• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    The distribution network is equipped with sensors. The steps are: - acquire for each sensor a time series of physical measurements separated by time intervals, - define time windows each corresponding to several successive time intervals, - extract operational characteristics of each time series in each time window for each time window, forming at least one current vector having, for coordinates, operational characteristics, structural data relating to the network and cyclical characteristics specific to the time window; comparing the current vector with previous vectors, corresponding to time windows previous, and whose cyclical characteristics and structural data are similar to those of the current vector, and - report an anomaly in cases where the current vector is significantly dissimilar to said previous vectors.
    公开号:FR3024260A1
    申请号:FR1457209
    申请日:2014-07-25
    公开日:2016-01-29
    发明作者:Francis Campan；Abel Dembele；Guillaume Cussonneau
    申请人:Suez Environnement SA；
    IPC主号:
    专利说明:

    [0001] FIELD OF THE INVENTION The present invention relates to a method for detecting anomalies in a distribution network, in particular distribution of a Newtonian fluid, and more particularly to a method for detecting abnormalities in a distribution network, in particular distribution of a Newtonian fluid, even more particularly a drinking water supply network.
    [0002] Operational performance is at the heart of the management of drinking water supply systems. The level of performance can be significantly improved by means of tools for detecting and locating hydraulic anomalies on the transport and drinking water distribution networks. The invention can be used for all Newtonian fluid networks, for example urban heating and cooling networks.
    [0003] For the detection of anomalies, anomalies detection methods based on statistical techniques are known, such as the analysis of the minimum flow of a sector of the network, widely used by the operators of drinking water networks. These methods do not always make it possible to characterize or locate the anomaly.
    [0004] There are also known detection methods based on hydraulic models. These methods based on hydraulic equations use a network modeling, which is not always available. Detection and anomaly localization combine mathematical tools and physical measurements, which implies the establishment of an adequate instrumentation density. The object of the invention is to provide an anomaly detection method that is efficient while being low in input data and relatively economical in terms of computing power.
    [0005] According to the invention, the method for detecting anomalies in a distribution network, in particular distribution of a Newtonian fluid, more particularly a drinking water supply network, the distribution network being equipped with sensors, a method of in which one acquires for each sensor a time series of physical measurements separated by time intervals, is characterized by the following steps: - defining time windows each corresponding to several time intervals, - extracting operational characteristics from each time series 10 in each time window, - forming at least one current vector for each time window whose coordinates are the operational characteristics, structural data relating to the network and the cyclical characteristics peculiar to the time window, 15 - comparing the current vector with vectors precedes nts, corresponding to previous time windows, and whose cyclical characteristics and structural data are similar to those of the current vector, - report an anomaly in cases where the current vector is significantly dissimilar to said previous vectors. For the comparison of the current vector with the preceding vectors, a sensitivity parameter corresponding to a minimum degree of dissimilarity outside of which the current vector is described as significantly dissimilar is defined.
    [0006] The time interval between two measurements of a time series is, for example, a few minutes, for example 3 minutes. A time window is typically a day or a week.
    [0007] Thus a time series typically includes many measurements. The method can advantageously be implemented simultaneously for time windows of different durations. During a time window, the parameters measured by the sensors, and constituting the time series, undergo variations, for example as a function of cyclical parameters such as the time, the day of the week, the season, the weather, etc. ., network structural data, such as closing a valve, stopping a pump, and also depending on one or more anomalies on the network.
    [0008] According to the invention, a vector having dimensions representative of the time series, representative dimensions of the conjunctural parameters, and representative dimensions of the structural data is formed.
    [0009] The idea underlying the invention is that if the representative dimensions of the cyclical parameters and the structural data have been appropriately chosen, the vectors having substantially the same coordinates for the cyclical and structural dimensions should also have substantially the same coordinates for the operational dimensions. Otherwise, an anomaly is reported. Given the typically large number of numerical values in each time series, the method should process vectors having a very large number of dimensions if all these numerical values each become a coordinate of a respective dimension of the vector. Thus, another idea underlying the invention is to extract characteristics of each time series, and then to use these characteristics for vector coordinates.
