Method for monitoring the fluid consumption of a building to be monitored by classifying fluid consumption events by means of supervised learning

Abstract

The invention relates to a method for monitoring the consumption of a fluid of interest by a building to be monitored by classifying, by means of supervised learning, fluid consumption events from among multiple classes of consuming elements. The classification is carried out using a prediction model that has been configured from a database. The classification comprises, for each of the identified fluid events: event parameters PEv(i); time parameters PT(i); and the corresponding class LEV(i) of consuming element EC.

Description

[0001] METHOD FOR MONITORING FLUID CONSUMPTION OF A BUILDING TO BE MONITORED, BY SUPERVISED LEARNING CLASSIFICATION OF FLUID CONSUMPTION EVENTS TECHNICAL FIELD [1] The field of the invention is that of monitoring the fluid consumption of a building to be monitored, for example the consumption of liquid water by a residential house, a building, a school, a hospital, a factory, etc. [2] The invention relates more specifically to monitoring by classification, here by supervised learning, of fluid consumption events, present in a measurement signal representative of the temporal evolution of the overall consumption flow rate of the fluid of interest by the building to be monitored, and in particular by different consuming elements, such as for example taps, showers, flush toilets in the case of a residential house.[3] The classification amounts to determining the class (label, category, or label in English) of consumer element which is to be associated with each of the fluid events. The invention makes it possible in particular to detect excessive consumption of consumer elements of one or other of the predefined classes, and then makes it possible to indicate to the user if consumer elements are overused or have a flow rate above a predefined reference. STATE OF THE PRIOR ART [4] A building to be monitored can be a house, a building, a school, a hospital, a factory, a business hotel, etc. The fluid of interest can be, depending on the type of building to be monitored, a gas or a liquid such as water, hydrogen, oxygen, nitrogen, etc. [5] The building to be monitored comprises several devices connected by a fluid circuit to a fluid source and which consume the fluid of interest.These devices will be referred to here as "consumer elements". Consuming elements can be classified into different predefined classes. For example, in the case of a residential house where the fluid of interest is liquid water, the consuming elements may fall into the class of taps, showers, toilet flushers, dishwashers, or even washing machines. [6] The building to be monitored is usually equipped with a flow sensor adapted to measure a signal representative of the temporal evolution of the overall flow rate of the consumption of the fluid of interest by the building to be monitored, and more precisely by its consuming elements. This measurement signal includes fluid consumption events, which are defined as being moments of non-zero flow rate.[7] There is then a need to be able to determine with improved precision the class of consumer elements to be associated with each of the fluid events linked to the consumption of the fluid of interest by the building to be monitored. This then makes it possible to monitor the consumption of the fluid of interest by the building to be monitored, and to be able to facilitate the detection, for example, of unusual fluid consumption (excessive or insufficient) of such or such class of consumer elements. DISCLOSURE OF THE INVENTION [8] The invention aims to propose a method which makes it possible to classify with improved precision the fluid events linked to the consumption of the fluid of interest by the consumer elements of the building to be monitored, that is to say to determine the class of consumer elements to be associated with each of the fluid events, in order to monitor the fluid consumption of the building to be monitored.[9] For this, the object of the invention is a method, implemented by computer, for monitoring the consumption of a fluid of interest by a building to be monitored BS, for classifying fluid events. ^ Ev (^,^) consumption of the fluid of interest by EC consumer elements of the building to be monitored BS, among several predefined classes L ^ ^^(^,^) of EC consumer elements, the method comprising the following phases.

[0010] First of all, a phase of generating at least one database (BD1, BD2), comprising the following steps: - acquiring measurement signals ^Dmes ^^^^(^) ^ ^ (t)^ ^^^;^ representative of the consumption flow rate of each of the EC consumer elements of N real reference buildings {Bref (n)} n=1 ;N, with N>1, the measurement signals then being said to be classified, the EC consumer elements being of the same classes as those of the building to be monitored BS; - define consumption profiles of M so-called simulated buildings {Bsim (m)} m=1 ;M , with M>N, from the acquired classified measurement signals, each simulated building comprising EC consumer elements of the same classes as those of the building to be monitored BS; - generate, by digital simulation, simulated signals ^Dsim ^!"#($) ^ ^ (t) representative of the consumption flow rate of each of the EC consumer elements of the simulated buildings {Bsim (m)} m=1 ;Mhaving the defined consumption profiles, the generated simulated signals then being said to be classified; - determining, for each of the fluid events {Ev(i)}i=1;NA identified in the generated classified simulated signals: so-called event parameters {PEv(i)}i=1;NA, representative of a duration and an elapsed volume of each of the fluid events; and so-called temporality parameters {PT(i)}i=1;NA, representative of a number of so-called similar fluid events, having a flow rate substantially identical to that of the fluid event considered and located in predefined successive time slots located before and after the fluid event considered; - generate the database comprising, for each of the fluidic events {Ev(i)}i=1;NA identified: the event parameters {PEv(i)}i=1;NA, the temporality parameters {PT(i)}i=1;NA, and the corresponding class {LEV(i)}i=1;NA of consumer element EC.

[0011] Then, a phase of parameterization of at least one prediction model by supervised automatic learning from the database.

[0012] Finally, a classification phase, comprising the following steps: - acquire a measurement signal S. ^ my ^( (t) representative of the overall consumption flow rate of the building to be monitored; - determine, for each of the fluid events { ^ Ev (^,^)} j=1 ;NP identified in the acquired measurement signal, event parameters {P ^ ^^(^,^)} j=1 ;NP and temporality parameters {P ^ )(^,^)}j=1 ; corresponding NP ; - predict, by the prediction model, for each of the fluidic events { ^ Ev (^,^)}j=1 ;NP whose event parameters {P ^ ^^(^,^)}j=1 ;NP and the temporal parameters { ^ P )(^,^)}j=1 ;NP form input data to the prediction model, the corresponding class {L^ ^^(^,^)}j=1 ;NP of EC consumer element.

[0013] The acquisition step during the generation phase is carried out by means of several fluidic measurement sensors each connected to an EC consumer element of the N real reference buildings. Furthermore, the acquisition step during the classification phase is carried out by means of at least one fluidic measurement sensor connected to a fluidic inlet of the building to be monitored, and preferably by means of a single sensor.

[0014] The monitoring method comprises a phase of monitoring the building to be monitored, comprising the following steps: identifying a class of consumer elements, from at least a portion of the predicted classes, for which a monitoring parameter associated with the corresponding fluidic events has a deviation from a reference value greater than a threshold deviation; then communicating to a user the identified class of consumer elements.

[0015] Some preferred but non-limiting aspects of this monitoring method are as follows.

[0016] The temporal parameters P. T(i) and P ^ )(^,^) may also include, for each of the time slots, the cumulative volume of the fluid of interest flowed and the flow duration.

[0017] The event parameters P Ev(i) And ^ P ^*(^,^) may include, for each fluid event, the initial instant of the fluid event, the flow duration, the elapsed volume and / or the average flow rate.

[0018] The database and the input data of the prediction model may include so-called comparison parameters, noted respectively P Ev|Bsim(i) and P ^ ^*|^((^,^"→^)), and defined, for each fluidic event Ev (i) , E ^ v (^,^) , as the ratio of an average flow rate of the fluid event considered to an average overall flow rate of all fluid events Ev (i) , E ^v (^,^) of the building considered.

