Digital Twin Refinement for Renewable Plant Prediction Accuracy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Inaccurate predictions from digital twins of renewable energy systems lead to inefficient control and increased degradation of system components, reducing the overall efficiency and consistency of power delivery.

Innovation Solution

A machine-learning model is used to refine the parameters of a digital twin by comparing predicted and actual outputs, adjusting parameters to minimize discrepancies, and identifying issues such as component degradation and sub-optimal control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If digital twin parameters are used without refinement, then the system is simple to operate, but prediction accuracy deteriorates

Engineering Contradiction:

Improveprediction accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system implements feedback by continuously comparing actual system output with digital twin predictions, using machine learning to identify discrepancies, and automatically refining digital twin parameters to reduce prediction errors, creating a closed-loop improvement mechanism

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The digital twin system performs self-refinement through machine learning algorithms that automatically adjust parameters based on observed discrepancies between predicted and actual outputs, eliminating the need for manual parameter tuning and enabling the system to self-correct over time

Inventive Principle:

Principle #25Self-service

2Productivity

If digital twin predictions are inaccurate, then control signal generation is simplified, but system efficiency deteriorates

Engineering Contradiction:

Improvesystem efficiencyVSAvoidprediction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system uses feedback from actual performance data to continuously improve prediction accuracy, ensuring that control signals generated from the digital twin become progressively more effective at optimizing system efficiency rather than deteriorating

Inventive Principle:

Principle #23Feedback

3Reliability

If digital twin parameters are not refined, then computational resources are conserved, but component degradation detection is delayed

Engineering Contradiction:

Improvecomponent degradation detectionVSAvoidcomputational resource usage

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system applies partial refinement by focusing computational resources on refining only those digital twin parameters that show significant discrepancies between predicted and actual outputs, rather than uniformly refining all parameters, thus balancing detection capability with resource conservation

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS12360502B1Using a machine-learning model to refine a digital twin of renewable energy system

Publication Date: 2025.07.15 BRIGHTNIGHT LLC

AI summary

A method may include initializing a simulation of a renewable energy plant with an initial set of plant parameters, simulating a first output of the renewable energy plant based on first environmental factors and the initial set of plant parameters, executing a machine-learning model using as input the first simulated output of the renewable energy plant and an actual output of the renewable energy plant to generate an updated set of plant parameters, and simulating a second output of the renewable energy plant based on second environmental factors and the updated set of plant parameters.