Water Pump Coolant Leak Detection Using Image Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining coolant leaks in water pumps are subjective and prone to variations due to worker judgment, leading to inconsistent results in assessing coolant leakage from the pump housing.
Innovation Solution
An anomaly determination device utilizing image data and machine learning-based maps to objectively determine coolant leaks by acquiring input variables from image data, executing provisional determination processes, and finalizing the determination based on majority results from multiple maps, eliminating subjective judgment and enhancing accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If workers observe the outer surface of the pump housing to determine coolant leakage, then the determination can be performed manually, but subjective judgment variations occur leading to inconsistent results
Solution Approach 1:
The patent replaces the manual mechanical observation system with an automated image processing system. A camera captures images of the pump housing outer surface, and an image processing device automatically analyzes these images to detect coolant leakage, eliminating human subjective judgment and ensuring consistent, objective determination results.
Solution Approach 2:
The patent creates a digital copy (image) of the pump housing outer surface and analyzes this copy rather than requiring direct physical inspection. The image serves as a reproducible data source that can be processed consistently by the image processing device, ensuring that the same leakage condition is identified uniformly across different assessments.
2Measurement precision
If multiple maps are used for determination, then accuracy is improved through majority voting, but device complexity increases
Solution Approach 1:
The patent divides the determination task into multiple independent maps, each specialized for detecting specific types or patterns of coolant leakage. Each map processes the image data independently and provides a provisional determination, which are then aggregated through majority voting. This segmentation allows for improved accuracy while keeping each individual map relatively simple.
Solution Approach 2:
The patent combines multiple map results into a final determination through a merging process where the outputs of all maps are aggregated and majority voting is applied. This merging step integrates the strengths of multiple individual maps to produce a more accurate and reliable final result, balancing the added complexity with improved performance.
Data Source
AI summary
An anomaly determination device, an anomaly determination method, and a memory medium for a water pump are provided. An acquisition process acquires an input variable of a map from image data obtained by capturing an outer surface of the water pump. Execution circuitry obtains provisional determination results from maps, respectively. The provisional determination results are respectively obtained by executing the provisional determination processes for output variables output from the maps. A determination finalizing process treats, as a majority of the provisional determination results, a final determination result indicating whether coolant has leaked out of the water pump.