Monitoring Indoor Air Quality in Buildings: An Overview of Measuring Devices and Main Challenges for a Correct Operation

Abstract

This research delves into the foundational elements of thermal comfort, crucial alongside visual comfort, acoustic comfort, and air quality for ensuring the quality and sustainability of living environments. With increasing recognition of thermal comfort’s implications across various human activities, from energy management for efficiency to considerations of environmental impact and economy, its comprehensive understanding is paramount. The study scrutinizes the prevailing methodology of evaluating comfort via true thermal sensation rating, dissecting the involved variables and their relative significance in determining comfort levels. Following this analysis and parameter definition, a comparative assessment of diverse sensors was conducted to gauge measurement accuracy concerning key variables of interest, thereby identifying the most suitable sensor for real-world applications. Conducted at the ENEA research center in Rome, the study executed an experimental setup within one of the center’s offices. Subsequently, reflections were made on the feasibility of providing indoor comfort indications amidst variable data availability, exploring potential simplifications and approximations to streamline comfort index evaluations.

1. Introduction

Indoor air quality (IAQ) monitoring within buildings is a critical endeavor, impacting the health, productivity, and comfort of occupants. This article provides a comprehensive overview of the various measuring devices employed for monitoring IAQ, along with an examination of the primary challenges encountered in ensuring their effective operation. Additionally, we present a detailed case study where both commercial and experimental sensors are analyzed, showcasing their respective characteristics and performance. With a growing emphasis on healthy indoor environments, understanding the landscape of IAQ measurement tools and addressing associated challenges, as evidenced by real-world applications, is paramount for maintaining optimal indoor air quality standards [1].

EN ISO 7726 [2] standard is a document in the series of international standards that specifies the minimum requirements of instruments for measuring physical quantities describing an indoor environment, providing recommended measurement approaches considering pragmatic options and performance of the available measuring instruments. The specifications contained in this standard are divided into two classes, according to the magnitude of the monitored quantities. The first is defined as “Type C” describing specifications and methods for measurements carried out in moderate environments, for instance how to record thermo-hygrometric comfort conditions following the indices of EN ISO 7730 [3]; the other class is “Type S” containing specifications and methods for environments exposed to thermal stress by the indications of EN ISO 7933 [4] for hot environments and EN ISO 11079 [5] for cold environments. The suitable parameters for determining thermo-hygrometric comfort or thermal stress indexes can be divided into basic and derived physical quantities. The firsts are used to define comfort or thermal stress indices such as air temperature (in K or °C), air velocity (in m/s), mean radiant temperature (in K or °C), and absolute humidity expressed as the partial pressure of water vapor (Pa). The derived physical quantities are evaluated as a function of the sensor features and are employed for determining empirical indexes related to thermo-hygrometric comfort conditions or thermal stress avoiding deterministic methods that rely on the thermal exchanges between the human body and surroundings. Methods of measuring the environmental physical properties must take into account characteristics changeable with position and time, for instance, a human body considered sitting or standing.

2. Physical comfort variables

Physical comfort is a multifaceted construct encompassing various environmental factors that influence an individual’s subjective experience of well-being and satisfaction. These factors, referred to as physical comfort variables, play a pivotal role in shaping the overall comfort level within indoor environments. Understanding and effectively managing these variables is essential for optimizing occupant comfort, productivity, and overall quality of life. Key physical comfort variables include air temperature, relative humidity, air velocity, and mean radiant temperature, among others. Each of these variables interacts dynamically to create a thermal environment that can either promote or hinder comfort. Additionally, factors such as clothing insulation, metabolic rate, and individual preferences further modulate the perception of comfort, highlighting the intricate interplay between environmental conditions and personal characteristics.

2.1 Air temperature

Among the quantities involved in microclimate evaluations, air temperature is often used to measure the dry-bulb temperature, most easily measurable in light of the several consolidated technologies available, both by traditional or innovative probes. Traditional technologies include liquid or solid expansion thermometers, resistance thermometers, thermocouples, and platinum resistance thermometers. Innovative technologies involve the employment of optic fibers and infrared thermometers. However, for this sensor typology, it is required to shield it adequately to prevent a measurement affected by thermal radiation emitted from surrounding bodies at different temperatures than the air temperature. In order to give a quantitative indication, a not shielded sensor or inappropriately shielded installed inside an indoor environment with an average radiant temperature 10°C higher than the air temperature, errors of even more than 1°C could be detected. The EN ISO 7726 [2] standard establishes the required features of sensors employed for air temperature measurements. For class C measurements, instruments with a maximum desirable error of less than ±0.2°C and required ±0.5°C are recommended, other parameters instead are given for class S. Regarding the response time, the standard states that this should be as small as possible for both classes, but without specifying a limiting value. However, it is recommended that the measurement duration should be at least 1.5 times the instrument response time, because a high response time may be critical for the monitoring of parameters in non-steady-state conditions.

2.2 Mean radiant temperature

Different solutions are available for the measurement of the mean radiant temperature in a confined space, therefore also different measuring instruments can be employed for this purpose as indicated in the standard EN ISO 7726 [2]. The mean radiant temperature assumes a significant role in the evaluation of thermal environment [6] features because it represents the uniform temperature of a fictitious black cavity in which a subject would exchange the same amount of radiant energy as that could be exchanged in the real non-uniform environment (the correspondent symbol is denoted by tr). The standard outlines three techniques for measurement: using a globo-thermometer, employing a two-sphere radiometer, and utilizing a constant air temperature sensor. Additionally, it introduces two calculation approaches grounded in the view factors between the cavity’s surface and the radiant temperatures of the plane.

