IT202100005918A1 - METHOD AND SYSTEM OF COOLING THROUGH VENTILATION OF A BUILDING - Google Patents

Description

Method and system of cooling by ventilation of a building

DESCRIPTION TECHNICAL FIELD

The present invention relates to a method for cooling by ventilation the internal environments of a building, for example internal environments used as residences or premises for use in the tertiary sector, schools and in general where the building is located. quality control applicable? of the internal atmosphere considering at the same time a low energy impact. In particular, cooling by ventilation is based on natural or mechanically forced circulation, e.g. through at least one fan, of air currents not previously treated through a heat exchanger and heat transfer fluid subjected to the action of a compressor. In ventilation cooling, the air introduced into the building is not treated in such a way as to reduce energy consumption and the environmental impact of the building.

STATE OF ART

AND? known to control the temperature inside a building by natural air currents, i.e. not cooled in an artificial way through a heat exchanger possibly combined with a refrigerating battery, arranging thermostats and a control unit in which temperature thresholds are memorized, once reached which the unit activates one or more? actuators for? opening of windows or the like, or one or more? fans in order to allow access of air from the external environment.

OBJECTS AND SUMMARY OF THE INVENTION

The purpose of the present invention ? to provide a cooling method by ventilation capable of considering a greater number of parameters to increase the living comfort of the building and, at the same time, have a relatively low environmental impact, and in particular energy consumption.

The purpose of the present invention ? achieved through a cooling method by ventilation of a building having openings for the access of air from the external environment, comprising the phases of:

- To acquire:

o data representative of the solar radiation of the building; or at least a data representative of an internal temperature of the building and of the external environment; And

- Using an electronic control unit on the basis of an algorithm that receives input data on irradiation and internal and external temperature, control at least one of the following cooling devices for ventilation:

o An actuator configured to open/close one of said openings via a movable element so that when the opening? open, an air flow passes between the eternal environment and the inside of the building;

o A fan configured to generate an air flow between the external environment and the inside of the building;

in combination with a device with artificially cooled air through a circuit, having previously acquired and sent to the electronic control unit an operating parameter of said device,

In order to obtain a target value, e.g. pre-set by a user, of an output parameter of the algorithm, the algorithm being such as to calculate a test value of the output parameter based on at least one pre-set actuation of the ventilation cooling device while the device at artificially cooled air through a circuit? worn out;

and obtaining said target value either by generating an adjustment command for the at least one ventilation cooling device while the artificially cooled air device ? off when the test value ? equal to or greater than the target value; or by generating an adjustment command for the at least one cooling device for ventilation and switching on the artificially cooled air device when the test value ? lower than the target value.

In this way, a combined control of the air inside the building ? possible at relatively low cost through the combination of a passive and natural system, e.g. for the exchange between inside and outside of the building of air at the temperature of the external environment through the spontaneous generation of air currents circulating in the building, and active, e.g. by means of a fan to generate a flow of untreated air or by means of an artificially cooled air device through a circuit with a heat transfer fluid and a compressor acting on the heat transfer fluid disposed in heat exchange with the cooled air. Also, temperature and actuators and/or fans are relatively inexpensive. Solar position data is very readily available and allows for subsequent estimations which improve the accuracy of the algorithm. AND? for example, it is possible to derive irradiation data, for example on the basis of a database of meteorological data of the area in which the building is located, therefore without the great use of measuring infrastructures. In particular, isn't it? strictly necessary to provide sensors applied to external structures of the building to directly or indirectly measure the radiation. The meteorological data can be purchased for subsequent processing at relatively low costs and, above all, they allow you to program the electronic control unit without setting up sensors on the building, which require maintenance as well as installation due to their exposure to atmospheric agents.

It should be noted that an indoor dry bulb temperature and an outdoor one collected from meteorological databases, e.g. by means of a meteorological data survey station close to the building, they implement a particularly simple and economical embodiment. AND? It is possible to use input and output data of the algorithm other types of temperature to increase accuracy, e.g. mean radiant temperature estimated using a globe thermometer, and subsequently processed to obtain an average operating temperature of the building, defined as the 24-hour average of the various operating temperatures of the building. The artificially cooled air device through a circuit, e.g. a traditional air conditioner operating through the change of phases of a refrigerant, pu? for example, be activated to counteract peak loads and/or solar radiation in the building. In off-peak conditions such a device ? switched off and the ventilation devices are active, e.g. the fan, which passive and natural, e.g. air currents generated by the opening of passages between the external and internal environment of the building as happens for example due to the chimney effect, and therefore reduce consumption.

