agreement with experimental data—showing an average deviation of less than 5% on key performance outputs—the same model was adopted here, and further re-validation was considered unnecessary.

Building and plant model import in DB. DB enables detailed modelling of building thermal properties, internal and external heat gains, and a precise representation of HVAC systems. Time-dependent behaviour and energy performance of the equipment are made possible through dynamic simulations.

For this kind of application, the principal applied capabilities include zone-level thermal modelling, control logic customization, temperature and humidity tracking, and HVAC system scheduling. Key input parameters included the building envelope characteristics (e.g., geometry, transmittance values, and thermal mass), HVAC equipment specifications (e.g., boiler nominal efficiency, chiller COP, fan and pump performance curves), indoor temperature and humidity setpoints, and domestic hot water (DHW) consumption profiles. In this study, all input data were sourced from the “Relazione di Legge 10”, original architectural and mechanical drawings, onsite surveys, and technical office documentation.

The modelling process began with the development of the building geometry, including surrounding obstructions, in order to accurately simulate shading effects. Thermal properties were then imported, followed by the definition of HVAC systems. These were selected from the DB component library by choosing the most appropriate modules and adjusting operational schedules, performance parameters, and control strategies as needed.

AWG implementation in DB. The AWG functionalities were implemented in DB using standard library components such as chillers, condensation coils, and evaporation units. These elements were carefully sized and their operational schedules configured to realistically replicate the AWG thermal and moisture behaviour.

Unlike the previous case study, where only the cooling capacity of the AWG was modelled, the current scenario required a more comprehensive and detailed representation. As a result, special attention was dedicated to its schematic setup, which is described in further detail in the next section.

Building-plant and AWG DB model validation. After importing into DB the geometrical and thermal features of the building and the configuration of both the plant layout and the AWG, according to the previously described approach, the model was validated in two stages. The first addressed the performance of the existing HVAC system, while the second focused on the behaviour of the integrated AWG system.

For the first stage, real energy consumption data from the hospital HVAC system for the year 2023 were available, allowing for a data-driven validation process. To ensure consistency between the simulation results and observed performance, a calibration procedure was implemented through an iterative simulation process. This involved the progressive refinement of model parameters to minimise the deviation between simulated outputs and the scaled real-world energy consumption data. Simulations conducted throughout the calibration process employed the same weather dataset referenced earlier, corresponding to the monitoring period associated with the 2023 energy bills.

In the second validation stage, the AWG model implemented in the DB environment was assessed by comparing its hourly simulation outputs with those obtained from AWGSim. Both simulations used environmental input data from the ARPA database for the year 2023, covering the period from May to September. This comparison enabled the validation of the integrated AWG model against the reference behaviour simulated via AWGSim under varying climatic conditions.

The results of both validation phases are presented in the following section. Figure 2 presents a flow chart of the methodology.

5. Case Study

The building selected for the current study, as previously mentioned, is a significant section of the Mondino Hospital, designated to accommodate both patients and hospital staff throughout the day. Figure 3 illustrates the entire structure, with the analysed section—DNE—enclosed in the dashed red line; this section can host up to 69 patients. The building dates back to the early 2000s and reaches six storeys at its highest points.

5.1. Building Characteristics

The analysed building portion has a net internal area of about 1845 m².

Table 1. Room types and floor area.

Room Type	Quantity	Internal Floor Area [m2]
double	25	19
single	19	11
bathroom	44	4
toilet	8	3
nurse’s station	4	17
kitchen	3	9
doctor’s office	3	11

shows the room types and their internal floor area:

Additionally, there are medical consultation rooms, corridors and storage rooms.

The building is made of masonry and has reinforced concrete supporting structures. The main characteristics of the external opaque components and windows are summarized in

Table 2. External structure thermal characteristics.

Element	Transmittance
Insulated masonry walls	0.42 W/(m2K);
Reinforced concrete supporting walls	0.45 W/(m2K)
External doors	3.23 W/(m2K)
Windows	3.23 W/(m2K)

.

Figure 4 shows the first-floor plan of the DNE wing. It is worth noting that the internal area is mainly composed of patient bedrooms and corridors. The layout of the other floors is largely similar, with the only difference being that the second, third, and fourth floors have, respectively, one, two, and three fewer bedrooms, as can be seen in Figure 2.

5.2. HVAC System Description

The real-life HVAC system of the hospital serves the DNE area by performing the following functions:

• Supplying 11,015 m³/h of primary fresh air;
• Heating the indoor environment to maintain a constant temperature of 22 °C from November to March;
• Cooling the indoor environment to maintain a constant temperature of 26 °C from May to September;
• Providing domestic hot water, heating 22,500 L per day from 12 °C to 40 °C.

