Reshaping the boundaries between physical and digital world: from Industry 1.0 to Industry 4.0

Categories: Innovation

20 Dec 2018

Reshaping the boundaries between physical and digital world: from Industry 1.0 to Industry 4.0

Steam and water power characterized the 1st Industrial Revolution. The 2nd Revolution introduced electricity to mass produce things. With the 3rd Revolution, the Internet digitized the connection between people and devices. The 4th Industrial Revolution is currently in progress. It is reshaping the boundaries between physical and digital world, blurring into a seamless and interconnected ecosystem.

Splitting Industry 3.0 technologies from 4.0 ones is not always easy. Simply what sets aside Industry 4.0 solutions is not the technological complexity, but the compliance with two key principles:

 

  • Connectivity enabling collaboration between the physical and digital spheres. Each physical item (equipment, people, raw material, etc.) have a unique digital identity exchanging data to enable collaboration at different levels.
     
  • Automatization. While automation refers to a technology that perform a repetitive task (i.e. mixing the same compounds in the same quantity every day). Automatization defines a system equipped with intelligence able to take decisions on which task to perform and the best way to do it. For example, Industry 4.0 machines intelligently combine actual sales and demand forecast to optimize which compound to mix.

 

Data driven machines are helping humans to take better decisions

 

In Industry 4.0, data are the lifeblood of decision making processes. Data are neither logically related nor structured. When processed by humans or machines, data translates into rationally organized information that drive decision making processes. To use an industrial metaphor data are raw materials, information are components and decisions are final products. Today, machines capability to analyze data is unparalleled, and these are the top reasons:

 

  • Machines proficiently collect and analyze large amounts of data (from digital and physical assets) that human brains cannot even conceive.
  • Machines can work as microscopes or telescopes by showing small details and big pictures of datasets. This double perspective reveals correlations that were previously unknown by decision makers.
  • Machines collect and analyze data in real time; their impact on decision making processes has a completely new strategic value from retro perspective data analysis that were available in the past. 

These new analytical capabilities are due to a paradigmatic revolution of digital machines. Industry 3.0 devices address each contingency through long receipts (codes) written by software engineers. For instance, an iPhone works thanks to a 14 million lines receipt. If iOS software crashes, only engineers can rewrite the receipt and fix the iPhone. Industry 4.0 devices work differently: these write and rewrite their own code starting from examples they analyze (data). Like humans, devices learn from new experiences and failures. Powerful analytical capabilities allow machines to analyze many more experiences (data) than human beings and take extremely complex decisions.

 

Overall, humans are better decision makers than machines because of our ability to adapt to many different domains. Machines can make us more effective and efficient! We can take advantage of their ability to understand complex data and take better decisions in specific domains, for example driving cars or forecast customer demand. As we leave machines taking specific decisions for us and start making sense of complex data, we will have more time and opportunities to apply our uniquely valuable capabilities, such as creativity, leadership, morality and emotional intelligence.

 

Application: autonomous cars

 

Autonomous cars are a great example to understand the paradigmatic shift of Industry 4.0 devices. To drive this disruptive innovation, we are changing the nature of the problem. Industry 3.0 software engineers would have tried to teach cars how to handle every single situation (through long receipts of code). In Industry 4.0 engineers provide cars with data about surrounding environments; cars are independent in analyzing data and finding the best solution to move around. Cars gather data from different sources. Occasionally, the environment is pre-mapped. In real time, car sensors update pre-maps (i.e. road deviations), gain data about moving shapes (i.e. pedestrians) and share car localization with other vehicles (with centimeter accuracy). As cars acquire more data about what's happening in our streets, their autonomous driving system become more reliable and will soon overcome humans' driving ability. Car accidents due to humans’ mistakes are among the first causes of death: every day 3,287 people die while driving (asirt.org). As autonomous cars learn from data, we hope this dramatic figure will drop down. 

 

Business case: PRY-CAM

 

PRY-CAM is a great Industry 4.0 case to understand how Machine Learning can make electrical assets smarter. PRY-CAM collects partial discharges (PD) data innovatively by not being directly connected to electrical fields: a revolution of the PD data acquisition process. Indeed, PRY-CAM is built on data stored on the cloud. As, Roberto Candela, Managing Director of Prysmian Electronics, explains: “all connected devices can transfer to the cloud, either Prysmian’s or the client’s, a whole set of information that can be accessed in real time and from any location both by our technicians and the client’s”. Today the cloud database hosts more than one million PD measurements. The analysis of this huge amount of data would have required hundreds of human experts in the past, nowadays PRY-CAM algorithm alone performs this task and generates hundreds of electric field graphs daily.

Applying the industrial metaphor to PRY-CAM: PD data are raw materials (illogically related data) and graphs of electric fields designed by the device are components (logically structured info). Now, humans look at those graphs to find previously identified PD patterns, anticipate future ones and take maintenance decisions. PRY-CAM has developed a computer that is learning PD patterns. As we collect more PD data the computer is getting better and will most probably overcome humans at interpreting graphs and taking decisions. Prysmian’s goal is embedding electronic solutions into cables: these smart electrical assets will continuously monitor their status and autonomously take maintenance decisions only when and where is needed.

 

To learn more about Prysmian Electronics’ solutions please follow this link

 

 Camillo Dalloglio - International Graduate

Davide Fassone - Innovation Specialist