Trained models for discovering target device presence
Embodiments generate machine learning predictions to discover target device energy usage. One or more trained machine learning models configured to discover target device energy usage from source location energy usage can be stored. Multiple instances of source location energy usage over a period of time can be received for a given source location. Using the trained machine learning model, multiple discovery predictions for the received instances of source location energy usage can be generated, the discovery predictions comprising a prediction about a presence of target device energy usage within the instances of source location energy usage. And based on the multiple discovery predictions, an overall prediction about a presence of target device energy usage within the given source location's energy usage over the period of time can be generated.
1. A method for generating machine learning predictions to discover target device energy usage, the method comprising:
storing one or more trained machine learning models configured to discover target device energy usage from source location energy usage;
receiving, for a given source location, multiple instances of source location energy usage over a period of time, wherein the multiple instances correspond to windows of time that span the period of time;
generating, using the trained machine learning model, multiple discovery predictions for the received instances of source location energy usage, each of the discovery predictions comprising a prediction about a presence of target device energy usage within the instances of source location energy usage over each of the windows of time; and
generating, by combining the multiple discovery predictions, an overall prediction about a presence of target device energy usage within the given source location's energy usage over the period of time, wherein the overall prediction is used to perform energy grid demand planning, the energy grid being controlled at least in part based on the demand planning.
2. The method of claim 1 , wherein the discovery predictions comprise a confidence data value that indicates the presence of target device energy usage within the instances of source location energy usage, and the overall prediction comprises an overall confidence data value that indicates the presence of target device energy usage within the given source location's energy usage over the period of time.
3. The method of claim 2 , wherein generating the overall prediction comprises using a mathematical function to combine the confidence data values corresponding to the multiple discovery predictions to generate the overall confidence data value.
4. The method of claim 3 , wherein the overall confidence data value is compared to a criteria or threshold, and the presence of target device energy usage is positively discovered when the overall confidence data value meets or exceeds the criteria or threshold.
5. The method of claim 2 , wherein,
the target device comprises at least one of an electric vehicle, a water heater, a washer, a dryer, a pool pump, an electric heating device, an electric cooling device, a heating, ventilation, and air conditioning (HVAC) device, and a photovoltaic device, and
the trained machine learning model is trained using data comprising features of the target device energy usage.
6. The method of claim 5 , wherein the target device comprises an electric vehicle and the target device energy usage comprises a first type of electric vehicle charging and a second type of electric vehicle charging, the first type of electric vehicle charging using a 120-volt standard charger and the second type of electric vehicle charging using a 240-volt fast charger.
7. The method of claim 2 , wherein the source locations comprise households, the instances of source location energy usage comprise household energy usage at an hourly granularity over a week of time, and the period of time comprises at least four weeks.
8. The method of claim 2 , wherein the source locations comprise households and the instances of source location energy usage comprise one of a monthly granularity, weekly granularity, and bi-monthly granularity.
9. The method of claim 2 , wherein the trained machine learning model comprises one or more recurrent neural network layers and one or more convolutional neural network layers.
10. The method of claim 2 , wherein each discovery prediction generated by the trained machine learning model is based on a corresponding instance of source location energy usage and weather conditions associated with the corresponding instance of source location energy usage.
11. The method of claim 10 , wherein the weather conditions are relative to the source location and further comprise one or more of an average temperature, a maximum temperature, a minimum temperature, a dew point, a humidity value, and solar irradiance.
12. A system for generating machine learning predictions to discover target device energy usage, the system comprising:
a processor; and
a memory storing instructions for execution by the processor, the instructions configuring the processor to:
store one or more trained machine learning models configured to discover target device energy usage from source location energy usage;
receive, for a given source location, multiple instances of source location energy usage over a period of time, wherein the multiple instances correspond to windows of time that span the period of time;
generate, using the trained machine learning model, multiple discovery predictions for the received instances of source location energy usage, each of the discovery predictions comprising a prediction about a presence of target device energy usage within the instances of source location energy usage over each of the windows of time; and
generate, by combining the multiple discovery predictions, an overall prediction about a presence of target device energy usage within the given source location's energy usage over the period of time, wherein the overall prediction is used to perform energy grid demand planning, the energy grid being controlled at least in part based on the demand planning.
13. The system of claim 12 , wherein the discovery predictions comprise a confidence data value that indicates the presence of target device energy usage within the instances of source location energy usage, and the overall prediction comprises an overall confidence data value that indicates the presence of target device energy usage within the given source location's energy usage over the period of time.
14. The system of claim 13 , wherein generating the overall prediction comprises using a mathematical function to combine the confidence data values corresponding to the multiple discovery predictions to generate the overall confidence data value.
15. The system of claim 14 , wherein the overall confidence data value is compared to a criteria or threshold, and the presence of target device energy usage is positively discovered when the overall confidence data value meets or exceeds the criteria or threshold.
16. The system of claim 13 , wherein,
the target device comprises at least one of an electric vehicle, a water heater, a washer, a dryer, a pool pump, an electric heating device, an electric cooling device, a heating, ventilation, and air conditioning (HVAC) device, and a photovoltaic device, and
the trained machine learning model is trained using data comprising features of the target device energy usage.
17. The system of claim 13 , wherein the source locations comprise households, the instances of source location energy usage comprise household energy usage at an hourly granularity over a week of time, and the period of time comprises at least four weeks.
18. The system of claim 13 , wherein the source locations comprise households and the instances of source location energy usage comprise one of a monthly granularity, weekly granularity, and bi-monthly granularity.
19. The system of claim 13 , wherein the trained machine learning model comprises one or more recurrent neural network layers and one or more convolutional neural network layers.
20. A non-transitory computer readable medium having instructions stored thereon that, when executed by a processor, cause the processor to generate machine learning predictions to discover target device energy usage, wherein, when executed, the instructions cause the processor to:
store one or more trained machine learning models configured to discover target device energy usage from source location energy usage;
receive, for a given source location, multiple instances of source location energy usage over a period of time, wherein the multiple instances correspond to windows of time that span the period of time;
generate, using the trained machine learning model, multiple discovery predictions for the received instances of source location energy usage, each of the discovery predictions comprising a prediction about a presence of target device energy usage within the instances of source location energy usage over each of the windows of time; and
generate, by combining the multiple discovery predictions, an overall prediction about a presence of target device energy usage within the given source location's energy usage over the period of time, wherein the overall prediction is used to perform energy grid demand planning, the energy grid being controlled at least in part based on the demand planning.