For this parameter, the most widespread instrument is undoubtedly the globo-thermometer because it is also the cheapest sensor, which, however, suffers from different drawbacks, such as the high response time (normally more than 20 minutes), which leads to evident issues when numerous measurements are requested for a limited period, and to overestimation due to the perfect spherical shape that does not consider the radiant contributions of horizontal surfaces adequately. Moreover, in moderate environments, the globe-thermometer does not allow calculation of the radiant temperature asymmetry, indispensable to determine thermo-hygrometric comfort conditions in large glazed environments (panes have a surface temperature much lower than opaque surfaces). The globe-thermometer is made of a very thin and opaque black metal sphere with an assigned diameter of 15 cm whose emission coefficient is about 0.95. A temperature sensor is instead placed inside the sphere so that the globe temperature (that after a proper time reaches the thermal equilibrium with the cavity) is determined by including the effects of the body’s radiative and convective exchanges. Once the convective share is determined through temperature and air velocity, it is possible to extrapolate the average radiant temperature of the environment. In particular, the mean radiant temperature is evaluated using the equations introduced in Annex B of EN ISO 7726 [2]. The following empirical equation, for instance, is used assuming natural convection and when the globe temperature tg refers to a standard instrument with a diameter of 0.15 cm.

$$t_r={\left[{\left(t_g+273\right)}^4+0.4·{10}^8·{\left|t_g-t_a\right|}^{1/4}·\left(t_g-t_a\right)\right]}^{1/4}-273$$E1

with ta is the air temperature. Another issue related to measurements carried out by a globe-thermometer is due to the measurement conducted at a single point that is not representative of the entire radiative field relative to the subject.

Regarding the available theoretical models, the view-factor method is more accurate and it determines the average radiant temperature starting from the measurement of the surface temperatures for each wall inside the cavity, in turn detected by contact or distance methods. Anyway, the estimation of the view factor between the person and the surrounding surfaces is required and its calculation is quite difficult. Indeed, the view factor is a function of shape, size, and relative positions concerning the person and the surface. Compared with the other methods, the view factor method benefits from better accuracy in determining radiant contributions and the ability to estimate asymmetric conditions as well. However, problems such as calculating the view factors in complex spatial geometries and measuring the mean surface temperature arise.

An alternative measurement can be conducted through a radiometer, which consists of a flat thermal element with high emissivity measuring the flux of the incident radiant energy by a thermopile, representing the sensing element, as a function of absorbed heat. The sensor temperature Ts is then directly related to the radiant plane temperature Tpr according to a balance equation. The average radiant temperature can then be evaluated by measuring the radiant plane temperature in the six space directions and the projected area factors. Radiant asymmetry can also be assessed by using the same approach.

2.3 The relative air velocity

The measurement of air velocity is quite complex, both because of its rapid temporal fluctuation and its vector nature with great variability in intensity and direction. Omnidirectional probes allow measurement of the modulus of velocity regardless of their placement, while one-way probes must be placed orthogonally to the flow and must be more than one in number to allow measurement of individual velocity components. The velocity modulus can be determined using a single omnidirectional probe, such as a hot bulb probe, or in the case where the flow is unidirectional with a single one-way probe, such as a hot-wire anemometer. The Annex E of the standard EN ISO 7726 [2] considers the use of wire/hot film anemometers, whirlwind anemometers, and ultrasonic anemometers in the measurement of wellbeing, but the direction sensitivity and response time of the instrument should be evaluated in advance.

In wire and hot-film anemometers, velocity monitoring is based on measuring the exchange of thermal energy between the sensing element and the surrounding air. The anemometer consists of a solid body, which in the case of wire sensors has the shape of a cylinder or with different shapes for film sensors, such as that of a sphere. The sensor body is heated electrically by a higher temperature than the surrounding air, giving up thermal energy mainly by convection. The heating power, the temperature of the element, and the air temperature allow for calculating the air velocity by the convective heat transfer coefficient. Hot-sphere sensors are isotropic, namely, they show equal sensitivity to flows from all spatial directions, but the significant mass makes these sensors slower in response and unsuitable for tracking the rapid changes in a turbulent airflow. Hot-wire sensors are directional and therefore their sensitivity varies significantly between the perpendicular plane to the wire and the parallel direction to the wire anyway they are fast in response. Whirlwind anemometers consist of a small impeller with blades or cups suspended in the flowing stream, with their axis of rotation coaxial or perpendicular to the flow direction. They are therefore unidirectional sensors and it is needed to know in advance the velocity vector track, moreover, they are typically less sensitive. Ultrasonic anemometers use special phenomenology related to the propagation of ultrasonic waves through a moving fluid for air velocity detection. Based on the physical principle used, they can be divided into transit time meters and meters and the Doppler effect. In both, the propagation of pressure waves at frequencies higher than those audible by the human ear and propagating within the fluid current is exploited. Hot-wire anemometer normally produces negligible errors and is indicated in the context of PMV calculation. If the velocity is very low and no dominant direction of flow is identified, the effect of air velocity is inevitably small. If, on the contrary, there is appreciable air flow, its direction becomes obvious and the sensor can be oriented accordingly to maximize its sensitivity. More accurate velocity meters such as Pitot tubes can be used to periodically check anemometers, and the instrument can be calibrated at least every 2 years.

