WO2020084299A1 - Procédé de détection d'un état de fonctionnement d'un appareil - Google Patents

Procédé de détection d'un état de fonctionnement d'un appareil Download PDF

Info

Publication number
WO2020084299A1
WO2020084299A1 PCT/GB2019/053011 GB2019053011W WO2020084299A1 WO 2020084299 A1 WO2020084299 A1 WO 2020084299A1 GB 2019053011 W GB2019053011 W GB 2019053011W WO 2020084299 A1 WO2020084299 A1 WO 2020084299A1
Authority
WO
WIPO (PCT)
Prior art keywords
appliance
data
measures
power consumption
determining
Prior art date
Application number
PCT/GB2019/053011
Other languages
English (en)
Inventor
Neal Coady
Sean STEPHENSON
Phuong Pham
Nicholas O'MALLEY
Original Assignee
Centrica Plc
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Centrica Plc filed Critical Centrica Plc
Publication of WO2020084299A1 publication Critical patent/WO2020084299A1/fr

Links

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B15/00—Systems controlled by a computer
    • G05B15/02—Systems controlled by a computer electric
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F25—REFRIGERATION OR COOLING; COMBINED HEATING AND REFRIGERATION SYSTEMS; HEAT PUMP SYSTEMS; MANUFACTURE OR STORAGE OF ICE; LIQUEFACTION SOLIDIFICATION OF GASES
    • F25D—REFRIGERATORS; COLD ROOMS; ICE-BOXES; COOLING OR FREEZING APPARATUS NOT OTHERWISE PROVIDED FOR
    • F25D29/00—Arrangement or mounting of control or safety devices
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B17/00—Systems involving the use of models or simulators of said systems
    • G05B17/02—Systems involving the use of models or simulators of said systems electric
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring
    • G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
    • G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
    • G05B23/0224—Process history based detection method, e.g. whereby history implies the availability of large amounts of data
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
    • H04L12/2803—Home automation networks
    • H04L12/2823—Reporting information sensed by appliance or service execution status of appliance services in a home automation network
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F25—REFRIGERATION OR COOLING; COMBINED HEATING AND REFRIGERATION SYSTEMS; HEAT PUMP SYSTEMS; MANUFACTURE OR STORAGE OF ICE; LIQUEFACTION SOLIDIFICATION OF GASES
    • F25B—REFRIGERATION MACHINES, PLANTS OR SYSTEMS; COMBINED HEATING AND REFRIGERATION SYSTEMS; HEAT PUMP SYSTEMS
    • F25B2700/00—Sensing or detecting of parameters; Sensors therefor
    • F25B2700/15—Power, e.g. by voltage or current

