Monitoring system

Abstract

A system (2) for monitoring resource flows at a number of devices (D1, D2, D3) includes a receiving unit (4) which receives data from a number of sensors (S1, S2, S3, S4, S5) configured to detect the flow rate and/or the change in the flow rate at device level, and a number of meters (M1, M2, M3) configured to measure the flow for at least a part of the devices (D1, D2, D3). The system (2) has a calculation module (8) configured to receive information from the sensors (S1, S2, S3, S4, S5) and the meters (M1, M2, M3), and the calculation module (8) includes a mathematical statistical model (38) configured to estimate and/or predict flow of resource and/or performance (e.g., activity) of at least a selection of the devices (D1, D2, D3).

Claims

1. A system useful for monitoring resource flows at a number of devices characterized by a latent stochastic process, the system comprising: at least one sensor selected from the group consisting of piezoelectric sensors, induction sensors, temperature sensors, and lux sensors, configured to detect a flow rate and/or a change in a flow rate at said devices, wherein said at least one sensor is attached non-intrusively to said devices, and wherein said at least one sensor is configured to wirelessly communicate data; a receiving unit configured to wirelessly receive said data from said at least one sensor; at least one meter configured to measure an overall flow for at least a part of the devices; and a calculation module configured to receive information from the at least one sensor and the at least one meter, wherein the calculation module is in communication said at least one sensor; wherein said calculation module comprises a mathematical statistical model, wherein the mathematical statistical model is configured to estimate and/or predict flow of resources and/or performance of at least a selection of the devices based on data from the at least one sensor and from the at least one meter; wherein the calculation module is further configured to calibrate said at least one sensor based on said mathematical statistical model operating on said information from said at least one sensor and said at least one meter; wherein the data from the at least one sensor and from the at least one meter at time t are given as a vector (Y.sub.t), wherein the latent stochastic process at time t is a vector (.sub.t), wherein the mathematical statistic model uses a noise vector (.sub.t) of the data from the at least one sensor and from the at least one meter at time t and a noise vector (.sub.t) of the latent stochastic process at time t, and wherein a statistical distribution of the noise vectors (.sub.t, .sub.t) is estimated based on the data from the at least one sensor and from the at least one meter.

2. A system according to claim 1, further comprising: a user interface configured and arranged to provide access to the estimated and/or predicted resource flows of at least a selection of the devices.

3. A system according to claim 1, further comprising: at least one router; wherein the at least one router comprises the calculation module.

4. A system according to claim 1, wherein at least one of the at least one sensor comprises an energy harvester.

5. A system according to claim 4, wherein: the at least one sensor is configured to use energy harvested by the energy harvester to detect resource flow; or the system is configured to use energy harvested by the energy harvester to change resource flow rate of at least one of the devices.

6. A system according to claim 1, wherein the calculation module and the mathematical statistic model are integrated in an apparatus of the system.

7. A system according to claim 1, wherein the calculation module is a cloud service comprising a mathematical statistic model and a data storage.

8. A system according to claim 1, further comprising: a control unit configured to change the activity of one or more of the devices.

9. A system according to claim 1, wherein the calculation module is configured to estimate the actual and/or future activity of at least a selection of the devices.

10. A system according to claim 2, wherein the user interface is a remote user interface.

11. A system according to claim 1, wherein: vector Y.sub.t is given by
Y.sub.t=f.sub.t(.sub.t)+.sub.t; vector .sub.t is modelled by an underlying process
.sub.t=g.sub.t(.sub.t-1)+.sub.t; and f.sub.t and g.sub.t are mathematical functions.

12. A system according to claim 1, wherein the at least one meter is configured to measure a total power flow for a group of electrical devices, or to measure a total water flow, or to measure energy consumption.