2.4 Relative humidity

The humidity content in air has a significant influence on an individual’s thermo-hygrometric balance, especially on evaporative transmissive share. High air humidity drastically reduces the evaporation of sweat, and in the summer regime, this rate can account for as much as 70–80% of the entire heat transfer. In thermo-hygrometric air conditioning problems, atmospheric air is considered a binary mixture made of dry air and water vapor called “moist air.” In turn, dry air is a mixture of gases consisting mainly of nitrogen and oxygen, therefore this component is approximated as a single component with ideal gas behavior with invariable composition. So, moist air is considered as a mixture of perfect gases, however with different properties than mixtures of perfect gases because the water vapor can be subjected to phase exchange processes, altering the composition of the mixture. Moist air can contain an amount of vapor varying from zero (dry air) to a maximum value (saturated humid air) that depends on the temperature and pressure levels of the mixture. At any given temperature, air cannot contain more than a certain amount of water vapor; above that amount, water vapor condenses forming a liquid phase. In these conditions, the saturation point at a precise temperature and pressure is reached. As the temperature of the air increases, the maximum amount of water vapor it can contain also increases, consequently humid air mixtures at high temperatures contain more water vapor than the same mixtures at low temperatures. To quantify the magnitude of the water vapor within the mixture, the concept of absolute humidity (or humidity ratio) is defined, to consider how much the most air is distant from the saturation point, relative humidity instead is introduced.

Absolute humidity refers to the quantity of water vapor, measured in grams, present within one cubic meter of air under specific temperature and pressure conditions. As noted earlier, absolute humidity typically rises as temperature increases. To better highlight its physical meaning, it is expressed in kg of vapor (kg^v) per kg of dry air (kg^a) [kg^v /kg^a or g^v /kg^a, the last more used considering that the quantity of water vapor is very smaller than the mass of dry air]. Absolute humidity is measured directly, using hygrometers with lithium chloride salts, hair, and dew point. Its value varies between zero (complete dry air) and an infinity value, namely when the mixture is composed exclusively of water vapor.

Relative humidity (ϕ) is the ratio of the mass m^v of water vapor present in a certain volume V of moist air to the mass m^s of vapor containable under saturated conditions at the same temperature in the same volume V of moist air. It is also expressed as the ratio of the vapor pressure to the saturation pressure of vapor at the temperature ta of the moist air mixture.

$$\phi =\frac{p_v}{{\left|p_s\right|}_{\mathit{ta}}}$$E2

Relative humidity is the most widely used hygrometric parameter for practical reasons. Since the capacity of air to hold water vapor increases with temperature, it follows that, leaving the pressure and the amount of water vapor contained in a given volume of air unchanged, relative humidity decreases as temperature increases and vice versa. This is because as temperature increases, more water vapor is required to reach saturation. Relative humidity is measured indirectly with a psychrometer (Figure 1). This instrument consists of two temperature probes: the first probe is in direct contact with the air, measuring the “dry bulb temperature,” the second probe is wrapped in moistened gauze and equipped with a ventilation system that facilitates the evaporation of water, measuring the “wet bulb temperature.” Due to the cooling effect of evaporation (the water absorbs heat from the air to vaporize) the wet probe measures a lower temperature and the humidity of the air increases in light of the greater vapor quantity. By comparing the two temperatures using appropriate diagrams (called psychrometric) or algorithms, the relative humidity value is detected.

Figure 1.

Psychrometric diagram with ET* curves, constant for M/ab = 1.2 met, Icl = 0.60 clo, var. = 0.20 m/s.

Relative hygrometers, such as hair hygrometers, capacitive and resistive electrical hygrometers can measure relative humidity directly. The principles of measurement (mechanical, electrical, resonance, etc.), production technologies (thin-film, thick-film, solid-state, etc.), and materials used (polymeric, ceramic, etc.) are the most varied allowing this parameter to be measured over a wide range (10–100%), with an uncertainty typically of 2–3% and at best less than 0.5%.

2.5 The operating temperature, equivalent temperature, and effective temperature

To characterize a confined thermal environment using fewer parameters, and avoiding measuring the mean radiant temperature (the monitoring of which is laborious), built-in parameters have been introduced. The three most important are operating temperature (to), equivalent temperature (teq), and effective temperature (ET*). The integrated parameters combine the influence of individual parameters on energy dissipation in the following way:

1. to integrate the effect of ta + tr

2. teq integrated effect of ta + tr + va

3. ET* integrated effect of ta + tr + pa

The dependence of PMV on air temperature and mean radiant temperature, which determines convective and radiative exchanges, respectively, is traced back to that from a single variable, a linear combination of these two temperatures and is named operating temperature too. The operating temperature is rigorously defined as:

$$t_0=\frac{h_r·t_r+h_c·t_a}{h_r+h_c}$$E3

where

Ta is the air temperature, in [°C],

Tr is mean radiant temperature, in [°C],

hc is the convective heat transfer coefficient, in [W·m⁻²·K⁻¹],

hr is the radiative heat transfer coefficient, in [W·m⁻²·K⁻¹].

The operating temperature corresponds to the temperature actually felt by people that takes into account the main heat exchange mechanisms affecting the human body on the feeling of well-being. Alternatively, t0 can be calculated, in a simplified way, by the equation:

$$t_0=A·t_a+\left(1-A\right)·t_r$$E4

where A, listed in Table 1, represents a parameter related to the relative air velocity. It can be appreciated that the higher the air velocity, that is, convective heat transfer coefficient, the greater the weight of air temperature and the lower the weight of mean radiant temperature.

Var	Var < 0.2	0.2< Var < 0.6	0.6 < Var < 1.0
A	0.5	0.6	0.7
Note: ${\mathrm{V}}_{\mathit{ar}}={\mathrm{V}}_a+0.0052·\left(\frac{M}{\mathit{Ab}}-58.15\right)$

Table 1.

Values of parameter a as a function of relative velocity expressed in m·s⁻¹.