Definitions

  • the present invention relates to a method and system for identifying the operating state of a compressor based appliance from power series data.
  • the invention also relates to methods for detecting faults in the operation of the power consuming device.
  • the power consumption of a domestic appliance can be measured via a device installed between the plug of the appliance and the wall socket or by disaggregating an energy consumption signature measured for the whole house, e.g. using a smart meter. Such information is commonly used to understand a household’s energy consumption in detail.
  • consumption data can also be used to identify the operation of an appliance (e.g. to detect when a washing machine is operating).
  • appliances are quite unpredictable in their energy consumption, making it difficult to exploit the available consumption data.
  • compressor-based appliances such as refrigeration appliances, commonly include a compressor control processor which controls the repeated cycling of the compressor. These compressor cycles generally do not follow a fixed pattern of power consumption but rather they vary with operation over time in order to maintain a set temperature.
  • the compressors of compressor-based appliances are not activated by the user and do not follow a fixed cycle of operation but instead come on and turn off as required, automatically. This makes it more difficult to extract useful information from their power consumption signatures.
  • the present invention provides a method for determining a state of operation of an appliance comprising a compressor, the method comprising: receiving a time series of power consumption data relating to the operation of the compressor appliance over at least one cycle of operation of the compressor; deriving one or more measures characteristic of the at least one compressor cycle from the time series of power consumption data; and determining the state of operation of the compressor appliance based on comparing the one or more measures to a stored model of power consumption for the appliance.
  • Determining the state of operation may comprise identifying, based on the comparing, one of: a normal operating state of the appliance; and an abnormal operating state (e.g. a suboptimal performance state) or fault state of the appliance.
  • the method may further comprise receiving context data; and determining the state of operation of the appliance also in dependence on the context data, preferably wherein the context data is data that is not indicative of or related to power consumption of the appliance.
  • the method may further comprise identifying a potential fault state based on comparing the one or more measures to the stored model of power consumption for the appliance; and in response to identifying the potential fault state, evaluating operation of the appliance against the context data, and confirming identification of the fault state or disregarding the fault state, in dependence on the evaluation.
  • the context data comprises one or more of: weather data; temperature data, optionally interior or exterior temperature for a user environment at which the appliance is installed; occupancy data (e.g. current or typical number of occupants) for a user environment at which the appliance is installed; user behaviour data for one or more appliance users; data identifying usage frequency or high usage periods for the appliance; power usage data for a power distribution grid supplying power to the user environment at which the appliance is installed.
  • the appliance may be installed in a first user environment, and the context data may comprise context data associated with one or more further user environments, wherein the user environments optionally comprise residential dwellings (alternatively user environments could comprise any form of building of part thereof, commercial properties such as offices and factories etc.)
  • the context data may comprise power consumption data for one or more further appliances at respective user environments, the method comprising identifying a potential fault state based on identifying a deviation of the one or more measures from corresponding measures of the stored model of power consumption for the appliance; and disregarding the potential fault state in response to identifying one or more similar or related deviations identified for one or more of the further appliances.
  • the one or more further user environments may be selected based on a similarity criterion, the similarity criterion optionally comprising one or more of: location, socio-economic classification, size of property, number of occupants.
  • Evaluating the operation of the appliance against the context data may comprise determining whether a detected deviation of one or more measures from the model is indicative of a fault state based on the context data, e.g. based on whether the context data indicates that the deviation is caused by context factors other than the fault state.
  • evaluating the operation of the appliance against the context data may comprise comparing the derived measures, or deviations in one or more such measures from the model, to corresponding measures or corresponding deviations determined for other appliances / other user environments, e.g. to assess whether similar deviations are observed which may be taken to indicate that a local fault of the appliance is unlikely.
  • the time series of power consumption data may be received from one of: the appliance; a power metering device connected to the appliance arranged to measure the power consumption (specifically/only) of the appliance; and a power metering device for monitoring overall power consumption at a user environment comprising a plurality of power-consuming devices including the appliance, the method preferably further comprising processing a power consumption signal for the plurality of devices to determine the time series of power consumption data for the appliance and/or the compressor of the appliance, optionally by disaggregating the power consumption signal to identify a contribution to the signal from the appliance and/or compressor.
  • the method comprises deriving from the time series, optionally by disaggregation, power data relating specifically to the compressor of the appliance, the deriving step performed based on the derived power data.
  • the time series of power consumption data may relate to power consumption of the appliance during a first time period, the stored model comprising power consumption data relating to a second, past (earlier), time period.
  • the method may comprise monitoring power consumption of the appliance during the second time period and deriving the stored model based on the monitored power consumption (e.g. by deriving one or more comparison measures of the model).
