System-level Prognostics and Health Management for Complex
Industrial Systems: An Application to Pressurized Water Reactors
Mattia Zanotelli
1
, Jamie Coble
2
1,2
University of Tennessee, Knoxville, TN, 37916, USA
mzanotel@vols.utk.edu
jamie@utk.edu
ABSTRACT
Prognostics and health management (PHM) has become es-
sential to guarantee aware and safe system operation and to
inform economic decision-making. However, due to the na-
ture of detection, diagnostics, and prognostic methods, ap-
plications have mainly been limited to the component level.
In practice, most industrial systems consist of multiple inter-
acting components whose partial degradation could lead to
system’s failure (or subsystems). This research addresses the
limitations of traditional component-level PHM techniques
by proposing a novel system-level framework. By imple-
menting a hierarchical structure of components and subsys-
tems, we will select an optimal method for each subsystem
to aggregate its component health assessments. The over-
all system health can then be estimated by further combining
the obtained estimates. The research considers simplified and
holistic modeling techniques, margin-based methods, and hy-
brid graphical models. This approach aims to provide reliable
system health predictions and online components’ sensitiv-
ity measures to enhance maintenance decision-making. We
consider an application in the context of the nuclear indus-
try, characterized by strict safety and economic requirements.
Using a SIMULINK model to approximate a Pressurized Wa-
ter Reactor (PWR) with real industrial inputs, we plan to add
component degradation modules and use simulated sensor
data and reliability information to test the proposed frame-
work. Initial results on artificial case studies show the feasi-
bility of integrating component-level health predictions.
1. PROBLEM STATEMENT
Modern industrial systems have gained a high level of com-
plexity. One of the main interests within the current systems
is the assessment of their health conditions and the mitiga-
tion of possible failure consequences. This can be accom-
plished by PHM techniques (Hu, Miao, Si, Pan, & Zio, 2022)
Mattia Zanotelli et al. This is an open-access article distributed under the
terms of the Creative Commons Attribution 3.0 United States License, which
permits unrestricted use, distribution, and reproduction in any medium, pro-
vided the original author and source are credited.
that have become essential in maintaining the reliability and
safety of such systems. The traditional PHM approach fo-
cuses primarily on individual components, assessing their
health, identifying potential fault modes (diagnostics), and
predicting future degradation behaviors (prognostics). While
this component-level approach is beneficial, it has significant
limitations when applied to complex systems composed of
numerous interacting components whose partial degradation
could lead to the system’s failure. Estimating or predict-
ing the health of a system as a structure of components can
give maintenance operators more insight than a simple collec-
tion of components’ assessments. However, it is particularly
challenging to aggregate the health assessments of individ-
ual components to form an accurate and reliable system-level
health estimate.
The interest in estimating the current and (forecasted) future
health of a whole system has led to the emergence of an in-
novative subfield of PHM, often identified as system-level
prognostics (SLP). This methodology considers the interac-
tions, dependencies, and cumulative effects of all components
within the system, aiming to provide a comprehensive assess-
ment of system performance at the current time and in the
future. By integrating data from various components and ac-
counting for environmental conditions, operational profiles,
and non-linear degradation mechanisms, SLP enables more
accurate and reliable predictions of system failures. Several
literature reviews clearly describe the intentions of SLP, sys-
tematically categorize the proposed approaches, and address
challenges and research gaps (Tamssaouet, Nguyen, Medja-
her, & Orchard, 2023), (Kim, Choi, & Kim, 2021).
Another critical aspect that PHM approaches must consider
and tackle is the prioritization of maintenance actions. In
a component-level framework, maintenance decisions are
made based on the individual health status of components
without considering the system-wide implications. This can
lead to suboptimal maintenance strategies where critical com-
ponents that significantly affect the system’s performance
might be overlooked. Therefore, there is a need for an ap-
1
