Abnormal Situation Prevention Algorithm for Plugged Impulse Lines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing abnormal situation prevention algorithms in process plants face challenges in detecting plugged impulse lines due to limitations in regression models developed during specific operating conditions and the need for prolonged learning phases, which may not adapt quickly to changing conditions or detect abnormalities in shorter time frames.
Innovation Solution
A system and method that calculate statistical signatures of process variables over sample windows, generating a function modeling these signatures to detect abnormal conditions by comparing predicted and actual values, with a learning function to update the model when new data points exceed existing values, and a monitoring function to identify deviations beyond a predetermined threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional regression analysis is used to model monitored variable as a function of load variable during a learning phase, then the model can predict values of the monitored variable, but the learning phase becomes too long relative to the time required for abnormal situation to develop
Solution Approach 1:
The patent changes the parameters used in regression analysis from raw process variables to statistical signatures (mean, standard deviation, skewness, kurtosis) calculated over sliding windows. This transformation allows the model to capture essential patterns with fewer parameters, enabling faster convergence during the learning phase while maintaining prediction accuracy for detecting abnormal conditions such as plugged impulse lines.
Solution Approach 2:
The patent extracts key statistical features (signatures) from the raw process data, separating the essential predictive information from the redundant details. By using only these extracted statistical signatures as model inputs, the system reduces the dimensionality of the problem, allowing the regression model to learn the relationship between load variable and monitored variable much faster without sacrificing detection capability.
2Measurement precision
If regression models are developed during specific operating conditions, then the model can accurately predict under those conditions, but the model becomes invalid when operating conditions change
Solution Approach 1:
The patent implements a dynamic model that continuously adapts to changing operating conditions. The regression model is updated in real-time as new data becomes available, allowing it to track changes in the relationship between load variable and monitored variable. This dynamic approach ensures the model remains valid across varying operating conditions while maintaining prediction accuracy through continuous learning from incoming statistical signatures.
Solution Approach 2:
The patent creates a universal model that can handle multiple operating conditions through the use of statistical signatures that are invariant to normal operating variations. The model learns the fundamental relationship between load changes and monitored variable responses, making it applicable across different operating scenarios rather than being tuned to a specific condition, thus achieving both accuracy and adaptability.
Data Source
AI summary
Systems and methods are provided for detecting abnormal conditions and preventing abnormal situations from occurring in controlled processes. Statistical signatures of a monitored variable are modeled as a function of the statistical signatures of a load variable. The statistical signatures of the monitored variable may be modeled according to an extensible regression model or a simplified load following algorithm. The systems and methods may be advantageously applied to detect plugged impulse lines in a differential pressure flow measuring device.