    [0010] The characteristics extracted from the time series may comprise maximums, minima and / or means, and / or even elementary frequencies, obtained in particular by decomposition of the time series in Fourier series. Certain cyclical parameters or structural data, for example the ambient temperature or the state of a valve, may also be available as a time series which is processed to extract characteristics which will constitute coordinates of the vector in place and place measurements or raw data.
    [0011] For the extraction of characteristics, it is even possible to aggregate time series of measurements. For example, it is possible to add the known consumptions according to the remote reading, and thus to obtain a total single measurement, or else a time series of the total consumptions, for example a measurement per day in the case of a time window of one week.
    [0012] Advantageously, certain conjunctural characteristics are at least partly based on hypotheses derived from experience. For example, trends in water use by an individual may be based on time, day of the week, season, outdoor temperature, rainfall, number of occupants, and housing etc. In an advantageous embodiment, in the event of an anomaly, the current vector is compared with at least one preceding vector having cyclical characteristics and similar structural data, and operational characteristics as close as possible to the current vector, and at least one operational characteristic is indicated for which the current vector has a large deviation from said at least one preceding vector.
    [0013] Preferably, abnormal characterization software deals with situations that have resulted in the signaling of at least one anomaly. Such software works much more efficiently when it only deals with situations sorted as abnormal with further information already available on the parameters (or vector coordinates) having an abnormal value. The result, a vector anomaly or network anomaly determined after processing by characterization software as indicated above, is preferably provided by reference to a gravity scale of the anomaly. Having quantified the severity of the anomaly, in case of several simultaneous anomalies, one can prioritize the anomalies relative to each other according to their urgency and / or according to the magnitude of the corresponding corrective intervention.
    [0014] In a preferred embodiment, a non-anomalous vector is classified in the same memory compartment as other substantially equipolent vectors, and each time a current vector has been calculated the compartment of the compartment is searched for. memory containing the previous vectors which resemble it the most, and the current vector is only compared with the previous vectors of this compartment. This reduces the processing power required and the time required to obtain the results. A new memory compartment is created for a vector whose conjunctural coordinates and / or coordinates corresponding to structural data of the network do not correspond to any existing compartment. For example, an exceptionally cool meteorology for a summer month can lead to the creation of a new compartment.
    [0015] It is advantageous to analyze the evolution over time of the vectors of the same compartment, and to provide information on the evolution of the network. It is thus possible, for example, to reveal leaks initially weak but tending to become worse and which may become important, whereas the simple search for the most similar vector in the compartment will not reveal the increasing leak, since the vector preceding the more recent will appear very similar and lead to the conclusion that there is no anomaly. In an even more sophisticated version, the changes in the different compartments are compared, and information is provided which differentiates the evolution of the state of the network and the evolution of the consequences of the cyclical parameters on the network. For example, a consumption that only increases on hot days indicates a change in consumer habits rather than an increasingly leaky network.
    [0016] In order to simplify the processing and to refine the results, it is very advantageous to provide for each time window several vectors each corresponding to a respective subnetwork forming part of the network. Thus, smaller vectors are treated and additional chances of locating an abnormality are more readily available.
    [0017] In the event of detection of an anomaly, the current vector is compared with comparatively abnormal recent vectors, to provide indications of a speed of evolution of the anomaly and / or a link of the anomaly with at least one anomaly. cyclical parameter and / or structural data. In the event of an anomaly, it is also possible to search in the preceding vectors for comparable anomalies that have given rise to a diagnosis, to provide a pre-diagnosis of the cause of the current anomaly.
    [0018] During an initialization step, it is advantageous according to the invention to load a memory with vectors reconstituted from archives relating to the network.
    [0019] The components of a vector preferably include at least one component relating to the complaints of the consumers served by the network, for example on the flow, the pressure, the taste, etc. Other features and advantages of the invention will emerge again. of the description below, relating to non-limiting examples, and with reference to Figure 1 which shows a flowchart of the main steps in the example described. The following description is a description of any particularity it contains, whether it is taken separately from the other features, even if they are part of the same paragraph or the same sentence, and is a description of any combination of such features. that such particularity or combination of features is distinctive of the state of the art and offers a technical effect, whether alone or in combination with the concepts presented above, and that this feature is expressed in the very terms of this description or in more or less generalized terms.
