AI Set to Turbocharge the Factory Floor

Artificial Intelligence will transform manufacturing, allowing customers to attain previously unreachable levels of efficiency and quality.

Manufacturing and the industrial automation industry have always been quick to embrace innovation. Modern factories are now capturing rich data at every step of the manufacturing process. When this data is fed into new technologies that enable machine learning (ML) and artificial intelligence (AI), manufacturers can quickly achieve previously unattainable levels of efficiency and quality.

AI is fundamental to survival

By 2022, IDC predicts that 75% of enterprises will embed intelligent automation into technology and process development, using AI-based software to guide innovation. This has implications for industrial solution builders, who need to urgently rethink their products and services, or risk being left behind by more agile competitors. Technology and solution architecture must evolve quickly to support this transformation. You can learn more by reading our information brief: Smart Industry Meets Artificial Intelligence: A Modern Renaissance.

The edge and the cloud

We’re all aware that it has become increasingly difficult to store, manage, process, and secure the ever-increasing tsunami of data. Savvy manufacturers are responding by performing analysis at the edge and only moving the most relevant data to the cloud for further analysis. Apart from benefiting from real time decision-making and insights at the edge, this reduces the cost of transport and storage.

Making faster decisions with AI at the edge

As a result, we’re now entering an era where compute is following the data and where AI has become the main technology for data-driven innovation. Certainly, AI has the potential to deliver tangible benefits, significantly improving all aspects of manufacturing, including quality control by identifying potential issues earlier in the process. For example, applying AI to images and measurements taken during the various stages of a manufacturing process can help identify problems with a product as it is being assembled and before it’s finished. If AI can identify early on that the product is no longer viable, it follows that you get to save many further steps that involve time, material and resources. According to Forbes, AI has already increased defect detection in quality control by up to 90%.

AI enabling predictive maintenance

AI at the edge is also transforming predictive maintenance. Picture a motor in a factory line that continuously produces telemetry data as it runs and rotates. This motor creates vibration patterns in three axes and can be measured by computer. In the past, a summary of the data, including averages or outliers, would have been sent to the data center or Cloud for analysis. Today, with edge technologies, the compute resides right at the motor, where an AI inference algorithm can monitor and instantly send alerts, when something out of the ordinary occurs. Having this immediate access to unfiltered raw data is the key to identifying subtle trends that can indicate or predict failure. This would simply not be possible by performing analytics with consolidated or aggregated data.

Of course, while initial inference is happening at the edge, data will continue to be valuable in the cloud as the primary place for deep analytics, new rules and training models. As early warning symptoms for failures are identified by cloud-based training, these new rules will then be updated and redistributed out to the edge. I predict that this combination of a “data first” culture, backed by AI insights and unprecedented levels of automation, will transform everything from supply chain operations, materials, and order management through to production processes, factory maintenance, order fulfillment, logistics and services.

AI use cases enable transformational outcomes

Ultimately, success with AI means gaining new insights and capabilities from a pool or stream of data as well as acting on those findings in the most efficient and effective way. For example, the partnership between Noodle.ai, Dell Technologies, and Big Steel is revolutionizing Industry 4.0 with an Edge-to-Edge AI network while OTTO Motors has worked with Dell Technologies to create self-driving robots that use AI and LIDAR to navigate through facilities, move materials and enhance productivity for the world’s largest manufacturers.

Looking ahead, I believe that acceleration technologies, which perform parallel computation and faster execution of AI decisions than traditional CPUs, will play an increasingly bigger role in managing AI and ML workload. When accelerators are applied to the correct workload, you can expect to see a phenomenal increase in compute performance for training or inferencing. The proper gearing of offload accelerator to the correct type of AI workload ensures peak efficiency. While efficiency is appreciated everywhere to control costs, efficiency at the edge is critical. Proper efficiency lowers the thermal and cost challenges that enable compute, and thus AI, to extend further out onto the edge, where data is born.

All of these changes present huge opportunity for solution builders to drive innovation and identify new ways to build products more efficiently as well as create new revenue streams through providing new service offerings. Think about the potential to improve product quality, increase productivity, and enhance customer satisfaction. However, it’s not all smooth sailing. According to McKinsey, many organizations struggle to create AI value at scale, even though 99 out of 100 companies that have deployed AI report positive business value. You need to work with the right technology partner, who can help you design AI-powered solutions for your unique needs.

Connect the right technology with your workload

Dell Technologies OEM Solutions can help you and your customers connect the right technology with your workload at any stage in the AI journey. We enable solution builders to make compelling products that are designed on either industry-standard or customized hardware platforms. Both are capable of operating in the field in remote and harsh environments, with the same quality, reliability and security applied to data centers. Importantly, our customers benefit from access to our secure global supply chain, global services, and broad partner ecosystem.

What has been your experience of AI to date? Do join the conversation.

About the Author: Alan Brumley

Alan Brumley is Chief Technology Officer of Dell Technologies OEM Solutions, a specialist division helping customers to grow their businesses by leveraging Dell’s hardware, software and services as part of their own solutions as well as helping organizations take advantage of the rapid evolution in the Internet of Things. With 20-years’ leadership experience in the IT industry, Alan liaises with Dell EMC’s global business units to act as the voice of the customer. He is responsible for leading engagement activities with OEM customers, channel partners and technology partners. Collaborating with his CTO peers in the partner network, Alan and his team evaluate emerging technologies and promote early adoption and integration of breakthrough solutions with customers as well as designing new procurement options. Alan has over 18 years’ experience as an innovator with Dell EMC, previously designing Enterprise Solutions products in the Server Management Firmware group with a focus on modular enclosures that converge storage, networking, and compute capabilities, as well as server management solutions for these products. With a hardware and BIOS background, he holds over ten U.S. patents for the company and has delivered many first offerings in the Enterprise product family, including designing the Dell EMC PowerEdge M1000 enclosure. Alan started his career in a technology startup and has also enjoyed a five-year career as an innovator with Data Performance and Warehousing Solutions at NCR. Alan is excited that the Internet of Things has now merged his work focus with his personal passion for autonomous embedded systems and wireless communications, feeding data into Enterprise solutions for deep analytics.