The Role of Mechanical Ventilation in Indoor Air Quality in Schools: An Experimental Comprehensive Analysis

Abstract

Indoor air quality (IAQ) plays a vital role in supporting both the physical and mental well-being of individuals in enclosed spaces, and the role of mechanical ventilation systems has gained increasing attention due to building design’s focus on energy efficiency and thus airtight constructions. This study investigated the pre–post effects of installing a heat recovery mechanical ventilation system (MVHR) on indoor air quality in a high school classroom in Ferrara, Italy. The analysis focused on experimental measurements of temperature (T), relative humidity (RH), and carbon dioxide (CO₂) both inside and outside the same classroom, which had constant occupancy (17 students) for an entire school year, allowing a comparison between natural and mechanical ventilation. With a comprehensive approach, particulate matter (PM_2.5), volatile organic compounds (VOCs), and radon levels (Rn) were monitored as well, after the installation of the MVHR. By comparing natural and mechanical ventilation, the study highlights the strengths and limitations of the ventilation system implemented, together with an evaluation of the system’s energy consumption, including a 2 kW post-heating battery. In terms of results, the MVHR demonstrated clear benefits in managing CO₂ levels and improving sensory, olfactory, and psychophysical well-being, as well as the attention levels of students. In detail, under natural ventilation, peak concentrations exceeded 4500 ppm, while mechanical ventilation kept them below 1500 ppm. The average CO₂ concentration during occupancy dropped from 2500 ppm to around 1000 ppm, achieving a 62% reduction. However, beneficial effects were not observed for other parameters, such as PM_2.5, VOCs, or radon. The latter displayed annual average values around 21 Bq/m³ and peaks reaching 56 Bq/m³.

1. Introduction

Indoor air quality (IAQ) is a crucial factor for the well-being and mental and physical health of people in enclosed environments [1,2]. This aspect has become increasingly important over the years, especially as building envelopes are designed to be more energy-efficient and airtight, as required, for example in Europe, by progressively stricter European EPBD Directives [3,4]. Buildings with highly airtight envelopes, if not complemented by appropriate ventilation systems—either mechanical or natural—experience a deterioration in indoor air quality, leading to an increased risk of developing Sick Building Syndrome (SBS) [5,6,7]. Rostron in [6] highlights SBS symptoms, including dryness and irritation of mucous membranes, skin irritation, mental fatigue, nausea, tiredness, and respiratory issues. Furthermore, Rostron [6] and Takeda et al. [7] emphasize the correlation between SBS and the rise in increasingly airtight and efficient buildings, both globally [6] and specifically in Japan [7].

A commonly used parameter for assessing indoor air quality is the concentration of CO₂ [8,9]. This metric is essential in evaluating occupant well-being, especially in crowded spaces like daycares and schools, where monitoring becomes increasingly relevant [10,11]. When carbon dioxide concentration reaches critical thresholds of 1000 ÷ 1500 ppm(v), it may become a health concern, leading to a decline in cognitive abilities [12] and having a narcotic effect; for this reason, particular attention is given to settings such as schools and hospitals, which have been the focus of numerous analyses in the scientific literature [13,14,15,16,17]. Within this framework, low-cost sensors were employed in the international project by Sahme [18]. Recommended indoor CO₂ concentrations typically range between 1000 ÷ 1500 ppm, as noted by Dimitroulopoulou et al. and by Lowter et al. [19,20]. In particular, according to Lowter et al. [20], in schools, indoor air quality is deemed good if CO₂ remains below 1000 ppm and acceptable if it is below 1500 ppm. Standards such as EN 16798:2019 [21] categorize indoor air quality by CO₂ concentration levels. For instance, in residential environments, a Category I rating (the highest quality) requires CO₂ levels to stay within 500 ppm above outdoor levels in living rooms and 380 ppm in bedrooms. This aligns closely with findings from Lowter’s review [20], considering the global average outdoor CO₂ concentration of 419 ppm, as reported in the NOAA’s (National Oceanic and Atmospheric Administration) 2023 annual report [22].

Although CO₂ concentration is widely regarded as an indicator of indoor air quality, it is not the only parameter to consider [23,24]. Lucialli et al. [24], highlight this in their analysis of VOC (Volatile Organic Compound) concentrations in various Italian schools, showing that VOC levels are generally higher inside classrooms than outside. Similarly, in residential settings, VOC levels are typically elevated, underscoring the need for multiple indicators when assessing indoor air quality. In residential environments, gas stoves are a significant source of particulate matter and NO₂ (nitrogen dioxide) emissions, calling for adequate ventilation strategies [25]. Another crucial aspect, emphasized by the World Health Organization (WHO) [26,27], is that poor air quality, both indoors and outdoors, contributes to millions of deaths annually. This is especially important given that people spend most of their time indoors. The concentrations of PM_2.5 and PM10 also play a crucial role in determining indoor air quality [28,29]. The WHO has issued pollutant-specific limits in its WHO Global Air Quality Guidelines [27], discussed by Settimo et al. in [30], emphasizing the importance and need for legislation on this topic at both the European and Italian levels. The same study highlights that this issue is highly regarded within the scientific community, as most IAQ-related publications come from Italy, France, Germany, and the United Kingdom.

