Skip to content

Harnessing the power of data will help weather the chip shortage storm

Technologies like AI and HPC are allowing brands to be more intelligent about vehicle production in the long run, writes Christian Ott

The global chip shortage has put the brakes on automotive recovery, with automakers unable to keep up with demand due to the silicon supply shortage. Car sales in the UK, for example have experienced a phenomenal bounce, with SMMT recording 32 times the number of new vehicles registered in May 2021 compared to 2020. Electric vehicle sales are also on the up, with Ofgem research finding that 6.5 million households plann to buy an electric car by 2030. But what use is demand if there is no supply?

There is no silver bullet for offsetting the impact of the global chip shortage. Once again, we find ourselves using the word ‘unprecedented’. Silicon chips take between six and eight weeks to manufacture—plus shipping and distribution. Understandably, the demand for silicon nosedived as lockdowns across the world plunged industries, including automotive, into a cloud of uncertainty. So, chip manufacturers reduced production. In a world of just in time (JIT) production, stockpiling silicon wasn’t an option. As the automotive industry recovers and people start to buy cars again, chip production will ramp up, but there will be a lag between this happening and OEMs receiving the chip supplies they so urgently need.

The future revenues in the automotive industry will not come from selling cars but from selling additional services around cars, and all those will be data driven

While this is yet another challenge which the COVID-19 pandemic has posed for automotive manufacturers, it’s a good time to take a step back, and look at where automakers can make up for lost time. Technologies such as artificial intelligence (AI) and high performance computing (HPC) enable OEMs to increase production efficiency, reduce costs, and produce higher quality vehicles, while data is helping them understand their customer better, so they can produce the right vehicles for the right people. The future revenues in the automotive industry will not come from selling cars but from selling additional services around cars, and all those will be data driven.

From device to car to factory 

Given the excitement around autonomous vehicles, it’s easy to forget that connected vehicles have been around since the mid-1990s. Despite this, the data explosion that came with software-defined cockpits has accelerated rapidly in the past five years. This has been a game-changer for automakers in the sense that it has allowed them to learn more about their customers, such as how and where their cars are driven, after they’ve been purchased. Furthermore, this data is being collected in real-time.

Gone are the days where brands’ opportunities to learn more about their customer were limited to the point of purchase, annual service, replacement, and/or renewal. Connected cars are producing data all the time which can tell OEMs whether their customers’ driving needs are changing. And we’re not talking about the type of ‘black box’ data which is used by insurers to determine your premium, although this is also important. Automakers can now find out whether drivers are performing more long journeys. Are they carrying more passengers? What types of roads do they drive on? All of this information is highly relevant when it comes to renewal, enabling OEMs to offer more personalised deals and discounts based on usage.

Bosch software defined car
Technologies like AI and HPC are allowing OEMs to be more intelligent about vehicle production in the long run

Similar datasets, married with sophisticated AI, is enabling Formula One teams to analyse in-race performance based on data collected from over 200 vehicle sensors in real-time. This can help teams shave all-important milliseconds off their lap times, as well as improve fuel economy and inform racing strategy.

While singular nuggets of data are important, data sets are more powerful when viewed collectively. Automakers are also able to view and analyse colossal data sets which have been anonymised and provide macro-level information which can be used to enhance vehicle performance. For example, OEMs can get a sense of their vehicles’ fuel economy levels across the board based on how efficiently every single one of their cars is being driven. This is incredibly useful as there may be design tweaks which they can make to improve fuel economy and help their drivers to achieve more miles to the gallon. This type of predictive maintenance is just one example of how Big Data can help automakers improve vehicle performance by responding to data collected in real time.

Data and digital twins

Digital twins and design simulations are invaluable techniques that help OEMs put into practice the learnings from the data they have at their disposal. Given the ubiquity of CAD/CAM, OEMs are increasingly looking to create digital twins of every vehicle to ensure no stone is left unturned in the production process. These digital twins can then be tested via simulations and computation fluid dynamics (CFD) to see how the vehicle drives, corners and brakes. Manufacturers can even use acoustic modelling software to find rattling or minor noise which needs to be engineered out of the car design before it is produced. This process involves thousands of simulations being conducted simultaneously on an HPC platform, producing huge volumes of data.

Technologies such as artificial intelligence and high performance computing enable OEMs to increase production efficiency, reduce costs, and produce higher quality vehicles, while data is helping them understand their customer better

This data must then be analysed to provide the insights that engineers need to make the necessary design tweaks to produce the performance the driver needs, which is where we come full circle. OEMs, therefore, need a means of marrying up and visualising in one place data that has been collected from their vehicles in the field and their customers, with the data being produced by the CAD/CAM and CFD processes. The digital twin is going to become a digital representation of physical objects. It could be the customer, the car, a production facility or a machine in a production facility. The key challenge will be to join all those dots through a common descriptive language, so all these objects can interact in a digital world.  It’s about obtaining a single source of truth so that the sales and marketing teams are working from the same information as the mechanics and engineers. The way to achieve this is through a data fabric, which provides portability, visibility, and seamless access of data across all types of storage: from physical, to virtual, to cloud.

While there is no answer to the global chip shortage other than to ramp chip production back to pre-pandemic levels and accept that disruption will persist temporarily, technologies like AI and HPC are allowing OEMs to be more intelligent about vehicle production in the long run. Combined with a more sophisticated use of the vast data sets at their disposal, this will allow automakers to produce better quality products based on more insightful information about what their customers really need from their vehicles.


About the author: Christian Ott is Director, Solution Engineering, Global Automotive at NetApp

Welcome back , to continue browsing the site, please click here