Experimental Assessment of Demand-Controlled Ventilation Strategies for Energy Efficiency and Indoor Air Quality in Office Spaces

Abstract

This study investigates the performance of different demand-controlled ventilation strategies for improving indoor air quality while optimizing energy efficiency. The experimental research was conducted at the Indoor Live Lab at the University of Coimbra using a smart window equipped with mechanical ventilation boxes, occupancy sensors, and a real-time CO₂ monitoring system. Several occupancy-based and CO2-based ventilation control strategies were implemented and tested to dynamically adjust ventilation rates according to real-time indoor conditions, including (1) occupancy period-based control, (2) occupancy level-based control, (3) ON-OFF CO₂-based control, (4) multi-level CO₂-based control, and (5) modulating CO₂-based control. The results indicate that intelligent control strategies can significantly reduce energy consumption while maintaining indoor air quality within acceptable limits. Among the CO₂-based controls, strategy 5 achieved optimal performance, reducing energy consumption by 60% compared to the simple ON-OFF strategy, while maintaining satisfactory indoor air quality. Regarding occupancy-based strategies, strategy 2 showed 58% energy savings compared to the simple occupancy period-based control, but with greater CO₂ concentration fluctuation. The results demonstrate that intelligent DCV systems can simultaneously reduce ventilation energy use by 60% and maintain compliant indoor air quality levels, with modulating CO₂-based control proving most effective. The findings highlight the potential of integrating sensor-based ventilation controls in office spaces to achieve energy savings, enhance occupant comfort, and contribute to the development of smarter, more sustainable buildings. Future research should explore the integration of predictive analytics and multi-pollutant sensing to further optimize demand-controlled ventilation performance.

1. Introduction

In modern society, people spend the majority of their time indoors, whether at home, in offices, or in other enclosed environments [1]. The COVID-19 pandemic further increased indoor occupancy time, highlighting the critical importance of maintaining good indoor air quality (IAQ) to ensure occupant health, well-being, and productivity [2,3]. Poor IAQ has been associated with various health problems, including respiratory issues, headaches, and decreased cognitive function, which can significantly impact occupants’ overall performance and comfort [4], and may result in Sick Building Syndrome (SBS) [5,6].

Ventilation plays a crucial role in maintaining IAQ by diluting and removing indoor pollutants. While natural ventilation is a passive, energy-efficient, and sustainable approach, it is highly dependent on external factors such as wind pressure and temperature differences, making it unreliable and difficult to control [7]. Furthermore, fluctuating weather conditions as well as dissimilar occupant behavior can lead to inconsistent ventilation rates, which may either result in inadequate fresh air supply or excessive energy loss due to over-ventilation [8]. To address these limitations, mechanical ventilation systems provide a controlled means of ensuring adequate air exchange. However, mechanical ventilation requires energy input, making it essential to develop smart and efficient control strategies to optimize ventilation performance while minimizing energy consumption [9]. This is particularly important given the European Union’s ambitious energy efficiency targets and the global push towards more sustainable building operations [10].

Several strategies can be employed to enhance IAQ, including source removal, localized extraction, air purification, and ventilation. Among these, ventilation remains one of the most effective approaches for maintaining a healthy indoor environment, especially in office spaces where occupant density can vary significantly throughout the day [11]. Over the past two decades, advancements in sensor technology and smart control systems have enabled the development of intelligent ventilation strategies that dynamically adjust air exchange rates based on real-time indoor conditions [12,13]. By integrating demand-controlled ventilation (DCV) systems that respond to occupancy levels and indoor pollutant concentrations, buildings can achieve a balanced trade-off between IAQ improvement and energy efficiency.

DCV systems dynamically regulate ventilation rates based on real-time measurements of indoor pollutants, such as carbon dioxide (CO₂) [14,15], and occupancy levels [16,17,18], ensuring optimal energy use while maintaining acceptable IAQ. These systems operate using either constant air volume (CAV) or variable air volume (VAV) approaches, depending on the control strategy employed. CAV systems maintain a steady airflow rate, adjusting ventilation only through ON-OFF control mechanisms, while VAV systems modulate airflow in response to fluctuating indoor conditions, providing a more adaptive and energy-efficient solution [19,20]. By integrating advanced control algorithms [13], DCV systems can optimize ventilation performance, reducing energy waste while ensuring a healthy indoor environment [21].

