Operational Opportunity Discovery Using LSTM Mode Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Complex production processes like oil sand extraction and steelmaking require optimal balance to achieve multiple objectives, but existing methods lack real-time opportunities for productivity enhancement, cost reduction, and energy saving within short time windows.
Innovation Solution
A computer-implemented method using a neural network, specifically an LSTM auto-encoder, to extract features from time series data, identify operational modes, and compare current states to historical data for discovering operational opportunities, recommending control and non-control variables to enhance productivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional production monitoring methods are used, then operational stability is maintained, but real-time productivity enhancement opportunities are not identified
Solution Approach 1:
The system performs preliminary encoding of historical operational data into compressed representations before real-time analysis. This pre-processing allows the system to quickly compare current operational states against historical patterns, enabling real-time identification of productivity opportunities without time loss.
Solution Approach 2:
The patent replaces traditional mechanical monitoring methods with neural network-based pattern recognition. The encoder-decoder architecture automatically identifies operational patterns and anomalies, substituting manual analysis with intelligent algorithms that operate in real-time to capture fleeting productivity opportunities.
2Loss of information
If detailed time series data is analyzed, then operational insights are improved, but data processing complexity increases
Solution Approach 1:
The encoder portion of the neural network extracts essential features from detailed time series data, separating critical operational information from redundant details. This extraction process retains meaningful operational insights while eliminating unnecessary data complexity for efficient processing.
Solution Approach 2:
The system transforms detailed time series data into compressed latent representations through the encoder, changing the parameter space from high-dimensional raw data to low-dimensional encoded features. This parameter transformation preserves operational information while dramatically reducing processing complexity.
3Productivity
If historical operational data is used for comparison, then productivity opportunities are identified, but real-time responsiveness is reduced
Solution Approach 1:
Historical operational data is pre-encoded into compressed representations during offline processing. When real-time analysis is needed, the system only needs to compare current states against these pre-prepared encoded historical patterns, enabling both accurate opportunity identification and rapid responsiveness.
Solution Approach 2:
The system creates compressed encoded copies of historical operational data that capture essential patterns. These encoded copies serve as efficient reference models for real-time comparison, allowing the system to leverage historical insights without the computational burden of processing full historical datasets in real-time.
Data Source
AI summary
In an approach for real-time opportunity discovery for productivity enhancement of a production process, a processor extracts a set of features from time series data, through autoencoding using a neural network, based on non-control variables for the time series data. A processor identifies one or more operational modes based on the extracted features including a dimensional reduction with a representation learning from the time series data. A processor identifies a neighborhood of a current operational state based on the extracted features. A processor compares the current operational state to historical operational states based on the time series data at the same operational mode. A processor discovers an operational opportunity based on the comparison of the current operational state to the historical operational states using the neighborhood. A processor identifies control variables in the same mode which variables are relevant to the current operational state.