Dynamic Reference Period for Anomaly Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing anomaly diagnosis systems for machine facilities face challenges in accurately predicting remaining operable hours and long-term performance trends due to the complexity of analyzing large quantities of sensor data and maintenance history information, often resulting in inadequate maintenance and reduced operational efficiency.
Innovation Solution
An anomaly diagnosis system that utilizes a time series data receiver, state measure calculator, approximation formula calculator, and state measure estimating unit to calculate anomaly and performance measures using statistical methods and polynomial expressions, allowing for accurate long-term predictions by extending the reference period as new data is acquired.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a long reference period is used to calculate the approximation formula for state measure estimation, then the long-term prediction accuracy is improved, but the responsiveness to recent changes in machine facility state deteriorates
Solution Approach 1:
The reference period is made dynamic by adjusting its length based on the variance of state measures. When variance is high (indicating recent changes or anomalies), the reference period is shortened to improve responsiveness. When variance is low (indicating stable operation), the reference period is extended to improve prediction accuracy. This dynamic adjustment resolves the contradiction between long-term accuracy and responsiveness to recent changes.
Solution Approach 2:
The system changes the parameter of reference period length based on the calculated variance of state measures. By monitoring variance and adjusting the reference period accordingly, the system adapts to different operational conditions, achieving both high prediction accuracy during stable periods and rapid responsiveness during changing conditions.
2Loss of time
If a short reference period is used to quickly respond to recent changes, then the responsiveness is improved, but the long-term prediction accuracy deteriorates
Solution Approach 1:
The reference period dynamically adjusts its length based on operational conditions. During stable operation, it extends to capture long-term trends for accurate prediction. During periods of change or anomaly, it contracts to quickly adapt to new conditions, thus resolving the contradiction between responsiveness and prediction accuracy.
Solution Approach 2:
The system changes the reference period parameter in response to variance calculations. When state measure variance exceeds thresholds, the reference period is reduced to improve responsiveness. When variance is within acceptable ranges, the reference period is increased to enhance prediction accuracy, effectively managing the trade-off between these two competing requirements.
3Measurement precision
If complex statistical methods are applied to analyze large quantities of sensor data, then the analysis precision is improved, but the system complexity increases
Solution Approach 1:
The system extracts only the essential feature - the variance of state measures - from large quantities of sensor data. By focusing on this single key metric rather than analyzing all raw sensor data directly, the system achieves high analysis precision while keeping the processing complexity manageable. This extraction approach filters out unnecessary information and retains only what is critical for prediction.
Solution Approach 2:
The system transforms complex sensor data into simplified state measure parameters through statistical processing. By changing the data representation from raw sensor readings to derived state measures with calculated variances, the system maintains analytical precision while reducing computational complexity and making the data more tractable for prediction algorithms.
Data Source
AI summary
The anomaly diagnosis system includes the state measure calculator acquiring sensor data from sensors in a machine facility as time series data; an approximation formula calculator calculating a state measure being an index indicating a state of the machine facility, such as anomaly and a performance by a statistical method in which the time series data is used as learned data; and a state measure estimating unit estimating the state measures until future time using the approximation formula. Whenever the latest time series data is acquired, the reference period in which the time series data corresponding to the state measure referred to calculate the approximation formula by the reference period setting unit, is successively extended by addition of time when the latest time series data is acquired. The approximation formula calculator calculates the approximation formula using the state measure of the time series data acquired in the reference period.