Théry-type heat ux sensors
The measurement of this type of sensors is based on the Seebeck eect. Several and very small thermocouple junctions are connected in series, i.e., as a thermopile, distributed all over the surface of an insulated support (based on the printed circuit board technique). This support is then covered by two plates of copper, on each side. When both plates of this embedded element are submitted to dierent temperatures, each of the thermocouple junctions generates a potential dierence due to the temperature gradient. This potential dierence (U) is proportional to the heat ux traversing the sensor (‘), and they are related by a thermoelectric coecient [23, 99, 110] (cf. g. 1.26). These sensors are also able to measure the temperature of the surface where they are being placed.
The thermoelectric modules, also known as Peltier modules, are composed of several thermocouple junctions connected electrically in series and thermally in parallel, integrated between two ceramic plates  (cf. g. 1.27). These thermocouple junctions consist of a n- and a p-type semiconductor materials connected by small and thin copper tabs. These modules are widely available in the market, normally used in electronics for cooling purposes, where the most common semiconductor materials employed is a quaternary alloy of bismuth, tellurium, selenium, and antimony, e.g., Bi2Te3. In the case of cooling applications when higher performance is needed, multistage Peltier modules can be found in the market (cf. g. 1.27 (b)).
However, these modules can also be used for power generation . For heat ux measurement or heat ux detection, they work in an open circuit conguration when connected to a multimeter or an acquisition system for data collection . In this case, they follow a similar relation between the potential dierence generated when a heat ux is traversing the module (cf. eq. 1.16). Even though they may have the particularity of aK coecient dependent on temperature due to the semiconductor materials of the thermocouples junctions, and the time response is about one minute. The dimensions availability for this type of module is limited, but for the same dimensions a TEM presents a sensibility value of about a hundred times larger than the sensibility value of a conventional heat ux meter; their cost lays around the twentieth of euros.
Implementation of heat ux meters for measuring the supercial heat exchanges
The measurement of supercial heat exchanges, i.e., convection and radiation, using FGT sensors
and thermoelectric modules, have been a topic of interest of various researchers [103, 51, 66, 30, 29, 23, 110, 10, 113]. The interest lays in the possibility of splitting the convection and radiation parts from the heat ux measurement. This brings us to the question: can convection and radiation heat exchanges be separated from a heat ux measurement in a mixed environment? Thus, this review is focused on the implementation of such sensors for the estimation of both supercial heat exchanges; any other case was excluded. However, other research works have been found regarding the implementation of such sensors in the thermal characterization of walls, among which we have [65, 69, 103].
In the late 80s, a technique to estimate the convection and radiation heat exchanges on a surface was implemented, consisting of using two FGT sensors, where one was to be coated with a black surface and the other with a shiny surface. Then, under the premise that the black and shiny surfaces had emissivities close to 1 and 0 respectively, the former was said to estimate the total heat ux (convection + radiation) and the latter to estimate the convection heat ux on the rigid surface where they installed the sensors, e.g., heavyweight and lightweight walls, and isolate-type wall. Herin in 1988  worked on the design and characterization of convective and radiative heat ux meters applied to in situ measurements. The heat ux meters (25 x 25 x 0,02 cm and 30 Vm2W1), coated with foils with contrasted emissivities, were tested with the objective of separating the convective and radiative components when installed inside a climatic chamber (cf. g. 1.28). The heat ux and temperature measurements were validated, allowing the separation of the convective and radiative components on heavyweight and lightweight walls, and also the determination of the thermophysical properties of such walls.
Analytical approaches for splitting the convective and radiative heat exchanges
Cherif et al.  proposed a way to separate the convective and radiative parts from the heat ux measurement. The principle is based on a heat energy balance on the surface of the heat ux sensor, well-placed onto a wall surface, which is submitted to convection (‘C) and radiation (‘R) heat exchanges.
The total heat ux (‘T) on the sensor’s surface is equal to the sum of ‘C and ‘R. Also, this is equal to the heat ux traversing the sensor (‘). This before could be expressed for any surface i as follows: ‘i(t) = ‘Ci (t) +’Ri (t) h Wm2 i (1.18).
It was stated that the heat ux measurement, on both sensors (black and shiny), would have a convection and radiation part, owing to the emissivity value of the black and shiny coating they used: a black paint (0,98) and a thin aluminum foil (0,1), respectively. Thus, the convection heat ux was determined by subtracting the radiation part from the shiny sensor measurement. This radiation part was computed using the “classical radiosity method” and also by numerical simulation on the Fluent software.
Going further on the emissivity issue, they analyzed numerically if the results depended on the emissivity value of the black coating. Changing the emissivity value from 0,9 to 1 (a perfect black body), it was found that increasing the emissivity yield to an increase in the total heat ux. Finally, it was concluded that the experimental procedure allows uncoupling the convection and radiation parts from the measurement since the relative error obtained was 5%maximum between the experimentation and numerical results. It is worth mentioning that the heat ux levels in this experiment reached up to 350 Wm2, as it will inuence later the permissible error range for the present work.