    [0020] 3024260 7 Definitions Entity: Potable water network or component of the drinking water supply system, eg hydraulic sectors or measuring devices / sensors. An entity is associated with one or more time series. Time series: finite sequence of scalar data indexed by time, usually spaced by a constant duration. Classification: Without further specification, refers to the process of assigning to a state of an entity a known class giving rise to predetermined actions on the part of the network operator. Clustering: Without further specification, refers to the process of associating a group of previous states with a state of an entity to determine the abnormal or new character.
    [0021] Measurement: This is the estimation of the value of a corresponding unit quantity, these values forming a time series provided by a data acquisition system. A measurement is associated with a particular elemental component of the drinking water system (arc or section for flow measurement, node for pressure measurement, reservoir for level measurement). Remote reading data: series of consumption indices for a meter, measured at a given periodicity, and TV transmitted for example at least once a day.
    [0022] Characteristic: ("feature") scalar or vector constituting information significant for the process studied, here the state of the network over a given time window. Principles In this embodiment the invention implements the following principles: Pretreatment of the signal by the methods of the state of the art to complete and clean the noise. The production of a feature vector of the state of operation of an entity for a given time window. This is done by aggregation: o characteristics extracted mainly from time series of the entity by methods of decomposition / signal processing, o characteristics from business performance indicators and 5 business data. - The use of classification and "clustering" algorithms, derived from machine learning techniques, applied to feature vectors of entities for a given time window. This then allows the transcription of the states of the entities (as represented by the characteristic vectors) into categories known to the network operators, in order to classify the situations and to prioritize the corrective actions to be carried out. The classification / clustering algorithms are based on objective functions and network management operational constraints, for example the number available for a simple maintenance operation, the delay to have a team for a more complex operation. , the delay between the date of occurrence of an anomaly and the date of its detection, etc. - The optional use of context data influencing the functioning of the entity, to specify the description of its state. 20 - The use of detection / scoring algorithms for anomalies resulting from machine learning techniques, calibrated / optimized using criteria that meet the operational constraints of the machine. network operators. These algorithms are applied to feature vectors.
    [0023] 25 Input Data and Parameterization: Three Data Sets: The structural data consists of descriptive data of the network infrastructure and installed equipment (valves, sensors, pumps, etc.).
    [0024] The implementation parameters of the method are calibrated automatically in a preparatory phase, itself automated to allow a registration when the system detects a loss of performance or an evolution of its infrastructure. The operational data are from all measurement systems present on the network. When available, this data will include, among other things, also the telecom consumption data collected, customer complaints and the interventions having an impact on the network behavior. Basic methods (See Figure 1) Pretreatment of the time series of the measurements These methods make it possible to obtain time series of the possibly completed measurements, smoothed / cleaned of the noise. The resulting time series are then ready to be used as input data of feature extraction algorithms. Combinations of the series into aggregated signals are also performed. For example, the algebraic sum of time series of input / output flows of a hydraulic sector is converted into a series of consumption of the hydraulic sector. The series can also be transformed (for example centered reduced) for the purposes of some of the algorithms used in the feature extraction phases.
    [0025] Extraction of Characteristics Resulting from Signal Decomposition This method consists in using the previously cleaned / smoothed time series as described above, in order to extract the relevant information making it possible to characterize in an operational manner the state of an entity. It is a question of producing the information summarizing the series structure (Fourier decomposition, in wavelets, principal components, ...) while decreasing the dimension in order to concentrate the relevant part of the signals. The different bases of decomposition are evaluated periodically in order to monitor the performance of the algorithms and to update the bases when they are judged to be too little parsimonious (supervision). The outputs of these algorithms are therefore characteristic vectors summarizing each signal in a time window defined by the network operator.