Another aspect to consider when addressing IAQ is radon. Radon (commonly referring to the isotope Rn-222) is a heavy, colorless, odorless gas belonging to the noble gas family. It occurs naturally as part of the uranium decay chain. Rn-222 has a half-life of 3.8 days and can enter buildings through infiltration from the ground, due to construction materials (some materials like tuff contain higher amounts of Rn-222 progenitors), and potentially through water [31]. Radon poses a significant health risk, as it is estimated to cause more than 100,000 deaths worldwide annually and is the second leading cause of lung cancer deaths after cigarette smoking [31,32]. According to Gaskin et al. [32], the total number of deaths due to radon exposure accounts for 3% of all cancer deaths in Italy. To address the radon issue, the EU has developed a radon concentration map at the European level with a cell resolution of 10 × 10 km [33,34]. A strategy to keep radon concentrations under control includes using mechanical ventilation (MVHR) [35,36], which helps ventilate indoor spaces and maintain low radon levels. In Italy, according to Legislative Decree D.Lgs 101/20 [37], which transposes the European Directive 2013/59/EURATOM [38], radon concentration must be below 300 Bq/m³ in existing buildings (or 200 Bq/m³ for new constructions).

Previous studies [17,35] have shown that controlled mechanical ventilation generally improves indoor air quality, regarding CO₂ concentrations and other pollutants, in new buildings with high airtightness and limited infiltration rates.

Moreover, the scientific literature mainly presents analyses devoted to a single aspect of the very complex issue of IAQ. For instance, in [39], attention is paid to the presence of a crossflow ventilation system, while in [40], reference is made to a temperature control model used in heating, ventilating, and air-conditioning (HVAC) systems in school spaces. A comprehensive approach is missing.

Also concerning standards, due to the complexity of IAQ and to its connections with many topics, many scientific standards are present, highlighting differences between the various approaches and issues that may influence future revisions, especially concerning ventilation [41,42]. The review literature on this subject highlights the need for the provision of evidence in Indoor Air Quality Standards [42], confirming the importance of experimental analyses such as the one discussed here.

The aim of the present analysis is to experimentally assess the impact of adopting a controlled mechanical ventilation system—in place of natural ventilation through the opening of windows—in a high school classroom located in Ferrara, Northern Italy, on indoor air quality conditions, focusing on CO₂, particulate matter, VOCs, and radon concentrations. Unlike many other studies in the literature, this analysis considers a broad range of IAQ parameters. The study will highlight the strengths and limitations of the implemented ventilation system, alongside an evaluation of the system’s energy consumption. This analysis forms part of a larger, multicenter study aimed at the sustainable regeneration of school spaces (Symbien—School Symbiotic Environments) [43], intended to enhance symbioses between microorganisms, processes, consumption, and ecosystems.

The field data were collected as part of a broader research project (SYMBIEN—www.cias-ferrara.it/projects/symbien, accessed on 10 November 2024) whose ultimate goal is the creation of a prototype classroom with the following features:

• Zero energy: achieved through the installation of a photovoltaic system capable of covering all the annual electrical, thermal, and cooling needs of the classroom;
• Controlled physical (dust) and chemical (CO₂ and VOC) contamination: ensured by the installation of a mechanical ventilation system;
• Controlled microbiological contamination: implemented through a protocol for sanitizing furniture and surfaces using probiotic microorganisms and eliminating chemical disinfectants, with a reduction in pathogenic microorganisms;
• Introduction of various types of plants and green walls;
• Optimization of the classroom’s interior design.

The research was conducted in Classroom 13 of the Liceo Classico Ariosto in Ferrara between March 2023 and June 2024 and also involved the students of Class IV N in several monitoring activities.

2. Materials and Methods

The main purpose of this analysis is to evaluate the indoor air quality (IAQ) and CO₂ concentration within a classroom located in Ferrara, Northern Italy, based on two ventilation approaches: natural ventilation and mechanical ventilation using a dual-flow heat recovery unit with post-heating support (MVHR, mechanical ventilation with heat recovery). Data were recorded over an entire school year period, specifically focusing on temperature, humidity, and CO₂ concentration levels within the classroom, as well as in the adjacent corridor and outdoors. Measurements were collected over nine months during the season from March 2023 to June 2024. The first period, from March to June 2023, was utilized to check the sensors installed in the classroom and external environment, as well as to verify the data collection system. The two months of data—from October to December 2023—correspond to conditions without mechanical ventilation, while the remaining months correspond to periods with active mechanical ventilation (from January 2024 to June 2024). During the latter phase, the power consumption of the installed ventilation system was recorded using an accessible power meter, including the electric power required by the post-heating battery.

After the installation of the mechanical ventilation system, additional sensors were installed near the air return inlet to monitor air quality in terms of particulate matter (PM_2.5), radon concentration, volatile organic compounds (VOCs), and atmospheric pressure, as well as CO₂, temperature, and humidity levels at the ventilation unit. These parameters were tracked between February 2024 and November 2024.

