Manufacturers are increasingly drawn to Industry 4.0, aiming to enhance profitability, safety, productivity, agility and adaptability. Yet, the transition is slow, with a low adoption rate of technologies required for a smart factory transformation. 

According to a Manufacturing Leadership Council study, only 4.5% of respondents currently consider their factories “very smart” — although aspirations are high, with 11.3% expecting to reach this level by 2026. Meanwhile, the majority (53.4%) see their factories as evolving but still in progress.

The factories of the future are envisioned as highly automated, powered by intelligent systems that process vast datasets with minimal human input to significantly increase efficiency and production flexibility. Artificial intelligence (AI) especially has vast potential to drive smart manufacturing once adoption barriers are overcome. 

Examining the role of AI in smart manufacturing

Madhu Gaganam is an engineering technologist specializing in the manufacturing domain at Dell Technologies, which collaborates with industry-leading partners like Intel to provide innovative, expert-driven AI solutions to customers. When asked about the evolution of AI in smart factory transformation, Gaganam indicated, “A smart factory needs to be flexible and adaptable to deliver operational efficiencies. This is where AI adoption can provide greater agility, productivity, safety and predictability of operational outcomes.”

When discussing AI use cases, he noted that opportunities can be segmented into three key areas — Factory Floor, Facilities and Enterprise Intelligence. Top examples for each include: 

• Improved Asset Reliability on the Factory Floor: AI can improve asset reliability against agreed-upon performance and reliability measures for assets, applications, servers and networking topology by predicting and prescribing recommendations.
• Efficient Utilities Management in Facilities: AI enables the monitoring of water, air, gas, electricity and steam from asset-to-facility level to monitor and manage wastage to meet sustainability goals. For example, AI-based monitoring can reduce, balance, or stop usage of utilities during unplanned downtimes, non-use of production and facilities assets and cellular manufacturing.
• Improved Planning and Forecasting for Enterprise Intelligence: With AI, factories and plants can perform predictive analytics by interpreting historical data to improve the planning of resources on the shop floor to meet future demand. For example, just-in-time inventory management, production and workforce scheduling.

Other AI use cases include Improved Yield and Throughput (e.g., identifying bottlenecks based on real-time constraints); Enhanced Safety (e.g., personnel monitoring in restricted areas); and Increased Workforce Productivity (e.g., using generative AI to streamline task execution during maintenance and repair).

Legacy pitfalls and new necessities for AI at the Manufacturing Edge

According to Gaganam, traditional approaches to deploying heterogeneous vendor solutions have improved efficiencies and delivered positive business outcomes. However, today, this approach could pose greater long-term operational challenges when integrating AI-based, diverse vendor solutions from a cost and effort perspective. Existing technologies and applications must be rationalized, and newer ones must be onboarded to deliver AI-based outcomes. 

For example, edge computing enables manufacturers to access, analyze and act on data right where it is created to uncover and generate immediate value. However, given that legacy compute devices at the edge are not commonly built to support AI-based applications, manufacturers looking to deploy an AI solution that leverages them must consider the following: 

• Computing power and storage capacity must support the vast data set requirements. 
• Data and systems must be safeguarded to prevent and protect against security threats. 
• Proliferation of AI use cases must be managed with a platform that helps rationalize and modernize existing applications. 

“By addressing these critical areas, maintenance and operations teams can develop efficient platforms for the management and orchestration of edge resources, thereby overcoming pitfalls and setting the stage for successful deployment of AI at the edge,” Gaganam explained. 

NativeEdge is an enabling technology

Enabling data-driven edge computing for transformation is like walking on a balance beam. Manufacturers want to move forward, but at the same time, they must remain steady and ensure connected, secure and robust operations. Many find that existing or legacy infrastructure makes progress difficult and hinders modernization effort, and they require an operational platform to help ease this effort.

Showcased recently at several key industry events, Dell Technologies has introduced a solution that can serve as that bridge for manufacturers. NativeEdge is an edge operations platform designed to securely and efficiently orchestrate smart manufacturing applications, including AI, across multiple domains. 

The foundational, horizontal platform empowers manufacturers to: 

• Run multiple applications in a technology-agnostic, protocol-agnostic manner.
• Make updates and fixes with zero-touch from human input.
• Gain centralized operational control for a holistic view of your operations.
• Collect and connect data across silos and domains.

“The capabilities of NativeEdge enable manufacturers to take essential steps toward modernization, providing them with the necessary computing power and advanced security needed to succeed,” Gaganam indicated. “For example, NativeEdge features Zero Touch Provisioning and Zero Trust, which operate in air-gapped environments. Manufacturers receive secure hardware that will not work if it is tampered with prior to delivery, as well as FDO secure device onboarding.”

Best practices for NativeEdge adoption

As organizations look to leverage NativeEdge technology, understanding and implementing best practices is crucial for success. Gaganam advises manufacturers to begin by identifying edge use cases that yield significant operational benefits, such as computing close to data sources to enhance performance. 

For manufacturers considering Far Edge deployment, evaluate which architecture best meets those needs: a single AI-based inferencing application on one compute device, one application with multiple microservices across several devices, or a single application spread across multiple devices for enhanced redundancy. Similarly, for Near Edge, they should consider whether their needs best align with model training solutions with data repositories, a solitary application on a single device, or more complex setups involving multiple devices. 

Along with the infrastructure to support flexibility and security, change management is also important. According to Gaganam, “As edge and AI technologies continually evolve, upskilling your workforce is essential. By focusing training on the latest in edge computing solutions for manufacturing and harnessing of NativeEdge, organizations can advance operational efficiencies and innovation. 

How NativeEdge intersects with the Industry 4.0 journey

The journey to Industry 4.0 presents an evolving range of concepts and technologies. Although these have enormous potential to deliver manufacturing outcomes, they can become siloed deployments and disparate systems.

Every manufacturing organization must reduce the risk of losing the efficiency gained by these technologies by accounting for the maintainability and operations involved. 

Gaganam noted that many organizations opt to start by standardizing the computing and storage infrastructure, which acts as the foundation or backbone for deploying industry 4.0 use cases and realizing economies of scale when it comes to maintenance and operations throughout the journey.

“This infrastructure needs to reflect self-secure features for easing critical risk from the operations point of view,” he explained. “NativeEdge adoption provides this option right out of the gate. It provides a structure and, thereby, a best practice to onboard compute devices and applications. Once development and operations teams become familiar with the NativeEdge process, they can govern their applications and devices with a standard operating procedure with minimal upskilling.”

NativeEdge is a platform that can be used at any point in a manufacturer’s journey to Industry 4.0.

“Organizations that are slow adopters due to resource constraints can expedite their modernization,” Gaganam indicated. “On the other hand, organizations that take a slower path due to new value-creation initiatives could also adopt NativeEdge as their foundational management and orchestration platform for ease of scaling various technology-oriented applications in the future.”

For more information on simplifying the path to smart manufacturing outcomes with the Edge, check out Dell Technologies Manufacturing Edge Solutions page.