How are companies meeting the need to innovate as the products they create get increasingly complex and interconnected? Innovation undoubtedly gives a company competitive advantage but there is no one prescription for innovation. Going digital, however, can make a company more agile, quicker to respond to market needs and customer demands.
As businesses evolve and adjust, so does the workforce. Businesses have gone from single locations to many, with management, sales, research and development, management, procurement and other company functions happening in different locations. The current COVID-19 pandemic has fractured business functions even further, with many workers working while physically isolated from coworkers.
Dr. Jan Leuridan, Senior Vice President, Simulation and Test Solutions at Siemens
With multiple locations, offices and homes spread across the globe, being competitive even with such drastic changes make the need for digital solutions even more pressing, says Dr. Jan Leuridan, Sr VP at Siemens Digital Industries Software, Simulation and Test Solutions (STS).
We connect with Dr Leuridan via phone at his home near Antwerp, Belgium, to understand why companies in all industries need to embrace a digital transformation.
Dr Leuridan, thanks for meeting with us. Let’s start by telling us which customers Siemens Digital Industries Software serves?
The customers that we're serving are first and foremost companies in the discrete manufacturing industries. These include leaders in automotive, aerospace, heavy equipment, industrial machinery, marine, energy, electronics, consumer goods, and medical industries. We also have companies in process industries that rely on our solutions both to design sophisticated manufacturing equipment and to optimize their processes.
Regardless of the industry, there are some essential challenges these companies are faced with across the globe.
What are these challenges and what is technology’s role as businesses evolve?
Any industry that looks toward the future is driven by innovation, striving to outdo their competitors. If not, they simply cannot compete. At the same time, everybody continues to be challenged relentlessly to reduce time to market while at the same time catering to consumer needs of more personalized products.
Innovation has a number of different dimensions as companies are increasingly making smart, intelligent products. That requires fusing various domains of engineering to design and manufacture the product. These intelligent products are often part of a larger system, even systems of systems. So, these products can no longer be designed in isolation, which makes development even more complex to manage.
Other topics companies need to consider are the use of new materials, new production methods and use of composites and metals. Sustainability and traceability throughout the lifecycle are important to consumers.
On top of that, there is an increasing synthesis between materials engineering and the actual engineering of products, where in the past that used to be separate. These novel techniques substantially increase the degrees of freedom one has to develop a product, while at the same time reducing the possible reliance on past experience.
Another challengehas to do with change in business models. This might mean companies are no longer simply producers of products anymore but instead offer the services that are associated with the product, relying on increasing connectivity for smart monitoring to receive feedback. Take a look at Rolls Royce, for instance. They’re not so much selling aircraft engines as they are the use of the propulsion the engine provides, whether it is billed on flying time or maintenance.
Sounds complicated.
Complexity comes with innovation. However, trying to reduce complexity is not an option, because that means limiting the ability to innovate. The goal must be to turn complexity into a competitive advantage. Those who are best organized to manage and deal with complexity are the ones best positioned to drive innovation quicker than their competitors.
What do companies need in order to rise to this level of complexity?
There are three focus areas I feel are worth mentioning to maintain a competitive advantage.
The first is to utilize a comprehensive digital twin. It is needed for all the different stages in the lifecycle of a product—i.e.product design, product manufacturing and performance of the product. It is imperative to have the best possible capabilities to virtually represent the product and manufacturing process that is as close as possible to what the real product or the real manufacturing installation will be. This virtual representation or digital twin then becomes the basis for making a range of decisions that impact user experience, performance, manufacturing, in fact the entire product lifecycle.
Secondly, companies need digitalization solutions that are personalized, adaptable, and modern. This means the journey of digital transformation is unique to every company. However, it's a journey where everybody in the company must be involved. Digital technologies must be made available in a way that can be consumed by many different people, with different roles in the company, no matter if that’s a small or medium business or a global enterprise. The key here is to make design and manufacturing solutions as accessible and easy to use so they will be as widely adopted as possible.
Thirdly, industrial companies need an interoperable, highly accessible and open ecosystem in which they do business. We support this by offering open solutions that are interoperable with existing customer flows requires standards but also underlying technology kernels, such as Siemens Parasolid with over four million users. However, this ecosystem goes much further, as the future of industry will increasingly rely on a network of partners, suppliers, consumers, app developers and others.
What is Siemens doing to promote these areas?
Siemens has developed and is constantly enhancing our portfolio of solutions to what we call Digital Enterprise. It comprises a unique combination of software and automation solutions, as the virtual and physical worlds of design and manufacturing are moving closer and closer together. Software is what speeds up this digital transformation. We summarize this under our Xcelerator portfolio. This portfolio combines tools for mechanical design, automation and electronic design automation and powerful simulation capabilities. Additionally, we offer manufacturing planning software, manufacturing operations management (MOM) software, analytics tools and an app development platform. With this, we feel confident to serve our customers as a trusted partner for their digital transformation.
