Interfacing to resource consumption management devices

A non-transitory computer storage medium encoded with instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising:

• maintaining, from two or more operation modes for a premises, an operation mode for the premises that includes a set of consumption endpoints that consume a resource at the premises;
• receiving, from a resource provider entity that provides the resource at the premises, control data for a direct load control event from the resource provider entity;
• in response to receipt, from the resource provider entity, of the control data for the direct load control event that indicates a shortage of the resource to be consumed at the premises at a future time;
• selecting, using the operation mode for the premises, a set of one or more policies i) to use in handling the direct load control event from the resource provider entity and ii) that define a resource priority and actions for at least one consumption endpoint in the set of consumption endpoints based on the shortage of the resource to be consumed at the premises identified in the direct load control event, wherein, for each of the two or more operation modes, a different set of policies is selected and the selected set of one or more policies is different from the operation mode for the premises, and
• predicting, using machine learning techniques related to historical analysis, feedback, and previous inferences, one or more inferences a) related to operation of the set of consumption endpoints at the future time of the direct load control event and b) that indicate a forecast of consumption of the resource by one or more of the set of consumption endpoints at the future time of the direct load control event;
• based on the selected set of one or more policies to use in handling the direct load control event from the resource provider entity, the operation mode for the premises, and the one or more inferences related to operation of the set of consumption endpoints at the future time of the direct load control event, determining an adjustment for operation of the set of consumption endpoints to implement the direct load control event;
• and
• using the determined adjustment for operation of the set of consumption endpoints to implement the direct load control event, determining one or more commands to control operation of the set of consumption endpoints in accordance with the selected set of one or more policies to use in handling the direct load control event from the resource provider entity and the one or more inferences related to operation of the set of consumption endpoints at the future time of the direct load control event.

A method comprising:

• maintaining, from two or more operation modes for a premises, an operation mode for the premises that includes a set of consumption endpoints that consume a resource at the premises;
• receiving, from a resource provider entity that provides the resource at the premises, control data for a direct load control event from the resource provider entity;
• in response to receipt, from the resource provider entity, of the control data for the direct load control event that indicates a shortage of the resource to be consumed at the premises at a future time;
• selecting, using the operation mode for the premises, a set of one or more policies i) to use in handling the direct load control event from the resource provider entity and ii) that define a resource priority and actions for at least one consumption endpoint in the set of consumption endpoints based on the shortage of the resource to be consumed at the premises identified in the direct load control event, wherein, for each of the two or more operation modes, a different set of policies is selected and the selected set of one or more policies is different from the operation mode for the premises, and
• predicting, using machine learning techniques related to historical analysis, feedback, and previous inferences, one or more inferences a) related to operation of the set of consumption endpoints at the future time of the direct load control event and b) that indicate a forecast of consumption of the resource by one or more of the set of consumption endpoints at the future time of the direct load control event;
• based on the selected set of one or more policies to use in handling the direct load control event from the resource provider entity, the operation mode for the premises, and the one or more inferences related to operation of the set of consumption endpoints at the future time of the direct load control event, determining an adjustment for operation of the set of consumption endpoints to implement the direct load control event;
• and
• using the determined adjustment for operation of the set of consumption endpoints to implement the direct load control event, determining one or more commands to control operation of the set of consumption endpoints in accordance with the selected set of one or more policies to use in handling the direct load control event from the resource provider entity and the one or more inferences related to operation of the set of consumption endpoints at the future time of the direct load control event.