× Paper submission deadline extended to July 23, 2021.

Panel Sessions

The PHM Society provides an opportunity to hear and interact with recognized industry leaders in relevant areas for our PHM work. These 90 minute panel sessions will consist of presentations and open discussion by 4-6 panelists directly engaging with the conference audience on the different topics listed below.

These sessions add an enriching dimension to the conference experience and a welcome networking alternative to traditional paper presentations, which dominate some conferences. We believe balancing the conference time in this fashion provides participants a much more engaging experience and increased opportunity to gain unique knowledge.

Panel Session Topics:

  1. Applying Artificial Intelligence and Machine Learning for Predictive Maintenance and Analytics
  2. PHM for Manufacturing: Assessing Operations to Advance PHM Capabilities
  3. Qualifying Data and Data Use – Assuring Data Capability for Intelligence Systems and Beyond
  4. Unlocking the Potential of Automotive PHM
  5. Digital Twin
  6. Cybersecurity and PHM: Securing the OT and PHM Data Streams
  7. Autonomous Systems
  8. Wind Energy
  9. DoD/Fielded Systems

Panel Committee Chair:

Brian A. Weiss (National Institute of Standards and Technology)

Panel Session Schedule:

No.Panel NameDate/Time
1Applying Artificial Intelligence and Machine Learning for Predictive Maintenance and AnalyticsTBD
2PHM for Manufacturing: Assessing Operations to Advance PHM CapabilitiesTBD
3Qualifying Data and Data Use – Assuring Data Capability for Intelligence Systems and BeyondTBD
4Unlocking the Potential of Automotive PHMTBD
5Digital TwinTBD
6Cybersecurity and PHM: Securing the OT and PHM Data StreamsTBD
7Autonomous SystemsTBD
8Wind EnergyTBD
9DoD/Fielded SystemsTBD

Panel Session Details:

Panel 1: Applying Artificial Intelligence and Machine Learning for Predictive Maintenance and Analytics
Lead: David Alvord, Georgia Tech Research Institute
With the increasing prevalence of Artificial Intelligence and Machine Learning (AI/ML), and the widening adoption of Model Based Systems Engineering practices (MBSE), applied AI/ML and MBSE are having a significant impact on the PHM community. From predictive maintenance planning through neural net data training and digital twin development to distributed enterprise level systems engineering, these cutting edge capabilities are impacting all operational domains across the public and private sectors. Panelists will discuss lessons learned and best practices leveraging these emerging
technologies. Topics will include integrating and leveraging SME expertise jointly with data science, system level challenges to real world MBSE implementation, and demystifying considerations when applying AI/ML to fielded challenges in the field.
List of Panelists:
Panel 2: PHM for Manufacturing: Assessing Operations to Advance PHM Capabilities
Lead: Brian A. Weiss, National Institute of Standards and Technology
Manufacturing has evolved over the last few decades to leverage emerging and advanced technologies. Many of these technologies enable the growth of PHM capabilities including the advancement of monitoring, diagnostics, and prognostics to enhance decision-making and maintenance strategies. Manufacturers recognize that these emergent PHM capabilities can enhance their maintenance strategy – optimize planned downtime and minimize unplanned downtime – to achieve more reliable, and ultimately, more profitable operations. For manufacturers to realize advanced PHM within their facilities, they face a challenging reality – How do they assess their PHM capabilities and the value it obtains? And, more importantly, what is the value they want to achieve and the corresponding PHM capabilities to be added? This panel will focus on how manufacturers can assess their current PHM capabilities and how they can determine what levels of PHM are most desired by their organization. This will be paired with individual value propositions in terms of the expected return on investment of additional PHM capabilities along with a discussion of current maintenance expenses.
List of Panelists:
Panel 3: PHM for Manufacturing: Assessing Operations to Advance PHM Capabilities
Lead: Michael Sharp, National Institute of Standards and Technology 
Reliable information and quality data are the foundations of the PHM philosophy. Qualifying that data for a range of applications can build trust in end users by providing expectations and limits to how the data should be used. This can also aid developers and solution providers who need understanding of the data to make best use of its capabilities. Understanding information, such as where the data comes from and how it can be used, is integral to the creation of optimal intelligent systems, viable models, and trustworthy information capable of providing actionable decision support.
This panel seeks to discuss the mechanisms for qualifying data collection, documentation, and use as it applies to specific domain applications within the PHM community. Although some qualifications of data are agnostic of application, other questions such as ‘how much data do I need’ or ‘is this an acceptable level of uncertainty’, can only be answered within the context of the end goals and application. Some data collection and storage methods may also dictate the capabilities of that data. Ex. just because a data set is appropriate to build a time series model – it may not work for frequency. Can metadata or data provenance help to communicate this type of information? The goal of this panel is to present and discuss mechanisms for measuring quality of collection, use, and return on investment for data and any associated models primarily with current goals in mind, while leaving room for potential expansion in the future. The panel is aiming to explore a subsect of the topic areas listed below. The scope will become clear as specific panelists are identified.