Another simplified formula can be adopted:

$$t_0=\frac{t_r+t_a}{2}$$E5

but only when precise conditions are respected. In particular, the human metabolism (related to the activity and measured in met following the Fanger theory, see EN ISO 7730) must range between 1.0 and 1.3 met, air velocity lower than 0.2m·s⁻¹, absence of direct solar radiation and difference between air temperature and mean radiant temperature below 4°C.

Starting from the consideration that relative humidity has a negligible effect on thermal comfort conditions, Dufton in 1929 had the idea of combining the other three physical parameters (ta, tr, and va) into a single index defined equivalent temperature teq, and defined as the temperature of the thermally uniform fictitious environment (ta = tr) with tinned air in which a person would exchange the same dry heat power as in the real environment. In residential environments, the use of the operating temperature is more recommended, considering that va generally takes on very low values. In different cases where va takes on values that are not considered negligible, such as analysis inside vehicles, the equivalent temperature comes into play. One of the last equations suggested for calculating teq is that of Madsen (1979):

When v_a > 10m/s

$$t_{\mathit{eq}}=0.55·t_a+0.45·t_r+\left(0.24-0.75·\sqrt{v_a}\right)·\left(36.5-t_a\right)/(1+I_{\mathit{cl}}$$E6

$$\text{When}{\mathrm{v}}_{\mathrm{a}}\le 10\mathrm{m}/\mathrm{s}{t}_{\mathit{eq}}=0.5·\left(t_a+t_r\right)$$E7

The ET* index, known as new effective temperature, considers the temperature of a fictitious environment at a uniform temperature (ta = tr) and with a hygrometric degree of 0.5 in which a human body would exchange by convection (C), radiation (R), and evaporation (E) the same amount of heat that could exchange in the real environment, at parity of skin temperature tsk and the same percentage of wet skin w as in the real environment. The calculation of ET* is carried out by using a thermoregulation model, and knowing the values of the six variables on which the global thermal discomfort depends, and the values of tsk, w, and (R + C + E) are calculated. Then, the temperature of the uniform fictitious room ta* = tr*, at ϕ = 0.5 in which, with real tsk and w, the subject would exchange a heat output (R* + C* + E*) = (R + C + E), is determined. The ta* thus calculated is precisely ET*. A psychrometric diagram (Figure 1) allows for determining ET*, as a function of the partial pressures on the abscissae and the operating temperatures on the ordinates, by setting the values of air velocity, metabolic rate, and clothes thermal resistance.

To summarize, Table 2 lists the main characteristics of measuring instruments according to the EN ISO 7726 prescriptions.

Quantity	Symbol	Class C (comfort)			Class S (thermal stress)
		Measuring range	Accuracy	Response time (90%)	Measuring range	Accuracy	Response time (90%)
Air temperature	ta	10–40°C	Required: ±0.5 °C Desirable: ±0.2 °C These levels shall be guaranteed at least for a deviation |tr-ta| = 10°C	The shortest possible. Value to be specified as characteristic of the measuring instrument.	−40°C to +120°C	Required: −40–0°C: ±(0.5 + 0.01|ta|)°C > 0–50°C: ±0.5°C > 50–120°C: ±[0.5 + 0.04 (ta−50)]°C Desirable: required accuracy/2; These levels shall be guaranteed at least for a deviation|tr−ta| = 20°C	The shortest possible. Value to be specified as characteristic of the measuring instrument.
Mean radiant temperature	tr	10–40°C	Required: ±2°C; Desirable: ±0.2°C; These levels are difficult or even impossible to achieve in certain cases with the equipment normally available. When they cannot be achieved, indicate the actual measuring precision.	The shortest possible. Value to be specified as characteristic of the measuring instrument.	40°C to +150°C	Required: −40–0°C: ±(5 + 0.02|ta|)°C; >0–50°C: ±5°C; >50–150°C: ±[5 + 0.08 (tr−50)]°C; Desirable: −40–0°C: ±(0.5 + 0.01|tr|) °C; >0–50°C: ±5°C; >50–120°C: ±[0.5 + 0.04 (tr−50)]°C	The shortest possible. Value to be specified as characteristic of the measuring instrument.
Air velocity	Va	0.05 m/s to 1m/s	Required: ±(0.05 + 0.05 Va) m/s; Desirable: ±(0.02 + 0.07 Va) m/s; These levels shall be guaranteed whatever the direction of flow within a solid angle.	Required: 0.5 s; Desirable: 0.2 s	0.2 m/s; to; 20 m/s	Required: ±(0.1 + 0.05 Va)m/s; Desirable: ±(0.05 + 0.05 Va)m/s; These levels shall be guaranteed whatever the direction of flow within a solid angle.	The shortest possible. Value to be specified as characteristic of the measuring instrument. For measuring the degree of turbulence, a small response time is needed.
Absolute Humidity as partial pressure of water vapor	Pa	0.5 kPa; to; 3.0 kPa	±0.15 kPa; This level shall be guaranteed for a difference |tr-ta|of at least 10°C.	The shortest possible. Value to be specified as characteristic of the measuring instrument.	0.5 kPa; to; 6.0 kPa	±0.15 kPa; This level shall be guaranteed for a difference |tr-ta|of at least 20°C.	The shortest possible. Value to be specified as characteristic of the measuring instrument.

Table 2.

Characteristics of measuring instruments EN ISO 7726.

3. Evaluation of measurements and discomfort

Currently, there is no specific standard in the European landscape that defines the criteria to be followed when conducting monitoring aimed at assessing thermal comfort in an existing building [7]. Some guidance is provided by the standard EN ISO 7726, which indicates how methods for measuring the physical characteristics of the environment consider the parameter variation with time and space.