13. A method for monitoring resource flows at a number of devices with a receiving unit which wirelessly receives data from a number of sensors selected from the group consisting of piezoelectric sensors, induction sensors, temperature sensors, and lux sensors, wherein said number of sensors are attached non-intrusively to said devices, wherein said number of sensors are configured to wirelessly communicate data to said receiving unit, wherein said number of sensors are configured to detect the flow rate and/or the change in the flow rate at device level, and a number of meters configured to measure the flow for at least a part of the devices characterized by a latent stochastic process, the method comprising: estimating and/or predicting loss and/or flow of resources of at least a selection of the devices by using a calculation module configured to receive information from the sensors and the meters, wherein the calculation module is in communication with and configured to calibrate said number of sensors by applying a mathematical statistical model, whereby the mathematical statistical model is configured to estimate and/or predict flow of resources and/or performance of at least a selection of said number of devices based on data from said sensors and said meters; wherein data from said sensors and from said meters at time t are given as a vector (Y.sub.t), wherein the latent stochastic process at time t is a vector (.sub.t), wherein the mathematical statistic model uses a noise vector (.sub.t) of data from the sensors and from the meters at time t and a noise vector (.sub.t) of the latent stochastic process at time t, and wherein a statistical distribution of the noise vectors (.sub.t, .sub.t) is estimated on the basis of the data from the sensors and from the meters; and calibrating said sensors based on said mathematical statistical model operating on said data from the sensors and from the meters in said calculation module.

14. A method according to claim 13, wherein the resource flows at the number of devices is characterized by a latent stochastic process, and wherein said estimating and/or predicting is performed with a system comprising: at least one sensor configured to detect a flow rate and/or a change in a flow rate at said devices; a receiving unit configured to receive data from said at least one sensor; at least one meter configured to measure an overall flow for at least a part of the devices; and a calculation module configured to receive information from the at least one sensor and the at least one meter, wherein the calculation module comprises a mathematical statistical model configured to estimate and/or predict flow of resources and/or performance of at least a selection of the devices; wherein the data from the at least one sensor and from the at least one meter at time t are given as a vector (Y.sub.t), wherein the latent stochastic process at time t is a vector (.sub.t), wherein the mathematical statistic model uses a noise vector (.sub.t) of the data from the at least one sensor and from the at least one meter at time t and a noise vector (.sub.t) of the latent stochastic process at time t, and wherein a statistical distribution of the noise vectors (.sub.t, .sub.t) is estimated based on the data from the at least one sensor and from the at least one meter.

15. A method according to claim 13, further comprising: controlling the activity of one or more of the devices based on said estimating and/or predicting.

16. A method according to claim 14, wherein: vector Y.sub.t is given by
Y.sub.t=f.sub.t(.sub.t)+.sub.t; vector .sub.t is modelled by an underlying process
.sub.t=g.sub.t(.sub.t-1)+.sub.t; and f.sub.t and g.sub.t are mathematical functions.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will become more fully understood from the detailed description given herein below. The accompanying drawings are given by way of illustration only, and thus, they are not limitative of the present invention. In the accompanying drawings:

(2) FIG. 1 shows a schematic view of a first system according to the invention;

(3) FIG. 2 shows a schematic view of a second system according to the invention;

(4) FIG. 3 shows a schematic view of a third system according to the invention;

(5) FIG. 4 shows a device being monitored by a system according to the invention; and

(6) FIG. 5 shows a portion of a system according to the invention.

DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

(7) Referring now in detail to the drawings for the purpose of illustrating preferred embodiments of the present invention, a system 2 according to the present invention is illustrated in FIG. 1.

(8) FIG. 1 is a schematic view of a system 2 according to the invention. The system 2 includes a first sensor S1 configured to monitor a first device D1 to which the first sensor S1 is attached. The system 2 also includes two further sensors (S2 and S3) which are configured to monitor a second device D2. The system 2 moreover includes two further sensors (S4 and S5) which are configured to monitor a third device D3. The sensors S1, S2, S3, S4, S5 are configured to communicate wirelessly with a router 4.

(9) The system 2 includes meters (M1 and M2) that are configured to communicate with the router 4 that is configured to communicate wirelessly with a cloud service 8. The router 4 is communicating wirelessly with a user interface 10 represented by a smartphone 10. The router 4 is also configured to communicate wirelessly with control unit 6. The control unit 6 may be an actuator capable of regulating the activity of one or more of the devices D1, D2, D3, or a fourth device D4. Regulation of the activity of one or more of the devices D1, D2, D3, D4 may be carried out by changing a flow (e.g., by using a valve), changing the speed (of a pump or a motor), or by shutting down one or more of the devices D1, D2, D3, D4, by way of example.

(10) The sensors S1, S2, S3, S4, S5 may monitor the devices D1, D2, D3 continuously or in selected time periods such as one minute every hour, one second every minute or another frequency and measurement duration.