[0019] The classification phase can be repeated, the measurement signals S ^ my ^( (t) of the building to be monitored being then acquired over successive predefined monitoring durations.

[0020] In the comparison parameters P ^ ^*|^((^,^"→^)present in the input data of the prediction model, the average global flow rate of all fluid events ^ Ev (^,^)can take into account the average overall flow rate of the previous iterations.

[0021] The generation phase can comprise the following steps: - determination of a first database where each fluidic event is defined by at least said event parameters and by said temporality parameters; - determination of a second database where each fluidic event is defined by at least said event parameters, by said temporality parameters, and by so-called comparison parameters by class defined, for each fluidic event of at least one class considered, as the ratio of an average flow rate of the fluidic event considered to an average overall flow rate of all the fluidic events of the class considered.

[0022] The parameterization phase can then comprise the following steps: - parameterization of a first prediction mode by supervised automatic learning from the first database; - parameterization of a second prediction mode by supervised automatic learning from the second database;

[0023] The classification phase can then comprise the following steps: - determination, for fluid events present in the measurement signal S. ^ my ^((t), event parameters and corresponding temporality parameters; then - prediction by the first prediction model, for each of the fluidic events, which are defined by the event parameters and the temporality parameters which form input data of the first prediction model, of the corresponding consumer element class; then - determination, from the previously predicted classes, of corresponding comparison parameters by class; then - prediction by the second prediction model, for each of the fluidic events, which are defined by the event parameters, the temporality parameters, and the comparison parameters by class which form input data of the second prediction model, of the corresponding consumer element class.

[0024] During the step of acquiring the measurement signals ^Dmes ^^^^(^) ^ ^ (t)^ ^^^;^, the N real reference buildings can be connected to a device for determining at least one prediction model, comprising fluid measurement sensors adapted to acquire the measurement signals for each of the consumer elements of the real reference building considered, and a computer connected to the fluid measurement sensors of the reference building considered and adapted to carry out the parameterization phase.

[0025] During the step of acquiring the measurement signal S ^ my ^( (t) of the building to be monitored, it can be connected to a classification device comprising a fluidic measurement sensor adapted to acquire the measurement signal S ^ my ^((t), and a computer connected to the fluid measurement sensor of the building to be monitored and adapted to carry out the classification phase.

[0026] During the monitoring phase, following the communication step, the method may comprise a correction step by the user so that the consuming elements of the identified class then have a value of the monitoring parameter having a deviation less than the threshold deviation. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Other aspects, aims, advantages and characteristics of the invention will appear better on reading the following detailed description of preferred embodiments thereof, given by way of non-limiting example, and with reference to the appended drawings in which: FIG. 1A schematically illustrates an example of a building to be monitored, here a residential house,equipped with a device for classifying fluid events linked to the consumption of the fluid of interest by the building to be monitored; Figure 1B illustrates an example of a measurement signal representative of the temporal evolution of the overall consumption flow rate of the fluid of interest by the building to be monitored in Fig. 1A; Figure 2 illustrates a flowchart of a method for monitoring fluid events linked to the consumption of the fluid of interest by the building to be monitored, according to one embodiment; Figure 3A schematically illustrates an example of a real reference building, here a residential house,equipped with a device for determining a prediction model by supervised learning; Figure 3B illustrates examples of measurement signals representative of the local consumption flow rate by each of the EC consumer elements of the real reference building of Fig. 3A; Figure 3C illustrates time slots located before and after the initial instant of a fluidic event, making it possible to define so-called temporality parameters representative of each of the fluidic events; Figure 4 illustrates steps implemented during the database generation phase, from simulated signals for so-called simulated buildings; Figure 5 illustrates a flowchart of a method for monitoring fluidic events linked to the consumption of the fluid of interest by the building to be monitored, according to another embodiment which implements two prediction models; Figures 6A to 6C illustrate examples of cumulative distributions of the flow rate,of the flow duration and the elapsed volume, associated with the fluid events from simulated signals for a simulated building, and compared to those from a real building; Figures 7A to 7C respectively illustrate an example of the precision rate, recall rate and F1-score, associated with the classification of the fluid events of a building to be monitored carried out by classification methods according to different embodiments. DETAILED DESCRIPTION OF PARTICULAR EMBODIMENTS

[0028] In the figures and in the remainder of the description, the same references represent identical or similar elements. In addition, the different elements are not shown to scale so as to favor the clarity of the figures. Furthermore, the different embodiments and variants are not mutually exclusive and can be combined with each other. Unless otherwise indicated, the terms “substantially”, “approximately”,"of the order of" means to within 10%, and preferably to within 5%. Furthermore, the terms "between ... and ..." and equivalents mean that the terminals are included, unless otherwise stated.

[0029] The invention relates to the monitoring of fluid consumption within a building which consumes a fluid of interest by means of EC consumer elements of different classes L, Ev . It concerns more precisely the determination of class L Ev(label or tag) of consumer elements EC which is to be associated with each of the fluid events Ev linked to the consumption of the fluid of interest by the consumer elements EC of the building to be monitored BS.

[0030] The monitoring method is carried out using a computer (processing unit) which integrates at least one prediction model f1, f2 by supervised machine learning. The prediction model f1, f2 can be a decision tree model, weak classifier boosting, linear or quadratic discriminant analysis, decision tree forest, k nearest neighbor method, among others.

[0031] As detailed below, to obtain improved prediction performance, the monitoring method comprises a phase of learning the prediction model f1, f2 which is carried out from at least one database BD1, BD2 preprocessed in a particular manner.

[0032] This database BD1, BD2 is not formed, as it could be according to a natural approach, by classified (labeled) flow measurement signals coming from each of the EC consumer elements of several real reference buildings Bref. Indeed, this approach cannot be sufficiently representative of the diversity of the actual fluid consumption in real buildings. On the contrary, according to the invention, the database is formed from simulated and classified signals, representative of the flow rate of each of the EC consumer elements of simulated buildings Bsim, and preferably of a number M of simulated buildings Bsim much greater than the number N of real reference buildings Bref. The consumption profiles Pstat. Bsimsimulated buildings, from which the numerical simulation was carried out, were determined from the classified flow measurement signals coming from the real reference buildings, so that the simulated signals remain consistent with the actual consumption of the real buildings. Thus, it appears that the prediction quality is greatly improved.

[0033] In addition, the database is formed from the different fluidic events Ev of the fluidic consumption of the consumer elements EC of the simulated buildings. More precisely, it is formed from their representative parameters, including: - so-called event parameters P Ev, representative of the event itself (flow duration, elapsed volume or equivalent (average flow rate for example), or even initial time, etc.), and - so-called temporality parameters PT which indicate, for each of the fluid events Ev, the number of similar fluid events (in terms in particular of average flow rate, or even flow duration), included in predefined time slots and located before and after the fluid event considered.

[0034] It appears that learning the prediction model from such a database improves the prediction performance, in particular its precision, its recall, and therefore also its F1-score (harmonic mean of the precision and recall).

