Adaptive Fault Diagnosis Using Dynamic Learning Data Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fault diagnosis methods in machine systems face challenges in accurately diagnosing system faults due to changes in system status, as they often rely on pre-created learning data that may not be relevant to the current system state, leading to potential erroneous or non-detections.
Innovation Solution
A fault diagnosis device that acquires diagnosis target data, determines an extraction condition using condition setting information, and extracts a limited number of data sets from stored data sets based on this condition, creating learning information suitable for pattern recognition methods to dynamically adapt to the system's status, thereby improving diagnostic accuracy and reducing processing time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pre-created learning data is used for fault diagnosis, then the diagnostic process is simple and fast, but the diagnostic accuracy deteriorates when system status changes
Solution Approach 1:
The patent implements dynamic learning data extraction by determining extraction conditions based on current diagnosis target data and condition setting information. The system dynamically selects and extracts relevant data sets from stored extraction source data according to the current system state, making the learning data adaptive to changing system conditions rather than static and pre-fixed.
Solution Approach 2:
The patent changes the parameters of learning data by extracting data sets that satisfy dynamically determined extraction conditions. The extraction conditions are determined based on parameter values in the diagnosis target data and condition setting information, allowing the system to select learning data with appropriate parameter ranges that match the current system state, thereby improving diagnostic accuracy.
2Reliability
If all stored data sets are used as learning data, then comprehensive coverage is achieved, but processing time increases
Solution Approach 1:
The patent extracts only the necessary learning data sets from the stored extraction source data by determining extraction conditions based on current diagnosis target data. Instead of using all available data, the system selectively extracts data sets that satisfy the extraction conditions, reducing the volume of learning data while maintaining diagnostic accuracy and decreasing processing time.
Solution Approach 2:
The patent applies partial action by extracting a subset of data sets that are sufficient for accurate fault diagnosis rather than using all available data. The extraction conditions are determined to obtain just enough relevant learning data to achieve high diagnostic accuracy, avoiding the excessive processing time that would result from using the complete data set.
3Reliability
If learning data is dynamically selected based on current system state, then diagnostic accuracy improves, but data processing complexity increases
Solution Approach 1:
The patent applies local quality by determining extraction conditions specifically tailored to the current diagnosis target data and condition setting information. Rather than using a uniform approach for all diagnostic situations, the system locally adapts the extraction conditions to match the specific system state being diagnosed, improving accuracy while managing complexity through targeted rather than universal processing.
Data Source
AI summary
A fault diagnosis device includes circuitry configured to acquire diagnosis target data including parameter values from the diagnosis target system, to store extraction source data including a plurality of data sets that includes the parameter values of the diagnosis target data representing a normal state of the diagnosis target system, to determine an extraction condition based on the diagnosis target data and on condition setting information, to extract a group of data sets satisfying the extraction condition among the plurality of data sets, to select, as learning data, a number of first data sets of the group, in which the data sets of the group are sorted according to a sorting criterion, to generate learning information from the learning data, and to determine whether the diagnosis target data associated with the diagnosis target system is faulty based on the learning information.