Adaptable and Generic Methods for Monitoring and Prognostics of
Energy Assets
Mohammad Badfar
1
, Ratna Babu Chinnam
1
, and Murat Yildirim
1
1
Industrial & Systems Engineering Department, Wayne State University, Detroit, MI, 48202, USA
mohammadbadfar@wayne.edu
ratna.chinnam@wayne.edu
murat@wayne.edu
ABSTRACT
Monitoring and prognostics of energy assets are crucial for
maintaining their reliability and efficiency. Effective monitor-
ing ensures that potential issues are identified early, prevent-
ing unexpected failures and optimizing maintenance sched-
ules. However, several challenges complicate this process in
real-world scenarios, including poor data quality, low-fidelity
and sparse data, the influence of external environmental fac-
tors, and diverse operating conditions and asset types. These
challenges highlight the need for adaptable and generic solu-
tions that can handle variability and complexity across differ-
ent energy systems. This Ph.D. project aims to address these
challenges by developing scalable, data-driven approaches
for monitoring and prognostics. By focusing on creating
adaptable and generic frameworks, the research seeks to pro-
vide robust solutions for real-world monitoring and prognos-
tic problems for energy assets.
1. PROBLEM STATEMENT
Monitoring and prognostics of energy assets are crucial for
maintaining their reliability and efficiency. Effective monitor-
ing ensures that potential issues are identified early, prevent-
ing unexpected failures and optimizing maintenance sched-
ules. By continuously tracking the condition and perfor-
mance of assets, it becomes possible to detect anomalies and
signs of wear and tear before they escalate into critical fail-
ures. This proactive approach to asset management not only
extends the lifespan of the equipment but also minimizes
downtime, reduces maintenance costs, and ensures the un-
interrupted operation of energy systems. In the context of
energy systems, where operational continuity and efficiency
directly impact economic and environmental outcomes, the
importance of robust monitoring and prognostics cannot be
overstated. Ensuring that these systems operate optimally
Mohammad Badfar 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.
requires a comprehensive understanding of their health and
performance, which can only be achieved through effective
monitoring and timely prognostics.
However, in real-world scenarios, several challenges compli-
cate this process. Data quality issues, such as low-fidelity
and sparsity, can obscure critical information. Additionally,
the influence of external factors, such as environmental con-
ditions, can confound sensor data, hiding vital signs of asset
degradation or impending failure. In addition, energy systems
operate under diverse conditions and include various types of
assets, each with unique characteristics and operational re-
quirements. This diversity necessitates monitoring and prog-
nostics solutions that are both generic enough to apply across
different scenarios and adaptable enough to cater to specific
conditions. The need for such adaptable and generic solu-
tions is paramount to address the aforementioned challenges
effectively.
This Ph.D. project focuses on developing such adaptable and
generic solutions for the monitoring and prognostics of en-
ergy assets. By integrating advanced machine learning tech-
niques with real-world data, my work aims to create robust
frameworks that can handle the variability and complexity
of actual operating environments. These frameworks are de-
signed to be versatile, capable of processing large-scale data
efficiently, and adaptable, adjusting to new data patterns and
operational conditions as they emerge.
To ensure the practical applicability and effectiveness of these
solutions, my research incorporates real-world case studies.
Each case study serves as a testbed to develop and refine the
frameworks, ensuring they can address the unique challenges
presented by different types of energy assets and operating
conditions. The case studies include preemptive failure de-
tection in photovoltaic (PV) inverters, cross-battery state-of-
charge estimation for lithium-ion batteries, and novelty detec-
tion in connected vehicle systems. These studies highlight the
versatility and adaptability of the proposed solutions, demon-
1
