Remote diagnosis of energy or resource-consuming devices based on usage data
Systems and methods are provided to retrieve or analyze usage data collected from a device or a facility where the device, optionally with devices are located, and identify useful features for making a diagnosis of the device. The diagnosis can be made before a system failure to reduce down time and inefficient use of the device, or after the system failure to expedite and facilitate diagnosis and repair. In addition to the usage data, such as energy and resource consumption, the system can also obtain information relating to the facility and the device's external environment which can be used for normalizing the usage data. Further, based on the diagnosis, the system can make suitable recommendations for repair, replacement, maintenance and upgrade.
1 . A computer-implemented method for determining an efficiency of a physical household appliance, comprising:
sensing, by a computing system, appliance usage data associated with the physical household appliance from one or more sensors, wherein the appliance usage data comprises data of different sources, wherein the data of different sources comprises messages or signals including codes;
transforming, by the computing system, the appliance usage data into a standardized format compatible with object based data structures, wherein the transforming comprises normalizing the appliance usage data into a time series format;
storing the transformed appliance usage data in the standardized format in a data source;
determining, by a machine learning model, one or more attributes, selected from candidate attributes identifying a property of a material, an environmental parameter, or an energy parameter associated with the physical household appliance or a facility at which the physical household appliance is located, that distinguish an efficient operation of the physical household appliance from an inefficient operation, wherein the determining of the one or more attributes comprises;
selecting a first attribute having a highest individual diagnosis capability compared to other candidate attributes;
iteratively scoring other attributes besides the first attribute based on respective combined diagnosis capabilities of each of the other attributes when combined with the first attribute; and
selecting a second attribute from the other attributes based on the iterative scoring, wherein a combined diagnosis capability of the second attribute is different from an individual diagnosis capability of the second attribute by itself;
generating, by the computing system based on the selected first attribute and the selected second attribute, an indication of whether to perform a repair or a maintenance on the physical household appliance; and
in response to generating an indication to perform a repair or a maintenance, selectively implementing a physical shutdown of the physical household appliance.
2 . The computer-implemented method of claim 1 , wherein the one or more attributes comprise a strength of sunlight.
3 . The computer-implemented method of claim 1 , wherein an area under curve (AUC) of a receiver operator characteristic (ROC) is indicative of a rate of false positives.
4 . The computer-implemented method of claim 1 , further comprising determining an efficiency of the physical household appliance based on the determined attributes, wherein the efficiency is filtered based on an intensity of the physical household appliance.
5 . The computer-implemented method of claim 1 , wherein the indication of whether to perform a repair or a maintenance on the physical household appliance corresponds to a declining relative efficiency of the physical household appliance prior to a failure of the physical household appliance.
6 . The computer-implemented method of claim 1 , further comprising:
generating, by the computing system based on an efficiency of the physical household appliance, an indication of whether to perform an upgrade or a replacement on the physical household appliance.
7 . A computing system for determining an efficiency of a physical household appliance, comprising:
a processor; and
a memory storing instructions that, when executed by the processor, cause the system to perform:
sensing, by the computing system, appliance usage data associated with the physical household appliance from one or more sensors, wherein the appliance usage data comprises data of different sources, wherein the data of different sources comprises messages or signals including codes;
transforming, by the computing system, the appliance usage data into a standardized format compatible with object based data structures, wherein the transforming comprises normalizing the appliance usage data into a time series format;
storing the transformed appliance usage data in the standardized format in a data source;
determining, by a machine learning model, one or more attributes, selected from candidate attributes identifying a property of a material, an environmental parameter, or an energy parameter associated with the appliance or a facility at which the physical household appliance is located, that distinguish an efficient operation of the appliance from an inefficient operation, wherein the determining of the one or more attributes comprises:
selecting a first attribute having a highest individual diagnosis capability compared to other candidate attributes;
iteratively scoring other attributes besides the first attribute based on respective combined diagnosis capabilities of each of the other attributes when combined with the first attribute; and
selecting a second attribute from the other attributes based on the iterative scoring, wherein a combined diagnosis capability of the second attribute is different from an individual diagnosis capability of the second attribute by itself;
generating, by the computing system based on the selected first attribute and the selected second attribute, an indication of whether to perform a repair or a maintenance on the physical household appliance; and
in response to generating an indication to perform a repair or a maintenance, selectively implementing a physical shutdown of the physical household appliance.
8 . The system of claim 7 , wherein the one or more attributes comprise a strength of sunlight.
9 . A non-transitory computer-readable medium of a computing system for determining an efficiency of a physical household appliance, the non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the computer system to perform:
sensing appliance usage data associated with the physical household appliance from one or more sensors, wherein the appliance usage data comprises data of different sources, wherein the data of different sources comprises messages or signals including codes;
transforming the appliance usage data into a standardized format compatible with object based data structures, wherein the transforming comprises normalizing the appliance usage data into a time series format;
storing the transformed appliance usage data in the standardized format in a data source;
determining, by a machine learning model, one or more attributes, selected from candidate attributes identifying a property of a material, an environmental parameter, or an energy parameter associated with the physical household appliance or a facility at which the physical household appliance is located, that distinguish an efficient operation of the physical household appliance from an inefficient operation, wherein the determining of the one or more attributes comprises;
selecting a first attribute having a highest individual diagnosis capability compared to other candidate attributes;
iteratively scoring other attributes besides the first attribute based on respective combined diagnosis capabilities of each of the other attributes when combined with the first attribute; and
selecting a second attribute from the other attributes based on the iterative scoring, wherein a combined diagnosis capability of the second attribute is different from an individual diagnosis capability of the second attribute by itself;
generating, by the computing system based on the selected first attribute and the selected second attribute, an indication of whether to perform a repair or a maintenance on the physical household appliance; and
in response to generating an indication to perform a repair or a maintenance, selectively implementing a physical shutdown of the physical household appliance.
10 . The non-transitory computer-readable medium of claim 9 , wherein an area under curve (AUC) of a receiver operator characteristic (ROC) is indicative of a rate of false positives.
11 . The computer-implemented method of claim 1 , wherein the normalizing, by the computing system, the appliance usage data is based on any three or more of: a strength of sunlight to which the facility is exposed, an atmospheric pressure, an altitude, a latitude, and a wind speed.
12 . The computer-implemented method of claim 1 , wherein the normalizing, by the computing system, the appliance usage data is based on a type of insulation associated with the facility and energy utilization associated with a different appliance device in the facility.
13 . The computer-implemented method of claim 4 , wherein the efficiency is based on a rate of deterioration corresponding to the one or more attributes relative to an expected rate of deterioration.
14 . The computer-implemented method of claim 1 , further comprising:
determining an extent of a predicted improvement resulting from the repair or the maintenance; and
selectively implementing, by the computing system, the repair or the maintenance on the physical household appliance based on the extent of the predicted improvement.
15 . The system of claim 7 , wherein the physical household appliance comprises a HVAC device.
16 . The system of claim 7 , wherein the physical household appliance comprises an air conditioning or heating device.
17 . The system of claim 7 , wherein the physical household appliance comprises a washer, a dryer, an oven, a toaster, a refrigerator, or a freezer.
18 . The system of claim 7 , wherein each of the other attributes is independent from the first attribute, and selecting a second attribute from the other attributes based on the iterative scoring comprises selecting a second attribute that has a score indicating a highest combined diagnosis capability compared to remaining other attributes, wherein the remaining other attributes are unselected.