[0035] Figure 1A illustrates an example of a building to be monitored BS comprising several consumer elements EC of different classes, and equipped with a classification device adapted to classify the fluid events E ^v linked to the consumption of the fluid of interest by the EC consumer elements, i.e. adapted to determine class L ^ ^*of each of the fluidic consumption events by the building to be monitored BS.

[0036] The building to be monitored BS is one or more structures which consume a fluid of interest via its consumer elements EC, and whose consumption is to be monitored in order to detect possible anomalies such as excessive consumption by the consumer elements EC of a certain class. Thus, the user is able to know the consumption of each class of consumer elements EC, and can thus know, for example, if the average flow rate of a class of consumer elements EC is greater than a predefined value. By consumption, it is meant that the building to be monitored BS receives the fluid of interest from a fluidic source SF, and uses it (“consumes” it) according to different uses, which may be personal and / or professional.Such a building to be monitored BS may be, for example, a house, an apartment, a building comprising several apartments, a school, a factory, a hospital, a campsite, or other, and generally speaking, is any type of structure or assembly for personal and / or professional use. The building to be monitored BS may comprise several separate buildings, as for example in the case of a factory.

[0037] The fluid of interest may be a liquid or a gas, such as water, hydrogen, oxygen, nitrogen, helium, etc. Thus, the building to be monitored BS may be a factory which consumes for example hydrogen, oxygen or nitrogen, in liquid or gaseous phase. It may also be a training building (school) which uses liquid water. In the remainder of the description, the building to be monitored BS is a residential house, and the fluid of interest is liquid water.

[0038] The building to be monitored BS includes several consumer elements EC which ensure the effective consumption of the fluid of interest. The consumer elements EC are of different L classes, for example a class R for the taps (toilets, kitchen, bathroom, etc.), a class D for the showers, a class C for the toilet flushes, another class LV for the dishwashers, and another class LL for the washing machines. In this example, the classes are noted: L = {R, D, C, LV, LL}.

[0039] The building to be monitored BS also includes a fluid circuit which ensures the distribution of the fluid of interest from the fluid source SF to the consumer elements EC. These are distribution conduits, possibly equipped with valves. The fluid source SF of the fluid of interest can be a supply network, for example from the city, a reservoir, or other.

[0040] The building to be monitored BS is equipped with a classification device, which includes a fluidic measurement sensor CM. BS to acquire a measurement signal S ^ my ^( (t) representative of the temporal evolution of the overall flow rate, at the entrance to the building to be monitored BS. The classification device also includes a processing unit UT (computer) to determine the classes L ^ ^^ fluidic events ^ Ev linked to the consumption of the fluid of interest by the consumer elements EC and present in the measurement signal S ^ my ^((t). The processing unit UT integrates at least one prediction model f1 by classification whose parameterization was carried out by supervised automatic learning from a predefined database BD1.

[0041] Note that the classification device is not able to know the temporal evolution of the flow rate of each of the consumer elements EC, but only to know the overall flow rate at the entrance to the building to be monitored BS. It is because of this absence of sensors dedicated to each of the consumer elements EC, and therefore the fact that only the overall flow rate information is accessible, that the invention provides for using a prediction model, here by supervised learning, to estimate the class L ^ ^^ of consumer element EC which is to be associated with such and such fluidic event E ^ v of overall flow rate.

[0042] The CM fluidic measurement sensor BSis connected to the fluid circuit, and is located between the fluid source SF and the consumer elements EC. It is suitable for measuring a measurement signal S ^ my ^( (t) representative of the overall consumption flow rate of the fluid of interest by the building to be monitored BS. This CM fluid measurement sensor BS may be a flow meter water meter, a pulse water meter, among others.

[0043] In the remainder of the description, the CMBS fluid measurement sensor is considered to be a turbine flow meter water meter, which acquires a measurement signal S ^ my ^( (t), whose data are representative of the different fluid events ^ Ev. A fluid event is defined as a non-zero flow duration temporally delimited by zero flow durations.

[0044] Figure 1B illustrates an example of a measurement signal S ^ my ^((t), here in the form of a temporal evolution of the overall flow rate linked to the water consumption of a house to be monitored BS. The measurement signal S ^ my ^( (t) presents several fluidic events ^ Ev (^) , where j is an increment ranging from 1 to NP, with NP the total number of fluid events E ^ v over a monitoring period Δts (for example 1 day). The measurement signal S ^ my ^( (t) can be presented as a succession of electronic messages relating to each of the fluidic events: S ^ my ^(( t ) = . ^ P ^*(^) / ^^^,^0 . Each fluidic event E ^ v (^) is characterized here by so-called event parameters ^ P ^*(^) , for example here its instant of start of flow t e(j) , the flow duration Δt e(j) , and the average flow rate, or even the elapsed volume V (i) :P ^ ^*(^) = .t^(^) ; ∆t ^(^) ; V (^) / . These parameters are given here as an example, and other parameters are possible. The measured information can be transmitted in real time to the processing unit, for example at a predefined frequency or as soon as a fluidic event ^ Ev (^) is completed, in the form of one or more electronic messages each containing the parameters ^ P ^*(^) of the fluidic event ^ Ev (^) considered.

[0045] Alternatively, the measurement signal S ^ my ^( (t) can be a vector whose values ​​correspond to the flow rate measured at regular frequency. In the case of a pulse water meter, the measurement signal S ^ my ^((t) can be a vector comprising only the instant at which a pulse is emitted. This pulse is emitted when a predefined volume of the fluid of interest has flowed (e.g. 1 pulse per liter).

[0046] The processing unit UT is connected to the fluidic measurement sensor CM BS wired or wireless. It can be placed in the building to be monitored BS or can be located remotely from it. It is a computer that includes a calculator and at least one memory. It allows the implementation of the operations of the monitoring process to determine the class L ^ ^^(^) fluidic events E ^ v (^) present in the measurement signal S ^ my ^((t). The computer comprises a programmable processor capable of executing instructions recorded on an information recording medium. The memory contains instructions for implementing the monitoring method. It is also adapted to store the information received by the fluidic measurement sensor CM BS , and here comprises at least one prediction model f1 (or f1, f2) having been parameterized during a supervised learning phase from at least one database BD1 (or BD1, BD2).

[0047] Figure 2 is a flowchart of a monitoring method, according to one embodiment, of the classes L ^ ^^(^) of EC consumer elements to be associated with each of the fluidic events ^ Ev (^) present in the measurement signal S ^ my ^((t). The method comprises a phase 100 of generating a database BD1, a phase 200 of learning, followed by a phase 300 of prediction.

[0048] In this example, the building to be monitored BS is a residential house with several inhabitants. The monitoring method is obviously not limited to this example. The building to be monitored BS comprises consumer elements EC of different known classes L, namely here classes: R (taps), C (flush toilets), D (showers), LL (washing machines) and LV (dishwashers).

[0049] Phase 100: Generation of the database BD1.

[0050] The database BD1 is generated on the basis of measurement signals Dmes ^^^^(^) ^ ^^4,^,5… (t) representative of the actual flow rate (i.e. the measured and not simulated flow rate) of each of the EC consumer elements of N real reference buildings Brief (n), with n ranging from 1 to N>1. These are therefore classified (i.e. labeled) measurement signals, given that we have a measurement signal for each of the EC consumer elements. The real reference building Bref(n) is identical or similar to the building to be monitored BS, in the sense that it has the same use (here a residential house) and includes EC consumer elements of the same class L (here: tap R, shower D, flush C, washing machine LL, and dishwasher LV).

