EP2784742A1 - Verfahren und Vorrichtung zur Identifizierung der Verbrauchs- und/oder Erzeugungsquelle - Google Patents

Verfahren und Vorrichtung zur Identifizierung der Verbrauchs- und/oder Erzeugungsquelle Download PDF

Info

Publication number
EP2784742A1
EP2784742A1 EP14162908.9A EP14162908A EP2784742A1 EP 2784742 A1 EP2784742 A1 EP 2784742A1 EP 14162908 A EP14162908 A EP 14162908A EP 2784742 A1 EP2784742 A1 EP 2784742A1
Authority
EP
European Patent Office
Prior art keywords
sources
consumption
individuals
probability
activity
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
EP14162908.9A
Other languages
English (en)
French (fr)
Inventor
Lionel Caudron
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Synapses
Original Assignee
Synapses
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Synapses filed Critical Synapses
Publication of EP2784742A1 publication Critical patent/EP2784742A1/de
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply

Definitions

  • the present invention relates to the field of energy and in particular the consumption and / or the production of energy or fluids, in particular electricity.
  • the present invention more particularly relates to a device and a method for identifying sources of consumption and / or production.
  • a general problem in the field concerns the knowledge of the consumption and / or production of fluids (or sources of consumption or energy production). This problem is widespread in most sectors of human activity since knowledge of the consumption and / or production of fluids is essential to control their volume and cost. Indeed, fluids such as electricity, gas, oil, water, steam or any other solid particles, liquid or gaseous in circulation, are generally consumed for various types of use, and especially in the industrial field, the field of services, in buildings for personal use or profession, tertiary, rental or individual.
  • the representative characteristics of the consumptions can correspond to various types of physical values, such as volume, flow rate, power, etc.
  • the present application refers to the term consumption, any type of use of any type of fluid, whether it is actually rather a production of energy (heat pump, photo panels). voltaic, hydro-electric power station, thermal power station, etc. all cited in an illustrative and non-limiting manner) or energy consumption (and that the fluid is actually consumed or simply measured or distributed, for example in terms of volume , flow, power, etc.).
  • the term fluid means in fact any solid particles, liquid or gas in circulation, which can be consumed (eg, the flow rate can be measured).
  • the term "source of consumption” means in this application any type of place or device where the consumption takes place.
  • Fluid consumption characteristics are generally measured by devices that read these characteristics on-site and / or transmit digital or analog signals transmitted to other devices. Consumption is usually determined over predefined time intervals (ie, with predefined time steps) during a specified period of time (ie, for a specified duration). The time steps generally range from the fraction of a second to several hours. This gives overall consumption data over time, for example in the form of tables or curves.
  • a particular problem in the field concerns the fact that these global consumption data measured on a site do not make it possible to know the sources of consumption, that is to say the distribution of the use of fluids in the different uses for which they were consumed (eg, distribution of consumption in different devices or places using fluids).
  • the uses may be such as motive power, thermal energy, cooling energy, lighting, heat, ventilation, physical or chemical treatments, etc.
  • a meter measuring a total power consumption eg, energy consumed, expressed in Watts-hour or KiloWatts-hour
  • this meter does not allow to know the distribution of the consumption in the various electrical appliances of the installation.
  • additional measuring devices are used but it is desired to restrict their number, it is necessary to identify the sources for which the addition of an additional measuring device will be relevant to allow better identification. the distribution of consumption.
  • a general problem related to the above problems concerns cost management, especially when the cost of the fluid increases according to the total consumption, as for example in the case of electricity. Indeed, the cost of electricity increases according to the kilowatt hours consumed. Because of suppliers' price schedules, a momentary peak of consumption therefore costs more than an identical consumption in power but distributed over time (the power is identical but the energy is different and it is the latter which is billed) . This problem therefore provides a particular motivation for identifying the sources of consumption in order to be able to predict a more homogeneous distribution over time and thus avoid peaks in consumption.
  • the aim of the present invention is to overcome certain disadvantages of the prior art by proposing a method of identifying sources of consumption making it possible to identify in an installation, from a global consumption measured over predefined time intervals and at During a fixed period of time, the consumption sources within the installation for each of the predefined time intervals of the determined period.
  • the present invention also aims to overcome certain disadvantages of the prior art by proposing a device for identifying sources of consumption for identifying in an installation, from a global consumption measured over predefined time intervals and during a fixed period of time, the sources of consumption within the installation for each of the predefined time intervals of the determined period.
  • a device for identifying sources of consumption comprising data processing means and storage means containing data representative of at least one overall consumption measured in at least one installation on at least one device. defined time interval, characterized in that the storage means contain the data necessary for the implementation, by the data processing means, of the method of identifying sources of consumption according to the invention.
  • the present invention relates to a device and a method for identifying sources of consumption and / or production (emission, generation), generally of energy or fluids, in particular electricity.
  • consumption refers in this description to both consumption and production, as explained in the preamble of the present application.
  • the term "fluid” is used in the present description in a nonlimiting manner, particularly for its analogy to a flow, and can be applied for example to electricity but also to fluids such as gas, oil, water, steam or any other solid, liquid or gaseous particles in circulation.
  • the representative characteristics of the consumptions can correspond to various types of physical values, such as volume, flow rate, power, etc.