These functions are achieved through the following equipment and systems:

• Fresh air supply is managed by a dedicated system consisting of air ducts, fans, and an air treatment unit (UTA1) exclusively serving the DNE sector. UTA1 includes two heating coils with capacities of 136.9 kW and 56.6 kW, and one cooling coil with a capacity of 157.8 kW;
• Space heating is provided by three methane boilers, each with a thermal capacity of 581 kW and an efficiency of 92%, according to their technical data. Heat is distributed through fan-coil units across the building;
• Cooling is delivered by two chillers, each with a cooling capacity of 567 kW and an Energy Efficiency Ratio (EER) of 3.23. These chillers, in conjunction with fan-coil systems, serve the entire building;
• Domestic hot water (DHW) is also heated using the same methane boilers and stored at 60 °C in dedicated reservoirs to ensure hygiene standards. Each day, approximately 22,500 L of hot water are distributed to users at a temperature of 40 °C.

A simplified schematization of the existing HVAC system is shown in Figure 5.

It is worth noting that the hospital maintained full occupancy during 2023.

Based on technical schematics, on-site inspections, and available records, the DNE section was estimated to account for approximately 13.5% (±2.5%) of the hospital’s total energy use. This assessment was obtained by rescaling global energy consumption data according to the building’s geometric features and HVAC operating schedules. Specifically, the DNE area represents about 11% of the hospital’s total opaque envelope surface—relevant to transmission heat losses—and around 16% of the gross indoor volume—impacting ventilation loads. Taking into account both the thermal characteristics and spatial distribution, heat losses due to envelope transmission and ventilation within the DNE were estimated to contribute, on average, the aforementioned share of total energy demand for both heating and cooling. Consequently, electricity consumption for cooling in the DNE is estimated at 13.5% ± 2.5% of the total cooling system electricity use, based on monthly utility bills. Likewise, methane gas use for heating in this section represents the same proportion of the hospital overall heating gas consumption. Finally, all DHW consumption was assigned to the DNE sector, since usage in other hospital areas was deemed negligible.

5.3. Building Modelling in DB

Following the methodology described in the previous section, the DNE building and its existing HVAC system were imported in DB taking into account its geometry and envelope features, as well as the geographical and environmental conditions of Pavia by means of the ARPA database.

5.3.1. External–Internal Frame Modelling

Figure 6 presents the three-dimensional representation of the building model developed in DB. While the model is visualised using false colours for illustrative purposes, the actual radiative envelope properties are fully accounted for in the simulations. The specific sector under investigation is shown in grey, whereas the other areas—displayed in brown and pink—are included in the model to represent thermal exchanges and shading effects accurately. Figure 6C shows a model of the indoor configuration of one of the DNE floors.

5.3.2. HVAC System Modelling

As mentioned in the methodology section, the existing plant serves the entire hospital; thus, for the HVAC modelling, it was re-scaled to represent the equivalent system, serving only the DNE sector. Specifically, the real-life heating configuration was scaled down to a single 250 kW methane boiler. Likewise, the cooling configuration was re-scaled to a single chiller with a cooling capacity of 144 kW. It is worth noting that the air handling unit was not rescaled, as the real-life one is specifically dedicated to the DNE area.

Figure 7 shows the plant schematics used to model the configuration of the real-life equivalent system, which includes the following components.

Figure 7a: An air handling unit (UTA1) dedicated to the primary air supply, comprising the following:

• Two heat exchangers connected to the methane boiler for air pre-heating and post-heating;
• A cooling coil linked to the chiller for air cooling and dehumidification;
• An air humidification system;
• Two fans for supplying fresh air to the DNE area and extracting return air.

Figure 7b: A DHW system, consisting of a methane boiler, a 3000-L hot water storage tank, and a water distribution unit.

Figure 7c: Fan-coil units used for space heating and cooling, comprising the following:

• A fan for air circulation;
• A heating coil connected to the methane boiler;
• A cooling coil connected to the chiller.

To model the thermal and hygrometric loads within the building, the following key assumptions were made, in compliance with the DNE documentation:

• An indoor heating setpoint temperature of 22 °C;
• An indoor cooling setpoint temperature of 26 °C;
• A target indoor relative humidity range of 40–50%;
• The number of occupants per bedroom (one or two) and their daily activities;
• The presence and activities of staff (e.g., doctors and nurses) in each room;
• A natural air infiltration rate of 0.5 air changes per hour;
• The characteristics of lighting systems;
• The use of electrical and clinical equipment by staff in various rooms;
• Cooking activities carried out in the nurses’ stations.

