Industry 4.0

Towards an autonomous and intelligent “Smart Factory”

Photo by Samuel Zeller on Unsplash

A couple of years ago we’d grab a phone book, pick up our phone and call a cab which would bring us from A to B. Nowadays you’d order an Uber or Lyft with just a touch on your Smartphone.

A nearly similar disruption happened to the hotel industry when Airbnb took over and made it possible for everyone to join the sharing economy. Renting ordinary rooms, apartments or whole houses became mainstream and made it possible to earn some money on the sideline.

Technology has greatly transformed our world throughout the last couple of decades. It always will. The only difference is the pace we’ll see upcoming technological changes.

How technology drove all industrial revolutions

The transportation and hotel industries are just two examples which highlight how disruptive technologies can take over and redefine the status quo. Let’s take a look how technology helped shape and pave the ways for the last three industrial revolutions.

In 1780 the first industrial revolution started when the steam engine was invented. Using steam engines made it possible to automate manufacturing processes which resulted in improved speed, quality and efficiency. More and more customer could be served while still reducing prices and increasing revenue. The first era of mass-productions just started.

At the end of the 19th century manufacturing was reinvented again. Motors, electricity and assembly lines made it possible to further automate the production process, yielding to yet another cost reduction and revenue increases. The emerging car industry was one of those industries which greatly benefitted from these changes.

Nearly at the end of the 20th century the first computers took over. Introducing the start of the third industrial revolution.
The digital era made it possible to define processes through algorithms and further optimize and automate most of the repetitive work. Computers greatly increased the quality of products, speed of delivery and therefore value creation while still reducing costs.
The invention and rise of the internet was another important milestone in the third industrial revolution. This new, interconnected world made it possible for businesses to easily get in touch with suppliers, customers and other businesses.

Entering Industry 4.0

It may seem like that’s where we are right now.

“Are we stuck?”
“What’s next?”
“What technology will drive the next industrial revolution?”

The answer to those questions is:

“It’s already happening!”

We’re already participating in and writing the next chapter in the book of industrial revolutions.

It’s just that this time there’s not one dominant technology which drives it. This time it’s the composition of multiple, different technologies which were invented and refined throughout the last couple of years.

The name and term „Industry 4.0“ differs slightly from the other industrial revolutions since this time we’re using multiple technologies to compose different technological frameworks we can (re)use. Given that it might feel that Industry 4.0 is more like an enhancement of the third industrial revolution, rather than an industrial revolution of its own.

The part 4.0 refers to software versioning, putting more emphasis on the huge impact software has this time.

The main goal of Industry 4.0 is to compose and orchestrate the new technologies and build an interconnected system which will transform old factories into so-called „Smart Factories“.

The Smart Factory is automated, customizable, intelligent, self-learning and self-healing.

The technologies

Since Industry 4.0 is driven by many different technologies it’s worthwhile to take a look at each technology individually to understand its capabilities. In addition to that it’s important to see how it fits into the whole picture, paving the way towards a Smart Factory.

IoT

Technologies from the realm of the „Internet of Things“ (IoT) are a main building block in every Smart Factory.

IoT can be rolled-out via sensors which are installed in all machines in the respective factory. Once implemented, those machines can be connected to the intranet / internet and send and receive data.

Autonomous Robotics

When machines have access to the intranet / internet it’s possible to connect them with one another so that they can share information.

This way machines can react to events in an autonomous way. Furthermore they can coordinate their work without any kind of human interaction.

Cloud Computing

Making the machines more powerful through sensors and processing units while connecting them with each other introduces challenges of its own:

“How much computing power should each machine have?”
“How will software updates and security fixes be rolled-out on those machines?”
“How would a hardware upgrade plan look like?”
“Where and how should the processed data be stored so that it’s shareable with other participants?”

A solution for those problems is „Cloud Computing“.

Machines should only gather sensor data. This data will be send into a nearby data-center where its stored and processed. Once ready, results such as machine instructions are sent back to the machine.

Using Cloud Computing for data storage and processing has different advantages:

  • Computing capabilities can be optimized for the machines purpose
  • All data is stored and processed in one central place
  • The data-center becomes the single source of truth
  • The hard- and software in the data-center can be easily updated without the need for a complex machine-based rollout plan
  • Data can easily be made accessible (e.g. to other factories, business partners or even customers)

It’s important to note that other technologies such as „Edge Computing“ emerged in the last couple of years. Edge Computing in particular makes it possible to process data in geographically nearby locations to save expensive round-trips and therefore greatly improve latency and overall system performance.

Cyber Security

There’s this saying that “data is the new gold”. The more data, the better algorithms will perform.

Lots of different data is gathered by all machines in a Smart Factory. This data needs to be protected against corruption, loss or theft.

Top-notch cyber security techniques need to be implemented to save this valuable business asset.

Furthermore it’s important to define data-related requirements (e.g. versioning, permissions, privacy, etc.) so that different Cloud Computing providers and their service offerings can be evaluated to decide where and how to operate the businesses data-center.

Big Data, AI & ML

Given this massive amount of data it’s important to define and understand how to get the most out of it.