  • the stored model comprises power consumption data for a second or representative appliance, optionally of a type or model corresponding to the appliance.
  • the stored model may have been created for a similar appliance (e.g. same make/model) in another user environment or in a laboratory / controlled test environment, or may be generated based on manufacturer data for the appliance.
  • the stored model may comprise respective comparison measures for the one or more derived measures, the method comprising comparing the measures derived from the times series of power data to corresponding measures of the model.
  • identifying the operating state comprises identifying a fault state based on detecting one or more deviations of one or more of the derived measures from corresponding comparison measures of the model.
  • the method may comprise computing one or more difference metrics indicative of a difference between one or more derived measures and one or more comparison measures, and determining the state of operation based on the one or more difference measures (e.g. individual differences may be computed for different measures or a single combined distance metric, e.g. Euclidean distance, could be computed).
  • the method may comprise, for at least one measure: determining a difference between the derived measure and the corresponding comparison measure in the model; comparing the difference to a threshold; and determining the state of operation in dependence on the comparison; the method preferably comprising determining the state of operation by comparing respective differences for a plurality of measures to one or more thresholds.
  • the method may comprise identifying a fault condition in response to one or more determined differences exceeding a threshold (or each measure difference exceeding a threshold), optionally for a predetermined duration.
  • the thresholds for different measures may be the same or different.
  • the method may also comprise determining a rate of change of a difference between one or more derived measures and one or more corresponding comparison measures, and determining the operating state and/or identifying a fault condition in dependence on the rate of change, the method optionally comprising identifying one of a plurality of operating states and/or fault conditions in dependence on the determined rate of change.
  • deriving one or more measures comprises detecting one or more operating cycles of the compressor from the series of power consumption data (e.g. by detecting periods during which power consumption exceeds some threshold).
  • deriving one or more measures comprises determining one or more cycle durations of the at least one operating cycle; and/or determining a representative cycle duration for a plurality of operating cycles detected over a given time period, optionally a mean cycle duration.
  • Deriving one or more measures may comprise determining a number of compressor operating cycles over a given time period. Deriving one or more measures may comprise determining a metric based on a representative cycle duration value and a number of cycles for a given time period or a plurality of time periods; preferably comprising determining the linear correlation between representative cycle durations and numbers of cycles for a plurality of time periods and/or determining the variance around the linear correlation.
  • the method comprises performing a frequency decomposition, optionally a Fourier analysis, to identify one or more frequency components in the time series of power consumption data, and determining one or more measures based on the frequency decomposition, preferably including determining one or more of: a peak in the frequency decomposition, a peak frequency at which the peak has maximum power, the maximum power of the peak, a peak width (e.g. determined as a width of the peak at half of the maximum power).
  • the method may comprise generating a user alert, optionally comprising an electronic message for delivery to a user device, in dependence on the identified operating state or fault state of the appliance, the user alert preferably generated only if an identified fault state is not disregarded based on evaluation of the appliance operation against context data.
  • the method comprises sending a control signal to the appliance, optionally to deactivate the appliance or alter an operating mode or configuration of the appliance, in dependence on the identified operating state or fault state of the appliance, the control signal preferably sent only if an identified fault state is not disregarded based on evaluation of the appliance operation against context data.
  • the appliance may comprise one of: a refrigeration appliance, for example a domestic refrigerator and/or freezer, a heat pump, and an air conditioning or other heating, ventilation and air conditioning (HVAC) appliance.
  • a refrigeration appliance for example a domestic refrigerator and/or freezer, a heat pump, and an air conditioning or other heating, ventilation and air conditioning (HVAC) appliance.
  • HVAC heating, ventilation and air conditioning
  • fault states that may be identified for a refrigeration appliance include, for example, a refrigerant leak or a door left open for a period of time.
  • the present invention also provides a computer-readable medium comprising software code adapted, when executed on a data processing apparatus, to perform any method as set out above.
  • the present invention also provides a system having means, optionally in the form of a processor with associated memory, for performing any method as set out above.
  • Figure 1 shows an overview of a system for evaluating operation of an appliance
  • Figure 2 shows an overview of a central processing system
  • FIG. 3 shows an overview of an energy monitoring device
  • Figure 4 shows a flow diagram of a method for evaluating operation of an appliance
  • Figure 5 shows a disaggregated or otherwise isolated power series for a compressor of a compressor based appliance
  • Figure 6 shows a graph of the number of compressor cycles plotted against the average duration of the compressor cycles over a given time
  • Figure 7 shows a frequency spectrum of the isolated power series data of a compressor based appliance.
  • Figure 1 shows a system 101 for determining a state of operation of a compressor based appliance 106 (for example a refrigerator) from a disaggregated or otherwise isolated power series according to an embodiment of the present invention.
  • the system includes a central processing system 102 (e.g. cloud-based) and one or more energy monitoring devices 104 installed at appliance locations, for example in the homes or other user environments where the appliances are installed and used.
