Helicopter Engine Residual Analysis for False Alarm Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current helicopter engine health monitoring systems often generate false alarms due to unrecorded pilot actions, which are misinterpreted as engine state changes, leading to unnecessary overhauls.

Innovation Solution

A system that differentiates between unrecorded pilot actions and engine state changes by analyzing the residuals between real data and a reference model, using a method that calculates instantaneous and overall differences, intrinsic residuals, and deviating portions to identify pilot actions, and then corrects these to determine if the engine state has changed.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If pilot actions are not recorded, then the monitoring system is simpler, but false alarms increase due to misinterpretation of pilot actions as engine state changes

Engineering Contradiction:
Improvemonitoring system complexityVSAvoidengine state detection accuracy
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent segments the residual signal into different components by analyzing its temporal structure. It divides the monitoring approach into detecting overall engine state changes versus detecting transient pilot actions, allowing the system to distinguish between permanent deviations (engine issues) and temporary deviations (pilot actions) without requiring additional sensors

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies preliminary signal processing to the residual data by calculating both instantaneous and global deviations before making diagnostic decisions. This preliminary analysis of the residual signal structure allows the system to preemptively identify and filter out pilot action patterns, preventing false alarms before they occur

Inventive Principle:
Principle #10Preliminary action

2Ease of operation

If the monitoring system uses simple comparison with reference model, then the system is easier to operate, but measurement precision decreases due to inability to distinguish pilot actions from engine changes

Engineering Contradiction:
Improvemonitoring system operationVSAvoidengine state change detection precision
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent introduces dynamic analysis of the residual signal by examining both instantaneous deviations and global deviations over time. This dynamic approach allows the system to adapt its interpretation based on the temporal characteristics of the deviation, distinguishing between transient pilot actions and persistent engine state changes while maintaining ease of operation

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent uses feedback from the residual signal analysis to improve detection precision. By continuously monitoring the deviation pattern and comparing instantaneous residuals with global residuals, the system receives feedback that helps distinguish pilot actions from engine changes, enhancing measurement precision without complicating operation

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3956646B1System for monitoring the health of a helicopter
Publication Date: 2023.03.22 SAFRAN SA
  • EP3956646B1 patent drawingFigure 1~2
  • EP3956646B1 patent drawingFigure 3

AI summary

The invention relates to a system for monitoring the health of a helicopter, comprising a helicopter and a device for determining a change in state of the engine that is configured to collect data measured by engine sensors and external conditions during a stable flight phase and to process said measured data in the following way: comparing said measured data with a reference model of the engine, determining, at each time interval, an instantaneous discrepancy between each item of data measured and each item of data estimated by the reference model of the engine, determining, throughout the stable flight phase, an overall discrepancy between the measured data and the data estimated by the reference model of the engine, determining, at each time interval of the stable flight phase, an intrinsic residual corresponding to a difference between the instantaneous discrepancy and the overall discrepancy, determining a deviating portion corresponding to the part of the intrinsic residual that does not satisfy a predetermined criterion, each deviating portion containing an item of information relating to a pilot action that is not recorded, and determining a corrected residual corresponding to the instantaneous discrepancy from which the deviating portion has been subtracted, the corrected residual being analysed in order to determine whether the state of the engine has changed.