The comprehensive dataset on environmental conditions and IAQ thus provides a thorough overview of indoor air quality within the public building, both before and after MVHR installation. For the first dataset (October 2023—June 2024), the sensors used were two factory-calibrated Testo 160 IAQ CO₂ meters [44] (serial number 51616618 for the one installed in the classroom and serial number 51616624 for the one placed outside, Symbien, Espoo, Finland), with their main features listed in

Table 1. Measurements range, accuracy, and resolution for Testo 160 IAQ sensor. “m.v.” refers to the measured value.

Variable	Range	Accuracy	Resolution
Temperature	0–50 °C	±0.1 K	0.1 K
Humidity	0–100%	±2% (range 20–80% at 25 °C)	0.1%
CO2	0–5000 ppm(v)	±(50 ppm(v) + 3% m.v.)	1 ppm(v)

. The characteristics of the power meter used are specified in its datasheet, indicating a measurement uncertainty of ±5% of the recorded value. Table 1 also includes the accuracy and the resolution of the various measured quantities and the measurement ranges.

The Testo sensors used were placed in the classroom and outdoors/in the adjoining corridor at a height of 1.70 m (Sensor 1, Figure 1b). In addition to measuring CO₂ concentration in ppm(v) (parts per million by volume), they also allow for the monitoring of temperature and humidity in the surrounding environment, as shown in

Table 2. Measurements ranges and accuracy for AIRTHINGS sensor [45].

Variable	Range	Accuracy
Temperature	4–40 °C	0.5 K
Humidity	Up to 85%	3%
Pressure		0.6 hPa
Radon	0–20,000 Bq/m3	±10% m.v.
Particulate (PM2.5)	0–500 μg/m3	±5 (μg/m3 +15% m.v.), range 0–150 μg/m3
VOCs	0–10,000 ppb	-
CO2	400–5000 ppm	±(50 ppm(v) + 5% m.v.)

.

The additional data on CO₂, temperature, humidity, and the general IAQ at the entrance of the heat recovery unit were recorded (Sensor 2 in Figure 1b). These measurements were collected starting from February 2024 utilizing a View Plus AIRTHINGS sensor [45]; the measurement ranges and uncertainties are reported in Table 2 for each variable. For these collected data, there are different time granularities, and the time between a record and another is in the range of 2–6 min.

2.1. Building and Heating System Description

The classroom analyzed is part of the “Liceo Ludovico Ariosto” school complex (located in Ferrara, a municipality in Northern Italy). During lessons, the classroom is consistently occupied by a single class of 17 students and 1 teacher. Figure 1 and Figure 2 show the classroom layout and sensor and MVHR position, in terms of a sketch and photos, while Figure 3 presents a photograph taken inside the room in the absence of people and before the installation of the controlled mechanical ventilation system that will be analyzed. Figure 1b shows that only one wall, shown on the lefthand side, separates the classroom from the unheated environment. The key characteristics of the analyzed environment are provided in

Table 3. Main characteristics of the classroom analyzed.

Length	8.47 m
Width	5.75 m
Ceiling height	3.23 m
Maximum height	5.15 m
Floor area	48.7 m2
Volume	170 m3

. Specifically, Figure 3 shows that the classroom features large windows facing outward, as well as a ceiling shield characterized by additional windows, reaching a maximum height of 5.15 m at this point. Moreover, additional internal windows are facing the main corridor (shown on layout sketch in Figure 1).

Winter heating is currently provided by a wall-mounted fan coil with a thermostat, while summer cooling is achieved through an air split system with a condenser unit installed on the building’s roof.

Additionally, like all other classrooms in the school complex, the classroom under examination was not equipped with any air exchange control system, meaning ventilation was ensured manually by opening doors and windows.

2.2. Controlled Mechanical Ventilation Unit

The controlled mechanical ventilation machine is a commercial unit produced by Aldes (model VEX 380, Aldes Italia, Modena, Italy) [46], and the unit installed inside the classroom is visible in Figure 3. The machine is designed vertically and features a maximum airflow of 1000 m³/h. It also includes a 2 kW post-heating coil. This post-heating unit is necessary, as otherwise, the air introduced into the environment could have a temperature lower than 20 °C, resulting in thermal discomfort for the occupants. The fans have variable speeds, and the heat exchanger operates on the counterflow principle. The main technical specifications, as well as the performance efficiency curves of the exchanger, are presented in

Table 4. Main parameters of mechanical ventilation system.

Heat recovery efficiency 1	80%
Nominal power (excluding electrical post-heating)	377 W
Nominal external pressure	50 Pa
Fan efficiency	49.3%
Specific power (SFPint)	1189 W/(m3/s)
Filters	F7/ePM1 70% for extraction and fresh air
Sound power level 1	56 dB(A)

¹ According to EU n1253/2014.

and Figure 4 based on volumetric flow rate. The post-heating coil ensures that the air supplied is at a fixed temperature of 20.8 °C at nominal flow, assuming an external air temperature of 15 °C.