Perturbations and uncertainties introduced by the heat ux meter
Herin  compared its measurements with simulation results for two dierent cases, to estimate the errors in steady state: (i) the heat ux sensors (25 x 25 x 0,02 cm and 35 Vm2W1) placed on a vertical concrete wall (thermally heavyweight), (ii) the heat ux sensors placed on a polystyrene wall (thermal isolation). For the rst case, the convection heat ux was encountered to be overestimated by near 26% and the radiation heat ux to be underestimated by practically 35% (cf. g. 1.37 (a)). For the second case, the error on the radiation heat ux was around the -79% and on the convection heat ux was between 250 and 350% (cf. g. 1.37 (b)). For the case of isolate wall types, it was concluded that the estimation of the convective and radiative components is complexed and a temperature correction under the sensors is needed (in the sensor-wall interface, T0 s ), where errors were said to be mainly caused by the local temperature modications on the wall surface, due to the presence of the sensors, which also modies the heat ow distribution through the wall thickness (cf. gs. 1.37 (b) and 1.38); this was also pointed out in  and .
Table of contents :
1 Theoretical background and state of the art
1.1 Natural ventilation in buildings
1.1.1 The role of natural ventilation
1.1.2 The natural ventilation concept
1.1.3 Assessment techniques for the natural ventilation airow rate
1.2 Thermal behavior modeling and identication approaches
1.2.1 Thermal behavior of a passive building: the role of the envelope
1.2.2 Modeling approaches for forecasting in buildings
1.3 Coupling between natural ventilation and thermal mass
1.4 Heat ux measurement techniques for rigid surfaces
1.4.1 Principles of the heat ux measurement
1.4.2 Implementation of heat ux meters for measuring the supercial heat exchanges 31
1.4.3 Perturbations and uncertainties introduced by the heat ux meter
1.4.4 Calibration techniques for at-plate type of heat ux sensors
1.5 Concluding remarks and perspectives
2 Characterization of the coupling of the energy charge-discharge process with natural ventilation
2.1 Description of the experimental platform
2.1.1 Passive strategies for the charge-discharge process in summer
2.1.2 Natural ventilation system and openings
2.2 Experimental approach and protocol
2.2.1 Instrumentation of the thermal mass domain
2.2.2 Instrumentation of the building
2.2.3 Measuring system and calibration
2.3 Semi-empirical models for data processing
2.3.1 Decoupling the convection and radiation heat exchanges
2.3.2 Mean radiant temperature of surrounding surfaces
2.4 Experimental plan implemented in the measurement campaigns
2.5 Results and discussion on the coupling between the energy charge-discharge and natural ventilation .
2.5.1 Classication of the experimental data
2.5.2 Thermal behavior of the platform under dierent opening congurations and outdoor conditions
2.5.3 Coupling of the charge-discharge process with natural ventilation within the indoor environment
2.6 Characterization of the energy charge and discharge processes
2.6.1 Impact of the outdoor environment upon the indoor environment via correlation analysis .
2.6.2 Identication of the heat transfer phenomena involved
2.7 Thermal comfort assessment of a simple opening-closing strategy
2.7.1 Indicators employed in the assessment
2.7.2 Impact of a passive night ventilation strategy on internal thermal comfort
2.8 Concluding remarks
3 Characterization of the natural ventilation airow rate
3.1 Initial considerations and airow pattern recognition
3.2 Modeling of the natural airow rate in the platform
3.3 Airow characteristics of the openings
3.3.1 Recovery of the envelope tightness and airtightness tests
3.3.2 Discharge coecient of the openings
3.3.3 Identication of C and n for each opening
3.4 Results and analysis
3.4.1 Airtightness tests and discharge coecient identication
3.4.2 Comparison of the air change rate per hour using airtightness tests
3.4.3 Airow simulation results
3.5 Concluding remarks on the airow characteristics of the platform and perspectives
4 Identication of a thermal model to describe the behavior of the platform
4.1 Identication of a thermal behavior model for the indoor air using experimental data .
4.1.1 Energy balance for the indoor air
4.1.2 Estimation of the natural airow rate via heat ux and temperature measurements
4.1.3 Identication of Model A for all opening conguration
4.2 Choice of a thermal behavior model for the thermal mass
4.2.1 Lumped-capacitance approximation
4.2.2 Spatial discretization via nite dierences
4.2.3 Choice of a Model B for all opening conguration
4.3 A complete model for describing the coupling between the energy charge-discharge of the thermal mass and natural ventilation
4.3.1 Evaluation of the model with data inside the identication process
4.3.2 Evaluation of the model with data outside the identication process
4.4 Concluding remarks
General conclusion and future work