(11) The cloud service 8 includes a data storage 36 that can be used to store information received from the router 4 or data modified or calculated by the cloud service 8. The cloud service 8 moreover includes a mathematical statistic model 38 that is configured to estimate and/or forecast the activity of one or more of the devices D1, D2, D3 or the flow (e.g., of a fluid, gas, electricity or heat) through the one or more of the devices D1, D2, D3. The mathematical statistic model 38 may be a mathematical statistic model 38 of any suitable type and the mathematical statistic model 38 may conduct inference on combined data from one or more of the sensors S1, S2, S3, S4, S5 and from one or more of the meters M1, M2, M3 to estimate and/or forecast the activity of one or more of the devices D1, D2, D3 or the flow (e.g., of a fluid, gas, electricity or heat) through the one or more of the devices D1, D2, D3.

(12) The cloud service 8 may receive inputs from a number of other systems (not shown) and by this way information may be gathered from several systems of equal or different types. By this way it is possible to provide an updated mathematical statistic model 38 according to flow profiles (e.g., resource flows) of the devices D1, D2, D3.

(13) It is preferred that the meters M1, M2, M3 are meters of high precision, however, the sensors S1, S2, S3, S4, S5 may be cheap sensors. The data from the meters M1, M2, M3 and from the S1, S2, S3, S4, S5 may be used to predict the flow rate(s) of the D1, D2, D3 by using the mathematical statistic model 38.

(14) The sensors S1, S2, S3, S4, S5 may be configured to measure flow (e.g., of a fluid in a pipe), electricity, temperature, lux (if the sensor is a lux meter for measuring illuminances in a room or outside by way of example) for example.

(15) The devices D1, D2, D3 may be various devices having a flow of resources such as power, fluid (e.g., a liquid like a coolant, beverage or water, or a gas such as natural gas). Accordingly, the devices D1, D2, D3 may be HVAC devices such as heating devices, ventilating devices and air conditioning devices or refrigerators, lamps, radiators, floor heating or electronic devices in general.

(16) The system 2 is easy to retrofit at existing installations. The sensors S1, S2, S3, S4, S5 may be added to existing installation, e.g., simply by attaching a wireless temperature sensor S1 to a heating surface or an induction sensor S2 to an electrical device D2 by way of example.

(17) The meters M1, M2, M3 may an existing meters M1, M2, M3 in an existing installation. Thus, some of the elements in the system 2 may already exist in the installation, to which the system 2 is to be retrofit.

(18) It is possible to use a repeater in case that the system 2 has to be installed in a large area. One or more repeaters may be configured to receive a signal and retransmits it at a higher level or higher power, or onto the other side of an obstruction, so that the signal can cover longer distances.

(19) FIG. 2 illustrates a schematic view of a system 2 according to the invention. The system 2 includes a wireless router 4 that is configured to communicate wirelessly with a group of sensors S1, S2, S3, S4, meters M1, M2, M3, a cloud service 8, a user interface 10 in the form of a smartphone 10, and three control units 20, 22, 40.

(20) The system 2 includes a first meter M1 configured to measure the total power flow of group of electrical devices including an air conditioner 24. It is important to underline that the system may include a large number of electrical devices even though this is not indicated in FIG. 2.

(21) The first meter M1 is configured to communicate wirelessly with the router 4 and send information about the total power flow to the router. The first meter M1 also includes a power switch 18 that can be controlled wirelessly by the router 4. In case that a power alert is generated the router 4 may shout down the main power line by using the power switch 18.

(22) The system 2 further more includes a main pipe (for water supply) 14. The main pipe 14 is equipped with a meter M2 configured to communicate wirelessly with the router 4 and send information to the router 4. The meter M2 is configured to continuously measure the total water flow through the building into which the system 2 is installed.

(23) A valve 20 is mounted on the main pipe 14 and the valve is configured to communicate wirelessly with the router 4. The router 4 can control the valve 20 and hereby regulate the flow through the main pipe 14. The valve 20 can reduce (e.g., by closing the valve 20) the flow through the main pipe 14, increase the flow through the main pipe 14, or maintain a constant the flow through the main pipe 14.

(24) The system 2 also includes a third meter M3 that is configured to communicate wirelessly with the router 4. The router 4 receives wireless information about the total energy flow measured by the energy meter M3.

(25) The first sensor S1 is attached to a first device 16 that is a water pipe 16. The water pipe 16 is provided with a control unit formed as a valve 40. The first sensor S1 is configured to detect the flow through the pipe 16 or if there is flow through the pipe 16 or if the flow through the pipe 16 is changing and then to send the detected information wirelessly to the router 4. The router 4 is configured to send wireless signals to the valve 40 and hereby control the valve 40, e.g., in order to regulate the flow through the pipe 16. The valve 40 may be closed in order to stop the flow through the pipe 16 in case of leakage (detected by the router 4). The router may compare the flow measured by the first sensor S1 with the total flow measured by the meter M2 at the main pipe 14.