2.3. Set-Up of the Analysis and Sensors’ Position

As previously indicated, the experimental analysis related to thermohygrometric conditions and CO₂ concentration (the first phase of the analysis) was divided into two sub-phases:

-

The first phase (from mid-September 2023 to 31 December 2023) focused on using only natural ventilation in the classroom (Sensor 1 active);

-

The second phase (from 6 January 2024 onwards) involved mechanical ventilation (Sensor 1, Sensor 2, and MVHR active).

For both phases, the CO₂ levels detected are reported, as well as the temperatures and humidity measured inside and outside for the characteristic sample days.

For the first sub-phase, the openings of the windows and the main door facing the corridor were recorded as well. It is also noted that, for the scenario involving controlled mechanical ventilation, the airflow rate of the unit was limited to a constant value of 400 m³/h, estimating a necessary renewal rate of 22 m³/h per person (17 Students and 1 Professor—6.1 L/s per person). Subsequently, from 8 April 2024, the airflow was increased by 10% to 440 m³/h (24.4 m³/h per person, i.e., 6.7 l/s per person) to analyze its effects on daily CO₂ concentration; the airflow was not increased further to keep noise levels within acceptable limits in the classroom. The daily operating schedule of the unit was set to 6 h, from 7:00 a.m. to 1:00 p.m., so it would only be used during periods of actual occupancy and for 6 days a week (excluding Sundays, as there are no classes on Sundays). Additionally, during analyses with controlled mechanical ventilation, the fresh air from the recuperator entering the classroom was heated by the post-heating coil to achieve an outlet temperature of 20 °C.

Moreover, during this phase, an additional extended measurement campaign from February 2024 to November 2024 involved an indoor air quality analysis and measurement of values at the MVHR outlet, aiming to assess the influence of the system’s operation on a wider range of parameters that describe IAQ within the environment (radon, VOCs, and PM_2.5 concentration, basically). Moreover, during the period of February–April 2024, the energy demand of the MVHR was also recorded.

3. Results and Discussion

In this section, the experimental results obtained during the experimental campaign will be presented and discussed.

3.1. Thermohygrometric and CO₂ Concentration Analysis (October 2023–April 2024)

In Figure 5, carbon dioxide concentration, temperature, and relative humidity values, recorded for two Sundays (29 October 2023, and 17 December 2023), are presented. Focusing on CO₂ (shown with measurement uncertainty according to the technical datasheet), we observe that, on both days, within the selected monitoring period from 7:00 a.m. to midnight, the indoor CO₂ concentration in the classroom exhibits a decreasing trend, dropping from about 650 ppm(v) to approximately 550 ppm(v). In contrast, outdoor CO₂ levels are lower than indoors and show more fluctuations. This trend (CO2_INT) is typical in spaces where, due to external infiltration and no internal generation (as the room is unoccupied), CO₂ concentration tends to decrease. Additionally, we note that the indoor concentration does not reach the outdoor level; this equilibration is expected to occur in the early hours of Monday morning (not reported here for brevity).

Regarding relative humidity, we observe stability within the classroom on both days: around 70% on the Sunday in November and between 55% and 60% on the Sunday in December. Indoor temperatures also remained stable over the two days, but were lower in December (16–17 °C) compared to October’s 19–20 °C. Outdoor temperatures, however, were notably different, with December’s minimum around 7 °C compared to October’s minimum of 14 °C. The temperature trend in December highlights how the heating system is correctly turned off when the school is unoccupied.

In Figure 6, the trends of the CO₂ concentrations recorded over the two analyzed days in October and December are shown, in the absence of internal CO₂ production due to the lack of anthropogenic activities. It is observed that at 7:00 a.m. on both days, the concentrations are very close (642 ppm(v) on the October day and 641 ppm(v) on the December day). Additionally, the two curves are perfectly superimposable, with deviations within the accuracy of the internal sensor.

Let us now consider the measurements taken without controlled mechanical ventilation, i.e., when ventilation and air exchange are manually managed by opening and closing windows during class hours. Figure 7a,b, refer to 23 October, Monday, a day when the heating system was not yet active, as the external temperature combined with internal loads was sufficient to maintain a comfortable climate inside the classroom (the average indoor temperature during student occupancy from 8:00 a.m. to 1:00 p.m. was 20.5 °C, with a minimum of 19.3 °C and a maximum of 21.7 °C). From the CO₂ graph in Figure 7a, significant variations in CO₂ levels in the environment are observed (the average value during the usual period of occupancy is 1334 ppm(v), with extremes of 618 and 2205 ppm(v)). Rapid changes occur in correspondence with the opening of windows and/or doors to carry out natural ventilation.