[0051] Phase 100 of generating the database BD1 is illustrated in more detail in Figure 4. This phase 100 includes the following steps: o a step 110 of acquiring, by experimental measurement, measurement signals Dmes ^^^^(^) ^ ^^4,^,5… (t), representative of the actual flow rate of each of the EC consumer elements of the N real reference houses Bref(n); o a step 120 of determining profiles, called statistics, of consumption Pstat ^!"#($) ^ ^^4,^,5…of each simulated building Bsim(m), with m ranging from 1 to M>N; o a step 130 of generation, by digital simulation, of simulated signals (t), representative of the flow rate of each of the EC consumer elements of the M simulated houses Bsim (m) ; o a step 140 of determining, for each of the fluid events Ev identified in the simulated signals, at least the event parameters PEV and the temporality parameters PT; o a step 150 of generating the database BD1.

[0052] During step 110, measurement signals Dmes are acquired ^^^^(^) ^ ^^4,^,5… (t) representative of the actual flow rate of each of the EC consumer elements of the N actual reference houses. Brief (n), over at least one monitoring duration Δts (here at least one day). The number N may be equal, for example, to a few units, or even to a dozen. It is distinguished from the number M of simulated houses which may be equal to a few hundred, thousands, tens of thousands or even more.

[0053] Figure 3A schematically and partially illustrates a real reference house Bref identical or similar to the house to be monitored BS (here a residential house). The reference house is said to be real insofar as it is occupied by real users, and is therefore not a simulated house. It includes consumer elements EC whose classes are L = { R, C, D, LL, LV}. These are the classes of taps R, flush toilets C, showers D, washing machines LL, and dishwashers LV. The reference house Bref may have a number of inhabitants identical or not to that of the house to be monitored BS.

[0054] The N real reference houses Bref(n) are connected, directly or indirectly, to a device for determining at least one prediction model f1, f2. This device comprises fluidic measurement sensors CM. EC and a UT processing unit. In this example, the CM fluidic measurement sensors EC acquire Dmes measurement signals ^^^^(^) ^ ^^4,^,5… (t) which correspond to the temporal evolution of the flow rate of each of the EC consumer elements of the N real reference houses, and transmit them to the UT processing unit. In other words, each EC consumer element is equipped with a CMEC fluidic measurement sensor. The measurement signals Dmes ^^^^(^) ^ ^^4,^,5… (t) are here vectors indicating the value of the flow rate measured at a predefined frequency, for example every second. They are noted: Dmes ^^^^(^) ( ^^^^(^) ^ ^^4,^,5… t) = { Dmes4(t) ; Dmes ^^^^(^) ^ (t) ; Dmes ^^^^(^) 5 (t); Dmes^ 7; Dmes ^^^^(^) 7 ^ (t)}. They can obviously present other formats, such as in particular a succession of electronic messages relating to each of the measured fluid events.

[0055] Figure 3B illustrates examples of measurement signals from a Dmes valve R (t), of a Dmes shower D (t), of a flush toilet Dmes C (t), of a Dmes washing machine LL(t), and a dishwasher DmesLV(t). We note that it is possible to extract the fluidic events Ev and the corresponding event parameters PEv, which relate in particular to the initial instant te, the flow duration Δte, the elapsed volume V, the average flow rate, etc. We can also determine a statistical consumption profile associated with each reference house Bref(n), the maximum flow rate, etc., but also the number of uses, the average frequency and the standard deviation of uses, etc.

[0056] Then, during step 120, we determine statistical consumption profiles ^Pstat ^!"#($) ^ ^ ^ #^^;% of the M simulated houses, from information from the acquired measurement signals ^Dmes ^^^^(^) ^ ^ (t)^ ^^^;^ , and more precisely, from statistical consumption profiles ^Pstat ^^^^(^) ^ ^ ^ ^^^;^associated with the N reference houses which are determined from the acquired measurement signals.

[0057] Thus, a statistical profile of a building groups together information related to this building and its occupants / users, such as the number of occupants, the number and class of the different EC consuming elements. It also includes statistical information related to the fluid consumption habits of each of the occupants, such as the typical times of use of the different EC consuming elements, the flow times, the breaks between each fluid event (in particular for showers). Finally, it includes statistical information related to the EC consuming elements themselves: maximum flow rate, average flow rate, etc. This consumption information is called statistical insofar as it can include an average value and a variability (standard deviation) associated with a given distribution (normal, log-normal, exponential, bimodal, etc.).

[0058] Thus, for each reference house Bref(n), we obtain the information related to the house and its occupants. In addition, from the measurement signals acquired for each of the consumer elements EC, we determine the statistical consumption information. We thus define the corresponding statistical consumption profile Pstat. ^^^^(^) ^ ^ .

[0059] The statistical consumption profiles Pstat are then determined. ^!"#($) ^ ^ for each of the M simulated houses Bsim (m) , from the statistical consumption profiles of the reference houses Bref(n). As indicated previously, the number M can be very high compared to the number N, for example being around 100000 while N would be around 10. The statistical consumption profiles Pstat ^!"#($) ^ ^are determined by varying the different parameters of the profile. This will avoid subsequently parameterizing the prediction model f1 with a database BD1 whose values ​​would not be sufficiently representative of the possible effective variability of the fluid consumption of real houses, and therefore of the house to be monitored BS.

[0060] Thus, concerning the parameters linked to the house and its occupants, houses are generated by varying the number of occupants, the number of consumer elements of each class, the number of classes present (thus some houses may not have a washing machine or a dishwasher). Statistical parameters linked to the consumption habits of each of the occupants are also varied, for example the typical schedules and the frequency of use of the different consumer elements EC. Thus, for example, for some occupants, the shower is taken in the morning, for others in the evening.For some, the shower flow time will be long, for others short, etc. Finally, we vary the parameters linked to the EC consumer elements themselves in terms of average flow rate, maximum flow rate, standard deviation, etc. We thus obtain the statistical consumption profiles Pstat. ^!"#($) ^ ^ of the M simulated houses.

[0061] Then, during a step 130, the so-called simulated signals ^Dsim are generated ^!"#($) ^ ^ (t)^ #^^;% representative of the consumption flow rate of each of the EC consumer elements of the M simulated buildings, from the corresponding statistical consumption profiles. These simulated signals are said to be classified to the extent that a consumer element class is associated with each of the signals.

[0062] Thus, for each simulated house (and for each occupant), and taking into account the parameters of the corresponding statistical consumption profile, the temporal evolution of the flow rate Dsim is simulated ^!"#($)5 (t) of each shower during a monitoring period Δt s (here at least one day), as well as that of Dsim ^!"#($) ^ (t) of each flush and that Dsim ^!"#($) 4 (t) of each tap. We can also simulate that Dsim ^!"#($) 77 ( t ) washing machines and the Dsim one ^!"#($) 7 ^ ( t ) dishwashers. Of course, filters are present to prevent simulated signals of class D 'shower', for several occupants of the same simulated house, from presenting Ev fluid events D simultaneous… while the simulated house Bsim (m) has only one shower.