(26) It may be an advantage that the sensor S1 only is required to detect if there is flow through the pipe 16 or if there is change in the flow through the pipe 16. Hereby a simple a cheap sensor S1 may be applied to the system 2.

(27) The second sensor S2 is attached to a heater 12. The second sensor S2 is configured to detect the temperature of the heater 12 and/or flow of water through the heater 12 or to detect if there is flow through the heater 12 and send the detected information wirelessly to the router 4. It may be an advantage that the sensor S2 only is required to detect if there is flow through the heater 12 or if there is change in the flow through the heater 12 since it then would be possible to apply a simple, cheap sensor S2 in the system 2.

(28) The third sensor S3 is attached to the power cable 26 of an air conditioner 24. The third sensor S3 is configured to detect the temperature of the heater 12 and/or flow of water through the heater 12 and send the detected information wirelessly to the router 4. It may be an advantage that the sensor S3 only is required to detect if there is power at the cable 26. In this way it would be possible to use a simple a cheap sensor S3 in the system 2.

(29) The air conditioner 24 includes a control unit 22 provided on the top of the air conditioner 24. The control unit 22 is capable of regulating the activity of the air conditioner 24. The control unit 22 is controlled wirelessly by the router 4. This means that the router 4 can switch off turn on, and regulate the activity of the air conditioner 24 by communicating wirelessly with the control unit 22.

(30) The system 2 also includes a sensor S4 that not is attached to any particular device. The sensor S4 may be a temperature sensor place in any suitable location, e.g., in order to detect the room temperature, by way of example. The sensor S4 may also be a lux sensor configured to measure the inflow of light, since this information may be important when deciding whether to turn on the heater 12 or not.

(31) The system 2 includes a cloud service 8 including a data storage 36 that can be used to store information received from the router 4 or data modified or calculated by the cloud service 8. The cloud service 8 also includes a mathematical statistic model 38 that is configured to estimate and/or forecast the activity (flow of water, power, or heat energy) of one or more of the devices 16, 24, 12 of the system. Any suitable mathematical statistic model 38 may be applied and the mathematical statistic model 38 may conduct inference on combined data from one or more of the sensors S1, S2, S3, S4 and from one or more of the meters M1, M2, M3 to estimate and/or forecast the activity of one or more of the devices 16, 24, 12.

(32) The user interface 10 that is a smartphone 10 communicates wirelessly with the router 4. Hereby the user of the system 2 can have access to information about the estimated or predicted resource flow of at least a selection of the devices 16, 12, 24. The smartphone 10 may also be used to control the control units 22, 40, 20 through the router 4 or to change settings in the router 4 (e.g., upload new structures or parameters to the mathematical statistic model 38).

(33) The system 2 may be used to generate an alert and shut down or regulate the activity of the any of the devices 16, 24, 12, e.g., in order to reduce the total energy flow. By way of example, the system 2 will automatically detect if the heater 12 and the air conditioner 24 are activated at the same time. The system 2 will then either switch off the heater 12 (in case the room temperature is within a predefined comfort zone). It the room temperature is higher than the preferred room temperature, the system 2 may increase the activity of the air conditioner 24. K on the other hand, the room temperature is lower than the preferred room temperature, the system may switch off the air conditioner 24 and turn on the heater 12.

(34) FIG. 3 illustrates a schematic view of a system 2 that is almost similar to the system 2 illustrated in FIG. 2. The wireless router 4 is, however, replaced with a user interface 10 in the form of a smartphone 10. The smartphone 10 is configured to communicate wirelessly with a group of sensors S1, S2, S3, S4, meters M1, M2, M3, a cloud service 8, and three control units 20, 22, 40. The function of the system 2 corresponds to the function of the system 2 illustrated in FIG. 2. It is important to point out that the user interface 10 may be any other type of suitable user interface, e.g., a laptop.

(35) FIG. 4 illustrates a schematic view of the air conditioner 24 illustrated in FIG. 2 and in FIG. 3. FIG. 4a) is a schematic close-up cross sectional view of a sensor S1 attached the power cable 26 of the air conditioner 24. FIG. 4b) is a perspective view of the air conditioner 24.

