5 min read

AI for Green Manufacturing

Mike Mohseni, William Malkes

Competitiveness in manufacturing is often synonymous with production efficiency. However, the environmental consequences of manufacturing processes, particularly welding operations, are frequently overlooked. This article sheds light on the environmental implications of defects and reworks in welding processes, emphasizing the urgent need for sustainable practices in manufacturing.

Production efficiency and loss reduction are closely tied with clean manufacturing. The US Qualifying Advanced Energy Project Credit [1] program specifically includes an eligible category under “Energy Efficiency and Reduction in Waste from Industrial Processes”.

Greenhouse Gas Emissions from Welding

Sources of Greenhouse Gas Emissions. Source: United States Environmental Protection Agency.

Cost of Bad Quality

The MIG welding process is integral to the modern manufacturing landscape, underpinning the construction of essential infrastructure such as bridges, buildings, ships, and automobiles. However, welding has hidden environmental costs, particularly when it involves rework or scrap due to faulty welds.

Rework involves measures such as grinding, cleaning, and re-welding, with costs sometimes exceeding 2-3X the original budget [2]. It is estimated that 5% of the cost of all welding operations is the cost of defect and rework [3]. In automotive production, for instance, 2-4% of parts may have welding issues, with up to 1% deemed unsalvageable and scrapped.

Material Loss and Its Environmental Impact

Welding defects contribute to material loss, particularly in industries that rely on metal fabrication processes, such as automotive manufacturing. Even a low scrap rate of 1% due to welding defects can result in significant steel wastage. This increases the environmental burden of material extraction, processing, and transportation, and leads to additional carbon emissions. For instance, a 1.5% scrap rate in automotive manufacturing can release over 131,000 tons of CO2 annually. Considering higher scrap rates in other industries, the total CO2 emissions due to welding defects could exceed 524,000 tons annually.

Energy Loss and Its Environmental Impact

Rework processes in welding consume additional energy, contributing to higher overall energy consumption and greenhouse gas emissions. For instance, the electricity required for grinding faulty welds, re-welding, and production of welding consumables adds to the environmental impact. In the US alone, over 87,500 units of MIG welding equipment operate for at least 8 hours a day, consuming over 1.75 billion kWh of electrical power annually. With a rework rate of 3%, approximately 52.5 million kWh of energy is wasted each year, leading to about 58,000 tons of additional CO2 emissions annually.

A typical workflow in automotive parts manufacturing.

How AI Can Enhance Green Manufacturing

Due to the complexity of real-time monitoring of MIG welding processes, manufacturers mainly rely on post-production defect inspection for quality control. Although safeguarding production output quality, this approach does not address the financial and environmental costs of rework, scrap, and operational downtime associated with late discovery.

Proactive Detection with AI

Employing AI for proactive prediction during the manufacturing process can mitigate these inefficiencies. Early detection of welding quality issues prevents the progression of faulty welds to a point where salvaging the product becomes impossible, thus eliminating costs associated with scrapped materials.

AutoMetrics’ approach in leveraging machine learning to learn and monitor the process quality instead of product defect enables a proactive measure to avoid scraps and find and address defects at the source. This ultimately increases production efficiency and through saving over 582,00 tons of CO2 emissions, it significantly improves sustainability. AutoMetrics’ Inspection 4.0 platform is proven to predict over 13 types of variabilities, and sources of weld defects, in real-time with minimal false alarms.

Industry Applications

In industries with continuous production lines, like automotive parts manufacturing, detecting the initiation of weld defects early can prevent the production of defective parts. For example, a truck chassis comprised of over 100 welding joints could face significant rework costs if a defect forming on the 50th weld goes unnoticed, potentially leading to defects in subsequent welds.

Ensuring consistent welding quality across manual and automated stations is crucial in shipbuilding,

where cutting and MIG welding are primary processes. Many manual and automated/robotic welding

stations work together to place hundreds of meters of joints in building ship structures. Variations in

welding quality at a single station can result in meters of defective weld joints, posing significant

fabrication challenges. Proactive monitoring of welding quality can prevent such costly mishaps.

Conclusion

AI technologies offer a transformative potential for green manufacturing by improving production

efficiency and significantly reducing environmental impacts. By integrating AI for proactive quality

control in welding processes, manufacturers can enhance their competitiveness and contribute to

global sustainability efforts.

Appendix: Derivation of GHG Emission Figures

Material Loss

Automotive manufacturing in the US consumed 25% of the overall 98-million-ton national steel reserve [4]. The majority of structural components of a car are currently made of steel and except at the final assembly, are joined together using MIG welding process. Reasonably, the weight of the parts in a car with involvement of this process closely represents the total weight of the steel used to build the car.

Considering a low scrap rate of 1% due to welding defects, the loss due to welding defects can be up to 245,000 tons of steel. For critical structural components such as seating, the post-process quality inspection of welds includes destructive testing of 0.5-1% of the built components wherein these samples are cut upon for evaluations hence scrapped afterwards. This adds at least another 122,500 tons of steel scrap.

Either scrapped due to defects or destructive testing, these wastage increase the overall consumption of raw materials, heightening the environmental burden associated with material extraction or recycling, processing, and transportation. The carbon footprint of steel recycling is 0.357 tons per ton of steel [5]. This indicates only a 1.5% scrap rate in automotive manufacturing cycle, can release an additional 131,197 tons of CO2, annually.

Automotive manufacturing consumes a quarter of the US overall steel quota and follows high execution standards relative to shipbuilding and construction industries, which leads to lower scrap rates. Therefore, a realistic figure for CO2 emission due to lack of defect prevention and material loss is over 524,000 tons per year.

Energy Loss

The rework process involves grinding out the faulty weld, re-welding, and potentially re-inspecting the welds, each step consuming additional energy. The electricity required for power tools, grinding equipment, and the production of additional welding consumables adds to the environmental impact, contributing to higher overall energy consumption.

A welding machine consumes 10,000 W of power, in average, and over 87,500 units of MIG welding equipment [Footnote 1], manual and robotic, are in operation in the US, i.e., perform at least 8 hours a day over a 250-day work-year. This volume of machinery consumes over 1.75 billion kWh of electrical power annually in the US. With a rework rate of 3%, 52.5 million kWh of energy is wasted due to repeating the same welding process. Considering other steps of rework including grinding, surface prep, cutting, scrap, and usage of consumables, this figure realistically increases up to 78.3 million kWh per year. This is equivalent to about 58,000 ton of additional CO2 emission, annually [Footnote 2].

Footnotes

There are 366,000 operational industrial robots in the US [6] of which est. 12% are used for MIG welding [7]. Assuming one welding machine used for manual welding setup per welding robot (actual figure may be higher given low robotic system density [8]), there are roughly 87,840 units of MIG welding machines in operation in the US.
Generating every kWh in US roughly emits, 1.04 kg of CO2 with coal and 0.44 kg of CO2 with natural gas [9], average 0.74 kg.

Additional Notes

A typical passenger vehicle emits about 4.6 metric tons of carbon dioxide per year [10]
10% of global GHG emission is from cars and vans (all transportation 28%)
20% of global GHG emission is from manufacturing [11] that also consumes 54% of the world’s total energy resources [12] (2022 stats)
Welding can account for 20 to 30% of the cost of manufacturing automotive equipment and large pipeline projects [13].