All News & Insights

ML Journal October 2021

Weaving the Silos Together

  

One of the major aspirations of Manufacturing 4.0 is to digitally unite a manufacturing company. The idea is that, using transformative digital technologies, all the key functions of a manufacturing company – design, production, sourcing, warehousing, service, support, etc. – can be integrated together using a seamless flow of data that provides one, accurate picture of operations.

The theory is that such a common stream, or thread, of data will create far greater efficiencies, speed of execution, and shared knowledge of operations than the current alternative – silos of disparate information, often in different and incompatible systems, that result in more mistakes, more time, more confusion, and more cost.

It is a compelling theory and one with obvious advantages. What company wouldn’t want a system that creates cross-functional integration and compatibility? What leadership team would not want what some call “one version of the truth” across the enterprise?

But the path to creating digital unification is anything but obvious or easy. Most manufacturing companies are collections of functionally-oriented, disparate systems that were put in place to serve a particular discipline or role. Moreover, experts in particular domains often don’t speak the same language or relate well to individuals in other domains; witness the now decades-old struggle to bring together IT and OT teams.

Many in the industry talk about tackling the problem by using phrases such as “breaking down the silos”, meaning companies need to tear down cultural borders and truly collaborate to realize the benefits of digitization. This is certainly necessary to achieve the benefits aspired to, but the choice of language may unnecessarily inject an obstacle on the path forward and stoke resentment and even fear in people working diligently in those so-called silos.

A better way is to first acknowledge that functional domains grew up and remain important for a reason: companies need the domain expertise to make whatever they are selling. Instead of talking about knocking down the silos, let’s describe what needs to be done as weaving the silos together. This way the proper respect is accorded to the domain and its experts, the benefits of integration are advanced, and language some might consider scary is avoided.       M

 

ML Journal

Democratizing a Data Driven Workforce

In an M4.0 world, democratization not only means making data available to more people, but giving them the tools and guides to use that data effectively. 

Since the 5th century BCE, democracies have flourished in some form throughout the world. Most people agree that giving people the right to govern themselves – with either the authority to select their leaders, or decide on legislation, or both – is a right and noble cause. As companies pursue their digital journeys, can the concept of democratization now also be applied to manufacturing businesses that have traditionally operated with a top-down leadership style?

Over the last few decades, and recently accelerated by the global pandemic, workers have taken an ever more active and influential role in organizations. Employees have become an important driver of corporate policy and success. Yet, it’s also become increasingly more difficult to find and retain great people. Workers today have many options as to where, and when, to work. That’s why empowering employees is now essential to improving the levels of job satisfaction that will ensure a company’s increasing profitability.

A meta-analysis published in the Harvard Business Review1 states that leaders who empower their employees are more likely to be trusted by their people compared to leaders who do not. This is not to say that empowering employees involves pushing work onto underlings that managers don’t feel like doing themselves. Leaders who empower their employees act as coaches, pushing their employees to do their best work and supporting them along the way. But just like the mechanic that is asked to work on a new vehicle without training, guides, or instructions on the vehicle’s new systems, asking employees to improve the business without first providing them the data to understand where changes need to be made, will lead to frustration, anger, and doubt.

For manufacturing businesses, democratization means not only making data available to more people but giving them the tools and guides to use the data effectively. When a frontline manager knows that unscheduled overtime has a negative impact on their department’s profitability, they may suggest ways to reduce it. When an employee is incentivized to improve productivity and can see where they and their team are with easy-to-understand metrics that update in real-time, making positive changes that are easily measurable and readily visible becomes gamified, making improving productivity both fun and lucrative.

Democratizing a data driven workforce can have many positive results. It allows companies to be much more effective at influencing employee creativity and improving corporate citizenship behavior to better align with the company’s vision and goals. It helps build a culture of trust that delivers additional benefits including better employee retention and improved employee / manager relationships. It enables managers to optimize operations by leveraging advanced people analytics to build nimble teams that are scheduled effectively to address product planning, order requirements, and equipment constraints. It also allows for more personalized training scenarios and learning plans that workers can build themselves in a Netflix-style experience to take ownership of their own career advancement, and so consumerize the employee experience (EX) to drive engagement for both established and younger workers.

According to McKinsey & Company in their report on “Building a more competitive US manufacturing sector,”2 a critical factor for advancing M4.0 is the need for manufacturers of all types to find, retain, and train more specialized manufacturing talent. Doing so, “can boost future output growth and competitiveness…and upgrading existing jobs to attract the next generation of workers, perhaps from other industrial sectors,” notes the report. This in turn can help US manufacturers, “boost the [manufacturing] sector’s annual GDP by more than 15% above baseline forecasts while adding up to 1.5 million jobs,” suggests McKinsey.

Democratization also means making powerful Human Capital Management (HCM) and Work from Home (WFH) technology available in the cloud to manufacturers of all sizes. This allows smaller manufacturers to take advantage of the same economies of scale that larger organizations have access to. While this may not help them gain access to capital markets to finance growth, smaller manufacturers can be better positioned to attract and retain top talent, gain better control of their labor cost, and remove potential barriers to growth typically experienced in the market, while making their operations nimbler at the same time. It also enables them to scale quickly to meet unique demand increases.3

One thing that hasn’t changed over the course of the pandemic is that the second largest expense, and the most valuable asset, for manufacturers is their people. To succeed in M4.0, manufacturers will need to invest in an employee experience that reflects a modern workforce, one that is empowered, engaged, and more in control of their destiny, and by extension, that of the company they work for. Today, data drives decision-making at companies across geography and sectors. For manufacturers, it’s time to put some of that power, and trust, in the hands of your people.     M

Footnotes:

1    When Empowering Employees Works, and When It Doesn’t; Harvard Business Review, March 2018:  https://hbr.org/2018/03/when-empowering-employees-works-and-when-it-doesnt

2    Building a more competitive US manufacturing sector; McKinsey & Company: https://www.mckinsey.com/featured-insights/americas/building-a-more-competitive-us-manufacturing-sector

3    DeRoyal Industries shifts labor and production to supply essential protective equipment; Ceridian: https://www.ceridian.com/resources/deroyal-industries-shifts-labor-and-production-to-supply-essential-protective-equipment

ML Journal

Streamlining Supply Chains

How advanced technologies can drive supply chain visibility and efficiency. 

This year has had no shortage of events that underscore the importance of supply chain visibility: shifting business impacts as COVID-19 case numbers have started to rise again, extreme weather events around the country and the world, and a return to a tight labor market. Manufacturing companies need to achieve situational awareness by gaining visibility and synchronization of their supply chains in order to navigate these fluctuations efficiently and with minimal disruption to customers.

Harnessing the power of enabling technologies is one of the most effective ways to improve visibility and drive efficiencies across the supply chain. Manufacturers might consider using AI-driven simulation tools and digital twins to better anticipate supply chain disruptions and help them understand when to adjust their plans accordingly. Because interconnectivity will drive such efforts, businesses should also look for opportunities to further integrate smart device capabilities across their organization.

Just as important to the implementation of these AI-driven tools is a leadership team committed to supporting their integration throughout the business on a holistic level.

Transformative Tools

For manufacturers in the early stages of using advanced technologies to streamline visibility into their supply chains, achieving better situational awareness of these supply networks may seem daunting. At a basic level though, the effort boils down to three main components:

  1. The what — What types of disruptions and events can your organization identify that will adversely affect your supply chain?
  2. The so what — What would the impact of those disruptive events be?
  3. The now what — What solutions can you implement to mitigate and/or resolve disruptions that do happen?

The skyrocketing cost of shipping that many companies have had to navigate this year is one example of the type of disruption that should compel manufacturers to prioritize improving their supply chain visibility. Several technologies can assist companies in achieving situational awareness of their supply chain. A complex event processing engine — a computing tool that leverages internal and external data streams with artificial intelligence — can help organizations understand how various geopolitical, climate and economic conditions affect operations around the world. Such processing engines can answer the first question by identifying events that have the potential to impact the supply chain.

If there is a hurricane churning in the Gulf of Mexico, for instance, an event processing engine can help a company that has manufacturing facilities in Alabama to be proactive about how the hurricane will eventually affect operations. Businesses can similarly use event processing engines to understand how surging COVID-19 cases in specific regions will affect operations in those locales.


“Harnessing the power of enabling technologies is one of the most effective ways to improve visibility and drive efficiencies across the supply chain.”