  • a central processing system 102 e.g. cloud-based
  • energy monitoring devices 104 installed at appliance locations, for example in the homes or other user environments where the appliances are installed and used.
  • a data log 302 of a metering device 104 receives and stores power series data from an energy sensor 306 and clock 305 of the metering device.
  • Power series data is data which directly relates to the rate of energy consumption over time of one or several devices or appliances that consume electrical power. The data is obtained by frequent sampling of electrical power consumption. For example electrical power consumption may be sampled once every 10 seconds or several times a second, in accordance with the clock. The change in power consumption over time can then be visualised on a graph as an energy signature 501 as depicted in Figure 5.
  • the power series data typically comprises a time series of power (or alternatively energy) consumption values measured at regular time intervals, and specified in a suitable unit such as Watt if (average) electrical power is measured at each time interval or joule or kWh if total energy consumption over each time interval is measured.
  • Time information may be implicit (due to constant measurement intervals) or each power/energy value may be associated with an explicit time value (e.g. from clock 305).
  • the energy monitoring device 104 (also referred to as a metering device) may for example be a whole house power monitoring device, a smart meter or a smart plug. The energy monitoring device may also be integrated into the appliance 106 being monitored.
  • the data log 302 of the metering device also receives and stores identification information 307 from an internal memory 301 of the metering device.
  • This identification information may be uploaded by the user or the provider to the internal memory of the metering device before or during its installation via a user interface 105.
  • the identification information 307 may include information in relation to the nature of the monitoring device and whether the power series data is aggregate data or relates to a single domestic electrical appliance.
  • the identification data 307 may also include information in relation to the user, the house and the household such as the number of occupants in the household, the address of the house, the size of the house, and the socio-economic group of the household. This information may be input by the user via the user interface 105.
  • the location or address of the house may be obtained via a geo-location sensor 303 connected to or part of the energy monitoring device 104.
  • the data log of the central processing system 202 receives the stored power series data and the identification information 307 from the data log of the metering device 302 via a network 103.
  • the processor of the system 203 carries out a disaggregation algorithm 401 ( Figure 4), for example as disclosed in GB2479790, to obtain an isolated signature for the stored power series data 501 , an example of which can be seen in Figure 5.
  • a disaggregation algorithm 401 Figure 4
  • the stored power series data is in relation to a single domestic electrical appliance 106, no disaggregation is necessary.
  • the stored power series data may be aggregated power series data of, for example, the whole house (including other domestic appliances 107), or the power series data may relate to a single appliance for example wherein the metering device is a smart plug.
  • the stored power series data is aggregated power series data, it may be preferable to store data during periods when usage of other appliances is likely to be low, particularly at night (e.g. from 01 :00 to 05:00), when general household power use is low.
  • Use of the compressor based appliance would also likely be low during these hours making disaggregation of the power series data of the compressor based appliance easier, more reliable or more accurate since other power consuming elements of the appliance (other than the compressor) would not be in use.
  • the power series data for example, in the case of refrigerators, there would be fewer instances of the door being opened and activating the internal light, therefore only the power consumption of the compressor and compressor controller would contribute to the aggregate power consumption signal.
  • the central processing system is preferably located on one or more cloud-based servers 201.
  • the metering device may be connected to the system via the network 103 as depicted in Figure 1 , for example via the Internet through a wired or wireless connection.
  • Identification information 307 may also be provided by the user or provider directly to the central processing system data log 202 via the user interface 105.
  • the isolated signature of a compressor based appliance 106 may be further disaggregated 401 to obtain the energy trace 501 of the compressor alone.
  • the processor 203 of the central processing system After receiving the aggregated energy data from the energy monitoring device 104, the processor 203 of the central processing system processes the data in order to identify an operating state of the appliance.
  • An example method for processing the data and identifying operating states, including fault states, is outlined in Figure 4.
  • the aggregated data is disaggregated (401 ) by the processor 203 to produce a time series of power consumption data specific to the appliance or even to the appliance compressor.
  • a frequency decomposition / transform may then performed over a given time interval of the data, e.g. 24 hours (402).
  • the feature set comprises various features or measures derived from the power data which are characteristic of the power consumption of the appliance, and in particular of compressor cycles of the appliance, and may e.g. include various statistical averages or peaks in the consumption data, examples of which are described in more detail below.
  • These extracted features may be recorded and stored over a period, for example at least 7 days, to determine a set of normal or baseline values for the features (404).
  • the baseline values form a model of normal operation of the compressor appliance and are used to set one or more thresholds (405) used to identify normal activity or deviations from normal activity.
  • the extracted features may be recorded and stored continuously for each rolling 24 hour period.
  • the feature values for a given 24 hour period are then compared to the corresponding features of the previously determined 7-day baseline model of activity (406).
  • the system identifies a potential fault.