An integral part of this journey is to make use of a comprehensive digital twin. The digital twin is a virtual representation of the product and the manufacturing process, thereby covering the entire lifecycle. It is enriched over time by analyzing data coming from machines in the manufacturing process or the product being used out in the field. By feeding these insights back into the development process, one can improve the fidelity of such a digital twin. All of this is what we refer to as the comprehensive digital twin. It is used to predict behavior, optimize performance and validate the most complex products and processes.
There are different types of models that enable the digital twin representation. One way is to think of a very highly detailed geometric model for the rendering of a product before it is produced. Or another aspect is a model that represents how a product behaves when in usage: How much energy will it consume? How much noise will it make? Or how much emissions will it release? In other words, performance engineering.
A lot of what we do with our Simcenter portfolio of simulation and test solutions is focused on creating models for studying the performance of a product, which then enables the optimization of a product for performance requirements. When designing a product, there are requirements that you try to work with to get to a design. Simulation and testing are critical to help design the right product to meet these performance targets. Our investments in this portfolio are designed to enable five key initiatives related to engineering that allow our customers to innovate and bring revolutionary new products to market faster.
What are the five areas?
Front loading engineering, hybrid engineering, generative engineering, model-based engineering and continuous engineering. It’s worth going through them one-by-one.
What do you mean by frontloading engineering?
When a customer first starts engineering a product, they want to determine as early as possible, meaning the concept phase, whether the approach they are considering will work, before converting to the requirements for the design. This is what we refer to as frontloading.
Computer aided design (CAD) embedded analys is tools are now available for design engineers to rapidly evaluate possible design alternatives and quickly innovate to find the best design. Being fully embedded in the design environment, the tools use the same native geometry for analys is, eliminating the need to export the data and heal it for analysis. Accelerating simulation opens the door for new capabilities to seamlessly transition from one design to hundreds.
Early use of systems simulation is another example of frontloading. It allows to do a lot of “what-if” analysis based on 1D models even before a CAD model has been created.
Let’s use an automobile as an example. In this industry,system simulation is important to test drive requirements. If a car must meet a certain standard for fuel economy, what does that mean in terms of requirements for different parts of the car? What does it mean for the transmission or the engine? Should the vehicle be a hybrid platform or an electric car? All these considerations need to happen early in the development of a product, and that requires system simulationto cascade down requirements for components and subsystems.
Finally, companies are increasingly deploying sophisticated computer-aided engineering (CAE) tools for multi-discipline and multi-physics simulation that reduce the reliance on testing and further enable the push towards frontloading of engineering decisions.
This is why frontloading is so important — the earlier that you can check that a new design meets requirements, the less corrections are needed downstream in the development. It's always going to be more cost effective to implement corrections earlier in a design process.
What do you mean by “hybrid engineering?”
I mentioned reducing the reliance on testing, but it is also true that our customers continue to invest in testing. Furthermore, testing is more and more used as a key element of a hybrid engineering approach that combines test with simulation. Using this approach enables simulation models to be validated with real loads, enables some simulation models to be augmented with test-based models, and to enable simulation into testing itself. Siemens is uniquely positioned to help our customers progress this area due to our significant expertise and on-going investments in both simulation and testing methodologies that complement each other.
Another area where customers benefit from a hybrid approach is when you implement various XiL (X-in-the-loop) strategies. I am referring here to hardware-in-the-loop, software-in-the-loop, model-in-the-loop and even person-in-the-loop approaches to verification and validation. Such approaches rely heavily on a combination of simulation and test to speed-up the validation process.
By generative engineering, do you mean generative design?
It is more than that. Generative engineering is really a paradigm shift that relies on simulation to generate a design that meets requirements. Let me explain: The established way for designing a product is to start with requirements, then develop concepts and designs and systems, and deploy simulation to check whether it's a good concept, a good system or a good design. In contrast, generative engineering uses simulation to generate designs that meet requirements. It accelerates the pace of innovation by creating a systems architecture of mechanical, electrical, electronic and software domains, using simulation and AI-driven concepts.
Instead of starting with a design and then verifying its performance against requirements, the process starts with a specification of requirements, constraints, and context. Frontloading simulation and AI-based methods are then used to generate innovative designs, allowing for shorter validation and verification cycles. It is now possible to apply generative engineering to system-level architecture providing the ability to optimize across multiple engineering domains.
At each step of the design process, from requirements to fully operational system, there is a continuous loop of generating design options, verifying those options against key criteria and validating the design meets increasing complex design requirements.