 

While complex event processing engines can help identify events that may turn into disruptions, a discrete event simulation tool will allow teams to understand the impact of these disruptions when they occur. By answering the second question, companies will gain insights around which customer orders might be delayed, the various types of inventory shortages they might face, and which production lines may go down. Such insights allow for companies to be more proactive in heading off any supply chain issues, which leads to more consistency and better service for customers.

After an organization has identified disruptions and understands their impact, the next step would be to mitigate and/or resolve those disruptions and answer the third question. To that effect, a discrete event simulation tool along with a digital twin — a digital replication of a company’s network on a computer or platform — can provide potential resolutions. The digital twin concept uses real-time updates and data integration to provide a holistic picture of the supply chain. In tandem with the simulation tool, this allows for timely analysis and scenario planning to understand how to mitigate disruptions and their downstream impacts.

These elements combined together provide the framework of an operational control tower, which uses real-time data to understand how current situations might affect the supply chain and how best to resolve those problems. All of this can dramatically improve a company’s situational awareness.

Connectivity Is Key

These tools and technologies use information from a variety of data feeds and align that data with a company’s supply chain network, including plants, suppliers, and warehouses. Because data is essential to these efforts, manufacturers need to have foundational information technology and operational systems and architecture in place to support the collection and analysis of huge amounts of data.

“Connectivity and communication across a manufacturer’s supply chain are essential to any efforts around improving supply chain visibility.”

 

Many major-name manufacturers are investing heavily in these advanced supply chain technologies and even creating microservices around them. For midmarket and smaller companies, though, investment and adoption vary. Manufacturers should assess whether they have the right IT and OT systems in place to support such efforts.

Connectivity and communication across a manufacturer’s supply chain are essential to any efforts around improving supply chain visibility. Such connectivity can come in the form of enterprise resource planning systems, the use of various Internet of Things-enabled devices, and making sure IT and OT systems are effectively integrated.

Beyond connecting systems to be able to use data harmoniously, manufacturers prioritize three key aspects of interconnectivity: the consumption of data, insight generation from that data, and making decisions empowered by that data. Those three elements are the backbone to any truly connected, integrated supply chain.

Organizational Support

The level of holistic integration required to build a more nimble supply chain doesn’t just apply to technological systems; it also extends to the people and teams that head up adoption efforts around advanced technologies.

Manufacturers need to have a center of excellence (COE) in place to support the adoption of artificial intelligence-driven technologies for supply chain transformations. Projects will likely fail if they are left in the hands of a couple of siloed leaders, but a company that stands up a COE comprised of subject matter experts across various functions will have a built-in support network to troubleshoot and navigate issues more effectively as they work on this integration.


“Manufacturers need to have a center of excellence in place to support the adoption of artificial intelligence-driven technologies for supply chain transformations.”

 

Leadership teams should create a smart technologies framework that also helps to drive user adoption across the organization, including training employees and getting them comfortable with using these new systems. Automation and artificial intelligence will present opportunities to move workers away from repetitive tasks toward more critical problem-solving roles, and will allow companies to reimagine what other roles look like beyond the shop floor as well.

Best Practices

Especially for midmarket and smaller manufacturers, it may be overwhelming to determine where to start on the journey of improving supply chain visibility. Leadership teams should assess the company’s current position and figure out where the biggest pain points are. This will help to prioritize areas of focus, from suppliers to shop floor operations to a range of other internal or external factors.

Here are some best practices that manufacturers should consider as they adopt advanced technologies to improve their supply chains:

  • When implementing new technologies, start small to prove out the benefits before taking anything to scale.
  • Identify the top supply chain challenges to help your team prioritize.
  • Understand which technologies will fit your business — just because another company is implementing something doesn’t mean it will benefit your company’s manufacturing, supply chain or warehousing operations the same way.
  • Assess how third parties can help you implement these technologies, how they can help with considerations around data integration and accelerate the time frame for adoption. This is especially important for companies that may not have robust IT departments in house.

Given the ever-evolving nature of artificial intelligence and related advanced technologies, manufacturers will need to regularly revisit and update their approaches to supply chain connectivity and visibility, but the areas outlined above are a foundational place to start.  M

ML Journal October 2021

POV: Transformative Technologies

One of the few positive things to emerge from the pandemic has been a greater sense of urgency on the part of manufacturers to embrace Manufacturing 4.0 technologies. Spurred by the need for greater flexibility, agility, and speed to deal with disruption to everything from the front office to supply chains, manufacturers have reacted by accelerating their adoption plans for a range of technologies, from remote working platforms to plant floor monitoring systems.

MLC’s new Transformative Technologies survey, results of which are published in this issue, reinforce and expand upon the trend to accelerate adoption, first reported by MLC in June of 2020. The new survey shows that 51% of respondents plan to accelerate their adoption of operational and information technologies as a direct result of the pandemic, up from 43.7% in last year’s survey.

But what has been perhaps an unexpected consequence of this greater sense of urgency on the part of manufacturers is their perception that the competition for digital capability, and, hence, competitive advantage, has heated up. One of the key findings of the new survey is that some manufacturers now feel they are falling behind competitively in the digital race. This year, only 5.5% felt that they were substantially ahead of their competitors with M4.0 technologies, compared with 11% last year. And 21.8% saw themselves as slightly behind competitors, up from 13.5% in 2020. Those feeling parity with competitors, at about one-third of the sample, was pretty much unchanged with last year.

Perhaps it is only natural that, when the stakes rise in any competition, those involved feel the pressure and question the speed of their progress. Despite best efforts, competitors could move faster and more effectively no matter what a particular company does. It is also possible that a company’s expectations for progress by a certain time were frustrated by internal difficulties in evaluating, selecting, deploying, and realizing value from M4.0 technologies, resulting in feelings that they had slipped competitively.

Whatever the actual reason or combination of factors is, however, the fact that companies are feeling the heat is a good thing. The heat will create more energy for the industry as a whole to build digital strength. And that will benefit all companies. – David R. Brousell

ML Journal

How Manufacturers Can Amplify Intelligence with AI

Manufacturers struggling to scale value from investments in digital technologies and AI need to take a more holistic approach that ensures C-suite buy-in, addresses new skills requirements, and creates an AI-friendly culture for the future 

For many manufacturing companies, the potential of artificial intelligence (AI) is easier to envision than the steps needed to integrate it into their business. Leaders can see the transformative power of the technology as it continues to evolve and they begin to imagine what it would mean for their business, from end-to-end supply chain visibility to powerful new insights from predictive analytics to the ability to respond to sudden demand shifts more quickly. It’s all out there, and yet, the strategy and execution needed to bring these tools to life continue to be elusive.

As companies struggle to scale the value from their investments in digital technology, many often find themselves in a state of pilot purgatory. They experiment with AI and machine learning, the Internet of Things (IoT), augmented reality, virtual reality, and other solutions that may have no clearly defined purpose. It’s a fundamental flaw that keeps many manufacturing companies from leveraging the benefits that new technology can bring to the table. They could have the best data scientists in the world and a whiteboard full of bold ideas that would forever reshape the company’s future. But if they don’t have an empowering business vision tied to clear business outcomes and a viable business case, it’s going to be very difficult to find success.

Leaders need the right mindset, the right skill set, and the right culture to effectively scale digital innovation in a company and create a functioning unit that can maximize the advantages that it brings to an organization. This effort requires time and careful planning, and there will, no doubt, be setbacks along the way. Do it right, however, and the payoff for the company and its customers can be substantial.

Opportunities and Hurdles

It’s easy to see why manufacturers are enthusiastic about AI. In factories, smart sensors, IoT, and AI enable predictive maintenance¹ to save costs and extend the lifespan of important assets. Then there are digital twins², which are virtual replicas of a product, process, or piece of equipment, which they can use in simulations to help make supply chains more resilient³. AI can play a role on the other side of the value chain as well by enabling chatbots to respond to inquiries quickly through text analysis. Cybersecurity intrusion identification is a popular response as well.

Other use cases are more nascent, but also powerful. For example, AI can help forecast customer demand and manage inventory for seamless fulfilment. Analytics4 can also drive better decision-making and more effective utilization of labor, and AI visual analytics can be used in maintenance for faster inspections and verifications.

“If companies don’t have an empowering business vision tied to clear business outcomes and a viable business case, it’s going to be very difficult to find success.”

