Robotic arm welding
Industry 4.0 scalability and flexibility

01

5 min read

Introduction - The blueprint for scalability and process flexibility
Robotic arm

The challenges and rapid changes that characterize the modern business environment lay bare the inflexibility of many manufacturers’ operations. For more than a decade, Industry 4.0 and smart manufacturing concepts and initiatives have arisen to help combat these issues. New technologies,such as cloud, AI and edge computing, which have completely changed the IT landscape have accelerated their impact. Yet their full potential has yet to be realized as manufacturers continue to struggle to scale these advances throughout their operations.

Most industrial enterprises have implemented pilots or proofs of concept with Industry 4.0 technologies but have been unable to meaningfully expand these projects through out their plants and across locations. In many cases, these projects, intended to address specific issues, may even add more complexity to an already complicated heterogeneous environment.

A more holistic, technology platform approach is required. Enterprises must begin with the end in mind, with a vision for an open, flexible, integrated and AI-infused architecture that’s designed to scale. This approach doesn’t mean that high return on investment (ROI), quick-win technology projects shouldn’t be pursued. But rather, that they should be implemented with a broader understanding of architectural decisions within the context of a platform approach. In this way, isolated initiatives will contribute to the big picture: a fully modern digital manufacturing environment.

Worker in hardhat with tablet looking at two robotic arms

IBM brings hybrid cloud technologies to the factory floor

IBM’s extensive work with industrial clients has taught us that manufacturers need a guiding architecture. This blueprint for digital transformation provides the foresight to pull together the patch work of good intentions that exists across manufacturing plants today into a bold, unified vision for modern manufacturing. IBM’s architecture for manufacturing is driven by the goal of improving operations’ flexibility and guided by the principle of acting where the action is.

IBM’s hybrid cloud approach is uniquely open and provides a security-rich, industrial-ready foundation for manufacturers. It’s grounded on a common, industry standard platform for data, applications and cybersecurity, built on the Red Hat® OpenShift® Platform with Kubernetes containers. Our open hybrid cloud enables you to manage and run applications on your choice of public cloud, on-premises systems in your plant or on edge computing devices. IBM uniquely offers an on-premises cloud option, so you get all the modern tech advantages while keeping manufacturing data under your direct control.

IBM’s architecture brings hybrid cloud technologies to the action in each level of manufacturing.

Organize data
Rethink work with AI and automation
Modernize applications
Extend cyber security
Deploy open, flexible plan IT infrastructure

IBM hybrid Cloud Platform

Enterprise level ISA95-levels 4, 5

Wide area network

Plan levelISA95-level 3

local area network

Shopfloor ISA-96-levels 0, 1, 2

Process control netwrork

Three-Tier Manufacturing Topology

Our open architecture embraces the manufacturing ecosystem of operational technology (OT) vendors, industry independent software vendors (ISVs) and systems integrators(SIs). This architecture protects your current investments and allows you to continue to grow your business with the offerings and experience of leading OT vendors and others in the manufacturing ecosystem.

IBM Services® helps clients co-create intelligent workflows with preconfigured industry solutions and tools for automation and integration. With the IBM Garage™ methodology for Industry 4.0, IBM can help you define your vision and co-create a blueprint to achieve it. With a focus on using your existing OT and systems investments, use cases with the most important metrics and highest ROI are targeted first.

02

5 min read

Operate smarter with predictive insights

The advent of AI offers new possibilities to derive predictive insights and learnings which can improve plant-wide operations, but only if plant data can be unlocked from within these heterogeneous silos.

Worker in hardhat using computer equipment

Modern manufacturing creates terabytes of data, but only a fraction of itis acted upon. The typical plant operates across a labyrinth of multigenerational equipment, controllers and plant systems, each carrying out specific tasks.

Many manufacturers are analyzing data from individual machines or systems for equipment monitoring and predictive maintenance. But this analysis is only scratching the service of the potential of using data from across all equipment and systems. The advent of AI offers new possibilities to derive predictive insights and learnings that can improve plant-wide operations, but only if plant data can be unlocked from within these heterogeneous silos.