Applying generative engineering during system architecture design maximizes the design space exploration in concept development. The process begins with identifying various building blocks that could make up the system architecture. The combination of these blocks can create a large amount of possible system architectures. Until recently, this approach could only evaluate a maximum of 10 different system configurations, limiting design performance and innovation.
However, recent advancements in generative engineering now support the simulation and evaluation of more than 1,000 systems, dramatically opening the design space for exploration.
We apply these methods to, for example, help companies in the very early stages of development generate system architectures, such as designing a new hybrid vehicle. There are so many ways to design a hybrid vehicle, but frequently companies will keep to established design approaches. When you go about doing things the way they’ve always been done, that stifles the freedom to come up with a better design.
Generative engineering principles also apply to detailed design as it includes topology optimization. Some of the interesting applications we have seen our customers address include combining topology optimization with design space exploration and using generative design in combination with CFD to optimize flow related behavior.
What do you mean by “continuous engineering?”
There is an increasing need as well as opportunity for our customers to use the simulation models, built when designing the product, downstream in other parts of the lifecycle of the product.
Think about a bottling machine. When a machine builder is developing a next generation bottling machine, they are dealing with increased complexity while reducing time to develop their next generation equipment. Simulation models make sure the machine does what is intended before committing to the final design. When the machine is eventually installed, it will need to connect into certain controls on the manufacturing floor. Using the simulation model for the bottling machine, the purchaser can start doing some of that calibration of the controls, or virtual commissioning, before the actual machine is installed.
Then as it operates, real sensors on the machine continuously generate information on critical measurements. This data can then be fed into simulation models, providing richer insights into how to best optimize the machine. It is even possible to use the simulation to probe into areas where there are no actual sensors, with virtual sensors
Virtual sensors use information from physical sensors in combination with the simulation model to simulate the response as if it had been measured in real-time. Neural networks allow users to convert these models in real-time as well. The simulation model is used to generate lots of training information for these neural networks. Then you can efficiently feed that neural network with data from real sensors on the machine. Now it is possible for the neural network to create information on virtual sensors for areas on the machine where there is no physical sensor. This can be used, for example, to simulate virtual diagnostics to better understand what's happening in the machine.
That is how simulation models are part of the lifecycle of the product. We sometimes talk about the executable digital twin: a simulation model that’s been embedded in the final product, so it becomes an IP within the product.
Can you give another example where a digital twin could provide an advantage?
Sure. A traditional anti-lock braking system (ABS) system in a car only kicks in when the car already starts slipping. This is too late. It would be much better if that ABS system could start just before the car is slipping. Let’s now assume that we have an exact model of the car running inside the car while driving. That model takes all the same sensor information as the real car, such as speed, steering wheel angle, camera showing the upcoming turn in the road, and could calculate upfront the potential of running into problems in the next seconds and take corrective actions before the event takes place. That’s a real value that the simulation model can bring once embedded in the final product.
Can you tell us how model-based engineering plays into how products are designed and manufactured?
Yes, model-based systems engineering (MBSE) integrates multi-domain engineering with considerations like cost, reliability, and manufacturability; all to create a support system that models requirements, design, analysis, verification and validation: from concept through development and into later lifecycle phases.
The underlying trend: Companies want to organize how they develop the product by using models that represent the different domains of the system or the product, such as electronics, mechanics or software. They are designing and combining these models like little LEGO bricks, if you want to think of it like that, to build models for assembled products, systems, and so on.
The MBSE approach uses digital threads to weave together crucial capabilities, including requirements management, continuous verification, closed loop validation, multi-domain and simulation-driven design engineering all based on a common platform and user interface. It also needs to integrate the application lifecycle management for a wide range of software solutions both developed internally, as well as those that will be supplied by external partners.
Do you have an example of model-based system engineering?
Of course. Let’s take autonomous driving vehicles. To simulate an autonomous driving vehicle, you need models for the sensors to simulate how they perceive the world. You then need to combine these kinds of models with the vehicle dynamics, with control functions, with the input coming out of those sensors and even AI methods for processing the sensor data. You can simulate what the camera sees, but also simulate the analysis of the image to see if it correctly identified whether it's person, car or a bike. All this before moving it into the control function.
This is really pushing the industry to start thinking about organizing how to do engineering work with models, for different parts of the system and products, and being able to have solutions where these models can be combined to build those system to improve the performance of these products.
Dr. Leuridan, thanks for your time. What would you like our readers to take away from this interview?
At the end of the day, only companies that know how to manage complexity will remain innovative and thereby gain a competitive advantage. By embracing digital twin technologies and solutions that are personalized, adaptable and modern and by making use of an interoperable, highly accessible and open ecosystem, they won’t just make their products better, but bring them to market faster and more cost-effectively.