System, detector unit and method for detecting a leak in a water pipe system

Abstract

System for detecting a leak in a water supply system, in particular a water supply system in a building, wherein several consumers can be supplied with water from the water supply system (8) and wherein each of the several consumers, when drawing water from the water supply system (8) - active consumer - emits sound into the water supply system (8), comprising: a flow measuring unit (3) designed to be attached to a supply line (5) of the water supply system (8) and to measure an actual flow rate through the supply line (5), a sound measuring unit (4) designed for attachment to the supply line (5) of the water supply system (8), for measuring sound in the water supply system (8) and for outputting sound measurement values ​​based on measured sound, wherein the sound measuring unit (4) is designed for detecting structure-borne sound, a sound evaluation unit (11) designed to detect active consumers based on sound measurement values, an estimation unit (12) designed to determine an estimated flow rate from the active consumers detected by the sound evaluation unit (11), and an evaluation unit (14) which is trained to compare the actual flow rate with the estimated flow rate and to detect a leak in the water supply system (8) based on a comparison result.

Description

[0001] The invention relates to a system, detector unit and method for detecting a leak in a water supply system, in particular a water supply system in a building, wherein several consumers can be supplied with water from the water supply system and wherein each of the several consumers emits sound into the water supply system when drawing water from the water supply system - active consumer.

[0002] Leaks in water pipe systems cause significant damage every year. In 2019, leaks in Germany alone resulted in more than 3,000 water damage incidents per day, with average damages exceeding €2,800. The damage increases with the amount of water that escapes and the longer it remains exposed to the building, its contents, or other items that come into contact with water. These individual incidents add up to considerable sums annually. Even micro-leaks, from which only small amounts of water escape, for example, 10 milliliters per minute, can lead to costly repairs, as affected wall sections remain persistently damp due to the escaping water, potentially resulting in significant mold growth and structural damage. Therefore, there is a need to detect leaks of all sizes and to respond to them promptly.

[0003] Several approaches addressing this need are known from practical experience. One approach uses smart flow control valves that are flanged into a water pipe, for example, directly after a water meter. These valves monitor individual water consumption. If typical consumption levels are exceeded, the water system is shut off and a warning message is sent. While this approach offers significant benefits, it is also very expensive. Furthermore, installation is complex, as the water system must be shut off for the installation, and the work can only be carried out by qualified personnel.

[0004] Another approach is based on the "smart home" principle. Multiple sensors and actuators are networked together. The sensors detect moisture at various points within a building, and additional sensors can be integrated and added to the system relatively arbitrarily. If elevated moisture levels are detected, an actuator can shut off the water supply and issue a warning. This allows such systems to be adapted very flexibly to a wide variety of installation scenarios and generate significant benefits. However, sensors currently available on the market only react to contact with water. This means that considerable amounts of water must be leaking, and a puddle must already have formed, before a sensor reacts to a leak. Damage from micro-leaks goes undetected. Furthermore, the large number of sensors and actuators generates considerable costs.

[0005] Another approach involves monitoring the flow rate through a supply line of a water supply system. Solutions like these have been described by companies such as Smart Wins Technologies GmbH, Berlin, Germany, and HomeServe Labs, Walsall, England. In this approach, the actual flow rate is measured non-invasively by a measuring unit and analyzed in the cloud using artificial intelligence (AI). This is intended to detect pipe bursts and micro-leaks, as well as dripping faucets. However, these approaches require high bandwidth to the cloud and / or relatively significant computing resources in the measuring unit to implement the AI-based methods.

[0006] US 2021 / 0215569A1 describes the detection of a water leak using pressure measurements. This involves recording both the flow rate and the pressure. Non-cyclical pressure events are used to infer the presence of a leak.

[0007] US patent 2014 / 0165731A1 discloses a pipe leak detection system. This system uses an acoustic sensor attached to the pipe, the signals from which are evaluated by a monitoring system. A flow sensor and a pressure sensor can also be used. Leaks are detected by a continuous tone or a characteristic sound spectrum. A pipe rupture is detected by a pressure wave.

[0008] US 2016 / 0146648A1 relates to the detection of events that affect fluid flow in a fluid distribution network. For this purpose, pressure is measured at one or more points in the distribution network. An analysis of the pressure readings, along with the characteristics of individual consumers, allows conclusions to be drawn about active consumers. Furthermore, the flow rate is estimated. If a flow meter detects a low flow rate over an extended period without identifying an active consumer, a leak is inferred. Another approach utilizes the temporal behavior of the measured values.