Sensors
TB of data for month

Each line could have 100 to 200 sensors that collect data continuously, generating 2,200 terabytes of data each month. Insights based on the aggregation, classification, and analysis of this data through AI and predictive analytics could increase productivity and reveal opportunities for data monetization.

Data must be organized into an “information architecture” that aligns with the various levels of manufacturing systems. From edge equipment to line, plant and enterprise systems, minimizing unnecessary data movements is key to reducing cost, security risk and latency. Specific use cases and automation objectives should drive the decisions about where various data should reside.

AI is applied to create actionable insights, leading to overall equipment effectiveness (OEE) improvement. Edge operations require near real-time analysis for time-sensitive activities like quality inspection and worker safety. At the manufacturing line, throughput is maximized using insights drawn from data integrated across equipment and controls systems, manufacturing execution systems (MES), maintenance and other systems. Optimizing production across lines and across plants requires broader visibility into enterprise systems, customer demand and the supply chain.

Architectural Decision

Where will you manage and organize your various manufacturing data? At the enterprise level in the cloud, in the plant or at the edge with the equipment?

Learn more
0
5
10
15
20
25

25% Predictive maintenance / condition-based maintenance

22%Assembly line creation and optimization

21%Quality assurance

21%Employ safety (including social distancing

20%Warehouse management, distribute, and logistics

19%Digital/data security

Q: Looking ahead, which of the following machine learning use cases that your organization has not yet deployed to you plan to deploy in the next two years? Base: Manufacturing industry respondents (n=154) Source: 451 Research’s Voice of the Enterprise: AI & Machine learning Use Cases 2021

With IBM’s hybrid cloud for manufacturing, you operate on the data where you need it, without requiring moving it to a central repository or out of the plant.

Our hybrid cloud approach lets you manage and use data across edge devices, on-premises systems and in public and private clouds in a security-rich environment. This approach allows you to bring AI to the data versus sending the data to AI in the cloud. IBM Cloud Pak® for Data software simplifies data collection, organization and analysis.

A flexible architecture means you can perform faster and less costly with more secured analytics to achieve your automation objectives. For instance, image capture from AI-driven visual inspection for quality generates huge amounts of data. It would add time, cost and security risk to move all the data to the cloud, so this analysis is best performed at the edge where the data is produced. Key data such as inspection results and associated exception data can be transferred to plant or enterprise databases. The data can then be used for further training of machine learning (ML) models, and analysis with other data to pinpoint root causes or for communication to engineering.

03

5 min read

Rethink how work gets done with AI and automation

The goal is to augment humans with the power of AI technologies, allowing both humans and machines to focus on tasks for which they’re uniquely suited.

Worker with tablet looking at machinery

Reduces production defects by 97% by applying AI/Machine Learning at the edge for quality monitoring. The Total Economic Impact™ Of IBM And Red Hat For Manufacturing, Source: Forrester 2021

Today’s workforce brings a new set of expectations to the job, as well as a digital-native skill base. It wants fewer rote tasks and more creative problem-solving. At the same time, manufacturing processes are under pressure for continued efficiency gains and improved flexibility. These forces lead manufacturers to rethink how work gets done.

Digital transformation is enabling new ways of working. Automation of repetitive tasks improves consistency and reduces costs. AI serves up insights to employees for improved decision-making, delivered through traditional computers, hardened handheld devices or even through augmented reality (AR). Augmenting humans with the power of AI technologies allows both humans and machines to focus on tasks for which they’re uniquely suited.

These new technologies enable “intelligent workflows” which integrate data insights across applications and domains to follow a more natural flow of work. This process doesn’t just improve the employee experience and job satisfaction by eliminating repetitive manual tasks, it also empowers employees all along the organizational ladder to make better, data-driven decisions.

The transformation program must address the skills and cultural challenges for existing employees on the shop floor who need to transform their skills and learn new ways of working.

Architectural Decisions

How will you orchestrate and automate new intelligent workflows? How will you access data from the plant, edge and enterprise? Where will you run these workflows—at the edge, plant or in the cloud?