Topic Areas:
Data Qualification: Establishing trust and expectations within the data
Data Collection
What is considered ‘Big Enough’ data in context of a given PHM application
Data Provenance
Documenting data for use beyond its original intent
Managing information from unreliable sources
Human agents, Natural Language Processing (NLP), Etc.
Quantifying and propagating uncertainty within a data stream
Sensor noise, process noise, model misspecification, etc.
Data Utilization Qualification: Understanding ROI for data applications
Qualifying Data Driven AI and Other ‘Black-Box’ Solutions for PHM
Designing actionable KPIs
Identifying Measurable Outcomes of Data Driven Solutions
Justifying continued support for PHM technologies
Public Benchmark Datasets
Utilizing and contributing open information for competitive advantage
List of Panelists:
Panel 4: Unlocking the Potential of Automotive PHM
Lead: Steve Holland, VHM Innovations, LLC
The automotive industry has proven to be one of the most fertile application domains for PHM technology in terms of financial impact, analytics sophistication and sheer scale. Successful examples have been implemented for both manufacturing systems and the automotive vehicles themselves. The case has been made for even greater opportunity in coming decades as the continuing electrification of vehicles takes place. Similarly, the potential impact for fleets is anticipated to be huge. This applies to conventional automotive and trucking fleets as well as for future autonomous fleets. But, the pace of PHM introduction continues to lag behind what it might be. This panel seeks to understand the key enablers for recent industry successes as well as the barriers that have limited more rapid progress. The discussion will be centered on strategies that effectively exploit those enablers while mitigating the barriers.
List of Panelists:
Panel 5: Digital Twin
Lead: Antonios Kontsos, Drexel
The objective of this panel is to invite the PHM community in a focused discussion related to the concept of digital twins. Experts from industry, government and academia will provide their views while there will be sufficient time for a live discussion and exchange of ideas. The motivation for this panel is two-fold. First, the number of applications where the digital twin concept is used increases continuously. The main reason for this trend can be found on the growing and in some cases mandated process of digitally threading an increasing number of systems in which sensor data is interfaced with storage, processing, computing and decision-making tools. Second, the associated hardware and software components used to design such digital twins provides increasing capabilities. From Internet of Things (IoT) workflows to artificial intelligence methods coupled also with models, the range of options that are available to construct digital twin paradigms is practically vast. However, there are still fundamental questions that need to be answered related to: a) efforts to standardize digital twin processes or at least validate parts of it using traceable, repeatable and effective ways; b) issues related to creating validated dataset repositories which could assist model and processing approach development, independent of a given application and of case-specific data acquisition, and contributing towards demonstrations of successful alignments of the physical and digital spaces; c) benchmark problems tailored to a PHM-related hierarchy of detection, classification and prediction which presents a structure appropriate for related taxonomy and generalization, similar to other domains e.g. in material characterization, nondestructive evaluation, as well as inspection and maintenance; d) modeling including knowledge-based, deep learning, probabilistic, analytical, physics-based etc., which could be leveraged in digital twin workflows; e) efforts to use digital twinning not only in forward flows that involve steps from data to decision but are also creating dynamic adaptations capable of evolving as monitoring occurs, providing feedback to sensors as data is processed and ultimately even creating real autonomy via e.g. data and model driven adaptive control.
List of Panelists:
Panel 6: Cybersecurity and PHM: Securing the OT and PHM Data Streams
Lead: Radu Pavel, TechSolve
The COVID-19 pandemic has led to an accelerated digitalization of the work environment and adoption of remote supervision of manufacturing assets and production. In this new digital manufacturing ecosystem, the Prognostics and Health Management (PHM) approach is becoming the strategy of choice for the advanced manufacturing enterprise. The value of real-time data from various functions, and the benefits of new technologies fuel the desire to connect production and non-production devices on the factory floor. However, the appetite for advanced technology is rapidly exceeding the organizations’ ability to protect it, and this connectivity and data rich environment raise significant concerns and challenges associated with cybersecurity.
This panel will explore the latest trends regarding standards, regulations, strategies and technologies aiming to secure the operational technology (OT) and PHM data and information. The panel also aims to reveal perceived challenges faced by the developers, implementers and providers of PHM technology, and their current strategies for mitigation.
List of Panelists:
Panel 7: Autonomous Systems
Leads: Karl Reichard, The Applied Research Laboratory – Penn State University; George Vachtsevanos, Georgia Tech
List of Panelists:
Panel 8: Wind Energy
Lead: Shawn Sheng, National Renewable Energy Lab (NREL)
List of Panelists:
Panel 9: DoD/Fielded Systems
Lead: Andy Hess, The Hess PHM Group
List of Panelists: