Stress testing design for digital engineering

In this article, Amit Visrolia, Chief Engineer for Digital, at the National Composites Centre (NCC), discusses the evolution of design, and if it will survive the stress test when applied to manufacturing at scale.

What does an engineer do? It's a simple enough question and for centuries the answer has been, ‘make  things’, or ‘solve problems’. If we’re feeling very confident, we might even go as far as saying, ‘engineers apply scientific discovery and make it useful’.

The rise of digital, where complex systems are seamlessly linked, data shared and laborious processes automated, challenges us to consider another, more nuanced question: ‘what can engineers be now?’

In essence we are wired to make things better.  Engineers want to apply their creativity and skills to design and produce products with ever higher levels of performance, fully exploiting the capabilities of the materials and processes at our disposal as we go.

In reality, how often do we actually do this?  As engineering has evolved, so has it fragmented, reflecting different – sometimes conflicting priorities: we’ve had ‘design for cost’, ‘design for manufacture’ and increasingly ‘design for sustainability’.  However, with digital the divides come down. We’re no longer designing ‘for’, we’re designing ‘from’ – with data shaping and refining every stage of the process.  The promise is that engineers will be able to apply their imaginations as never before, freed from the tyrannies of validation and checking, and uncowed by the ever-growing complexity of modern products. 

Maybe!

Big data certainly has the potential to be game-changing for manufacturing. The digital toolkit, from Machine Learning and Artificial Intelligence to High-Performance Computing and Virtual Reality, presents huge potential. But, right now,  it feels like the hype is running ahead of hard evidence.

The terms “AI” and “machine learning” are sprinkled so liberally across any technology offering that its easy to overlook that the question of what, exactly, they can do for a physical industry such as manufacturing is still pretty unknown.  Will digital survive the stress test when applied to manufacturing at scale?

This is what we’re investigating at the Digital Engineering Technology Innovation (DETI) Programme.   

We are all familiar with the Systems Engineering V model which has formed the structure of engineering manufacturing for the past 20 years: the drawing up of requirements and specifications, moving to integration and production to validation and verification in a linear, sequential pattern of work.  Each stage is engineered in its own silo of specialisms before moving onto the next one – shifting from “Are we building the right thing?” to “Are we building it right?”.  It’s tried and tested - it’s just that we know, we can work better.

Digital engineering flattens the V, as all the parameters need to be considered in parallel.  It requires a multidisciplined approach, where teams work concurrently, with the majority of effort focused on the design and development stage where most cost (and compromise) are locked in.  This is about integrated design driven by data – from the machines making the product to customers and back again, virtual testing to design-out problems early and fast manufacturing cycles.

It’s a model already successfully used in software engineering.  In mobile phones for example, teams work in parallel, designing the operating system, camera and casing as one, holistic action; given the nature of the product, where space, power and capability are so finitely linked, it is the only way to do it. 

The challenge comes in applying this approach to bigger, more complex products, which may have hundreds or thousands of components, each with their own set of impacts. Few companies have the bandwidth, especially in the current climate, to take a forensic look at the digital toolkit and work-out, from first principles, how their operations and structures might need to change in order to extract the most value from it.

Fortunately for them, industry came together with the support of the West of England Combined Authority and our region’s leading universities to create DETI – a unique initiative designed to ‘road test’ digital and find out what works in the real world.  

We are applying the tool kit to a series of test cases, the first being the production of a design system for a hydrogen pressure vessel. The majority of these (about 90% of the market) are constructed using metals, but with increasing focus on hydrogen powered transportation, comes growing demand for composites to create strong but lightweight alternatives.

The project team includes engineers from design, analysis, materials, and manufacturing automation working together to bring forward the complexity of detailed design and manufacturing data at an earlier stage in the design process.  We will use a model based systems engineering (MBSE) framework and a multi-disciplinary optimisation (MDO) platform and toolkit, which we will have down-selected based on the outcomes of trials being conducted in one of DETI’s core ‘enabling capabilities’. We will model the numerous combinations of design decisions that go into a product to create the initial design concept.  As an interconnected, multidisciplinary design system, we will rapidly be able to identify the impact on each choice on the overall performance of the product and optimise to get the ‘best’ output – for whatever we consider the best.  The equipment used for the filament winding of the carbon fibre will be fitted with multiple sensors to then gather and feedback the data into the platform to verify the choices, and to inform the next iteration of the design.

We’re also investigating how products can be optimised, not just for performance and cost, but sustainability too, exploring the practicalities of embedding lifecycle analysis seamlessly into the design process.  All too often the question is asked towards the end of the design process, when the incentive to make significant change may be crushed under the weight of costs already accrued.

In-process verification is one of the greatest challenges for complex products. Virtual testing could be a major enabler but is currently stuck at the component and sub-element level; predicting what will happen at the system scale is a way off yet.  For now, we are working to design-out problems that could cause re-calls or reduce iterations of products, two elements that are costly in both time and money. 

In our role as path-beater we will encounter some dead-ends and bumpy roads, but that’s the point: we do the work so UK businesses don’t have to.  Already we are seeing industry engaging with DETI and keen to know more about how digital can work for them.  The current economic climate has put processes and costs under particular scrutiny, of course. The dramatic shift to remote working, with the new routine use of digital tools, may have altered perspectives too.  

But I believe there is more to this than just practical consideration. In the years ahead society is going to have to reconcile a series of extraordinary, complex, interlinked demands, bound by the need to do better while using less.  In other words, there are problems which engineers will be called on to solve – and that means getting to grips with digital now.  The role of engineers will not be diminished by digital technology: it will be enhanced.  The right questions must be asked of the data. Domain knowledge will remain critical. Machines, however smart, do not push boundaries. Only we can do that.

Sign up for DETI news and updates