An environment can be considered bio-climatically homogeneous if, at any given time, air temperature, radiation, air velocity, and humidity can be considered uniform around the subject. This condition is frequently met for temperature, air velocity, and humidity, but more rarely in the case of radiation [8]. In monitoring, it is aimed at determining the comfort level, measurements should be made at several positions representative of the average and/or worst conditions around the subject. The determination of measurement locations is done to capture the heterogeneity of thermal environments. In a homogeneous indoor space, a single measurement could be made at a conventional point, such as the center of the room, or multiple measurements that correspond to the spatial gradient of thermo-hygrometric parameters.

The EN ISO 7726 [2] standard indicates at which heights the basic quantities have to be measured and the weighting coefficients to use about the class of the environment (Table 3). A space is said to be stationary when the physical quantities used to describe the thermal environment are time-independent. This condition is reached when the parameter fluctuations, about their time average, do not exceed the values obtained by multiplying the measurements with the factors given by EN ISO 7726 (Table 4) [2].

Locations of the sensors	Weighting coefficients for measurements for calculation mean values				Recommended heights (for guidance only)
	Homogeneous environment		Heterogeneous environment		Sitting	Standing
	Class C	Class S	Class C	Class S
Head level			1	1	1.1 m	1.7 m
Abdomen level	1	1	1	2	0.6 m	1.1 m
Ankle level			1	1	0.1 m	0.1 m

Table 3.

Measuring heights for the physical quantities of an environment.

Quantity	Class C (comfort) Factor X	Class S (thermal stress) Factor X
Air Temperature	3	4
Mean radiant temperature	2	2
Radiant temperature asymmetry	2	3
Mean air velocity	2	3
Vapor pressure	2	3

Table 4.

Criteria for a homogeneous and steady-state environment.

It should be noted that individual quantities used to describe the level of exposure (metabolism, heat insulation) may also be time-dependent.

Measurements should also be carried out concerning weather conditions representative of hot and cold seasons. The duration of individual monitoring should be sufficient to comprehensively represent the reference condition, also considering the response times of different probes. The results of the measurements are inevitably influenced by the methodological choices made, which are decisive to ensure the highest level of repeatability and to produce an accompanying report that contains useful information to contextualize the results.

A description of the sample environments, a description of the test conditions, the criterion used for the choice of measurement points, the duration of the measurements, the external meteorological conditions, and an accurate description of the sensor used are indicated in the following.

The conditions −0.5 < PMV < +0.5 and 5% < PPD < 10% represent necessary but not sufficient conditions for comfort in average comfortable environments. To attain actual comfort there must also be zero discomfort due to the unevenness of environmental variables, that is, there must be no local discomfort [9]. Along with PMV, which condensates the energy balance of the human body, the EN ISO 7730 standard contains some indices to describe “local” discomfort considering local fluctuations or disconformities of microclimatic quantities, such as to induce discomfort conditions in the subject. The main causes causing local discomfort are mainly due to temperature unevenness and air velocity fluctuations, which affect heat exchange with the surroundings. For this, four main causes can be considered.

4. Investigated case study

The experimental campaign was conducted in a part of the building hosting the ENEA Casaccia Research Center in Italy. The Casaccia Research Center is ENEA’s largest laboratory and facility complex and is located 25km northwest of Rome, near Bracciano Lake. The building considered is identified as Building F40 of the Department of Energy Technologies (DTE). The data were collected in an operational office located on the second floor belonging to the smart city and smart community laboratory, identified as room 105 and highlighted in the 3D model representation of Figure 2.

Figure 2.

3D model representation of the building and plan.

Room 105 has a net area of 17.5 m², dimensions 3.88 × 4.51 m and faces east with a window opening of 1.70 × 1.40 m. The room is equipped with office furniture set up for two operating desks (Figure 3). The room is heated and air-conditioned by a single fan coil. In the winter period, the fan coil is supplied with hot water from a district heating system, in summer, the chilled water comes from an autonomous chiller located in the building. The room is illuminated by six OSRAM Lumilux T8 LEDs 26 mm L 36 W/840 lamps not equipped with either dimmer or modulating ignition devices.

Figure 3.

Geometrical representation of the room 105.

4.1 Installed sensors

Table 5 shows schematically all the sensors considered in the study.

Table 5.

Sensors installed in the case study office.

4.2 Sesto senso

Sesto Senso is a patented multisensor system by ENEA consisting of a central unit that acts as a gateway, sensors that communicate with the gateway via a Z-Wave protocol, and a Z-Wave people-counting sensor whose computational algorithm was developed by ENEA itself. This sensor is derived from the optimization of a prototype (Figure 4), and the upgrade consisted of integrating new sensor modules, particularly microphones, optimizing the previous people-counting module, and integrating new hardware and software.

Figure 4.

Sesto senso prototype.

The sensors included in Sesto Senso are:

Fibaro “door/window” sensor 2 (Figure 5). In particular, two sensors were installed; one for assessing the door status and another for the window status. The detection of the opening occurs when the sensor body and the magnet are separated. Furthermore, the Fibaro door/window sensor 2 includes an integrated temperature sensor and is powered by a 3.6 V DC battery. It has dimensions of 71 × 18 × 18 mm and a temperature measurement accuracy of ±0.5°C, which is within regulatory limits.

Figure 5.

Fibaro door/window sensor 2.

Multisensor 6 from Aeotec (Figure 6), a small device for detecting motion, temperature, humidity, illuminance, UV, and vibration. Regarding accuracy concerning temperature, it does not meet the necessary limit dictated by EN ISO 7726 [2].

Figure 6.

Multisensor 6 Aeotec.

An ad-hoc people counting device was realized for the monitoring, with 2 IR TF mini LIDAR sensors, based on time-of-flight technology. The maximum detection range is 12 m and the acceptance angle is 2.3°. The object has dimensions 150 × 60 × 30 mm, with two holes for laser output, power supply is 12VDC and is via cable (Figure 7).

