NL2020228B1

NL2020228B1 - Detecting inefficient appliances

Info

Publication number: NL2020228B1
Application number: NL2020228A
Authority: NL
Inventors: Basu Kaustav; Toma Tudor; John Galsworthy Stephen
Original assignee: Quby B V
Priority date: 2018-01-03
Filing date: 2018-01-03
Publication date: 2019-07-10
Also published as: WO2019134861A1; EP3735759A1

Abstract

Method for detecting inefficient appliances, comprising: receiving (5 1) resource consumption data (10) from a resource consumption meter (1), wherein appliances (11— 13) are connected to the meter, the resource consumption data comprising information about aggregated resource consumption of the appliances in time; detecting (52a) a use of and obtaining (52b) resource consumption characteristics (520) of one appliance from the resource consumption data; determining (53) a characteristic (530) of a component (111— 115) of the appliance for a plurality of moments in time from the resource consumption characteristics; calculating (54) a representative characteristic (540) for the component based on the determined characteristic of the component; and comparing (55) the representative characteristic With further representative characteristics (20, 30) for similar components of similar appliances connected to further resource consumption meters (2, 3) to obtain a comparison result (550), the comparison result being indicative of an efficiency of the appliance.

Description

TECHNICAL FIELD [0001 ] The present invention relates to a computer-implemented method for detecting inefficient appliances. More specifically, the invention relates to the detection of inefficient appliances or inefficient use of appliances regarding the consumption of resources such as electricity, water, natural gas and/or natural oil.

BACKGROUND ART [0002] In a typical residential, commercial or tertiary building there can be various appliances present with different resource consumption patterns. Depending on the appliance, resources such as electricity, water, natural gas or natural oil may be used by the appliance during operation.

[0003] For environmental and financial reasons there is a strong focus on resource efficiency of appliances. A resource inefficient appliance - this is typically an appliance that uses more than average resources compared to other, similar appliances - may be considered undesirable and replaced by another more efficient appliance. Inefficient use of resources may also indicate a defect in an appliance, which may justify replacing the appliance, or an inefficient operation of the appliance by the end user, which may be correctable by changing the operation of the appliance.

[0004] The detection of inefficient appliances typically involves monitoring the use of resources at the appliance. An appliance may for example be plugged into a wall socket via an energy meter to measure the electricity consumption over time. The measurement results may be collected in a data file and compared to other measurements results from other, similar appliances. Another example may involve a water consumption meter attached to a water supply of a water consuming appliance. Again, the measurements results may be collected and compared to other, similar water consuming appliances.

[0005] hi a building including many appliances is can be cumbersome and expensive to monitor the resource consumption of all appliances individually.

[0006] There is a need for a solution to detect inefficient appliances in buildings, without having to monitor the appliances individually.

-2SUMMARY [0007] The present invention provides a solution for detecting inefficient appliances, without having to monitor the appliances individually.

[0008] According to an aspect of the invention, a computer-implemented method for detecting inefficient appliances is proposed. The method can comprise receiving resource consumption data from a resource consumption meter for a resource being one or more of electricity, water, natural gas and natural oil. One or more appliances can be connected to the resource consumption meter. The resource consumption data can comprise information about aggregated resource consumption of the one or more appliances in time. The method can further comprise detecting a use of an appliance from the resource consumption data and obtaining resource consumption characteristics of the appliance from the resource consumption data. Herein, the appliance is one of the one or more appliances. The method can further comprise determining a characteristic of a component of the appliance for a plurality of moments in time of the use of the appliance from the resource consumption characteristics. The method can further comprise calculating a representative characteristic for the component based on the determined characteristic of the component of the appliance for the plurality of moments in time. The method can further comprise comparing the representative characteristic with one or more further representative characteristics for similar components of similar appliances connected to one or more further resource consumption meters to obtain a comparison result. The comparison result can be indicative of an efficiency of the appliance.

[0009] An appliance may be considered inefficient if its resource consumption is above an expected or average consumption compared to other, similar appliances. An appliance may alternatively be considered inefficient if it is being operated or used inefficiently by an end user, compared to other, similar appliances. In the determination of the inefficiency of the appliance, a characteristic of a component of the appliance is determined and further analyzed.

