System and method for decentralized energy production

Patent metadata

Field	Value
Publication number	US-10504195-B2
Application number	US-201715810262-A
Country	US
Kind code	B2
Filing date	Nov 13, 2017
Priority date	Nov 13, 2017
Publication date	Dec 10, 2019
Grant date	Dec 10, 2019

Abstract

A power generation system of a power producer for generating and providing electricity to an electric power system operated by an independent system operator (ISO). A processor determines a residual demand curve, at each time step of a future energy generation period, relating to a price and a quantity of a demanded energy for that time step. Determine a residual reserve curve at each time step of the future energy generation period, relating to a price and a quantity of a reserved energy for that time step. Determine a schedule of operations of a set of generators for each time step of the future energy generation period, by the power producer. A controller to control the set of generators to produce energy according to the schedule of operations, when the receiver receives an acceptance of the schedule of operations from the ISO.

First claim

What is claimed is:

1. A power generation system of a power producer for generating and providing electricity to an electric power system operated by an independent system operator (ISO), wherein multiple power producers supply the electricity to the electric power system and multiple power consumers demand the electricity from the electric power system, the power generation system of the power producer comprising:
a set of generators of the power producer;
a receiver to receive a forecast of a demanded energy quantity for the electric power system from the ISO for a future energy generation period, and to receive a forecast of a reserved energy quantity for the electric power system from the ISO for the future energy generation period;
a processor in communication with the receiver, is configured to
determine a residual demand curve, at each time step of the future energy generation period, relating to a price and a quantity of the demanded energy for that time step, based on the received forecast of the demanded energy for each time step, wherein the residual demand curve includes pluralities of sections, such that a section includes a most likely quantity of the demanded energy for the power producer;
determine a residual reserve curve at each time step of the future energy generation period, relating to a price and a quantity of the reserved energy for that time step, based on the received forecast of the reserved energy for each time step, wherein the residual reserve curve includes pluralities of sections, such that a section includes a most likely quantity of the reserved energy for the power producer;
determine a schedule of operations of the set of generators for each time step of the future energy generation period, by optimizing a value function of a difference between a revenue from generating the demanded energy and the reserved energy, and a cost of producing the generated energy, by the power producer; and

a controller to control the set of generators to produce energy according to the schedule of operations, when the receiver receives an acceptance of the schedule of operations from the ISO.

2. The power generation system of claim 1 , wherein the residual demand curve for each time step is determined by subtracting a combined demanded energy supply offer curve from other power producers of the electric power system from the received forecast of the demanded energy.

3. The power generation system of claim 2 , wherein the combined demanded energy supply offer curve for each time step for the future energy generation period, is determined by
access a memory by the processor, wherein the memory has stored thereon a set of neural networks trained using historical data of demanded energy supply offer curves from other power producers of the electric power system and historical data of demanded energy and market clearing demand prices of the electric power system, such that each trained neural network output corresponds to a range of market clearing demand prices;
select a neural network from the ranges of market clearing demand prices of the set of stored neural networks, based on a most likely market clearing demand price estimated from the received forecast of the demanded energy of the electric power system; and
produce the combined demanded energy supply offer curve using the selected neural network, based on an input of the received forecast of the demanded energy of the electric power system, and the most likely market clearing demand price.

4. The power generation system of claim 3 , wherein the most likely market clearing demand price of the electric power system, is determined by
access the memory by the processor, wherein the memory has stored thereon a neural network trained using historical data of cleared demanded energy and market clearing demand prices of the electric power system; and
produce the most likely market clearing demand price of the electric power system using the neural network, based on an input of the received forecast of the demanded energy of the electric power system and a stored demanded energy and a stored market clearing demand price of the electric power system for the current energy generation period in the memory.

5. The power generation system of claim 1 , wherein the residual reserve curve for each time step is determined by subtracting a combined reserved energy supply offer curve from other power producers of the electric power system from the received forecast of the reserved energy.

6. The power generation system of claim 5 , wherein the combined reserved energy supply offer curve for each time step for the future energy generation period, is determined by
access a memory, by the processor, wherein the memory has stored thereon a set of neural networks trained using historical data of reserved energy supply offer curves from other power producers of the electric power system and historical data of reserved energy and market clearing reserve prices of the electric power system, such that each trained neural network output corresponds to a range of market clearing reserve prices;
select a neural network from the ranges of market clearing reserve prices of the set of stored neural networks, based on a most likely market clearing reserve price estimated from the received forecast of the reserved energy of the electric power system; and
produce the combined reserved energy supply offer curve using the selected neural network, based on an input of the received forecast of the reserved energy of the electric power system, and the most likely market clearing reserve price.

7. The power generation system of claim 6 , wherein the most likely market clearing reserve price of the electric power system, is determined by
access the memory by the processor, wherein the memory has stored thereon a neural network trained using historical data of cleared reserved energy and market clearing reserve prices of the electric power system; and
produce the most likely market clearing reserve price of the electric power system using the neural network, based on an input of the received forecast of the reserved energy of the electric power system, and a stored reserved energy and a stored market clearing reserve price of the electric power system for the current energy generation period in the memory.

8. The power generation system of claim 1 , further comprising:
determining a most likely demanded energy quantity, a most likely reserved energy quantity, and a most likely reserve used ratio for the power producer for each time step of the future energy generation period, by
access the memory by the processor, wherein the memory has stored thereon a neural network trained using historical data of the demanded energy, the reserved energy, a market clearing demand price and a market clearing reserve price for the electric power system and historical data of the demanded energy, the reserved energy and the reserve used ratio for the power producer; and
produce the most likely demanded energy quantity, the most likely reserved energy quantity and the most likely reserve used ratio, using the neural network, based on an input of the received forecasts of the demanded energy and the reserved energy for the electric power system, the most likely market clearing demand price and the most likely market clearing reserve price for the electric power system.

9. The power generation system of claim 8 , wherein the sections of the residual demand curve and the sections of the residual reserve curve for each time step of the future energy generation period are reduced to a set of combinations of a state of likely demanded energy for the power producer, and a state of likely reserved energy for the power producer

Assignees

• Mitsubishi Electric Res Laboratories Inc

Inventors

• Sun Hongbo
• Keerthisinghe Chanaka

Classifications

• H02J3/003

  Load forecast, e.g. methods or systems for forecasting future load demand · CPC title

• H02J3/00

  Circuit arrangements for AC mains or AC distribution networks · CPC title

• H02J3/008

  Circuit arrangements for power supply or distribution technologies responsive to energy trading · CPC title

• G06Q30/0206

  Price or cost determination based on market factors · CPC title

• G06Q50/06 Primary

  Energy or water supply · CPC title