[0009] German patent application DE 10 2017 005 499 A1 describes a method for detecting fluid loss in a fluid supply network. A metering device is installed on a supply line to record the quantity of fluid supplied. Usage sensors detect consumers to determine usage data. Based on this usage data and the quantity of fluid supplied, the system determines whether fluid loss has occurred.

[0010] The invention is based on the objective of providing a system, a detector unit, and a method of the type mentioned above with which a leak in a water pipe system can be detected reliably, cost-effectively, and with minimal effort. A non-invasive installation is desirable so that an end customer can also put the system into operation.

[0011] This task is solved by a system, a detector unit and a method having the features of the dependent claims.

[0012] Further advantageous embodiments are disclosed in the respective subclaims and the following description.

[0013] It should be noted that the features listed individually in the claims can be combined with one another in any technically meaningful way (even across category boundaries, for example between method and apparatus) and demonstrate further embodiments of the invention. The description further characterizes and specifies the invention, particularly in conjunction with the figures.

[0014] It should also be noted that the conjunction “and / or” used herein, which stands between two characteristics and links them together, is always to be interpreted in such a way that in a first embodiment of the described object only the first characteristic can be present, in a second embodiment only the second characteristic can be present, and in a third embodiment both the first and the second characteristic can be present.

[0015] It has been recognized that reliable leak detection does not require costly sensor-actuator networks. Instead, it is sufficient to gather various pieces of information from a water supply line and combine them in a specific way. Every appliance supplied with water through the system generates sound when drawing water and transmits this sound into the system. Since water is a good conductor of sound, the sound waves also reach the water supply line. Metallic pipes in the system further enhance sound transmission. By measuring and analyzing the sound, it is possible to determine whether an appliance is active and—given sufficient analysis—which specific appliance is in use. Knowing which appliances are active allows for an estimation of the expected flow rate.If, in addition, the flow rate through the supply line into the water supply system is measured, a reliable deviation of this actual flow rate from the estimated flow rate can be determined. This, in turn, can indicate a possible leak. It can be advantageous to incorporate knowledge from the system's previous operation, thereby enabling the system to be self-learning to a certain extent. Machine learning and neural networks can also be used in this process.

[0016] A system according to the present disclosure, implementing this approach, comprises a flow meter, a sound meter, an estimating unit, and an evaluation unit. The flow meter can be installed on a supply line of the water supply system. There, it measures the actual flow rate, i.e., the amount of water flowing through the supply line into the water supply system. During system operation, the sound meter can measure sound in the water supply system and generate sound measurement values. These sound measurement values ​​are transmitted to the sound evaluation unit, which, based on the sound measurement values, can detect active consumers and transmit the information thus obtained about active consumers to the estimating unit. The estimating unit determines an estimated flow rate from the detected active consumers, i.e., the amount of water that is typically expected to be drawn by the detected active consumers.This estimate can incorporate empirical data from the system's previous operation. This estimated flow rate, along with the actual flow rate, is entered into the evaluation unit, which compares the two flow rates. A comparison provides insights into whether a leak might be present. By combining sound and flow rate measurements, the reliability and accuracy of leak detection can be significantly increased.

[0017] This type of sensor data combination can also be considered sensor data fusion. This is because sensor data from the acoustic measurement unit and the flow measurement unit can be combined. Additionally, temporal profiles and / or analyses of the sensor data can also be incorporated into the sensor data fusion.

[0018] A "water supply system" to which this disclosure applies can be constructed in a relatively arbitrary manner. What is important is that the water supply system can supply one or more consumers with water and thus conveys water from a supply line to the consumers. The supply line, on the side facing away from the water supply system, can be connected, for example, to a public water network, a spring, a well, a cistern, or a water treatment plant, to name just a few conceivable examples. How and by what means water is conveyed from the supply line to the consumers is of secondary importance. Accordingly, it is also not decisive whether the pipes of the water supply system are installed in a wall, on the outside of a wall, in a ceiling, in the ground, and / or in some other way.The specific installation method of the pipes or pipe sections is only of interest insofar as it could, at least to some extent, affect the potential coupling of additional noise into the water supply system. However, these boundary conditions can usually be addressed without difficulty.

[0019] The water supply system can also fulfill different functions. In one configuration, the water supply system is part of a supply network that, for example, provides water to several buildings. In another configuration, the water supply system is installed within a building and supplies various consumers within the building and on the surrounding grounds, such as a garden or lawn.

[0020] The material from which the pipes and / or pipe sections of the water supply system are made is also of secondary importance. Plastic pipes can be used just as well as pipes made of metal or other materials. Only the supply line might be advantageously made of a metal, such as stainless steel, as this design can facilitate the transmission of sound from the water supply system to the sound measurement unit.

