Cimdata Logo

CIMdata Blog

Thursday, December 14, 2017

Questions from the recent “What does Semantic Technology do for Robust Design?” webinar

Written by 

Q-AYou may be aware that CIMdata has been hosting a series of educational webinars. We have seen record attendance and many of the past webinars are available for replay from the CIMdata on-Demand page of our website. I recently hosted a webinar on Semantic Technology which generated lots of questions that we did not have time to answer live, here are some of the questions and their answers. For those of you who were unable to join the webinar, you can watch it here.

In this webinar, the basics of Semantic Technology-based Ontology were presented along with an example of the open source application Protégé and OWL exercised in the context of Ontologies for FMEA and FTA of a simplified automotive brake-by-wire system. The webinar demonstrated, at a high-level, how to reuse knowledge about developing FMEAs and FTAs to ensure that the new designs can be made robust by leveraging prior knowledge about product failure modes.

Hopefully, the questions and the responses will generate further discussion that enriches the understanding of the topic.

Best wishes,


Your example shows an ontology which is extremely domain-specific. Do you think that a minimal ontology for FMEA is possible to construct?

A minimal ontology for FMEA can certainly be constructed to standardize within individual industries and across industries by organizations, such as NIST, INCOSE, SAE, AIAG, etc. CIMdata is working towards the application of ontologies for capturing, preserving, and reusing knowledge within individual businesses as intellectual property for competitive advantage to ensure that dependable products are developed despite growing engineering complexity and hostile operational environment.

The title of the presentation is "What does Semantic Technology do for Robust Design?" The answer I got so far is "It provides a nice set of tools for formulating the failure model in FMEA." Is this oversimplified? Are there other advantages to reformulate an FMEA with OWL?

Semantic Technology does not just provide a set of tools for formulating the failure model in an FMEA. In the case of an FMEA, as discussed in the webinar, Semantic Technology provides a way of developing an ontology that captures how a SME in the domain (e.g., failure modes prevention in brake-by-wire system) would go about developing the framework of an FMEA. Essentially, it captures the thought process of SMEs in that domain. The detailed information contained in FMEA forms are then encapsulated in the ontology through instantiation.

The main idea is that when an engineer is working in the domain, the knowledge captured in the ontology can be prompted so that relevant thought process is not missed. This would be one way of avoiding repeat reliability issues in any company.

How would you design a robust ontology for a robust system?

For developing a robust system, one should know how the system is likely to fail and based on that knowledge either change the design to avoid the failure mode altogether or provide backup solution(s) that will take over immediately. Such design change is usually worked on if the engineer(s) can a priori guess the failure mode or have experienced it in the past. In a nutshell, this implies an iterative design process.

As mentioned in the webinar, like product design, the development of an ontology is also an iterative process. The failure mode of an ontology is that it does not help reuse and thus, is not used. Based on this metric, the ontology in a specific domain needs to have a governance group in place comprised of SMEs and product development personnel who will need domain knowledge.

A domain-specific ontology needs to be maintained and updated keeping pace with the changes in the systems and subsystems developed by the organization. The use of intra company social media to develop SME communities to develop and keep up domain ontologies would be an approach that could be practical.

Essentially, a robust ontology is a living system that evolves based on subject matter expertise in the domain and has proven to be useful.

How would you validate that semantic connections are rightly placed?

One way to validate semantic connections is through the reasoning feature of OWL ontologies. The reasoner tests whether or not a class is a subclass of another class. Such tests on the classes in an ontology enable the reasoner to compute the inferred ontology class hierarchy. Also, the reasoner helps with consistency checking. Based on the conditions describing a class, the reasoner can check whether a class can have any instances. If a class cannot have any instances, it is deemed inconsistent.

On the other hand, the semantic connection needs to be agreed upon between SMEs for high level ontologies as well as domain ontologies. The validity of these connections is likely to be only as good as the understanding amongst the SMEs.

If we are using some standard and adding our own ontology to serve the requirement how we can validate?

At the level of the industry standard, I presume the validation of the ontology is being taken care of by an industry organization such as NIST, INCOSE, SAE, etc. As to extending that ontology within your company, that would be the responsibility of your SME communities. It needs to be validated by the extent of its usefulness within your company.

As mentioned above, your company would need an SME team to develop and maintain the relevant ontologies and its validation would be based on the expectations from that ontology.

Essentially, the validation of upper level and domain ontologies should help businesses decide whether they want to invest in that effort. CIMdata’s joint incubation projects with industry should help to a large extent in making that decision.


Venki Agaram

Email This email address is being protected from spambots. You need JavaScript enabled to view it.
ipad background image

Featured Cimdata Reports

PLM Benefits Appraisal Guide

CIMdata’s PLM Benefits Appraisal Guide is designed to help potential PLM users evaluate the applicability and payoffs of PLM in their enterprise, and to help existing users of PLM monitor the impact it is having on their product programs.

PLM Market Analysis Reports

The PLM MAR Series provides detailed information and in-depth analysis on the worldwide PLM market. It contains analyses of major trends and issues, leading PLM providers, revenue analyses for geographical regions and industry sectors, and historical and projected data on market growth.

PLM Market Analysis Country Reports

These reports offer country-specific analyses of the PLM market. Their focus is on PLM investment and use in industrial markets. Reports cover Brazil, France, Germany, India, Italy, Japan, Russia, South Korea, the United Kingdom, and the United States.

Simulation & Analysis Market Analysis Report

This report presents CIMdata’s overview of the global simulation and analysis market, one of the fastest growing segments of the overall product lifecycle management market, including profiles of the leading S&A firms.

CAM Market Analysis Report

This report presents CIMdata’s overview of the worldwide CAM software and services market. It also includes a discussion on the trends in the CAM industry and updates on the top CAM solution providers.