Failure Prediction Support Using Actual-Simulation Trace Differences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulty in identifying abnormalities in time series data related to apparatus control, as changes in such data are complex and not easily discernible, making it challenging to predict failures effectively.
Innovation Solution
A failure prediction support device that synchronizes actual machine trace data with simulation trace data, detects differences, and notifies users through a display system when predefined conditions are met, allowing for easier identification of abnormalities and potential failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If users manually analyze time series data to predict failures, then failure prediction capability is maintained, but the complexity of identifying abnormalities increases and detection efficiency deteriorates
Solution Approach 1:
The patent introduces an intermediary system that includes a data acquisition unit to collect actual time series data, a simulation unit to generate expected time series data, and a comparison unit to detect deviations. This intermediary processing system mediates between the complex raw data and the user, automatically identifying abnormalities and presenting them in an understandable format, thereby maintaining failure prediction capability while significantly reducing the difficulty of abnormality identification.
2Quantity of substance
If complex time series data is analyzed without processing, then data completeness is maintained, but user comprehension and abnormality detection efficiency deteriorate
Solution Approach 1:
The patent extracts only the essential information from the complete time series data by comparing actual data with simulation data. The comparison unit identifies and extracts specific deviation points and abnormal patterns from the vast amount of complete data, presenting only the relevant abnormalities to the user. This extraction process maintains data completeness for analysis purposes while dramatically improving detection efficiency by eliminating the need for users to process the entire data set.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided are a failure prediction support device (7), a failure prediction support method, and a failure prediction support program (740), by which a user easily knows an abnormality in time series data relating to an apparatus. The failure prediction support device (7) includes: a difference detection part (780), acquiring actual machine time series data (750) being time series data relating to control of an apparatus and simulation time series data (751) being time series data relating to control of the simulated apparatus, and detecting a difference between the actual machine time series data (750) and the simulation time series data (751); a determination part (781), determining whether the difference satisfies a condition predetermined in relation to an abnormality in the actual machine time series data (750); and a notification processing part (782), notifying that the difference has been detected when it is determined that the difference satisfies the condition.