Batch Event Identification Using Time-Series Distance Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Batch control systems generate a mix of important and less relevant events during batch processes, making it difficult to align different batch executions for meaningful analysis, as arbitrary events can obscure the progression of the batch process.
Innovation Solution
A method for automatically identifying important batch events by determining the distance between time series of process variables across multiple batch executions, using metrics like multivariate or univariate distances, and selecting events with the smallest distance to align batch processes effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If all batch events generated by batch control systems are used for alignment algorithms, then the quantity of events available for analysis increases, but the quality and meaningfulness of alignment results deteriorates due to inclusion of arbitrary or harmful events
Solution Approach 1:
The patent extracts and selects only the relevant subset of batch events that are meaningful for alignment algorithms. The system identifies and extracts events that correspond to actual process phase transitions (filling, reaction, emptying) while filtering out arbitrary events (operator acknowledgments, supply tank status, cleaning notifications). This extraction principle resolves the contradiction by maintaining quantity of useful events while eliminating harmful ones.
Solution Approach 2:
The patent applies local quality by assigning different weights or relevance scores to different batch events based on their significance to process progression. Critical events like phase transitions receive higher weight while arbitrary events receive lower or zero weight. This allows the system to utilize all available events while giving predominant influence to those with local quality of importance.
2Quantity of substance
If arbitrary batch events are included in alignment algorithms, then more data points are available for analysis, but the ability to separate and compare different operational phases deteriorates
Solution Approach 1:
The system extracts events that specifically mark phase transitions (filling start/end, reaction start/end, emptying start/end) while removing events that do not contribute to phase progression information. This extraction ensures that the aligned data maintains clear separation of operational phases while still providing sufficient data points for statistical analysis.
Solution Approach 2:
The patent segments the batch process into distinct operational phases (preparation, reaction, off-loading) and selects events that mark the boundaries or critical points of these segments. This segmentation approach ensures that each phase can be independently analyzed and compared, preventing loss of phase progression information even when using multiple data points.
3Ease of operation
If batch executions are aligned using arbitrary events, then alignment can be performed with available data, but the meaningfulness and interpretability of alignment results deteriorates
Solution Approach 1:
The system automatically extracts and identifies meaningful batch events from the raw event stream, selecting only those events that provide interpretable process information. This automated extraction makes the alignment operation easy to perform while ensuring that the results remain meaningful and interpretable by removing arbitrary events that would otherwise compromise information quality.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Method for automatic identification of important batch events for a batch execution alignment algorithm, comprising: receiving (S10) historical batch data of a batch process, wherein the historical batch data comprises a plurality of batch executions (10, 20), and wherein each of the plurality of batch executions (10, 20) comprises a plurality of batch events (E11, E12, E21, E22), indicating a specific event of the batch process, and at least one time series of a process variable (V1, V2, V3), indicating a development of the process variable during the batch process; determining (S20) a distance between the at least one time series (V1, V2, V3) of the plurality of the batch executions (10, 20) for each of the plurality of batch events (E11, E12, E21, E22); and identifying (S30) at least one important batch event (MI) for a batch execution alignment algorithm with a smallest distance using the determined distances.