[0063] Each simulated signal Dsim ^!"# ^ ^ comprises one or more fluid events. For example, it has the format of a vector indicating the value of the flow rate at a frequency of, for example, one second over a period of one day, and preferably over a period of several days, or even weeks. The simulated signals can thus take the form of the signals illustrated in Fig. 3B.

[0064] Thus, a large number of realistic signals are obtained which are representative of the actual flow rate of the EC consumer elements in real houses, which will make it possible to generate a realistic database BD1 of large dimension. This great diversity of simulated signals will make it possible to avoid prediction biases which could be induced by a database which would not be sufficiently representative of the possible actual diversity of consumption habits in real houses.

[0065] Then, during a step 140, parameters associated with each fluid event associated with the fluid consumption of the simulated houses Bsim are determined. (m) , and more precisely event parameters P Ev and temporal parameters P T .

[0066] For this, simulated global flow signals Ssim are determined. Bsim(m) (t), which are each representative of the overall flow rate of the corresponding simulated house Bsim(m), from the simulated signals Dsim ^!"#($) ^ ^^4,^,5… ( t ) These are obviously classified (labeled) signals, since they are associated with an L signal. Bsim(m) (t) indicating the corresponding class. This amounts to obtaining, for each simulated house Bsim (m) a signal whose format is identical to the measurement signal S ^ my ^( (t) of the house to be monitored BS. In this example, the measurement signal S ^ my ^((t) is formed from a succession of fluidic events .P ^ ^^(^) / and more precisely of a succession of parameters representative of the fluidic events.

[0067] For this, we add all the simulated signals Dsim ^!"#($) 5 (t), Dsim ^!"#($) ^ (t), Dsim ^!"#($) 4 (t), Dsim ^!"#($) 77(t) and Dsim ^!"#($) 7 ^ (t) relating to the same simulated house Bsim (m) . We thus obtain, for each simulated house Bsim(m), the simulated global flow signal Ssim Bsim(m) (t) accompanied by the class L signal Bsim(m) (t) corresponding.

[0068] Thus, for each of the simulated global flow rate signals Ssim Bsim(m) (t), we identify the fluidic events Ev Bsim(m) present and we associate them with the corresponding class. Let us recall that fluidic events are events with non-zero flow rate framed before and after by a period with zero flow rate.

[0069] However, a fluidic event Ev Bsim(m)of the simulated global flow signal Ssim Bsim(m) (t) can be derived from the at least partial juxtaposition of several fluidic events from the simulated signals Dsim ^!"#($) ^ ^ ( t ) . Also, we can attribute the new class 'mixed' or 'mixture', noted M, to each fluid event Ev of the same simulated global flow signal Ssim Bsim(m) (t) and originating from at least two consumer elements.

[0070] Then, we characterize each of the fluidic events Ev of the same simulated signal of global flow rate Ssim Bsim(m) (t) by event parameters P Ev , namely for example the start time te, the flow duration Δte and the elapsed volume V or equivalent (average flow rate, etc.). We thus obtain a simulated overall flow rate signal Ssim Bsim(m)(t) comprising a succession of fluidic events and their characteristic parameters PEV, with the corresponding classes LEV: { PEv ; LEv}.

[0071] Furthermore, we determine so-called temporality parameters P T for each of the fluid events of the simulated global flow signals Ssim Bsim(m) (t). This involves enriching the information associated with the fluid events Ev of each simulated global flow signal Ssim Bsim(m) (t), with correlation information with similar fluid events located before and after each of the fluid events considered. Indeed, it appears that enriching the database with such temporal parameters makes it possible to improve the prediction performance of the model.

[0072] As illustrated in Figure 3C, a plurality of time slots C located before and after the initial time t are defined e of a fluidic event Ev. We can thus define NC avtime slots located before the initial instant t e and NC ap time slots located after the initial instant t e . In this example, NCav and NCap are identical but they could be different from each other, and here are each equal to 5. In addition, the upstream Cav and downstream Cap time slots have the same duration two by two, but they could have different durations. We can thus have: - two slots C(-1) and C(+1) going from [-tnc1 ; te] and from [te ; +tnc1], for example going from te to more or less 7min; - two slots C (-2) etc (+2) ranging from [-t nc2 ; -t nc1 ] and [+t nc1 ; +t nc2 ], for example going from -30min to - 7min, and going from +7min to +30min; - two slots C(-3) and C(+3) going from [-tnc3; -tnc2] and from [+tnc2; +tnc3], for example going from -60min to - 30min, and going from +30min to +60min; - two slots C (-4) etc (+4) ranging from [-t nc4 ; -tnc3 ] and [+t nc3 ; +t nc4], for example ranging from -120min to -60min, and ranging from +60min to +120min; - two time slots C(-5) and C(+5) ranging from [-tnc5; -tnc4] and from [+tnc4; +tnc5], for example ranging from -240min to -120min, and ranging from +120min to +240min;

[0073] We then determine the number of fluid events { NC(p)}p=-NCav;+NCap similar to that considered in terms of flow rate (for example, same average flow rate plus or minus a predefined tolerance, for example equal to 1L / min), whose initial instant is located in one or other of the time slots. This makes it possible to establish the correlation between the fluid events of the same class. Indeed, it appears that the use of a consumer element, for example a tap or a shower, can present a succession of fluidic events of the same flow rate, or even, in addition, of the same flow duration.This may be the case when using the shower, the tap for washing dishes, the cleaning program of a washing machine and a dishwasher, etc. The closest time slots allow, for example, the identification of showers, while the most distant time slots concern more the washing machines and dishwashers. On the other hand, certain consumer elements may not present such a temporal correlation between fluid events.

[0074] Note that it is also possible to identify similar “long” fluid events, for example, having a flow duration greater than a predefined value, such as for example 30 seconds, and / or similar “short” fluid events, for example, having a flow duration less than a predefined value, such as for example 10 seconds.It is also possible to calculate, for each of the time slots, the cumulative volume of all the fluid events identified in the time slot in question.

[0075] Thus, each of the simulated global flow signals Ssim. Bsim(m) comprises, per fluid event Ev, PEV event parameters representative of the fluid event in question (initial instant, flow duration, etc.) as well as PT temporal parameters.

[0076] Furthermore, it may be advantageous to determine, for each of the simulated global flow signals Ssim Bsim(m) , additional parameters called comparison parameters P Ev|Bsim , where a flow rate (average flow rate for example) of each of the fluidic events is compared to an overall flow rate (average flow rate) of the simulated house Bsim (m). We can thus calculate, for each fluidic event, a normalized flow rate defined as being equal to the ratio of the average flow rate of the fluidic event to the average overall flow rate of the simulated house Bsim(m).

[0077] Finally, we generate the database BD1 as being the collection of all the fluidic events of all the simulated signals of overall flow rate Ssim Bsim(m) , and more precisely event parameters P Ev , temporality parameters P T, and here comparison parameters PEv|Bsim. For each event, we also have the corresponding class LEv.

[0078] Thus, we obtain a set of NA fluidic events Ev(i) with i ranging from 1 to NA, each fluidic event Ev(i) being defined by the event parameters PEV(i), the temporality parameters PT(i) and the comparison parameters PEV|BR(i), and by the corresponding class LEv(i). This set then forms the database BD1. The number NA can thus be very high, especially since it comes from the large number M of simulated houses Bsim (m) .

