Identifying and categorizing power consumption with disaggregation

• US 10,175,276 B2
• Filed: 11/28/2017
• Issued: 01/08/2019
• Est. Priority Date: 11/26/2014
• Status: Active Grant

• Current Assignee: Sense Labs, Inc.
• Original Assignee: Sense Labs, Inc.
• Inventors: Fishburn, Matthew William, Phillips, Michael S.
• Primary Examiner(s): Bhat, Aditya
• Application Number:
  US15/824,174
• Publication Number:
  US 20180196094A1
• Time in Patent Office: 406 Days
• Field of Search: None
• CPC Class Codes: G01R 21/133 by using digital technique

Abstract

A power monitor in a building processes a power monitoring signal obtained from the power line to the building to identify electrical events corresponding to state changes of devices in the building. The electrical event may be processed with a classifier to determine if the electrical event corresponds to a state change of an identified device from a list of identified devices or corresponds to a state change of an ambiguous device that is not from the list of identified devices. The power monitor may also determine power consumption information about individual identified devices and about groups of devices, such as a group of always on devices, a group of ambiguous devices, and categories of the group of ambiguous devices. Information about state changes of devices and power consumption of individual devices and groups of devices may be transmitted to a user device for presentation to a user.

Claims

1. A system for providing information about devices in a building, the system comprising one or more devices, each comprising at least one processor and at least one memory, configured to:

obtain a first classifier, wherein the classifier processes an electrical event to generate scores indicating a match between the electrical event and state changes of a plurality of identified devices;

obtain a power monitoring signal corresponding to a measurement of a power line to the building;

identify a first electrical event in the power monitoring signal, wherein the first electrical event corresponds to a state change of a device;

determine that the first electrical event corresponds to a first state change of a first identified device by processing the power monitoring signal with the first classifier;

update a list of identified devices using the first state change of the first identified device;

identify a second electrical event in the power monitoring signal, wherein the second electrical event corresponds to a state change of a device;

determine that the second electrical event corresponds to an ambiguous device by processing the power monitoring signal with the first classifier, wherein the ambiguous device is not in the plurality of identified devices;

select one or more categories from a plurality of categories as corresponding to the second electrical event by processing the electrical event with a second classifier;

add an entry to a list of ambiguous devices corresponding to the second electrical event, wherein the entry comprises information about a change in power consumption caused by the second electrical event and information about the one or more categories corresponding to the electrical event;

compute an amount of power consumption for a sequence of time intervals for each identified device of the plurality of identified devices using the list of identified devices and one or more wattage models;

compute an amount of power consumption for the sequence of time intervals for each category of the plurality of categories using the list of ambiguous devices; and

transmit device information to a user device for presentation to a user, wherein the device information comprises (i) information about power consumption of the plurality of identified devices and (ii) information about power consumption of the plurality of categories of ambiguous devices.

2. The system of claim 1, wherein the system comprises:

a power monitor installed in the building and comprising at least one processor and at least one memory; and

one or more servers, each comprising at least one processor and at least one memory.

3. The system of claim 2, wherein the power monitor is configured to:

compute the amount of power consumption for the sequence of time intervals for each identified device of the plurality of identified devices; and

compute the amount of power consumption for the sequence of time intervals for each category of the plurality of categories of ambiguous devices.

4. The system of claim 1, wherein the one or more devices are configured to:

compute an amount of power consumption for a group of devices that are always on; and

transmit to a user device for presentation to a user information about power consumption for the group of devices that are always on.

5. The system of claim 1, wherein the device information transmitted to the user device corresponds to an amount of power consumption over a time period.

6. The system of claim 5, wherein the time period is a second, a minute, an hour, a day, a week, or a month.

7. The system of claim 1, wherein the device information is transmitted to the user device in real time.

8. A method for providing information about devices in a building, the method comprising:

obtaining a first classifier, wherein the classifier processes an electrical event to generate scores indicating a match between the electrical event and state changes of a plurality of identified devices;

obtaining a power monitoring signal corresponding to a measurement of a power line to the building;

identifying a first electrical event in the power monitoring signal, wherein the first electrical event corresponds to a state change of a device;

determining that the first electrical event corresponds to a first state change of a first identified device by processing the power monitoring signal with the first classifier;

updating a list of identified devices using the first state change of the first identified device;

identifying a second electrical event in the power monitoring signal, wherein the second electrical event corresponds to a state change of a device;

