Continuous Process Factor Analysis Using Similar Time Period Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In continuous processes, accurately predicting monitoring targets and identifying necessary process factors to adjust when exceptions occur is challenging, leading to potential operational disruptions if the correct factors are not known.
Innovation Solution
A method and system that capture and analyze process factor values and monitoring target values over time series, select similar time periods, and use local modeling to identify contributing process factors, employing algorithms like Orthogonal Least Squares and elastic net regression to determine significant factors affecting the monitoring target.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple process factors are monitored and analyzed to improve prediction accuracy of monitoring targets, then the reliability of continuous process operation is improved, but the device complexity and data processing requirements increase
Solution Approach 1:
The patent segments the continuous process data into multiple time periods and divides process factors into different groups for analysis. By selecting similar time periods and analyzing contributions of individual process factors separately, the system manages complexity through structured segmentation of monitoring data and factors.
Solution Approach 2:
The patent introduces an analysis system that acts as an intermediary between process factors and monitoring targets. This system captures data from multiple sources, processes it through contribution analysis, and provides insights on which process factors to adjust, thereby simplifying the complex relationship between numerous factors and targets.
2Measurement precision
If comprehensive process factor analysis is performed to identify all contributing factors, then the accuracy of exception resolution is improved, but the time required for analysis and response increases
Solution Approach 1:
The patent performs preliminary analysis by pre-identifying similar time periods and pre-calculating contribution relationships between process factors and monitoring targets. When exceptions occur, the system can quickly reference pre-established contribution data rather than performing complete analysis from scratch, reducing response time while maintaining accuracy.
Solution Approach 2:
The patent focuses analysis on local characteristics by selecting similar time periods that match current process conditions. Instead of analyzing all historical data uniformly, the system identifies and analyzes only the relevant local time periods that are most similar to current conditions, improving both speed and accuracy of exception resolution.
3Measurement precision
If all process factors are continuously monitored and adjusted, then the prediction accuracy of monitoring targets is improved, but the ease of operation decreases due to the large number of factors to manage
Solution Approach 1:
The patent extracts and identifies only the key process factors that have significant contributions to monitoring targets from the complete set of process factors. By using contribution analysis to determine which factors to focus on, the system reduces the number of factors operators need to manage while maintaining prediction accuracy.
Solution Approach 2:
The patent changes the operational parameters from monitoring and adjusting all process factors to monitoring only the identified key factors with significant contributions. This parameter change in the operational approach simplifies the operation while preserving prediction accuracy by focusing on the most influential factors.
Data Source
AI summary
A method and a system for analyzing a plurality of process factors affecting the trend of a continuous process are provided. The method includes the following steps: A plurality of similar time periods are selected from a time series. The trend of the continuous process in each of the similar time periods is similar to the trend of the continuous process in a current time period. A contribution of each of the process factors corresponding a monitoring target is analyzed according to the process factor values in the similar time periods and a plurality of monitor target values of the monitoring target. Part of the process factors are picked out according to the contributions.