  • environmental or context data (407) that may have been received from contextual data feeds such as the external temperature 408 or a corresponding set of average feature values for other similar appliances/locations (e.g. a regional average feature set 409).
  • the context data is used to filter potential faults identified by the baseline model comparison, by disregarding or confirming a potential fault depending on the context data.
  • the context data does not provide reasons to disregard the potential fault, the fault state of the appliance is confirmed and an alert is sent (410), e.g. via electronic message or otherwise, to the user or the energy supplier.
  • the central processing system 203 analyses the time series of power consumption data (optionally after disaggregation) to extract one or more features, in the form of a set of metrics (or measures) characteristic of the power consumption signature of the appliance and/or its compressor.
  • the features typically relate to aspects of the compressor cycle such as cycle durations and numbers. Examples of the analysis and the determined feature metrics are given below. Example 1 - Length of each compressor cycle
  • FIG. 5 A typical isolated signature of a compressor based appliance, such as a refrigerator, is shown in Fig. 5.
  • the isolated signature depicts a series of compressor cycles resulting from the repeated turning on and off of the compressor in the compressor based appliance.
  • a given compressor cycle is a distinct period during which the compressor is on essentially continuously.
  • the duration of each compressor cycle may be typically 5 to 40 minutes.
  • the compressor control processor monitors the internal temperature of the refrigerator. As the internal temperature of the refrigerator increases, from the opening of the refrigerator door or from general leakage of air, the compressor control processor activates the compressor to cycle refrigerant into the evaporator coil in the internal cavity of the refrigerator.
  • the central processing system determines the duration of each individual compressor cycle 502 from the isolated signature 501 of the compressor based appliance 501.
  • the cycles themselves may be detected by identified continuous periods of power consumption above zero or above some threshold level indicative of background consumption and/or signal noise.
  • the system may further derive a metric from the values of the durations of the individual compressor cycles 502 over an extended period of time.
  • the period of time may, for example, be between 4 hours and 1 day.
  • the metric may be a mean or another statistical average of multiple cycle durations. The mean or other representative cycle duration may then be used as one of the features for evaluating operation of the appliance.
  • the central processing system may also determine the number of individual cycles over an extended period of time from the isolated signature of the compressor based appliance.
  • the period of time for which the number of individual cycles is sampled may be 24 hours.
  • the number of cycles over the given period may be used as a further feature in the evaluation of the appliance operation.
  • an average number of individual cycles over multiple time periods may be used as a feature metric.
  • the system may further evaluate the number of individual cycles in a given time period against a representative (e.g. mean) cycle duration for the time period (as described above). This relationship is depicted in graph form in Figure 6, and may be used by the system to compute further feature metrics of the isolated signature such as the linear correlation 602, variance and/or variance around the linear correlation 603.
  • clustering 601 may be used to extract a statistical average. Such clustering techniques are known in the art.
  • the system may further transform the isolated signature of the compressor based appliance (in the form of the time series of power consumption data) over a given period of time into the frequency domain, to provide a frequency spectrum as shown in Figure 7.
  • the system preferably performs a Fourier transfer or other frequency decomposition on a given window (e.g. 24h) of the power data time series.
  • a frequency decomposition may typically produce a peak 704.
  • Features of this peak 704 can be calculated by the central processing system and stored as further feature metrics. Examples of features of the peak 704 may include the frequency 702 at which the peak reaches its maximum power, the maximum power 701 of the peak, and the width of the peak at half max height 703. These features may provide further metrics of the isolated signature which are stored in the data log of the central processing system 202.
  • the above metrics may be calculated over a period of time during normal performance of the compressor based appliance. This period of time may be referred to as an evaluation or training period for the system.
  • the data log stores benchmark values for the metrics obtained over this period of time as a baseline model of normal operation of the appliance.
  • the training period may be between 1 and 14 days. It may preferably be 7 days. It may also be longer than 14 days.
  • the training period may be fixed, however it may also be a moving period. In some cases, such as in the cases of refrigerators and freezers, night time readings may be more reliable. In these cases it is preferable for the training period to be limited to night time power series data only.
  • the baseline model may be determined from a different appliance (e.g. one of the same or similar type/make/model), e.g. by monitoring operation in a laboratory environment and/or based on manufacturer-provided data.
  • a different appliance e.g. one of the same or similar type/make/model
  • the feature metric(s) relating to the compressor appliance may be monitored for deviations from the baseline model in order to diagnose a possible fault condition. Deviations are determined by computing differences between one or more feature metrics derived from the power consumption data and one or more corresponding measures of the baseline model (e.g. individual metric differences may be computed, or combined difference metrics for multiple features such as Euclidean distance could be used). Deviations may only be identified as significant if they exceed one or more relevant thresholds.
  • a threshold for deviation from the baseline can be set by the central processing system 102, and ongoing readings compared to the threshold(s) (see operations 204, 406 in Figures 2/4).
  • the threshold may be set (step 405 in Figure 4) based on statistical models (Z test), distribution (modified z-score), distance (local outlier factor using kNN), density (relative density based outlier factor) or clusters of baseline data. Threshold may also be set simply based on prior knowledge (e.g. laboratory trials).
