Abnormality detector of a manufacturing machine using machine learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing abnormality detectors for manufacturing machines struggle to accurately predict anomalies due to varying usage environments and machine-specific conditions, leading to incorrect learning models and erroneous detection.
Innovation Solution
An adaptive abnormality detector that learns to recognize signs of anomalies post-deployment, using both positive and negative rewards, and classifies waveforms into abnormal, stable, and unrelated categories based on physical quantities from the machine and its environment, enabling accurate anomaly prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If uniform warning conditions are set at manufacturing time, then device complexity is reduced and ease of operation is improved, but measurement precision and reliability deteriorate due to environmental variations
Solution Approach 1:
The system performs preliminary learning during a learning period after installation to establish machine-specific baseline values before actual monitoring begins. This preliminary action captures the unique characteristics of each machine in its specific environment, enabling accurate future detection without requiring complex manual configuration.
Solution Approach 2:
The abnormality detection system automatically adapts to each machine's characteristics through self-learning during the learning period. The controller autonomously determines appropriate warning conditions by analyzing actual operational data, eliminating the need for manual configuration and ensuring accuracy specific to each machine's environment.
2Adaptability or versatility
If machine learning is used to learn coupling weight coefficients using physical quantity data, then adaptability to specific machines is improved, but reliability deteriorates due to incorrect learning of continuous physical quantities as abnormality signs
Solution Approach 1:
The system performs preliminary learning during a designated learning period to establish baseline values for normal operational variations before actual abnormality detection begins. This preliminary action separates the learning of normal variations from the detection of abnormalities, preventing incorrect association of continuous physical quantity changes with abnormality signs.
Solution Approach 2:
The learning process is segmented into distinct phases: a learning period for establishing baseline values and a detection period for identifying abnormalities. This segmentation prevents the system from incorrectly learning continuous physical quantity variations as abnormality indicators, maintaining reliability while achieving adaptability.
Data Source
AI summary
An abnormality detector includes a signal output unit for detecting a sign of an abnormality based on a physical quantity acquired from a manufacturing machine and outputting a signal; and a machine learning device including state observation unit for observing, as a state variable representing a present state of the environment, physical quantity data indicating the physical quantity related to an operation of the manufacturing machine from the manufacturing machine; a label data acquisition unit for acquiring, as label data, operation state data indicating an operation state of the manufacturing machine; a learning unit for learning the operation state of the manufacturing machine with respect to the physical quantity, using the state variable and the label data; and an estimation result output unit for estimating the operation state of the manufacturing machine using a learning result by the learning unit and outputting an estimation result.