Learn more
Illustration of a bar graph representing currency

Leading organizations expect to realize an average ROI of 24% within three years from edge computing investments. Source: IBM IBV

IBM co-creates with clients to automate processes and build intelligent workflows

With IBM’s preconfigured industry solutions, new intelligent workflows, infused with AI, extend to the closest point of action and integrate processes between machines, the plant and the enterprise. The IBM Garage methodology for Industry 4.0 collaboratively prioritizes development to the highest ROI use cases and takes a human-centered approach to facilitate employee buy-in.

IBM Cloud® Paks software automate business processes and integrate data flows. Key IBM technologies that bring workflows to life in a manufacturing environment include the IBM Plant Service Bus and IBM Watson® AI.

IBM Plant Service Bus includes manufacturing protocols to integrate across industrial equipment and automation tools, while IBM Watson AI draws insights from all types of data, including visual and acoustic data. Edge computing solutions bring the power of IBM Watson AI to workers and machines to capitalize on data and extend the experience of experts in the organization.

04

5 min read

Increase OEE by building and modernizing next-generation apps

Co-creation of modern applications pairs developers with plant subject matter experts, ensuring solutions are designed to meet manufacturing needs.

Welding sparks falling down

According to the director of digital transformation, there was always a risk of downtime, since systems were cobbled together over more than 30 years, spanning regivons and divisions within the interviewed organization. By implementing a platform approach, it eliminated downtime, while improving quality and agility. The Total Economic Impact™ Of IBM And Red Hat For Manufacturing, Source: Forrester 2021

The average plant relies on a portfolio of hundreds of aging, heterogenous plant applications. Some applications are proprietary or may be fraught with lack of documentation. Other apps may be back-level or running on unsupported operating systems. Such a patch work of siloed apps is costly to maintain and enhance and puts production continuity, quality and safety at risk. The advent of more edge computing applications coming alongside 5G is poised to expand the challenge even further. This issue grows exponentially for large organizations that are operating dozens or hundreds of individual plants.

Fortunately, there’s a path out of this jungle. Modern DevOps techniques apply modularized development in small chunks, so applications can be rapidly deployed and expanded over time. Cocreation of modern applications pairs developers with plant experts, ensuring solutions are designed to meet manufacturing needs.

Manufacturers need to adopt an architecture that grows with their business priorities, one that facilitates integration and can adapt as needs change. New applications should be aligned with the information architecture, with workloads performed at the edge, line and plant levels.

An open, container-based approach enables the creation of modular systems to speed deployment, increase reuse, simplify maintenance and easily scale. With modern application development technologies, you can be less reliant on specialized skills, enabling workforce flexibility and more nimble, efficient operations.

Architectural Decisions

How will you modernize plant and edge applications for more resiliency without risking stability and operational disruption? Will you use proprietary or open-source development techniques?

Learn more
IBM’s hybrid cloud approach builds cloud-native applications using Red Hat OpenShift with industry-standard Kubernetes containers.

This leading platform brings the advantages of open-source technology, but is “industrial-ready,” with the reliability, security and support needed for critical operations. IBM’s open platform enables new applications to run in essentially any operating environment—private clouds, on premises or a choice of public clouds.

IBM’s solution includes a reference semantic model that handles the mapping of integration points, such as data, OT, other applications and so on. This semantic model separates the complexity of the plant environment from the applications to simplify development and allow changes to the operating environment to be reflected in all affected applications.

With the principle of bringing the applications as close to the point of action as possible, IBM® Edge Application Manager lets you create, deploy, run, maintain and scale edge solutions basically anywhere.

IBM’s solutions help reduce costs; speed development; enable tuning new applications for edge, line and plant systems; and provide scalability and flexibility for the future.

05

5 min read

Secure assets, information and people

How will you identify cyber attacks within the plant and respond swiftly? How will you secure data and access points within the plant? How will you combine security monitoring in the plant with IT?

Man looking at a laptop holding a clipboard
A yearly research study conducted by IBM shows that in 2019 there was a 2,000 percent (20x) increase in OT cybersecurity attacks. Yes, you read that correctly — 2,000 percent. As operational organizations and industries experience innovation and connectivity, bad actors can take notice and execute security attacks. To help mitigate these risks, organizations can adopt an operational technology security strategy.