[0079] Note here that the fluidic events Ev (i)are mixed, in the sense that they are no longer associated with this or that simulated house Bsim(m) nor with this or that measurement day. On the other hand, temporal correlation information is present via the temporality parameters PT, which makes it possible to improve the prediction performance of the model.

[0080] Phase 200: Parameterization of the prediction model.

[0081] Phase 200 then consists of training the prediction model from the database BD1, that is to say, parameterizing the prediction model automatically so that from the input data which are here the parameters { P Ev(i) ; P T(i) ; P Ev|Bsim(i)} i=1 ;NA , it determines the output data which are the corresponding classes { L Ev(i)} i=1 ;NA. Thus we note: Lev(i) = f1( PEv(i); PT(i); PEV|BR(i) ), where f1 is the parameterized prediction model.

[0082] As indicated previously, different prediction model algorithms can be used, but the inventors have found that the algorithm based on boosting-type decision trees, and more precisely the gradient boosting type, has good performance.

[0083] Phase 300: Classification.

[0084] Phase 300 consists of performing the prediction of the class (classification) of the fluid events E ^ v (^) present in a measurement signal S ^ my ^( (t) representative of the overall flow rate of the house to be monitored BS, this signal having been acquired by the CMBS fluid measurement sensor (see fig.1A). More precisely, it is a question of estimating class L ^ ^^(^) of each of the fluidic events E ^ v (^) .

[0085] Also, during a step 310, the measurement signal S is acquired. ^ my^( (t) by at least one CMBS fluidic measurement sensor (here by a single sensor). It corresponds to a succession of event parameters ^ P ^^(^,^) representative of fluidic events E ^ v (^,^) that it contains: S ^ my ^( (t) = . P ^ ^*(^,^) / ^^^,^0. Event parameters ^ P ^*(^,^) are identical to the P event parameters Ev from the BD1 database, and include in this example, the initial moment of the fluid event, its flow duration and the elapsed volume. The measurement signal S ^ my ^( (t) is acquired by the CM fluidic measurement sensor BS over a monitoring period Δt s(k) , for example over a day. When the monitoring duration k is over, the fluidic measuring sensor acquires a new measuring signal S ^ my ^( (t) over the following monitoring duration Δt s(k+1), and so on.

[0086] Then, during a step 320, the additional parameters are determined, here the temporality parameters ^ P )(^,^) and the comparison parameters P ^ ^*|^((^,^"→^), so that the measurement signal S ^ my ^( (t) forms input data for the prediction model f1 having the same parameters (same dimensions) as during the learning phase 200.

[0087] The temporality parameters are determined ^ P )(^,^) in the same manner as that described previously, by identifying the number of fluid events similar to that considered and located in the upstream and downstream time slots.

[0088] Where appropriate, the comparison parameters P are determined ^ ^*|^((^,^"→^)in the same way as described previously, by calculating the normalized flow rate of each fluid event E ^ v (^,^) , i.e. the ratio of the average flow rate of each fluid event^ Ev (^,^) on the average overall flow rate of fluid consumption of the house to be monitored BS. Note that the average overall flow rate of the house to be monitored BS is advantageously the average overall flow rate of all the measurement signals S ^ my ^( (t) having been previously acquired, from the initial acquisition (index ki). Thus, the average overall flow rate of the house to be monitored BS will be more precise as the acquisitions of the measurement signal S ^ my ^( (t) will follow one another.

[0089] Finally, during a step 330, the classes (classification) of the fluidic events E are predicted (classified). ^ v (^,^) present in a measurement signal S ^ my ^( (t). The input data provided to the prediction model f1 are .P ^ ^*(^,^) ; ^ P )(^,^) ; ^ P ^*|^((^,^"→^) / , and the output data are the estimated classes L ^ ^*(^,^)for each fluidic event ^ Ev (^,^) , so that we have: L ^ ^*(^,^) = f1: .P ^ ^*(^,^) ; ^ P )(^,^) ; ^ P ^*|^((^,^"→^) / ;. When the prediction is made for the measurement signal S ^ my ^( (t) of the monitoring duration of index k, we repeat phase 300 for a new measurement signal S ^ my ^( (t) acquired over the monitoring duration of index k+1.

[0090] Thus, the monitoring method is able to determine the consumer element class L ^ ^*(^,^) for each fluidic event E ^ v (^,^) present in the measurement signal S ^ my ^( (t) representative of the overall consumption flow rate of the fluid of interest of the building to be monitored BS. By the fact that the input data includes the temporal parameters ^ P )(^,^), it appears that the prediction model exhibits improved accuracy.

[0091] Note that the prediction step 330 may include constraints which block the attribution of one or other of the classes to certain fluid events. ^ Ev (^,^)and assign a class called 'Other' when the estimated class is prohibited. Thus, the events of each class respect conditions, for example, on the flow duration, the average flow rate, the number of similar successive fluid events, etc. Also, for example the class 'Shower' could not be assigned to a fluid event whose flow duration would be less than 5 seconds or greater than 900 seconds.

[0092] Phase 400: Monitoring

[0093] Finally, note that the monitoring method can make it possible to detect unusual consumption (for example excessive or insufficient) of the consuming elements of one or other of the predefined classes, to then communicate it to a user.

[0094] As such, the method comprises a monitoring phase 400, which follows the classification phase 300. During this phase 400, we identify (step 410), among at least a part of the predicted classes {L ^ ^^(^,^)}j=1 ;NP, and preferably for each of the predicted classes, a class L ^ ^^ for which a PS tracking parameter 7^ associated with the corresponding fluidic events presents a gap E 7^ to a reference value PS ^7,^^^ greater than a threshold deviation E 7^,<= .

[0095] The PS tracking parameter 7^ can be derived from event parameters ^ P ^^ fluid events of the predicted class considered, such as for example an average or instantaneous flow rate, an average flow duration, etc. Then, we calculate a deviation E 7^ between the PS tracking parameter 7^ is a predefined reference value PS 7^,^^^This reference value is associated with the class considered. It can therefore be a reference value of the average flow rate, in the case where we consider for example the class of taps (or showers, washing machines, etc.). It can be a predefined value that is constant over time, or derived from a sliding value (average, minimum or maximum value, etc.) determined over a predefined time window (previous days, weeks, months, years, etc.). This value can be derived from consumption habits from the building to be monitored BS, or from real reference buildings In short, simulated buildings Bsim, or even from national statistical data, etc. The gap E 7^ can be the absolute value difference (or ratio) between the value of the tracking parameter and its reference value: > PS 7^ − PS 7^,^^^>.

[0096] Finally, during a step 420, if such a class is identified, this information is communicated to a user of the monitoring method. This communication can take different forms, and can be a display on a monitoring screen. The method can also include a step of correction by the user so that the consuming elements of the identified class then have a monitoring parameter having a deviation less than the predefined threshold deviation. For example, this can result in a replacement of the shower heads, or even a replacement of the consuming elements in question by less consuming elements.

[0097] Figure 5 illustrates a flowchart of a monitoring method according to another embodiment, which differs from the method of Fig. 2 essentially in that it uses two prediction models f1 and f2 successively.The principle is to use the first prediction model f1 to perform a first level of classification, and thus deduce additional parameters, which will then be added to the input data of the second prediction model f2 which will then perform a second level of classification. This makes it possible to further improve the performance of the monitoring method.

