Designing the Future

Share this on social media:

Issue: 

Gemma Church investigates how modelling and simulation tools are used to design new components and systems

When designing a component or system, where do you start? In days gone by, designers would literally go back to the drawing board, using their experience and expertise to develop a new concept. Now, simulation and modelling have unleashed a new paradigm, with a wealth of innovations continuing to push the design tools field forward.

This increase in demand for simulation and modelling techniques comes at a time when advancing digitisation and product complexity are already pushing the boundaries of our current design capabilities.

James Dagg, CTO at Altair, said: ‘Increasing product complexity at the part and assembly level is the overriding trend. Leading companies ensure collaboration between design, engineering and manufacturing bring their innovative products to market faster.’

This intrinsic need for simulation and modelling in the design tools space is a result of advancing digitisation and electrification. Paul Brown, senior marketing director of the product engineering group at Siemens Digital Industries Software, said: ‘If we think about how products have evolved, the role of electronics and software has increased substantially. This situation is the same for the majority of industries, where companies must deliver new concepts to market quickly and in the face of growing product complexity. Simulation and modelling must aid communication between the electrical and mechanical domains to achieve this.’

Simulation and modelling tools must also be available across the design lifecycle. Dagg added: ‘Because of this increased complexity and stringent time constraints, engineers and designers need to be able to evaluate their products’ manufacturability in the conceptual development phase, evaluating different processes based on their performance and production constraints.’

Such measures not only simplify the design process but also provide extensive cost and time savings for the company and its design teams. Dagg said: ‘Integrating Altair SimSolid into Altair Inspire eliminates the time-consuming and expertise-extensive tasks of geometry simplification and meshing and also provides a single, easy-to-use environment where design, simulation, and manufacturing engineers can make rapid design iterations.’

 ‘The same Inspire user experience now provides manufacturing process simulation of casting, extrusion, stamping, and additive manufacturing process. This ensures greater confidence in the early design direction,’ Dagg added.

 Siemens Digital Industries Software has also integrated machine learning into its software interfaces, guiding specialist and non-specialist users through the design process, suggesting the next step based on past habits. ‘This has scratched the surface of the capabilities of AI in the simulation space, but it is an important step forward,’ Brown added.

Additive manufacturing is another area driving the capabilities of simulation and modelling forward, according to Brown, who added: ‘Additive manufacturing is changing the way people think about design problems. We can now complete CAD at an industrial scale where we have batch quantities of thousands and iterate over different types of materials.’

This has led to further innovation at Siemens Digital Industries Software. ‘While topology and parameter optimisation tools are required to suggest additive geometries, they can also come up with wild shapes that work from a mathematical point of view but may not be feasible in the real world. So, we created Convergent Modelling,’ Brown explained.

 The company’s Convergent Modelling technology provides a unified 3D modelling capability for two data formats with different geometric characteristics – facet and B-Rep data. Facet data is characterised by triangular meshes which approximate surfaces and tend to generate larger data sets for applications including digital mock-up, animation and gaming. B-Rep data is characterised by mathematically defined surfaces that represent solid volumes using smaller data sets. Applications that use B-Rep data include 3D modelling for engineering and manufacturing.

A proliferation of facet data in high growth areas like 3D scanning, topology optimisation and 3D printing means that design engineers often need to move facet data into 3D modelling systems that were designed for B-Rep data. Convergent Modeling Technology enables 3D product modelling on both data sources in a single environment while eliminating the complexity, error and delay of converting between the two.

Brown added: ‘We are trying to get the models to model the design intent. This is blurring the lines between the data sources and negating the need for reverse engineering.’

Machines learning

Artificial intelligence (AI) is a rapidly growing area for design tool engineers and scientists, which is being integrated into workflows across multiple industries.

There is much machine learning can achieve in this space – but we first need to manage the disparate skills required to run successful AI-enabled simulations. Jos Martin, director of engineering at MathWorks, explained: ‘Engineers and scientists can greatly influence the success of an integrated AI project because of their inherent engineering knowledge of the data and domain. This is a significant advantage over a data scientist who is not as familiar with the domain area.’

‘However, traditional AI tools often require deep programming skills familiar to computer scientists or data scientists, but engineers and scientists need tools that allow less time writing code, more time exploring innovative design ideas, and more quickly improving the effectiveness of those AI models.’

A lack of high-quality data is another barrier to success for many fledgeling AI projects, according to Martin, who added: ‘Often engineers and scientists don’t have enough data that incorporates rare events with high enough frequency to effectively train the AI model to properly handle them.’

More data is also required for robust training in new areas, such as predictive modelling, where the availability of failure data, in particular, is essential. 

Martin explained: ‘Failure data is a crucial part of teaching algorithms to recognise the warning signs that trigger just-in-time maintenance. However, failure data is often not readily available, and producing failure data is time-consuming and expensive. However, it is possible to easily and cheaply simulate failure data and train models using that data to recognise warning signs from operational data.’