Dynamic simulations were conducted using a 15 min time step (i.e., four time steps per hour), which ensures a good balance between accuracy and computational efficiency. This resolution was sufficient to capture the transient behaviour of HVAC systems and moisture dynamics throughout the building. Primary simulation output variables were the time-dependent consumption of natural gas (boiler) and electricity (chiller), and the temperature, relative humidity, and mass flow rate of the treated airflow for the UTA1.

Table 3. DB model of the HVAC: equipment specifications and key settings.

Equipment		Technical Features and Control Parameters
UTA1 fan		Fan type: variable volume; airflow rate: 4.167 m3/s
UTA1 pre-heating coil		Control variable: T; heating capacity 136.9 kW; inlet waterflow T: 75 °C; outlet waterflow T: 60 °C; inlet airflow T: −5 °C; outlet airflow T: 22 °C
UTA1 cooling coil		Control variable: T; waterflow:7.53 kg/s; airflow rate: 4.167 m3/s; inlet waterflow T: 7 °C; inlet airflow T: 32 °C; outlet airflow T: 14 °C
UTA1 post heating coil		Control variable: T; heating capacity 56.6; inlet waterflow T: 45 °C; outlet waterflow T: 40 °C; inlet airflow T: 14 °C; outlet airflow T: 26 °C
Chiller		Chiller type: Electric EIR Chiller Scroll Trane CGWD; condenser type: cooled by water; maximum cooling power: 144 kW; reference EER:4; regulation scheme: modulation following the setpoint
Heat disposal fan		Fan type: single speed; nominal capacity: autosize; waterflow rate: autosize; waterflow inlet T: 80 °C; airflow inlet T: 38 °C
Boiler		Type: new style low temperature boiler; fuel: natural gas; maximum heating power: 250 kW; reference efficiency 92%; heating circuit waterflow rate: autosize
DHW tank		Tank type: mixed; volume: 3 m3; water set point T: 40 °C; user side waterflow rate: 0.26 kg/s; recovery circuit side waterflow rate: 10 kg/s
Fancoil		Type: variable waterflow rate fun; waterflow rate: autosize; airflow rate: autosize
Fancoil cooling coil		Type: water; inlet waterflow T: 7 °C; inlet airflow T: 27 °C; outlet airflow T: 19 °C; waterflow rate: autosize; airflow rate: autosize
Fancoil heating coil		Type: water; inlet waterflow T: 80 °C; outlet waterflow T: 70 °C; inlet airflow T: 20 °C; outlet airflow T: 35 °C; performance definition method: nominal capacity; heating power capacity: autosize

presents a summary of the equipment selected in DB to represent the main HVAC components, including their key features and settings used for the model.

5.3.3. Existing Building and HVAC Plant Model Validation

After importing the building and HVAC system characteristics, the model was calibrated and tested as described in the methodology section. For the year 2023, the month of July was selected as representative of summer conditions, while January was used to assess winter performance. The calibrated model was then tested against additional months to evaluate its generalisation capacity and accuracy across different conditions. Subsequently, the validation process yielded the following results.

With regard to heating, the calculated consumption ranged from 11.7% to 13.5% of the actual real-life values. This outcome falls within the acceptable range, which was estimated as 13.5% ± 2.5% of the real-life values, as described in Section 5.2. Figure 8 shows the monthly differences between the calculated methane consumption and the estimated range of the real-life consumptions.

It is worth noting that during the autumn–winter months, heating is required for both the fan coil systems and the operation of the UTA1. Conversely, in the summer months, heating is only needed for post-heating of the air stream in the UTA1. The absolute deviation from the average values ranges between 0% and 1.8%, indicating that the model heating performance is well within the acceptable range.

Regarding electricity consumption for cooling, the simulation results ranged from 12.1% to 12.8% of the actual measured values, thus remaining well within the acceptable range. It is worth highlighting that, during the summer period, cooling is required both for the fan coil systems and for the operation of the UTA1. Figure 9 displays the monthly discrepancies between the simulated and real-life electricity use. The absolute variation, ranging from 0.7% to 1.4%, confirms that the model cooling performance is well within the expected tolerance.

5.4. AWG Modelling in DB

One of the main difficulties in integrating AWG modelling into DB for the current case study concerned the heating capability of the integrated machine, as it needed to serve a dual purpose: DWH and fresh air post-heating. Therefore, the AWG condenser had to be connected both to the methane boiler and to the air handling unit.