Moreover, comparing the measured CO₂ values with temperature and indoor humidity, it is observed that an increase in CO₂ corresponds to an increase in relative humidity in the classroom, due to the presence of people; however, relative humidity quickly decreases following the opening of doors and windows. This effect is also visible in the indoor temperature, although to a lesser extent, due to the position of the installed sensor within the classroom, which is installed 1.70 m above the floor. These data show that natural ventilation was carried out, but the exchanges were still insufficient to ensure adequate CO₂ levels for the presence of people in the environment.

Now, in the case of Figure 7c,d, referring to 23 November 2023, the heating system was on, and during the time interval of classroom occupation by students, an average temperature of 21.5 °C was recorded, with extremes of 18–23.5 °C. Even in this case, from the CO₂ graph, it is observed that natural ventilation was carried out, and the corresponding trends related to humidity and indoor temperature discussed previously are observed. Also, in this case, natural ventilation was entirely insufficient to ensure adequate air quality in the environment during the hours of occupation (with an average concentration of 1504 ppm(v), with a maximum of 2485 ppm(v) and a minimum of 442 ppm(v), the latter corresponding to the opening of windows).

Considering the carbon dioxide concentration shown in Figure 7a,c in the absence of mechanical ventilation, a characteristic “sawtooth” pattern can be observed. This pattern is typical of the Lotka–Volterra model [47], commonly used in ecology to estimate the dynamics of two interacting populations, prey and predator, over time. This time-dependent model describes population trends through two differential equations, where key factors include the prey’s growth rate (in this case, the increase in CO₂ when windows are kept closed), the prey’s mortality (CO₂‘s decreasing rate due to infiltration), and the predator’s growth rate (represented by the opening of windows).

Finally, analyzing Figure 7e,f, referring to 13 December, Wednesday, which was characterized by a more severe external climate (with an average of 11.2 °C and minimum of 8.4 °C), there were very limited air exchanges with the outside and other environments: this situation led to poor air quality inside the classroom, reaching a peak CO₂ level at 12:15 of 4063 ppm(v), corresponding to a peak in relative humidity of 62%. Furthermore, an abnormal trend is observed when attention is focused on indoor temperature. Even in this case, the heating system was operational. Still, unlike the previously analyzed day, when the system was turned off at the end of classroom use, in this case, the system remained on until 6:00 p.m., leading to excessively high internal air temperatures of over 27 °C. This abnormal behavior of the system was also detected on other days of the heating season.

In Figure 8, the graphs relating to the CO₂ concentration, temperature, and relative humidity for two days when the controlled mechanical ventilation system was operational are shown. From Figure 8a,b, referring to 19 January, a day when both the heating system was operational and controlled mechanical ventilation was set at a fixed flow rate of 400 m³/h from 7:00 a.m. to 1:00 p.m., a rising CO₂ trend is observed, with a peak at 11:30 a.m. of 1552 ppm(v) and an average during the occupancy period of 1204 ppm(v).

In fact, if carbon dioxide is seen as gas tracker, Figure 8 highlights that an efficient air renew occurs. These values are lower than those obtained with natural ventilation; moreover, during the occupancy period, they are comparable to the values detected in the adjacent corridor (the Testo sensor previously positioned outside was moved to the adjacent corridor). It is also observed that after the students leave the school, CO₂ shows a decreasing trend both in the corridor and the classroom; however, in the classroom, the CO₂ levels are close to those normally found outside. During the occupancy period, the temperatures were maintained at around 23 °C in the classroom while ranging between 20 and 23 °C in the corridor; the humidity also remained stable at values between 50 and 55% in the classroom.

Figure 8c,d refer to 4 April, when the heating was off, but the controlled mechanical ventilation was operating at a fixed rate of 440 m³/h. Also, in this case, the peak of CO₂ increased during the occupancy period, with a peak of 1250 ppm(v) and an average during the occupancy period of 844 ppm(v). It is also noted that during the period from 11:30 a.m. to 1:00 p.m. a plateau was reached; the values obtained are still lower compared to 19 January, due to the 10% increase in the air renewal rate. The sensor placed in the corridor, however, measured a lower concentration, close to what should be present outside. It is likely that some corridor windows were open, while the classroom door to the corridor was closed. Inside the classroom, temperatures and humidity were maintained constantly at ranges of 21–23 °C and 45–50%, respectively. It is clearly visible that the concentration of CO₂ decreased following the implementation of controlled mechanical ventilation, as well as the stabilization of the relative humidity level.

Moreover, Figure 9 shows the trend of thermohygrometric parameters when using only scheduled natural ventilation, achieved by opening the windows for 5–10 min at each hour change. The aim was to experiment with a simple procedure carried out by the students to reduce CO₂ concentration.

This behavior highlights that the vast majority of schools in Italy, as well as in many other European countries, do not have classrooms equipped with mechanical ventilation systems. Therefore, it is essential to identify indirect criteria for controlling IAQ and CO₂ concentration.

A distinct sawtooth pattern can be observed. The presented results demonstrate that, although such an intervention is theoretically feasible and can lead to improvements in both average and peak CO₂ concentrations, temperature control becomes more challenging. This results in complaints from students and teachers about temperature fluctuations. In this case, the ambient temperature ranged between 18 and 23 °C, compared to much smaller fluctuations (21–23 °C) in the case of Figure 8d (mechanical ventilation).