  • the values correspond to a flow rate or are proportional to a flow rate (the power, for example, is proportional to the intensity that is comparable to a flow) because consumption is usually measured over time. Quantities that do not depend on time itself (such as volume for example) will therefore generally be related to a given time interval since the invention identifies the time distribution of the consumption of several sources of an installation.
  • the present application refers to the term “consumption” as any type of use (consumption or production) of any type of fluid (whether the fluid is actually consumed or simply discharged, measured or distributed, for example in terms of volume, flow, power, etc.).
  • source of consumption means in this application any type of place or apparatus where machine where takes place “consumption” (or production).
  • the non-limitative term of consumption used in the present application therefore covers consumption and / or production, that is to say that one can identify sources of consumption or sources of production or both ( and at the same time depending on the type of input data).
  • production is understood to mean any type of emission or generation of one or more fluids or set of particles coming from one or more origins mixed in a common flow (for example electricity from a pump heat and photovoltaic panels or others).
  • installation refers to any type of network or location in which a set of “sources of consumption” is installed. It is therefore understandable that all of these terms should not be interpreted in a limiting way. The present description therefore refers to electricity for simplicity and because the invention is particularly suitable for this example, but it is understood that this example is illustrative and not limiting.
  • the present description illustrates the example of electricity which often has an increased complexity compared to other energies, in particular by the diversity of the sources of consumption and production (various types of electrical appliances), but also by the does a given source may have a consumption that varies over time.
  • the devices do not necessarily operate at their nominal power continuously, but may have a consumption that fluctuates over time, according to a programmed cycle or other types of parameters. We therefore distinguish the nominal power of the average power in the present application, in particular for electricity but also for gas for example.
  • the present application sometimes refers to a "nominal consumption” or a “nominal power” (e i ) of the sources of consumption, or sometimes to an "average consumption” or “average power” , but that the term power must not be interpreted in a limiting way since the invention can be applied to other physical quantities.
  • this term of power must be interpreted in the sense of consumption and may in fact designate any value proportional to a rate or even a volume.
  • the maximum filling volume of a tank will correspond to its nominal "power" in terms of consumption, while the maximum gas flow rate used by a boiler will correspond to the nominal power of the boiler (which may have, for a given time interval, an average power different from this nominal power).
  • the present invention aims to identify the various sources of consumption which are responsible for the overall consumption measured over an entire installation (2), using as little as possible (ideally none) to devices ( 20) additional measurement of consumption (often called “sub-meters” in the case of electrical installations).
  • the invention also makes it possible to determine whether at least two different sources can not not to be discriminated and therefore to identify whether at least one additional measuring device (20) is required and where it is to be placed.
  • the invention must therefore be interpreted as also relating to a method and a device for determining the implementation (or deployment) of consumption measurement devices (20) in at least one installation (2).
  • the invention is particularly advantageous in the case of fluids whose cost increases as a function of flow, such as for example electricity (which will be noted that it is measured in power, by watts or kilowatts, but billed in energy, by watts-hour or kilowatt-hours).
  • electricity which will be noted that it is measured in power, by watts or kilowatts, but billed in energy, by watts-hour or kilowatt-hours.
  • the invention may therefore also relate to a method and / or a device for assisting the adjustment of consumption.
  • the figure 2 schematically represents an example of a source identification device (1) which is connected to at least one measuring device (20) ("meter” for example) of at least one installation, or which receives the consumption data recorded by this measuring device (s) (20).
  • the identification device (1) can be located in the installation or remotely, thanks to the known communication techniques of which no detail is necessary here.
  • the figure 2 represents the device (1) for identifying sources of consumption in the form of a computer, which is naturally the first implementation that comes to mind, but it is clear from the present description that this form is only illustrative and by no means limitative. In particular, this device can be implemented in many forms and can even be deployed within the installations (2), in the form of a housing for example, integrated or not in a device (20) for measuring the consumption.
  • the overall consumption of an installation is measured over time, by a meter (20), generally in power (watt or kilowatt) reported over at least a determined time interval (allowing energy billing in kilowatt hours even if the energy has been consumed for a shorter period).
  • the measuring devices (20) record the consumption over time intervals (t x ) of the order of minutes (usually 10 minutes), which is generally sufficient for the invention to determine the distribution of the sources of consumption.
  • modern meters (20) are often able to measure consumption over fairly short time intervals (t x ) (up to a fraction of a second in some cases) and the invention can take advantage of time measurement accurate (less than 10 minutes for example) in some embodiments detailed below.
  • the figure 3B represents an example of consumption within an installation, with intervals (t x ) of time of 10 minutes, over a period (T) of determined time of 24 hours (thus x ranging from 1 to 144).
  • the installation in this example includes 19 sources of consumption (electrical appliances) whose (nominal) powers are known.
  • the present invention is based on an inventory of the sources of consumption present in the installation, to list at least the nominal powers of the sources and their probability of activity (operation for example) during at least one time interval (in general over several intervals to cover an entire period (T) of time).
  • the method may therefore, in some embodiments, include at least one inventory step for recording this information (or even activity profile information as detailed below).