Overall, therefore, ? possible to obtain a control for the purpose of cooling the temperature in the hot seasons with low energy consumption and an improvement in thermal comfort since? the target value is achieved through air flows exchanged with the external environment through relatively inexpensive and simple to install devices.

According to a preferred embodiment of the present invention, the step of checking ? such that the? algorithm ? predictive to calculate the test value as an estimate referred to the end of a time interval starting at the acquisition time of at least one of the input data, preferably the internal temperature, the electronic control unit being programmed to perform a plurality? of times said algorithm and, preferably, generate said signals during said time interval.

Through the use of a predictive model based on data relating to the sensors placed on the building and meteorological data that may be purchasable, e.g. via API (Application Program Interface) ? It is possible to obtain a particularly efficient and low energy impact control of the environmental comfort of the building according to an approach which, preferably, adapts to a configuration automatically adapted to the specifications of the same. Furthermore, the control can be considered open-loop since? the calculated value, e.g. the expected core temperature, is not compared to calculate an error. Through the repetitive execution of the forecast, an adaptive effect to the climatic conditions outside the building is obtained, e.g. of the environment exposed to atmospheric agents surrounding the building. Based on the embodiments, this environment surrounds the building within a few meters if the input data of the algorithm is detected by means of corresponding sensors installed on the building. Alternatively, for example when one or more? input data is collected via databases, e.g. meteorological databases, the external environment also extends for a few kilometers around the building.

According to a preferred embodiment, the method comprises a calibration or learning phase of the algorithm comprising the phases of:

- Apply to? Building one or more? known thermal loads;

- applying known operating parameters to said devices and at least one of the actuator and the fan;

- acquire the internal temperature of the building;

- Store representative data of the dynamic and/or hourly response of the building to the application of known thermal loads and known operating parameters; said algorithm being configured to calculate said output parameter on the basis of said stored data.

In this way, ? possible to avoid costly surveys, for example with reference to both the instrumentation to be used and the time necessary to carry out the measurements and to adapt the said mathematical model also to period buildings or in any case without detailed documentation. AND? it is also possible that the calibration or learning phase is performed in the presence of an existing building or that a mathematical model of the building is prepared and the thermal loads and operating parameters are input data from a numerical simulation based on the mathematical model of the building and the temperature is the result of this simulation and possibly recalibrate the model after a sufficient time for acquiring and storing data in the operating phase.

According to a preferred embodiment, the algorithm ? configured to receive in input, also dew point temperature, average air temperature, a representative parameter of the air change e.g. flow rate of air exchanged with the outside or air exchange rate ACH and humidity? relative; and the sun position data includes an azimuth angle and an elevation angle. For example, data relating to the solar position with respect to the building is present in libraries or tables such as Pysolar, Python.

Through these parameters alone ? possible to generate a particularly precise model for the control of the internal atmosphere of the building that also takes into account comfort parameters e.g. as defined by the standard EN 16798-1 of 2019 or ISO 7730. Furthermore, the sensors for the measurement are particularly cheap and/or the estimation algorithms are precise. Other advantages of the present invention are discussed in the description and referred to in the dependent claims.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention? described below on the basis of non-limiting examples illustrated by way of example in the following figures, which refer respectively to:

- Fig. 1 a schematic view of a cooling control system according to the present invention;

- Fig. 2 a schematic view of a flow diagram relating to the operation of the system, such as for example the system of figure 1; and - Fig. 3 a correlation matrix identifying the operating parameters of the system and related impact referred to an exemplary building. DETAILED DESCRIPTION OF THE INVENTION