Another complication was related to the cooling effect. The integrated AWG machine was expected to provide a constant air flow of 11,015 m³/h, which, during certain periods, also contributed to covering the internal heat loads.

To incorporate all these functionalities into DB, the scheme shown in Figure 10 was developed.

An explanation of Figure 8 is provided below.

The white rectangle—outlined in a dotted yellow line—encloses the AWG representation in DB, which comprises the following:

• The AWG heat recovery system, whose effect is simulated in DB by means of the following:

  ➢

  Coil A, which models the air pre-cooling effect;

  ➢

  Coil C, which represents the air post-heating effect;

  ➢

  Chiller D, which is needed to model the cooling effect of A.

• Coil B, which models the real-life AWG evaporator, where the air flux is cooled and dehumidified;
• Chiller E, which represents the AWG Vapour Compressor Refrigeration Cycle VCRC, including the condensation plate coil used for DHW and for air post heating in the UTA1;
• Ancillary air cooled condenser F, intended for the disposal of condensation heat excess (when the thermal flux excess that required for DHW).

The rectangle outlined with a dotted red line is the DB representation of the existing DHW system. As aforementioned, it is composed of a domestic water tank and a methane boiler The potentiality of using the AWG heating capacity is simulated in DB throughout its connection to the condensation coil of chiller E and an ancillary hot water tank, called G in the schema.

The rectangle marked with a dotted blue line contains the UTA1, which is the existing air handling unit, composed of the aforementioned pieces (pre-heating coil I, cooling coil J, post-heating coil K, air humidification system L, and distribution fan M). Integration of the AWG cooling capacity into DB is achieved via its connection to coils A, B, and C of the AWG system. Air post heating through the AWG system is schematised by the connection of coil K to the ancillary hot water tank H connected to the chiller E.

It is worth mentioning that the condensation heat recovered from the AWG is used to meet both DHW and air post-heating demands. In the DB model, the system is composed of chiller E, which simulates the AWG cooling effect; tank G, which stores thermal energy for DHW purposes; and tank H, which stores heat for the post-heating of the air handling unit UTA1.

A priority logic regulates the distribution of the recovered heat:

• Heat is first transferred to tank G to satisfy DHW requirements;
• Any remaining heat is then directed to tank H for air post-heating;
• Surplus heat not used for either function is dissipated via component F.

This control strategy is implemented in DB through the configuration of component connections and operational settings, ensuring correct prioritization in line with the intended system behaviour.

Table 4. DB model of the AWG: equipment specifications and key settings.

Equipment		Technical Features and Control Parameters
Coil A		Control variable: T; waterflow:15 kg/s; airflow: 4.167 m3/s; inlet waterflow T: 1 °C; inlet airflow T: 40 °C; outlet airflow T: 2 °C
Coil B		Control variable: T; waterflow:15 kg/s; airflow: 4.167 m3/s; inlet waterflow T: 1 °C; inlet airflow T: 35 °C; outlet airflow T: 2 °C
Coil C		Coil type: electric; maximum heating power: 100 kW; efficiency: 1
Chiller D		Chiller type: DOE-centrifugal; condenser type: cooled by water; maximum cooling power: 200 kW; reference EER:3; regulation scheme: modulation following the setpoint
Chiller E		Chiller type: DOE-centrifugal; condenser type: cooled by water; maximum cooling power: 200 kW; reference EER:2; heating recovery circuit waterflow rate: 9 kg/s; regulation scheme: modulation following the setpoint
Fan F		Fan type: single speed; nominal capacity: 200 kW; waterflow rate: 5 kg/s; waterflow inlet T: 60 °C; airflow inlet T: 40 °C
Tank G		Tank type: mixed; priority in heating recovery: 1; volume: 15 m3; water set point T: 40 °C; user side waterflow rate: 0.3 kg/s; recovery circuit side waterflow rate: 9 kg/s
Tank H		Tank type: mixed; priority in heating recovery: 2; volume: 15 m3; water set point T: 40 °C; user side waterflow rate: autosize; recovery circuit side waterflow rate: 9 kg/s

presents a summary of the equipment selected in DB to represent the main AWG components, including their key features and settings used for the model.

After defining the model in DB, AWGSim was used with the weather dataset to provide the following hourly outputs:

• Water production;
• Energy consumption;
• t_A, temperature of the pre-cooled air;
• x_B, hygrometric degree, and t_B, temperature after the evaporator;
• t_C, temperature of the exiting air flux;
• Air flux;
• Heating and cooling power.

