System for estimating individual water consumption

Abstract

The present invention relates to a system for estimating the individual water consumption of a plurality of devices supplied by the secondary fluid distribution network of a user. The system comprises a sleeve able to be fitted around the supply pipe of the secondary network and comprises an electromechanical sensor placed against the outside wall of the supply pipe, and a processor for analyzing the signals delivered by the electromechanical sensor with a view to extracting information characterizing the individual consumption of the devices supplied by the secondary network.

Claims

1. A system for estimating an individual water consumption of a plurality of devices supplied by a secondary fluid distribution network of a user, comprising: a sleeve configured to be fitted around a supply pipe of said secondary network said sleeve comprising a passive electromechanical sensor placed against an outside wall of said supply pipe, and a processor for analyzing the signals delivered by said electromechanical sensor with a view to extracting information characterizing the individual consumption of at least part of the devices supplied by said secondary network.

2. The system according to claim 1, wherein said sleeve comprises means for the remote transmission of signals delivered by the electromechanical sensor.

3. The system according to claim 1, wherein said sleeve further comprises an electronic circuit for pre-processing signals delivered by the electromechanical sensor.

4. The system according to claim 1, wherein said processor is associated with a memory for saving a library of signatures of consumption of at least part of the individual devices.

5. The system according to claim 1, wherein said processor is programmed to perform a process of estimation and classification based on signals supplied by said electromechanical sensor.

6. The system according to claim 1, wherein said electromechanical sensor comprises at least one MEMS acceleration sensor, a piezoelectric sensor, an electret microphone and a MEMS microphone.

7. The system according to claim 1, further comprising at least one valve controlling the supply of at least one of said devices based on said information characterizing the individual consumption of at least part of the devices.

8. A computer program stored in a non-transitory storage media for implementing a system according claim 1, for controlling a processor for analyzing signals delivered by an electromechanical sensor with a view to extracting information characterizing the individual consumption of the devices supplied by a secondary network.

9. The system according to claim 1 in which the sleeve includes a microphone to detect the ambient sound signals in order to eliminate spurious signal.

10. A sleeve for acquiring signals with a view to estimating the individual water consumption of a plurality of devices, comprising two tubular half-shells linked by a joint enabling the opening thereof so that it can be inserted around a pipe, and the closing thereof so that a passive sensor can be applied against a wall of said pipe.

11. The sleeve according to claim 10, further comprising means for adjusting the load pressure of the sensor on an outside wall of the supply pipe.

12. The sleeve according to claim 10, further comprising a connector for receiving a peripheral device containing an electronic memory wherein a computer file for updating the computer program controlling the processor for analyzing signals is saved.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will be better understood upon reading the following detailed description, relating to a not restrictive exemplary embodiment referring to the accompanying drawings wherein:

(2) FIG. 1 shows a perspective view of a sleeve according to the invention;

(3) FIGS. 2 to 5 represent the signals detected by various types of sensors; and

(4) FIGS. 6 and 7 show two other types of signals detected by an accelerometer.

DETAILED DESCRIPTION

(5) FIG. 1 shows a view of a sleeve according to the invention, consisting of a housing 1 made of plastic material through which a cylindrical bore 2 goes, the section of which is complementary to the section of a water supply pipe.

(6) The sleeve 1 is formed of two parts 3, 4 linked by a joint 5. Each one of the parts 3, 4 has a semi-tubular inside wall, and a locking spoiler respectively 6, 7.

(7) More precisely, the first part 3 has a tile or semi-tubular trough shape with a radius increasing from the joint 5 to the opposite edge.

(8) It has a joint area along one of the longitudinal edges thereof to form a joint with the second part 4. The other longitudinal edge has a shoulder 7 extending radially towards the inside of the part 3. The inner radius of curvature of the first part 3 measured at the spoiler 7 corresponds to the outer radius of curvature of the second part 4 measured at the additional spoiler 6 forming a shoulder adapted to link with the shoulder of the first spoiler 7.

(9) The spoiler 7 has a chamfered longitudinal edge, on the side opposite the shoulder, to enable the introduction of a blade when dismounting the collar.

(10) The sleeve can be opened to allow the adjustment on a water supply pipe, requiring neither water supply shut-off nor any other action thereon.

(11) When the sleeve 1 is positioned around the water supply pipe, the two parts 3, 4 are closed and locked to surround the pipe, with the semi-tubular surfaces of the two parts coming into contact with the outside surface of the pipe.

(12) One of the parts 4 contains an electronic circuit for executing the pre-processing of the signals which are transmitted to a remote processor using a remote wired or radio frequency connection, using the WiFi standard, for instance. Such signals are also saved in a local memory to enable a reading from a device via a USB port 8 or using a memory card 9 of the SIM type for example.

(13) In the example described, the second part 4 incorporates a vibration or acoustic sensor which can be pressed against the pipe wall when closing and clipping the two complementary parts 3, 4.

(14) An adjustment member 10 comprises, for example, a system comprising an adjusting screw interacting with a thread provided in the wall of the first part 3. This screw acts on a cradle positioned inside the first part 3 in order to adapt the section defined by the cradle on the one hand and the second part 4 on the other hand, to the section of the pipe.

