Throughout history, customer demand has always been a driving factor in companies’ product portfolios around the world. Today, the demand for a cleaner earth is unparalleled, with individuals wanting products and services that minimize harm to the environment.
This amazing transition in customer demand is apparent everywhere. The food industry is experiencing a rise in vegan requests, the government is cracking down on businesses for carbon emissions, and the automotive industry is developing more and more vehicles based on electric and battery-powered engines.
In fact, the demand for electric vehicles (EVs) is notably large. Today, roughly 30-45% of vehicle buyers in the U.S. and Germany consider purchasing an EV when getting a new car. This high demand to create more green mobility solutions is backed by government regulations and customers’ desire to reduce their carbon footprint.
With such a drastic change comes a few challenges, especially for those developing and creating EVs. While customers want to heal the planet and move towards electrification, they also want to drive their electric vehicles the same distance as their internal combustion engine vehicles—without a charge—and they want the vehicles themselves to be affordable.
Electric vehicle manufacturers need to take these demands into consideration when designing and manufacturing. The battery pack inside the vehicle must be high performing, lightweight, and inexpensive to create—requirements that generally don’t go hand-in-hand.
The challenge: optimizing powertrain design, production, and performance
At the heart of EV design is the development and optimization of a complete electric powertrain, Although the powertrain is an essential component in all vehicles, shifting from an ICE to an electrified powertrain comes with its own set of pressing challenges. The powertrain is a central component in EV performance, and, at the same time, it can dangerously drive up production costs. Given the fierce competition in the market, EV manufacturers must find ways to optimize their powertrain production process. Why is this such a challenge?
The manufacturing process of powertrains is extremely complex and expensive to rework. This is the case for a few reasons. First and foremost, the powertrain has many parts, including the battery pack, cooling systems, sensors, and motor. Only after you have installed all of these subsystems together within the vehicle can you know if it was manufactured successfully. This poses a problem, as it’s best to find problems within the manufacturing line before the powertrain is even assembled.
Another reason powertrains are difficult to manufacture is that they rely on many different technologies that are incorporated through irreversible processes, such as welding in power modules. Managing all of this is challenging, as it produces a lot of data, requires continuous analysis, and is sensitive to a wide variety of influences, such as incoming material variation, process drift, and environmental factors.
So how can you manufacture powertrains that are high quality, safe, and inexpensive? By collecting, aggregating, and analyzing your manufacturing data in order to find discrepancies in the powertrain manufacturing process as early as possible.
Let’s go over some tips on how to get this done.
The solution: bringing product data analysis into focus
“Collect, Clean, Harmonize, and Correlate.” That’s your new EV manufacturing motto! All parametric test data should be collected at every assembly level throughout the product hierarchy to enable real-time process optimization and predictability. Here’s how you can make this happen.
Collect: centralize and structure your data
All of your data (from the past and present) should be available in one place. To do this, you can create and form a data lake that will break down silos within the powertrain production line. Every piece of data associated with the EV production line should be in the data lake, including data from the on-board charger, battery pack, electric drive module, PWR module, EV inverter, and more.
Clean: normalize your data so it can be easily processed
Make sure that all your data is normalized, structured, and formatted so that the best and most accurate data analytics can be conducted. This is one of the most important, and unglamorous, parts of the process, as it is where GIGO (garbage in, garbage out) will come back to haunt you. Without the data being properly formatted and cleaned, your analyses won’t get anywhere.
Harmonize: aggregate and analyze the data
Harmonizing data means tearing down silos so that your data is easily accessible and can be compared across operations, products, factories, and infield use. Take, for example, battery packs, where data from the following departments must be included in your data lake and analyzed:
- Total EV drivetrain
- In Use
This is crucial, as information from one department can greatly help another. For example, data on why an electric vehicle was returned could bring to light a powertrain manufacturing or design error.
You can use advanced machine learning algorithms to determine problematic material and predict which components or subsystems of the powertrain will create problems down the line. For example, this technology can be used to find cracks in welds through machine vision technology, and it’s more accurate than human inspection, as it is not susceptible to training variances or emotional states such as feeling tired.
Correlate: find and apply the insights in real life
Now that your data is ready to go, you can use it to derive insights into your products and manufacturing processes that are not only meaningful, but actionable. Being able to do this in real time so that you’re not needlessly wasting/scrapping good products is all the more crucial given the cost of these systems. Additionally, you can use past test results as the base to determine which parts of the powertrain are problematic and which perform well every single time.
The result: creating high quality products with minimized scrap
As things stand in the EV market today, to develop high-quality products at competitive prices, manufacturers must embrace the Industry 4.0 paradigm shift and adopt a Smart Manufacturing strategy. As we detailed above, this entails putting product data analytics under the center spotlight. Manufacturers should strive to have full visibility across the supply chain, aggregating and correlating data at every assembly level throughout the product hierarchy, as well as capturing failure signatures.
This data can then be analyzed with AI algorithms offering real-time optimization and prediction, minimizing scrap, improving NPI ramp, boosting sub-system quality and reliability, and reducing RCA to solve problems more quickly—all the while driving higher quality overall. This approach can be invaluable when it comes to powertrain production. Optimizing the production of powertrains can lead to lower production costs, ultimately lowering the price for consumers and allowing manufacturers to gain a competitive edge.