How AI and IoT can become game changers for manufacturers amid COVID-19: Part 1

It is well known that AI and IoT are the levers for the new industrial revolution, and companies that embrace these technologies produce goods at a lower cost, improve customer experience, and offer better quality. Now we’ve used these technologies to help maintain a safe working environment for employees as they get back to work. In this blog you can read why AI and IoT are crucial for the health of your employees and your competitive position.

AI reduces manual work and improves quality control

One area where AI and IoT can mitigate the negative impact of COVID-19 is in quality control, which is a highly labor-intensive process. In most cases, quality inspectors follow classic statistical process controls (SPC) and select only a few samples of the manufactured products that go through the comprehensive lab tests to determine whether products meet the initial requirements. Design of experiments (DOE) may then be used to determine the root cause of detected quality defects. Both methods are often manual and time consuming, and require deep domain expertise in the current process design of the specific product – we’ve seen situations where running experiments within the DOE work took months, and was still unable to locate the problem. This has a profoundly negative effect on the value chain and ultimately leads to dissatisfied customers and higher costs, which could be avoided with automatic quality inspection via IoT and AI.

Some of the results we’ve seen of leveraging these tools in the manufacturing industry are compelling:

• Product Defects: down 2.8%
• Plant Uptime: +1.3 %
• Plant Productivity: +5.7%
• Energy, Wattage: – 6.5%
• Premium quantity: +20%
• Yield: +1.3 %
• Material Waste: – 4%
• Operation Stability: +9.2%

There are multiple quality assurance techniques that leverage AI. One of the most effective has proven to be visual inspection powered by AI algorithms. Instead of sampling, a camera continuously analyzes manufactured products – all of them – for the characteristics that are important. Cameras have greater visual accuracy than human labour and also work 24/7. AI tools analyze the data from the cameras and detect the smallest deviations. The quality manager receives a signal if products deviate from the standards or the production process stops immediately. This not only saves a lot of manual work, but prevents problems during production, reduces rejected batches and improves customer satisfaction. This technology also gives manufacturers more control over production processes and interventions are faster. Modern AI-based visual control systems don’t require expensive cameras and utilize cost-effective edge-to-cloud processing: machine learning models that detect defects are trained in the cloud taking full benefit of storage and compute power, and then executed in real time at the edge (factory floor). We’ve seen situations where just 30 defective items were sufficient to train accurate models, making this technology very affordable and scalable.

Data analysis identifies the causes of production errors

Thanks to IoT and AI, manufacturers have significantly more control over the quality of their products. They can now take proactive steps to see what abnormalities exist in the process and remedy them as they occur by taking a data-driven approach. As this critical data is collected and analyzed, managers can precisely determine the deviation ratio with higher reliability and validity. The data is then automatically compared with numerous other datasets such as production time, type of raw material, and thousands of process parameters, which gives managers the transparency to locate deviations. With this information at hand, adjustments in production processes can be made in near-real time, thus reducing the margin of error and costs.

Automate social distancing on the factory floor

Increasing efficiency and reducing costs are always high on the agenda of manufacturers. Since the outbreak of COVID-19, new safety measures have also been at the top of the priority list. In many cases, they are even conditions for restarting the production process. But how do you arrange social distancing in the workplace? We have already applied this in our offices at Capgemini and it is also easy to use in factories.

Guarantee better work safety

There are several IoT options to support manufactures in reopening operations. One option leverages infrared sensors that monitors worker presence at specified spots; with analytics for managers to observe enterprise policy compliance while preserving privacy. Another technology uses BLE (Bluetooth Low Energy) beacons, worker badges, or smartphones. With this data, AI analyzes the distance between colleagues in real time and if the threshold is exceeded, a signal is sent. This can be a message to the manager or an automated audio warning that keeping distance is in everyone’s favor. With this, managers also see whether something is wrong in the workplace because of the sudden gathering. Dependent on the administration requirement, such analytics may be completely anonymous. At the same time, provided that employees share consent to use the beacons data to warn them of risk, the same technique can be applied for automated contact tracing if someone tested positive to COVID-19. Colleagues from the same team or shift that were exposed to the risk of contamination will be notified, and disinfection teams can clean up specific areas.

Start the factory of the future now

If you want to learn how to set up your factory more efficiently with IoT and AI, we are offering an operational assessment. Please reach out to smartcovidresponse.nar@capgemini.com.

For more information, watch the webinar we recently ran with Microsoft, “How AI and IoT are Changing Manufacturing and help to control COVID-19 on the factory floor.” In this webinar, Patrick van Loon (Microsoft) and Sergey Patsko (VP AI & Analytics, Capgemini) show you how to set up the factory of the future. Register now for the webinar here.

This blog is the first part of a two-part series. Read the second part here.