[0021] Furthermore, it is advantageous for the "supply line" to be the point where all water supplied to consumers via the water supply system is fed into the system. If multiple supply lines are present, it may be necessary to measure the actual flow rate at each of them.

[0022] A "consumer" can be any device that can be connected to the water supply system and can, in principle, draw water from it. Examples of consumers include a faucet, a shower head, a toilet flush, a washing machine, a dishwasher, a sprinkler system, an ice cube machine, or a coffee machine with a water connection, to name just a few possible, non-limiting examples.

[0023] At the same time, it can be advantageous if the number of consumers is not too large. A large number of consumers means that the evaluation becomes increasingly complex. This is because, firstly, the sound mixture arriving at the sound measuring unit becomes significantly more complex, and secondly, estimating the flow rate becomes increasingly time-consuming. Preferably, in an application scenario of the system disclosed herein, the water supply system provides water to fewer than 200 consumers, particularly preferably a maximum of 100, most preferably a maximum of 50 consumers, and further preferably a maximum of 25 consumers.

[0024] For the purposes of this disclosure, an "active consumer" is defined as a consumer connected to the water supply system who draws water from the system at any given assessment date. Therefore, any consumer who is active at a first assessment date may no longer be an active consumer at a second assessment date, and vice versa. Conversely, active consumers may remain active at different assessment dates. Active consumers are a subset of all consumers connected to the water supply system.

[0025] This reveals that every active consumer generates and emits characteristic noises, which can depend on the type of consumer, its design, material, manufacturer, flow rate, specific installation situation, and / or other environmental factors. These characteristic noises differ, for example, in their level and spectrum. The frequencies of the noises range from a few hertz to 20 kilohertz and above. Based on these characteristic noises, conclusions can be drawn about the specific active consumer from the sound waves received by the sound measurement unit.

[0026] An analysis of the overall noise is sufficient, as this noise is created by a mixture of characteristic sounds. Even more detailed information can be obtained by breaking down the overall noise into individual "sound components," as this allows for better isolation of the individual characteristic sounds. In this way, a wide variety of active consumers can be identified, right down to a dripping tap or a running toilet flush.

[0027] The evaluation of recorded sound measurements is the task of the "sound evaluation unit." This task can be accomplished by the sound evaluation unit in various ways. For example, the unit can further process the sound measurements, such as through appropriate filtering. In this way, the unit can isolate and / or eliminate background noise from a spectrum of the sound measurements. Such background noise can include, for example, noise pulses caused by a pipe in the water supply system striking the ground or vibrations coupled in from other devices, such as a fan. This background noise is best detected during times when no consumer is active.

[0028] The sound analysis unit is designed to identify active water consumers from sound measurements. It can specifically determine which of the consumers connected to the water supply system are currently active. However, the analysis may also reveal that none of the consumers are currently active, in which case a measured actual flow rate is very likely due to a leak.

[0029] The sound analysis unit can also provide initial indications of a leak, as leaks – depending on their size and design – can also generate noise. For example, a leak can cause a constant "hissing" sound that increases rather than decreases in intensity.

[0030] The "estimation unit" estimates the "estimated flow rate" from the detected active consumers, i.e., the amount of water delivered to the active consumers via the supply line. If no active consumers are detected, the estimated flow rate is zero. Depending on the detected active consumers, a non-zero estimated flow rate would be determined. For example, analyzing the recorded sound might indicate that a toilet is flushing. Since toilets are usually either on or off, the flow rate can be estimated relatively easily. With a detected active faucet, different turbulences occur depending on the flow rate, which in turn produce different sounds. Therefore, the sound emitted by a faucet allows conclusions to be drawn about the flow rate.

[0031] The flow meter can be designed in various ways. The important thing is that the flow meter can reliably measure the actual flow rate with sufficient accuracy. In principle, the flow meter can be installed in the supply line, similar to a water meter. Due to its very simple installation, one design of the flow meter operates non-invasively, meaning it is attached to the outside of the supply line. It can be advantageous if the flow meter can be fitted around the supply line like a sleeve.

[0032] The "sound measuring unit" can also be designed in different ways. As long as the sound measuring unit is capable of detecting sound from the water supply system and generating sound measurement values ​​from it, it can be used in connection with the present disclosure. The sound measuring unit is attached to the supply line.

[0033] Furthermore, it can be advantageous if the sound level meter is positioned at a sufficient distance from other sound sources (i.e., sound sources not originating from an active consumer). A shut-off valve or a water meter, for example, can act as such a sound source. If the distance to these other sound sources is too small, the sound level meter might only be able to detect the sound from these other sources and no longer be able to adequately resolve the sound from active consumers.

