Methods, systems, and apparatus, including medium-encoded computer program products, for configuring location-specific electrical load models while preserving privacy. A first machine learning model configured to predict electrical load curves of an electrical utility grid can be obtained from a server. Load values associated with a particular region of the electrical utility grid can be obtained. The load values can be applied as calibration input to the first machine learning model to produce first adjustment parameters for the first machine learning model. The first adjustment parameters can be provided to the server.

Systems/methods for managing a load center uses a thermal model of an overcurrent protection device to adjust the load center energy level. The systems/methods provide an energy management system that manages the amount of load energy in the load center to allow the load center to operate in an overloaded state without inducing thermal tripping of the load center. The energy management system determines a thermal shedding time that provides a margin or threshold against the thermal tripping, then determines, for a given load center energy level, whether the load center has been operating in an overloaded state for a time that equals or exceeds the thermal shedding time. The energy management system thereafter determines any load shedding that may be needed to reduce the load energy level. This arrangement allows the load center to remain in an overloaded state without thermal tripping.

Systems/methods for managing a load center uses a thermal model of an overcurrent protection device to adjust the load center energy level. The systems/methods provide an energy management system that manages the amount of load energy in the load center, adjusting as needed for panel temperature in the load center, to allow the load center to operate in an overloaded state without inducing thermal tripping. The system determines a thermal shedding time that provides a margin or threshold against the thermal tripping, then determines, for a given load center temperature and energy level, whether the load center has been operating in an overloaded state for a time that equals or exceeds the thermal shedding time. The system thereafter determines any load shedding that may be needed to reduce the load energy level. This arrangement allows the load center to remain in an overloaded state without thermal tripping.

An electrical network includes a plurality of distribution transformers, a plurality of network protectors each electrically connected to one of the distribution transformers, a secondary bus electrically connected to each of the plurality of network protectors, and a control unit. The control unit is configured to receive data from each of the plurality of network protectors associated with electricity flowing therethrough, analyze the data to determine whether an electrical imbalance exists at any of the network protectors, and operate any imbalanced network protectors to disconnect the distribution transformers connected to the imbalanced network protector from the secondary bus. The control unit is further configured to permit beneficial backflow through the distribution transformers that may be created by generators coupled to the secondary bus.

The energy management system and method provide for the control of electrical loads within a group. In one embodiment, a plurality of electrical loads are connected to an electrical grid and, upon detection of a predetermined condition, a user-defined sub-set of the electrical loads are disconnected from the electrical grid. The predetermined condition may be, for example, a blackout, thus when an interruption in power from the grid is detected by one or more sensors, the user-defined sub-set of the electrical loads is electrically disconnected from its electrical connection to the grid power. Following this disconnection, at least one of the electrical loads from the user-defined sub-set of the electrical loads is connected to an alternative source of power, such as a backup battery, a solar power system, a generator or the like.

A grid-forming energy storage system includes a data interface and transfer module, a state analysis and prediction module, an event response and strategy adjustment module, and an energy scheduling and management module. In the present disclosure, the energy management efficiency in the smart grid is significantly improved through real-time data formatting and efficient information transfer, real-time transmission and accuracy of data are ensured by using a JSON formatted data interface and a TCP/IP protocol, the data processing flow is optimized in conjunction with Apache Kafka, and the throughput capacity is improved and the response time is shortened, such that the grid manages resources more efficiently. Through state analysis and demand prediction by Spark Streaming, the grid responds in real time and predicts energy demand changes to reduce the energy waste.

A method identifies asymmetrical vibrations during the operation of an electric device which is connected to a high-voltage grid. Vibrations of the electric device are detected using vibration sensors which provide measured values on the output side. The measured values and/or values derived from the measured values are transmitted to a communication unit via a close-range communication connection. The measured values and/or the values are transmitted by the communication unit to a data processing cloud via a far-range communication connection. The measured values are separated into frequency components by the data processing cloud using a Fourier transformation, thereby obtaining a frequency spectrum. Odd and even frequency components of the frequency spectrum are ascertained based on a base frequency of the high-voltage supply grid and put into a ratio R relative to one another. The presence of asymmetrical vibrations is indicated if the ratio R exceeds a specified threshold.

A system for determining a linear state estimation of a power distribution network. The system includes a plurality of metering devices, a plurality of power sources, each of the plurality of power sources coupled to a respective metering device, and a plurality of collection devices, each of the plurality of collection devices in electronic communication with a respective metering device. The plurality of collection devices measure an Advanced Metering Infrastructure (AMI) dataset and calculate a phase angle of the power source coupled to the metering device. The system includes a host device in electronic communication with the plurality of collection devices. An electronic processor of the host device receives the AMI dataset and the phase angle of the plurality of metering devices, determines the linear state estimation of the power distribution network based on the phase angle and the AMI dataset, and identifies information associated with the power distribution network.

A meter socket adapter (MSA) for connecting a battery energy storage system (BESS) to a household microgrid through a meter combo load center. The MSA includes: a housing matching an electricity meter and a meter socket in the meter combo load center; a plurality of power conductors that connect among the electricity meter, the meter socket, and BESS; a meter interconnection device (MID) switch for connecting a utility grid or one or more distribution energy resource (DER) devices to supply power to household loads in the household microgrid; a MID driving circuit for driving the MID switch to close or open; and a MID connection detection circuit to verify an electrical connection of a closed MID switch.

Systems and methods for estimating energy consumption data for a building. A system for estimating an energy use of a building uses a dilated convolutional neural network architecture to receive time-series data for the building and to predict one or more time-series data points representing an estimated energy consumption for the building. A method for estimating an energy use of a building includes obtaining time-series data for the building, providing the time-series data as input to a dilated convolutional neural network architecture, and predicting one or more time-series data points representing an estimated energy consumption for the building using the dilated convolutional neural network architecture. The systems and methods may be used to help users and building controllers reduce energy use within a building.