Abnormality Detection Display for Time-Series Parameter Tuning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting operational abnormalities in mechanical facilities, such as robots and machine tools, using simple thresholds are inadequate as they fail to detect deviations within predetermined ranges, requiring users to repeatedly adjust parameters to accurately identify abnormal data points.
Innovation Solution
An abnormality detection parameter adjustment device that visually supports users in adjusting parameters by displaying normal and abnormal data samples, allowing intuitive adjustment of parameters like data window width, abnormality threshold, and number of data points for abnormality calculation, facilitating accurate abnormality detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple threshold method is used for abnormality detection, then device complexity is reduced, but measurement precision deteriorates because deviations within predetermined ranges cannot be detected
Solution Approach 1:
The patent changes the detection parameter from fixed threshold values to dynamic reference values based on historical data. By calculating the average value and standard deviation of historical data, the system adapts the detection criteria to actual operational patterns, enabling detection of deviations within predetermined ranges while maintaining reasonable system complexity.
2Measurement precision
If k-nearest neighbor algorithm is introduced to improve abnormality detection accuracy, then measurement precision is improved, but ease of operation deteriorates due to complicated parameter adjustment
Solution Approach 1:
The system performs self-adjustment by automatically calculating optimal parameters (window width, abnormality threshold, number of data points) based on historical data statistics. The control device computes the average and standard deviation of historical values, then uses these to dynamically set detection parameters without requiring manual user adjustment, thereby maintaining high detection precision while simplifying operation.
Solution Approach 2:
The system incorporates feedback mechanisms where detection results and historical data are continuously fed back to refine parameter settings. The control device monitors operational data, compares it against dynamically adjusted thresholds, and uses the outcomes to further optimize detection parameters, creating a self-improving system that reduces manual intervention.
3Measurement precision
If manual parameter adjustment is required to match user's feeling on abnormality detection, then measurement precision can be improved, but loss of time increases due to repeated monitoring and adjustment
Solution Approach 1:
The system performs preliminary actions by pre-calculating statistical parameters (average value, standard deviation) from historical data before actual abnormality detection begins. This preliminary analysis establishes the baseline for detection parameters, eliminating the need for time-consuming manual trial-and-adjustment cycles during actual operation.
Solution Approach 2:
The control device automatically adjusts detection parameters based on historical data statistics without requiring repeated manual intervention. The system self-calibrates by computing optimal window widths, thresholds, and data point counts from accumulated historical information, thereby eliminating the time loss associated with manual parameter tuning while maintaining detection precision.
Data Source
AI summary
An abnormality detection parameter adjustment display device adjusts a parameter for determination for detecting operation abnormality of a mechanical facility based on an input from a user, collects operation information on the mechanical facility, creates a dataset based on the operation information, determines whether or not the dataset indicates a normal operating state, based on the parameter, displays the dataset in a graph, and makes an adjustment for display in such a position that the relationship between the data of the graph and the parameter is visually ascertainable.