In figure 1 ? illustrated with reference numeral 1 as a whole a system for controlling the internal atmosphere of a building by means of ventilation generally comprising at least a fan 2 for generating a flow of untreated air towards or from the building, an actuator 3 for opening a window or another passage of air between the outside and inside of the building in order to create an air flow e.g. due to the chimney effect with a significant impact for an individual inside the building, a sensor 4 of a physical parameter of the air inside the building, for example the temperature and advantageously also the humidity? and an electronic control unit 5 connected in data exchange with the fan 2, the actuator 3 and the sensor 4. The control unit 5 furthermore receives data representative of the radiation of the building. Such data can be received in the form of a database coming from a meteorological data source, for example a website S, and stored in a memory of the control unit 5. Alternatively, ? It is possible to place a sensor on the building and connect it to exchange data with the electronic control unit 5. Examples of sensors are a pyranometer near the main transparent surfaces of the building; or a solar meter. AND? it is also possible, through a mathematical model, to estimate the irradiation through the position of the sun, e.g. azimuth and altitude angle of the sun, which can? be calculated on the basis of the geographical location of the building and the time and weather indications, e.g. outside temperature, overcast sky etc. For example, the radiation estimate also includes, using geometric formulas, the effect of obstacles on the various windows for the various hours of the year in order to bring the direct and diffuse radiation values assumed by the weather station back to the position of the building more nearby.

The IoT system? preferably designed to be scalable, easily integrated with new components. For example, in a complex use case, one or more devices can also be connected in data exchange. of the following sensors/devices: a radiant heating element 6, for example electric, with a relative sensor of absorbed energy, for example a current sensor C, a device for artificially cooled air through a circuit 7, for example an air conditioning system of air comprising a heat transfer fluid circuit which changes state in the circuit and a compressor acting on the fluid for closed-cycle operation of the device, a sensor for measuring the energy consumption of the device 7, for example a current sensor if the compressor ? electric, a differential pressure sensor P to measure the pressure difference between the outside and inside of the opening controlled by the actuator 3 so as to estimate the flow of air passing through when the opening is? open. For example, there? can? be done using the formula:

where v ? the speed? of the air at the height of the window with the pressure coefficient pi? high; the two Cp are the pressure coefficients of the two openings; Cd are the discharge coefficients due to the openings and A are the net areas of the various openings through which the flow passes in sequence.

With reference to openings in which the air flow rate is mistaken for chimney effect, ? For example, you can use the formula:

Where g ? the acceleration of gravity?, H the height differential between two openings, the T in the numerator are internal and external temperatures in absolute value and that in the denominator ? the desired temperature; cd ? the average discharge coefficient due to the openings and A ? the average of the net areas of the two openings through which the flow passes in sequence.

The control unit 5 ? programmed to collect input data via sensors 4, ?C?, ?P? and the database S, using respective mathematical models to estimate additional parameters such as irradiation, if it has not been measured using a sensor installed on the building, based on the input data, and balance a ventilation cooling with a traditional mechanical one using the activation of at least one of the fan 2, the actuator 3, and the artificially cooled air device via circuit 7, e.g. an air conditioner, also in order to reduce energy consumption.

One or both of the above objectives, e.g. target value of temperature and quality? of the target air, are achieved through the control or balancing of the air flows exchanged between the inside and outside of the building through windows or other air passages of the building and the relative regulation of the fan(s) 2 and/ or of the actuator(s) 3.

According to an embodiment, a mathematical formula for calculating the air balance thanks to the flow rates calculable using equations 1 and 2 above and to the solar thermal input loads e.g. obtainable based on the position of the sun?:

differential Requested = Solar thermal loads / (air flow rate * air density * air thermal capacity) EQ. 3

By combining equations 1-3, ? possible to calculate a limit or test temperature obtained on the basis of a predefined setting of the ventilation cooling devices 2, 3 e.g. maximizing the exchange of air with the outside. In case the temperature is equal to or lower than an indoor temperature target value set by a user, the control box 5 keeps the air conditioner 5 off and operates the ventilation cooling devices 2, 3 in a pre-defined configuration e.g. the same considered to calculate the test value. Conversely, if the test value of the internal temperature ? lower than that set by the user, the control unit 5 operates the air conditioner 7 in a pre-set configuration e.g. at maximum and, at the same time, the ventilation device/s 2, 3 are set for the maximum exchanged air flow rate. By successive measurements of the internal temperature, the air conditioner 7 is switched off, e.g. when the indoor temperature target value is detected.