Additionally, considering Figure 8c,d, when mechanical ventilation is employed, not only is a stabilization of CO₂ concentration observed during the occupancy period, remaining below 1330 ppm(v), but also a stabilization of the indoor temperature between 8:00 a.m. and 1:00 p.m., ranging from 22.0 to 23.3 °C. The same behavior was not observed in the case of natural ventilation (Figure 9a,b). Although periodic air changes kept the CO₂ concentration below 2200 ppm(v), greater temperature variability was recorded in the classroom, ranging from 23.3 to 18.9 °C during window openings. This resulted in an important increase in thermal discomfort associated with window openings for ventilation.

In Figure 10a,b, the daily values of average and maximum CO₂ inside the classroom for when students are present in the classroom of are reported, respectively. Focusing on Figure 10a, it is observed that the average maximum values occur in the presence of natural ventilation, with particular values reaching up to 2500 ppm(v). In the case of controlled mechanical ventilation, however, the average values are much lower and rarely exceed 1200 ppm(v). Furthermore, the effect of increasing the renewal rate of the machine in the final part of the measurement campaign is observed: in the last part, the average CO₂ values resulted in an average concentration almost always below 1000 ppm(v). Similar considerations can be made for the maximum daily concentration during the occupancy period reported in Figure 10b: it is particularly of note that the peaks of maximum concentration in the presence of natural ventilation reach up to 4500 ppm(v), while in the case of controlled mechanical ventilation the maximum peaks rarely exceed 1500 ppm(v). Consider Equation (1):

$$\mathsf{\Delta }\mathrm{C}{\mathrm{O}}_2(\%)=100{\displaystyle \frac{C_{NV}-C_{MV}}{C_{NV}-500}}$$

(1)

In expressing the percentage reduction in CO₂ concentration following the adoption of mechanical ventilation, C_MV represents the average concentration recorded during occupancy periods under mechanical ventilation, while C_NV refers to the average concentration recorded during occupancy periods under natural ventilation. Analyzing the data reported in Figure 10a, excluding non-occupancy periods such as Easter and Christmas holidays and Sundays, a significant reduction in CO₂ concentration is observed. Specifically, the daily average reduction is −62% for mean daily values (Figure 10a). More precisely, under natural ventilation, the daily average CO₂ concentration reaches 1576 ppm, while it reaches 900 ppm under mechanical ventilation. Furthermore, considering the maximum concentrations obtained during the same observation period, a C_NV value of 4573 ppm(v) was recorded on 3 November 2023, and a C_MV value of 1802 ppm(v) was recorded on 12 January 2024. This results in a concentration reduction of 68%, as calculated using Equation (1).

The CO₂ concentration can also be seen as a tracker to measure the air renewal rates, n, inside the classroom. Details on the estimation of this parameter, in non-occupancy conditions, are reported in Appendix A.

3.2. CO₂ Emission Rate per Person Inside the Classroom

It is possible to express the CO₂ emission rate inside the classroom, starting from the concentration in ppm relative to volume y_CO2 (ppm(v), as read from Testo sensors) and then expressing it as a weight concentration (x_CO2, ppm by mass, Equation (2)). Additionally, the CO₂ concentration in mg/m³ of air (C_CO2) can also be calculated within a given environment:

$$x_{\mathrm{C}\mathrm{O}2}=y_{\mathrm{C}\mathrm{O}2}{\displaystyle \frac{{\mathsf{\mu }}_{\mathrm{C}\mathrm{O}2}}{{\mathsf{\mu }}_{air}}}$$

(2)

$$C_{\mathrm{C}\mathrm{O}2}={\displaystyle \frac{n_{\mathrm{C}\mathrm{O}2}}{n_{air}}}{\displaystyle \frac{{\mathsf{\mu }}_{\mathrm{C}\mathrm{O}2}}{v_{mol,air}}}=f_{\mathrm{C}\mathrm{O}2}\cdot {\displaystyle \frac{{\mathsf{\mu }}_{\mathrm{C}\mathrm{O}2}}{v_{mol,air}}}$$

(3)

In these equations, the variables are as follows:

• μ_CO2 and μ_air are the molar masses of CO₂ and air, which are 44 g/mol and 29 g/mol, respectively;
• v_mol,air is the molar volume of air, which depends on temperature. At 20 °C, it is 24.05 × 10⁻³ m³/mol;
• f_CO2 is the molar (or volume) fraction expressed as a percentage, representing the ratio of n_CO2/n_air divided by 10⁶. For example, at a concentration of 1000 ppm(v) in the environment, the volume fraction is 0.1%, as shown in Equation (4):

$$f_{\mathrm{C}\mathrm{O}2}={\displaystyle \frac{y_{\mathrm{C}\mathrm{O}2}}{{10}^6}}={\displaystyle \frac{1000}{{10}^6}}={\displaystyle \frac{0.1}{100}}=0.1\%$$

(4)