  • the sources of consumption present in the installation with their power and possibly their consumption signatures, ie the curve (or the curves if there are several modes of consumption) are determined. operation or activity profiles) that describes the power of the device over time when operating. Indeed, it is common for a source of consumption (or production) to have a consumption energy signature, such as for example a power spectrum, harmonics or any form of consumption or production curve or histogram. To cite a simple example, some electric convectors operate by power peaks at the nominal value of the device. Frequency and duration peaks can then be representative of the consumption of the convector.
  • a dishwasher has various cycles (short, long, economical, etc.) and the power of the appliance varies over time according to a profile that depends on the type of cycle.
  • the curve that describes this power variation for each of the cycle types thus represents an activity profile and the set of profiles represents a signature of the device, here called “consumption signature" in an illustrative and nonlimiting way (on will use indifferently the terms signatures or profiles to designate the power variation or variations).
  • a convector less simple than that described above can operate in peaks but at variable power values (depending on the setting of the device by the user for example).
  • the figure 3A represents, on the left chart, an example of a signature of a simple convector (with nominal power peaks) and, on the graph to the right, an example of a dishwasher's signatures (a "curve” of power or "Activity profile” of a long cycle and a “curve” of power or “activity profile” of a short cycle).
  • the manner in which the inventory is carried out is not specifically the subject of the present application, but it may still be included and detailed in the list of equipment and appliances that consume the energy ( or energies) studied and their nominal power is listed.
  • the consumer signatures can also be obtained, for example by means of the technical specifications provided by the manufacturers of this equipment and apparatus.
  • these probabilities can also be enriched from objective parameters according to the type of apparatus.
  • convectors for heating have a probability that is correlated with the outside temperature, or even an operating time that is correlated to the size of the room in which they are located.
  • the data representative of these probabilities can therefore be derived from a pre-established table but also from a calculation based on this type of objective information (eg, measured).
  • the method for identifying sources of consumption is implemented by at least one device (1) for identifying sources of consumption.
  • This device (1) comprises data processing means (11) and storage means (10) containing data representative of at least one overall consumption (s) measured in at least one installation (2) on at least one defined time interval (t x ).
  • the method is reiterated for a plurality x of time intervals (t x ) defined to identify the distribution of the sources of consumption during at least one period (T) of determined time, but it is understood that the The invention can be applied to a single time interval if necessary.
  • the invention can be implemented over time intervals of different durations. Indeed, it is possible to choose what is the time interval on which the invention is to be implemented, as long as it has a duration greater than or equal to the time interval provided by the measuring device ( 20) in the installation (2).
  • the storage means (10) therefore stores data representative of the consumption sources present in the installation, as well as their probability of operation over time. Moreover, these storage means (10) can also store data representative of consumption signatures from each of the sources, as indicated above.
  • the storage means (10) also contain the data necessary for the implementation of the method for identifying sources of consumption by the data processing means (11), in particular thanks to at least one optimization algorithm (A). multi-objective.
  • the device stores data enabling the execution, on the data processing means, of at least one application implementing the algorithm (A).
  • the device, the application and the algorithm are designated by the expression "at least one" because it is obvious that the invention can be implemented in a system comprising several devices (as described here or comprising other additional devices) and that the various variants of the invention can be supported by several applications and / or algorithm. It is therefore understood that, in general, the designations used in the present application must be interpreted in a non-limiting manner.
  • the multi-objective optimization algorithm (A) which corresponds to an algorithm of the type often referred to as "genetic algorithms” uses the data stored in the storage means (10) and which are representative of each of the N sources (m i ) of consumption present within the installation (2).
  • these data include, for each of the sources (m i ), at least one information relating to the nominal consumption (e i ) and at least one information relating to the probability (p i ) of consumption during at least one time interval, or even for a plurality of time intervals (t x ) defined within at least one period (T) of time studied.
  • the discrete variables are actually binary variables (v i ) representative of the activity or inactivity of the sources (m i ).
  • each of the discrete variables (v i ) is representative of the percentage of the nominal power at which the source operates over the time interval (t x ) studied.
  • an average power of the source over the time interval (t x ) studied is 50% of the rated power.
  • the consumption of the apparatus is correctly represented by the product (1000 x 0.5) of the nominal power (e i ) of a value of 1000 kW and of a discrete variable (v i ) whose value is 0.5.
  • the sources whose signature shows an activity of a duration less than or equal to the duration of the time slots used it is possible to represent such a source by computing a power average (full power over the duration of the interval) and use the binary variables since the activity of the source is then correctly represented by the product of this average power with the value 0 or 1 of the binary variable.
  • the processing means can therefore assign the value of the average power to the source, replacing the nominal power, so as to take into account that the source has a profile of activity shorter than the duration of the time interval. studied. We then always look for the nominal power (but it corresponds in fact to the average power), with an activation / inactivation which in fact provides binary values to the variables, but while keeping the discrete variables so as not to have to mix the two types variables (for the sake of simplicity even if it is possible to mix them).
  • the algorithm must be able to represent the average consumption of the device by the successive values of this product over 10 successive time intervals (an interval at 1000 x 0.9, then 8 intervals at 1000 x 0.5, then an interval at 1000 x 0.1).
  • a succession of discrete variable values is used to represent and / or retrieve the activity of the source.