[0034] The flow meter and / or the sound meter can each be controlled by an electronic control unit. This control unit can provide power to the respective sensing element and process sensor signals. The control unit can amplify, filter (e.g., with high-pass, low-pass, band-pass, or notch filters), convert analog to digital, and / or process the sensor signals in other ways. The control unit can be connected to the sensing element, acquire a sensor signal based on the physical quantity acting on the sensing element, and convert it into a measured value representative of that physical quantity. This measured value can be output analogously or digitally. The design of the control unit is likely to depend on the design of the respective sensing element and its sensor technology.

[0035] The system can be implemented in various ways. In one configuration, the system is implemented entirely in hardware. In another configuration, the system is implemented through a combination of hardware and software. The hardware can include sensors, analog-to-digital converters, filters (e.g., high-pass, low-pass, band-pass, notch filters), a processor (e.g., a microcontroller, digital signal processor, or ASIC (application-specific integrated circuit)), and / or a programmable logic circuit (e.g., an FPGA (field-programmable gate array) or CPLD (complex programmable logic device)). Furthermore, one or more memory modules can be present, such as RAM (random-access memory), ROM (read-only memory), or flash memory, which other hardware components can access.Software can control the individual components of the system, such as the processor or parameters of an analog-to-digital converter. The software can be stored in one of the memory locations and loaded from memory for execution. This approach allows the system to be flexibly adapted.

[0036] In one embodiment, the flow meter includes a temperature sensor, which is designed to be attached to a pipe in the water supply system, preferably a supply pipe, and to detect the pipe's temperature. This enables non-invasive, cost-effective, and reliable measurement of the flow rate. This approach is based on the fact that the incoming water is usually colder than the surrounding environment of the supply pipe. As a result, the supply pipe cools down more significantly with increasing flow rate. At lower flow rates or without flow, the surrounding environment warms the supply pipe, so that there is an overall correlation between temperature and flow rate. By measuring the temperature, the flow rate can be determined with remarkable accuracy. Furthermore, this embodiment offers another advantage.Because if the temperature of the supply line drops below a certain threshold, there is an increasing risk of icing, which in turn increases the risk of leakage. Therefore, the system disclosed here can also be used as an icing warning device.

[0037] In a further development, this flow metering unit comprises a heat source, from which a defined heat flow can be emitted, and from which the temperature sensor detects the effect of the heat flow on a measured temperature. In this way, the actual flow rate can be determined even more precisely. This is because flowing water carries away the heat introduced by the heat flow. The more water flows through the supply line, the more heat is carried away, resulting in a lower temperature measured by the temperature sensor. Conversely, the less heat is carried away by the water, the higher the measured temperature will be. The heat source preferably emits less than 20 watts, more preferably less than 10 watts, and most preferably less than 5 watts to the supply line. Preferably, the emitted heat is greater than 1 watt.

[0038] In a further refinement, a temperature sensor can be positioned both before and after the heat source. This allows one sensor to measure the temperature of the incoming water, while the other measures the temperature increase caused by the heat source. This enables a more reliable determination of the heat source's effect on the water temperature in the supply line.

[0039] The sound measurement unit is designed to detect structure-borne sound. In one embodiment, the sound measurement unit comprises a piezo-based or a MEMS microphone. Detecting structure-borne sound allows for relatively low-loss sound acquisition, as structure-borne sound can be very effectively coupled into microphones. Piezo-based microphones offer a robust and cost-effective sound measurement unit. MEMS (Micro-Electro-Mechanical System) microphones achieve high sensitivity while being highly miniaturized. This allows for the creation of a sensitive yet compact sound measurement unit.

[0040] In one embodiment, the sound evaluation unit is designed to analyze sound measurements and / or the temporal progression of sound measurements with respect to a level, multiple level values, a frequency, a specific time profile, a duration, and / or a spectrum. Analyzing sound measurements provides immediate results without requiring any waiting time. Analyzing the temporal progression of sound measurements allows for a more in-depth analysis of the sound. This makes it possible to determine a spectrum. Furthermore, changes in the recorded sound over time can be reliably detected, potentially indicating the existence of a problem. This is because most appliances are only active for short periods; even appliances that are active for very long periods, such as a bathtub faucet or a lawn sprinkler, are not continuously active.This allows a persistent sound component to point to a problem, such as a constantly running toilet cistern or a leak. Accordingly, an analysis of a specific time course or duration can be helpful. An analysis of a spectrum allows for the isolation of specific active noise sources. An analysis of one or more levels can easily provide information about active noise sources, for example, that no noise source is active.