Figure 7.

Counting people Sesto senso.

For the measurement of CO₂ concentration, no specific sensor was used, but an indirect value was derived according to an algorithm, developed by the same research organization that owns the system. The algorithm that calculates the percentage of CO₂ has as input five environmental parameters of the room: door and window status, temperature, humidity, and number of people present; even if just one of the variables is missing, the value of CO₂ is not returned.

The temperature value reported on the control unit is that of the center of the room to have a more indicative value of the temperature distributed in the room. The room-center temperature value is also used for the calculation of CO₂, and also for the subsequent analyses covered in this discussion, only the latter will be considered referring to the Aeotec sensor.

Most of the sensors, except the people-counting sensor, are wireless and battery-powered so that the problem of connecting them to the control unit was overcome.

Table 6 lists the main characteristics of Sesto Senso measurements.

Table 6.

Technical characteristics of Sesto senso measurements.

The control unit consists of a 7″ touchscreen display whose system is developed from a Raspberry Pi 3 board. The visual interface allows the user to view all the detected parameters also through intuitive and user-friendly icons. The data-sending routine runs every two minutes and sends the last data recorded within two minutes. The touch screen, on the other hand, updates every 30 seconds (Figure 8).

Figure 8.

Sesto senso control screen.

Since Sesto Senso consists of multiple sensors, each requires specific placement according to its function. Fibaro’s door/window sensor 2, in the case of the door, was placed at a height of 1.6 m, although this is irrelevant for the relief of opening and closing, while a high height could distort the temperature reading. In the case of the window, on the other hand, considering the double side and vasistas opening, it was necessary to place the sensor on the top. These problems in sensor positioning, and also the peripheral location for the working positions, led to disregarding the temperature values detected by this sensor and focusing only on the Aeotec one. The latter was placed on the south wall at a height of 1.60 m (Figure 9). Finally, the people-counting sensor was placed on the inside of the door. This sensor presented several problems in its placement, the first of which is that it detects the opening and closing of the door as the passage of a person, thus distorting the number of people inside the room and consequently also the calculation of CO₂.

Figure 9.

Sensors position and location in room 105.

4.3 ERS CO2

ERS CO2 (Figure 10) is a sensor for measuring indoor environment conditions, developed by ELSYS. It is designed to be wall-mounted at a height of 1.6 m. It is completely wireless and powered by two 3.6 V AA lithium batteries. The sensor box with a rectangular-shaped case, measuring 86 × 86 × 28 mm, contains sensors for measuring CO levels, air temperature, humidity, illuminance, and presence. As for temperature, it falls within the “desirable” values of the EN ISO 7726 standard (Table 7).

Figure 10.

ERS CO₂ sensor.

Table 7.

Technical characteristics of ERS CO2.

The sensor was placed on the south wall and at a height that meets the datasheet requirement (Figure 11). The range of the sensor covers the area fairly accurately (Figure 12), although some areas remain uncovered. The main issue related to the sensor position concerns the presence detection. All other quantities are evaluated punctually at the sensor location.

Figure 11.

Sensor positioning.

Figure 12.

Sensor radius of action.

4.4 Motion sensor FIBARO

The motion, light, and temperature sensor; motion sensor is a product of Fibaro, a Polish brand operating in the IoT (Internet of Things) sector and a home automation company. The sensor is small in size (Figure 13), easily installed, non-invasive in indoor environments, and wireless because battery-powered. In addition to detecting motion, temperature, and light intensity, the sensor has a built-in accelerometer that detects changes in position or evidence of device tampering. Detected motion, temperature, and vibration are reported through the LED embedded in the object.

Figure 13.

Fibaro motion sensor.

The sensor, in terms of temperature, is within the range of required accuracy reported by EN ISO 7726 (Table 8) [2]. A height of 2.4 m is recommended for installation; however, the motion sensor was placed on the south wall at a height of 1.6 m near the rest of the sensors (Figures 14 and 15) to make an accurate comparison. Nevertheless, a lower position was chosen because a higher location would have distorted the temperature monitoring, which would not have been representative of the position of the office users.

Table 8.

Motion sensor specifications.

Figure 14.

Positioning of the sensor.

Figure 15.

Sensor radius of action.

4.5 Text 480: Air conditioning meter

Testo 480 is an instrument for measuring climatic parameters and is particularly suitable for thermal comfort analysis (Figure 16). Testo began as a small thermometer manufacturing company in 1957 and today is a company with a worldwide presence as a provider of measurement solutions. The instrument takes the form of a handheld with six connections for different probes, all placed on a stand. For each connected probe, the handheld displays a tab with the different quantities. The instrument has an autonomy of 17 hours thanks to the built-in battery, but it also has a power supply connection, data are saved in an internal memory and are downloadable via USB connection and text software to a personal computer.

Figure 16.

Testo 480.

The probes included for comfort assessments are:

• Globometer probe Ø 150 mm, used to measure radiant heat following ISO 72461, ISO 7726, DIN EN 277262, and DIN 334033. It has a type K thermocouple in the center to measure temperature and has an adaptation time of about 30 minutes.

• IAQ probe, integrated probe that measures CO2, humidity, temperature, and absolute pressure.

• Luxmetric probe, for measuring illuminance.

• Well-being probe, used to measure air velocity and determine the risk of drafts. It also allows for measuring the degree of turbulence, which corresponds to the magnitude of temporal fluctuations in air velocity, an important parameter for calculating the risk of drafts. This probe makes it possible to determine air velocity and draft risk as well as ambient temperature and pressure. It is a hot-wire technology probe, and with low air velocity, the displayed temperature is slightly higher, justifying the reason for using the temperature values returned by the IAQ probe for analysis.

