The digital transformation of industry

Recycling is one of the fundamental concepts of the circular economy, but so is efficiency in the use of raw materials and production facilities. The digitalization of the industry makes it possible to keep these two parameters under control and to significantly improve performance and energy consumption, reducing waste and therefore CO2 consumption.

At a structural level, the most important challenge for the chemical industry is to use data science to increase the overall effectiveness of a plant (Overall Equipment Effectiveness or OEE). The entire value chain is involved in the digital transformation:

• production: digital operations exploit the data collected to improve safety and increase the efficiency of the industry
• customer experience: it is possible to support dynamic customer decisions on multiple digital touchpoints such as websites, social networks, and chats
• business models: web technology is used to offer new advantages to customers through digital marketplaces and online customer services.

The challenges of data science in the chemical industry

At the large Baytown Texas site, Covestro is pursuing advanced digitalization that has the ambitious goal of reducing unplanned outages by 50%. With the use of all available technologies, the aim is to increase safety at work and the environment, improve energy efficiency, reduce maintenance costs with the consequent achievement of business objectives in terms of increased revenue.

The adoption of the digital systems of the Californian company OSIsoft has allowed the US headquarters in Covestro to create a large-scale digital monitoring and data science platform in the space of a few months. A global hierarchy of assets has been organized to monitor the health of company resources in real- time.

The collected data will be used in a predictive way to guide future choices. Digitization does not only mean monitoring, with sensors connected to computers, breaks ,and interruptions as they occur but above all collecting data over time to have a general look at the company's operation. The calculated models can estimate maintenance before failures occur and provide the right information at the right time so that technicians can make quick and informed decisions.

How to design digitalization to improve the OEE

Digitalization was structured on three levels: the sensors, placed in the relevant points of the plant, the assets organized into categories and the actual production.

Sensors collect data in critical areas of a facility or production line. The choice of Covestro and OISsoft was to make all the detection points redundant and then proceed with the calibration of the data transmitted through the experience of the technicians. Rules have been decided to calculate and validate the conflicts that each measurement made with two different devices brings with it.

Stability, accuracy, and validity of the data collected, are organized by production resources, to establish the functioning and status of the assets. The amount of data is made understandable at a glance and functional through a graphic stylization to visualize the operation of each production sector.

This dynamic data collection system has led to:

• improved and simpler maintenance
• early detection of sensor failures or alterations in production processes
• the addition of missing sensors to define new asset models and their hierarchy
• the definition of further information on the production process and the improvement of the algorithms.

Everything is updated in real time and can be consulted remotely.

How to organize the amount of data in an industrial digitalization

The real challenge for data scientists who deal with digital transformation in the industrial sector is to organize and hierarchy the numerical sequences coming from the sensors in models and objects that are centered and consistent with the production infrastructures. Every single industrial asset is a complex and layered set of fundamental information.

For example, if we analyze a pump, we can classify at least six groups of functional, physical, or service information.

A Covestro production site can have about 1,000 pumps inside and if you multiply this value by the 30 sites around the world, you get an enormous amount of data that must be organized to be used in decision-making processes. We must face what specialists call the hurricane of data, a veritable hurricane of information.

The phases of the industrial digital transformation.

The experience of the Texan plant is paradigmatic for concretely observing how digitalization takes place over time within an active and important production site such as that of Baytown. The engineers have defined five operational phases starting from the choice and monitoring of industrial assets up to the creation of analysis dashboards, useful for control, and above all for decision-making in perspective.

1 - The first phase was to connect any significant equipment placed in the production line through recording systems to visualize how it works, how it is managed and maintained. This part lasted about 3 months.

2 - A second moment was reserved for the standardization of data, organized, for example, with tags and categories. In this way, resources are grouped on the one hand and processes on the other, and templates, and libraries are defined for organizing the databases.

3 - To make the information collected more readable and useful, the data is transformed through the definition of queries (database query) that come from the work practice and real needs within the plant. In this third phase, attention is paid to solving the problems of interfacing with the user, with his consolidated practices, knowledge and operational needs.

4 - At this point, it is necessary to maintain the synchrony and reliability between the different source systems to coordinate the integrity and governance of the data with a persistence strategy that also protects them from any deletions by users

5 - Thanks to all the previous actions, it is possible to analyze the entire network of company assets with increasingly specific and modular research for all maintenance or production needs. Proper digitization is the key to the success of the chemical industry and increases advantages in terms of its operational reliability, maintenance and safety.

How to make digital transformation in industry sustainable

People often wonder if digital transformation can replace human beings in the management of production plants which are thus made fully automated. For Covestro, which as an industry manages complex and high-tech chemical products, this is not possible: individuals and their vast experience can never be replaced by machines.

Information and data must be available to people but it is up to them to decide the path to follow: machines give advice but it is humans who choose.

Corporate sustainability is created if the production processes are defined by paths that put the human being as a point of departure and arrival.

To assess the impact that the mathematical models defined by the digitization system could have on workers, volunteer groups of Covestro plant technicians in China and the United States have tested and evaluated them.

Jane Arnold, Head of Global Process Control Technology at Covestro, says:

When you involve people in the project and don't just show them how it's done, then you see real results. The operation of roles changes and new levels of responsibility is created.

People are the key to success also in the digital transformation which must serve to create reasonable workflows, roles and responsibilities redefined with a view to sustainability, safety and the enhancement of skills.