Pump Condition Detection via Video Sequence Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting the state of pump units, such as acoustic analysis, are unreliable in noisy environments and cannot effectively monitor pump conditions in real-time, leading to potential failures and downtime.
Innovation Solution
An optical method using video sequence analysis to determine the state of pump units by recording and processing surface movements, allowing for the identification of critical states like wear and tear through speed changes between images, which can be compared to a database for accurate status determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If acoustic analysis is used to detect pump unit state, then bearing damage or cavitation can be detected during operation, but the method cannot be carried out reliably in noisy environments
Solution Approach 1:
The patent replaces acoustic analysis with optical analysis. Instead of using microphones to capture sound waves that are contaminated by environmental noise, the invention uses a camera to capture video sequences of the pump unit's surface. The optical method analyzes vibrations through image processing, which is not affected by acoustic noise, thereby resolving the contradiction between detection reliability and noise interference.
2Productivity
If video sequence analysis is used to monitor pump surface movements, then real-time monitoring during operation is enabled, but complex signal processing and database comparison are required
Solution Approach 1:
The patent stores reference vibration patterns and their corresponding diagnostic information in a database before actual monitoring occurs. During real-time operation, the processed video data is quickly compared against these pre-stored patterns to identify anomalies. This preliminary preparation enables rapid diagnosis without complex real-time analysis, resolving the contradiction between monitoring speed and processing complexity.
Data Source
Figure 1
Figure 2A~2D
Figure 3~5
AI summary
A video sequence of a pump unit or pump unit component is generated, which is evaluated using digital image processing with database support, whereby the condition of the pump unit is determined by this evaluation, in particular whether the pump unit is running as intended or whether impermissibly high vibrations are detected in the range that indicate a defect, for example a bearing damage or an unfavorable speed setting.