Pump Monitoring Using SVM Error Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Centrifugal pump failures in critical infrastructure, such as water supply systems, are challenging to detect preventively, leading to costly equipment damage, technical hazards, and system interruptions.
Innovation Solution
An apparatus and method that utilize a control module to estimate output quantity data based on operational parameters, combined with a support vector machine (SVM) model for enhanced prediction, and an error detection unit to compare estimated and measured values, outputting an error status signal when deviations exceed a threshold, thereby improving pump failure detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional pump monitoring methods are used, then the system is simple and easy to operate, but the failure detection accuracy is insufficient leading to costly equipment damage and system interruptions
Solution Approach 1:
The monitoring system is segmented into distinct functional modules: a control module that receives operational parameters and generates estimated output quantities, and a separate error detection unit that compares these estimates with actual sensor measurements. This segmentation allows each module to specialize in specific tasks, improving overall detection accuracy while maintaining manageable complexity through modular design.
Solution Approach 2:
The control module acts as an intermediary that processes operational parameters and generates predicted output quantities. This intermediary component bridges the gap between raw sensor data and failure detection, enabling the error detection unit to compare theoretical predictions with actual measurements, thereby significantly improving failure detection accuracy without requiring direct complex analysis of all operational parameters.
2Reliability
If advanced monitoring methods with multiple parameters are implemented, then the failure detection accuracy improves, but the device complexity and cost increase
Solution Approach 1:
The error detection unit is designed with multi-functionality, capable of handling various operational parameters (flow rate, pressure, temperature, power consumption) and generating comprehensive failure detection across different failure modes. This universal design improves pump operation reliability by covering multiple failure scenarios while avoiding the need for separate specialized monitoring systems for each parameter or failure type.
Solution Approach 2:
The system implements feedback by continuously comparing the control module's estimated output quantities with actual sensor measurements and using this comparison to detect errors. This feedback mechanism enables real-time monitoring and immediate failure detection, significantly improving reliability by allowing preventive maintenance before actual failures occur, while maintaining manageable complexity through a straightforward compare-and-detect approach.
3Loss of time
If real-time monitoring of operational parameters is performed, then early failure detection is enabled, but the data processing requirements and system complexity increase
Solution Approach 1:
The control module performs preliminary action by pre-calculating expected output quantities based on operational parameters before actual failures occur. This preliminary estimation creates a baseline for comparison, enabling the error detection unit to quickly identify deviations indicating potential failures. This approach reduces the time to detect failures by having predictions ready in advance, while keeping data processing complexity manageable through straightforward predictive calculations.
Solution Approach 2:
The system replaces complex mechanical monitoring and analysis methods with computational approaches. Instead of using sophisticated mechanical sensors and analysis systems to detect subtle failure indicators, the invention uses computational models to predict output quantities and compare them with actual measurements, substituting mechanical complexity with simpler computational processing that achieves the same or better detection speed and accuracy.
Data Source
AI summary
An apparatus for monitoring of a pump includes a control module, and an error detection unit, wherein a support vector machine based module is provided that receives an estimated output quantity data value from the control module, processes the estimated output quantity data value to provide a processed estimated output quantity data value via the support vector machine, and supplies the processed estimated output quantity data value to the error detection unit instead of the estimated output quantity data value of the control module.