 Using MATLAB and Simulink, specialist engineers can generate sample failure data, which is then labelled and used to build an AI model to accurately predict the remaining useful life of the equipment in question, according to Martin.

 Virtual commissioning is another growing technique in the industrial automation and machine design sector, according to Chris Harduwar, vice president of automation at Maplesoft. ‘Typical machine commissioning involves hooking prototypes up to control hardware, and this is often when engineers and integrators find a variety of performance issues they need to address – this process can iterate for weeks or months,’ he explained. ‘With virtual commissioning, the engineers developing control code can use machine simulation models as virtual test platforms for validating their control code.’ 

This can help customers ‘reduce costs and time to market tremendously,’ according to Harduwar, who said: ‘Customers can use MapleSim to create a functional, dynamic model of their machine, which can be used to help size motors, optimise cycle times, and so on. This model can be exported for use in a variety of different automation tools, so when the engineer wants to test their control strategies, they can get realtime feedback, and even see a 3D visualisation of their machine in action.’

Harduwar added: ‘While new techniques like virtual commissioning require internal skills development, they continue to offer benefits that can’t be denied by competitive organisations. To make sure that we can keep pace with the increased demand for simulation-based design processes, we’re constantly adding new usability features to our simulation tool MapleSim, and we’re offering streamlined options for project consulting that allows companies to benefit from these new techniques while learning these skills at the same time.’

Maplesoft recently worked with a customer to validate their motor sizing choices for a new injection moulding machine. They also needed to test their control strategies virtually, since physical machine testing posed a high risk of damage to their machine – a cost they wanted to avoid.

By using results from virtual commissioning, the company could identify the precise loading requirements for its new motors and motion profile, eliminating the added costs of oversized motors. ‘Furthermore, the machine’s control strategy was thoroughly tested against the dynamic model, preventing the risks of damaging the actual machine during testing. With our help, the new machine could be offered at a lower price than before, while still offering the same high standard of reliability required in the injection molding industry,’ Harduwar added.

Next-generation trends

What’s next for design tools? We can expect many of these trends to continue growing and expanding, as Dagg explained: ‘We anticipate greater cooperation between departments and specialised domains. To support this, we will be continuing to extend our toolset to drive increased early simulation at more organisations, regardless of who performs it. Our goal is to enable greater product innovation through better collaborative decision making across product teams of designers, analysts, and manufacturing engineers.’

‘The ‘democratisation of simulation’ isn’t just about putting some simulation functionality into tools for designers. It must address the larger goal of making simulation an effective tool to perform rapid ‘what if’ studies at the speed of concept design,’ he added.

Simulation and modelling will continue to lower a primary barrier to successful AI adoption – lack of data quality – according to Martin.

‘We know training accurate AI models requires lots of data. While you often have lots of data for normal system operation, what you really need is data from anomalies or critical failure conditions. This is especially true for predictive maintenance applications, such as accurately predicting remaining useful life for a pump on an industrial site,’ he explained. 

‘Since creating failure data from physical equipment would be destructive and expensive, the best approach is to generate data from simulations representing failure behaviour and use the synthesised data to train an accurate AI model. Simulation will quickly become a key enabler for AI-driven systems.’

Another emerging form of machine learning – reinforcement learning – will also move into the mainstream. Using reinforcement learning, a computer effectively learns to perform a task through repeated trial-and-error interactions in a dynamic environment, allowing it to autonomously make decisions.

But reinforcement learning requires a lot of training in order to reach acceptable performance. ‘Even for relatively simple applications, training time can take anywhere from minutes to hours or days,’ Martin added.

That won’t stop reinforcement learning from extending its reach in the design tools space. Martin explained: ‘This year and beyond, reinforcement learning will go from playing games to enabling real-world industrial applications particularly for automated driving, autonomous systems, control design, and robotics. We’ll see successes where reinforcement learning is used as a component to improve a larger system.’

‘Key enablers are easier tools for engineers to build and train reinforcement learning policies, generate lots of simulation data for training, easy integration of reinforcement learning agents into system simulation tools and code generation for embedded hardware,’ he added.

 An example is improving driver performance in an autonomous driving system. AI can enhance the controller in this system by adding a reinforcement learning agent to improve and optimise performance – such as faster speed, minimal fuel consumption, or response time. ‘This can be incorporated in a fully autonomous driving system model that includes a vehicle dynamics model, an environment model, camera sensor models, and image processing algorithms,’ Martin added.

Remote working is another trend, which will transform not only how we work but also our simulation and modelling tools. Brown added: ‘We need to not only provide take-home licenses but also ensure the security of our software solutions as people increasingly work remotely. To protect our customers, vendors have to ensure the software is free from any trapdoors or entry points for malicious players. At Siemens, every business line has a security officer to provide this diligence.’

 As the years pass, simulation will continue to provide cost savings, according to Harduwar, who said: ‘Fixing issues with mouse clicks is always going to be cheaper than fixing physical machines. For that reason, we’re expecting that virtual design techniques will continue to grow, and adoption will increase across industries.’

Exclude from view: