Abstract: A method of controlling assets connected to an electricity distribution grid is disclosed. For each of a plurality of assets to be configured, a trained neural network model is provided, which takes an asset configuration as input, the asset configuration specifying a response of the asset to variations in one or more operating conditions detected at the asset, and outputs one or more performance indicators relating to the operation of the asset when operated using the asset configuration. The outputs of the neural network models are provided as inputs to an optimization function. A search process optimizes the optimization function by varying asset configurations for the assets. The final asset configurations are then used to control the assets to provide a coordinated demand response service.

Abstract: Energy loads, sources or batteries exchange mathematical models with each other to form clusters of devices that together provide a service (self-reliance, frequency control, etc.) to a grid operator. Models are exchanged before or after forming clusters; a particular model is used to control its own device and is also used by another load/source to influence its control policy. Heuristics and an optimization technique (using models) are used to form a cluster of devices. Exchanging models obviates the need for a central entity to directly control loads/sources, and the need to exchange real-time data between loads/sources, providing resilience against communication failure. A service manager (demand-response aggregator) sends a service or technical constraints to loads/sources to form clusters on their own. Negotiation between manager and clusters occurs to form consensus on a response.

Abstract: Energy loads, sources or batteries exchange mathematical models with each other to form clusters of devices that together provide a service (self-reliance, frequency control, etc.) to a grid operator. Models are exchanged before or after forming clusters; a particular model is used to control its own device and is also used by another load/source to influence its control policy. Heuristics and an optimization technique (using models) are used to form a cluster of devices. Exchanging models obviates the need for a central entity to directly control loads/sources, and the need to exchange real-time data between loads/sources, providing resilience against communication failure. A service manager (demand-response aggregator) sends a service or technical constraints to loads/sources to form clusters on their own. Negotiation between manager and clusters occurs to form consensus on a response.