As the connectedness between manufacturing equipment in the factory—and across factories— increases, so does vulnerability to attacks. The average plant runs hundreds of apps. It’s no wonder that cyberattacks on industrial facilities are up 20 times in the last two years. The outcomes can be catastrophic financially or, in some cases, even life-threatening.

In this increasingly risky environment, manufacturers need to be sure that all software and control systems can be monitored and protected in a clear framework to avoid gaps. End-to-end visibility allows for a more consistent security practice, including threat response and continuous learning. Security needs to be embedded within all processes and at every point of data transmission.

Architectural Decisions

How will you identify cyberattacks within the plant and respond swiftly? How will you secure data and access points within the plant? How will you combine security monitoring in the plant with IT?

Learn more
IBM helps you define and develop your comprehensive security blueprint for the IT and OT infrastructure.

Building applications and workflows with IBM’s hybrid cloud comes with built-in security in all IBM environments, and across third-party ecosystems. The IBM QRadar® architecture and the IBM Cloud Pak for Security solution help you manage threat monitoring and response from the enterprise to the edge, for full visibility of cyberattacks. IBM is collaborating with OT vendors like ABB to extend the security perimeter into OT equipment and the Internet of Things (IoT) platforms.

With 8,000 dedicated professionals around the world, IBM is one of the largest enterprise security vendors and brings the experience of managing security for leading banks, retailers and other global enterprises to industrial environments.

06

5 min read

Deploy an open, flexible plant IT infrastructure

The infrastructure choice between manufacturing and cloud doesn’t need to be either-or. A multi-cloud approach enables manufacturers to achieve the best of all worlds.

Person working on an assembly line
It has been demonstrated that the value derived from a full hybrid, multicloud platform technology and operating model at scale is 2.5 times the value derived from a single platform, single cloud vendor approach.

Traditionally, manufacturing has exerted direct control over the IT infrastructure to run plant applications and data management in systems on premises, and often outside the realm of IT. As cloud computing has emerged to offer attractive new technologies, manufacturers have the option to shift workloads to public clouds. Yet plant managers express concerns that public clouds introduce a loss of control, increase the potential for security threats and create latency issues. Additionally, some operations are restricted from using public clouds due to data residency and privacy regulations that restrict movement of data beyond national boundaries, or lack of accessible data centers in some regions.

Manufacturers need the flexibility to move workloads between on-premises systems, including edge devices, and to different public clouds, depending on latency requirements, costs, security and other considerations. The infrastructure choice between manufacturing and cloud doesn’t need to be either-or. A multicloud approach enables manufacturers to achieve the best of all worlds.

A modern infrastructure uses a software-defined approach to optimize computing, networking and storage resources for maximum control and flexibility. At the same time, it allows the enterprise to share solutions and rapidly scale new capabilities globally, across hundreds of plants, with significantly reduced IT operating costs.

Architectural Decisions

How will you design the technology architecture for your plant? Consider options for connectivity, data management and processing? Consider the level and scope of IT skills to preserve within the plant?

Learn more
Illustrations of electronics

According to the director of digital transformation, there was always a risk of downtime, since systems were cobbled together over more than 30 years, spanning regivons and divisions within the interviewed organization. By implementing a platform approach, it eliminated downtime, while improving quality and agility. The Total Economic Impact™ Of IBM And Red Hat For Manufacturing, Source: Forrester 2021

IBM’s open, hybrid cloud architecture gives you choices instead of limitations.

Adopt IBM’s hybrid cloud approach, based on Red Hat OpenShift, and your manufacturing infrastructure can run on essentially any public cloud, reducing the risk of vendor lock-in. Your infrastructure can also run on private on-premises systems, so you can retain control within your operations.

IBM’s open approach embraces OT providers and public cloud vendors to allow you choice of the best infrastructure for particular workloads. We work with network equipment and telecommunications providers to enable edge computing and low latency operations.

IBM brings cloud technologies into your manufacturing environment, so you can locally apply AI, ML, 5G edge computing and software-defined networking and storage. Our approach brings the tech to the action, rather than sending your data outside of the plant. Run your workloads wherever is most appropriate for the job, whether that’s at the edge, an on-premises server or a public cloud —basically any public cloud. The IBM Cloud Satellite™ solution lets you scale these deployments across multiple types of infrastructures and across plants and the world.