    [0026] Extraction of the business characteristics The construction of the business characteristics of a site is automated and is based on: a reference system of characteristics established from the experience of networks of different types, an observation of the structure and the behavior of the targeted network. The construction can be based, for example, on a calculation of the minimums of a parameter over a given time window, the observation of the periodicity of the filling / emptying cycles of the reservoirs, the average consumption level of the remote meter readings in function of their consumer segment, etc.
    [0027] Algorithms for Evaluating the State of Operation of the Entity - Operating Regime According to an important feature of the invention, to describe the operating state of the network, data from the business expertise is combined with those resulting from the methods of decomposition of the signal. These come mutually enriching by giving for each entity a vector completely characterizing a state of operation of this entity for the considered network.
    [0028] With the aid of the feature vectors, the operating state of the considered entity at a given time in a given time window, or "regime", can be characterized. It can then be compared with the previous states, or the states of the other entities, and classified according to operational criteria involving actions to be taken. To achieve this ranking, the 30 machine learning tools such as classification and clustering can be implemented. In the case of classification, a learning process will have previously made it possible, by means of historical data marked with the various possible states and constituting groups, to cause a discriminating function ("classifier") which Allows you to automatically give the group to which the new state belongs. This discriminant function can be obtained by optimizing a quality criterion, for example an inertia function in unsupervised mode or the proxy for a classification error in supervised mode. In the case of a lack of marked historical data, clustering techniques make it possible to group the states 10 according to similarity criteria and thus to discriminate those belonging to the least represented categories. These are of interest to the network operator since they indicate behavior that is out of the ordinary. The operator can then focus his attention on this entity. Algorithms for Detecting Anomalies The state characterization vectors are applied anomaly detection algorithms. It is about characterizing an event in progress or recently completed. The type of event corresponds to a category of events monitored by the operators of drinking water networks (leaks, pressure drop, sensor failure, consumption anomaly, etc.).
    [0029] The detection algorithms are started in parallel and their results are aggregated for the discrimination of the entities exhibiting abnormal behavior. They were previously calibrated on historical datasets, in order to adapt their settings to the operational constraints of the network.
    [0030] 30 Corollary particularities The assembly constituted by the components exposed above can be connected to the technical information system of the operator of a drinking water supply system. Each available data source is then connected to the feature extraction algorithm dedicated thereto. The set can be activated regularly, according to the acquisition frequency of the 3024260 12 data. The time window on which the analysis is done is adjustable by the user. It is nevertheless relevant to use first time 24-hour and 7-day time windows.
    [0031] Under these conditions the characterization of the operating state of the network and the detection of associated anomalies are much more relevant than with the methods usually used, and the operational efficiency is considerably improved.
    [0032] In the diagnostic phase of the past events of a network, for purposes of operating balance, for example, the method makes it possible to save considerable time by discriminating against past events. The combination of signal related characteristics and business features improves the robustness of the anomaly detection. The adjustment of the sensitivity of the algorithm, that is to say the balance between the number of correct detections and the number of anomalies for each entity, allows, for example, to adapt the detection of each type of detection. anomaly in the ability of the operator to plan and initiate corrective actions. Embodiment example: Chlorine example and signal decomposition The use of a signal decomposition method, such as wavelet or Fourier decomposition, on signals relating to a drinking water network makes it possible to isolate the different components (intra and inter day) of these signals. An algorithm for the cognitive analysis of these components, in relation to the business expertise, leads to identify the domain of definition of the normal functioning of the network. This allows, when new signals are available, to detect significant changes in the nature of these components. Thus, a significant difference observed can be interpreted as an indicator of abnormal behavior. A second set of search algorithms can then be executed. For example, a classification algorithm based on feature vectors (including customer complaints for example) can help define the level of risk achieved. This approach can be applied for water quality monitoring using all the available measurement points on a drinking water network to identify hidden structures and detect, for example, anomalies on the water. residual concentration of chlorine in conjunction or not with other quality parameters. The customer complaints used are then related to the taste of the water. 10
    权利要求:
    Claims (20)
    [0001]
    REVENDICATIONS1. Method for detecting anomalies in a distribution network, in particular distribution of a Newtonian fluid, more particularly a drinking water supply network, the distribution network being equipped with sensors, in which process for each sensor is acquired a time series of physical measurements separated by time intervals, characterized by the following steps: - defining time windows each corresponding to several time intervals, - extracting operational characteristics from each time series in each time window, - forming for each temporal window at least one current vector whose coordinates are the operational characteristics, structural data relating to the network and the cyclical characteristics specific to the time window, - comparing the current vector with previous vectors, corresponding to time windows previous, and whose cyclical characteristics and structural data are similar to those of the current vector, and - report an anomaly in cases where the current vector is significantly dissimilar to said previous vectors.
