Cognex Corporation

06/01/2023 | News release | Distributed by Public on 06/01/2023 07:44

Driving Affordability: How AI and Machine Vision Lower Electric...

Driving Affordability: How AI and Machine Vision Lower Electric Vehicle Cost

Electric vehicles (EVs) are here to stay. More manufacturers are committed to reducing vehicle emissions, going carbon neutral, or introducing more electric vehicles (EVs) into their lineups.

Consumer demand for EVs is growing, but the biggest speedbump on the road to mass adoption is the price tag. While EVs have reduced maintenance and fuel costs compared to their gasoline-powered counterparts, and governments are incentivizing purchases through tax rebates, EV prices are still cost-prohibitive for many motorists. (Bartlett, Preston)

Manufacturers Acknowledgment and Pursuit for Affordable EVs

The high costs of EVs are inhibiting widescale acceptance, and ultimately limiting the number of vehicles and batteries on the market. For years, automotive industry leaders have been grappling with how best to lower battery and ultimately vehicle prices. More recently, Tesla lowered prices as much as 20% in an attempt to capture more market share. Ford followed suit, slashing prices on numerous models.

Reducing EV prices often starts with the battery itself, which can account for up to 60% of a vehicle's cost. The confluence of expensive raw materials, opportunities for error in manufacturing, high demand of batteries and limited supply has caused battery prices to surge. Reducing costs per battery and scaling operations without sacrificing quality are all top manufacturing priorities.

In a 2020 Wall Street Journal article, Toyota's president, Akio Toyoda, acknowledged the high cost of sourcing raw battery materials as a significant challenge for EV adoption stating, "Converting entirely to EVs could cost hundreds of billions of dollars and make cars unaffordable for average people." Although Toyota plans to step up its electrification efforts, the cost of batteries needs to come down for EVs to become more affordable. The company is exploring alternative technologies to reduce battery costs and impact on supply chains.

In addition, there are other hidden costs of EV ownership, such as the complexity of their systems, which may require more battery, electronics, and charger maintenance than traditional cars. EVs also have higher insurance rates and the cost of electricity to charge an EV can be higher than the cost of gasoline, depending on the location and time of day. These factors could discourage consumers from buying EVs when they are weighing their options.

Domestic Lithium Refinement: Increasing Raw Material Supply

Although there are a few challenges, there are many reasons for optimism about the future of EVs. Elon Musk, the CEO of Tesla, recently urged entrepreneurs to help refine the production of lithium, a critical component of EV batteries, to reduce the cost and environmental impact of EV batteries. As more car manufacturers switch to EVs, economies of scale will kick in, leading to lower production costs.

When it comes to refinery, most of the world's lithium refining happens in China and is catered towards growing their own EV market. Fortunately, the USA is pushing for green energy and large companies like Piedmont Lithium, who's customer portfolio includes LG and Tesla, is trying to increase the amount of refinement done in the United States to meet the massive demand for lithium. According to Piedmont Lithium's CEO Keith Philips, "We currently produce around 20,000 tons of lithium hydroxide refined in the US. We think we need over 700,000 by the second half of this decade. So, 35 times more."

Piedmont Lithium is currently in the process of building a domestic integrated mine and refinery in North Carolina along with another facility in Tennessee in hopes they can drive up the number of raw materials refined to lower overall battery costs.

Automation in Battery Production: Lowering Costs and Ensuring Quality

Another way manufacturers are looking to lower costs and increase consumer confidence in EV technology is through automation. By deploying machine vision and AI-based technology for battery production, they can overcome labor shortages, run 24/7 manufacturing, ensure 100% quality control resulting in faster and higher production yields and safer, more reliable batteries. Below are some examples of how this technology has been applied.

Cap Welding Inspection

A critical aspect of battery production is the assessment of low-heat welds in battery cell caps. Poorly manufactured cells not only decrease efficiency but also create an uneven load between cells, making battery management challenging and reducing the overall lifespan of the battery pack. Once the electrodes and separator are enclosed within a cylindrical cell and filled with electrolyte, the housing is sealed with a precise low-heat welding method, to ensure a secure seal around the cap. However, it is crucial to examine and pass these welds before integrating the cell into a battery module or using it as a single cell, as any electrolyte leakage through a flawed weld can lead to reduced efficiency and potential short circuits within the battery.

To address this challenge, proper assessment of cap welds is essential for ensuring the functionality and longevity of the entire battery. Cognex's Deep Learning defect detection and classification tools have been trained on a wide range of weld variations. Through training, the system gains the ability to accurately classify and differentiate between different defect types, despite variations in the object and weld characteristics.

Injection Seal Inspection

Injection seal inspection involves examining the welds used to seal the filling hole after adding the liquid electrolyte to the cell. To avoid heat damage, a low-heat laser weld is implemented for this process. However, the presence of electrolyte contamination or welding flaws can lead to reduced cell efficiency. While an electrical test can help detect problems before installation, it is not entirely reliable. Therefore, proper assessment of the injection seal welds is vital for ensuring the functionality and lifespan of the entire battery. AI-based Deep Learning technology offers defect detection and classification tools that are trained on a wide range of weld variations, enabling accurate identification and differentiation between functional flaws and cosmetic variations, ultimately ensuring the quality of the welds.

Pouch Surface Inspection

During the degassing process, batteries take on an irregular shape, it is essential that pouches are smooth, unwrinkled, and unbent. To detect surface defects, such as bubbles and wrinkles, manufacturers employ automated inspection systems. However, the complex texture of the pouches presents challenges in identifying issues like wrinkles and bubbles. VisionPro Deep Learning is an ideal solution for this application. By training the deep learning model on a range of good examples of the standard pouch surface, the system reliably and repeatably identifies any irregularities that divert from this model, eliminating the need for extensive defect libraries.

Side and Top Panel Welding

In prismatic battery cells, a rectangular case is welded around an electrode sheet and the top panel is welded closed. It's important for the top panel or lid to accommodate the cell's expansion and contraction during temperature changes, especially considering that prismatic cells can be tightly stacked for efficient use of space. Before installing the battery cell in a module, it is crucial to inspect the seam welds on both the side and top panels for any defects. Proper assessment of these welds is essential for the overall functionality and lifespan of the battery. The appearance of these welds can vary significantly, with a wide range of possible defects and variations that affect performance. A combination of 2D and 3D vision systems along with deep learning technology accurately identifies any anomalies. Defect detection and classification tools are trained on good and bad weld seams allowing deep learning to identify and differentiate between functional flaws and cosmetic imperfections.

The high cost of EVs remains a significant barrier to their adoption. However, car manufacturers are stepping up their cost reduction efforts by refining the production of EV batteries with innovative machine vision technology. It may take several years for EVs to become affordable for the mass market as consumers weigh the upfront costs and hidden costs of EV ownership against the potential long-term benefits, such as reduced emissions and lower fuel costs.

Learn more by checking out the Electric Vehicles Solution Guide

Pete Sopcik | 06-01-2023

More Posts on