  • a given source reaches its nominal power, at least for a moment, during at least one of its activity profiles. It is therefore sufficient to extract the nominal power of the signature.
  • the maximum power reached during a given activity profile is preferably considered to be the nominal power (whether it is actually equal to the nominal power or not the value of this maximum power is assigned to the data representative of the nominal power). So, in some modes of realization, the maximum power peak of a given activity profile, which is more easily detectable than lower powers, can advantageously be used for the integration of the signature in the calculations, as detailed below. This use of the maximum power achieved advantageously reduces the complexity of the calculations (but is not absolutely necessary, even if it is preferred).
  • the discrete variables will generally have decimal values, but it is possible, at least in simple cases, to go down to several decimal places if necessary, even if it is better to stop at the decimal place. because it is clearly sufficient in general and the complexity of the mathematical problem (eg, the number of possibilities) increases rapidly with the number of discrete variables.
  • these discrete variable embodiments are not exclusive of the previous ones using binary variables, in particular because, over a given time interval, one source can operate at its nominal power while another operates at an average power. . For example, if the signature of a source shows that it operates at nominal power but for a shorter duration than that of the time interval studied, its average power must be considered.
  • An instantiation here corresponds to a vector comprising a sequence of N variables, index i, whose value is either 0 or 1 in the case of binary variables, or whose value is between 0 and 1 in the case of variables discrete values taking into account average powers, for example in decimal values (from 0.1 to 0.1).
  • index i whose value is either 0 or 1 in the case of binary variables, or whose value is between 0 and 1 in the case of variables discrete values taking into account average powers, for example in decimal values (from 0.1 to 0.1).
  • the number of binary values (0 or 1) is 2
  • the number of possible instantiations is 2 N and this simplistic solution can not be envisaged.
  • the number of values (0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1) is 11 and the number of possible instantiations is then 11 N.
  • the present invention proposes to enrich the input data with the probabilities (p i ) activation devices (m i ) at time t.
  • the pair (e i , p i ) is specified.
  • This second equation represents the search for an average probability of operation (activity) of the sources (m i ) that is maximal for the solution of the error minimization problem. Indeed, we seek here to maximize the average operating probability of the sources (m i ) identified as minimizing the error between the calculated sum and the total power measured.
  • the invention uses a methodology based on techniques stochastic and in particular so-called genetic algorithms.
  • the invention proposes to solve the problem by means of a variant of conventional genetic type algorithms, to solve this problem of minimization in a robust manner in a reasonable time.
  • Genetic algorithms are indeed known in the field, but various embodiments of the present application advantageously provide to take into account the functional probability of activation of all devices of the facility studied.
  • probabilistic discrete variables to be maximized these embodiments provide a unique solution that has never been considered, even when genetic algorithms have been used for this type of problem.
  • the probability maximization functional (G in the examples detailed in the present application) is advantageously adapted for the use of the discrete variables and makes it possible to take into account the probabilities (p i ) of activation of each of the different sources of the installation.
  • the genetic algorithms used in these advantageous embodiments are clearly distinguishable from those conventionally used since the algorithms used here take into account the activation probability functional and make it possible to represent all the sources as possible instantiations of the discrete variables (integrating the probability of activation), which provides a robust solution that no method allowed to approach so far.
  • the method may be reiterated for a number x of time intervals (t x ) defined to cover said period (T) of determined time.
  • T time intervals
  • the algorithm is reiterated over the 144 time intervals of 10 minutes covering the 24 hours of the survey. It is naturally possible to define the period of time on which the algorithm must be implemented, for example if one wishes to analyze the consumption only over the diurnal period.
  • the identification device (1) has in some embodiments a user interface (12) for such a period selection (T) and / or a plurality of intervals (t x ).
  • the user interface (12) may comprise, in a manner known per se, display and input means, for example with a graphical user interface (GUI).
  • GUI graphical user interface
  • this interface allows a user of the invention (method, device) to enter and / or select at least one parameter relating to the implementation of the algorithm.
  • the number L of iterations is determined in advance and stored in the storage means (10) or selectable using the user interface (12) of the device (1). .
  • the number L of iterations can be selected using the user interface (12) of the device (1), from among a plurality of maximum numbers of iterations stored in the means (10). ) of storage determined according to the number N of sources (m i ).
  • the number L of iterations is determined by the algorithm (A) determining whether the function (F) is descended below a threshold (SL) determined and stored in the means (10) of storage or selectable using the interface (12) user device (1).
  • the number L of iterations is defined by the algorithm (A) determining whether the function (F) has fallen below a threshold (SL) calculated by the processing means (11) as a function of the number N of sources (m i ) and / or if the function (G) exceeds a determined threshold (close to the maximum).
  • the number L of iterations is determined by the algorithm (A) determining whether the best individual or a plurality of best individuals of a current generation has no weaker function (F).
  • the user interface (12) may be used to enter (or select from a prerecorded list) the number m of the best individuals to be selected at each iteration of the selection step.
  • the number m must be small to keep only a very small part of the population, in order to avoid reaching a local optimum. Selection makes it possible to guide the process towards a selection of the best individuals, but it should be avoided that these best individuals are in fact particular cases less optimal than other individuals who have not yet been generated.
  • m can be defined as 3 or equal to a low percentage (less than 5%) of the population. The number m is therefore preferably fixed by the algorithm (A) even if it is possible to leave the choice to the user.