In order to increase the precision of the calculation, ? It is also possible to consider the internal thermal loads such as the heat emitted by machines or devices present in the building both through usage estimates and through localized current absorption meters e.g. the S sensors, the average number of people present in the building, in addition to the external radiation sensors. If so, a numerical example of some parameters of the above equations?:

eg. 20W/m<2 >(internal loads) and 30W/m<2 >(solar loads) considering an office with eight hours of occupation ?

density air = [1.2kg/m3]]; And

capacity? thermal air = [0.28W/kg]

It should be noted that the above calculations are performed on the basis of the measurement or estimation of a building?s thermal load. In the first case the sensors are dry bulb temperature, humidity, irradiation (or solar coordinates), (pressure differentials), in the second the estimation is performed through mathematical models capable of dynamically simulating the thermal behavior of the building. For example, climate data is collected and a model of the building's own structures, installations and devices impacting the internal temperature is generated via a program, e.g. EnergyPlus?, what? also configured to receive in the unit? measurement common climate data e.g. of temperature, humidity, radiation on the axes, zenith angle, so that these data are the inputs of the simulation models. In the initial setup (first network training(s) and initial inputs) ? possible to estimate reference values similarly to what would be done to make an energy diagnosis or an energy certification. In case there are measurements of transmittance values e.g. U-value ? possible to estimate the actual ones, otherwise we proceed as for certification referring to standard materials or even standard wall packages by type and year of construction (e.g. TABULA database for various EU countries, or those directly included in the energy analysis software). With regard to the efficiency of the building's systems, simplified models or formulas can be used, assuming the manufacturer's reference data from technical sheets (e.g. boiler booklets, etc.). It should be noted that, at least in part, the errors would still be compensated by the predictive model, in particular if re-configured on the basis of historical data as briefly mentioned below.

AND? also estimated as an input a value of occupancy, i.e. average number of people in each environment over time e.g. a static or dynamic model of occupancy, also taking into account what is suggested through specific functions in EnergyPlus? based on e.g. of the type of building. Via an additional simulation program, e.g. Python, do you run pi? simulations with the same input data indicated above to obtain temperature values as the air flow rate exchanged with the outside varies, e.g. ACH 10%, 20% etc. In particular, using Python, the simulations obtained present the temperature value inside the building at a future time interval with respect to that of the input climatic data, e.g. 1 hour. In this way, ? It is possible to collect data relating to the behavior of the building by varying the flow of air exchanged with the outside and, by changing the climatic data, also when the external conditions vary.

According to a preferred embodiment, the control unit 5 is programmed to implement a predictive control algorithm with learning, for example based on neural networks, deep learning, Kalman filter and the like, and in which the learning phase is such as to generate a short-term forecast through said mathematical model, e.g. 1 hour, of one or more? operating parameters of the system 1. According to one embodiment, the control unit ? programmed to implicitly learn the effects of natural ventilation on the specific building located and improve the potential? of exploitation of natural forces when present.

For example, the data obtained via the numerical simulation described above is processed, e.g. selected in terms of the greatest / least impact on the temperature and sorted by time to carry out a learning phase of a predictive algorithm. In particular, the learning of the predictive algorithm ? performed for such a number of epochs as to make it negligible, i.e. below a predefined threshold, an error parameter with respect to test data for which e.g. from the literature or from field measurement campaigns, both the input data and the corresponding temperature values. For example, the error parameter ? the mean squared error and ? it has been verified that after 30,000 epochs, intended as training cycles with input data and output data assigned and corresponding for the purpose of calculating the parameters of the hidden nodes, the mean squared deviation values between one epoch and the previous one are lower at 0.05.

According to the embodiment illustrated in Figure 2, historical data (time series) relating to the behavior of the building over time and on the basis of different input values are measured or estimated. For example, the building is monitored in order to establish how a target parameter that can be set by the user varies, such as the expected internal temperature and/or other thermo-hygrometric comfort parameters such as the internal operating temperature (i.e. based on the bulb temperature dry and average radiant temperature), on the basis of input parameters including data on solar radiation e.g. altitude angle and solar azimuth, and the current internal temperature. Optionally but preferably, other input parameters that improve the precision of the achievement of the objectives are: external temperature, dew, humidity? relative, measured exchanged air flow e.g. in ACH in a m<3>/s.

Alternatively, through mathematical models of dynamic energy simulation of buildings based on finite elements and / or computerized fluid dynamics (CFD) are performed more? simulations in which the input parameters mentioned above have different numerical values between the various simulations and the relative results are collected in relation to the observed parameters.

In this way ? generated through numerical simulation a set of data that relate a plurality? of set of data of entry with a relative set of results and the ci? constitutes a database for training the mathematical control model. In particular, through these data, a relationship is established between various quantities of air exchanged with the outside and a forecast of the internal temperature, i.e. target core temperature, within a predefined time frame, e.g. 1 hour.