[0010] In the determination of the inefficiency of the appliance, the representative characteristic for the component of the appliance can be compared with one or more further representative characteristics for similar components of similar appliances connected to one or more further resource consumption meters. It will be understood that these further representative characteristics may be and are in fact preferably obtained in a manner similar to the representative characteristic as defined in claim 1.

-3[0011] The resource consumption data is typically measured by the resource consumption meter. Such resource consumption meter is typically present in a building for measuring the overall resource consumption of appliances in the building and enabling charging of the resource consumption.

[0012] In case one appliance is connected to the resource consumption meter, the information about the aggregated resource consumption includes resource consumption data for the one appliance. In case more than one appliances are connected to the resource consumption meter, the information about the aggregated resource consumption typically includes the sum of resource consumption of all the appliances. Typically the resource consumption meter polls the aggregated resource consumption of the appliances at a regular time interval, resulting in a data set including the resource consumption at different moments in time. Thus, information about aggregated resource consumption of the one or more appliances in time can be collected by the resource consumption meter.

[0013] The resource consumption data is typically transmitted to a remote server for further processing. Hence, the steps of the method are typically performed by a server remote from the resource consumption meter. The resource consumption meter may be communicatively connected to the remote server via the Internet or any other data network.

[0014] A component of the appliance as defined in claim 1 may be a single component or may include further sub-components. The characteristic of the component may include a specific state of the component, such as an on/off state, state of movement, state of heating, etcetera.

[0015] The representative characteristic for the component can be calculated from the obtained data set with the determined characteristic at multiple moments in time. The representative characteristic is the calculated characteristic that is representative for the component. In the calculation of the representative characteristic the multiple data values for the multiple moments in time may be analyzed, filtered to remove incorrect or potentially unreliable data values, averaged, and etcetera.

[0016] Advantageously, the inefficiency of a single appliance can thus be detected from resource consumption data containing aggregated information of multiple or all appliances connected to the resource consumption meter. Measuring the resource consumption of the individual appliance is thus not needed. Moreover, a specific component within the appliance may be pinpointed that contributes to the inefficiency.

-4[0017] In an embodiment wherein the resource consumption data comprises information about aggregated resource consumption of more than one appliances in time, the detecting of the use of the appliance from the resource consumption data can comprise the step of applying a disaggregation algorithm to the detect the resource consumption characteristics of the appliance in the resource consumption data. This enables reliable detection of individual appliances from the aggregated information. [0018] In an embodiment multiple characteristics of the component can be determined. The method can further comprise applying a mixture model to determine a distinguished characteristic from the multiple characteristics of the component. The representative characteristic can be calculated for the distinguished characteristic. This advantageously enables a component that is used in different use cases to be analyzed for the particular use case. A non-limiting example of such component is a heater element of a washing machine that may be used for a 30°C short washing cycle or for a 60°C longer washing cycle. The distinguished characteristic may then be related to one of 30°C short washing cycle and 60°C longer washing cycle, for which the representative characteristic can be calculated. The determination of the inefficiency of the appliance can thus be related to the particular use case of the appliance.

[0019] In an embodiment, the resource can be electricity. The detecting of the use of the appliance can be based on a non-intrusive appliance load monitoring technique. This advantageously enables accurate detection of a single appliance from the resource consumption data in case of multiple appliances being monitored with a single electricity consumption meter.

[0020] In an embodiment, the characteristic of the component can be determined based on the non-intrusive appliance load monitoring technique. This advantageously enables accurate detection and determination of a characteristic of a component in the appliance from the resource consumption data in case of multiple appliances being monitored with a single electricity consumption meter.

[0021] In an embodiment, the appliance can be one of: a refrigerator; a dishwasher; a washing machine; a tumble dryer; a hot water tap; a boiler; a central heating boiler; an air conditioning system; an electrical heater; an oven. These are appliances that typically consume large amounts of resources and are thus most interesting to analyze for inefficiency. It will be understood that the invention is not limited to these appliances.