  • a single threshold may be set, or multiple thresholds, where multiple feature metrics are available.
  • a fault condition e.g. one or more or all feature metrics deviating from the model by less then the relevant thresholds
  • the rate of deviation from the baseline and hysteresis or smoothing may be included to determine when a trigger condition for identifying a fault state is met, or to distinguish different types of fault states. For example, in the case of a refrigerator, very gradual changes may indicate a slow refrigerant leak, while sudden changes in a feature metric may indicate a component failure, or door left open. In some cases a fault condition may only be triggered if the trigger conditions (e.g. metric deviations exceeding one or more thresholds) are fulfilled for a continuous period of time.
  • the identification of the operating state and/or fault states of the appliance may also be based on contextual information input to the central processing system, as shown in Fig. 1 , to provide additional context to the fault diagnostics.
  • the contextual data may include environmental factors such as external and/or internal temperature 1 10.
  • the internal temperature may, for example, be detected by a temperature sensor of the energy monitoring device 304 ( Figure 3) or via smart thermostat or the like.
  • Human behaviour may also be recorded, such as occupancy data derived from cameras or passive infrared sensors 109. Patterns of human behaviour may be established to help identify deviations from the benchmark values, for example high usage periods such as morning and evening meal times.
  • the context filter may allow a fault condition to be distinguished from a change or sub-optimal user behaviour such as an increased number of house occupants for a period of time, or an empty house while the occupants are on holiday or, in the case of the refrigerator, the door being left open accidentally.
  • the context filter 407 may also take into account more general data from outside the home 408, such as weather data 108, as shown in Fig. 1 or overall grid power usage, time of day, or data from other comparable households 409.
  • metrics from other households may be monitored, and can be used to assess an individual metric in question.
  • Information relating to the households may be input into the central processing system by the users or provider via the user interface shown in Fig. 1.
  • Information such as postcode, socio-economic group, size of house, number of occupants, may be used to determine a degree of similarity of other households to the house in question where the monitored appliance is located.
  • a specific subset of households may then be selected based on similarity (e.g. those most similar), and data pertaining to those households is then used as context data in step 407 of the Figure 4 process.
  • a deviation of the measured metric from the baseline model is observed, however a similar deviation is observed for a group of households considered to be similar or representative, then it is less likely to be a fault. For example, half-time in a World Cup game may see the door of a refrigerator being opened for a longer than usual period. Further, the beginning or end of school holidays may trigger a change in user behaviour.
  • step 407 involves evaluating any potential fault conditions identified by comparing the feature metrics to the baseline model against the context data. If the context data indicates factors that may contribute to the identified deviation from the model other than appliance-specific faults (e.g. similar deviations detected in other households; unusually high temperatures explaining increased cycle durations in refrigeration appliances etc.), then the potential fault may be disregarded.
  • the context data is thus used to provide a context filter for filtering potential fault states identified by the baseline comparison. If, on the other hand a potential fault (due to a detected baseline deviation) has been validated against the context filter 407 and is not accounted for by the context data, then the fault state is classified as a confirmed fault or other sub-optimal operating behaviour (for example leaving a refrigerator door open).
  • an electronic message or other notification or alert is sent by the central processing system to the user or another entity such as an energy or service provider for resolution (see operations 205 / 410), with the timestamp of the fault and the classification of the fault / sub-optimal behaviour and any appropriate analysis and context.
  • the notification may be sent, e.g. as an electronic message via SMS or e-mail to a user device of the user.
  • the central processing system sends a control signal directly to the appliance itself to deactivate the appliance (e.g. in the case of a serious fault) or to modify an operating mode or configuration of the appliance (e.g. to optimize operation in case of sub-optimal performance).
  • a refrigeration appliance may be deactivated in the case of a refrigerant leak being detected.
  • Archetypes could be created for certain compressor based appliances in order to accelerate the benchmarking process, or make it more accurate. This could be achieved by monitoring specific compressor based appliance types (e.g. specific makes/models) in the lab or in the field. This information would then be stored on the data log of the central processing system and would be used by the central processing system during baselining to provide more reliable baseline data and therefore more sensitive thresholds. The system could also be improved and tuned by taking information from faults predicted and then found in the field (or not) and then fed back into the central processing system to increase the accuracy of its output, based on past results.
  • specific compressor based appliance types e.g. specific makes/models
  • compressor based appliances examples include Heating, Ventilation and Air Conditioning (HVAC) systems, ground-source and air-source heat pumps, refrigerators, freezers, fridge-freezers and chillers.
  • HVAC Heating, Ventilation and Air Conditioning
  • the compressor based appliances may be for domestic, commercial or industrial use. For example, they may be used for the storage or processing of food and drink (e.g. in homes, restaurants or food processing plants).
  • the appliances may also be used in laboratories, manufacturing plants, chemical processing plants, or other industrial settings.
  • the system may simultaneously monitor many such appliances across different user environments (e.g. households).
  • performance data may also be supplied to e.g. appliance manufacturers to assist with appliance support and quality management.

