Grid Energy Storage Control Using Prediction Reliability Windows
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing power management systems struggle to accurately predict power demand and generation, leading to improper energy management, such as charging or discharging energy storage devices when actual power generation exceeds demand.
Innovation Solution
A power management system that charges or discharges energy storage devices during time periods with high reliability of power generation and demand prediction, minimizing unnecessary charging and discharging by setting charge and discharge instructions based on reliability, and adjusting these instructions for low reliability periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If power management is performed based on predicted power generation and demand, then energy management can be planned in advance, but prediction accuracy is insufficient leading to improper charging/discharging operations
Solution Approach 1:
The patent divides the prediction period into multiple time periods with different reliability levels. By segmenting the prediction horizon and assigning different reliability weights to each segment, the system can make more accurate charging/discharging decisions for high-reliability periods while avoiding unnecessary operations in low-reliability periods, thus reducing energy loss from improper operations.
Solution Approach 2:
The patent introduces a reliability parameter that varies by time period to modify the charging/discharging control strategy. By changing the reliability parameter based on prediction confidence levels for different time periods, the system adjusts its operational behavior dynamically, performing energy management actions only when prediction reliability is sufficient, thereby reducing unnecessary energy exchanges.
2Productivity
If charge and discharge operations are performed frequently to maintain reference SOC, then energy management responsiveness improves, but unnecessary operations increase when predictions are inaccurate
Solution Approach 1:
The patent makes the charging/discharging control dynamic by incorporating time-varying reliability assessments. Instead of applying a fixed control strategy, the system dynamically adjusts its responsiveness based on the predicted reliability for each time period, enabling aggressive management when reliable and conservative management when uncertain, thus optimizing both productivity and energy efficiency.
Solution Approach 2:
The patent implements a feedback mechanism where prediction reliability information feeds back into the charging/discharging control decisions. By continuously assessing prediction reliability and using this information to modulate operational intensity, the system creates a closed-loop control that adapts to prediction quality, reducing unnecessary operations while maintaining responsiveness when needed.
Data Source
AI summary
A power management system for managing at least one power generation system connected to an electrical grid and at least one energy storage device connected to the electrical grid, wherein the energy storage device is charged or discharged in a time period in which the reliability of the prediction of the amount of power generation and the power demand per time period is high based on a predicted amount of power generation which is an amount of power generation of the power generation system predicted for each time period and a predicted power demand which is a power demand for the electrical grid. Thus, more proper energy management can be performed.