[0098] The monitoring method comprises a phase 100 of generating two databases BD1 and BD2, then a phase 200 of learning the two prediction models f1 and f2, and finally a phase 300 of two-level prediction.

[0099] During phase 100 of generating the databases BD1 and BD2, steps 110 to 140 are identical or similar to those described previously and are not described again here.

[0100] Step 150 consists of determining (sub-step 151) the event parameters P. Ev, the temporality parameters PT (and here the comparison parameters PEv|Bsim) to obtain the first database BD1. Here, we operate in an identical or similar manner to the process in fig.2. The database BD1 therefore includes these parameters PEv, PT and here PEv|Bsim, as well as the classes L EV assigned to each fluidic event Ev.

[0101] Step 150 also consists of determining (sub-step 152) additional parameters P Ev|L so-called 'class comparison' (while the P parameters Ev|Bsimare comparison parameters with respect to the simulated house Bsim), and then to obtain the second database BD2. Thus, the database BD2 includes the event parameters PEv, the temporality parameters PT, the comparison parameters PEv|Bsim, and the comparison parameters by class PEv|L. It also includes the classes LEv assigned to each fluid event Ev.

[0102] Thus, for each of the simulated global flow signals Ssim Bsim(m) , we determine the additional parameters P Ev|L, where the average flow rate of each of the fluid events is compared to the average overall flow rate of the fluid events of each of the classes. It is thus possible to calculate, for each fluid event, the normalized class flow rate defined as being equal to the ratio of the average flow rate of the fluid event to the average overall flow rate of the fluid events of the same class, for the simulated house considered.

[0103] Then, during phase 200, the two prediction models f1 and f2 are parameterized by supervised machine learning. More precisely, during step 201, the prediction model f1 is parameterized from the database BD1, and during step 202, the prediction model f2 is parameterized from the database BD2.

[0104] Finally, during phase 300, the acquisition steps 310 and preprocessing 320 of the measurement signal S are carried out. ^ my ^((t) in an identical or similar manner to the steps described in connection with fig.2.

[0105] Then, during a step 331, a first level of prediction of the fluidic events E is carried out ^ v (^,^) present in a measurement signal S ^ my ^( (t), using the first model f1. The input data provided to the model f1 are . ^ P ^*(^,^) ; ^ P )(^,^) ; ^ P ^*|^((^,^"→^) / , and the output data are the estimated classes L ^ ^*(^,^) for each fluidic event ^ Ev (^,^) , so that

[0106] Then, during a step 332, the comparison parameters are determined by class. ^ P ^*|7(^,^) . This step is possible to the extent that the preceding fluidic events E ^v(^,^"^^→^@^) were classified during the previous iterations, either by the model f1 or by the model f2. It is then possible to determine the characteristic flow rate (e.g. average, median flow rate, etc.) of each class of consumer elements of the house to be monitored BS, and then to compare the determined flow rate of the fluidic event to be classified with the characteristic value of each class (ratio of the average flow rate of the fluidic event considered to the average (or median) flow rate of all the fluidic events that fall within the class considered). And so on for all classes. Of course, it is possible to determine the comparison parameters by class P ^ ^*|7(^,^) for each of the classes, or for only one or the other of them. We could thus limit ourselves to considering only the class of toilet flushes.

[0107] Finally, during step 333, we proceed to the second level of prediction of fluidic events ^ Ev(^,^) present in a measurement signal S ^ my ^( (t), using the second model f2. The input data provided to the model f2 are . ^ P ^*(^,^) ; ^ P )(^,^) ; ^ P ^*|^((^,^"→^) ; ^ P ^*|7(^,^) / , and the output data are the estimated classes L ^ ^*(^,^) for each fluidic event ^ Ev (^,^) , so that It appears that the monitoring method according to this embodiment has improved performance.

[0108] In this respect, Figures 6A to 6C illustrate changes in the distribution function (also called cumulative distribution function) of the flow rate (Fig. 6A), the flow duration (Fig. 6B) and the elapsed volume (Fig. 6C) for the fluid events resulting from the simulated global flow signal Ssim BSim whose statistical consumption profile Pstat ^ ^ ^!"# is identical to the Pstat profile ^ ^ ^ ^^^ of a real reference house. These curves are compared with those from the real house considered.

[0109] Note here that the flow distribution function CDF D (d) corresponds to the probability of obtaining a value of the average flow rate less than or equal to d: CDF D (d) = P(D≤d). This probability P(D≤d) is equal to the proportion %Δts|D of the cumulative flow duration Δtec|D during the monitoring duration Δts (here the day) for which the average flow rate D is non-zero and at most equal to d. The distribution function CDFD(d) is between 0% and 100%, and the average flow rate D varies between zero and the maximum flow rate measured by the measurement signals DmesEC(t).

[0110] Thus, it appears that the simulated signals Dsim EC (t), generated during step 130 described previously, show a very good correspondence with the measurement signals Dmes EC(t). As a result, the simulated global flow signal Ssim(t), generated from the simulated signals DsimEC(t), is particularly realistic. This is therefore also the case for the database.

[0111] Finally, Figures 7A, 7B and 7C illustrate the prediction performances for different classification methods, and here respectively the precision, the recall and the F1-score. The precision is defined as being equal to the ratio of the number of fluid events correctly classified in a given class to the total number of fluid events classified in the class in question. The recall is defined as being equal to the ratio of the number of fluid events correctly classified in a given class to the total number of fluid events actually belonging to the class in question. Finally, the F1-score corresponds to the harmonic mean of the precision and the recall.

[0112] We consider here the results of a model f0 which would have been parameterized by the PEV event parameters alone and not also by the PT temporality parameters nor by the P parameters. Ev|Bsim . We also consider the results of the model f1 described with reference to fig.2, and those of the two-level model f1 and f2 described with reference to fig.5. The classes considered here are those of the flush toilets C, that of the showers D, that of the washing machines LL, those of the dishwashers LV, and finally that of the consumer elements R* which it is not possible to distinguish from the others (here essentially the taps R and the 'mixed' class M with other consumer elements). It appears that the performances of the models f1 and f1+f2 are indeed better than those of the model f0.

[0113] Particular embodiments have just been described. Different variants and modifications will appear to those skilled in the art.

