Jean-Rémi Massé

A combined sensitivity analysis and kriging surrogate modeling for early validation of health indicators

Abstract To increase the dependability of complex systems, one solution is to assess their state of health continuously through the monitoring of variables sensitive to potential degradation modes. When computed in an operating environment, these variables, known as health indicators, are subject to many uncertainties. Hence, the stochastic nature of health assessment combined with the lack of data in design stages makes it difficult to evaluate the efficiency of a health indicator before the system enters into service. This paper introduces a method for early validation of health indicators during the design stages of a system development process. This method uses physics-based modeling and uncertainties propagation to create simulated stochastic data. However, because of the large number of parameters defining the model and its computation duration, the necessary runtime for uncertainties propagation is prohibitive. Thus, kriging is used to obtain low computation time estimations of the model outputs. Moreover, sensitivity analysis techniques are performed upstream to determine the hierarchization of the model parameters and to reduce the dimension of the input space. The validation is based on three types of numerical key performance indicators corresponding to the detection, identification and prognostic processes. After having introduced and formalized the framework of uncertain systems modeling and the different performance metrics, the issues of sensitivity analysis and surrogate modeling are addressed. The method is subsequently applied to the validation of a set of health indicators for the monitoring of an aircraft engine’s pumping unit.

An approach to the health monitoring of the fuel system of a turbofan

This paper focuses on the monitoring of the fuel system of a turbofan which is the core organ of an aircraft engine control system. The paper provides a method for real time onboard monitoring and on-ground diagnosis of one of its subsystems: the hydromechanical actuation loop. First, a system analysis is performed to highlight the main degradation modes and potential failures. Then, an approach for a real-time extraction of salient features named indicators is addressed. On-ground diagnosis is performed through a learning algorithm and a classification method. Parameterization of the on-ground part needs a reference healthy state of the indicators and the signatures of the degradations. The healthy distribution of the indicators is measured on datas whereas a physical model of the system is utilized to simulate degradations, quantify indicators sensibility and construct the signatures. Eventually, algorithms are deployed and statistical validation is performed by the computation of key performance indicators (KPI).

Prognosis and Health Management in system design

Several airframe systems are considered as case studies. It is assessed that there is a benefit to develop concurrently systems and the attached PHM (Prognosis and Health Management).

Improving Aircraft Engines Prognostics and Health Management via Anticipated Model-Based Validation of Health Indicators

The aircraft engines manufacturing industry is subjected to many dependability constraints from certification authorities and economic background. In particular, the costs induced by unscheduled maintenance and delays and cancellations impose to ensure a minimum level of availability. For this purpose, Prognostics and Health Management (PHM) is used as a means to perform online periodic assessment of the engines’ health status. The whole PHM methodology is based on the processing of some variables reflecting the system’s health status named Health Indicators. The collecting of HI is an on-board embedded task which has to be specified before the entry into service for matters of retrofit costs. However, the current development methodology of PHM systems is considered as a marginal task in the industry and it is observed that most of the time, the set of HI is defined too late and only in a qualitative way. In this paper, the authors propose a novel development methodology for PHM systems centered on an anticipated model-based validation of HI. This validation is based on the use of uncertainties propagation to simulate the distributions of HI including the randomness of parameters. The paper defines also some performance metrics and criteria for the validation of the HI set. Eventually, the methodology is applied to the development of a PHM solution for an aircraft engine actuation loop. It reveals a lack of performance of the original set of HI and allows defining new ones in order to meet the specifications before the entry into service.

Diagnostics of an aircraft engine pumping unit using a hybrid approach based-on surrogate modeling

This document introduces a hybrid approach for fault detection and identification of an aircraft engine pumping unit. It is based on the complementarity between a model-based approach accounting for uncertainties aimed at quantifying the degradation modes signatures and a data-driven approach aimed at recalibrating the healthy syndrome from measures. Because of the computational time costs of uncertainties propagation into the physics based model, a surrogate modeling technic called Kriging associated to Latin hypercube sampling is utilized. The hybrid approach is tested on a pumping unit of an aircraft engine and shows good results for computing the degradation modes signatures and performing their detection and identification.

Towards an integrated development of PHM systems for aircraft engines: In-design selection and validation of health indicators

This paper proposes an approach for the in-design selection and validation of health indicators (HI), based on a virtual prototype, which is part of a larger development scheme named integrated development of PHM. Physics-based modeling is combined with sensitivity analysis to select the most relevant HIs and then with uncertainty propagation to estimate the probability density functions of health indicators for both healthy and degraded states. The validation is then performed through the computation of original performance indicators. The methodology is finally applied to the selection and validation of health indicators for an aircraft fuel system and demonstrates its interest by assessing the performance that the user can expect in terms of detection, identification and localization.

Numerical Key Performance Indicators for the Validation of PHM Health Indicators with Application to a Hydraulic Actuation System

In order to perform Prognostic and Health Management (PHM) of a given system, it is necessary to define some relevant variables sensitive to the different degradation modes of the system. Those variables are named Health Indicators (HI) and they are the keystone of PHM. However, they are subject to a lot of uncertainties when computed in real time and the stochastic nature of PHM makes it hard to evaluate the efficiency of a HI set before the extraction algorithm is implemented. This document introduces Numerical Key Performance Indicators (NKPI) for the validation of HI computed only from data provided by numerical models in the upstream stages of a PHM system development process. In order to match as good as possible the reality, the multiple sources of uncertainties are quantified and propagated into the model. After having introduced the issue of uncertain systems modeling, the different NKPI are defined and eventually an application is performed on a hydraulic actuation system of an aircraft engine.

Methodology for the diagnosis of hydromechanical actuation loops in aircraft engines

This document provides a method for on-board monitoring and on-ground diagnosis of a hydromechanical actuation loop such as those found in aircraft engines. First, a complete system analysis is performed to understand its behaviour and determine the main degradation modes. Then, system health indicators are defined and a method for their real time on-board extraction is addressed. Diagnosis is performed on-ground through classification of degradation signatures. To parameterize on-ground treatment, both a reference healthy state of indicators and degradations signatures are needed. The healthy distribution of indicators is obtained from data and a physics-based model is used to simulate degradations, quantify indicators sensibility and construct the signatures database. At last, algorithms are deployed and a statistical validation of the performences is conducted.

PHM functions maturation

Snecma has been developing Prognostic and Health Monitoring (PHM) functions to monitor different sub-systems of an aircraft engine. To gain maturity and to take more accurate decisions, algorithms need a reality check on the significance of their results. Algorithms have been deployed on in service fleets of engines, getting access to large amounts of data. Nevertheless, probabilities of failure are very low. Thus there are not enough degradation cases collected to compute with accuracy performance metrics, such as Probability of False Alarm (PFA) and Probability of Detection (PoD), for each algorithm. To address this issue, healthy indicators distributions are used to set detection alarm thresholds. Those thresholds are first checked on healthy data and on rare degradations. Algorithms shall indeed raise no alarm on healthy cases and detect all rare degradations. This allows to alarm the operator only when a health indicator is changing. On the base of following observed failure mechanisms, simulations can help to compute with accuracy performance metrics.