    [0002]
    2. Method according to claim 1, characterized in that for the comparison of the current vector with the preceding vectors, a sensitivity defining a minimum degree of dissimilarity is set so that the current vector is qualified as significantly dissimilar.
    [0003]
    3. Method according to claim 1 or 2, characterized in that for the extraction of characteristics, time series of measurements are aggregated.
    [0004]
    4. Method according to one of claims 1 to 3, characterized in that the cyclical characteristics are at least partly based on assumptions from experience.
    [0005]
    5. Method according to one of claims 1 to 4, characterized in that in case of anomaly, the current vector is compared with at least one previous vector 3024260 15 having similar cyclical characteristics and operational characteristics as close as possible current vector, and signal at least one operational characteristic for which the current vector has a large deviation with said at least one previous vector.
    [0006]
    6. Method according to one of claims 1 to 5, characterized in that the result is provided by reference to a scale of gravity of the anomaly.
    [0007]
    7. Method according to one of claims 1 to 6, characterized in that in case of several simultaneous anomalies is prioritized the abnormalities according to their urgency and / or depending on the magnitude of the corresponding corrective action.
    [0008]
    8. Method according to one of claims 1 to 7, characterized in that a vector free of anomalies is classified in the same memory compartment as other substantially equipollent vectors, and each time a current vector has It is calculated that the memory compartment containing the most similar anterior vectors is searched for, and the current vector is compared only with the previous vectors of this compartment.
    [0009]
    9. The method as claimed in claim 8, characterized in that a new memory compartment is created for a vector whose cyclical coordinates and / or coordinates corresponding to structural data of the network do not correspond to any existing compartment.
    [0010]
    10. Method according to one of claims 1 to 9, characterized in that one analyzes the evolution over time of the vectors of the same compartment, and provides information on the evolution of the network.
    [0011]
    11. Method according to claim 10, characterized in that the changes in the different compartments are compared, and information is provided which differentiates the evolution of the state of the network and the evolution of the consequences of the cyclical parameters on the network. . 30
    [0012]
    12. Method according to one of claims 1 to 11, characterized in that provides for each time window several vectors each corresponding to a respective sub-network network part.
    [0013]
    13. Method according to one of claims 1 to 12, characterized in that the characteristics extracted from the time series comprise maxima, minima and / or averages. 3024260 16
    [0014]
    14. Method according to one of claims 1 to 13, characterized in that the characteristics extracted from the time series comprise elementary frequencies, obtained in particular by decomposition of the time series in Fourier series. 5
    [0015]
    15. Method according to one of claims 1 to 14, characterized in that in case of detections of an anomaly, the current vector is compared with the preceding vectors to provide indications on a speed of evolution of the anomaly. and / or a link of the anomaly with at least one conjunctural parameter and / or a structural datum. 10
    [0016]
    16. Method according to one of claims 1 to 15, characterized in that during an initialization step a memory is loaded with vectors reconstructed from archives relating to the network.
    [0017]
    17. Method according to one of claims 1 to 16, characterized in that it implements it simultaneously for time windows of different durations.
    [0018]
    18. Method according to one of claims 1 to 17, characterized in that the components of a vector include complaints from consumers served by the network.
    [0019]
    19. Method according to one of claims 1 to 18, characterized in that the situations giving rise to the signaling of at least one anomaly are processed by an abnormality characterization software.
    [0020]
    20. Method according to one of claims 1 to 19, characterized in that in the event of an anomaly, it is sought in previous vectors of comparable anomalies that have led to a diagnosis, to provide for the cause of the anomaly. a pre-diagnosis similar to the previous diagnosis.