-5[0022] In and embodiment, the component can be one of: a heater component; a heating block; a drum; a compressor block; a component involved in rinsing cycles. These are components that typically consume large amounts of resources and are thus most interesting to analyze for inefficiency. It will be understood that the invention is not limited to these components.

[0023] In an embodiment, the characteristic can comprise at least one of: a duration of use of the component; an average power consumed by the component; a maximum power consumed by the component; a temperature of the component.

[0024] In an embodiment, the calculating of the representative characteristic for the component can be further based on thermostat data collected by a thermostat device. The thermostat data can comprise temperature measurement data and/or boiler control data. Thus, a secondary source of information, namely data from the thermostat device, may be used to fine tune or aid in the determination of the inefficiency of the appliance. Temperature measurement data may include data about when, how, under what conditions and/or how long a temperature is being controlled in the surroundings of the thermostat device. Boiler control data may include data about when and/or how long a heater boiler of a central heating system is being activated by the thermostat device. [0025] In an embodiment, the calculating of the representative characteristic for the component can be further based on user data collected via a questionnaire. Thus, a further source of information, namely the date provided by the user, may be used to fine tune or aid in the determination of the inefficiency of the appliance. Examples of how a questionnaire may be implemented are a form on a web page or an app on smartphone or tablet.

[0026] According to an aspect of the invention, a data processing system is proposed comprising a processor configured to perform one or more of the above described steps of the method.

[0027] According to an aspect of the invention, a computer program product is proposed, which is typically implemented on a computer-readable non-transitory storage medium. The computer program product comprises computer executable instructions which, when executed by a processor, cause the processor to carry out one or more of the above described steps of the method.

[0028] According to an aspect of the invention a computer-readable non-transitory storage medium is proposed, comprising computer executable instructions which, when

-6executed by a processor, cause the processor to carry out one or more of the above described steps of the method .

[0029] Hereinafter, embodiments of the disclosure will be described in further detail. It should be appreciated, however, that these embodiments may not be construed as limiting the scope of protection for the present disclosure.

BRIEF DESCRIPTION OF DRAWINGS [0030] Embodiments will now be described, by way of example only, with reference to the accompanying schematic drawings in which corresponding reference symbols indicate corresponding parts, and in which:

[0031] FIG. 1 shows a network architecture of an exemplary embodiment of the invention;

[0032] FIG. 2 shows an example of an appliance and internal components;

[0033] FIG. 3 shows a flow chart of an exemplary method for detecting inefficient appliances;

[0034] FIG. 4 shows an exemplary graph of resource consumption data and the result of a non-intrusive appliance load monitoring technique to detect the use of appliances and components;

[0035] FIG. 5 shows another exemplary graph of resource consumption data over time;

[0036] FIG.6 shows an exemplary graph of a characteristic of a component of one appliance for a plurality moments in time; and [0037] FIG.7 shows an exemplary graph indicating a representative characteristic for a component.

[0038] The figures are meant for illustrative purposes only, and do not serve as restriction of the scope or the protection as laid down by the claims.

DESCRIPTION OF EMBODIMENTS [0039] In FIG. 1 a distinction is made between solid lines and dashed lines. Solid lines in between appliances indicate the use of the same resource. Dashed lines are indicative of data connections for the transmission of data. Data is indicated in between brackets. The arrows next to the data indicate a typical direction of the data, but this is not to be construed as the data connections being uni-directional.

[0040] FIG. 1 shows a building 14 that houses a number of resource consuming appliances 11-13. The building 14 is to be understood as an exemplary location of the

-7appliances; the appliances may be located at any location. In this example the appliances 11-13 are electricity consuming appliances, but the invention is similarly applicably to appliances using other resources, such as water, natural gas and natural oil. The appliances 11-13 are connected to a resource consumption meter, in this example an electricity meter 1 that measures the combined electricity consumption of the appliances 11-13 over time. Another group of appliances 21-26 are connected to a second electricity meter 2 and may be located within the same building 14 or at any other location. A further appliance 31 is connected to a third electricity meter 3, which may be located within the same building 14 or at any other location.

[0041] The electricity meters 1, 2, 3 may have a build-in modem or may be connected to a modem for transmitting resource consumption data 10, 20, 30 to a remote server 5. Typically the resource consumption data 10, 20, 30 will be transmitted via the Internet 8.