(36) In FIG. 4b) it can be seen that the air conditioner 24 has a power cable 26 to which a sensor S1 is attached. The sensor S1 is configured to send a wireless signal 34 to a router or a user interface (not shown).

(37) In FIG. 4a) it can be seen that the sensor S1 includes an energy harvester 28, arranged next to a centrally arranged power manager 30 that is arranged next to a wireless transducer 32. The energy harvester 28 may be any suitable type of energy harvester. The power manager 30 may be any unit capable of generating a signal 34 that can be send by using the wireless transducer 32.

(38) It may be an advantage that the harvested energy is used to detect activity of the air conditioner 24. It would then be possible to provide a sensor S1 capable of generating a wireless signal 32 when energy is harvested. When the air conditioner 24 is activated the power cable 26 will be current-carrying and thus energy will be harvested.

(39) The sensor S1 may, however, be provided with a battery and or a number of capacitors or any other suitable energy storage if necessary.

(40) The system 2 according to principles of the present invention may be used to detect early warning of emerging problems (e.g., leakage of water in a pipe network or emerging disease in an animal herd).

(41) The system 2 may also be used to detect overconsumption of resources, e.g., if a heater 12 and an air conditioner are activated at the same time. In this case the system 2 may be configured to take care of the situation and switch off or regulate the activity of at least one of the devices.

(42) The system 2 may be used to provide general information about the actual flows of resources over time and to generate alarms in case of sudden failures. Besides the system may be used to estimate or predict future resource flow of the devices.

(43) The system 2 may also be used to simulate the future resource flow of a new constellation of devices (e.g., if a company considers to increase the number of devices due to an increased production capacity demand).

(44) The system 2 may be used to benchmarking similar devices, to perform service surveillance on technical devises, and to allow for remote surveillance.

(45) The sensors S1, S2, S3, S4 and the meters M1, M2, M3 may each contain a unique identifier (e.g., number, signature, description, or ID). In this way the router 4 or the user interface 10 is capable of coupling each incoming wireless signal 34 with a specific device. In other words, the use of a unique identifier secures that the wireless signal 34 can be used in an optimum way by the route 4 or the user interface 10.

(46) The sensors S1, S2, S3, S4 may be configured to detect when the devices D1, D2, D3, 12, 16, 24 are activated (when the devices are switched on or when there is a flow of a resource, e.g., water, gas, or current, through the device).

(47) FIG. 5 illustrates a portion of a system 2 according to principles of the present invention. FIG. 5 illustrates how a system 2 can be used to measure water flow in a main pipe 14 and in a number of smaller distribution pipes 16.

(48) The water supplied to the installation is metered by one main meter M1 that provides measurements of high precision (e.g., approved by authorities or within standards for flow measurements). The main pipe 14 is in fluid communication with a plurality of smaller distribution pipes 16, while a number of sensors S1, S2 and S3 are attached to the outside of some of these smaller distribution pipes 16, 16, 16.

(49) In the following, we take a look at the metering at sensor S1. The sensor S1 is a piezoelectric sensor that includes simple piezo elements. Thus, the sensor signal changes with the flow in the pipe 16. The signal from the sensor S1 depends, however, on the material and the dimension of the pipe 16, the sensor attachment, and the medium (in this example water). Accordingly, and the sensor signal must be transformed or adjusted/calibrated to give the actual flow (the flow that would be achieved by using a high precision meter). In a prior art system it would be necessary to apply a high quality sensor configured to measure flow in the specific environment (the dimensions of the pipe 16, the material of the pipe 16, and attachment type).

(50) Since the present invention applies a mathematical statistic model it is possible to use a simple sensor S1 to determine the flow at the pipe 16. The use of a mathematical statistic model makes it possible to compensate for the low precision of the sensor S1.

(51) The sensor S1 is attached to the outside of the pipe 16 and hereafter it is possible to sample a series of measurements from the sensor S1, over a timespan.

(52) In the same timespan a series of measurements is sampled from the main meter M1 in the same timespan.

(53) The mathematical statistic model is used to estimate the part of the main meter flow that flows through the pipe 16 having the sensor S1 attached to it. It is possible to carry out a time-dependent transformation of the measurements of sensor S1 in order to correct for the actual environment of the sensor S1.

(54) A less computer-demanding approach can be obtained by using a mathematical statistic model to estimate if a time-independent transformation can be used.