 

One key is understanding that doing AI is not just a matter of implementing the technology. A focused approach on business outcomes first, followed by a robust data quality and governance process, are critical to drive business value at scale. Hurdles, such as how data must be collected and cleansed to be easily connected to AI solutions in production, must be addressed. Too often, siloed functions and unintegrated platforms don’t forge the links needed to make AI effective.

Then, there are manual processes to digitize and sensor-based data points to be linked. Governance is also crucial for company-wide implementation and utilization to bridge AI’s trust gaps6. Technology is just as much about humans as it is about computers and digitization. How people work, the tasks they fulfil, and the culture that enables that work will all take on new dimensions.

From Concept to Implementation

AI is not a single technology, but a set of methods and tools with subdomains applied to countless situations. One way to look at it is that technology can be implemented, bots can be built, but AI must be applied. Value from AI doesn’t come from putting it in, at least not yet. But AI is maturing and being embedded in enterprise systems, as well as becoming more accessible for nontechnical users.

It’s an evolving platform with plenty of room to keep growing. The lesson for manufacturers struggling to develop a plan for AI is to begin with the end goal in mind, but that end goal needs to be much broader and longer-term than companies typically think of in a business setting. It’s not always looking at where the company is now and what the next two steps it needs to take should be. It’s also about where the company wants to be 5, 10, or 20 years from now. What does that world look like?

It may not be easy to see, especially in a world where everything can change so suddenly. But companies still need to develop a road map, set priorities, and make a plan so they can become more agile and more responsive to changes in consumer wants and needs. They acknowledge that they may need to revisit this plan from time to time, and perhaps make big changes to it. But it’s still a plan, and it gives the company valuable direction to help it move to where it wants to go. In Manufacturing 4.0, companies need a different mindset about the pace of change and how they’re going to adapt to it.

“Corporate cultures that have become rigid and narrowly focused on the needs of today rather than the possibilities of the future must be challenged.”

 

One lesson that has been validated again and again across all industries, including manufacturing, is the idea that simply spending money is almost never the answer for what ails the business. There are companies that have spent millions of dollars on new equipment, new factories, and new software to expand their ability to serve customers. But when the leader of that company brings a product to the conference room and asks his or her team to explain how much it would cost to make more of it, or explain the path that product took from inception to its delivery to the customer, there is often no answer. Digital can solve so many problems in business, but only if there is a clear implementation strategy and a defined purpose to it.

As companies develop pilots and embark on proof-of-concept initiatives with different technologies, they need to keep the big picture in mind. It’s great that they’re doing a pilot on blockchain, or how AI could dramatically change how they use warehouse space, but how do these fit into what they’re trying to accomplish from a business outcomes perspective? Here’s what they’re doing, here’s how they’re doing it, and here’s the outcome they’re ultimately going to achieve. That’s the piece many companies struggle with because many still operate in silos, and they are never able to bring that full picture together.

By combining human and AI understanding and interpretation, they can find whole new ways of approaching challenges, developing ideas, and driving growth. Rather than having to begin with a blank canvas, they can use AI to direct their creative judgement and decision-making processes to suggest new routes for innovation that are more likely to be successful, based on the available data7.

EY recently worked with one manufacturer that saw an opportunity to make a bigger version of its signature product. The company had spent $100 million on process improvement and assembled a top-notch team of data scientists but struggled to develop a cohesive plan of attack. The company’s CFO had become very skeptical about the value of digitization through this effort. Every time the company tried to change its product mix, throughput dropped significantly, and the production schedule was disrupted.

“Without an engaged C-suite, it will be a struggle to have a dialogue about how best to use AI, how to allocate resources, and how to set priorities across all business units and functions.”

 

The problem was planning and execution. The CFO needed a baseline understanding of throughput across the company’s factories, including when new products were introduced. He also needed a way to simulate how the plan might work before any capital allocation decisions were made. The solution was to create a business case and pilot a digital twin capability that enabled the plan and led to significant savings for the company.

The AI Road Map

Shaping, accelerating, and optimizing an AI journey requires six steps, regardless of whether a company is just starting out or trying to strengthen its plan to make it holistic.

1. Acknowledge AI potential

Without an engaged C-suite, it will be a struggle to have a dialogue about how best to use AI, how to allocate resources, and how to set priorities across all business units and functions. It’s a good idea to pick company AI agents who know about the potential of the technology and will keep it on the agenda by helping to hone robust business cases, develop metrics for a proof of concept and then move any AI solutions into production. Without leadership from the top, AI initiatives can get lost in the shuffle amid other priorities and disruptions in the market.

2. Transform and plan

An agile and open culture is a baseline need for the business to be able to effectively leverage new technologies, not just AI. A plan should include key performance indicators aligned with the organization’s business strategy, and finance allocations should be clearly set. A data unit should be established, working in tandem with AI agents and a digital committee or center of excellence, to address requirements in the current state and support the journey to the future state around items such as data collection and cleansing.

3. Data foundation and structure

The data unit or owner is vital for asserting oversight across all of the data points across the supply chain, involving many customers and processes. Non-digital data must be converted, other data sources should be cleaned, and structure should be added to boost the quality of the data, and ultimately, its effectiveness in the AI solution. Data storage through databases such as data lakes help guide the data flow and strengthen the ability to perform analytics. Data governance, processing, explainability, and transparency are all components of a successful solution that should be addressed up front.

4. External partnership ecosystem

Manufacturing companies have developed many talented resources with varied skill sets, but AI know-how can be in short supply. Thankfully, a robust ecosystem of external parties, including start-ups, academia, consultancies, and other tech leaders, can be tapped, adding perspective to the company’s understanding of the business and use cases.

5. In-house AI expertise

Even with partners, the existing workforce will need to learn new skills and fulfil new responsibilities. AI experts, data scientists, and engineers are crucial personnel to hire, but an understanding of data science must be spread throughout the organization. Corporate cultures that have become rigid and narrowly focused on the needs of today rather than the possibilities of the future must be challenged, because AI works only when skills and experiences from many disciplines unite.

6. Architecture and infrastructure

Algorithms are a core part of AI solutions. In fact, these rarely pose a struggle for most organizations to build. The complexity arises when the time comes to integrate them with the technological architecture. Smaller modules with clear guidelines and principles make this process simpler for running proofs of concept and scaling the solutions. And standardized infrastructure service offerings on the market provide agile and robust ways to enable these AI solutions with flexibility.

With great challenges come even greater opportunities. Manufacturers that create an AI-friendly culture today are positioning themselves to boost customer and employee satisfaction as costs decline and helping drive their competitive edge in a challenging and complex moment for businesses across the world.   M

Footnotes:
1    How preventive maintenance can backfire and harm your assets, https://www.ey.com/en_gl/consulting/how-preventive-maintenance-can-backfire-and-harm-your-assets
2    Can a supply chain digital twin make you twice as agile? https://www.ey.com/en_gl/advanced-manufacturing/can-a-supply-chain-digital-twin-make-you-twice-as-agile
3    How to navigate supply chain disruption with digital process mining and digital twins, https://www.ey.com/en_gl/advanced-manufacturing/how-to-navigate-supply-chain-disruption-with-digital-process-mining-and-digital-twins
4    How forward-thinking organizations are becoming data-driven, https://www.ey.com/en_us/consulting/how-forward-thinking-organizations-are-becoming-data-driven
5    Why contactless field service presents an opportunity beyond COVID-19, https://www.ey.com/en_gl/consulting/why-contactless-field-service-presents-an-opportunity-beyond-covid-19
6    Bridging AI’s trust gaps: Aligning policymakers and companies, https://assets.ey.com/content/dam/ey-sites/ey-com/en_gl/topics/ai/ey-bridging-ais-trust-gaps-report.pdf
7    Is AI the start of the truly creative human? https://www.ey.com/en_us/ai/is-ai-the-start-of-the-truly-creative-human

The views expressed by the authors are not necessarily those of Ernst & Young LLP or other members of the global EY organization.

Data supporting the content of this article is derived from the Microsoft and Ernst & Young LLP report to explore how AI can transform manufacturing: https://www.ey.com/en_gl/advanced-manufacturing/how-manufacturers-can-amplify-intelligence-with-ai

 

 

 

ML Journal

Digital Twins: The Key to Unlocking Value and Innovation

By turning the physical into the computable, manufacturers can reach new heights in visibility, performance, and business value.
  