  • the number M of individuals generated is determined as a function of the number N of sources (m i ).
  • the algorithm then provides for choosing a number greater than 7 and preferably less than 2 N (so as not to take all possible instantiations).
  • this number M can be entered using the user interface (12) of the device (1) or selected from among a plurality of values proposed via the user interface (12) of the device (1).
  • the multi-objective optimization algorithm (A) also includes, prior to the updating (55) of the current population and following the crossing (53), a step of mutation (54) of the cross population in which the crossed individuals are modified with a probability (pm) of mutation.
  • this mutation we therefore select; with a given probability (the same for all individuals), some individuals whose one modifies at least one value of the discrete variable of the sources.
  • the modification relates to an inversion of the values of the binary variables (v i ) of the crossed individuals.
  • v i binary variables
  • the mutation probability (pm) is reduced at each iteration so as to limit the mutation as generations progress.
  • the pseudo-random generation step (50) of an initial population of M individuals As explained above, a number M is defined (or selected) so that the initial population comprises M instantiations of the discrete variables (v i ) of each of the sources (m i ). To obtain these instantiations, for each of the N sources (m i ), a uniform random draw (501) of a real number between 0 and 1 is realized then this real number is compared with the value of the probability (p i ) of the source (m i ) in question.
  • the selection step helps guide the process to the best solutions to the optimization problem. We therefore seek to select the best individuals, that is to say those that minimize the error function F. In addition, it is generally sought to maximize the average probabilities of activity of the selected sources.
  • the selection step (51) comprises a selection (511) of the weakest functions (F) among the M individuals and, when the difference between the functions (F) of two individuals is less than at a determined threshold (SF), a selection (512) of the individual whose function (G) is the highest.
  • the census step (52) includes, for each of the sources (m i ), a calculation of a probability inversely proportional to the value of their function (F).
  • a probability inversely proportional to the value of their function (F) will be weighted by the probability (p i ) of activity of the source (m i ), to obtain the census probability (q j ).
  • This census of the population makes it possible to apply an artificial selection to the population for the crossing stage.
  • the generation (53) of a crossed population of Mm crossed individuals corresponds to a cross of individuals taking into account their census probabilities (q j ).
  • the crossed individuals are each obtained by the crossing of a pair of individuals from the census population, selected in a pseudo-random manner taking into account their census probability (q j ).
  • the interval [0,1] is partitioned into segments of lengths q j .
  • each segment occupies a part [ q j ] of the interval [0,1].
  • the m best individuals selected during the selection step (51) of the previous population, can drive the generation of populations to the best possible.
  • the invention provides for detecting errors in the data and alerting the user. Indeed, if the data are erroneous, the invention can lead to solutions (of the best individuals) that are unsatisfactory (below the fixed error threshold). In this case, the invention can provide an alert message, for example indicating the time interval that is problematic. Thanks to the signatures, as detailed below, it is even possible to refine the method, possibly to the point of being able to correct erroneous data. For example if a source has a probability of activity of 0 while it does not make it possible to find a satisfactory optimal individual, it is possible to indicate that this probability is called into question. In addition, if the signature makes it possible to determine that the source actually has a probability of non-zero activity during the time interval studied, it is possible to alert the user to the fact that the probability is erroneous.
  • the invention makes it possible to determine the optimal individuals, that is to say those for which the function (F) is the smallest, but especially for which the function (G) is the highest since it is this function G which is the most discriminating.
  • the optimal individuals that is to say those for which the function (F) is the smallest, but especially for which the function (G) is the highest since it is this function G which is the most discriminating.
  • the invention provides for improving the method by using additional data, for example from the inventory, as defined in the present application.
  • the installation has 3 convectors of identical power (and identical signature), it is very likely that they can not be discriminated even by correlating their operation at the outside temperature since they will have the same probabilities of operation. .
  • the information sufficiently influences the operating time of the convectors and that the invention is put into operation (or put back into effect) over time intervals of relevant duration to discriminate the operating times of these convectors.
  • the invention also provides for the classification of sources by grouping them by categories.
  • the invention can then provide to determine the source category that has worked rather than the source itself.
  • the example of the above 3 convectors is realistic and can be used, it is possible not to have to implement it for this type of example and to reserve this type of optimization of the invention.
  • the signatures to devices that it is really desirable to discriminate, for example because they do not belong to the same category of sources. Nevertheless, the method and the device, in some embodiments, make it possible to refine the results of the identification and / or to accelerate their obtaining.
  • At least one algorithm comprises a search (56), in the storage means, of data representative of the consumption signature of the source, that is to say at least one profile (PL).
  • at least one source (m i ) has an activity profile that indicates that it operates in general at least for a specified duration, which is greater than the duration of the time interval over which the process has been performed.
  • At least one algorithm comprises a search (58), in the storage means, of data representative of at least one profile (PC) of the activity of the sources (m i ) over a period of time less than that of said defined time interval (s) (t x ), then a comparison (59) of the measured consumption (s) during this period with the activity profiles, to select the best individual (s) whose discrete source variables (m i ) are compatible with these activity profiles.
  • the algorithm in this case can even search for a kind of "consumer signature". For example, some devices can operate by successive power peaks relatively short compared to the time interval (no time) on which the identification is performed. It is therefore useful to look in the consumption data for fluctuations that correspond to such peaks.