A time step of 0.1 s was used in the simulations to achieve high-precision results.

Accordingly, in DB the AWG contribution was implemented imposing the following hourly values:

➢

t_A, after coil A;

➢

x_B and t_B after coil B;

➢

t_C after coil C;

➢

v, air flux moved by the fan of the UTA1, equal to 11,015 m³/h;

➢

Temperature value of 40 °C at the condenser coil of chiller E.

It is worth mentioning that the condenser coil temperature of chiller E, representing the AWG condenser, was set at 40 °C. This value was selected to ensure consistency between the DB outputs and AWGSim results. In actual operation, the AWG produces domestic hot water at an average temperature of approximately 40 °C. By setting the condenser to this fixed temperature, the model was able to replicate the expected DHW outlet conditions. Although this assumption may reduce the ability to account for dynamic condenser behaviour, it allowed for a reliable comparison between the two simulation environments, which was a key objective of the study.

The aforementioned air thermo-hygrometric conditions were imposed in DB by means of “Setpoint Managers” (SMs), which allow users to program the output conditions of the various coil sections.

An example of data simulation results coming from AWGSim and intended for DB implementation is shown in

Table 5. Example of output results coming from AWGSim used for DB implementation.

t0	r.h.	tA	tB	xB	tC	v
°C	%	°C	°C	kg/kg	°C	m3/h
19.4	67.4	12.8	4.2	0.0061	12.5	11,015
21.9	73.9	16.0	6.8	0.0071	13.7	11,015
29.7	79.6	24.4	13.7	0.0088	24.6	11,015
29.5	71.3	22.8	12.6	0.0081	23.3	11,015

, where t₀ and r.h. refers to the environmental conditions of temperature and relative humidity, respectively, coming from the ARPA database.

It is worth mentioning that the water production and the energy consumption of the AWG cannot be imported into DB with the current methodology.

AWG Model Validation

After importing the aforementioned scheme into DB, validation of the model was carried out by means of comparison of the hourly output of the two software—AWGSim and DB itself—given the same input file.

Figure 11 shows the hourly difference, in percentage, between the AWGSim and DB results concerning the temperature at the outlet section of the integrated machine. The average error is completely negligible (<0.01%).

Similarly, Figure 12 illustrates the hourly difference, in percentage, between the AWGSim and DB results concerning the hygrometric degree at the outlet section of the integrated machine. Again, in this case, the average error is also negligible (<2%).

The above analysis demonstrates a very good agreement between the DB and the AWGSim results. The low discrepancies confirm that the approximation is acceptable, and the AWG model developed in the DB workspace is well-suited for accurately simulating the thermal behaviour of the advanced integrated AWG system.

Although the integration of the AWGSim and DB currently represents a design simulation that cannot be physically verified, it is worth noting on the one hand that the AWGSim model itself has been previously validated against experimental data with satisfying agreement. On the other hand, the building energy model in DB has been carefully calibrated using real energy consumption data from the existing hospital system.

5.5. Photovoltaic Field

In the proximity of Mondino Hospital, there is a Photovoltaic Field of 3.04 MWp. The field is composed of 11,260 polycrystalline silicon panels (dimensions 1.65 m × 1.00 m), with a peak production of 0.27 kW each. Such panels are mainly oriented towards the southwest direction. The annual production, in perfectly cleaned conditions, was estimated as 3789 MWh. The soiling efficiency loss, after 180 days, was estimated as 7%, due to the particulate presence in the city’s air.

Demineralised water is one of the best cleaning substances for panels [52]. To clean the entire solar field, considering 0.5 L of water for each m² of photovoltaic panel [53], about 9725 L of water are needed.

Figure 13 shows the efficiency loss trends due to the soiling effect:

• Without any cleaning activities (red trend);
• With monthly cleaning activities (blue trend).

These efficiency loss trends were calculated applying the DIrt method, which is accurately described in [52]. Supposing the cleaning activity, performed with demineralised water with characteristics of purity similar to those achievable with the integrated AWG [52], the yearly energy saving was calculated as equal to 191 MWh, with a water consumption of about 116.7 m³.

6. Simulation Results

The possibility of importing the advanced AWG into the simulation environment of DB enables designers to compare different plant configurations. In the present study, the authors analysed the following scenarios:

• Configuration A, representing the real-life building served by the actual HVAC system;
• Configuration B, in which the same building and HVAC system were integrated with the advanced AWG, and part of the produced water was used to enhance the efficiency of the photovoltaic field.