As decision-makers in manufacturing embrace digital transformation to improve business visibility, increase asset and product reliability, optimize operations, and unlock new business models and revenue streams, it is imperative to consider a digital twin as a key piece of this transformation. A connected, virtual replica of a physical product, asset, or system, digital twins turn the physical world into computable objects.

Digital twins live in the present but learn from the past to predict future states and create new opportunities and insights to perform rewind, replay of product or asset operations, and get answers to “what if” questions all in the digital world.

Digital twins provide a way to combine technologies, such as streaming sensor data with IoT, advanced data analytics, and machine learning to enable a knowledge graph representation of the world. Although most of the technologies already exist as stand-alone capabilities, digital twins provide a common framework to bring them together and more rapidly innovate.

The digital twin is based on massive, cumulative, real-time, real-world data measurements across an almost endless array of dimensions. These measurements can create an evolving profile of the object or process in the digital world that can provide important insights on system performance, leading to actions in the physical world such as a change in product design or manufacturing process.

Agility Across the Value Chain

During the design and development phase, digital twins allow optimal decisions for design alternatives and system properties. They also help to ensure the best usage and performance and to realize advanced service or maintenance solutions of an industrial asset or product. They even support end-of-life-management such as for airframe service time life or a gas turbine MTBF. This means a digital twin can accompany an industrial product over the whole lifecycle, seamlessly linking all stages of the value chain. All the data collected during its lifecycle can inform the design, production, operation, and use for the next generation of a product.

As products and systems increase in complexity, almost all industrial or technical products will come with such a digital twin, enabling detailed simulations and their management over the whole lifecycle.  They start as twins of a product type with the original design information, engineering models, and testing data. They continue after manufacturing as representations of specific instances of a product using the data collected during operation to support maintenance and service along with other activities. And finally, the digital twin helps facilitate end-of-life-usage and product obsolescence, which could help a circular economy.

IDC research outlines a Digital Twin Maturity Model1 (Figure 1) that begins with digital visualization and moves through digital development, digital enterprise, digital ecosystem, and finally digital twin orchestration. Digital twin technologies can have a positive impact on management and operations across the entire value chain, moving from ideation and visualization of processes to service and management at a greater scale. 

In manufacturing today, there is a focus on the operational phase of products and assets, allowing for digital twins to be adjusted when the real product gets modified. Another focus area is using simulations to predict the need for maintenance and service. This allows for better product service, product quality, time and cost savings, and unprecedented flexibility for operations. A digital twin as a synchronous image of a system in operation enables advanced solutions like extended monitoring, simulation of changes, and simplified control. Since production in nearly all industries is required to be flexible, these digital twins will also support real-time adjustments during operation.

For any industrial product to come to life it must pass through several stages that often are only loosely interconnected today in a digital sense: from data in the design and development phase to manufacturing, implementation, operation, end-of-life, and recycling. Digital twins in conjunction with digital threads have the potential to build connectivity within and between different business processes such as product lifecycle management, manufacturing planning and execution or supply chain management.

Key Considerations for Digital Twins

It is important to note that digital twins are not complete representations of a product – they are made up of various models and data for specific outcomes and apply suitable precision for the required solution. Another important distinction is the difference between a digital twin representing an asset type – the information of which may be in the cloud – and its instances, which are additionally characterized by operational data, its operational context, technical configurations, and environmental parameters.

Key to the digital twin is focus on the various information and data points needed at various stages across the lifecycle of the product. It is important to structure data in a reusable way, typically through a canonical data model, which is a common, enterprise-standard data structure that enables different systems and applications to connect and exchange enterprise information. A canonical structure can allow various systems to communicate in a simple agreed-upon framework. This can reduce the amount of data that must be stored outside of the systems of record, eliminate the need to manage large master data structures, and allow a company to use the digital twin in multiple ways with more flexibility, with continual updates as it is integrated further within the enterprise.

Arguably the most important enabler for digital twins are digital threads, which is a communication framework that links all elements of a product’s data, from design to obsolescence. Harnessing the enormous amount of data in the product lifecycle with digital threads reduces the complexity of implementing digital twins and increases their accuracy.  Manufacturers start by implementing digital threads that make sense for them, or incrementally drive specific outcomes without having to undertake massive investments to develop a holistic digital thread up front.

Benefitting from the insights that digital threads provide is only possible if data is accurately identified, classified, and correlated across the various sources it resides in. Relevant security and access controls must also be applied to data delivered across internal and external applications. A critical success factor for digital twin deployment is this transparent and automated information-identification processing, especially in distributed enterprises such as aerospace and defense organizations and highly regulated industries such as medical devices.

For digital twins to fully realize their potential it is important that they have an open format, allowing it to be easily updated, scaled, and extended when new models and data representing new outcomes is added. Communication modules ensure a secure and stable connection to the assets or other integration services. For modern open protocols like OPC UA, MTConnect, and MQTT there are already various modules available, whereas proprietary protocols require vendor-specific communication modules. Additionally, communication modules actively connected to the cloud are a vital feature of the edge level.

Additional modules handle other functions that are necessary for an integrated solution: buffering, filtering, distribution of the messages to further endpoints, or other edge modules (e.g., AI modules), and preprocessing (e.g., message enrichment). The edge layer can run complex logic modules and act as a recipient of trained or updated machine learning models, applying the new logic to incoming data and making determinations from it. Device management modules for monitoring, updating, or security activities on the device are essential to managing large-scale edge ecosystems.

Digital twins can have a positive impact across the entire value chain, moving from ideation and visualization of processes to service and management at a greater scale.

Standards facilitate an efficient use of digital twins, and this in turn enables the creation of economic value. To date, there are no genuine standards developed solely for digital twins. The reason for this is that digital twins are still created within the context of established standards in different domains, such as the automotive, industrial machinery, or aerospace industries; different disciplines such as mechanical, electrical engineering, or software; or with respect to specific uses or outcomes, be it testing, production, or maintenance

Leveraging cloud-based computing, storage, analytics, and AI/ML services which are secure, feature-rich, and highly scalable from anywhere in the world provide a unique advantage in building digital twins. In today’s globally integrated networks of factories and supply chains, these features enable OT and IT managers to build, deploy, and grow solutions quickly and at low costs. These efficiencies are coupled with the tradeoff of control or freedom. Careful architectural planning (e.g., defining a business domain model) and the application of cloud principles, such as containerization, enables manufacturers to realize the benefits while minimizing the tradeoffs.

As the digital twins evolve in such a manner, they morph into continuously newer evolutions of the digital twin. Cost is obviously a concern – digital twins will only be widespread if they have a positive benefit-cost-ratio. And while today a digital twin may be still expensive to build and maintain, the technical and economic advantages they enable will help with scaling so that digital twins will become more commonly incorporated in industrial business models.

Achieving Success through Increasing Value

True success in achieving early milestones on a digital twin journey rely on an ability to grow and sustain the digital twin initiative in a fashion that can demonstrate increasing value for the enterprise over time. To help ensure such an outcome, one may need to integrate digital technologies and the digital twin into the complete organizational structure — from R&D to sales and field service — continuously leveraging digital twin insights to change how the company conducts business, makes decisions, and creates new revenue streams.

Developing a digital twin strategy always begins with the intended business outcomes and impact on people, practices, technologies, and processes. Often many digital strategies fail to meet expectations because they are focused on proof-of-technology vs. proof-of-value. Once value is well articulated, the technology can be proven through use cases with measurable results.

As an example, if the goal of a digital twin strategy is to increase equipment availability in the field, the scope of the use case can span equipment health insights through to optimal scheduling of parts and field engineers. The KPIs can then take the form of total service availability by the number of maintenance incidents mitigated, elapsed time for replacement part availability, product enhancements with digital services through OTA updates, and reduction in time taken to deploy resources in the field.

With the large adoption of 5G wireless network, edge connectivity gets a major boost in terms of data transfer and bandwidth. The 5G core network controls the data flows and connections between the 5G device endpoint and the data network endpoint (the user plane functions or UPFs).  The 5G system is able to communicate the network’s status and events and obtain configurations from outside via a standardized API.

“Data management is an important function that needs to be operationalized and
included in regular processes.”

 

 

The approach to this use case can be the analysis of equipment telemetry such as onboard diagnostics or analytical processing of historical performance data but represented as a relative health measurement easily identifiable by a field engineer. This representation could be purely data driven and displayed through a dashboard or a holographic indicator shown within the context of the actual equipment as it operates.

