Top 5 Smart Factory Technologies
In today's article, I’ll be discussing the Smart Factory and the top technologies that support a new way of doing business in manufacturing. I'd love to hear if you agree or disagree with the list.
Also, join us for our webinar this Wednesday (April 20th) on Streamlining the Performance of Design of Experiments within Manufacturing with Smart Factory Solutions .
Introduction
To get started, I’ll take a quick look at a couple of the primary definitions of “Industry 4.0” and “Smart Manufacturing” from the World Economic Forum and from the Boston Consulting Group. Then I’ll talk about the focus for Visual Decisions and what areas we’ll cover today.
Industry 4.0 Technologies – World Economic Forum
Let’s start with the World Economic Forum definition of industry 4.0 technologies since they were the group that initially came up with that term.
The very small print in the diagram is their definition of the different technologies that fall under the Industry 4.0 banner. The shading represents whether the technologies are already in the mainstream, are currently maturing, or if they are just now emerging.
They then put each of the technologies into the following categories:
When it comes to the “Smart Factory”, we at Visual Decisions focus on those last three categories. As an aside, we also have a focus on the Production Philosophies and how to enable or blend those with the Smart Factory Solutions.
Nine Pillars of the Factory of the Future: BCG
In 2015, the Boston Consulting Group put out their study on the Factory of the Future and the nine pillars or technologies that would serve as key enablers.
This article shaped a lot of the following discussion of Industry 4.0 and the Smart Factory here in the USA. Here is a list of the nine technologies they listed:
For our purposes in this article, we will be focusing on the computer systems or software technologies. The Advanced Robotics and Additive Manufacturing are considered out of scope for our consideration.
The Top Five
All right, with that, let's get to the top five.
Industrial Internet of Things (I-IoT)
What is it?
I-IoT is a set of systems and platforms that connect most or all types of devices from the shop floor with the purpose to collect information from those, persist that data, make it available to other systems, and enable various workflows. Connection options for the shop floor include sensors, PLCs, and other devices.
There are several types of systems in this category, or that overlap functionality with I-IoT systems. Some of these systems include true I-IoT platforms such as PTC Thingworx; but they also include OPC servers, SCADA systems, and automation platforms. They all perform roles similar to or overlapping with the I-IoT systems.
How Prevalent is it?
The true I-IoT platforms are still in the early stages in the marketplace and starting to ramp up towards mainstream adoption. Other systems listed above with overlapping capabilities have extremely widespread adoption within manufacturing.
What Does it Do for Manufacturing?
The most common use cases for manufacturing involve asset monitoring and alerting, quality control, OEE improvements, and visual controls for the shop floor. When looking at the totality of what can be provided by I-IoT systems, there are truly too many use cases to list. I have a webinar coming in a few weeks where I’ll cover that topic in more depth when I look at the top ten IoT use cases. Given that I-IoT systems are built to connect to any device on the shop floor and provide a development architecture to create workflows around that data, the only true limit is imagination.
One last note for this section is that these systems can act as a central hub for manufacturing data and systems. I’ll talk more about integration later on in this list, but I-IoT systems can play a key role in coordinating and orchestrating shop floor information.
Why is it Valuable?
Why are I-IoT systems so valuable? Because of the breadth of use cases that I-IoT supports, it can help across many different improvement vectors. I’ve gone into great detail in other webinars and articles, but these systems can directly and indirectly improve productivity, cost and quality.
Big Data
What is it?
Big Data is a new set of data storage technologies that support different use cases than traditional relational databases. This is often differentiated by the five V's of big data:
In addition to the traditional five V’s, I have seen up to 10 V's in different definitions with variability, venue, vocabulary, vagueness, and validity also included.
Some of the different types of systems in this category are Hadoop systems, such as the one from Cloudera, and NoSQL databases like Cassandra or MongoDB.
How Prevalent is it?
These systems have become very common in many industries and use cases. Within manufacturing the data stores listed above are becoming increasingly frequent with the rise of new data sources such as I-IoT that can need to stream large amounts of data at very high speeds.
In addition, some of the traditional data historians such as OSI Pi could really be considered big data stores. Those systems are also designed to handle massive volumes of data being streamed at high velocity. So, the purposes of those traditional historians certainly align in purpose if not in design with modern big data systems.
What Does it Do for Manufacturing?
Manufacturing shop floors generate massive amounts of data every day. Most of this data is never collected from the sources. Most of the data that is collected in the PLCs and other temporary data stores is discarded at the end of the shift when the PLC drops that information to reset its registers. Then most of the data that does get saved is never used within any other system or process workflow. Given the cost and overhead of traditional data storage, this lack of use leads to the decisions not to collect or persist the information in the first place.
Big data systems are built in a way to mitigate the traditional issues with data storage. The first factor is that they make it easier and cheaper to store data from new data sources. Since they do not require the same structural definition as traditional relational databases, it is easier for operations to direct new data streams to the database. Since they are also built to ingest massive amounts of high speed data, there is also less need to involve IT database managers to administer the systems and optimize the system performance. So it becomes very simple to make the decision to save new information “just in case” it will be needed later.