Landscapes

  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Combustion & Propulsion (AREA)
  • Chemical & Material Sciences (AREA)
  • Mechanical Engineering (AREA)
  • Thermal Sciences (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Air Conditioning Control Device (AREA)

Abstract

Procédé de détermination d'un état de fonctionnement d'un appareil comprenant un compresseur. Le procédé consiste à recevoir une série chronologique de données de consommation d'énergie relatives au fonctionnement de l'appareil de compression sur au moins un cycle de fonctionnement du compresseur ; dériver une ou plusieurs mesures caractéristiques du ou des cycles de compresseur à partir de la série chronologique de données de consommation d'énergie ; et déterminer l'état de fonctionnement de l'appareil de compression sur la base de la comparaison de la ou des mesures à un modèle de consommation d'énergie stocké pour l'appareil. Cet appareil de compression peut être, par exemple, un appareil de réfrigération.
PCT/GB2019/053011 2018-10-24 2019-10-22 Procédé de détection d'un état de fonctionnement d'un appareil WO2020084299A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
GB1817322.9 2018-10-24
GB1817322.9A GB2578332B (en) 2018-10-24 2018-10-24 Method of detecting an operating state of an appliance

Publications (1)

Publication Number Publication Date
WO2020084299A1 true WO2020084299A1 (fr) 2020-04-30

Family

ID=64453888

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/GB2019/053011 WO2020084299A1 (fr) 2018-10-24 2019-10-22 Procédé de détection d'un état de fonctionnement d'un appareil

Country Status (2)

Country Link
GB (1) GB2578332B (fr)
WO (1) WO2020084299A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2479790A (en) 2010-04-23 2011-10-26 Alertme Com Ltd Identifying an electrical appliance from an aggregate power series
WO2012106709A2 (fr) * 2011-02-04 2012-08-09 Myenersave, Inc. Systèmes et procédés d'amélioration de la précision de désagrégation de niveau appareil dans techniques de surveillance de charge d'appareil non intrusives
GB2524033A (en) * 2014-03-11 2015-09-16 British Gas Trading Ltd Determination of a state of operation of a domestic appliance
US20160299038A1 (en) * 2015-04-09 2016-10-13 Palo Alto Research Center Incorporated System and method for remotely inferring characteristics of thermostat-controlled appliances
WO2017004416A1 (fr) * 2015-06-30 2017-01-05 Emerson Climate Technologies Retail Solutions, Inc. Maintenance et diagnostic pour systèmes de réfrigération