[0041] In one embodiment, the sound evaluation unit includes a comparator, which compares sound measurements with a threshold value and detects active consumers based on the comparison result. This embodiment allows for particularly simple analysis. In this way, it can be detected, for example, when no consumer is active or when particularly noisy consumers are active, such as pressure flush valves.

[0042] In one embodiment, the sound evaluation unit includes a frequency analyzer, which is configured to determine a spectrum of sound measurements. This allows for more comprehensive analysis of the sound measurements. The spectrum can be determined, for example, using an FFT (Fast Fourier Transform). If only individual frequencies are of interest, other approaches can be used, such as the Goertzel algorithm. The specific operation of a frequency analyzer depends on the information to be obtained, the available resources, and / or other constraints, and is not essential to the present disclosure.

[0043] In one embodiment, the sound evaluation unit comprises a classifier, wherein the classifier is configured to analyze sound measurements and / or a spectrum of sound measurements using at least one sound signature, and wherein each of the at least one sound signature characterizes a sound emitted by an active consumer. In this way, detailed information about active consumers can be obtained. As previously mentioned, every consumer emits sound when drawing water from the water supply system, and this sound depends on various environmental factors. This sound, specific to the respective consumer, can be summarized in a sound signature. A sound signature can, for example, be a spectrum of a specific active consumer. The classifier can analyze the sound measurements and / or their temporal evolution using a correlation, such as cross-correlation.However, the classifier can also use a neural network or other methods in the field of "artificial intelligence".

[0044] In one embodiment, the system additionally includes a memory containing multiple sound signatures, each characterizing a sound emitted by an active consumer, and / or the memory being configured to store sound measurements and / or extracted characteristics of sound measurements and / or actual flow rates. Storing sound signatures allows for flexible adaptation to changing installation situations. The sound signatures can be stored in a variety of ways. For example, they can be in the form of spectra or parameters of a neural network capable of identifying a specific consumer. The information in the memory can be stored during the manufacturing of the detector unit. However, it is also conceivable that the information could be loaded from an external memory during a parameterization process.This external storage can, for example, be located in a cloud system connected via a communication interface. The external storage can also be located on or accessible via a mobile device configured for parameterization. This mobile device could, for example, be a smartphone with a suitable app.

[0045] Storing previous sound level measurements, characteristics, actual flow rates, and / or other information generated during system operation, forming "historical values," allows for targeted analysis of changes. These historical values ​​can be stored in various ways. It is advisable to include a timestamp alongside the stored values ​​to provide a chronological context for their acquisition.

[0046] Memory can be implemented in various ways. Both volatile and non-volatile memory can be used. Examples include RAM (Random Access Memory), MRAM (Magnetoresistive Random Access Memory), EEPROM (Electronically Erasable Read Only Memory), and flash memory. When using volatile memory, its contents can be protected by a buffer element, such as a gold cap (or other capacitor) or an accumulator.

[0047] In one embodiment, the system additionally features a signaling unit. Upon detecting a leak and / or a potential leak, this unit generates and transmits a warning message. The signaling unit can issue a visual and / or audible warning, for example, by flashing a light and / or emitting an alarm tone. Furthermore, in line with the Internet of Things (IoT), the signaling unit can send a message to a monitoring unit or a central monitoring station, which then sends a warning via other services such as email, an app, SMS, an instant messaging service, or similar channels. This allows users to be flexibly informed of a detected leak and / or a highly probable leak.The warning message can also be received by a maintenance system, which triggers an inspection of the possible leak, for example by a caretaker, an installer or another tradesperson.

[0048] In one configuration, the system includes a higher-level control unit that manages and / or monitors the system's functions. This control unit is designed to put the system into a learning mode, during which previously unknown active consumers connected to the water supply system can be identified and / or named. The use of a higher-level control unit provides added value by making the system even more versatile and controllable. A learning mode enables the identification of consumers connected to the water supply system. In this mode, for example, one consumer can be selectively activated after another, and the respective sound level measurements and actual flow rates can be recorded and analyzed by the system. A user can utilize a mobile device connected to the system via a communication interface to assist with the learning process.

[0049] In one embodiment, the system additionally features a preferably radio-based communication interface, which can be configured to connect a higher-level control unit, a cloud system, a mobile device, a display unit, and / or an operating unit, and / or which can be based on WLAN (Wireless Local Area Network), NB-IoT (Narrow Band Internet of Things), LTE-M (Long Term Evolution for Machines), and / or LoRaWAN (Long Range Wide Area Network). A communication interface enables flexible communication with the system, for example, the output of warning messages and monitoring of current consumption. WLAN is an interface that is already available in many buildings and is universally usable.NB-IoT is a wireless technology that enables Internet of Things devices to efficiently exchange data, offering high building penetration, low costs, and low energy consumption. LTE-M is a technology that enables higher data rates and low latency. LoRaWAN offers a long range with extremely low energy consumption.