Also, preferably, the predictive algorithm ? trained to take into account the dynamics of energy phenomena, receiving as training inputs sets of data relating to successive times in order to calculate a forecast.

For example, the predictive algorithm includes a ?feed-forward? having an entry level, a hidden level, an exit level, 21 nodes in the entry level, 43 in the hidden level and 1 node in the exit level, cio? the internal temperature of the building between one hour for 10 air exchange values ACH or other indicative parameter of the air flow exchanged with the outside and a sigmoid activation function for the input and hidden levels and a identity activation for the output level in order to avoid losing negative values of the target parameters. For example, a sub-zero internal temperature in the output must be predictable. If you use activation functions with Domain [0,+inf) you lose negative output.

The 21 nodes have been identified as a set of 7 parameters repeated in three instants of time prior to time 0 starting from which ? calculated the forecast, for example 1 hour from each other, e.g. for integer multiples of the forecast time interval, in this specific case 1 hour. Can you? express this relationship as follows:

The 7 identified parameters are preferably selected on the basis of the correlation matrix of figure 3 obtained through the numerical simulations of modeled and subsequently simulated buildings, e.g. via EnergyPlus? released through the U.S. Department of Energy's Building Technologies Office, similarly as described for collecting predictive algorithm training data. It should be noted that the results of Figure 3 can reasonably be extended to any type of building. In fact, during the learning phase, the specific structural and environmental characteristics of each building are included in the predictive algorithm.

On the basis of this matrix, 7 parameters are identified to fully describe the energy dynamics of the building, also including the parameters relating to the quality temperature, i.e.: dry bulb temperature, dew point temperature, azimuth angle of the sun, height angle of the sun above the horizon, mean air temperature, air exchange rate ACH and humidity? relative to air. These parameters are then subdivided, for example through cost criteria, among those that can be measured directly, e.g. through a dedicated sensor and those that can be estimated through mathematical models, data acquisition from providers, etc.

After training the neural network for predictive purposes, when System 1 is operating for control purposes, a user establishes a target internal building temperature. The predictive algorithm is subsequently applied by the control unit 5 indicating, among the input parameters, in some cases, e.g. internal temperature of dry bulb, values measured in real time, in other estimated forecast values or acquired through meteorological data provided, such as for example in the case of? humidity? relative, dew point temperature, average hourly dry bulb temperature, and solar parameters which can be calculated from tables on the basis of the time (azimuth and angle of altitude of the sun). With reference to the flow of air exchanged, e.g. the air exchange rate ACH, is instead assigned a group of progressively increasing values in order to obtain as many short-term forecasts of internal temperature. For example, the following formulas are implemented:

Each of said increasing values corresponds to a specific setting of the fan 2 and actuator 3 to achieve this air exchange rate. It should be noted that the correspondence between the setting of the ventilation devices 2, 3 and the flow rate of air exchanged with the outside is established i.e. ?learned? during the learning phase of the predictive algorithm. The control unit 5 ? it is also programmed to automatically select the ACH value, which impacts, as described in the previous paragraphs, on the operation of fan 2, actuator 3, etc. whose temperature forecast pi? approaches the user-selected target value. However, if the maximum flow rate value exchanged in the forecast time fails to satisfy the desired temperature assigned by the user, the control unit commands the switching on of the air conditioner 7. For example, the air conditioner 7 is switched on and set to the target value of the temperature: on the basis of the successive predictions it will be able? subsequently be switched off e.g. following a sudden drop in outside temperature e.g. for the approach of a summer storm, the forecast of the internal temperature reaches the target value.

For example, once the expected temperature has been assigned by the user, the predictive algorithm is launched with a greater frequency than the unitary frequency in the forecast time interval: for example, when the forecast interval ? of 1 hour, the algorithm is performed once a minute or every 10 minutes with input data updated in real time. In this way, since the expected temperature value hopefully remains constant for a time comparable to if not longer than the forecast time interval, an effect is obtained on the regulation of the fan 2, the actuator 3 and possibly the air conditioner analogous to that of a feedback control since, in particular if the expected temperature remains unchanged for at least the duration of the forecast interval, the relative value converges. Furthermore, the predictive algorithm reacts according to the difference between the expected temperature and the measured temperature, therefore as the measured temperature approaches the expected one, the impact of the regulation gradually becomes more effective. Sweet.