[0042] In the example of FIG. 1, the resource consumption data 10 measured by the electricity meter 1 includes the electricity consumption of appliances 11-13 over time. The resource consumption data 20 measured by the electricity meter 2 includes the electricity consumption of appliances 21-26 over time. The resource consumption data 30 measured by the electricity meter 3 includes the electricity usage of appliance 31 over time.

[0043] The resource consumption data may be analyzed by the server 5 to detect the presence of an inefficient appliance or an inefficient use of an appliance. This analysis is typically performed for appliances connected to one resource consumption meter and may be repeated for other resource consumption meters.

[0044] The electricity consumption data 10 may thus be analyzed by the server 5. FIG. 3 shows a flow chart of steps that may be followed in this analysis. The flow chart begins with receiving the electricity consumption data 10 from the electricity meter 1. [0045] From the electricity consumption data, which includes the aggregated electricity consumption over time of all appliances 11-13, a use of a single appliance is detected in step 52a and resource consumption characteristics 520 for this appliance may be obtained in step 52b. Herein, steps 52a and 52b maybe performed in a single step and may be implemented by applying a disaggregation algorithm 52c, as will be explained in more detail with FIG. 4. The single appliance of which the resource consumption

-8characteristics 520 may thus be obtained is for example a washing machine 11, such as shown in FIG. 2.

[0046] With the aggregation of the electricity consumption in the resource consumption data 10, if multiple appliances 11-13 are operating at the same time it may not always be possible to detect the use of a single appliance such as washing machine 11 from the aggregated consumption data. And even if the use of the single appliance may be detectable, then a further analysis of the use of components within the appliance may not always be possible. Therefore, as illustrated in FIG. 5, the resource consumption data, here visualized as electricity consumption data 10b, may be taken for a long period of time, e.g. in the order of months or any other time frame that provides enough data to discard some of the usages of the washing machine 11. In FIG. 5 a use of the washing machine 11 has been detected three times 531, 532, 533. Two more instances in time the washing machine has been used, but this use is undetectable - indicate as 1 la - from the electricity consumption data 10b.

[0047] The resource consumption characteristics 520 may be further analyzed in step 53 to determine a characteristic 530 of a component of the washing machine 11. Herein, the same disaggregation algorithm 52c may be used to detect the use of the component of the appliance. Mathematically, steps 52a-c and 53 may be combines into a single step or otherwise combined.

[0048] In FIG. 6 an example is shown for a component in the form of a heater block

111. Other components of the washing machine 11 may be analyzed in a similar way, such as a motor 112 for turning the drum, a drain pump 113, a door lock 114, control electronics 115, or any other electrical component. For the heater block 111 the power usage and duration of being powered on is plotted for the detected moments in time 531533 as shown in FIG. 5 and further moments in time 534-537 not shown in FIG. 5. Herein, the power versus duration represents a characteristic of the heater block 111. [0049] The combined set of data points 531-537 representing the characteristic of power versus duration may be jointly referred to as a characteristic 530 of the heater block component 111 of the washing machine appliance 11 for a plurality moments in time. From the determined characteristic 530 of the component a representative characteristic 540 may be calculated in step 54. In FIG. 7, for the data points 531, 532, 534, 535 and 536 a representative characteristic 541 is determined in the form of an area covering power versus duration values that is representative for the heater component

111. Any suitable mathematical algorithm may be used in the calculation of the representative characteristic, which typically results in an average or most relevant data area of the characteristic of the component.

[0050] Multiple characteristics of a component may be determined, in which case the representative characteristics may be distinguished, for example by applying a mixture model. In the example of FIGs. 6 and 7 two representative characteristics are shown. The data point 533 and 537, after applying a Gaussian mixture model, turned out to form a further representative characteristic 542. It will be understood that in practice the number of data points will be much larger than shown in FIG. 6, enabling the mixture model to determine a relative occurrence of different characteristics. In this example the distinguished representative characteristic 541 may be the power versus duration of the heating block 111 for a 30°C short duration wash and the representative characteristic 542 may be the power versus duration of the heating block 111 for a 60°C long duration wash.