The quantity of CO₂ in mg/m³ emitted by each person in the room can be estimated using a mass balance (Equation (5)) in steady state, which considers the air supply rate into the room through mechanical ventilation (${\dot{m}}_{air}$), the external concentration (C_ext), the internal concentration (C_int) measured at the machine’s intake, and the internal CO₂ production (${\dot{m}}_{\mathrm{C}\mathrm{O}2,tot}$) due to people in the room. In this analysis, air change rate due to infiltration is neglected:

$${\dot{m}}_{air}{C}_{ext}+{\dot{m}}_{\mathrm{C}\mathrm{O}2,tot}={\dot{m}}_{air}{C}_{int}$$

(5)

The infiltration airflow has been considered negligible compared to the airflow of the Aldes machine. This can be inferred from Figure 7a,c,e; in fact, the decay of the internal CO₂ curve within the class (with doors closed, no presence, and the machine turned off) is very slow.

Assuming an external concentration of C_ext = 760 ppm (500 ppm(v)), an internal concentration of C_int = 2050 ppm (1350 ppm(v)), and an air renewal flow rate of 480 kg/h, the total internal CO₂ production is calculated to be 622 g CO₂/h, applying Equation (3) to determine the CO₂ concentration per unit of volume of air at 20 °C, considering the molar volume of air to be 24.05 ⋅ 10⁻³ m³/mol. The previous result was obtained by applying Equation (5) to determine ${\dot{m}}_{CO2,tot}$. Dividing this value by the number of people in the room (assumed to be 18, 17 students and 1 professor) provides a per-person CO₂ production rate of 34.5 g CO₂/(h⋅person). This value is in accordance with the scientific literature [48,49]. By introducing a calculated average weight of 54.6 kg per person, the CO₂ emission obtained is 0.63 g CO₂/h per kg of weight. This field data are useful for predicting the airflow required to maintain a predetermined level of carbon dioxide in any environment where people are at rest: of course, higher values of airflow ensure lower CO₂ concentrations, but face other issues, such as noise emissions or simply energy consumption.

3.3. IAQ Analysis (February 2024–November 2024)

In this section, the indoor air quality conditions in the environment with and without mechanical ventilation will be discussed; consumption values are also presented accordingly. As is well known, electrical energy consumption is strongly related to carbon dioxide emissions [50], also regarding a dynamic consumption context [51].

Figure 11 shows the electricity consumption trends for the MVHR system during the measurement period from 18 February to 18 April 2024, alongside the average external temperature recorded in Ferrara by sensors from the Emilia-Romagna Regional Environmental Protection Agency (ARPAE) for a climate station located in the city center [52,53]. The reported values represent averages from 7:00 a.m. to 1:00 p.m., corresponding to the actual operating hours of the MVHR system. It is noted that, as expected, the ventilation system is turned off on Sundays. While an inverse trend between external temperature and daily energy consumption is observed, the correlation between external temperature and electricity demand is not particularly strong (R² = 0.425). This electricity consumption includes energy used by the system’s electric post-heating battery and fans. The low correlation may be attributed to multiple influencing factors, such as the varying amounts of energy supplied by the heating system, exchanges with the surrounding environment, and the potential openings of doors to the outside or adjacent spaces. Additional correlations were also investigated, particularly with the 24 h daily average external temperature and with the internal average temperature during both the 6 h of occupancy and the 24 h daily period. However, the correlation index values were found to be below 0.4.

A focus on energy consumption due to the mechanical ventilation system is reported in Appendix B.

In Figure 12a,b, the average daily values of VOCs and PM_2.5 concentrations in the classroom are presented. Figure 12a also includes the trend for PM_2.5 outside the classroom, as measured by a station located about 3 km from the school (Villa Fulvia station), with data obtained from the open ARPA Emilia-Romagna database [54].

Focusing on the first figure, a good correlation is observed between the indoor and outdoor particulate matter concentrations (R² of 0.65), with indoor values consistently lower than outdoor levels over the observation period, showing an average reduction of 38% compared to outdoor concentrations. A decrease in particulate concentration is also noted on Sundays, when the ventilation system is inactive. This phenomenon can be explained by several factors. One reason is the presence of PM from occupants; human activity and movement during school hours contribute to the resuspension of particles. Another factor involves the system’s operation mode: the ventilation system maintains neutral pressure between the interior and exterior, which likely allows for the infiltration of air with higher PM_2.5 concentrations; this effect is not observed for CO₂ concentrations, as the concentration of incoming air from infiltration is equal to the renewal air concentration in terms of CO₂. This finding suggests that further studies should investigate the impact of maintaining a positive pressure in the classroom to minimize outdoor influence. This approach could leverage the filtration capabilities of the installed F7/ePM1 70% filters, which can capture at least 70% of even finer particles, such as PM1.

Turning to Figure 12b, which displays two months of PM_2.5 levels alongside VOC levels in the classroom, there appears to be little correlation between the two. VOC levels remain lower during periods of non-occupancy, such as Sundays and school holidays, particularly in August. It should be noted that VOCs depend not only on external infiltration but also on occupant presence, cleaning products used, and materials composing the surfaces within the classroom.

