System and method for real-time distributed micro-grid optimization using price signals
A system and method for providing real-time distributed micro-grid optimization using price signals to the electrical grid system by allowing bi-directional electricity usage from a distributed network of energy storage stations to form a large, distributed resource for the grid. A machine learning optimization module ingests various forms of data—from grid telemetry to traffic data to trip-to-trip data and more-in order to make informed spatiotemporal decisions about optimal pricing signals as well as strategically placing and balancing energy stores across various regions to support optimum energy usage, risk mitigation, grid fortification, and revenue generation. Energy stores are then sent updated price signals and updated parameters as to the amount of energy to hold or release.
1 . A system for real-time distributed micro-grid optimization using price signals, comprising:
a computing device comprising a memory, a processor, and a non-volatile data storage device;
a pricing engine comprising a first plurality of programming instructions stored in the memory of, and operating on the processor of, the computing device, wherein the first plurality of programming instructions, when operating on the processor, cause the computing device to:
obtain a plurality of spatiotemporal information relating to an electrical grid, real-time traffic flow, and real-time energy demand;
spatially align the plurality of spatiotemporal information into one or more regions;
determine a load profile for each of the one or more specified regions, wherein the load profile is determined using real-time traffic flow and energy demand data;
receive an overall risk score for each of the one or more regions;
compute a state of charge and a maximum revenue for each of the one or more regions;
use the plurality of spatiotemporal information, the spatially aligned information, the load profile, the computed state of charge, the overall risk score, and the computed maximum revenue as inputs into a neural network configured to generate as output a predicted price signal for each of the one or more regions; and
distribute the price signal to each corresponding region of the one or more regions, wherein the price signal is used to configure the operation of an energy storage system within the corresponding region.
2 . The system of claim 1 , further comprising an optimization engine comprising a second plurality of programming instructions stored in the memory of, and operating on the processor of, the computing device, wherein the second plurality of programming instructions, when operating on the processor, cause the computing device to:
retrieve training data relating to the electrical grid and components of the electric grid for each specified region of the one or more regions;
retrieve training data relating to the climate and weather for a plurality of specified regions;
retrieve training data relating to socio-economic factors for a plurality of specified regions;
retrieve training data relating to local behavior patterns for a plurality of specified regions;
use the training data for each of the specified regions in the plurality of specified regions to calculate an overall risk score for each specified region; and
send the overall risk score for each of the one or more regions to the pricing engine.
3 . The system of claim 2 , wherein the optimization engine is further configured to:
retrieve a status on the current energy demands for each of the specified regions in the plurality of specified regions;
receive monitored health and status data from a controller;
use the overall risk scores and the current energy demands for each of the specified regions and the electronic health and status data from a controller to optimize the energy storage of a high-voltage battery pack; and
send a control signal to the controller, wherein the control signal causes the high-voltage battery pack to store more power, store less power, or maintain the current level of stored power.
4 . The system of claim 1 , wherein the pricing engine is further configured to:
construct and maintain a second neural network configured to determine the complex relationships between temporal and spatial data;
use the plurality spatiotemporal data, the spatially aligned data, the load profile, and overall risk scores as input into the second neural network, wherein the second neural network generates as output context-rich data relating temporal and spatial data together.
5 . The system of claim 1 , wherein the neural network is a deep reinforcement neural network.
6 . The system of claim 1 , wherein the one or more regions are represented as traffic analysis zones.
7 . The system of claim 1 , wherein the controller sets a time of charging of a high-voltage battery based on a grid condition.
8 . The system of claim 7 , wherein the grid condition is a time period wherein electricity rates are low.
9 . The system of claim 7 , wherein the grid condition is a time period when a proportion of energy supply to the grid from renewable energy sources is high.
10 . The system of claim 1 , wherein risk scores are calculated using one or more machine learning models.
11 . A method for real-time distributed micro-grid optimization using price signals, comprising the steps of:
using a pricing engine comprising a first plurality of programming instructions stored in a memory of, and operating on a processor of, a computing device:
obtaining a plurality of spatiotemporal information relating to an electrical grid, real-time traffic flow, and real-time energy demand;
spatially aligning the plurality of spatiotemporal information into one or more regions;
determining a load profile for each on the one or more specified regions, wherein the load profile is determined using real-time traffic flow and energy demand data;
receiving an overall risk score for each of the one or more regions;
computing a state of charge and a maximum revenue for each of the one or more regions;
using the plurality of spatiotemporal information, the spatially aligned information, the load profile, the computed state of charge, the overall risk score, and the computed maximum revenue as inputs into a neural network configured to generate as output a predicted price signal for each of the one or more regions; and
distributing the price signal to each corresponding region of the one or more regions, wherein the price signal is used to configure the operation of an energy storage system within the corresponding region.
12 . The method of claim 11 , further comprising the steps of:
using an optimization engine comprising a second plurality of programming instructions stored in the memory of, and operating on the processor of, the computing device:
retrieving training data relating to the electrical grid and components of the electric grid for each specified region of the one or more regions;
retrieving training data relating to the climate and weather for a plurality of specified regions;
retrieving training data relating to socio-economic factors for a plurality of specified regions;
retrieving training data relating to local behavior patterns for a plurality of specified regions;
using the training data for each of the specified regions in the plurality of specified regions to calculate an overall risk score for each specified region; and
sending the overall risk score for each of the one or more regions to the pricing engine.
13 . The method of claim 12 , further comprising the steps of:
using the optimization engine:
retrieving a status on the current energy demands for each of the specified regions in the plurality of specified regions;
receiving monitored health and status data from a controller;
using the overall risk scores and the current energy demands for each of the specified regions and the electronic health and status data from a controller to optimize the energy storage of a high-voltage battery pack; and
sending a control signal to the controller, the control signal causes the high-voltage battery pack to store more power, store less power, or maintain the current level of stored power.
14 . The method of claim 11 , further comprising the steps of:
using the pricing engine:
constructing and maintaining a second neural network configured to determine the complex relationships between temporal and spatial data;
using the plurality spatiotemporal data, the spatially aligned data, the load profile, and overall risk scores as input into the second neural network, wherein the second neural network generates as output context-rich data relating temporal and spatial data together.
15 . The method of claim 11 , wherein the neural network is a deep reinforcement neural network.
16 . The method of claim 11 , wherein the one or more regions are represented as traffic analysis zones.
17 . The method of claim 11 , wherein the controller sets a time of charging of a high-voltage battery based on a grid condition.
18 . The method of claim 17 , wherein the grid condition is a time period wherein electricity rates are low.
19 . The method of claim 17 , wherein the grid condition is a time period when a proportion of energy supply to the grid from renewable energy sources is high.
20 . The method of claim 11 , wherein risk scores are calculated using one or more machine learning models.