A method and an apparatus for reinforcement learning based energy bidding, adapted for an energy aggregator to determine the energy supply configuration between multiple energy suppliers and multiple energy demanders, are provided. In the method, a supply amount of each energy supplier and a demand amount of each energy demander are acquired. A total demand amount of the energy demanders is calculated and replied to each energy supplier, and a total supply amount of the energy suppliers is calculated and replied to each energy demander. An electricity purchase quotation determined by each energy demander according to respective demand amount and the total supply amount, and an electricity sale quotation determined by each energy supplier according to respective supply amount and the total demand amount are received. A linear programming method is adopted to determine the energy supply configuration between the energy suppliers and the energy demanders according to information.

A central plant includes storage devices configured to store resources purchased from a utility or generated by the central plant and to discharge the one or more resources. The central plant includes a controller configured to obtain a cost function comprising a peak load contribution (PLC) term representing a cost based on an amount of the one or more resources purchased during coincidental peak (CP) subperiods. The controller is configured to modify the cost function by applying a peak subperiods mask to the PLC term, wherein, for each subperiod, the peak subperiods mask modifies a portion of the PLC term corresponding to the subperiod based on a probability that the subperiod will be one of the CP subperiods. The controller is also configured to perform an optimization of the modified cost function.

A method and device for predicting a thermal load of an electrical system are provided. The method includes: S1: pre-processing historical daily data of the thermal load of an electrical system. S2: acquiring a data daily reference line according to pre-processed historical daily data. S3: dividing acquired data daily reference line into a plurality of time sections. S4: screening the historical daily data, and calculating a trend similarity value of screened historical daily data and the data daily reference line within each divided time section of the plurality of time sections respectively. S5: choosing the historical daily data corresponding to the trend similarity value greater than a preset reference value to form a similarity sequence matrix. S6: inputting the similarity sequence matrix into an extreme learning machine (ELM) for training, acquiring a prediction model, and predicting the thermal load of the electrical system.

A method for performing validation, estimation, and editing (VEE), including: displaying real-time electrical usage for a building on a controllable video display; executing VEE rules on each of the streams to generate and store a corresponding post VEE readings, the post VEE readings comprising tagged energy consumption data sets each associated with a corresponding one of the streams, each of the data sets comprising groups of contiguous interval values tagged as having been validated and corresponding to correct data; for the each of the data sets, creating anomalies having different durations using only the groups of contiguous interval values; generating estimates for the anomalies by employing estimation techniques; for each of the durations, selecting one of the estimation techniques for subsequent employment when performing VEE of subsequent energy consumption data for the corresponding one of the streams; and executing functions on the streams translated by the generating and directing the controllable video display to display a weather normalized usage baseline recommendation for action regarding current energy usage.

A grid distribution system aggregates energy resources of multiple distributed energy resources (DERs) and provides service to one or more energy markets with the DERs as a single market resource. The DERs can create data to indicate realtime local demand and local energy capacity of the DERs. Based on DER information and realtime market information, the system can compute how to provide one or more services to the power grid based on an aggregation of DER energy capacity.

A power management device has power consumption information in a building of a power customer, and stores history information in which the acquired power consumption information and time information indicating a time frame in which power has been consumed in the building are associated with each other. When having received a request based on power supply and demand, the power management device determines ease or difficulty in management of power consumption according to the request, on the basis of the stored history information, and power-management-required time frame information indicating a time frame for which power management according to the request is required, and notifies the power customer of a determination result. Then the power management device receives acceptance or rejection of the request, and transmits response information corresponding to a received content, to a request source.

A power flow schedule of a cluster is determined by calculating sensitivity of the net power exchange bounds. Each cluster includes a different section of the power system. The cluster provides to another cluster the power flow schedule and the net power exchange bounds for determination of a second power flow schedule by another cluster based on collective net power exchange bounds, a forecast power supply of the plurality of clusters, and a forecast power demand schedule. The clusters are hierarchically arranged such that the another cluster is higher in a hierarchy than the cluster. The cluster receives from the another cluster the second power flow schedule. The first power flow schedule is adjusted locally within the cluster based on the second power schedule and the net power exchange bounds of the cluster. The power output of the cluster is controlled using the adjusted first power flow schedule.

The present disclosure discloses a method for optimizing equipment capacity and equipment power of an energy hub system. The method includes establishing an energy hub model containing natural gas boilers, electric boilers, coolers and heat pumps, establishing a bilevel optimized upper model to solve the optimal heat pump capacity, and establishing a bilevel optimized lower model to solve the optimal power utilization of each energy device based on the binary search algorithm of the quadratic function solves the upper model by using the multi-objective evolutionary algorithm NSGA-II to solve the lower model. The optimization method of the present invention can solve the multi-objective bilevel model problem without the help of commercial optimization software. Obtaining a reasonable, efficient and green planning scheme makes the total operating cost and total exhaust gas emissions of the energy hub relatively optimal.

A system and method to join distributed energy resources (DER) to achieve common objectives is provided. The present technology organizes and/or aggregates DERs by routing a (DER) program request for resources to DER contributors capable of responding to and performing the request using a routing system. The system accesses a plurality of DER profiles, each profile associated with a DER contributor capable of contributing a resource to the request, and calculates an initial value for each DER profile based on request attributes and scoring metrics associated with the profile. The system then calculates a fitness metric for each DER profile based on the initial value using a neural network having weights based on the plurality of performance indicators and selects the DER profile and contributors to whom to route the request.

Provided are techniques for predictively generating demand response (DR) events based on grid operations and faults, based on optimization and self-learning routines. The techniques include obtaining grid status information, and determining a fault condition based at least in part on the grid status information. The techniques also include generating a DR event based at least in part on the fault condition, and responsive to generating the DR event, transmitting a notification of the DR event.