Method and system for predicting real plant dynamics performance in green energy generation utilizing physics and artificial neural network models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Dynamics performance in green energy generation systems is often missing or poor due to system faults, leading to uncertainties that affect monitoring and predictive maintenance, and manufacturing scheduling is not optimized for resource service allocation and management.
Innovation Solution
A digital twin system using a physics-constrained machine learning model, incorporating a Levenberg-Marquardt algorithm and artificial neural network, updates well dynamics behavior data to complete missing data and improve prediction accuracy, enabling better monitoring and predictive maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a physics-based model is used to emulate dynamics performance, then prediction accuracy is improved, but the system cannot handle missing or poor quality data from system faults
Solution Approach 1:
The patent introduces an intermediary data processing layer between the physics-based model and the incomplete measurements. This layer uses machine learning algorithms to interpolate missing data points and reconstruct complete dynamics behavior sequences, allowing the physics model to receive quality input data even when sensor measurements are incomplete or corrupted by system faults
Solution Approach 2:
The patent transforms the input data parameters by applying machine learning-based data reconstruction techniques. Missing or poor quality measurement parameters are converted into complete, reliable data sequences through intelligent interpolation and prediction algorithms, enabling the physics-based model to maintain high prediction accuracy despite data quality issues
2Reliability
If dynamics performance data is missing or poor quality, then monitoring and predictive maintenance are adversely affected, but collecting complete data increases system complexity
Solution Approach 1:
The patent replaces complex physical monitoring infrastructure with a computational approach. Instead of adding more physical sensors and data collection hardware to ensure complete data coverage, the system uses machine learning algorithms to computationally reconstruct missing data, thereby maintaining monitoring reliability without proportionally increasing physical system complexity
3Productivity
If manufacturing scheduling is optimized for resource service allocation, then productivity is improved, but incomplete dynamics performance data reduces scheduling effectiveness
Solution Approach 1:
The patent applies preliminary data reconstruction actions before the manufacturing scheduling process. By using machine learning algorithms to complete and quality-assess dynamics performance data in advance, the system ensures that optimized scheduling decisions are made based on complete, reliable information, thereby maximizing productivity without suffering from information loss
Data Source
AI summary
A method of managing a well system includes: obtaining, by a digital twin manager and based on a predetermined monitoring criterion, dynamics behavior data of the well system, where the dynamics behavior data includes a plurality of measurements including an incomplete measurement that is missing data for a time interval and a complete measurement that includes data for the time interval; obtaining modeled dynamics behavior data for the well system using a physics-based model; training a physics constrained machine learning model using one or more machine learning algorithms based on the dynamics behavior data and the modeled dynamics behavior data as inputs; updating dynamics behavior data based on the physics-based model and the physics constrained machine learning model; outputting updated dynamics behavior data for the well system. The updated dynamics behavior data completes the missing data in the incomplete measurement for the time interval.