Algorithms, models and other techniques from the fields of Big Data, Machine Learning (ML) and Artificial Intelligence (AI) should be used to extract important information and help the system train, learn and adapt autonomously.

Data-based training and learning can come in different flavors. One example is the ability to detect patterns in data sets to predict upcoming, well-known events.
This could e.g. be the detection of patterns which resulted in a maintenance scenario in the past. A human can be automatically informed to check the system and therefore reduce the risk of an unplanned downtime.
Another example could be that the system is intelligent enough to self-heal, making it unnecessary for a human to intercept.

Simulations, VR & AR

Most of the discussed technologies are aiming to build and establish a fully automated system in which machines are connected with one another, share data and get back feedback about their performance to learn how they can optimize their own work and collaborate with other machines more effective.

While it’s still the goal to build a Smart Factory where most of the heavy lifting is done by machines in an autonomous way it’s equally important to use other technologies to aid and accelerate the work humans are still involved with.

Useful approaches such as Simulations or Virtual Reality (VR) can be used to design, explore and create new product ideas. This makes it possible to discuss prototypes without the need to create something physical (and therefore expensive) in the real world.

VR headsets make it possible to virtually meet and discuss products with team members or even suppliers in different cities or countries around the globe. Change requests can be implemented easily and iteratively in a timely manner.

Another great use-case for Augmented Reality (AR) and Virtual Reality (VR) are virtual product demos for customers or interactive training sessions for employees.

Challenges

Since Industry 4.0 introduces lots of new technologies, concepts and implementations it’s worthwhile to take a look at the different challenges a rollout may entail.

IT Security

One of the main building blocks of Industry 4.0 is data. Data is one of the most valuable business assets since it’ll be used to feed the algorithms which in turn inform the system and its machines which actions to take. Such actions could be to e.g. optimize processes or inform a human if an incident is on the horizon.

All this data is coming from different IoT devices, which often run several softwares with lots of different versions. This makes it harder to protect every single system against fraud and potential security breaches.

The data-center which acts as the single-source for data storing and data processing is another critical piece of infrastructure. Time and again the data needs to be accessible from outside the factory to be shared with a 3rd party. Securing those entry points needs lots of attention and costly expertise.

Data Processing

Lots of reccurring planning and updates need to be performed to optimize the algorithms and models around different AI / ML processes. Those algorithms and models determine how the systems will learn and behave in the future. Even marginal errors on that front can result in massive problems during production and therefore put the whole organization at risk.

Reliability

System-wide reliability is crucial since different systems rely on each other to communicate and share data.

Reliability comes in lots of different flavors. Ranging from uptime, to short response-times, to correctness of data.

Legal situation

Moving towards an automated, Smart Factory will introduce new questions around legal issues, regulation and liability.

In a Smart Factory machines are almost fully-automated, communicate with one-another to streamline processes and make decisions on their own.

The following questions may come up:

“Who’s responsible if the certified product doesn’t meet the standards?“
“Who’s liable when an accident happens where humans and machines worked together?”

Training

Industry 4.0 introduces lots of new concepts and processes which makes it harder for existing and new employees to get on-boarded and learn the necessary skill sets required to work in this environment.

Upfront investments

Going all-in towards a Smart Factory means that there’s a huge money- and time-wise upfront investment.

There’s a necessity to install new hard- and software, establish new processes, onboard and train the workforce, legal paperwork, etc.

When kicking-off this huge endeavor it’s not clear whether this investment will ever pay off.

Given this it’s beneficial to analyze the status-quo and introduce new technologies one after another. This way the business can move slowly in the direction of a Smart Factory without making a huge bet by going all-in.

Opportunities

Given all these challenges there are huge upsides as well. Those upsides may alleviate some (if not all) of the risks described above.

Continuously process optimizations

Artificial Intelligence (AI) and Machine Learning (ML) make it possible to continuously optimize business processes. The data-driven self-learning system is always eager to optimize the current process to be more efficient without the need for a human supervisor.

Reliability

Given the fact that most of the systems are self-learning and self-healing this also means that the overall reliability increases drastically over time.

Customizations

Being data-driven makes it possible to dynamically store and load the machine instructions to manufacture slightly or even completely different products in a timely manner.

This low-cost context-switching makes it possible to offer individual or even fully customized products at much lower costs compared to the old-fashioned way.

Inner- and inter-connection

Establishing a strong connection within the factory as well as between different 3rd parties introduces new ways to interact with each other and react faster to new market trends.

Conclusion

We’re already entering the next industrial revolution. This revolution is called Industry 4.0.

The main difference between the previous three industrial revolutions is that this time the revolution is driven by lots of different cutting-edge technologies. Those technologies are composed to transform old factories into so called Smart Factories. Smart Factories are highly connected, autonomous, intelligent, and data / algorithm driven.

All these technological improvements introduce new challenges around security and legal situations as well as questions around practicality and the related costs for buying into Industry 4.0 and Smart Factories.

Opportunities around increased speed, adaptability, reliability and continuously process improvements makes embracing Industry 4.0 and Smart Factories especially interesting.


Make sure to click on the “Follow” button and leave a clap (or two) if you like this kind of content.
Do you have questions or feedback? Let’s connect via Twitter or E-Mail!