The manifestation of value can also be the creation of a customer-facing digital twin service that manages availability of their equipment through automated insights and expert assistance from the manufacturer. As manufacturers bring in this kind of intelligence, they can drive significant levels of growth and efficiency across the organization by:

  • Personalized service. Real-time data shows how your customers are operating and maintaining their products. This has the potential to completely change customer experience and support to more personalized service vs. a one-size-fits-all approach. New value-added digital services and revenue streams also become possible. They can extend capabilities such as customer support and field service, as well as develop upgradable opportunities tailored for optimal performance of specific equipment.
  • Driving product quality and innovation. Digital twin simulations allow manufacturers to experiment with design iterations, make more informed design and engineering decisions, and enhance the overall product roadmap. This data allows for influence and validation of digital prototypes while also bringing more innovative products to market faster.
  • Preventing breakdowns before they occur. Digital twin simulations can be used to adjust product performance, safely postpone non-regulated maintenance events, or even eliminate problems before they occur. Service technicians can manipulate equipment digitally in real-time or visualize problems ahead of a job to optimize operation and perform preventative maintenance – solving problems before they come to pass.
  • Reinventing knowledge-sharing for employees. Digital twins can allow experts to explain complex scenarios visually, with 3D and 2D representations of real-world operational parameters and characteristics. These tools are essential for manufacturers to democratize skills and knowledge-sharing when facing an aging workforce and an approaching skills gap between new or less experienced employees. A digital twin strategy also enables employees in both professional and trade roles to build skills, creating new possibilities through advanced and contextual knowledge sharing.

Better Outcomes with Smarter Decisions

Implementing a digital twin strategy allows manufacturers to fix products and processes faster, which in turn means equipment stays working longer.  Manufacturers can develop higher quality products with a more efficient workforce. The key is the ability to use real-world operational data to make smart, informed decisions that provide the best resolution.

Going forward, digital twin strategy will be non-negotiable for most manufacturers. As the foundation for connected products and services, it is a critical path to enabling asset or process lifecycle management, collaboration, flexible vertical integration, and end-to-end reengineering. To be truly successful, companies must thread digital twin capabilities into the value chain as a strategic asset to enable new opportunities and provide market differentiation.  The digital twin strategy is more than just about technology. No longer is it an IT or engineering overhead for organizations; it is a powerful business strategy that can accelerate digital transformation.     M

Footnotes:
1 IDC, Digital Twins and Digital Thread for Engineering and Operations, Doc # US47453021, February 2021

ML Journal

Navigating Digital Transformation

It’s not enough to just automate. To succeed in today’s environment, manufacturers need to hyperautomate. 

Setting manufacturing right is hard. Manufacturers deal with a lot of complexity, variability, and uncertainty to design, procure, produce, ship, and sell the products that their customers already love or hopefully will like to use. Manufacturers struggle to do more with less due to increased customer pressure to deliver quality products faster and increased competition to get to the markets sooner.

Getting manufacturing right also takes time. Manufacturers are also constantly striving to improve their bottom line through efficient operations and cost management while also growing their top line through market share gains and new offerings and business models. The current COVID-19 pandemic has highlighted these struggles in a new light and showed us in stark contrast how the pioneers have fared well above the laggards. Success hinges on effective digital transformation.

Automation Is Not Optimization

Manufacturing is at the intersection of many disciplines, and manufacturers are not new to adopting technology to improve operational excellence and profitability. Incremental innovations get deployed everywhere throughout the production and business processes to cut down costs or cycle times, enhance product quality or performance, provide new product capabilities and better supply chain execution, and ultimately gain an edge over the competition.

These essential incremental innovations in products and processes are undertaken most often by the individual line of business (LOB) departments. Typically, each department operates independently or with a couple of other departments, trying to solve the problems at hand. Driven by ever-higher demand for increasing productivity and bound by time and limited resources, teams in most enterprises try to find solutions that solve their immediate needs and then move on. In other cases, most teams work with the silos of information they have access to, so the efforts and results are limited in scope. These incremental innovations provide gains, but even more significant improvements lurk in every corner of the supply chain operations.

Manufacturers are also increasingly deploying factory automation to improve quality, consistency, capacity utilization, and throughput. With automation, manufacturers can mass-produce quality goods faster and become competitive. A significant majority of manufacturers have also digitized their back-office operations by deploying ERP, SCM, PLM, and other enterprise software solutions.

“While building a fully connected enterprise is an advantage, it also brings about some disadvantages unless done right.”

 

Digitized enterprise operations along with automated factory operations have provided significant gains for manufacturers. Digitized and automated productivity solutions are most likely implemented on one or more process steps or value chain activities to alleviate a bottleneck situation, increase collaboration, or automate a routine procedure to make it more cost-effective, repeatable, and reliable. When one bottleneck gets addressed, others pop up. Excessive automation or optimization in one area of the value chain could cause issues in others and re-surface one or more of the three dreaded monsters of lean manufacturing: namely Muda (waste), Muri (overburden), and Mura (unevenness).

Automation and digitization are not new to manufacturers. They have been deploying these for the past two decades. On the one hand, production lines have benefited from automation and have grown much faster and complex. On the other hand, digitizing business processes and supply chains has enabled manufacturers to handle and build complex global supply chains. This operational complexity, supply, demand, geopolitical, economic, and lately pandemic-related uncertainties pose new challenges to manufacturers.

Industry 4.0 Enables System of Systems

By applying next-generation exponential technologies such as cloud computing, IoT, big data, AI/ML, and AR/VR, Industry 4.0 offers unprecedented digital transformation capabilities in leapfrogging operational excellence and creating new business opportunities. Industry 4.0, or manufacturing 4.0, is all about building cyber-physical systems. The promise of industry 4.0 is to break down silos, enable pervasive digitization, bridge the OT-IT gap, and build a scalable system of systems. IoT helps connect physical assets easier. Big data helps store and analyze large volumes of data generated by complex and high-speed operations. AI/ML helps with uncovering deeper insights from multi-dimensional and multi-variate data. Cloud computing offers a scalable computing infrastructure, and digital thread connects and brings all value chain activities together. Going beyond traditional automation, Industry 4.0 brings everything together.

The digitized and connected supply chain entities and activities allow manufacturers to build systems of systems where information flows end to end, and decisions can be taken quickly or sometimes automatically. The ability to digitally connect with products, processes, and most importantly, customers directly, makes new service-oriented and subscription-based business models possible. While building a fully connected enterprise is an advantage, it also brings about some disadvantages unless done right.

Manufacturers today manage complex supply chains and produce goods where many activities are sequential and interdependent on many other activities. Decision-makers must deal with a web of interlinked events and decisions, where their decision is based on something that happened upstream and will influence what happens downstream. Given that almost all manufacturing organizations are not entirely digitized, enterprise workflows often span multiple departments, myriad touchpoints comprising of manual, digitized, and automated activities. As a result, workflows have inherent inefficiencies and often span longer than expected timelines. Customer delivery timelines, product lifecycles, and business cycles are shrinking, forcing manufacturers to improve their efficiencies further.

Hyperautomation to the Rescue

Gartner coined the term hyperautomation in 2019 and listed it as one of the top strategic technology trends for 2021. In its simplistic terms, hyperautomation is a framework to automate the automation, allowing processes to complete faster and more efficiently, and be less error-prone. Hyperautomation is an approach that enables organizations to identify, vet, and automate existing processes that themselves could be automated. Tools such as process mining, robotic process automation (RPA), low-code application platforms (LCAP), and artificial intelligence (AI) are some of the technologies that enable hyperautomation. Gartner predicts that through 2024, the drive towards hyperautomation will lead organizations to adopt at least three out of the 20 process-agonistic types of software that enable hyperautomation. Tools such as RPA, LCAP, and AI are considered process-agnostic software, and can be used in any organization across multiple IT and business use cases.

“Hyperautomation is an approach that enables organizations to identify, vet, and automate existing processes that themselves could be automated.”