This study experimentally evaluates several DCV strategies to optimize energy efficiency and IAQ in office environments. The primary objectives are as follows: (1) to compare the performance of occupancy-based and CO₂-based ventilation controls; (2) to quantify energy savings while maintaining IAQ thresholds (e.g., CO₂ <1000 ppm); and (3) to demonstrate the feasibility of real-time sensor-driven automated control. Experiments were conducted at the Indoor Live Lab (I2L) [22] at the Department of Mechanical Engineering (DEM) at the University of Coimbra (UC), using a smart window system integrated with mechanical ventilators, calibrated CO₂ sensors, and infrared occupancy detectors. Five control strategies—ranging from simple ON-OFF logic to fully dynamic modulation—were tested to identify the optimal approach for maintaining IAQ while minimizing energy consumption. The remainder of this paper is organized as follows: Section 2 describes the methodology, including the experimental setup and data collection approach, and presents the ventilation control strategies evaluated in this study. Section 3 discusses the results and provides a comparative analysis of energy efficiency and IAQ performance. Finally, Section 4 concludes the study and outlines future research directions.

2. Methodology and Experimental Setup

2.1. Method

There are several methods for measuring and evaluating the air exchange rate (AER) caused by infiltration and/or ventilation (both for mechanical and natural ventilation systems) [23]. The most accurate, common, and widely used method is the tracer gas method, in which tracer gas concentration is measured over time with an appropriate gas analyzer. Once the tracer concentration evolution with time is measured, it is possible to evaluate and calculate the air flow rate with an appropriate mathematical analysis. The experiment can be implemented using different methods, such as constant concentration, constant tracer emission, and decay methods [24,25,26]. In this work, the tracer gas decay method is used, being the most common method for single zones.

By means of the recorded concentration over time, the AER can be calculated using the following equation [27]:

$$C(t)-C_{ext}=(C_0-C_{ext})e^{-\lambda t}$$

(1)

where C(t) is the indoor CO₂ concentration at the time instant t, C_ext is the outdoor environment CO₂ concentration, C₀ is the indoor initial CO₂ concentration at time t = 0, and λ is the AER (symmetrical to the numerical constant in the exponent). Equation (1) can be simplified as follows:

$$Y=K{e}^{-\lambda X}$$

(2)

Alternatively, logarithms can be applied to Equation (1) to obtain Equation (3), which allows the linearization of the concentration evolution, as simply described in Equation (4).

$$\ln \left(C(t)-C_{ext}\right)=\ln \left(C_0-C_{ext}\right)-\lambda t$$

(3)

$$Y=mX+b$$

(4)

Here, the fitting can be performed through linear regression, and the AER is the symmetric of the slope of the obtained line. It should be noted that an outdoor ambient CO₂ concentration of 450 ppm is assumed for all tests in this research.

The objective of this research is to develop, test, and evaluate the performance of several mechanical DCV strategies. The mechanical DCV strategies are defined considering the indoor concentration of CO₂ and the occupancy level as the input signals and the fan speed/power as the output signal. The ventilation control strategies are settled and implemented using the applications developed in LabVIEW 2016, and are tested on the experimental prototypes of ventilator boxes at the I2L. The I2L has been developed as a new platform for research and technology demonstration related to indoor environmental quality (IEQ) and energy consumption at DEM, UC, enabling researchers to access continuously monitored data on IEQ and energy use for all aspects of building functioning.

2.2. Experimental Setup and Functionality Test

For this research, the first step was to design and build the mechanical ventilator boxes. Two ventilator boxes were built using both linear (tangential) and centrifugal blowers (ebm-papst, Mulfingen, Germany). One was equipped with a DC centrifugal blower with 14 watts of input power and a nominal voltage ranging from 16 to 28 volts, and the other was equipped with a DC linear blower with 8.7 watts of input power and a nominal voltage ranging from 8 to 14 volts. The mechanical ventilator boxes were installed in the I2L in order to be tested. Figure 1 presents the prototype installation process at the I2L office.

Moreover, an infrared occupancy counter sensor [28,29] was developed and installed at the entrance of the I2L. This was implemented with two units of SHARP infrared distance sensors, installed next to each other at the entrance and perpendicular to the occupants’ movement direction. The occupancy counter detects the direction of movement of the occupants when passing through the office door (entering/leaving). Then, the signals from the infrared sensors were treated using the application developed in LabVIEW, resulting in the real-time occupancy level being presented on the screen. Figure 2 shows the development phase of the occupancy counter and the installed sensor at the entrance of the I2L. This counter is used to collect information about the occupancy level in the room to pilot the ventilation strategy algorithms.

Furthermore, as shown in Figure 3, the I2l was equipped with an Extech (Pittsburgh, PA, USA) SD800 device which was used to record CO₂ concentration at a rate of one sample per minute. This sensor enabled testing the CO₂-based ventilation strategies using the accurate real-time concentration level recorded by the sensor.