[0050] In one embodiment, the sound evaluation unit, the estimation unit, and / or the evaluation unit are based on AI and preferably comprise one or more neural networks. The use of artificial intelligence (AI) enables particularly powerful analysis of the acquired information. In particular, AI can reliably classify sound measurements with regard to active consumers. Furthermore, AI can perform sensor data fusion very effectively, i.e., combine the measurement data from different sensors for joint evaluation. The neural networks used can be trained by known consumers in a known consumption state. This training can be carried out and / or refined during a learning mode. The training datasets can also be supplied by the manufacturers of consumers or generated in a laboratory by a service provider or the system manufacturer.The detector unit determines the values ​​and inputs them into the neural networks during system initialization.

[0051] In one embodiment, the system according to the present disclosure comprises a detector unit that implements components of the system, preferably the sound evaluation unit, the estimation unit, and / or the evaluation unit. The detector unit may additionally include or be connected to the flow measurement unit or the sound measurement unit, or components thereof, for example, via a cable. The detector unit may also have an antenna through which a communication interface can communicate with the outside. The detector unit may include a power supply, which may be, for example, a battery, a rechargeable battery, or a power supply unit. The detector unit may be arranged in a housing. This housing may have a mounting device with which the housing can be attached to the supply line of a water supply system.Overall, a detector unit enables a compact implementation of essential system components, easy installation, even by non-experts, and / or safe operation of the system.

[0052] In one embodiment of the method, a leak is inferred if the actual flow rate deviates from an estimated flow rate by a predefined difference value. This provides a simple means for evaluating the water supply system.

[0053] Further features and advantages of the invention will become apparent from the following description of exemplary embodiments of the invention, which are not to be understood as limiting and are explained in more detail below with reference to the drawing. This drawing schematically shows: Fig. 1 a block diagram of an embodiment of a system according to the present disclosure, Fig. 2 a simplified representation of an installation situation of a system according to Fig. 1 and Fig. 3 a flowchart of an embodiment of a method according to the present disclosure.

[0054] In the different figures, parts that are equivalent in function are always provided with the same reference symbols, so that they are usually only described once.

[0055] Fig. Figure 1 shows a block diagram of an embodiment of a system according to the present disclosure. The system 1 comprises a detector unit 2, a flow meter 3, and a sound meter 4. The flow meter 3 and the sound meter 4 are arranged on a supply line 5, which is in Fig. The supply line 5 conveys a flow rate (indicated by an arrow 6) from a water source 7 (for example, a public drinking water network, schematically represented as a block) to a water supply system 8 (also schematically represented as a block). The supply line 5 can also be considered part of the water supply system 8. The flow rate 6 is recorded by the flow meter 3 as the actual flow rate. The sound meter 4 detects sound generated—among other things—by several consumers supplied by the water supply system 8 (not shown) and transmitted via the water supply system 8, including the supply line 5, to the sound meter 4. The flow meter 3 inputs the recorded actual flow rate to the detector unit 2; the sound meter 4 inputs recorded sound values ​​to the detector unit 2.

[0056] The detector unit 2 comprises a control electronics 9 for the flow measurement unit 3, a control electronics 10 for the sound measurement unit 4, a sound evaluation unit 11, an estimation unit 12, a memory 13, an evaluation unit 14, a signaling unit 15 and a communication interface 16.

[0057] The control electronics 9 controls the flow meter 3, supplies it with power, and converts the sensor signals into digital measured values. The control electronics 10 performs the same function with respect to the sound meter 4. The control electronics 9 and 10 comprise one or more amplifiers, filters, an analog-to-digital converter, and other control elements. The control electronics 9 outputs a value for the actual flow rate, obtained by the flow meter 3, to the evaluation unit 14. The control electronics 10 outputs a sound measurement value to the sound evaluation unit 11.

[0058] The sound evaluation unit 11 receives and evaluates the sound measurements. Within the sound evaluation unit 11, the measurements are first analyzed to determine whether a device is active. For this purpose, the sound measurements, possibly after averaging over a period of time to eliminate impulse noise, are compared to a threshold value. If the sound measurements do not reach the threshold value, it is highly likely that no device is active. If the sound measurements exceed the threshold value, it can be concluded that at least one device is active.