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5495148B1 (ja) * 2013-06-17 2014-05-21 軍 楊 運転制御装置及び運転制御方法
WO2015149928A2 (fr) * 2014-03-31 2015-10-08 Basf Se Procédé et dispositif pour l'évaluation en ligne d'un compresseur

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2479790A (en) 2010-04-23 2011-10-26 Alertme Com Ltd Identifying an electrical appliance from an aggregate power series
WO2012106709A2 (fr) * 2011-02-04 2012-08-09 Myenersave, Inc. Systèmes et procédés d'amélioration de la précision de désagrégation de niveau appareil dans techniques de surveillance de charge d'appareil non intrusives
GB2524033A (en) * 2014-03-11 2015-09-16 British Gas Trading Ltd Determination of a state of operation of a domestic appliance
US20160299038A1 (en) * 2015-04-09 2016-10-13 Palo Alto Research Center Incorporated System and method for remotely inferring characteristics of thermostat-controlled appliances
WO2017004416A1 (fr) * 2015-06-30 2017-01-05 Emerson Climate Technologies Retail Solutions, Inc. Maintenance et diagnostic pour systèmes de réfrigération

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
COLE A I ET AL: "Data extraction for effective non-intrusive identification of residential power loads", INSTRUMENTATION AND MEASUREMENT TECHNOLOGY CONFERENCE, 1998. IMTC/98. CONFERENCE PROCEEDINGS. IEEE ST. PAUL, MN, USA 18-21 MAY 1998, NEW YORK, NY, USA,IEEE, US, vol. 2, 18 May 1998 (1998-05-18), pages 812 - 815, XP010281749, ISBN: 978-0-7803-4797-7, DOI: 10.1109/IMTC.1998.676838 *
LESLIE K. NORFORD ET AL: "Non-intrusive electrical load monitoring in commercial buildings based on steady-state and transient load-detection algorithms", ENERGY AND BUILDINGS, vol. 24, no. 1, 1 January 1996 (1996-01-01), CH, pages 51 - 64, XP055288518, ISSN: 0378-7788, DOI: 10.1016/0378-7788(95)00958-2 *

Also Published As

Publication number Publication date
GB201817322D0 (en) 2018-12-05
GB2578332A (en) 2020-05-06
GB2578332B (en) 2021-07-28

Similar Documents

Publication Publication Date Title
Rogers et al. A review of fault detection and diagnosis methods for residential air conditioning systems
US11739963B2 (en) HVAC analytics
US11205027B2 (en) Equipment fault detection, diagnostics and disaggregation system
US10309677B2 (en) HVAC system air filter diagnostics and monitoring
Chen et al. Non-intrusive occupancy monitoring using smart meters
Najafi et al. Data analytics for energy disaggregation: Methods and applications
US7716012B2 (en) Method for process monitoring in a utility system
Schoofs et al. Annot: Automated electricity data annotation using wireless sensor networks
US20110288793A1 (en) Event identification
CN106461252B (zh) 加热、通风或空气调节系统空气过滤器诊断和监视
US20140324240A1 (en) Method And System For Disaggregating Thermostatically Controlled Appliance Energy Usage From Other Energy Usage
WO2012016589A1 (fr) Système et procédé permettant d'identifier automatiquement des dispositifs/appareils électriques
GB2476927A (en) Event or appliance detection from utility data
GB2478166A (en) Determining background utility consumption levels using NILM
Barker et al. Empirical characterization, modeling, and analysis of smart meter data
JP2004169989A (ja) 異常診断システム
US10031501B2 (en) Apparatus and method for detecting degradation in climate control system performance
US20220337081A1 (en) An in-line device and a method for controlling an electrical appliance
US11125790B2 (en) Method for operating a power consumption metering system and power consumption metering system
GB2563860A (en) A system and method for identifying appliances under recall
WO2020084299A1 (fr) Procédé de détection d'un état de fonctionnement d'un appareil
EP3735759A1 (fr) Détection d'appareils électriques inefficaces
GB2618987A (en) System for detecting abnormal operating states of a heating system
Guo et al. Labeling modes of operation and extracting features for fault detection with cloud-based thermostat data
CN114487663A (zh) 用电异常分析方法和装置、电子设备、存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19797336

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19797336

Country of ref document: EP

Kind code of ref document: A1