Before running the tests for ventilation control strategies and in order to check the functionality of the developed testing platform, an example test was performed overnight. To undertake this, when all windows were closed, as shown in Figure 4, CO₂ was injected into the room environment using a fire extinguisher and mixed.

When the concentration reached an amount around 1800 ppm, the ventilator was switched on at maximum power and everyone left the room to assess the AER of concentration decay during a non-occupancy period. This example test enabled us to find out the possible errors in the experimental setup and troubleshoot them to achieve more precise results. Figure 5 illustrates the concentration decay observed during this test, along with the linear regression of the concentration difference between the indoor and outdoor environments. The red line represents the selected profile chosen for the calculation of the AER. The AER was determined to be 0.59 h⁻¹, and the coefficient of determination (R² = 0.9978) indicates a strong correlation between the data and the fitted regression line.

Additionally, a test was conducted with the ventilator switched off to evaluate the air exchange rate under natural ventilation conditions. Including both scenarios—maximum fan power and complete shutdown—provides a representation of two operational extremes. This approach offers a more comprehensive understanding of the full range of achievable AER values, as well as insight into the baseline ventilation performance of the space in the absence of mechanical support. In comparison, while the fan-assisted condition yielded an AER of 0.59 h⁻¹ and reduced the CO₂ concentration to 500 ppm in approximately 4 h, the latter case resulted in a considerably lower AER of 0.18 h⁻¹, requiring over 16 h to reach the same concentration level. This difference underscores the significant contribution of mechanical ventilation to air renewal in the room.

Although the example test was carried out using both centrifugal and linear ventilator boxes, only the results from the linear blower are presented above. This is due to better performance of the linear fan compared to the centrifugal one, as well as its lower energy consumption and reduced noise level, which enhances occupants’ acoustic comfort. For the same reason, the results from the DCV strategies using the linear ventilator box are presented in this article.

2.3. Control Strategies

2.3.1. Control Strategy 1 (Occupancy Period-Based)

In this strategy, regardless of the number of occupants, the linear fan operates at maximum power whenever the room is occupied, and when the room is unoccupied, the ventilation system remains off. This approach follows a CAV ventilation principle, ensuring a consistent airflow rate whenever the space is in use.

2.3.2. Control Strategy 2 (Occupancy Level-Based)

Unlike the previous strategy, this approach adjusts the fan speed based on the number of occupants in the room. The linear fan operates at a specific percentage of its maximum power, set at 20% per occupant, ensuring a more dynamic and efficient ventilation response. This strategy follows the VAV principle, modulating airflow in real time to match occupancy levels. Figure 6 shows the flowchart of the two control strategies.

2.3.3. Control Strategy 3 (Double Set-Point ON-OFF CO₂-Based)

This strategy regulates ventilation based on indoor CO₂ concentration levels, ensuring that fresh air is supplied only when needed. Two set-points are defined: a maximum threshold of 1000 ppm and a minimum threshold of 500 ppm. The ventilation system remains off until the CO₂ concentration exceeds the maximum set-point, at which point the fan is activated to reduce indoor pollutant levels. Once the concentration decreases to 500 ppm or below, the fan switches off, preventing unnecessary energy consumption. This approach also follows the CAV ventilation principle, as the fan always operates at full capacity when activated.

2.3.4. Control Strategy 4 (Multi-Level CO₂-Based)

This strategy dynamically adjusts ventilation rates based on indoor CO₂ concentration levels using three predefined set-points: 1000 ppm, 750 ppm, and 500 ppm. Unlike the ON-OFF approach, the ventilator operates at different power levels depending on the CO₂ concentration, following a VAV control principle.

-

Above 1000 ppm: The ventilator runs at maximum power to quickly reduce CO₂ levels.

-

Between 750 ppm and 1000 ppm: The fan operates at a moderate power level to maintain adequate IAQ.

-

Between 500 ppm and 750 ppm: The ventilator runs at a lower power setting, ensuring minimal energy use while still providing fresh air.

-

Below 500 ppm: The fan turns off.

2.3.5. Control Strategy 5 (Modulating CO₂-Based)

This strategy regulates ventilation by continuously adjusting the fan speed based on real-time CO₂ concentration levels, following a VAV control principle. Two set-points are defined: a maximum threshold of 1000 ppm and a minimum threshold of 500 ppm. The ventilator’s operation is modulated as follows:

-

Above 1000 ppm: The fan runs at maximum power to rapidly lower CO₂ levels.

-

Between 500 ppm and 1000 ppm: The fan operates at a variable speed, adjusting proportionally to the difference between the measured CO₂ concentration and the two set-points. The closer the concentration is to 1000 ppm, the higher the fan speed, and vice versa.