 

Hyperautomation builds on what Industry 4.0 enables and takes it to the next level. Industry 4.0 facilitates the creation of cyber-physical systems, where digitized entities connect to communicate freely and collectively to accomplish business objectives. The digitized and inter-connected machinery, factory automation, people, and business processes are critical for further automating some of the processes to increase efficiencies. For example, a hyperautomation script, upon receiving a predictive failure alert from a predictive maintenance model deployed, checks the automation cell machine’s inventory levels. If the part is not in inventory, the automation places a purchase requisition and schedules the maintenance with the part arrival date and machine failure timeline in consideration. As illustrated in this example, hyper-automation employs a systems approach to connect cross-functional activities, automates mundane tasks, and accelerates the processes allowing human experts to spend time on more important and higher-level tasks. Hyperautomation could also include automation tools, such as optical character recognition (OCR), natural language processing (NLP) for extracting text and understanding information in printed documents, or any AI/ML implementations to automate tasks.

Implementation of hyperautomation starts with the analysis of existing processes to identify tasks or processes that are bottlenecks or need automation to alleviate the pain of laborious and boring manual tasks. Various data, task, and process mining tools uncover operational characteristics and create a virtual map of the enterprise-wide processes. Process mining software analyzes all the data logs and transactional data stored in enterprise systems (ERP/SCM) to build a virtual map. Similar maps generated by other software tools or supplemented by experts help understand areas that need attention and the overall impact or value add through automation.

Once an automation scenario is identified, tools such as RPA, LCAP, and integration platform as a service (iPaaS) automate workflows and implement hyperautomation. AI augments and extends these tools to implement digitized processes and capture data in newer ways.

Hyperautomation offers many benefits beyond increased productivity, lower cost of operation, and speed to execution. It also allows manufacturers to implement multiple workflows to meet customers’ growing needs and come up with even more efficient workflows. Gartner expects that by 2024, organizations will lower operational costs by 30% by combining hyperautomation technologies with redesigned operational processes. Hyperautomation takes one step closer to fully lights-out and autonomous factories.

Back to the Future

Since year 2000, many Fortune 500 companies have gone bankrupt or ceased to exist due to increased competition from their digital native competitors. Today, the manufacturing industry is going through another inflection point because of Industry 4.0. At the same time, we are going through unprecedented times because of COVID. It is an understatement to say that the COVID pandemic has highlighted the importance of digitization. Yet again, in the face of dire challenges posed by the pandemic, digitized enterprises survived and fared well compared to the laggards. According to the Gartner Board of Directors Survey1, conducted in May and June 2020, 69% of directors say that the effects of the pandemic, economic, and social crises are accelerating digital business initiatives. Analysis of a recent McKinsey survey2 shows that 94% of respondents said that Industry 4.0 had helped them keep their operations running during the crisis, and 56% said these digital technologies had been critical to their crisis responses.


“Hyperautomation builds on what Industry 4.0 enables and takes it to the next level.”

 

Increasingly manufacturers are taking notice. The latest MLC survey research3 confirms that 54.8% of manufacturing companies believe COVID-19 has increased management’s focus on digital transformation. The MLC survey also highlights that those that have grasped the digital challenge, and an encouraging 26% say that they had now scaled their M4.0 efforts on a company-wide basis, more than double the result of two years ago when the figure was 12%. What’s more, a further 18% are now implementing M4.0 on a single project basis, again double the figure from 2019. Manufacturers across multiple sectors and sizes are at various stages of the digital transformation journey. It’s probably not too late for the laggards to embark on this journey. However, laggards need to make huge and concerted efforts to catch up and continue to be competitive.

Systematic Adoption

Digital transformation is a journey — it’s not easy, and it takes a long time to change. With this in view following approaches are suggested to embark on this journey successfully.

Corporate initiative — Digital transformation of an entire enterprise is a strategy that needs to be adopted and executed top-down. It’s an integrated execution as opposed to individual LOBs choosing what’s best for them. A dedicated steering committee or a governing body must oversee and ensure corporate-wide activities align with set strategic direction and goals.

Think system of systems — The end goal of an Industry 4.0 or Manufacturing 4.0 transformation is to create a digital enterprise that functions as one. It’s ultimately a fully integrated ecosystem of suppliers, people, machinery, partners, and customers. Enterprises and stakeholders need to view the implementation from system of systems point of view. Every entity in the ecosystem adds value; hence, interactions, integrations, and contributions need to be considered and measured.

Adopt a platform — Just as a strong house needs a solid foundation, the success of the digital transformation depends on the solutions used. Considering the complexity of the system of systems, adopting best-of-breed piecemeal solutions may not be ideal. Instead, consider a platform that supports Industry 4.0 capabilities, offers digital thread functionality, and provides standards-based open interfaces. It’s much easier when the foundation is built on a capable enterprise platform that offers maximum capability, extendability, and flexibility to realize the corporate vision.

Incremental approach — Digital transformation cannot be turned on at the flip of a switch. Furthermore, manufacturers cannot interrupt their production or other activities for longer periods of time. Organizations need to consider implementation pathways that are least disruptive and incrementally add value while delivering the desired results to run day to day business. Adopting a capable and scalable platform will make incremental adoption easier as you are able to start anywhere and expand gradually.

Skill-building — People are always an essential part of the digital transformation equation. While automation is seen as a job killer, it is in general, a productivity tool to improve working conditions and a competitive tool for the survival of the company. Enterprises must invest in employee education to be successful in this journey. Employees who understand the benefits of the technology and can use the technology once deployed become enthusiastic participants in the digital transformation journey.

“Gartner expects that by 2024, organizations will lower operational costs by 30% by combining hyperautomation technologies with redesigned operational processes.”

 

Given all the industry trends and the exciting productivity technologies and solutions available today, the laggards might still have a chance to embark on this digital transformation journey before it’s too late. The digital pioneers, at the same time, can build on their existing implementations to take their competencies to the next level. Manufacturers should not just settle for traditional automation. On the Industry 4.0 journey, enabled by exponential technologies, potential and possibilities are unlimited. Autonomous factories are not a distant future. Where are you playing?    M

ML Journal

Realizing Value from Artificial Intelligence and Machine Learning

Manufacturers see the potential for big payoffs, but attaining them requires a disciplined approach. 

Although the concepts of artificial intelligence and machine learning have been intriguing executives for decades, manufacturing has seen tangible impact only in the last 10 or so years. As with any emerging technology, there has been considerable excitement around the potential from these advanced capabilities.

But manufacturers have found the path leading to benefits paved with complexity and obstacles. That’s not to say the payoff isn’t worth it; current use cases prove there are benefits in the form of cost reduction, revenue growth, and enhanced profitability.

But delivering those benefits requires discipline. There are right approaches and principles for introducing AI and machine learning applications successfully into your operations with an eye toward maximizing their potential value.

Defining AI and Machine Learning

AI is a broad term that can encompass any simulation of human intelligence processes by machines. These processes include learning (the acquisition of information and rules for using the information), reasoning (using rules to reach approximate or definite conclusions), and inference or prediction. Applications of artificial intelligence perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.

Machine learning, a subset of artificial intelligence, revolves around the idea that machines can adapt and learn from experience, just as people do. Applications of machine learning perform a specific task without explicit instructions, instead relying on patterns and inference. Manufacturers can apply machine learning to execute certain activities humans could do — for example, going through large data sets to find patterns — albeit much more efficiently.

There are principles for introducing AI and machine learning successfully into your operations with an eye toward maximizing their potential value.

 

There are four basic types of machine learning:

Supervised learning: The most common form, this maps an input to an output based on examples of input-output pairs. This type of machine learning is often applied in areas such as predictive maintenance or quality.

Unsupervised learning: This is a type of self-organized learning that helps find previously unknown patterns in data sets without pre-existing labels. A common example is in customer segmentation, using revenue and customer data to determine where to focus effort in order to increase revenue.

Reinforcement learning: This gets into the area of cause and effect, or how software agents should take action in an environment to maximize some notion of reward.

Deep learning: This is a combination of machine learning methods based on artificial neural networks. Neural networks are specific algorithms that mimic behavior. A neural network has layers, each of which are part of the decision-making process and are designed to learn on their own. That learning can be supervised, semi-supervised, or unsupervised. An example is computer vision that can “see” microscopic defects at resolutions beyond human capabilities using an algorithm based on sample images. It then processes the information by sending an automated issue identification alert.

Applications and Benefits in Manufacturing Today

A cross industries, many companies are investing in AI, but few have moved initiatives past the production stage. In Gartner’s recent poll, 79% of respondents said their organizations were exploring or piloting AI projects, while only 21% said their AI initiatives were in production.