Why is it Valuable?
Many of the use cases that are talked about in other sections would not be possible without the ability to store this large quantity of data from the process.
For example, having detailed process information enables things like correlations of the manufacturing process data to warranty data for analysis of what happens in manufacturing that leads to downstream warranty issues. Predictive maintenance and predictive quality use cases also depend on having access to that detailed process information to identify leading indicators for downtime or quality issues.
Analytics
What is it?
The first two technologies are about getting information from the process through I-IoT and then being able to store it with Big Data systems.
Using that data leads us to the third technology, which is Analytics.
“Analytics” covers a lot of different ground. Some of the different types of analytics (not nearly an exhaustive list) are:
AI/ML systems use various forms of learning systems or machine intelligence to augment or automate analysis. A very common application of those systems within manufacturing is predictive analytics.
领英推荐
Data Mining analytics are solutions that perform various types of automated analysis of datasets to find structures, anomalies, and other elements within those data sets.
For numerical data (and other data), statistical analysis can be readily applied to learn key summary information about the data the key factors about the data, identify trending information and much more.
They say a picture is worth a thousand words. Good visualizations can help people understand complex data at a glance. As the founder of the company called Visual Decisions, I am obviously a big fan of data visualization!
There are many types of systems in this category. In fact, there are too many to be anywhere near complete in the space here. General purpose machine learning solutions, data mining solutions, business intelligence systems, predictive analytics, and predictive maintenance solutions all fall within this category.
How Prevalent is it?
These systems are everywhere. I guarantee you have at least one of these types of systems in your company. Given that this category includes popular system such as Tableau, PowerBI, and MiniTab, just about everyone has software capable of performing some type of data analysis.
What Does it Do for Manufacturing?
These systems provide the predictive part of predictive maintenance and quality that was mentioned earlier. They can automate or augment process analysis. Many visual controls systems are based on some of the data visualizations made possible by software in this category. And much more!
Why is it Valuable?
These systems are valuable within manufacturing because they lead to better machine uptime with the predictive maintenance solutions, better quality, higher throughput, identifying deviations from standards and much more.
Augmented and Virtual Reality
What is it?
We have been building up capabilities. The I-IoT solutions help us acquire the data. The Big Data stores allow that information to be persisted in volume. The Analytics allow it to be analyzed and gain various insights from the data.
Augmented and Virtual Reality (AR and VR) now helps to display those insights to the people on the shop floor. AR systems retain visibility and interaction with the real-world environment, where the objects that reside in the real-world are enhanced with additional information as shown in the picture above. In manufacturing, this is mainly accomplished on devices such as phones and tablets or on glasses such as Microsoft HoloLens or similar products.
Full VR is where the user is completely immersed in the scene with no remaining view of the real-world. This type of interface also has applications within manufacturing, primarily for things like training or remote operation of equipment.
How Prevalent is it?
Augmented Reality systems are rapidly becoming more common. Use cases around remote support with AR really exploded during the pandemic. Use cases for training are extremely common as this allows for people to be trained without taking experts out of the process and sometimes without bringing key pieces of equipment out of production for the training. This can also help significantly reduce scrap being produced while new operators are being trained on a process.
Other use cases are now gaining acceptance, as well. For example, vendors for complex equipment are beginning to ship their equipment manuals augmented reality format, along with glasses with which to view those instructions.
What Does it Do for Manufacturing?
AR has the potential to change how information is distributed within manufacturing. Training can change from something that happens in a classroom or with an experienced expert by your side, to having an interactive guide providing you feedback as you perform anything new within the plant. Standard work can be embedded within the glasses so that people can be guided on the standards, and exceptions to those standards can be immediately detected and addressed to reinforce desired behaviors. Visual Controls within the factory can be revolutionized by the capabilities of AR.
Why is it Valuable?
Proper deployment of AR will reduce the cost of training while making it much more effective. Time to productivity for new workers will be significantly reduced. Costs to cross-train employees can be decreased to enhance the flexibility or agility of the operation. There should also be improvements to productivity from a better trained and augmented work force. Quality rates should improve from that same augmentation plus enhanced error-proofing, inspections and much more.
Horizontal and Vertical System Integration
What is it?
The last of today’s Top Five Smart Factory Technologies is Horizontal and Vertical System Integration and the possibilities are enormous. This is a graphic that I put together for last week’s article on “Connecting Everything ” with IoT. The gist of the article was that once most or all of the assets on the shop floor are connected and collecting data, there are many different opportunities that open up to connect that data to additional systems to create new or enhanced workflows within the plant.
In this context, horizontal integration refers to bringing all the “always connected” machines together into a system where there is always access to all of them. With this connection established, their real-time status can be communicated, and they can respond autonomously to dynamic production requirements. This can be achieved in single systems such as I-IoT or through multiple systems with high levels of data transparency and automated collaboration between those systems.