Claims

CLAIMS 1. Computer-implemented method for monitoring the consumption of a fluid of interest by a building to be monitored BS, by classification of fluid events E ^ v (^,^) consumption of the fluid of interest by EC consumer elements of the building to be monitored BS, among several predefined classes L ^ ^^(^,^) of EC consumer elements, the method comprising the following phases: o a generation phase (100) of at least one database (BD1, BD2), comprising the following steps: ^ acquiring (110) measurement signals ^Dmes ^^^^(^) ^ ^ (t)^ ^^^;^ representative of the consumption flow rate of each of the EC consumer elements of N real reference buildings {Bref (n)} n=1 ;N , with N>1, by means of several fluidic measurement sensors (CM EC) each connected to an EC consumer element of the N real reference buildings, the measurement signals then being said to be classified, the EC consumer elements being of the same classes as those of the building to be monitored BS; ^ define (120) consumption profiles ^Pstat ^!"#($) ^ ^ of M so-called simulated buildings {Bsim(m)}m=1 ;M, with M>N, from the acquired classified measurement signals, each simulated building comprising EC consumer elements of the same classes as those of the building to be monitored BS; ^ generate (130), by numerical simulation, simulated signals ^Dsim ^!"#($) ^ ^ ( t ) ^ #^^;%representative of the consumption flow rate of each of the EC consumer elements of the simulated buildings {Bsim(m)}m=1;M having the defined consumption profiles, the generated simulated signals then being said to be classified; ^ determine (140), for each of the fluidic events {Ev(i)}i=1;NA identified in the generated classified simulated signals: ^ so-called event parameters {PEv(i)}i=1;NA, representative of a duration and an elapsed volume of each of the fluidic events; and ^ so-called temporality parameters {P T(i)} i=1 ;NA , representative of a number of so-called similar fluid events, presenting a flow rate substantially identical to that of the fluid event considered and located in predefined successive time slots located before and after the fluid event considered; ^ generate (150) the database (BD1; BD2) comprising, for each of the fluidic events {Ev(i)}i=1;NA identified: the event parameters {PEv(i)}i=1;NA, the temporality parameters {P T(i)} i=1 ;NA , and the class {L EV(i)} i=1 ;NA corresponding consumer element EC; o a parameterization phase (200) of at least one prediction model (f1, f2) by supervised automatic learning from the database (BD1; BD2); o a classification phase (300), comprising the following steps: ^ acquire (310) a measurement signal S ^ my ^( (t) representative of the overall consumption flow rate of the building to be monitored BS by means of at least one fluid measurement sensor (CMBS) connected to a fluid inlet of the building to be monitored BS; ^ determine (320), for each of the fluid events {E ^ v (^,^)}j=1 ;NP identified in the acquired measurement signal S ^ my ^((t), event parameters {P ^ ^^(^,^)}j=1 ;NP and temporal parameters { ^ P )(^,^)}j=1 ; corresponding NP ; ^ predict (330), by the prediction model (f1; f2), for each of the fluidic events {E ^ v (^,^)} j=1 ;NP whose event parameters {P ^ ^^(^,^)} j=1 ;NP and the temporality parameters { ^ P )(^,^)} j=1 ;NP train input data to the prediction model, the corresponding class consumers EC; o a monitoring phase (400) of the building to be monitored BS, comprising: ^ identifying (410) a class of consumer elements, among at least part of the predicted classes, for which a monitoring parameter PS 7^ associated with the corresponding fluidic events presents a gap E 7^ to a reference value PS 7^,^^^ greater than a threshold deviation E7^ ,<=; then ^ communicate (420) to a user the identified class of consumer elements.

2. Monitoring method according to claim 1, in which the temporal parameters P T(i) And ^ P )(^,^) also include, for each of the time slots, the cumulative volume of the fluid of interest flowed and the flow duration.

3. Monitoring method according to claim 1 or 2, in which the event parameters PEv(i) and ^ P ^*(^,^) comprise, for each fluid event, the initial instant of the fluid event, the flow duration, the elapsed volume and / or the average flow rate.

4. Monitoring method according to any one of claims 1 to 3, in which the database and the input data of the prediction model comprise so-called comparison parameters, denoted respectively PEv|Bsim(i) and P ^ ^*|^((^,^"→^) ), and defined, for each event fluidics Ev (i) , E ^ v(^,^) , as the ratio of an average flow rate of the fluid event considered to an average overall flow rate of all fluid events Ev of the building in question.

5. Monitoring method according to any one of claims 1 to 4, in which the classification phase (300) is repeated, the measurement signals S ^ my ^( (t) of the building to be monitored (BS) then being acquired over successive predefined monitoring durations.

6. Monitoring method according to claims 4 and 5, in which, in the comparison parameters P ^ ^*|^((^,^"→^) present in the input data of the prediction model, the average global flow rate of all fluid events E ^ v (^,^)takes into account the average overall flow rate of the previous iterations.

7. Monitoring method according to any one of claims 1 to 6, in which: o the generation phase (100) comprises the following steps: ^ determination (151) of a first database (BD1) where each fluid event (Ev (i) ) is defined by at least said event parameters (P EV(i) ) and by said temporality parameters (PT(i)); ^ determination (152) of a second database (BD2) where each fluidic event (Ev(i)) is defined by at least said event parameters (PEV(i)), by said temporality parameters (PT(i)), and by so-called comparison parameters by class (PEV|L(i)) defined, for each fluidic event (Ev(i)) of at least one class considered, as the ratio of an average flow rate of the fluidic event considered to an average overall flow rate of all the fluidic events of the class considered (Ev (i)); o the parameterization phase (200) comprises the following steps: ^ parameterization (201) of a first prediction mode (f1) by supervised automatic learning from the first database (BD1); ^ parameterization (202) of a second prediction mode (f2) by supervised automatic learning from the second database (BD2); o the classification phase (300) comprises the following steps: ^ determination (320), for fluid events (E ^ v (^,^) ) present in the measurement signal S ^ my ^( (t), event parameters ( ^ P ^*(^,^) ) and temporal parameters (P ^ )(^,^) ) corresponding; then ^ prediction (331) by the first prediction model (f1), for each of the fluid events ( ^ Ev (^,^) ), which are defined by the event parameters (P ^ ^*(^,^) ) and the temporality parameters (P ^ )(^,^)) which form input data of the first model of prediction (f1), of the class (L ^ ^*(^,^) ) of corresponding consumer element (CE); then ^ determination, from the previously predicted classes, of comparison parameters per class ( ^ P ^*|7(^,^) ) corresponding; then ^ prediction (331) by the second prediction model (f2), for each of the fluid events ( ^ Ev (^,^) ), which are defined by the event parameters (P ^ ^*(^,^) ), the temporality parameters ( ^ P )(^,^) ), and the comparison parameters by class ( ^ P ^*|7(^,^) ) which form input data of the second prediction model (f2), of the class (L ^ ^*(^,^) ) of corresponding consumer element (EC).

8. Monitoring method according to any one of claims 1 to 7, wherein, during the acquisition step (110) of the measurement signals ^Dmes^^^^(^) ^ ^ (t)^ ^^^;^ , the N real reference buildings are connected to a device for determining at least one prediction model, comprising the fluidic measurement sensors (CMEC) adapted to acquire the measurement signals for each of the consumer elements (EC) of the real reference building considered (BR), and a computer connected to the fluidic measurement sensors of the reference building considered (BR) and adapted to carry out the parameterization phase (200).

9. Monitoring method according to any one of claims 1 to 8, wherein, during the acquisition step (310) of the measurement signal S ^ my ^( (t) of the building to be monitored (BS), this is connected to a classification device comprising the fluidic measurement sensor (CM BS ) adapted to acquire the measurement signal S ^ my ^((t), and a computer connected to the fluid measurement sensor of the building to be monitored (BS) and adapted to carry out the classification phase (300).

10. Monitoring method according to any one of claims 1 to 9, in which the monitoring phase (400) comprises, following the communication step, a step of correction by the user so that the consuming elements of the identified class then have a value of the monitoring parameter lower than the predefined threshold value.

Images