determining that the second electrical event corresponds to an ambiguous device by processing the power monitoring signal with the first classifier, wherein the ambiguous device is not in the plurality of identified devices;

selecting one or more categories from a plurality of categories as corresponding to the second electrical event by processing the electrical event with a second classifier;

adding an entry to a list of ambiguous devices corresponding to the second electrical event, wherein the entry comprises information about a change in power consumption caused by the second electrical event and information about the one or more categories corresponding to the electrical event;

computing an amount of power consumption for a sequence of time intervals for each identified device of the plurality of identified devices using the list of identified devices and one or more wattage models;

computing an amount of power consumption for the sequence of time intervals for each category of the plurality of categories using the list of ambiguous devices; and

transmitting device information to a server, wherein the device information comprises (i) information about power consumption of the plurality of identified devices and (ii) information about power consumption of the plurality of categories of ambiguous devices.

9. The method of claim 8, comprising:

determining a first amount of power consumption before the second electrical event;

determining a second amount of power consumption after the second electrical event; and

determining the information about the change in power consumption caused by the second electrical event using the first amount of power consumption and the second amount of power consumption.

10. The method of claim 9, wherein the device information comprises states of the plurality of identified devices.

11. The method of claim 9, comprising:

determining that the first electrical event corresponds to the first state change of the first identified device by adding one or more nodes to a graph corresponding to the first electrical event and performing a beam search on the graph.

12. The method of claim 9, comprising:

determining that the second electrical event corresponds to the ambiguous device transitioning from an off state to an on state;

identifying a third electrical event in the power monitoring signal, wherein the third electrical event corresponds to a state change of a device;

determining that the third electrical event corresponds to the ambiguous device transitioning from the on state to the off state; and

removing the entry for the ambiguous device from the list of ambiguous devices.

13. The method of claim 12, comprising:

determining that the third electrical event corresponds to the ambiguous device transitioning from the on state to the off state by comparing features of the second electrical event with features of the third electrical event.

14. The method of claim 9, wherein the entry for the ambiguous device comprises statistics of previous state changes of the ambiguous device.

15. The method of claim 14, wherein selecting one or more categories corresponding to the second electrical event comprises using the statistics of previous state changes of the ambiguous device.

16. The method of claim 9, wherein the entry for the ambiguous device comprises one or more categories corresponding to the ambiguous device, a time of the second electrical event, or features computed from the second electrical event.

17. The method of claim 9, further comprising converting the ambiguous device to an identified device using statistics of state changes of the ambiguous device.

18. One or more non-transitory computer-readable media comprising computer executable instructions that, when executed, cause at least one processor to perform actions comprising:

obtaining a first classifier, wherein the classifier processes an electrical event to generate scores indicating a match between the electrical event and state changes of a plurality of identified devices;

obtaining a power monitoring signal corresponding to a measurement of a power line to a building;

identifying a first electrical event in the power monitoring signal, wherein the first electrical event corresponds to a state change of a device;

determining that the first electrical event corresponds to a first state change of a first identified device by processing the power monitoring signal with the first classifier;

updating a list of identified devices using the first state change of the first identified device;

identifying a second electrical event in the power monitoring signal, wherein the second electrical event corresponds to a state change of a device;

determining that the second electrical event corresponds to an ambiguous device by processing the power monitoring signal with the first classifier, wherein the ambiguous device is not in the plurality of identified devices;

selecting one or more categories from a plurality of categories as corresponding to the second electrical event by processing the electrical event with a second classifier;

adding an entry to a list of ambiguous devices corresponding to the second electrical event, wherein the entry comprises information about a change in power consumption caused by the second electrical event and information about the one or more categories corresponding to the electrical event;

computing an amount of power consumption for a sequence of time intervals for each identified device of the plurality of identified devices using the list of identified devices and one or more wattage models;

computing an amount of power consumption for the sequence of time intervals for each category of the plurality of categories using the list of ambiguous devices; and

transmitting device information to a server, wherein the device information comprises (i) information about power consumption of the plurality of identified devices and (ii) information about power consumption of the plurality of categories of ambiguous devices.

19. The one or more non-transitory computer-readable media of claim 18, wherein the power monitoring signal indicates a current or voltage of the power line.

20. The one or more non-transitory computer-readable media of claim 18, wherein the one or more categories comprise lighting, electronics, heating, cooling, laundry, kitchen, cooking, refrigerators, freezers, pumps, small appliances, or electric vehicles.

21. The one or more non-transitory computer-readable media of claim 18, wherein the classifier comprises a plurality of transition models.