    类似技术:
    公开号 | 公开日 | 专利标题
    EP3172548B1|2018-09-26|Method for detecting anomalies in a distribution network, in particular for drinking water
    EP2696344B1|2019-01-02|Method and system for detecting acoustic events in a given environment
    CA2746537C|2017-05-09|Standardization of data used for monitoring an aircraft engine
    CN107949812A|2018-04-20|For detecting the abnormal combined method in water distribution system
    Bach‐Andersen et al.2018|Deep learning for automated drivetrain fault detection
    EP2409167B1|2017-01-25|Method and device for filtering electrical consumption curves and allocating consumption to categories of appliances
    CN106650797B|2020-12-04|Power distribution network electricity stealing suspicion user intelligent identification method based on integrated ELM
    FR2957097A1|2011-09-09|SYSTEM AND METHOD FOR MONITORING RESOURCES IN A WATER DISTRIBUTION NETWORK
    CN110113226B|2021-03-12|Method and device for detecting equipment abnormity
    CA2888716A1|2014-05-01|System for monitoring a set of components of a device
    Phan2018|Housing price prediction using machine learning algorithms: The case of Melbourne city, Australia
    FR2940694A1|2010-07-02|METHOD AND SYSTEM FOR CLASSIFYING DATA FROM DATABASE.
    FR3082963A1|2019-12-27|SYSTEM AND METHOD FOR EVALUATING AND DEPLOYING NON-SUPERVISED OR SEMI-SUPERVISED AUTOMATIC LEARNING MODELS
    FR3032786A1|2016-08-19|DATA PROCESSING AND MODELING SYSTEM FOR ANALYZING THE ENERGY CONSUMPTION OF A SITE
    WO2018020181A1|2018-02-01|Method for determining a maximum allowable volume of water that can be removed over time from an underground water source
    EP3489831A1|2019-05-29|Method and device for monitoring a data generator process of a metric for the prediction of abnormalities
    CN110362612B|2022-02-22|Abnormal data detection method and device executed by electronic equipment and electronic equipment
    FR3035232A1|2016-10-21|SYSTEM FOR MONITORING THE HEALTH CONDITION OF AN ENGINE AND ASSOCIATED CONFIGURATION METHOD
    EP3871078A1|2021-09-01|Computing environment system for monitoring aircraft engines
    EP3846087A1|2021-07-07|Method and system for selecting a learning model within a plurality of learning models
    WO2015193502A1|2015-12-23|Method for creating a data structure representing fluid consumption of at least one system, device and corresponding program
    WO2021009364A1|2021-01-21|Method for identifying outlier data in a set of input data acquired by at least one sensor
    FR3067501B1|2019-08-02|PROCESS FOR TREATING THE LOAD CURVE
    FR3054904B3|2019-11-08|METHOD FOR PROCESSING DIGITAL INFORMATION CONSISTING OF COLLECTING INFORMATION FROM A PLURALITY OF INFORMATION SYSTEM
    EP3646255A1|2020-05-06|System and method for managing risks in a process
    同族专利:
    公开号 | 公开日
    CL2017000173A1|2018-01-05|
    AU2015293548A1|2017-02-02|
    WO2016012972A1|2016-01-28|
    ES2703573T3|2019-03-11|
    US20170212003A1|2017-07-27|
    EP3172548A1|2017-05-31|
    CN106796157A|2017-05-31|
    AU2015293548B2|2020-03-12|
    AU2015293548B9|2020-03-26|
    BR112017000076A2|2017-11-14|
    EP3172548B1|2018-09-26|
    US10571358B2|2020-02-25|
    FR3024260B1|2016-07-29|
    CA2954812A1|2016-01-28|
    MX2017001102A|2017-09-29|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US5866803A|1995-12-28|1999-02-02|Agency Of Industrial Science And Technology|Gas leakage detection system|
