Factory Noise Cause Estimation Using Machine-State Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise generation detection systems in factories face difficulties in identifying the cause of noise, especially when it is intermittent and influenced by the surrounding environment, as they struggle to observe and organize the complex relationships between noise generation and various environmental factors.
Innovation Solution
A noise generation cause estimation device that connects with multiple machines in a factory, acquiring noise and operation information, using machine learning to form clusters and identify principal components, thereby extracting and displaying the operation information causing noise, and providing maintenance information for quick countermeasures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If noise detection is performed in a factory environment with multiple machines, then noise generation can be detected, but it becomes difficult to identify the cause of noise generation especially when noise is intermittent and environmental
Solution Approach 1:
The patent segments the complex factory environment into individual machine units, each equipped with its own noise detection and operation information acquisition capabilities. By dividing the monitoring system into distributed node units that independently collect and analyze data from their respective machines, the system can identify noise sources without requiring a monolithic complex architecture.
Solution Approach 2:
The patent introduces a determination unit as an intermediary that receives noise information from multiple machines and correlates it with operation information. This intermediary component performs the complex analysis of matching noise patterns with operational states, thereby identifying noise causes without requiring direct complex interaction between all system components.
2Reliability
If long-term observation and manual organization of noise-related factors are performed, then causal relationships can be ascertained, but the process requires accumulation of extensive know-how and is difficult to implement
Solution Approach 1:
The patent performs preliminary action by pre-collecting and storing operation information from all machines before noise events occur. The system maintains a database of normal operational patterns, which allows it to quickly compare against noise event data and identify deviations, eliminating the need for time-consuming manual observation and know-how accumulation.
Solution Approach 2:
The patent implements feedback mechanisms where the determination unit continuously compares current noise and operation data against historical patterns, and automatically updates its understanding of causal relationships. This closed-loop feedback system enables the system to improve its noise cause identification accuracy over time without requiring manual intervention or know-how accumulation.
3Measurement precision
If comprehensive operation information from all machines is collected and analyzed, then noise causes can be identified even in complex factory environments, but the data processing and analysis complexity increases
Solution Approach 1:
The patent extracts only the relevant operation information that is actually associated with noise generation events. Rather than analyzing all possible operational parameters from all machines continuously, the system selectively extracts and analyzes only those data points that correlate with detected noise, significantly reducing analysis complexity while maintaining detection accuracy.
Data Source
AI summary
A noise generation cause estimation device capable of easily estimating the cause of noise generation in a factory is connected for communication with a plurality of machines in the factory. The noise generation cause estimation device is provided with a noise information acquisition unit configured to acquire noise information generated in the machines, an operation information acquisition unit configured to continually acquire operation information of all the machines, and a determination unit configured to learn the relevance between the noise information and the operation information. The determination unit is provided with a state observation unit configured to observe the noise information and the operation information as state variables indicative of a current state of the environment and a learning unit configured to form a plurality of clusters including the state variables.