-

Below 500 ppm: The fan turns off.

Figure 7 shows the flowchart of the three control strategies.

3. Results and Discussion

The two occupancy-based strategies were evaluated over two entire days at the I2L. Figure 8 and Figure 9 show the indoor CO₂ concentration and the occupancy level in the test room for strategies 1 and 2, respectively.

These figures show that during the occupancy period, the concentration slightly exceeded the popular value (1000 ppm) [30], but still maintained a good IAQ, while the fan energy consumption in the first strategy is more than twice that of the second strategy. While control strategy 1 is simple to implement, it does not account for variations in occupancy levels or indoor air quality fluctuations, potentially leading to over-ventilation and unnecessary energy consumption when fewer occupants are present. However, it guarantees that fresh air is supplied during occupancy, contributing to improved IAQ and occupant comfort. This emphasizes the importance of well designed and implemented DCV to avoid over ventilation. In control strategy 2, by scaling ventilation rates according to the number of occupants, this approach optimizes energy consumption while maintaining IAQ. Compared to CAV ventilation, it reduces over-ventilation during periods of low occupancy and enhances energy efficiency. However, the effectiveness of this strategy depends on accurate and real-time occupancy detection to ensure appropriate air exchange rates.

Table 1. Results of occupancy-based ventilation control strategies.

Control Strategy	Maximum CO2 Concentration (ppm)	Average CO2 Concentration (ppm)	Fan Energy Consumption (kWh/Day)	Occupancy Period (Hours)	Average Occupancy (Person)
CS_1	1077	822	0.081	9.3	1.8
CS_2	1094	948	0.034	9.2	1.9

Note: The cumulative CO₂ exposure is also calculated for strategies 1 and 2, resulting in 816 and 955 ppm·h, respectively.

represents the maximum and average CO₂ concentration and average occupancy level in the I2L during the occupancy period, as well as the energy consumption for each strategy. As the occupancy level is not the same in the two experiments, a definite comparison of the results cannot be made, but it is obvious that the second strategy performed at almost the same level considering IAQ, while delivering a noticeable energy saving of about 58%. Although the preliminary results suggest strategy 2′s efficiency, larger-scale studies are still needed to confirm robustness across diverse occupancy patterns.

The performance of the three CO₂-based control strategies (control strategies 3, 4, and 5) was evaluated during unoccupied periods over three consecutive nights. During these tests, all windows and doors remained closed to ensure a controlled environment. Unlike the occupancy-based tests, where metabolic CO₂ was the primary source, a fire extinguisher was used to inject CO₂ into the room, raising the concentration to approximately 2000 ppm before leaving. This setup allowed for a controlled and repeatable evaluation of the ventilation strategies by monitoring the concentration decay between 1500 ppm and 500 ppm, ensuring a fair comparison of each approach. Figure 10 presents the concentration decay curves, the corresponding linear regression fits, and the fan operation power levels for each control strategy. The data revealed key differences in how each strategy managed CO₂ levels and energy consumption.

The comparison of ventilation performance in the CO₂-based control strategies shows that the choice of control logic significantly impacts both energy efficiency and IAQ. The double set-point ON-OFF strategy (control strategy 3) effectively maintained IAQ, but at the cost of higher energy consumption due to its binary operation mode. The fan remained inactive until CO₂ levels exceeded 1000 ppm, at which point it switched on at full power until the concentration dropped below 500 ppm, leading to frequent activation cycles and noticeable fluctuations in IAQ. In contrast, the multi-level CO₂-based strategy (control strategy 4) introduced intermediate set-points at 1000 ppm, 750 ppm, and 500 ppm, allowing the fan to operate at three discrete power levels. This approach smoothed out concentration variations and improved energy efficiency, consuming 33% less energy than strategy 3.

Among the tested strategies, the modulating CO₂-based control (control strategy 5) demonstrated the best overall performance, dynamically adjusting the fan speed proportionally to real-time CO₂ concentrations between the two set-points. This method provided a continuous and adaptive ventilation response, preventing abrupt changes in IAQ while minimizing energy use. As a result, control strategy 5 achieved the lowest energy consumption, reducing energy consumption by 60% compared to control strategy 3 and 40% compared to control strategy 4, while maintaining stable CO₂ levels throughout the test period.

Although these three ventilation strategies were tested during the non-occupancy period, the results, summarized in

Table 2. Results of CO₂-based ventilation control strategies.