[0051] An analysis as shown in FIGs. 3, 6 and 7 may be performed for various components and for multiple characteristics per component, based on the same resource consumption data 10.

[0052] The representative characteristic 540, such as the representative power versus duration characteristic 541 of the heater block 111 of the washing machine 11, may be compared with one or more further representative characteristics of similar components in similar other appliances. For example a washing machine 21 may be detected from the electricity consumption data 20 similar to the analysis of the electricity consumption data

10. Appliance 31 may also be a washing machine, which characteristics may be obtained from electricity consumption data 30. By comparing the representative characteristics of the similar components of the similar appliances, it may be detected that one of the components and thus one of the appliances is operating less efficient compared to the other components and appliances. The comparison result 550 may thus be indicative of an inefficiency of the washing machine 11 in case for example the representative power versus duration of the heater element 111 is worse than that of similar washing machines in use elsewhere.

[0053] Such comparison result 550 is particular relevant for appliances that consume relatively large amounts of resources, such as a refrigerator, a dishwasher, a washing machine (11), a tumble dryer, a hot water tap, a boiler, a central heating boiler, an air

-10conditioning system, an electrical heater, an oven, and etcetera. In these appliances, typically those components that are known to consume relatively large amounts of resources may be most relevant to analyze, such as a heater component, a heating block (111), a drum, a compressor block, a component involved in rinsing cycle, and etcetera. Various characteristics of components may thus be analyzed, such as a duration of use of the component, an average power consumed by the component, a maximum power consumed by the component, a temperature of the component, and etcetera, or any combination thereof.

[0054] FIG. 4 shows a more detailed example of resource consumption data 10a, in this example electricity consumption data 10a. Individual uses of appliances and components thereof may be detected from the electricity consumption data 10a by applying 52c a disaggregation algorithm. In case of the resource being electricity, a nonlimiting example of such disaggregation algorithm is non-intrusive appliance load monitoring (NIALM).

[0055] NIALM is a process for analyzing changes in the voltage and/or current from the electricity consumption data and deducing what appliances or components in the appliances are used as well as their individual energy consumption. This known technique may result in the detection of e.g. on and off events of appliances or components within the electricity consumption data 10a. In FIG. 4, on events are indicated by arrows pointing up and off events are indicated by arrows pointing down. The following appliances may thus be detected: a refrigerator 221 turning on and off two times, an oven element 231 turning on and off two times, and a stove burner 241 turning on and off two times.

[0056] The NIALM technique may be enhanced in various manners. For example, instead of just looking for on/off transactions within the resource consumption data, the various power levels or states at which appliances operate and also the temporal correlation among them may be taken into account. Furthermore, NIALM may make use of pre-measurement data from individual appliances which may be available as a stored data set within the server 5. This pre-stored data of individual appliances may be used to correlate with patterns in the resource consumption data to find a use of a particular appliance in the resource consumption data.

[0057] In an exemplary embodiment an enhanced NIALM based technique may be used, wherein a sequencing technique is used to group similar characteristics (power,

-11duration (ON/OFF), frequency). The algorithm may use a Gaussian Mixture Model (GMM) to find a convergence of grouped features. The features are evaluated from the grouped characteristics. The Maximum Likelihood Estimate (MLE) for each of the representative characteristics may be used to determine appliance-wise inefficiency. The threshold of energy inefficiency may be determined from the European label standard for each appliance. The inefficiency can be both due to inefficiency in device or in usage (e.g. high temperature washing).

[0058] The calculation 54 of the representative characteristic 540 may be based on thermostat data 70 collected by a thermostat device 7. The thermostat data may provide further data related to e.g. room temperature and/or boiler control, which, in combination with the determined characteristic 530 of a component of an appliance, may improve the accuracy of the calculation 54 in case of any interaction between the appliance and the thermostat device or the room or appliances controlled by the thermostat device.

[0059] The calculation 54 of the representative characteristic 540 may be based on user data 60 collected via a questionnaire, e.g. via a data form on a tablet 6 or any other means resulting in the user data. The user data may provide further data related to e.g. typical use scenarios of appliances by the user, which, in combination with the determined characteristic 530 of a component of an appliance, may improve the accuracy of the calculation 54 in case the questionnaire addresses the appliance.