The simulation was performed on an hourly basis for the period from May to September 2023. This time frame was selected for two main reasons:

• Both cooling and heating demands are present in the hospital during these months;
• In Italy, water stress conditions are more severe during late spring and all summer.

6.1. Configuration A Simulation Results

The simulation results for Configuration A, obtained using DesignBuilder, are summarized as follows:

• The total methane consumption is equal to 15,467 Sm³. The gas is used to heat 22,500 L/day of domestic water and to provide post-heating in UTA1 to achieve the neutral temperature of 19.5 °C;
• The whole electricity consumption is equal to 40.4 MWh. The energy is used for the air treatments of the UTA1 and for the fan coils.

Figure 14 shows the heating power provided by the methane boiler. In this case, the graph displays a relatively constant trend from the second half of June to the first half of September. During the first 10 days of May, as well as during the final days of September, higher heating power is required due to lower outside temperatures compared to the rest of the period, resulting in greater external heat loss.

Figure 15 depicts the corresponding monthly methane consumption of the boiler. It is worth noting that the largest deviation from the average value occurs in May and amounts to 6.1%. This result reflects the fact that the majority of the consumption is due to domestic hot water production, which remains nearly constant throughout the period.

Figure 16 shows the hourly cooling power provided by the air conditioning system to maintain the indoor set-point temperature. The highest values are observed in August, followed by July.

The corresponding electricity consumption is shown in Figure 17, summarised as cumulative monthly values. According to the graph above, the highest consumption is in August.

6.2. Configuration B Simulation Results

For the configuration B, both AWGSim and DB were used. The first software, beside the values required for the DB model, provided data about water production and AWG energy consumption, which cannot yet be output by DB applying the described methodology.

The second simulation tool gave the energy behaviour of the entire building once the HVAC was integrated with the AWG.

6.2.1. AWGSim Simulation Results

The hourly step simulation of the AWG, carried out using AWGSim from May to September under the conditions defined by Configuration B for the use of the integrated machine thermal outputs, yielded the following overall results:

• A total of 257,057 L of produced water;
• A total of 91.81 MWh of electricity consumption;
• A total of 316.57 MWh of provided cooling energy;
• A total of 147.15 MWh of provided heating energy (37% of its full capability).

Figure 18 illustrates the hourly water production of the machine during the simulation period and its consumption.

It is worth noting that, on a few days in August and one day in July, water production was significantly lower compared to most other days. This anomaly is attributed to a sharp drop in the hygrometric level recorded in the weather data during those periods. It was not possible to determine whether these values were the result of a sensor error or an actual meteorological anomaly. In any case, their impact on overall performance—both in terms of water production and energy consumption—is negligible.

The water required for cleaning the photovoltaic field amounted to 45.4% of the total production.

Figure 19 shows the percentage of heating power available from the AWG actually used. It is worth noting that, due to the comparatively low needs of the hospital, the heating power of the advanced AWG is exploited only for 37% of its full capability, the latter corresponding to 397,702 kWh.

6.2.2. Configuration B: DB Simulation Results

The Configuration B simulation results obtained from DB are summarized as follows:

• Methane energy consumption set to zero because of the use of the advanced multipurpose AWG machine thermal power. The thermal power was employed for the following:
  • Covering the entire domestic water heating needs, amounting to 22,500 L/day of tap water to be heated from 12 °C to 40 °C;
  • Providing the post heating of the air flux.
• The condensation recovered thermal energy is equal to 167 MWh and the subsequent overall savings of methane is equal to 15,467 Sm³;
• Cooling energy for the most part is covered by the advanced multipurpose AWG, 317.4 MWh, and residually by the existing chiller, 34.9 MWh;
• The total electricity consumption of the integrated system, covering heating and cooling loads, can be calculated as equal to 101.9 MWh by means of the combination of AWGSim and DB outputs. Such an amount can be covered by 53.3% of the recovered energy coming from the photovoltaic panel cleaning.

Figure 20 depicts the heating power provided by the advanced AWG and by the boiler in Configuration B. The integrated machine covers all of the needs, and thus the boiler can be turned off.

Figure 21 represents the cooling power provided by the advanced AWG and by the existing chiller in Configuration B. The integrated machine covers the major part of the cooling. It is worth noting that the main support from the existing chillers occurs during July and August.