Another poll of 1,000 senior manufacturing executives by Google Cloud shows that the pandemic has accelerated the industry’s activity around artificial intelligence: 66% of manufacturers who use AI in their day-to-day operations describe their reliance on it as “increasing.” Primary AI use cases fall in the areas of quality control and supply chain optimization. Of those surveyed, 39% use AI for quality inspection and 35% for product and/or production line checks. With respect to supply chain optimization, 36% of respondents said they use AI for supply chain management, 36% for risk management, and 34% for inventory management.

Applications around protecting health and safety have the highest potential return in terms of EBITDA lift (Figure 1). Videos, sensors, drones, or similar capabilities can monitor the work environment and provide insight for avoiding or reducing injuries: for example, alerting forklift drivers when someone is in a blind spot, making sure employees use proper movements to avoid injury while performing tasks, or using sensors to detect a dangerous chemical buildup before it causes a catastrophe.

Another area that often offers high payback with (relative) ease is predictive maintenance. Machine learning applications can analyze sensor data to identify early signs of equipment failure, enabling preventative action to avoid costly unplanned downtime. For example, Nouryon Industrial Chemicals has piloted several types of AI/ML solutions to predict when to maintain and replace pumps and other equipment.

One application of machine learning that has seen acceleration in recent years is around advanced process controls. An operator uses instructions from a machine learning model to adjust the settings in order to optimize inputs for a specific outcome. For example, if sustainability is a goal, then the processes are programmed not only to make sure widgets are produced on time but also for minimal energy consumption. This is beneficial in cases where there are many variables, including human variables such as new or inexperienced operators.

A Practical Approach for Introducing AI / ML

The approach outlined here largely follows the CRISP-DM (cross-industry standard process for data mining) model, considered by many to be the gold standard for developing analytics models. It includes seven key elements, with particular focus on up-front discovery and delivery of tangible business outcomes through operationalizing the insight.

1. Business understanding

AI and machine learning answer questions, but they do not solve problems. For example, we can use these capabilities to predict when a piece of equipment is going to break and prescribe how best to fix it, but they won’t keep the equipment from requiring repair. AI/ML projects can’t move forward until there is a clear and common understanding of the question(s) to be answered.

It’s important to start with a deep examination of the potential impact, partnering with cost accountants and continuous improvement managers or engineers to understand base process data and look at equipment performance.

From this analysis, you can begin to identify potential use cases that generally align with one of two goals: increasing revenue or reducing costs. Develop a value hypothesis, a statement about a single, objective question the analytical model needs to answer and the expected result when successful. Success metrics to define what’s in it for the company.

Workshops with operations managers, production managers, and operators who have shop floor experience are valuable for understanding pain points and refining potential use cases.

 

This is the single most important exercise you can do—and it must happen up front.

2. Ideation
Once there is a clear understanding around the business impact, then you can start to bring a potential use case to life. It is critical to involve those who will be using the application or its output.

Workshops with operations managers, production managers, and operators who have shop floor experience are valuable for understanding pain points and refining potential use cases – for example, a machine that has a higher-than-expected rate of failure. If the operator had more up-to-date information, that could be used for planning. Predictive signaling could also help procurement obtain the necessary parts to minimize downtime.

3. Data understanding
Some of the biggest challenges center around data and the pipeline for accessing it, particularly with a mix of legacy and modern equipment. This can create huge variances in data quality. You’ll need to consider the maturity of current systems to understand input and output variables and what constitutes “good” and “bad” data. From an AI/ML perspective, key indicators of data quality may include timeliness, completeness, consistency, accuracy, and relevance.

You’ll also need to understand how to access, combine, and prepare data from multiple sources in order to apply machine learning. There are new tools, such as SORBOTICS, that can connect with a variety of factory machines, even older ones, to access data.


AI/ML projects can’t move forward until there is a clear and common understanding of the question(s) to be answered.

 

Finally, use of AI/ML also underscores the importance of optimizing the data historian to provide visibility to data at an asset level so that you understand process performance, rather than relying on manual observations.

4. Data preparation
Manufacturing data often has not been “cleaned” and prepared for AI/ML use. Therefore, you’ll need a process for filtering data that is inaccurate, incomplete, or irrelevant and then deleting or modifying it in order to produce data sets appropriate for analysis. Otherwise, you risk a garbage-in, garbage-out situation that not only diminishes the effectiveness but increases the risk of acting based on incorrect insight. Don’t underestimate the significance of this step nor the effort required to address it.

5. Data modeling
With the right “clean” data, you can begin using machine learning models to test the value hypothesis. This will require expertise in data science and analytical technologies in order to design and build the capabilities for modeling. Note that there are plenty of established solutions that can fast-track development using out-of-the-box capabilities.

6. Evaluation
To evaluate results, you will need clearly designed success measures and criteria for evaluating both the business impact and the technical analysis capability itself. From a business perspective, this means understanding the cost of a false positive and the savings from a true positive, using test data to determine if a model is saving money or costing money.

From a technical perspective, evaluation should look at how to continue improving the model’s results. It’s important to approach this process as an iterative one, where you are continuously evaluating the model and getting better results as you train the algorithm.

7. Operationalization
Insights from AI and machine learning won’t create any value if they’re left sitting on the shelf. You’ll need a well-designed business consumption mechanism that defines how the business receives and uses the insight, as well as capabilities for measuring and reporting value. One of the easiest and most effective ways to promote business consumption is to integrate the insight into a system in the form of a record of action to be taken by a system user or operator.

Involving people with domain knowledge early on when defining problems and developing use cases reduces surprise later and prepares them to embrace the solution going forward.

 

It is also impossible to overemphasize the importance of change management. Involving people with domain knowledge early on when defining problems and developing use cases reduces the element of surprise later and prepares them to embrace and own the solution going forward.

It’s also important to communicate value and expectations. For example, the solution will produce information that helps operators be more effective in meeting their key performance indicators, but it won’t replace their job. Be open in acknowledging these fears as you would do when introducing any new automation solution.

Moving Ahead with Purpose

There has always been hype surrounding emerging technologies, but it can take time for expectations and reality to meet. AI and machine learning are driving tangible value for manufacturers today, but you need to get past the hype and focus on introducing it in a practical way. This requires a disciplined approach that starts with value, employs domain expertise throughout, and has the right mechanisms in place to operationalize the insight gained.  M

Footnotes:
1. “How AI Builds a Better Manufacturing Process,” Forbes, July 17, 2018
2. Press Release, Gartner, October 1, 2020
3. “New research reveals what’s needed for AI acceleration in manufacturing,” Google Cloud, Dominik Wee, June 9, 2021
4. “Machine Learning Glows Brighter,” Chemical Processing, Séan Ottewell, August 5, 2020

Business Operations

Nexteer Displays Advanced Manufacturing in Action

Get the Latest News

Sign up here

Would you like to see the latest advanced technologies exhibited and explained for your benefit, all without leaving your office? The NAM’s Manufacturing Leadership Council’s virtual plant tours provide just such an opportunity, taking you inside cutting-edge processes and complex systems at manufacturing facilities across the country. Most recently, the MLC dropped in on Nexteer Automotive, where tour participants got to see its innovative Digital Trace Manufacturing™ (DTM) System in action.

Who they are: Nexteer specializes in electric and hydraulic power steering systems, steering columns and driveline systems, as well as advanced driver assistance systems and automated driving-enabling technologies. The company serves more than 60 customers around the world, including BMW, Ford, GM, Toyota and Volkswagen.

What is DTM? Nexteer’s DTM System connects and standardizes the company’s entire operations—including thousands of data-production components in 27 manufacturing plants around the world. To showcase the system’s capabilities, Nexteer took tour participants inside its Saginaw, Michigan, site, which includes six manufacturing plants comprising 3.1 million square feet of manufacturing floor space.

Tour highlights: Participants learned about the complexities of running a large-scale automotive component manufacturing plant, as well as how Nexteer uses the DTM System to maximize efficiency.

  • Nexteer team members explained how they design and program machines for data processing, showing how they determine where data will be sent and how they use barcode scanners and other methods to track components’ serial numbers.
  • Participants also got a virtual walk-through of Nexteer’s tracking system, which follows material from receiving and shipping through the production line with single-box precision. They also learned how Nexteer uses its Center of Analysis to correct any issues that arise.

Why it matters: It’s one thing to have a large system collecting data, and it’s another to be able to use that data effectively. The Nexteer virtual plant tour provided participants with practical takeaways, which will help them adopt similar innovations at their own facilities—for the benefit of employees, customers and shareholders alike.