This climatic parameter meter was chosen as the comparison item because it is the one whose technical specifications fall within all that to be standardized (Table 9).

Table 9.

Testo 480 specifications.

The probes were mounted on the provided tripod and placed in the center of the room. This choice was made to properly detect all the quantities by making them make sense within the panorama on comfort assessment. The height of the trestle was modulated about 1.3 m from the floor, a height similar to that of a seated person, and the center of the room was chosen basically for a correct measurement of the average radiant temperature. More correct measurements would have been made by placing an instrument near each workstation, so due to limitations related to instrumentation availability, an average position relative to the two workstations was chosen. The only probe that was placed differently from the others was the luxmeter. Regarding the measurement of illuminance, the difficulties encountered were different. It usually makes sense to calculate illuminance on a work surface and not on vertical walls. The problem arose because most of the sensors previously analyzed are integrated sensors, so the obvious and correct location for measuring one parameter is not necessarily correct for all the others. In light of this, the luxmeter of the control unit was placed on the south wall next to the other sensors, to at least have an illuminance value comparable with the others, and then choose the best position for the sensor in the demonstrators.

4.6 CO2-display

The CO2-display instrument, which can be installed on walls or tables, is an interesting solution when the simultaneous measurement of CO₂, humidity, and temperature is required (Figure 17). Humidity measurement is done through the ROTRONIC HYGROMER® IN-1 sensor. The instrument is configurable via the side keyboard, and the recorded data can be downloaded to a USB stick and analyzed with free software. The object has dimensions 330 × 250 × 50 mm, DC 12 VDC power supply type, and an internal memory capable of recording up to 18,000 measured values (Table 10).

Figure 17.

CO2-display.

Table 10.

Technical characteristics CO2-display.

The sensor was placed on the southernmost desk near the Testo system (Figure 18). This tool, too, should be an accurate control and comparison system for testing different commercial sensors that could be used in different indoor applications.

Figure 18.

Positioning of the sensor.

5. Analysis of winter case

After having fine-tuned the methodology, the data acquired from all the sensors installed during the winter period were compared to be able to choose the most suitable for field applications. The Sesto Senso sensor took over, and ERS-CO₂ replaced Apio-Sense. Assuming as reference the Testo 480 control unit, the data returned by Sesto Senso, in addition to having a completely similar trend, almost showed complete overlap. Even the ERS CO₂ sensor although it has a phase shift of half a degree, can be considered sufficiently accurate. The data that arouse much suspicion are those for the motion sensor from Fibaro, which returned more sparse data. Later, it was verified that the problem was not strictly with the sensor but related to the network that handled its communication with the data storage platform.

From the 12 days of data collected, Figure 19 shows detail on the monitored days, precisely Thursday, December 5, 2019 to Saturday, December 7, 2019. We can see that with almost all sensors we fall within that temperature range dictated by Italian regulations, in particular the Presidential Decree April 16, 2013 No. 74 (19–22°C). The peak in temperatures represents the hours when the room is occupied. This probably occurs due to the sum of rising outdoor temperatures, radiative effects, and indoor thermal inputs.

Figure 19.

Temperature recorded from 5/12/2019 to 8/12/2019.

As for relative humidity (Figure 20), it can be appreciated that in the hours of occupancy, there is a lowering of the monitored values. Thus, despite the internal loads, the rise in temperatures leads to a parameter reduction. We can also see that Sesto Senso and the ERS CO2 sensor, despite an overestimation of less than 5% in any case, are the ones that come closest to the reference values.

Figure 20.

Relative humidity recorded from 5/12/2019 to 8/12/2019.

Another key parameter for assessing the environment is CO₂ concentration. Three sensors were available for its assessment (Figure 21). The Testo 480 instrument and the CO₂-display agree with each other for the readings, although with a slight discrepancy. Sixth Sense, in this case, is extremely poor in its ratings. The underlying causes of the malfunction are to be found in the relief of all those quantities that contribute to the calculation of CO2.

Figure 21.

CO2 recorded from 5/12/2019 to 8/12/2019.

The first problem encountered is in the evaluation of people in the room. The sensor installed at the door, as presented in the paragraph, in addition to counting the actual entry and exit of people, also includes opening and closing the door. Another problem that distorted the evaluation was the detachment of the window opening and closing sensor. Therefore, the system, and probably the algorithm, will need an intervention.

In contrast, the most difficult quantity to detect and compare was the illuminance. The first difficulty was encountered in deciding where to place the sensor. All sensors return illuminance in lux; this quantity must be used to assess whether there is the right degree of illuminance on a work surface. Since almost all the sensors used are integrated, it was necessary to find the right compromise for surveying the different magnitudes. In addition, after an initial analysis, it was seen that the sensors returned conflicting data, so it was first thought to assess the data.

Figure 22 depicts that curves are quite different from each other, moreover, the recorded illuminance values are also significantly lower than the recommended ones (between 200 and 350 lux). Considering the Testo instrument as a reference, a deeper analysis was done to understand their trend.

Figure 22.

Illuminance recorded from 5/12/2019 to 8/12/2019.

From the Sesto Senso sensor, it is worth noting to detect a trend line and a possible corrective factor (Figure 23), but for the other two sensors, more in-depth calibrations would be needed.

Figure 23.

Sensors correlation.

