Pump Fault Monitoring Using Reduced-Dimension Operating Space
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pump monitoring systems require the identification of numerous parameters to associate a current operating state with fault scenarios, making it difficult to quickly identify and address issues in pump systems, especially those with multiple pumps.
Innovation Solution
A pump monitoring system that reduces the number of parameters needed by using an interface and processing module to determine a decision vector based on operational values, comparing them to a non-faulty model pump characteristic to identify deviations and associate the current state with fault scenarios, allowing for quicker fault detection and maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional pump monitoring systems monitor multiple performance-dependent variables over time, then fault detection reliability is improved, but the number of parameters to be identified increases significantly
Solution Approach 1:
The patent extracts only the essential operational values (flow rate, power consumption, differential pressure, speed) needed for fault detection, eliminating the need to identify and monitor numerous other performance-dependent variables. This selective extraction maintains reliable fault detection while significantly reducing the number of parameters that must be identified and monitored.
Solution Approach 2:
The patent transforms the monitoring approach by using an m-dimensional operating space where m is small (typically 2-4), comparing the actual operating point to a non-faulty model characteristic in this reduced dimensionality space. This dimensional reduction allows reliable fault detection without requiring comprehensive monitoring of all possible performance variables.
2Measurement precision
If comprehensive monitoring of multiple operational parameters is implemented, then fault detection accuracy is improved, but the complexity of the monitoring system increases
Solution Approach 1:
The patent segments the fault detection task into two parts: (1) monitoring a small set of essential operational values, and (2) comparing these values against a non-faulty model characteristic in m-dimensional space. This segmentation achieves accurate fault detection without requiring a complex system that monitors all possible parameters.
Solution Approach 2:
The patent changes the approach from monitoring many parameters to monitoring few parameters in a transformed m-dimensional operating space. By changing the dimensional representation and using distance-based comparison to model characteristics, the system achieves high detection accuracy with reduced complexity.
3Loss of information
If more operational parameters are monitored, then the ability to distinguish between different fault scenarios is improved, but the time required for analysis and decision-making increases
Solution Approach 1:
The patent uses m-dimensional operating space (where m is small) to represent the essential operational characteristics. By computing distances from the actual operating point to the non-faulty model characteristic and to fault scenario boundaries in this reduced space, the system quickly distinguishes between different fault scenarios without requiring extensive analysis of numerous parameters.
Solution Approach 2:
The patent pre-establishes the non-faulty model pump characteristic and fault scenario boundaries in the m-dimensional operating space before actual monitoring begins. This preliminary preparation allows for rapid real-time fault scenario discrimination by simple distance comparison, minimizing analysis and decision-making time during operation.
Data Source
AI summary
A pump monitoring system associates a current operating state of a pump system including n≥1 pumps with one or more of k≥1 fault scenarios. The pump monitoring system includes an interface module for receiving at least one set of m≥2 operational values from the pump system. The m operational values define a current operational point in an m-dimensional operating space. A processing module processes operational values received by the interface module and consults given or determined model parameters describing a non-faulty model pump characteristic in the m-dimensional operating space and determines a k-dimensional decision vector with k decision vector components indicative of a deviation between an actual differential volume in the m-dimensional operating space based on distances between the m operational values and the non-faulty model pump characteristic, and a modeled differential volume in the m-dimensional operating space for the respective fault scenario.