Control Strategy	AER (h−1)	Average CO2 Concentration (ppm)	Fan Energy Consumption (kWh)	Decay Period (Hours)
CS_3	0.61	791	0.04	4.6
CS_4	0.46	747	0.027	6.2
CS_5	0.31	716	0.016	8.2

, show that the strategies were well implemented and completely functional. Moreover, the results indicated that linking ventilation power to multiple set-points or continuously modulating it in response to air quality fluctuations (proportional to the real concentration level) is a more effective approach for optimizing both IAQ and energy efficiency in demand-controlled ventilation systems.

4. Conclusions and Future Work

This study experimentally evaluated different demand-controlled ventilation strategies to improve indoor air quality while optimizing energy consumption in an office environment. The results highlight the significant role of smart ventilation control in balancing IAQ and energy efficiency, demonstrating that integrating sensor-based control mechanisms can lead to substantial improvements compared to traditional ventilation approaches. By employing occupancy-based and CO₂-based ventilation strategies, the study investigated how real-time monitoring and automated control can enhance ventilation efficiency while ensuring occupant comfort.

The findings indicate that occupancy-based strategies, which rely on simple sensor triggers, provide an effective means of reducing ventilation energy use by 58% compared to CAV systems that operate continuously. However, more sophisticated CO₂-based strategies allow for better adaptability to indoor pollutant levels, thereby ensuring that ventilation rates dynamically respond to actual IAQ conditions rather than just occupancy presence. Among the tested strategies, modulating CO₂-based control (control strategy 5) proved to be the most effective, offering a dynamic and proportional response to pollutant levels, minimizing both over-ventilation and under-ventilation, and achieving the highest energy savings (about 60%).

Despite the advantages of intelligent ventilation strategies, their performance and effectiveness depend on several key factors, including accurate sensor calibration, real-time data processing, and responsive control algorithms. Reliable CO₂ sensors are essential for ensuring precise measurements, as any calibration errors can lead to over-ventilation or insufficient air exchange, compromising both energy efficiency and occupant health. Additionally, the adaptability of DCV strategies to dynamic occupancy patterns, environmental conditions, and system constraints further influences their efficiency. While simpler ON-OFF control strategies (such as control strategy 3) may be easier to implement, they tend to create more ventilation cycles that lead to unnecessary energy consumption. More advanced approaches (such as control strategies 4 and 5) require well-calibrated sensors and fine-tuned control mechanisms to prevent system inefficiencies and ensure stable IAQ, with average CO₂ concentrations of 747 and 716 ppm, respectively. Therefore, the choice of ventilation strategy should be carefully considered based on installation complexity, sensor reliability, and operational goals.

This study has several limitations: (1) As a single-site lab experiment in a 25 m² test room, the results may not fully represent real offices with diverse layouts or HVAC systems. (2) Focusing solely on CO₂ excludes other pollutants (VOCs/PM2.5) needing different ventilation responses. (3) The findings may not scale directly to larger spaces. (4) Occupant behaviors (window-opening, personal adjustments) were not assessed. We recommend community trials to verify energy savings, evaluate multi-zone contamination risks, and test maintenance requirements in real workplaces.

As the global demand for energy-efficient buildings grows, future research should focus on further enhancing DCV strategies through machine learning and predictive analytics. Real-time data-driven models could improve ventilation efficiency by anticipating occupancy patterns and adjusting air exchange rates proactively. Additionally, integrating multi-pollutant sensing (e.g., VOCs, humidity, and particulate matter) into ventilation control algorithms could provide a more comprehensive approach to IAQ management.

Future work should also explore the impact of climate conditions and building characteristics on the performance of DCV strategies. Expanding experimental studies to different building types, including residential and commercial spaces, could provide valuable insights into how these strategies perform under varying conditions. Furthermore, hybrid ventilation systems that combine natural and mechanical ventilation in an intelligent manner could be investigated to maximize both sustainability and operational efficiency [31]. Building on this research, the authors are developing EnviroSense—Smart Well-being for Every Space, an innovative project designed to develop a portable set of sensors monitoring IEQ and occupant health in real time, focused on enhancing well-being and energy efficiency. By leveraging advanced technologies for environmental monitoring and control, the project aims to propose smart solutions that promote healthier and more sustainable indoor environments. Moreover, complementary to this study, which demonstrated the technical feasibility of the proposed strategies, future research should include comprehensive cost–benefit analyses encompassing equipment, installation, and maintenance costs versus energy savings to facilitate real-world adoption by building operators.

In summary, this study reinforces the importance of smart control systems in buildings, demonstrating that linking ventilation power to real-time environmental conditions can significantly improve both energy efficiency and IAQ. As technology advances, the continued development of adaptive, data-driven control strategies will be essential for creating healthier and more sustainable indoor environments.