7. Results, Discussion, and Limits of the Research

By comparing Figure 14 and Figure 19 it can be observed that the cooling power provided by the combination of the existing system and the AWG is greater than that delivered in Configuration A, despite the fact that the achieved indoor conditions are nearly identical. This difference can be attributed to the distinct operational modes of the two configurations. The first minimises latent heat removal and does not produce high-quality water, generating only polluted condensate that must be disposed of. Configuration B, by contrast, is primarily designed for water production, thus focusing on vapour condensation, with its cooling effect constituting a by-product of the process. The achieved indoor conditions are almost the same, but the wet air transformations are profoundly different.

Another point worth noting is that the electricity consumption of the combined system amounts to 101.9 MWh, whereas Configuration A requires 15,467 Sm³ of methane, providing 147.150 MWh of heating energy and 40.4 MWh of electricity. In order to compare such consumptions, it is useful to convert all the terms in primary energy by means of conversion factors [54]. In particular, for the current analysis, the following conversion factors were used:

• 1.05 kWh _{primary energy}/kWh _methane [55];
• 1.9 kWh _{primary energy}/kWh _{grid electricity} [56];
• 1 kWh _{primary energy}/kWh _{photovoltaic electricity} [55].

It is worth mentioning that, on 15 December 2022, the European Community introduced a new conversion factor of 1.9 to appropriately account for the increasing share of renewable electricity production across its territory [56].

Applying the above coefficients, Configuration A has a primary energy consumption of 231.3 MWh, while Configuration B, although powered by the electrical grid, shows a 16.3% reduction, resulting in a consumption of 193.6 MWh. This reduction leads to even greater performance—55.9%—when considering that the electricity is sourced from the recovered efficiency of the photovoltaic panels, reducing the primary energy consumption to 101.9 MWh.

Figure 22 shows the above results.

However, a direct comparison based solely on energy terms is neither straightforward nor entirely fair, as Configuration B—besides providing cooling and heating—also produces over 257 m³ of high-quality water. In this case, a portion of the produced water is used for cleaning the photovoltaic panels, and the resulting recovery in efficiency can fully offset the energy consumption of the combined system. Configuration A does not offer the same advantage. In this context, it can be stated that the total energy savings achievable with Configuration B correspond to the full amount of electricity and methane required by Configuration A.

As for the photovoltaic cleaning operations, it can be said that the produced water can be effectively employed in other solar fields, as in the Pavia energy district there are currently about 270 MW of panels installed [57]. Therefore, should a water shortage occur exclusively during summertime, the produced water can be usefully employed to cover the cleaning demand of the other plants located in the proximity.

The remaining portion of the produced water can be used for drinking purposes, replacing bottled water, or for certain hospital activities. It should be noted that, in hospitals, highly purified water is of the utmost importance, as it is used for various applications ranging from instrument washing to laboratory analysis. When a reliable source is available, such as municipal drinking water, it can be produced on-site using dedicated and certified equipment, which treats large volumes of liquid, from which only a small portion reaches the required level of purity. When such a source is not available, this type of water must be purchased and transported to the site, resulting in high costs and uncertainties regarding supply consistency. In the future developments section, further considerations on this topic are presented.

At any rate, the analysis conducted by means of DB permits users to understand that if the focus is to maximise the energy savings and not the produced water, a smaller size of advanced AWG should be chosen. In particular, as the heating energy is the highest output of such a type of process, sizing the machine on such an output can assure the complete exploitation of the useful effects of the system and the energy maximum benefit.

As for the limitations of the current study, it should be noted that DB, in its customer version, is not able to provide data on water production or the energy consumption of the AWG. Comprehensive data can only be obtained through the combined use of both software tools. This methodological limitation will be addressed through a tailored script, currently under development, which aims to integrate the AWG into the DB library. This step will allow the authors to easily extend the evaluation to other climates and better assess the potential benefits of the solution in areas affected by water crises. This aspect of the ongoing research is described in the following section, which is dedicated to future developments.

Moreover, although the hospital DB model was validated against real-life consumption data, it was not possible to determine the actual single-zone cooling and heating demand. Consequently, based on billing data, an estimation of the consumption distribution was carried out using the building’s geometry as a reference.

Finally, economic evaluations of the system were not conducted. This is due to the multipurpose nature of the integrated machine, which not only provides energy savings but also yields a significant amount of water. As previously mentioned, the produced water has very promising purity characteristics, and further consideration of its value is needed. The use of atmospheric water for hospital purposes, such as laboratory analysis, represents a major advantage that requires deeper evaluation and further studies, which are currently underway.