[0059] If at least one active consumer is detected, the sound measurements are processed by a Fast Fourier Transform (FFT) to determine the spectral components of the sound measurements. For the classification of active consumers, the sound evaluation unit 11 includes at least one neural network (not shown). This at least one neural network is trained with sound measurements and / or their spectra generated by known consumers in a known state of operation. Furthermore, this at least one neural network may have been further trained in a learning mode of the detector unit 2 in the specific installation situation by activating individual consumers in isolation as much as possible and assigning the sound measurements and extractable sound signatures recorded by the sound measurement unit 4 to the specific consumer.The sound measurements and spectra obtained through the FFT are fed into this at least one neural network. This network classifies active consumers and outputs this information to the estimation unit 12. The output can include the specific active consumers and their identified consumption states.

[0060] The sound evaluation unit 11 can access the memory 13, in which, among other things, sound signatures of the consumers and / or a representation of the at least one neural network are stored.

[0061] Estimation unit 12 determines an estimated flow rate based on the detected active consumers and, if applicable, the detected consumption states. Estimation unit 12 can access data stored in memory 13 for this purpose.

[0062] The estimated flow rate is entered into evaluation unit 14, which compares and evaluates the estimated flow rate with the actual flow rate. If the two flow rates match or differ only slightly, no leakage is present. If the two flow rates differ by a predefined minimum amount, a leakage is assumed, and a corresponding warning message is triggered via signaling unit 15. This warning message can also include a measure of the deviation, thus indicating the leakage and the reliability of the evaluation. Signaling unit 15 can process the warning message and output it via communication interface 16.

[0063] The communication interface 16 is connected to an antenna 17, which enables a connection to a WLAN access point (not shown). The warning message can be sent to a cloud system via this antenna. The cloud system can then inform a user and / or a system monitor, for example, via email, SMS, or other messaging channels.

[0064] Memory 13 also serves as a log file, storing recorded actual flow rates, sound level measurements, detected active consumers, and / or other states and values ​​within the detector unit. Each value is saved along with a timestamp, allowing for chronological classification of the values. The logged data can be transferred, for example, to a cloud system for analysis, such as determining current and average water consumption, or output via a suitable interface (e.g., an encrypted website) to a user's mobile device (not shown).

[0065] Furthermore, detector unit 2 includes a power supply 18, which provides energy to the components of detector unit 2, the flow measurement unit 3, and the sound measurement unit. The power supply 18 is a battery.

[0066] Fig. Figure 2 shows an installation situation of a system according to the present disclosure. A detector unit 2 with an antenna 17 is clipped to a supply line 5 of a water supply system 8. The sound measurement unit can be integrated into the housing of the detector unit. The detector unit 2 is connected via a cable 19 to a heat source 20, which is part of the flow measurement unit 3 and emits a heat flow to the supply line 5. In the installation situation according to Figure 2, the sound measurement unit 2 is connected to the supply line 5. Fig. Figure 2 shows that the detector unit 2, and thus the sound measurement unit, is arranged at some distance from a shut-off valve 21.

[0067] Fig.Figure 3 shows a flowchart of an embodiment of a method according to the present disclosure. In step S1, the actual flow rate through a supply line of the water supply system is first measured. In step S2, sound that can be detected in the water supply system by a sound measuring unit is measured. Sound measurements are obtained that are representative of the sound. The sound includes, among other things, sound generated by active consumers when drawing water from the water supply system and transmitted into the water supply system. In step S3, the sound measurements are evaluated to identify active consumers. In step S4, an estimated flow rate is determined by estimating the consumption of the identified active consumers.In step S5, the actual flow rate is compared with the estimated flow rate, and if the flow rates deviate based on a predefined difference value, a possible leak is inferred. If a leak is detected, a warning message is generated and sent.

[0068] Regarding further advantageous embodiments, reference is made to the general part of the description and to the attached claims to avoid repetition.

[0069] Finally, it should be expressly pointed out that the exemplary embodiments described above serve only to illustrate the claimed teaching, but do not limit it to these exemplary embodiments. Reference symbol list 1 system 2 detector units 3 Flow measuring unit 4 sound measurement units 5 Supply line 6 flow rate 7 Water source 8 Water supply system 9 Control electronics 10 Control electronics 11 Sound evaluation unit 12 Estimation Unit 13 storage locations 14 Assessment Unit 15 Signaling unit 16 Communication interface 17 Antenna 18 Energy supply 19 cables 20 Heat source