Coming soon: Don’t miss the MLC’s upcoming tour of Johnson & Johnson’s facility on Wednesday, Dec. 1, from 11:00 a.m. to 1:00 p.m. EST. You will see how Johnson & Johnson uses mobility tools, advanced robotics and material handling and adaptive process controls to improve its operations. After the tour, stay for the panel discussion on how to scale advanced manufacturing technologies to create a sustainable, reliable and adaptable product supply. Sign up here.

Blogs

10 Reasons Why Semiconductor Firms Should Rapidly Embrace Cloud

Semiconductor Market Trends

 

The semiconductor industry has been growing at a rapid pace for the past few years. Market research firm IDC estimates that  the semiconductor industry grew at a rate of 10.8% in 2020 and will grow at 12.5% this year, resulting in a $522 billion market sector. IDC attributes much of this growth to the impact of COVID-19.

The increased demand for semiconductor chips is due to new generations of smartphones, tabs, laptops, and desktop computers used in industries such as healthcare for telehealth services; in the education sector for online teaching and instruction; and as more people worked remotely. At the same time, the automotive industry, a heavy user of semiconductors,  is packing more and more chips into vehicles as it attempts to offer all the creature comforts  consumers want as they embrace the connected car experience.

In the manufacturing sector, too, the pandemic has driven home the values and virtues of setting up connected factories that enable contactless manufacturing and uninterrupted operations in the face of a crisis. All these trends indicate  that the demand  for semiconductor chips will rise steadily in the  future. Despite the rosy growth projections, the semiconductor industry still faces challenges, chief among which is continuing to innovate even as it delivers expected price/performance improvements.

Therefore, it is imperative that the industry invest more in research and development to drive innovation while at the same time optimizing costs by leveraging technology such as cloud computing.

If one examines the key attributes and requirements of the semiconductor industry – skilled resources, high competition, complex automation tools, data and IP, differences in industry supply chains, and the brief shelf-life of  designed chips,  it is apparent that these factors are highly expensive and difficult to manage. Given the level of  investments and expertise required, there are very few players in this industry. The race for excellence is fierce, and a considerable effort and investment is dedicated to driving R&D to identify areas and avenues for innovation.

Faster time to market through the acceleration of design cycles, performance enhancements of chips through upgrades and updates, and IP protection through foolproof and flawless security systems are the top three business priorities of this industry. The chip companies invest most of their time, energy, and capital in fulfilling these priorities. However, operational priorities are equally important, such as driving efficiencies in the manufacturing process through data analytics; optimizing  operations, processes, and costs; and driving productivity through collaboration.

Cloud computing provides a reliable and seamless infrastructure to address both the business and operational priorities of the semiconductor industry.

Reasons to Embrace Cloud

1. Faster Time to Market and Quicker Design and Development

The ever-increasing demand from consumers for products with higher compute powers and processing abilities has resulted in shorter product lifecycles, requiring semiconductor manufacturing companies to bring products to market faster.

To this end, applying cloud computing in the semiconductor industry offers scalable storage, big data analytics capabilities, and enhanced productivity with collaboration tools for reviews and feedback that enable quick product launches.

Cloud also provides a flexible, scalable, elastic, and secure infrastructure for chip designing by providing on-demand compute for EDA tools. It enables semiconductor manufacturers to set up and access high-performance computing (HPC) power with virtual machines (VM) images, enabling quicker design and development cycles.

2. Improvement in Foundry Operations and Yield

Cloud offers a data lake or repository that enables storing, processing, analyzing, and inferring the foundry’s generated data. Manufacturers can use data insights for predictive performance and analytics as well as the management of resources in their supply chains, thereby improving  production uptime and yield. It also allows for specific artificial intelligence and machine leaning use cases for fault detection in the production line using imaging techniques and smart analytics tools.

3. Smarter Manufacturing Powered by Democratization of Data and  Analytics

Chip designs evolve with each release, and the chip design companies have families of chips in incremental progression/evolution cycles. The chip lifecycle data must be logged, analyzed, and processed for value generation. Cloud Service Providers (CSPs) like Amazon Web Services offer storage and analytics capabilities to chip design companies to apply AI and ML models for systematic data processing. They also provide the necessary infrastructure to integrate IoT and implement Industry 4.0 solutions for smart and connected manufacturing .

4. Improved Collaboration, Transparency, and User Productivity

The semiconductor manufacturing industry is highly competitive, and the success or failure of a chip manufacturer entirely depends on the ability of the manufacturer to collaborate effectively with an  eco-system that includes suppliers, OEMs, and internal teams for design reviews, feedback, and testing.  Cloud infrastructure provides a centralized system to track the productivity of the different stakeholders, enabling transparency and boosting efficiency, especially in the current times of COVID 19 using collaboration tools such as MS Teams, Google Workspace, and Google Meet.

5. High Operational Efficiency  

Unlike on-premise data centers managed by internal IT teams with constraints on skill, availability, and resources,  cloud infrastructure is managed by specialists such as GCP, AWS, and Microsoft. These service providers have made huge investments in R&D, infrastructure, and resources, and provide service-level agreements which ensure uninterrupted operations for semiconductor foundries.

6. Higher Service Levels Due to Better Availability

One of the primary reasons for the semiconductor industry to not adopt or scale cloud has been the business criticality of its operations. However,  modern-day CSPs provide SLAs that comply with industry requirements and, in some cases,  go beyond to ensure reliability. For example, GCP provides a robust architecture with high-bandwidth connectivity across 25 regions and 76 availability zones to deliver global services.

7. Organizational Agility and Flexibility to Scale-up 

The use cases for sensors, chips, computing, IoT, and Industry 4.0 are ever-increasing. It is thus imperative for the semiconductor industry to be extremely agile and offer unmatched on-demand scalability and flexibility to ramp up/down its compute infrastructure to accommodate R&D, design, testing, and validation of GTM activities. Analytical capabilities to draw insights and make quick decisions must also be in place in order for the industry  to deliver on its reputation of being agile. Cloud offers all these capabilities to the industry and at the same time drives home the cost benefits, security, and efficiency.

8. Backbone for Driving Innovation

There are several aspects of cloud infrastructure that can drive innovation for the semiconductor industry. To begin with, it can provide a leeway for the industry to squeeze in cost efficiency to a perceived rigid cost structure. The possibilities of leveraging IoT, AI,  ML, big data analytics for gaining visibility, and driving efficiencies throughout the chip manufacturing value chain are tremendous. It can provide EDA support, high-performance design, HPC, and High Volume Manufacturing (HVM) capabilities that will enable  better outcomes at lower costs.

9. Cost Efficiencies

Cloud offers instant scale and capabilities to perform and execute operations across the semiconductor value chain from design to yield without investing in physical on-premise data centers, reducing infrastructure development costs. It provides a collaborative infrastructure for value chain stakeholders to review and test the designs and offer feedback irrespective of the location of the stakeholders. Chip manufacturers can also drive the cost efficiencies on account of improved uptime owing to predictive maintenance capabilities and the security that cloud infrastructure offers.

10. More Secure Environment for IP Protection

 

The semiconductor industry powers a host of other industries, and several of these industries manage data categorized as highly sensitive, IP, business-critical, or compliance-driven. The dedicated investments by the CSPs in ensuring the security of their cloud infrastructure is an added advantage for the semiconductor industry to ensure data and IP protection for its clients. These CSPs provide more secure and reliable infrastructure at lower costs than the on-premise setup. For example, Google’s global-scale infrastructure protects billions of users with world-class security.

In Conclusion

The semiconductor industry has been a pioneer in enabling digitalization across industries. With Industry 4.0 and IoT gaining prominence, the use-cases of semiconductor chips have evolved rapidly from device-specific applications to sensorization, integration, and communication areas.

However, the irony of this industry is that despite being the transformation catalyst for all the other sectors to adopt digitalization, the industry on its own has been lagging when it comes to the adoption of technologies such as cloud computing for cost optimization, innovation, and streamlining operations. According to KPMG, even when most other technology industries have been adopting digital transformation at a rapid pace of 89%, the adoption rate of the semiconductor industry remains at a paltry 50%.

Considering the outlook for the semiconductor industry, utilizing the cloud for digital transformation is the only way the  industry can scale and position itself to meet  consumer demands for speed, accountability, security, innovation, and reliability.

 

View More