  • the signatures and in particular at least one activity profile that these signatures may contain for example, the 2 curves of the graph of the right of the figure 3A correspond to 2 different activity profiles of a dishwasher.
  • the activity profiles as soon as a source (m i ) present in an optimal individual (among the best m for example) is identified as listed in the library. of signatures of the storage means (10).
  • the search (56, 58) and comparison (57, 59) steps can be performed as soon as the iterations of the algorithm (A) are completed for a given time interval (t x ). In some embodiments, it may even be envisaged that these steps are implemented during the iterations of the algorithm (A) in order to converge the computations more rapidly towards the optimum individuals whose signatures do not invalidate their relevance.
  • this source will be privileged over the others and we will retain in the best selected for the following intervals, only those which also contain this active source (at the average power corresponding to its signature, preferably).
  • the invention provides for determining, when implemented on a succession of several intervals time (t x ), to determine what are the so-called confidence intervals, in which the best individuals differ by a G function value which is the highest (for example, the 2 best have G functions that differ a value greater than a threshold).
  • a G function value which is the highest
  • the invention provides for determining, when implemented on a succession of several intervals time (t x ), to determine what are the so-called confidence intervals, in which the best individuals differ by a G function value which is the highest (for example, the 2 best have G functions that differ a value greater than a threshold).
  • This type of improvement can be used "on the fly", that is, when going from one interval to another, or in "post-processing", that is, after the time intervals of interest have been processed (all or at least a predetermined number, for example selected via the interface).
  • the invention can provide that the signatures are used when a source of a better individual is identified in the signature data during the processing of an interval, to integrate the information of this signature into the treatment. following intervals ("on the fly” treatment).
  • the invention can provide a correction loop returning in the previous time intervals to improve the results with the signature data.
  • the use of signatures as described above may advantageously use the maximum power reached in the signature or the activity profile of interest in the signature.
  • the signature or the profile of The activity is then used to integrate the other power values of this identified source over the previous and / or subsequent time intervals.
  • the invention also makes it possible to assist in the decision on the deployment of additional sensors or measuring devices in the installation.
  • additional sensors or measuring devices in the installation.
  • the invention makes it possible to identify this utility as well as the place where this additional measuring device is to be installed.
  • the device (1) for identifying sources of consumption comprises data processing means (11) and storage means (10) containing data representative of at least one overall consumption (s). ) measured in at least one installation (2) over predefined time intervals (t x ) and during at least one period (T) of predetermined time.
  • the device can store the data in its storage means (a hard disk for example) but that it can also access it via a network (for example intranet or internet) and that this notion should not be interpreted. in a limiting way. Indeed, the same device can also access the total consumption data measured / read by several measuring devices, in order to centralize the data processing of several measuring devices (of the same installation or of several different installations).
  • the storage means (10) contain the data necessary for the data processing means (11) to implement the method for identifying consumption sources according to various embodiments of the invention.
  • the device may also include at least one user interface (12) for interaction with the user. It is understood, however, that the processing means (11) are in fact only defined by the fact that they implement the method and the data necessary for their execution can be physically stored elsewhere than in the storage means (10) which store the data exploited by the process.
  • the notions of processing means are defined in a functional manner and it is clear that structurally or physically, they must not be interpreted in a limiting manner (the consumption survey data can be in the measuring device while the data of the algorithm are elsewhere, or even directly loaded into the processor or the RAMs that execute it, while the data relating to sources, including probabilities and signatures, may still be in another device).
  • the term device (1) actually covers also a "system” comprising several different devices that cooperate with each other to form the functional assembly as defined herein.
  • the method and the device obtained according to the various embodiments detailed in the present application make it possible to identify robust and fast consumption sources. It is even possible to perform the identification as soon as a time interval (t x ) of measurement of the total consumption has been completed (for at least one following interval, depending on the duration of the intervals).
  • the device may also include warning means to indicate that the consumption exceeds a threshold, for example beyond which billing is greater.
  • the skilled person may, especially in a complex installation (such as a factory), provide that this type of detection, or the implementation of the present invention in general, allows to control the operation of different devices in the installation, so that the consumption is better distributed (by delaying the activation of a device when the other devices already consume too much and are identified as necessary to the activity of the factory during the interval of studied time).
  • the result of the identification according to the present invention is particularly advantageous in the field of energy.
  • it can be represented in various ways, and especially as for example on the figure 3C but the latter is not limiting.
  • the survey gives a succession of measurements for each of the 144 time intervals (t x ) of 10 minutes, during a period (T) of 24 hours but that one n here is the total power.
  • the invention may also relate to a method and / or a device for helping the adjustment of consumption.
  • the invention provides for presenting the consumption peaks and the sources that are responsible for it, which enables (as the case may be) the users to find the source of the additional costs of consumption.
  • the invention can also provide time ranges of use of sources of consumption to better distribute consumption over time and avoid peaks.