Claims

[1] System for detecting a leak in a water supply system, in particular a water supply system in a building, wherein several consumers can be supplied with water from the water supply system (8) and wherein each of the several consumers, when drawing water from the water supply system (8) - active consumer - emits sound into the water supply system (8), comprising: a flow measuring unit (3) designed to be attached to a supply line (5) of the water supply system (8) and to measure an actual flow rate through the supply line (5), a sound measuring unit (4) designed for attachment to the supply line (5) of the water supply system (8), for measuring sound in the water supply system (8) and for outputting sound measurement values ​​based on measured sound, wherein the sound measuring unit (4) is designed for detecting structure-borne sound, a sound evaluation unit (11) designed to detect active consumers based on sound measurement values, an estimation unit (12) designed to determine an estimated flow rate from the active consumers detected by the sound evaluation unit (11), and an evaluation unit (14) which is trained to compare the actual flow rate with the estimated flow rate and to detect a leak in the water supply system (8) based on a comparison result. [2] System according to claim 1, characterized by , that the flow measuring unit (3) includes a temperature sensor, wherein the temperature sensor is designed to be attached to a line (5) of the water supply system (8) and to detect a temperature of the line (5). [3] System according to claim 2, characterized by, that the flow measuring unit (3) comprises a heat source (20), wherein a defined heat flow can be emitted by the heat source (20) and wherein the temperature sensor detects an effect of the heat flow on a detected temperature. [4] System according to any one of claims 1 to 3, characterized by , that the sound measurement unit (4) includes a piezo-based or a MEMS microphone. [5] System according to any one of claims 1 to 4, characterized by , that the sound evaluation unit (11) is designed to analyze sound measurements and / or a temporal progression of sound measurements with respect to a level, several level values, a frequency, a specific time progression, a duration and / or a spectrum. [6] System according to any one of claims 1 to 5, characterized by, that the sound evaluation unit (11) includes a comparator, wherein the comparator compares sound measurement values ​​with a threshold value and detects active consumers depending on a comparison result. [7] System according to any one of claims 1 to 6, characterized by , that the sound evaluation unit (11) comprises a frequency analyzer, wherein the frequency analyzer is configured to determine a spectrum of sound measurement values. [8] System according to any one of claims 1 to 7, characterized by , that the sound evaluation unit (11) comprises a classifier, wherein the classifier is configured to examine sound measurements and / or a spectrum of sound measurements using at least one sound signature, and wherein the at least one sound signature characterizes each sound emitted by an active consumer. [9] System according to any one of claims 1 to 8, which additionally comprises a memory (13) wherein several sound signatures are stored in the memory (13) which each characterize a sound emitted by an active consumer, and / or wherein the memory (13) is configured to store sound measurement values ​​and / or extracted characteristics of sound measurement values ​​and / or actual flow rates. [10] System according to any one of claims 1 to 9, which additionally comprises a signaling unit (15) wherein the signaling unit (15) generates and sends a warning message when a leak and / or a possible leak is detected. [11] System according to any one of claims 1 to 10, comprising a superior control unit that controls and / or monitors the functions of the system (1), wherein the control unit is configured to put the system into a learning mode, and wherein previously unknown active consumers on the water supply system (8) can be detected and / or named in the learning mode. [12] System according to one of claims 1 to 11, which additionally has a preferably radio-based communication interface (16), wherein the communication interface can be configured to connect a higher-level control unit, a cloud system, a mobile device, a display unit and / or an operating unit and / or wherein the communication interface can be based on WLAN - Wireless Local Area Network -, NB-IoT - Narrow Band Internet of Things -, LTE-M - Long Term Evolution for Machines - and / or LoRaWAN - Long Range Wide Area Network. [13] System according to any one of claims 1 to 12, characterized by that the sound evaluation unit (11), the estimation unit (12) and / or the evaluation unit (14) are based on AI and preferably comprise one or more neural networks. [14] Detector unit for detecting a leak in a water supply system, in particular a water supply system in a building, wherein the detector unit (2) implements components of a system (1) according to one of claims 1 to 13, preferably the sound evaluation unit (11), the estimation unit (12) and / or evaluation unit (14). [15] Method for detecting a leak in a water supply system, in particular a water supply system in a building, preferably using a system according to one of claims 1 to 13, wherein several consumers can be supplied with water from the water supply system (8) and wherein each of the several consumers is active when drawing water - emits sound into the water supply system (8), comprising: Measuring (S1) an actual flow rate through a supply line (5) of the water supply system (8), Measuring (S2) sound in the water supply system (8) to obtain sound measurement values, Evaluation (S3) of sound measurement values ​​to detect active consumers, Determine (S4) an estimated flow rate by estimating the consumption of identified active consumers and (S5) Comparing the actual flow rate with the estimated flow rate to detect any leakage that may be present. [16] Method according to claim 15, characterized by , that a leak is inferred if an actual flow rate deviates from an estimated flow rate by a predefined difference value.

Images