8. Future Developments

Previous studies [27] have shown that water produced by the advanced AWG has high-quality characteristics. Due to its low conductivity, this type of water is particularly well-suited for technical applications. Hospital requirements for technical water—which can account for up to 14% of the facility total water consumption [58]—are critical, and ensuring a continuous supply is already a challenge in areas most affected by the water crisis. It should be noted that for this type of water, both quality and availability are of the utmost importance.

Future developments of the current research are focused on assessing the potential of atmospheric water as a viable solution for hospitals in regions where freshwater is scarce or highly polluted. The main focus of that future step is on quality, and it will include real-world laboratory tests on the produced water to identify the appropriate treatment processes required to meet specific purity standards.

Another important topic currently under development concerns the economic assessment, which remains one of the least explored areas in the field of atmospheric water generation. The authors are investigating the potential cost savings achievable through the integration of the advanced AWG system with the hospital HVAC plants. Clearly, an accurate evaluation of the produced water value is a crucial step; therefore, the economic analysis are closely linked to the aforementioned research step concerning the achievable water quality. Furthermore, a detailed cost analysis of the various energy vectors is essential, given the significant price fluctuations observed in recent years.

An additional development step, which is nearing completion, is the creation of a tailored script embedding specific code lines to be implemented in DB. This script will allow users to import the advanced AWG into DB, not only in terms of its energy behaviour, but also incorporating its water production and energy consumption data. The machine operation will be simulated through the interpolation of performance data—including yields, airflows, achieved temperatures and relative humidities in the treated air, and heating power—imported into the script as lookup tables. The results will be validated against outputs from AWGSim, which will serve as the benchmark.

Following the aforementioned step, it will be possible to carry out simulations of the hospital–AWG system under different climatic conditions, thereby obtaining more generalizable results. It will also be possible to conduct a more detailed analysis of critical areas characterised by prolonged droughts.

9. Conclusions

This study focuses on the development of building design strategies that address the water–energy nexus by integrating atmospheric water generation (AWG) technology into the building design process. A simulation workflow was implemented by coupling a tool based on EnergyPlus, connected to a tailored software focused on dynamic AWG simulation, AWGSim. The methodology was applied to a real-world case: the inpatient ward of the Mondino Hospital in Pavia, Italy.

The integration of an advanced multipurpose AWG system was assessed in synergy with the existing HVAC configuration, delivering cooling, dehumidification, domestic hot water, and air post-heating. The hospital model was based on detailed architectural and system data and validated against real energy consumption (average error below 2%). A real-life AWG unit was incorporated into the DesignBuilder simulation through its energy performance, derived from accurate AWGSim outputs (average error below 5%).

Using weather data from a meteorological station near the building, it was possible to carry out hourly simulations of the hospital energy performance in two configurations:

A

the current system with methane boiler, chiller, and air handling unit;

B

a retrofitted system including the AWG integrated with the existing HVAC.

Simulations were conducted for the period from May to September, a period particularly prone to drought in Northern Italy. The results showed that during these months, the existing system requires 15,467 Sm³ of methane and consumes 40.4 MWh of electricity. Configuration B, which includes the advanced AWG, eliminates methane consumption entirely, produces over 257 m³ of high-quality water, and consumes 101.9 MWh of electricity. However, the additional electricity demand can be fully offset by cleaning a nearby 3 MWp photovoltaic field, with AWG water, recovering around 190 MWh otherwise lost due to soiling. About 45% of the water produced is sufficient for this purpose; the rest can meet hospital needs or replace bottled water use.

In configuration B, primary energy demand drops from 231.3 MWh to 193.6 MWh when supplied from the grid, and to 101.9 MWh when part of the recovered photovoltaic energy is reused. The AWG cooling capacity was almost fully exploited, while only 37% of its heating potential was used, suggesting that AWG sizing may be guided either by water needs or energy savings, where heating becomes a key factor.

Ongoing developments aim to further validate and extend the work. These include laboratory testing of water quality for potential technical or medical uses, especially in water-stressed regions. An economic assessment of the integrated system is also in progress, aiming to quantify the financial value of both the energy savings and the high-purity water generated. A DB script is also under development to enable full AWG simulation—including water yield and energy use—within building performance models. This will support both general and site-specific analysis, offering a valuable decision-making tool for sustainable energy–water solutions in critical infrastructure.

In conclusion, this study confirmed the technical feasibility and multi-benefit potential of integrating advanced AWG systems into building design processes. It represents a further step towards providing designers with a practical tool for incorporating such systems among available HVAC options. This approach supports more sustainable management of energy and water resources, particularly in areas facing growing environmental and infrastructural challenges.