Landscapes

  • Business, Economics & Management (AREA)
  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Primary Health Care (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
EP14162908.9A 2013-03-29 2014-03-31 Verfahren und Vorrichtung zur Identifizierung der Verbrauchs- und/oder Erzeugungsquelle Ceased EP2784742A1 (de)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
FR1352930A FR3003981B1 (fr) 2013-03-29 2013-03-29 Procede et dispositif d’identification de sources de consommation et/ou de production

Publications (1)

Publication Number Publication Date
EP2784742A1 true EP2784742A1 (de) 2014-10-01

Family

ID=49054660

Family Applications (1)

Application Number Title Priority Date Filing Date
EP14162908.9A Ceased EP2784742A1 (de) 2013-03-29 2014-03-31 Verfahren und Vorrichtung zur Identifizierung der Verbrauchs- und/oder Erzeugungsquelle

Country Status (2)

Country Link
EP (1) EP2784742A1 (de)
FR (1) FR3003981B1 (de)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104616078A (zh) * 2015-02-03 2015-05-13 河海大学 基于Spiking神经网络的光伏系统发电功率预测方法
CN105335791A (zh) * 2014-08-07 2016-02-17 中国南方电网有限责任公司 一种交直流大电网协调优化自动控制方法及系统
CN105574325A (zh) * 2015-12-10 2016-05-11 华南理工大学 一种结合人口指标的中长期用电量预测方法

Non-Patent Citations (7)

* Cited by examiner, † Cited by third party
Title
AHMED ZOHA ET AL: "Non-Intrusive Load Monitoring Approaches for Disaggregated Energy Sensing: A Survey", SENSORS, vol. 12, no. 12, 6 December 2012 (2012-12-06), pages 16838 - 16866, XP055109922, ISSN: 1424-8220, DOI: 10.3390/s121216838 *
BARANSKI M ET AL: "Genetic algorithm for pattern detection in NIALM systems", SYSTEMS, MAN AND CYBERNETICS, 2004 IEEE INTERNATIONAL CONFERENCE ON, IEEE, PISCATAWAY, NJ, USA, vol. 4, 10 October 2004 (2004-10-10), pages 3462 - 3468, XP010773294, ISBN: 978-0-7803-8566-5 *
MICHAEL ZEIFMAN ET AL: "Viterbi algorithm with sparse transitions (VAST) for nonintrusive load monitoring", COMPUTATIONAL INTELLIGENCE APPLICATIONS IN SMART GRID (CIASG), 2011 IEEE SYMPOSIUM ON, IEEE, 11 April 2011 (2011-04-11), pages 1 - 8, XP031894550, ISBN: 978-1-4244-9893-2, DOI: 10.1109/CIASG.2011.5953328 *
MING DONG ET AL: "An Event Window Based Load Monitoring Technique for Smart Meters", IEEE TRANSACTIONS ON SMART GRID, IEEE, USA, vol. 3, no. 2, 1 June 2012 (2012-06-01), pages 787 - 796, XP011445358, ISSN: 1949-3053, DOI: 10.1109/TSG.2012.2185522 *
POWERS J T ET AL: "Using a rule-based algorithm to disaggregate end-use load profiles from premise-level data", IEEE COMPUTER APPLICATIONS IN POWER, IEEE INC. NEW YORK, US, vol. 4, no. 2, 1 April 1991 (1991-04-01), pages 42 - 47, XP011419355, ISSN: 0895-0156, DOI: 10.1109/67.75875 *
TSAI M S ET AL: "Development of a non-intrusive monitoring technique for appliance' identification in electricity energy management", ADVANCED POWER SYSTEM AUTOMATION AND PROTECTION (APAP), 2011 INTERNATIONAL CONFERENCE ON, IEEE, 16 October 2011 (2011-10-16), pages 108 - 113, XP032162334, ISBN: 978-1-4244-9622-8, DOI: 10.1109/APAP.2011.6180394 *
WILSUN XU ET AL: "Tracking energy consumptions of home appliances using electrical signature data", POWER AND ENERGY SOCIETY GENERAL MEETING, 2012 IEEE, IEEE, 22 July 2012 (2012-07-22), pages 1 - 5, XP032465276, ISBN: 978-1-4673-2727-5, DOI: 10.1109/PESGM.2012.6344624 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105335791A (zh) * 2014-08-07 2016-02-17 中国南方电网有限责任公司 一种交直流大电网协调优化自动控制方法及系统
CN104616078A (zh) * 2015-02-03 2015-05-13 河海大学 基于Spiking神经网络的光伏系统发电功率预测方法
CN104616078B (zh) * 2015-02-03 2017-12-22 河海大学 基于Spiking神经网络的光伏系统发电功率预测方法
CN105574325A (zh) * 2015-12-10 2016-05-11 华南理工大学 一种结合人口指标的中长期用电量预测方法
CN105574325B (zh) * 2015-12-10 2018-06-22 华南理工大学 一种结合人口指标的中长期用电量预测方法