In Figure 13, the radon concentration is shown, considering the hours when mechanical ventilation was operational (6 h) and the total daily value averaged over 24 h, for the period from 14 February to 14 April. The figure shows that the values during the use of mechanical ventilation are comparable to the average over 24 h, with average annual concentrations around 21 Bq/m³ (both for the 6 h and 24 h averages). The highest recorded values reached 56 Bq/m³, which is still well below the legal limit of 300 Bq/m³, as established by Italian Decree 110/2020 and European Directive 2013/59/Euratom [41,42]. This indicates the limited influence of this mechanical ventilation set-up on radon mitigation, suggesting that this effect needs to be further investigated.

In Figure 14, the internal trends for temperature, CO₂ concentration, and humidity are shown over a two-month period, comparing measurements from the Testo sensor (positioned 1.70 m above the floor) with those from a sensor located near the return of the controlled mechanical ventilation unit.

In Figure 14a, it can be observed that the temperature measured near the ventilation unit is close to that recorded by the Testo sensor, with average deviations of 0.8 K over the measurement period (excluding data points where values from either sensor were missing). Minimal discrepancies are also noted for relative humidity, shown in Figure 14b, with an average difference of 1% over the period. However, for CO₂ concentration, the average deviation is more significant at 147 ppm, though still within a reasonable range. This discrepancy in CO₂ levels may be attributed to the differing concentrations that can occur at various heights within a room, as previously investigated by Pei et al. in [55].

4. Conclusions

This study examined the impact of adopting controlled mechanical ventilation on indoor air quality in a high school classroom located in Ferrara, Northern Italy. The analysis was conducted before and after installing a heat recovery mechanical ventilation system. In addition to environmental parameters like CO₂, particulate matter, VOCs, and radon, the study also assessed the machine’s electricity consumption, which includes a 2 kW post-heating battery. The key findings are as follows:

-

CO₂, humidity, and temperature measurements were obtained from two sensors: one located 1.70 m above the ground and the other near the ventilation machine system’s intake. The average deviations between the two sensors were approximately 0.8 K for temperature, 1% for relative humidity, and 147 ppm(v) for CO₂ (Figure 14a–c).

-

CO₂ concentrations in the classroom dropped significantly with mechanical ventilation, with peak values reaching more than 4500 ppm(v) under natural ventilation and rarely exceeding 1500 ppm(v) with mechanical ventilation. The average concentration during occupancy decreased from around 2500 ppm(v) to levels close to 1000 ppm(v) and a reduction of 62% (Figure 10a, considering mean values during the occupation period), and the daily reduction reached −68% considering the maximum daily values (Figure 10b).

-

Energy consumption by the MVHR system varied due to the use of the post-heating battery, mainly depending on external air temperature (R² = 0.425), as the machine was set to operate at a fixed rate. Daily energy usage ranged from about 1 kWh to 11 kWh over 6 h of operation, indicating the need for variable-rate operation to reduce energy use and avoid excessive ventilation and consumption.

-

While the MVHR system demonstrated clear benefits in managing CO₂ levels, similar effects were not observed for PM_2.5, VOCs, or radon. Throughout the measurement period, indoor PM_2.5 levels (daily range 3–39 μg/m³) remained lower than those outside, regardless of whether the machine was on or off. No significant differences were observed in VOC levels between machine operation and downtime, with daily averages between 60 and 370 ppb. Radon levels also showed no significant change with the system on or off, with annual average values around 21 Bq/m³ and peaks reaching 56 Bq/m³ (Figure 12a,b and Figure 13). These effects may be partly attributed to the machine’s operation at neutral pressure, allowing potential pollutant infiltration from outside, even though the system is equipped with an F7/ePM1 70% filter that can capture at least 70% of finer particles, including PM1. The observed indoor PM_2.5 levels mirrored those from nearby ARPAE monitoring stations outside the building.

Future research will explore operating the machine at variable rates to adjust flow based on pollutant dilution needs and assess the effects of slight positive pressure on radon, VOC, and particulate matter concentrations. Moreover, in future analyses, attention will be paid to the role of operative temperature. Renewable energy sources are also planned to be integrated, aiming to make the classroom nearly carbon-neutral. This study is part of a multidisciplinary analysis [43] for the sustainable regeneration of school spaces, focusing on improved habitability with objectives that include:

-

Characterization of the environmental microbiome (microbial, bacterial, and fungal communities) and methods to support a balanced ecosystem, including compatible plant installations (green classrooms);

-

Control of thermohygrometric, physical, chemical, and energy parameters to enhance occupant health and reduce the built environment’s carbon footprint (decarbonization);

-

Improvement in psychological well-being and learning outcomes by enhancing the aesthetic, functional, and material qualities of spaces, promoting interaction among people and between individuals and their environment (interior design);

-

Development and prototyping of innovative educational spaces, using specific and quantitative knowledge of the various parameters that contribute to the livability of constructed spaces.