Control Valve Performance Detection With Fewer False Alarms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing valve monitoring systems in process control systems are inaccurate and generate too many false alarms due to issues like dead band, resolution, lost motion, internal friction, and variable response time, leading to performance degradation and inefficiencies in process plants.
Innovation Solution
A computing device uses a moving time window to analyze valve performance over multiple time periods, estimating performance metrics such as dead band, change in controller output, and valve travel, and compares these metrics across time periods to accurately distinguish between well-operating and suspect valves, reducing false alarm rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional valve monitoring systems are used to detect valve performance, then valve status can be monitored, but false alarm rates are high and measurement accuracy is poor
Solution Approach 1:
The patent segments the valve performance detection into multiple discrete metrics (dead band, resolution, linearity, hysteresis) and evaluates each separately using specific test sequences. This segmentation allows for more precise measurement of individual valve characteristics without the interference that causes false alarms in traditional monolithic monitoring systems.
Solution Approach 2:
The patent implements dynamic test sequences that adaptively adjust testing parameters based on initial valve responses. The system dynamically modifies excitation signals and measurement parameters to optimize detection accuracy for different valve conditions, thereby improving measurement precision while reducing false alarms through adaptive rather than static monitoring.
2Productivity
If traditional valve monitoring methods are used, then valve operation can be tracked, but the system generates too many false alarms due to dead band, resolution, lost motion, internal friction, and variable response time
Solution Approach 1:
The patent performs preliminary characterization tests to establish baseline valve parameters (dead band, resolution, linearity) before normal operation monitoring. These preliminary measurements create reference profiles that enable the system to distinguish between normal valve behavior variations and actual performance degradation, significantly reducing false alarms while maintaining monitoring efficiency.
Solution Approach 2:
The patent implements feedback mechanisms where valve response data from previous test sequences informs subsequent testing parameters and evaluation criteria. The system uses feedback from measured parameters like dead band and resolution to adjust monitoring thresholds and reduce false alarms, thereby improving overall monitoring reliability without sacrificing productivity.
3Measurement precision
If simple valve monitoring is implemented, then basic valve status can be observed, but measurement accuracy and distinction between well-operating and suspect valves is insufficient
Solution Approach 1:
The patent applies local quality by focusing measurement efforts on specific critical valve parameters (dead band, resolution, linearity, hysteresis) rather than attempting to monitor all possible valve characteristics. This targeted approach to local parameter measurement achieves high measurement precision for the most critical performance indicators while keeping the overall system complexity manageable through selective rather than comprehensive monitoring.
Data Source
AI summary
Techniques for detecting suspicious performance of a throttling control valve (also referred to herein as a “valve”) in a process plant are described herein. For each of N time periods, a computing device determines and analyzes process parameter values for process parameters related to a valve to determine a status of the valve for the time period. The computing device compares the valve statuses over the N time periods to determine whether the valve is operating well for at least a threshold portion of at least a subset of the N time periods. In response to determining that the valve is not operating well for at least the threshold portion of at least the subset of the N time periods, the computing device determines that the valve is suspected of performing poorly, and provides an indication of the suspect valve to a user interface for display to a user.