[0060] One or more embodiments of the disclosure may be implemented as a computer program product for use with a computer system. The program(s) of the program product may define functions of the embodiments (including the methods described herein) and can be contained on a variety of computer-readable storage media. The computer-readable storage media may be non-transitory storage media. Illustrative computer-readable storage media include, but are not limited to: (i) non-writable storage media (e.g., read-only memory devices within a computer such as CD-ROM disks readable by a CD-ROM drive, ROM chips or any type of solid-state non-volatile semiconductor memory) on which information may be permanently stored; and (ii) writable storage media (e.g., hard disk drive or any type of solid-state random-access semiconductor memory, flash memory) on which alterable information may be stored.

Claims

CONCLUSIONS

A computer-implemented method for detecting inefficient household appliances, comprising:

receiving (51) data (10) on consumption of a resource from a resource consumption meter (1) for a resource being one or more of electricity, water, natural gas and natural oil, one or more household appliances (11-13) being connected to the resource consumption meter, and wherein the resource consumption data includes information about aggregated resource consumption of the one or more household appliances over time;

detecting (52a) a usage of a household appliance (11-13) from the consumption data of a resource and obtaining (52b) characteristics (520) of consumption of a resource of the household appliance from the consumption data of a resource resource, wherein the household appliance is one of the one or more household appliances;

determining (53) a characteristic (530) of a component (111-115) of the household appliance for multiple moments in time of use of the household appliance from the consumption characteristics of a resource;

calculating (54) a representative characteristic (540) for the component based on the determined characteristic of the component of the household appliance for the multiple moments in time; and comparing (55) the representative characteristic with one or more further representative characteristics (20, 30) for comparable components of comparable household appliances connected to one or more further resource consumption meters (2, 3) to obtain a comparison result (550), wherein the comparison result is indicative of an efficiency of the household appliance.

The method of claim 1, wherein, in case the resource consumption data includes information about aggregated resource consumption of more than one home appliance over time, detecting the usage of the home appliance from the data of consumption of a resource resource includes applying (52c) a disaggregation algorithm to detect the consumption characteristics of a resource of the home appliance in the resource consumption data.

The method of claim 1 or claim 2, wherein multiple characteristics of the component are determined, the method further comprising applying a mixture model to determine a distinct characteristic (541, 542) from the multiple characteristics of the component, and wherein the representative characteristic is calculated for the distinguished characteristic.

The method of any one of claims 1 to 3, wherein the resource is electricity, and wherein detecting usage of the household appliance is based on a non-intrusive technique for monitoring load on the household appliance intrusive appliance load monitoring technique ”).

The method of claim 4, wherein the characteristic of the component is determined based on the non-intrusive technique for monitoring load by the household appliance.

The method of any one of claims 1 to 5, wherein the household appliance is one of: a refrigerator; a dishwasher; a washing machine (11); a dryer; a hot water tap; a boiler; a central heating boiler; an air conditioning system; an electric heater; an oven.

The method of any one of claims 1-6, wherein the component is one of: a heating component; a heating block (111); a drum; a compressor block; a component that is involved in flushing cycles.

The method of any one of claims 1-7, wherein the characteristic comprises at least one of: a service life of the component; an average power consumed by the component; a maximum power consumed by the component; a temperature of the component.

The method of any one of claims 1-8, wherein calculating the representative characteristic for the component is further based on thermostat data (70) collected by a thermostat device (7), wherein the thermostat data comprises at least one of temperature measurement data and boiler control data.

The method of any one of claims 1-9, wherein calculating the representative characteristic for the component is further based on user data (60) collected via a questionnaire.

A data processing system (5) comprising a processor adapted to perform the steps of the method according to any of claims 1-10.

A computer program product implemented on a computer-readable non-temporary storage medium, the computer program product comprising computer-executable instructions which, when executed by a processor, cause the processor to perform the steps of the method according to any of claims 1-10.

A computer-readable non-temporary storage medium comprising computer-executable instructions which, when executed by a processor, cause the processor to perform the steps of the method of any one of claims 1-10.

1/5