Method and system for demand-response signal assignment in power distribution systems

Patent metadata

Field	Value
Publication number	US-11303124-B2
Application number	US-201815912610-A
Country	US
Kind code	B2
Filing date	Mar 6, 2018
Priority date	Dec 18, 2017
Publication date	Apr 12, 2022
Grant date	Apr 12, 2022

Abstract

A method for providing a demand-response (DR) signal assignment includes initializing reinforcement learning (RL) agents, each having an exploration scheme and being assigned to a consumer, with initial models about respective RL environments of the consumers. The RL agents send the models to an optimizer. The optimizer computes a DR signal assignment using the models. The RL agents send DR signals in accordance with the DR signal assignment to the consumers so as to aggregately achieve a load reduction by the consumers over a time window. The RL agents monitor the consumers over the time window, and update the models based on the monitoring.

First claim

What is claimed is:

1. A method for providing a demand-response (DR) signal assignment, the method comprising:
a) initializing reinforcement learning (RL) agents, each having an exploration scheme and being assigned to a consumer, with initial models about respective RL environments of the consumers;
b) sending, by the RL agents, the models to an optimizer;
c) computing, by the optimizer, a DR signal assignment using the models;
d) sending, by the RL agents, DR signals in accordance with the DR signal assignment to the consumers so as to aggregately achieve a load reduction by the consumers over a time window;
e) monitoring, by the RL agents over the time window, the consumers; and
f) updating, by the RL agents, the models based on the monitoring.

2. The method according to claim 1 , wherein the RL agents monitor load curves of the respective consumers to update the models.

3. The method according to claim 1 , further comprising sending, by the RL agents, the updated models to the optimizer and repeating steps c)-f).

4. The method according to claim 3 , wherein a discrete load reduction for each of the consumers is determined for each time window such that multiple objectives are defined for the RL environments, the discrete load reduction representing a minimum load reduction achieved by the respective consumers for each time window.

5. The method according to claim 4 , wherein the discrete load reductions are summed to determine a lower bound on the aggregated load reduction.

6. The method according to claim 1 , wherein at least one of the RL agents is assigned to multiple consumers that are virtually grouped together.

7. The method according to claim 1 , wherein the models indicate respective confidence levels of the RL agents in the models which are used by the optimizer to compute the DR signal assignment.

8. The method according to claim 1 , wherein the RL agents are disposed decentralized at respective locations of the consumers.

9. The method according to claim 8 , wherein the consumers are individual households having electrical appliances capable of providing a load reduction over the time window.

10. A reinforcement learning (RL) agent for providing a demand-response (DR) signal assignment, the RL agent having an exploration scheme and being assigned to a consumer, the RL agent comprising one or more processors which, alone or in combination, are configured to provide for execution of the following steps:
a) sending an initial model about a RL environment of the consumer to an optimizer;
b) receiving a DR signal assignment computed by the optimizer using the initial model and further additional models of further RL agents;
c) sending a DR signal to the consumer in accordance with the DR signal such that the consumer contributes to an aggregated load reduction achieved over a time window together with further consumers;
d) monitoring the consumer over the time window; and
e) updating the initial model based on the monitoring.

11. The RL agent according to claim 10 , wherein the RL agent is configured to monitor a load curve of the consumer to update the initial model.

12. The RL agent according to claim 10 , being further configured to send the updated model to the optimizer.

13. The RL agent according to claim 10 , wherein the model indicates a confidence level of the RL agent in the model which is usable by the optimizer to compute the DR signal assignment.

14. The RL agent according to claim 10 , wherein the RL agents is disposed at a location of the consumer.

15. The RL agent according to claim 10 , wherein the consumer is an individual household having electrical appliances capable of providing a load reduction over the time window.

Assignees

• NEC Laboratories Europe GmbH
• Nec Corp

Inventors

• Jacobs Tobias
• Schmidt Mischa

Classifications

• H02J2105/42

  Home appliances · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• H02J2105/12

  supplying households or buildings · CPC title

• H02J3/14 Primary

  by switching loads on to, or off from, the networks, e.g. progressively balanced loading · CPC title

• G06N3/006 Primary

  based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] · CPC title