Mean radiant temperature and air velocity were evaluated only with the Testo 480 control unit aimed at parametric analysis on the PMV calculation and identify possible simplifications of the calculation in cases where these quantities are not available. Regarding the air velocity (Figure 24), the range accepted by the standard is highlighted in green, and it can be appreciated that the greatest fluctuations are in the hours of occupancy, perhaps due to both the opening of doors and windows and the passage and movement of people. The average radiant temperature (Figure 25), on the other hand, undergoes significant increases in the presence of occupants.

Figure 24.

Wind data recorded from 5/12/2019 to 8/12/2019.

Figure 25.

Mean radiant temperature data recorded from 5/12/2019 to 8/12/2019.

6. Evaluation of comfort indexes

The comfort indices evaluated refer to winter data only, because of the greater accuracy in the measurements and because there would be no substantial difference from the results obtained in summer. From the analysis of the sensors and the magnitudes they return, three possible scenarios were identified (Figure 26).

Figure 26.

Possible scenarios from comfort indexes evaluation.

In the summer case, the Humidex [10] was considered in addition to the conventional PMV index (Figure 27). The Humidex Index is the bioclimatic index initially used in Canada that attempts to describe the physiological discomfort caused by the combination of heat and humidity in the air. The numerical value represents an apparent temperature that provides indications of the severity of a climatic condition. It should be used only for temperature ranges between 21°C and 55°C, and reference is made to the scale in Table 11.

Figure 27.

PMV evaluated with testo 480.

To 27°C	No discomfort
From 27 to 30°C	Feeling of Discomfort
From 30 to 40 °C	Intense discomfort. Caution: limit heavier physical activities
From 40 to 55°C	Severe feeling of discomfort. Danger: avoid efforts
Over 55 °C	Danger of death: imminent heat stroke

Table 11.

Reference scale for humidex index [10].

This index is based on a simple empirical relationship that takes into account air temperature and vapor pressure. Actually, instead of relative humidity, this index uses a related parameter, namely vapor pressure, according to the formula:

$$\mathrm{H}={\mathrm{T}}_{\mathrm{a}}+\left(0.5555\left(\mathrm{e}-10\right)\right)$$E8

where precisely, Ta = temperature in °C, while e = air vapor pressure (hPa). Since it is easier to know relative humidity (RH) rather than vapor pressure, by knowing the relation between these parameters, another version of the Humidex index can be formulated:

$$\mathrm{H}={\mathrm{T}}_{\mathrm{a}}+\left(0.5555\left({0.06}^{\ast }{\mathrm{UR}}^{\ast }{10}^{0.03\mathrm{Ta}}-10\right)\right)$$E9

In Figure 28 can be seen that with the Humidex index, there is a slight overestimation of perceived discomfort.

Figure 28.

Humidex evaluated with testo 480.

As for the PMV index, it was calculated using an Excel spreadsheet (Figure 29) in which with the help of a proper macro, the program in basic provided by EN ISO 7730 [3] was implemented.

Figure 29.

Excel spreadsheet for calculation of PMV and PPD.

The first identified case involves the exact calculation of PMV, which can only be obtained if an instrument, such as the Testo 480 control unit, that returns all the necessary quantities is available. In (Figure 30) the trend of PMV for occupancy hours only and the percentage of dissatisfaction is depicted. From the processing of these data, it can be appreciated that for 74% values fall within −0.5 < PMV < 0.5 and for 25% this range is exceeded, but remains between −1 and + 1, still acceptable. As for the percentage of dissatisfied, on the other hand, it touches 15%.

Figure 30.

PMV e PPD evaluated with testo 480.

The second case involves an approximation of those variables that are more complex to calculate, and that commercial home automation sensors rarely measure, such as the mean radiant temperature and the air velocity. In rooms that do not have large glass surfaces, the difference between air temperature and mean radiant temperature is very small [11, 12]. Figure 31 shows that the difference between the two quantities rarely exceeds half a degree, and the mean radiant temperature disagrees by less than 3% with the air temperature (see Figure 32). It was thus thought to replace the average radiate temperature with the air temperature. Another quantity that is hardly measured in home or work environments is air velocity. Both mean radiant temperature and air velocity would also have to be measured at several points in the room, near each workstation, which is quite complex. Figure 33 shows that during the hours of occupancy, only in 0.07% of cases the air velocity exceeds 0.1 m/s, with an average of 0.05 m/s. In light of these considerations, the PMV slightly varies by replacing the average radiant temperature with the air temperature and placing the air velocity equal to 0.05 m/s (Figure 34).

Figure 31.

Mean radiant temperature and air temperature comparison.

Figure 32.

Evaluation of average radiant temperature/air temperature error.

Figure 33.

Air velocity during periods of occupancy.

Figure 34.

Approximate calculation of PMV and PPD.

Despite the approximations, (Figure 35) highlights that the error is minimal and R² is almost 1, so in rooms that do not have radiant floors/ceilings/walls and that do not have large windows, the reduction of variables is possible.

Figure 35.

Approximated vs. exact PMV.

7. Conclusions

Contemporary urban society often confines individuals to indoor spaces, whether residential or occupational, for significant portions of their lives. In recent years, there has been a growing interest in improving indoor environmental quality (IEQ), driven by heightened awareness of its impact on human health and productivity. This study leverages experimental data analysis to assess and identify key variables influencing comfort perception, aiming to streamline calculation methodologies. Special attention is directed toward the “Sesto Senso” sensor, for which a prototype was available. Evaluation of the sensor’s measurements revealed discrepancies, particularly in the algorithm calculating CO2 levels, traced back to issues with presence sensors and door/window opening/closing sensors. Repositioning and recalibration of these sensors were performed to address the discrepancies. Additionally, an upgrade to the Sesto Senso system was implemented, incorporating a predicted mean vote calculation for thermo-hygrometric comfort assessment, currently undergoing testing at the Casaccia center.