(55) It may be an advantage to use the time-independent transformation, rather than the full mathematical statistic model, on the running measurements from the sensor S1.

(56) It may be beneficial from time to time, to test if the precision of the transformation is acceptable, and update it if needed, using the mathematical statistic model.

(57) The necessary amount of samples needed can be obtained by estimating the measurement error of the transformation. In this manner, the results can be detained until the desired precision is obtained.

(58) In the following, one preferred method to estimate the latent stochastic process using a mathematical statistic model 38 is described. The latent stochastic process can be modelled by, e.g., a state space model defined by
Y.sub.t=F.sub.t.sub.t+.sub.t .sub.tN(0,V.sub.t)(1)
.sub.t=G.sub.t.sub.t-1+.sub.t .sub.tN(0,W.sub.t)(2)

(59) where Y.sub.t is a vector determines (e.g., describing, or defining) the observed process at time t, includes observed data from the sensor (S.sub.1, S.sub.2, . . . , S.sub.n) and/or the meters M.sub.1, M.sub.2, . . . , M.sub.k; .sub.t is a vector determine the latent stochastic process at time t, including latent process data, such as percent ventilation efficiency, percent overconsumption of resources, resource flow, device health status, etc; F.sub.t is the regression matrix which determine the linear relation between the latent process and the observed process at time t; G.sub.t is the evolve matrix which determine the linear transition from time t1 to time t in the latent process; .sub.t and .sub.t are zero mean multivariate Gaussian distributed noise vectors of the observed process and the latent process respectively; V.sub.t is the observation variance-covariance matrix; and W.sub.t is the evolution variance-covariance matrix.

(60) The model parameter matrices F.sub.t and G.sub.t may be estimated by, e.g., the Kalman filter, using prior data from the modelled system and/or similar systems, hereunder the said data provided by the user and/or experts in the field. Standard statistical methods can be used to conduct inference (e.g., estimate information) on the process. The information can be, e.g., an estimated signal (e.g., trend) and/or forecasts (e.g., prognosis) of the process, and the related distributions of the estimates, variance and/or confidence intervals. Using these kinds of estimates, it is easy to, e.g., raise warnings and/or alarms. For example an alarm can be chosen to appear if the probability of an observed deviation in the process is less than 0.1% probable to occur by change.

(61) The above model framework is a special case of the more general model framework
Y.sub.t=f.sub.t(.sub.t)+.sub.t .sub.t.sub.1(V.sub.t)(3)
.sub.t=g.sub.t(.sub.t-1)+.sub.t .sub.t.sub.2(W.sub.t)(4)

(62) where f.sub.t and g.sub.t are general functions, .sub.1 and .sub.2 are general statistical distributions, and all other terms are as described above.

(63) Inference on this more general model framework can be conducted by, e.g., the extended Kalman filter in cases where the relation between the latent process and the observed process is non-linear, and the Kalman-Bucy filter in cases where the time is defined (e.g., described) on a continuous scale.

(64) Other time series analysis method and/or multivariate data-analysis methods, such as Analysis of Variance (ANOVA), Markov models, Generalized Linear Models (GLM), and Multivariate Gaussian Models may as well be used to estimate the said latent stochastic process, and infer the said information.

LIST OF REFERENCE NUMERALS

(65) 2 System 4 Router 6 Control unit 8 Cloud service 10 User interface D.sub.1, D.sub.2, D.sub.3 Device S.sub.1, S.sub.2, S.sub.3, S.sub.4, . . . , S.sub.n Sensor M.sub.1, M.sub.2, M.sub.3 Meter 12 Heater 14 Main pipe 16, 16, 16 Pipe 18 Power switch 20 Valve 22 Control unit 24 Air conditioner 26 Cable 28 Energy harvester 30 Power manager 32 Wireless transducer 34 Wireless signal 36 Data storage 38 Mathematical statistical model 40 Valve

(66) While the invention has been described in detail with reference to exemplary embodiments thereof, it will be apparent to one skilled in the art that various changes can be made, and equivalents employed, without departing from the scope of the invention. The foregoing description of the preferred embodiments of the invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed, and modifications and variations are possible in light of the above teachings or may be acquired from practice of the invention. The embodiments were chosen and described in order to explain the principles of the invention and its practical application to enable one skilled in the art to utilize the invention in various embodiments as are suited to the particular use contemplated. It is intended that the scope of the invention be defined by the claims appended hereto, and their equivalents. The entirety of each of the aforementioned documents is incorporated by reference herein.