Actuator Abnormality Detection Using Predicted-Observed Noise Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to detect unstable actuator states indicative of potential failures, as they only monitor for abnormalities at specific instances rather than providing early detection based on continuous monitoring.
Innovation Solution
An information processing device employing data assimilation techniques to calculate noise values, which indicate deviations between predicted and observed system states, allowing for early detection of actuator abnormalities by determining when these noise values exceed predetermined thresholds or show significant temporal changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing abnormality detection devices monitor actuator current or control signals at specific instances, then they can detect obvious abnormalities, but they cannot detect unstable states that indicate potential failures before they occur
Solution Approach 1:
The system performs preliminary actions by continuously calculating noise values that represent deviations between predicted and observed actuator states before actual failures occur. This allows early detection of unstable states by monitoring noise value changes over time, enabling proactive maintenance before the actuator completely fails.
Solution Approach 2:
The system introduces feedback by continuously comparing predicted actuator states (from control signals) with observed actual states, calculating the noise (deviation) between them. This feedback loop enables real-time monitoring of actuator health by tracking noise value changes, allowing the system to detect degradation trends and predict failures before they occur.
2Reliability
If the system continuously monitors actuator noise values to detect unstable states, then early detection of abnormalities is achieved, but the complexity of the detection system increases
Solution Approach 1:
The system uses noise values as an intermediary metric to simplify complex actuator state analysis. Instead of directly analyzing multiple actuator parameters, the system calculates a single noise value representing the deviation between predicted and observed states. This intermediary approach reduces complexity while maintaining early detection capability by focusing on the net effect of all deviations.
Solution Approach 2:
The system replaces complex mechanical monitoring mechanisms with computational analysis. Instead of using additional physical sensors or complex mechanical detection devices, the system uses data assimilation techniques to compute noise values from existing control signals and observation data, substituting physical complexity with computational processing.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is an information processing device, etc., that provides information which is the basis for quick detection of abnormalities that occur in a device. An information processing device 601 has: a degree calculation unit 602 that calculates the degree of conformation between observation information observed in relation to a system undergoing action from a certain device and prediction information predicted using a model in relation to the state of said system; and a difference calculation unit 603 that calculates the difference between an operation amount designated by the device when the degree fulfills a prescribed calculation condition, and a predicted operation amount predicted in relation to the operation amount based on the model.