Also Published As

Publication number Publication date
FR3003981B1 (fr) 2015-04-03
FR3003981A1 (fr) 2014-10-03

Similar Documents

Publication Publication Date Title
US10728083B2 (en) Methods for orientation and tilt identification of photovoltaic systems and solar irradiance sensors
EP2000780B1 (de) Verfahren und System zur Erfassung und Abschätzung des Stromverbrauchs von elektrischen Anlagen eines Teilnehmers
FR2941328A1 (fr) Procede de prevision de la production electrique d'un dispositif photovoltaique
EP2460057A1 (de) Energieverwaltung in einem gebäude
EP2784742A1 (de) Verfahren und Vorrichtung zur Identifizierung der Verbrauchs- und/oder Erzeugungsquelle
FR3045898A1 (fr) Methode de prevision de la consommation d'eau chaude, methode et systeme d'optimisation de la production d'eau chaude
EP2976612B1 (de) Verfahren zur herstellung einer wärmediagnose eines gebäudes oder eines gebäudeteils
EP3648020B1 (de) Absicherung eines eingriffs in einem niedrigspannungszweig eines stromversorgungsnetzes
JP7352385B2 (ja) 電力管理システム、電力管理方法及び電力管理プログラム
EP3847520B1 (de) Verfahren und system zur steuerung und regelung eines heizsystems für gebäude
EP3291033B1 (de) System und verfahren zum abschätzen des thermischen verhaltens eines gebäudes zur optimalen heizungssteuerung
FR3028114A1 (fr) Procede de diagnostic d'un systeme photovoltaique
FR3131988A1 (fr) Prévision bayésienne de consommation individuelle et équilibrage d'un réseau électrique
WO2022008552A1 (fr) Procédé de détection de défaillance d'un onduleur photovoltaïque
FR2979451A1 (fr) Procede de determination de la part de la consommation electrique liee aux temperatures exterieures
FR3081570A1 (fr) Procede d’emission de commandes de desactivation d’appareils lors d‘un pic de consommation, serveur et programme d’ordinateur correspondants.
FR3103604A1 (fr) Procédé de prédiction de consommations d’une ressource distribuée dans un réseau de distribution
FR3072192A1 (fr) Estimation d'une consommation electrique d'une piscine d'agrement
WO2015092279A1 (fr) Procédé de localisation d'une anomalie détectée dans une structure arborescente
EP3045999A1 (de) Verbesserte thermische vorrichtung
FR2982031A1 (fr) Procede et systeme d'estimation de la consommation electrique d'un dispositif de chauffage electrique
FR2975521A1 (fr) Procede de prevision d'une quantite d'energie electrique produite par un equipement photovoltaique et systeme de production d'energie electrique associe
WO2024061868A1 (fr) Procede de surveillance d'une consommation fluidique d'un batiment a surveiller, par classification par apprentissage supervise d'evenements de consommation fluidique
WO2016188958A1 (fr) Procede d'identification d'une structure de donnees representative d'une consommation de fluide, dispositif et programme d'ordinateur correspondant
FR3019286A1 (fr) Procede et installation de comparaison de consommation d'effluents sans divulgation de donnees de consommations mesurees

Legal Events

Date Code Title Description
17P Request for examination filed

Effective date: 20140331

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

AX Request for extension of the european patent

Extension state: BA ME

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

R17P Request for examination filed (corrected)

Effective date: 20150331

RBV Designated contracting states (corrected)

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

17Q First examination report despatched

Effective date: 20170322

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION HAS BEEN REFUSED

18R Application refused

Effective date: 20190307