(15) The physical parameters measured on the water inlet pipe are selected among the following ones: the acoustic waveform, the deformation or vibration phenomena, passively and not through the emission of a modulated wave front by the flow of the fluid in the pipe.

(16) The parameters do not include the measuring of the interaction between an incident wave, for example a front of ultrasonic waves, and the liquid flowing in the pipe. The measured signals are not those supplied by an ultrasonic flow meter, based on the measuring of the modulation of a carrier wave by the ultrasonic flow rate of a fluid. The devices do not include an emission source associated with the measuring of the modulation of such emission source by the fluid flow.

(17) The sensor according to the invention acts passively, on the detection of the intrinsic signals emitted by the interaction of the fluid with the pipe, and not through the analysis of the modulation of a reference incident wave by the fluid flowing in the pipe.

(18) The sensor is chosen from: condenser microphones, electret microphones such as the microphone sold with reference POM-3044-R by the PROJECTS UNLIMITED company, MEMS microphones, such as the microphone sold with reference ADMP504 by the ANALOG DEVICES company, piezo-ceramic sensors, such as the component 27301 sold by the FERROPERM company, resistance strain gauges, and MEMS accelerometers, such as the accelerometer sold with reference ADXL103 by the ANALOG DEVICES company.

(19) The signals delivered by the sensor are operated by a processor controlled by software intended for automatically classifying and quantifying the categories recognized according to time, frequency, duration and/or intensity variables. The person skilled in the art knows the general probabilistic classification and estimation methods, which use Bayesian models, neural networks or Markow models.

(20) The processing performed for the individual characterisation consists in firstly saving a reference library containing the signature of each one of the devices to be monitored.

(21) Such saving can be predefined or performed upon powering the system according to the invention through the saving of the signal transmitted by the sensor upon powering on and off of a specific device.

(22) FIGS. 2 to 5 represent the signals detected by various types of sensors, respectively: An accelerometer sold with reference ADXL103 by the ANALOG DEVICES company (FIG. 2), A piezo-ceramic sensor sold with reference 27301 by the FERROPERM company (FIG. 3), An electret microphone sold with reference POM-3044-R by the PROJECTS UNLIMITED company (FIG. 4), and A MEMS microphone sold with reference ADMP504 by the ANALOG DEVICES company (FIG. 5).

(23) The signals used, in this example, are amplitude versus time.

(24) FIGS. 6 and 7 show other forms of signals acquired with an accelerometer, namely: power versus frequency (FIG. 6), and the spectrogram (FIG. 7).

(25) These signals are subjected to a pre-processing consisting in sampling the signals and if necessary in carrying out filtering to eliminate the spurious signals, such as ambient sound signals, for instance. For this purpose, the sleeve may further include a microphone to detect the ambient sound signals.

(26) The processing applied to the signals delivered by the sensor aims at assigning signals or signal segments to categories or classes.

(27) The phase of learning comprises the retrieval of descriptors as digital signatures from a base of reference signals.

(28) The classification of the signals delivered by the sensor implements the technique based on hidden Markov models (HMM for Hidden Markov Models), the Gaussian mixture models (GMM for Gaussian Mixture Models) or the dynamic time alignment (DTW for Dynamic Time Warping).

(29) A system of statistical classification models each class of signals using a random variable that is often a Gaussian distribution. The statistical classification is the calculation of the likelihood of the signal belonging to each one of the possible classes, which will determine the most probable belonging class. The calculation uses an acoustic setting of the signal and the class models to be identified.

(30) The Gaussian distribution mixture models (GMM) are used in the case of complex signals where more than one random variable has to be considered.

(31) The classification of sounds using a GMM model includes two steps: a phase of learning of the system using a set of files admitted as representative of a class and a second phase of verification of any sound belonging to this class.

(32) Learning aims at estimating the parameters of the Gaussian distributions that make up the model from the acoustic vectors of the sounds composing the class. The learning of a class can be broken down into two successive steps: first obtaining approximate values of the class distribution parameters using the K-means algorithm, then optimising the values of such parameters using an algorithm of the EM (Expectation Maximisation) type. The phase of classification makes it possible to determine the most probable class from the calculation of the likelihood for each acoustic vector of the signal. The likelihood of the sound composed of a time sequence consisting of several vectors is the geometric mean of likelihood of each one of such vectors. The belonging sound class is the one for which the average value of likelihood is maximum.

(33) The system according to the invention makes it possible to very simply obtain information on all the devices supplied by the water distribution network, without it being necessary to equip each device with an individual sensor.

(34) It may comprise one or more controlled valve(s), with each one being associated with one of the devices or a sub-set of devices. The valve is controlled using the information characterising the individual consumption.

(35) Such information triggers the operating cycle of the devices, for example the shut-off of the supply after a predetermined time.

(36) For example, for an installation comprising a plurality of showers, the centralised detection of the activation of a shower makes it possible to send information triggering the timer to the control circuit of the relevant shower. This solution makes it possible to simplify the design of the controlled valves, while avoiding the integration of a flow sensor in each one of the valves.