    EP2477020A1|2011-01-17|2012-07-18|Clevergas Holding S.A.|System for detecting a fluid leak|FR3110724A1|2020-05-25|2021-11-26|Suez Groupe|Improved detection of changes in the physico-chemical composition of a liquid|CN1938567A|2004-03-22|2007-03-28|鲁比康研究有限公司|Loss detection system for open channel networks|
    JP4822990B2|2006-09-07|2011-11-24|株式会社東芝|Water leakage monitoring system|
    US7920983B1|2010-03-04|2011-04-05|TaKaDu Ltd.|System and method for monitoring resources in a water utility network|
    CN101871834A|2010-04-23|2010-10-27|北京中科通视科技有限公司|Wireless remote water leakage detection device and system|
    ITRM20110151A1|2011-03-28|2012-09-29|Giovanni Gregori|METHOD FOR THE DYNAMIC QUANTITATIVE CHARACTERIZATION OF THE AGING OF A SOLID MATERIAL.|
    US8341106B1|2011-12-07|2012-12-25|TaKaDu Ltd.|System and method for identifying related events in a resource network monitoring system|
    JP5756767B2|2012-02-29|2015-07-29|株式会社日立製作所|Water leak detection device|WO2014151384A1|2013-03-15|2014-09-25|Mueller International, Llc|Systems for measuring properties of water in a water distribution system|
    US11041839B2|2015-06-05|2021-06-22|Mueller International, Llc|Distribution system monitoring|
    WO2019063648A1|2017-09-29|2019-04-04|Suez Groupe|Improved detection and characterization of anomalies in a continuum of water|
    法律状态:
    2015-06-11| PLFP| Fee payment|Year of fee payment: 2 |
    2016-01-29| PLSC| Search report ready|Effective date: 20160129 |
    2016-06-29| PLFP| Fee payment|Year of fee payment: 3 |
    2017-07-03| PLFP| Fee payment|Year of fee payment: 4 |
    2018-06-22| PLFP| Fee payment|Year of fee payment: 5 |
    2019-07-26| PLFP| Fee payment|Year of fee payment: 6 |
    2020-07-28| PLFP| Fee payment|Year of fee payment: 7 |
    2021-07-26| PLFP| Fee payment|Year of fee payment: 8 |
    优先权:
    申请号 | 申请日 | 专利标题
    FR1457209A|FR3024260B1|2014-07-25|2014-07-25|METHOD FOR DETECTING ANOMALIES IN A DISTRIBUTION NETWORK, ESPECIALLY DRINKING WATER|FR1457209A| FR3024260B1|2014-07-25|2014-07-25|METHOD FOR DETECTING ANOMALIES IN A DISTRIBUTION NETWORK, ESPECIALLY DRINKING WATER|
    CA2954812A| CA2954812A1|2014-07-25|2015-07-23|Method for detecting anomalies in a distribution network, in particular for drinking water|
    PCT/IB2015/055583| WO2016012972A1|2014-07-25|2015-07-23|Method for detecting anomalies in a distribution network, in particular for drinking water|
    AU2015293548A| AU2015293548B9|2014-07-25|2015-07-23|Method for detecting anomalies in a distribution network, in particular for drinking water|
    MX2017001102A| MX2017001102A|2014-07-25|2015-07-23|Method for detecting anomalies in a distribution network, in particular for drinking water.|
    ES15762729T| ES2703573T3|2014-07-25|2015-07-23|Procedure to detect anomalies in a distribution network, in particular drinking water|
    EP15762729.0A| EP3172548B1|2014-07-25|2015-07-23|Method for detecting anomalies in a distribution network, in particular for drinking water|
    BR112017000076A| BR112017000076A2|2014-07-25|2015-07-23|method for detecting anomalies in a distribution network, in particular, drinking water|
    US15/328,520| US10571358B2|2014-07-25|2015-07-23|Method for detecting anomalies in a distribution network|
    CN201580039816.8A| CN106796157A|2014-07-25|2015-07-23|Abnormal method in detection distribution network, particularly drinking water distribution network|
    CL2017000173A| CL2017000173A1|2014-07-25|2017-01-23|Method for detecting anomalies in a distribution network, in particular a drinking water distribution network|
    [返回顶部]