Towards a Capabilities Taxonomy for Prognostics and Health Management

This communication proposes the development by the PHM Society of a classification or taxonomy for the skills needed for the prognostics and health management (PHM) field. This taxonomy is a set of descriptors for each skill for a chosen range of competency levels: entry, working and mastery. This communication defines a structure and process to ensure application to real needs by employers, practitioners and training developers. Preliminary results of the development of Analytics, Test and Experiment Design and Cost Benefit Studies sub-domains within the PHM field are reported based on workshops at the PHM 2012 and 2013 Annual Conferences. Steps for a way forward are proposed and mechanisms for interested parties to participate are suggested. The PHM Society’s Education and Professional Development Committee would lead these efforts.


BACKGROUND
Prognostics and Health Management (PHM) is a diverse, multi-disciplinary domain with rapidly evolving capability needs. Initial PHM education and training from many discipline entry points must be complemented by specialized and professional development over a career.
The community has identified that benchmarks are needed for career planners, employers and training developers to facilitate transitions between mastery levels.
The community's stakeholders in academia, industry and government require a stream of qualified practitioners with lifelong professional development.
Aligned with this scenario, the PHM Society has objectives to:  promote the development, growth, and recognition of prognostics and health management (PHM) as an engineering discipline;  support PHM education by developing standard teaching curricula in the field; and  establishing, developing, or adopting standards, methods, and metrics in PHM; To accomplish these objectives the board of the Society has approved activities in this direction through a committee on Education and Professional Development.

PROPOSAL
Through the PHM Society the goal is to compile a PHM Taxonomy or classification that maps: Skill/capability areas e.g. signal processing, statistics, control systems, [as rows] Competency levels: Entry, Working or Mastery for each [as columns] This taxonomy should be usable by: Employers: A job description could be readily prepared to include a capabilities list from various areas with desired mastery levels. Evaluation schemes could reference mastery levels to be demonstrated or developed. Professional development transitions could be defined and matched to courses or in-house assignments.
The progression of complexity is typical within the hierarchies used. For example [3], a creative thinking skill aims to respond to challenges with innovative solutions, products or services by questioning conventional means, using intuition, experimentation and fresh perspectives. Five levels of competency with descriptors are provided [3]: Level 1: Understands how to solve problems Uses common sense and knowledge to identify underlying issues and to solve problems.
Uses appropriate problem-solving techniques.

Level 2: Identifies existing solutions
Reviews existing options, concepts and approaches, and identifies one(s) that will work in the situation at hand.
Stands back from a problem and observes patterns and interrelationships in data to see where problems or discrepancies arise.
Recognizes and assesses several likely causes or ways of interpreting available information.
Recognizes when a new approach is needed to solve a problem.
Solves complex problems through application of existing theories or explanations.

Level 3: Modifies existing solutions
Questions traditional solutions and uses unconventional methods to find solutions.
Thinks about problems from a new perspective.
Applies ideas from other disciplines or fields of research to solve problems.

Level 4: Creates new solutions/concepts
Thinks several steps ahead in deciding on best course of action (anticipates likely outcomes).
Creates, assesses and applies new concepts, theories, approaches and/or solutions.
Solves complex or inter-related problems through developing new theories, explanations or applications.
Level 5: Impacts a field of science/technology Develops new theories and applications with revolutionizing or wide-ranging impacts.
Alternatively, a technical skill can also be categorized by breadth of application [4]: 1. Engineering and technical contribution at the subsystem level is required.
2. The conversion of concepts to subsystems and systems, major system design and/or technical leader in multiple projects 3. A developer of engineering concepts, major system design, and/or technical leader in multiple programs. A principal contributor in systems design, systems engineering, concept development, and project reports. For this PHM taxonomy we have decided to use three levels that incorporate important aspects of these two above approaches : 1. Entry Level: Understands, selects, interprets and applies basic concepts and known methods. May work at the sub-system level. 2. Working Level: Analyses and solves complex problems by combining or extending existing methods. Integrates and validates at the system level. 3. Mastery Level: Evaluates, innovates, synthesizes and validates new methods through unique, often multidisciplinary insights. Application of skills may include analysis, design, build/implement and test for equipment, hardware and software.

PHM SKILLS AND CAPABILITIES
Compiling a list of possible skills for PHM practitioners is challenging because the field is multi-disciplinary. A generic PHM taxonomy has been proposed by Jennions [5]. He identifies some key skills related to business, systems design, architecture, analytics, technologies, applications and support. In addition, work in PHM taxonomies by SAE HM-1 is expected to provide the following insights because of the broad membership base.
This work aims to extend these taxonomies to meet the stakeholder needs defined above with a comprehensive list of skills. Firstly, we propose to subdivide skills and capabilities as follows:

SAMPLE CAPABILITIES-COMPETENCY MAPPING
Here we take a practitioner skills/capability viewpoint as opposed to the IVHM system components view. A sample taxonomy entry is provided in the following table to illustrate some descriptors for a selected specialty within the Analytics domain and diagnostics sub-domain.

DEVELOPMENT OF TAXONOMY
The PHM Society Education and Professional development Committee sponsored a workshop at the PHM 2012 annual conference to workshop the concept described above. Three significant domains were selected, each to have a broad based group examine the sub-domain list and also to draft competency descriptors.
While these drafts are neither complete nor final, it was felt that the developing a 'strawman' version would be helpful to the Society.

Domain 1-Analytics
The working group at the PHM 2012 conference agreed to proceed with the sub-domains in section 4 above. However, they considered the prognostics sub-domain separate and were not able to address it in the time available. They proposed working versions of the competency descriptions for most of the sub-domains in that list and added some specific specialties within some sub-domains. For a number of the specialties/analysis tools they chose the same generic descriptors. Annex 1 provides the results. In general, they proposed application of existing tools at the entry level, adaptation at the working level and customization or new methods at the mastery level.

Domain 2-Test and Experiment Design
The working group at the conference added additional subdomains: integration, fault injection/insertion, data validation, data reduction and data management. They chose not to add specialties, in the limited time available. The descriptors addressed a number of diverse skills and capabilities across the hardware/software/sensor scope of this domain. Details are provided in Annex 2.

Domain 3-Cost Benefit Analysis
The group who discussed the important Cost Benefit Analysis domain proposed adding sub-domains for asset management, maintenance and the effect of prognostics management, and business versus customer perspective. It was also suggested that the metrics, uncertainty/confidence and evaluation sub-domain might be considered as specialties within the risk analysis sub-domain. Details are provided in Annex 3.

CONCLUSIONS AND WAY FORWARD
The need for a broadly applicable PHM taxonomy has received a preliminary validation through a limited workshop process. The three classes of users were confirmed. The general form and a working set of definitions have been developed and reviewed. Draft taxonomies for three PHM domains have been proposed for further development through either workshops or specialists. The appropriate number and level of detail for the specialties is yet to be determined.