Gear Feed Cycle Analysis for Abnormality Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing feeding devices lack effective abnormality detection and prediction capabilities, particularly for gears, especially when the tape is in use, and existing solutions only detect individual differences without predicting abnormalities.
Innovation Solution
An abnormality prediction device that decomposes detection information into trend, cycle, and random components to obtain an abnormality level based on the cycle component, which is strongly affected by gear deterioration or wear, allowing for more accurate prediction of gear abnormalities, including breakage and misalignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detection information is decomposed into trend, cycle, and random components to obtain abnormality level based on cycle component, then prediction accuracy of gear abnormality is improved, but device complexity increases due to additional decomposition processing
Solution Approach 1:
The detection information is segmented into three distinct components: trend component (long-term changes), cycle component (periodic variations corresponding to gear rotation), and random component (noise). This segmentation allows the system to isolate and analyze the cycle component specifically for gear abnormality detection, improving prediction accuracy by focusing on the most relevant signal while filtering out trends and random variations.
Solution Approach 2:
The cycle component acts as an intermediary that mediates between the raw detection information and the abnormality level determination. By using the cycle component as an intermediate representation, the system can effectively capture gear-specific periodic variations without being influenced by overall feed amount trends or random noise, thus achieving accurate gear abnormality prediction.
2Reliability
If the cycle component is used to determine abnormality level, then reliability of gear abnormality detection is improved, but loss of information increases by excluding trend and random components
Solution Approach 1:
The system extracts only the cycle component from the detection information for gear abnormality detection purposes. This extraction is justified because the cycle component specifically contains the periodic variations caused by gear rotation and wear, which are the most relevant indicators for gear abnormality. The trend and random components are excluded as they do not contribute significantly to gear-specific abnormality detection.
Solution Approach 2:
Different components of the detection information are treated with different levels of importance. The cycle component is given high priority as it directly reflects gear condition, while the trend component (overall feed amount changes) and random component (noise) are treated as less relevant for gear-specific diagnostics. This local quality approach allows reliable gear detection without being distracted by less relevant information.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An abnormality of a gear is more appropriately predicted. An abnormality prediction device of the present disclosure is used in a feeding device including a driving section, a gear connected to the driving section, and a detection section configured to detect a position of a medium fed in accordance with driving of the gear, and intermittently feeding the medium. The abnormality prediction device acquires detection information regarding a feed amount of the medium based on the position of the medium detected over time by the detection section, decomposes the acquired detection information into a trend component regarding a moving average of the gear, a cycle component based on a cycle of the gear, and a random component obtained by excluding the trend component and the cycle component, obtains an abnormality level of the gear based on the cycle component obtained by the decomposition, and outputs the obtained abnormality level.