Numerical Controller ML for Time-Series Abnormality Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting abnormalities in numerical controllers using machine learning lack the ability to perceive temporal changes in data and do not effectively handle data from sensors, manual operations, and mode switching, making it difficult to automatically detect complex abnormalities.
Innovation Solution
A machine learning device is integrated into the numerical controller that finds approximate polynomials through regression analysis of historical data, generates feature vectors, clusters them, and predicts future abnormalities by determining which cluster current data belongs to, producing a ranking of potential issues based on past occurrences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual inspection methods are used for abnormality detection, then simple cases can be identified by skilled persons, but complex abnormalities involving large amounts of data cannot be effectively detected
Solution Approach 1:
The patent replaces manual inspection methods with an automated machine learning system that processes time-series data from multiple sensors. The system uses clustering algorithms to automatically identify abnormal patterns without requiring skilled personnel to manually analyze complex data sets, thereby maintaining high detection accuracy while eliminating the limitations of manual inspection.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between raw sensor data and abnormality detection. This intermediary processes and interprets complex time-series data from multiple sources, transforming it into actionable insights that automatically identify both simple and complex abnormalities without requiring manual intervention.
2Reliability
If traditional regression analysis is used for abnormality prediction, then statistical trends can be identified, but temporal changes in data patterns cannot be perceived
Solution Approach 1:
The patent transitions from static regression analysis to dynamic clustering analysis that processes time-series data. By continuously analyzing temporal patterns and state transitions in the data, the system perceives how system behavior evolves over time, capturing dynamic changes that traditional static methods miss while improving prediction reliability.
Solution Approach 2:
The patent adds a temporal dimension to the analysis by processing time-series data through clustering algorithms. This transforms the data from simple statistical values into multi-dimensional state sequences that reveal temporal patterns and evolution of system conditions, preserving temporal information that traditional regression analysis loses.
3Adaptability or versatility
If machine learning is applied to handle multiple data types from sensors and manual operations, then comprehensive abnormality detection is achieved, but calculation complexity increases
Solution Approach 1:
The patent segments the complex machine learning task into distinct processing stages: data collection from multiple sources, time-series analysis, clustering operations, and abnormality identification. This segmentation allows each component to handle specific data types efficiently, managing calculation complexity while maintaining comprehensive adaptability across different data sources and abnormality types.
Data Source
AI summary
To provide a numerical controller and a machine learning device that predict an abnormality, based on machine learning with perception of temporal change in data. The numerical controller includes the machine learning device provided with a learning unit that conducts machine learning of trends in operation of a machine on occasions of occurrence of abnormalities in the machine, based on time-series data acquired by a data logger device and relating to the operation of the machine and abnormality information relating to the abnormalities which have occurred in the machine and a prediction unit that predicts an abnormality which will occur in the machine, based on results of the machine learning in the learning unit and time-series data acquired by the data logger device and relating to current operation of the machine.