The data measured during operation of the method in the building, e.g. evolution over time of the internal temperature based on climatic conditions, can be used to cyclically calibrate the predictive model i.e. the input data and the corresponding internal temperature after 1h are provided in order to update the internal parameters of the predictive algorithm based on the history, e.g. in the case of the neural network the coefficients or weights of the hidden nodes. In particular, thanks to the historical data stored during use and used to update the predictive algorithm through a subsequent learning phase which, previously, e.g. 1 year earlier, had been performed on the basis of simulated data, the margin of error introduced by the simulation models is reduced by adapting the algorithm to the actual performance and characteristics of the building.

Finally, it is clear how the system described and illustrated here? It is possible to apply modifications or variations without thereby leaving the scope of protection as defined by the attached claims.

Claims (8)

1. Cooling method by ventilation of a building having openings for the access of air from the external environment, including the phases of: - To acquire: o data on the solar radiation of the building; o data representative of an internal and external temperature of the building, the external temperature being related to the environment subject to atmospheric agents in an area surrounding the building; or a target value of an operating parameter of the building (Indoor Temperature) - Calculate through an electronic control unit on the basis of an algorithm receiving input data on irradiation and internal and external temperature, a test value of said operating parameter - Check at least one of the following ventilation cooling devices: o An actuator (3) configured to open/close one of said openings via a mobile element so that when the opening is? open, an air flow passes between the eternal environment and the inside of the building; o A fan (2) configured to generate an air flow between the external environment and the inside of the building; in combination with an artificially cooled air device (7) via circuit with phase-changing liquid; to obtain said target value or by generating a control signal of the at least one ventilation cooling device (2, 3) while the artificially cooled air device (7) ? off when the test value ? equal to or less than the target value; or by adjusting the at least one ventilative cooling device (2, 3) and turning on the artificially cooled air device (7) when the test value is ? higher than the target value. The method according to claim 1, wherein said algorithm calculates said operating parameter test value based on a pre-defined test drive of the ventilative cooler (2, 3) while the artificially cooled air device by circuit ? worn out. 3. Method according to any one of the preceding claims, wherein the acquiring step comprises the step of acquiring said internal temperature from a sensor of the building (4) and acquiring from one or more temperature sensors (4). databases the said external temperature and the data representative of the solar radiation of the building. 4. Method according to any one of the preceding claims, wherein the algorithm ? predictive algorithm for calculating the test value as an estimate referring to the end of a time interval starting at the acquisition time of at least one of the input data, the electronic control unit (5) being programmed to execute said algorithm at a frequency equal to at least twice with respect to the time interval. 5. Method according to claim 4, wherein the algorithm calculates a plurality? of test values for corresponding different air exchange flow rates starting from the same input data, said ventilation cooling device (2, 3) being adjusted to obtain the air exchange value corresponding to the test value pi ? close to the target value. 6. Method according to one of claims 4 or 5, wherein the predictive algorithm ? a feed forward neural network. The method according to one of claims 4 to 6, wherein the input data further comprises at least one of a dew point temperature, an average air temperature, a representative parameter of the air change and humidity. relative. 8. Ventilation cooling system of a building having openings for the access of air from the external environment, comprising: at least one ventilative cooling device (2, 3); at least one sensor (4) for measuring an internal temperature of the building; And an electronic control unit (5) programmed for: - To acquire: o data on the solar radiation of the building; or data representative of an internal temperature from said sensor (4) and external of the building, the external temperature being related to the environment subject to atmospheric agents in an area surrounding the building; or a target value of an operating parameter of the building (Indoor Temperature) - Calculate on the basis of an algorithm receiving input data on irradiation and internal and external temperature, a test value of said operating parameter - Check at least one of the following ventilation cooling devices: o An actuator (3) configured to open/close one of said openings via a mobile element so that when the opening is? open, an air flow passes between the eternal environment and the inside of the building; o A fan (2) configured to generate an air flow between the external environment and the inside of the building; in combination with an artificially cooled air device (7) via circuit with phase-changing liquid; to obtain said target value or by generating a control signal of the at least one ventilation cooling device (2, 3) while the artificially cooled air device (7) ? off when the test value ? equal to or less than the target value; or by adjusting the at least one cooling device for ventilation (2, 3) and turning on the artificially cooled air device (7) when the test value is ? higher than the target value.