This is my concluding post related to the various aspects of the model-driven enterprise. We went through Model-Based Systems Engineering (MBSE) where the focus was on using models (functional / logical / physical / simulations) to define complex product (systems). Next we discussed Model Based Definition / Model-Based Enterprise (MBD/MBE), where the focus was on data continuity between engineering and manufacturing by using the 3D Model as a master for design, manufacturing and eventually service information.

And last time we looked at the Digital Twin from its operational side, where the Digital Twin was applied for collecting and tuning physical assets in operation, which is not a typical PLM domain to my opinion.

Now we will focus on two areas where the Digital Twin touches aspects of PLM – the most challenging one and the most over-hyped areas I believe. These two areas are:

  • The Digital Twin used to virtually define and optimize a new product/system or even a system of systems. For example, defining a new production line.
  • The Digital Twin used to be the virtual replica of an asset in operation. For example, a turbine or engine.

Digital Twin to define a new Product/System

There might be some conceptual overlap if you compare the MBSE approach and the Digital Twin concept to define a new product or system to deliver. For me the differentiation would be that MBSE is used to master and define a complex system from the R&D point of view – unknown solution concepts – use hardware or software?  Unknown constraints to be refined and optimized in an iterative manner.

In the Digital Twin concept, it is more about a defining a system that should work in the field. How to combine various systems into a working solution and each of the systems has already a pre-defined set of behavioral / operational parameters, which could be 3D related but also performance related.

You would define and analyze the new solution virtual to discover the ideal solution for performance, costs, feasibility and maintenance. Working in the context of a virtual model might take more time than traditional ways of working, however once the models are in place analyzing the solution and optimizing it takes hours instead of weeks, assuming the virtual model is based on a digital thread, not a sequential process of creating and passing documents/files. Virtual solutions allow a company to optimize the solution upfront instead of costly fixing during delivery, commissioning and maintenance.

Why aren’t we doing this already? It takes more skilled engineers instead of cheaper fixers downstream. The fact that we are used to fixing it later is also an inhibitor for change. Management needs to trust and understand the economic value instead of trying to reduce the number of engineers as they are expensive and hard to plan.

In the construction industry, companies are discovering the power of BIM (Building Information Model) , introduced to enhance the efficiency and productivity of all stakeholders involved. Massive benefits can be achieved if the construction of the building and its future behavior and maintenance can be optimized virtually compared to fixing it in an expensive way in reality when issues pop up.

The same concept applies to process plants or manufacturing plants where you could virtually run the (manufacturing) process. If the design is done with all the behavior defined (hardware-in-the-loop simulation and software-in-the-loop) a solution has been virtually tested and rapidly delivered with no late discoveries and costly fixes.

Of course it requires new ways of working. Working with digital connected models is not what engineering learn during their education time – we have just started this journey. Therefore organizations should explore on a smaller scale how to create a full Digital Twin based on connected data – this is the ultimate base for the next purpose.

Digital Twin to match a product/system in the field

When you are after the topic of a Digital Twin through the materials provided by the various software vendors, you see all kinds of previews what is possible. Augmented Reality, Virtual Reality and more. All these presentations show that clicking somewhere in a 3D Model Space relevant information pops-up. Where does this relevant information come from?

Most of the time information is re-entered in a new environment, sometimes derived from CAD but all the metadata comes from people collecting and validating data. Not the type of work we promote for a modern digital enterprise. These inefficiencies are good for learning and demos but in a final stage a company cannot afford silos where data is collected and entered again disconnected from the source.

The main problem: Legacy PLM information is stored in documents (drawings / excels) and not intended to be shared downstream with full quality.
Read also: Why PLM is the forgotten domain in digital transformation.

If a company has already implemented an end-to-end Digital Twin to deliver the solution as described in the previous section, we can understand the data has been entered somewhere during the design and delivery process and thanks to a digital continuity it is there.

How many companies have done this already? For sure not the companies that are already a long time in business as their current silos and legacy processes do not cater for digital continuity. By appointing a Chief Digital Officer, the journey might start, the biggest risk the Chief Digital Officer will be running another silo in the organization.

So where does PLM support the concept of the Digital Twin operating in the field?

For me, the IoT part of the Digital Twin is not the core of a PLM. Defining the right sensors, controls and software are the first areas where IoT is used to define the measurable/controllable behavior of a Digital Twin. This topic has been discussed in the previous section.

The second part where PLM gets involved is twofold:

  • Processing data from an individual twin
  • Processing data from a collection of similar twins

Processing data from an individual twin

Data collected from an individual twin or collection of twins can be analyzed to extract or discover failure opportunities. An R&D organization is interested in learning what is happening in the field with their products. These analyses lead to better and more competitive solutions.

Predictive maintenance is not necessarily a part of that.  When you know that certain parts will fail between 10.000 and 20.000 operating hours, you want to optimize the moment of providing service to reduce downtime of the process and you do not want to replace parts way too early.


The R&D part related to predictive maintenance could be that R&D develops sensors inside this serviceable part that signal the need for maintenance in a much smaller time from – maintenance needed within 100 hours instead of a bandwidth of 10.000 hours. Or R&D could develop new parts that need less service and guarantee a longer up-time.

For an R&D department the information from an individual Digital Twin might be only relevant if the Physical Twin is complex to repair and downtime for each individual too high. Imagine a jet engine, a turbine in a power plant or similar. Here a Digital Twin will allow service and R&D to prepare maintenance and simulate and optimize the actions for the physical world before.

The five potential platforms of a digital enterprise

The second part where R&D will be interested in, is in the behavior of similar products/systems in the field combined with their environmental conditions. In this way, R&D can discover improvement points for the whole range and give incremental innovation. The challenge for this R&D organization is to find a logical placeholder in their PLM environment to collect commonalities related to the individual modules or components. This is not an ERP or MES domain.

Concepts of a logical product structure are already known in the oil & gas, process or nuclear industry and in 2017 I wrote about PLM for Owners/Operators mentioning Bjorn Fidjeland has always been active in this domain, you can find his concepts at plmPartner here  or as an eLearning course at SharePLM.

To conclude:

  • This post is way too long (sorry)
  • PLM is not dead – it evolves into one of the crucial platforms for the future – The Product Innovation Platform
  • Current BOM-centric approach within PLM is blocking progress to a full digital thread

More to come after the holidays (a European habit) with additional topics related to the digital enterprise