Abstract: A method, system, and device for controlling energy storage devices are provided, the method including receiving a trained machine learning model from a centralized machine learning system, recording temporal data for a respective energy storage device, periodically transmitting the temporal data to the machine learning system, performing a mode prediction for controlling the energy storage device using the trained machine learning model and the temporal data, and sending a control signal to the energy storage device to operate in the predicted mode. The machine learning system aggregates the temporal data transmitted by each agent and uses the aggregated temporal data to update the machine learning model. By using aggregated temporal data, less data is needed from an individual energy storage device so that when a new energy storage device joins the machine learning system, the new energy storage device can benefit from increased performance with less computation.

Abstract: A method, system, and device for controlling energy storage devices are provided, the method including receiving a trained machine learning model from a centralized machine learning system, recording temporal data for a respective energy storage device, periodically transmitting the temporal data to the machine learning system, performing a mode prediction for controlling the energy storage device using the trained machine learning model and the temporal data, and sending a control signal to the energy storage device to operate in the predicted mode. The machine learning system aggregates the temporal data transmitted by each agent and uses the aggregated temporal data to update the machine learning model. By using aggregated temporal data, less data is needed from an individual energy storage device so that when a new energy storage device joins the machine learning system, the new energy storage device can benefit from increased performance with less computation.

Abstract: A method and an apparatus for battery modelling. The method includes acquiring battery information from a sensor attached to a battery, the battery information including at least a terminal voltage and a load current; classifying, the battery information as effective data or ineffective data based on a phase of a battery cycle during which the battery information is acquired; identifying one or more parameters of the circuit model associated with the battery based on the effective data; and generating an estimation of a state of the battery using the circuit model having the one or more parameters identified using the effective data. Further, a circuit model is identified using effective data.

Abstract: A method and an apparatus for battery modelling. The method includes acquiring battery information from a sensor attached to a battery, the battery information including at least a terminal voltage and a load current; classifying, the battery information as effective data or ineffective data based on a phase of a battery cycle during which the battery information is acquired; identifying one or more parameters of the circuit model associated with the battery based on the effective data; and generating an estimation of a state of the battery using the circuit model having the one or more parameters identified using the effective data. Further, a circuit model is identified using effective data.

Abstract: The invention relates to an electronic management system (5) for electricity generating cells (3), the system comprising: cell connection terminals to be connected to n associated electricity generating cells (3), n being a positive integer number, outputs to be connected to m associated static converters (9); m being a positive integer number and at least m=2, an energy routing module (13) adapted for routing energy flows from and between said cell connection terminals towards said outputs; and an electronic control unit (15) adapted for controlling dynamically the energy routing module (13).

Abstract: The invention relates to an electronic management system (5) for electricity generating cells (3), the system comprising: cell connection terminals to be connected to n associated electricity generating cells (3), n being a positive integer number, outputs to be connected to m associated static converters (9); m being a positive integer number and at least m=2, an energy routing module (13) adapted for routing energy flows from and between said cell connection terminals towards said outputs; and an electronic control unit (15) adapted for controlling dynamically the energy routing module (13).