Vertical Integration refers to connecting that shop floor information to different departments and systems within the organization, such as engineering, product management, IT, finance, field service, etc. Once again, this requires very high levels of data transparency and integration between different systems to facilitate those new workflows across groups.
How Prevalent is it?
This type of interoperability between systems is still a challenge. To achieve real transparency and coordination between multiple systems is something that the software industry has been working on for decades.
Many efforts have been made to make it easier for data to be shared between programs. The move to object-oriented programming with encapsulation was meant to allow objects to communicate without exposing the underlying data structures. REST API are meant to serve a similar purpose in web-based development. There have been many other attempts over the years to make this process easier. As a result, we are still progressing down the path to true transparency and interoperability between different systems. Point to point integration remains the most common method of sharing information across different systems.
What Does it Do for Manufacturing?
As I mentioned up front with the I-IoT platforms, they can act as a hub for sharing information and data for the shop floor. When these platforms are connected to other systems in the enterprise, there are many different capabilities that become possible. Looking at the graphic at the top of this section, I covered these in much more detail last week. I will just touch on a few items here.
For maintenance, connecting the information from the machines to the maintenance management system (MMS) enables a lot of additional capabilities within manufacturing.
One example is to present autonomous maintenance checklist to the operator each shift (or whenever the work needs to be done). This works best when there's a common process across machines and departments, no matter who's working at the machine that day. By including work instructions for the tasks, it can greatly simplify job rotations or having people step in for absent workers.
The I-IoT system and the MMS can also be combined to automate the downtime repair ticketing system. This can help minimize the mean time to repair downtime by immediately performing triage and sharing the information about the failure with the maintenance system to create the work ticket. This also means that the ticket can be issued with information about the nature of the failure and the right people can be assigned and get out to the machine as rapidly as possible.
For quality, connecting to the quality management systems (QMS) enables many additional workflows or enhances existing ones. For example, if you're researching a quality issue in the plant, all the data becomes quality data. This includes data about the machine downtime issues, production information, testing information, and much more. When somebody is researching a quality issue, they should be able to look at a single interface to analyze information across all of those different sources.
For planning and scheduling, combining information from the shop floor with the planning and scheduling systems creates the possibility of closed loop real-time planning systems that are completely adaptive to the changing conditions on the shop floor. These types of planning systems are not possible without having the complete shop floor wired into the smart factory solution.
For engineering, combining smart factory systems with product life cycle systems and other engineering systems allows for significantly enhanced design for manufacturing (DFM) or design for six sigma (DFSS) processes. One small example of this is the automation of product FMEA data collection from the shop floor and product telemetry information.
Why is it Valuable?
Implementing those advanced workflows between the shop floor and other systems can impact just about every area of manufacturing performance. These types of workflows can also break down existing barriers between manufacturing and other departments within the organization.
Technologies That Just Missed
Here are some of the technologies that did not make the top five. They are listed here in this section because they either just missed inclusion, or because they are very commonly discussed.
Blockchain has tremendous value in particular applications, but it hasn't really found use cases on the shop floor that are as compelling as the technologies in the top five.
Cloud Computing has huge penetration and significant mindshare. But it's not really bringing any new capabilities to manufacturing. It's mainly a revolution for the IT department, how much it costs to maintain these different systems, and how long they take to deploy.
Cybersecurity is an absolutely necessary component of any of these different initiatives, but it does not help produce more product at a better quality rate.
Simulation plays an important role in various efforts such as testing shop floor improvements prior to capital investments. But I do not feel simulation provides quite as much value as the technologies in the top five.
Digital Twins – my take on this “technology” may be a bit controversial. I believe that “digital twin” is more of a marketing label than anything describing an actual technology. When everything from an OEE system to an ERP system can call itself a digital twin, I do not believe it is a well-defined technology.
Conclusion
So what do you think? Agree with the top five? Disagree?
Do you think Blockchain has compelling use cases in manufacturing that would put it above one of the top five?
Do you think Digital Twin is more than a marketing label?
Let me know in the comments - Thanks!
Great article Tim, thanks for sharing. Agreed that you have identified the top 5, but what's missing for me is the 'next level down' for data management - in other words first, the need for real-time transparency and response - second, the role that a data lake architecture can play in rationalizing and managing data from all sources - and third building a data management strategy for the long term. My goal is not to criticize your top notch research (clearly a step ahead of anything else I have seen), but rather to suggest an area for further exploration and discussion. ?? .
Technology Sales Director | Problem Solver | Customer Collaborator | McCombs BBA & MBA
2 年This is thought provoking stuff, Tim. Would you say Simulation is more important at the "Product" level, as part of the product design process, if that makes sense?
Director of Manufacturing at Sentinel Connector Systems
2 年I believe that you are underestimating the potential of Digital Twins, but I do agree that there is much confusion in industry and the media today as to the definition of a Digital Twin. If we look at Digital Twins as not a single technology but a interweaving of Digital Manufacturing technologies (IIoT, Data Analytics, Data Visualization, AR/VR, Simulation, etc), the power comes from making data Contextual, Consumable and Actionable.