Adaptive Actuator Fault Detection via Dynamic Load Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing actuator health monitoring systems in aerospace applications face challenges in detecting degraded actuators operating below normal speed but above the fault detection threshold, leading to high false alarm rates due to measurement errors and mechanical loading, which can result in delayed detection of critical failures.
Innovation Solution
An actuator control system that continuously monitors the position difference between the actual and modeled actuator positions, calculates a dynamic threshold based on estimated load, and declares a fault if the position difference exceeds this threshold for consecutive intervals, allowing for early detection of degraded actuators before they reach a failed state.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If high thresholds are used for fault detection, then false alarm rate is reduced, but degraded actuators operating above FDA limit cannot be reliably detected
Solution Approach 1:
The patent applies dynamics by making the fault detection threshold adaptive rather than fixed. The threshold dynamically adjusts based on real-time measurement of actual versus commanded actuator position, allowing the system to distinguish between normal variations and actual degradation. This resolves the contradiction by enabling reliable degraded actuator detection without generating false alarms from normal operational variations.
Solution Approach 2:
The patent changes the detection parameter from a fixed position threshold to a dynamic threshold based on position error accumulation. By monitoring the integral of position errors over time and comparing against an adaptive threshold, the system can detect gradual actuator degradation while filtering out normal operational variations, thus resolving the contradiction between detection sensitivity and false alarm reduction.
2Device complexity
If fixed thresholds are used for fault detection, then implementation is simple, but degraded actuators operating between normal and failed states cannot be detected
Solution Approach 1:
The patent implements a dynamic threshold that adapts to actual actuator performance by continuously monitoring position errors. This dynamic approach enables detection of degraded actuators operating in the intermediate state between normal and failed, while maintaining reasonable system complexity through efficient computational methods.
Solution Approach 2:
The patent employs feedback by continuously measuring actual actuator position, comparing it to commanded position, and using this information to adapt the detection threshold. This closed-loop approach enables reliable detection of degraded actuators while keeping the system complexity manageable through systematic feedback processing.
3Speed
If high calculation rates are used for FDA logic, then response time is reduced, but measurement errors magnify due to small position changes between calculations
Solution Approach 1:
The patent applies preliminary action by accumulating position errors over multiple calculation cycles before making a fault determination. This integration approach filters out random measurement errors that occur at high calculation rates, while still maintaining fast response to actual actuator degradation through the cumulative error detection mechanism.
Data Source
Figure 1~3
Figure 2
AI summary
An actuator control system (10) includes an actuator (40) continuously movable between multiple positions. A controller (18) is configured to command the actuator to a desired actuator (40) position and to apply the command to an actuator model (24). The controller (18) is configured to compare the modeled and actual actuator positions to determine if the position difference exceeds a fault detection accommodation limit and if the position difference is within a band that is different than the fault detection accommodation limit. The controller (18) is configured to calculate a threshold (30), which is based upon an estimated load on the actuator (40). The controller (18) is configured to calculate a band comprised of the threshold (30) applied to the modeled position and to determine if the actual actuator position is within the calculated band. The threshold (30) is calculated at regular intervals, and a fault is declared if the actual actuator position is outside the band for a number of consecutive intervals.