Industrial ConsumersInvestors

Artificial intelligence to reduce maintenance-time by 50% – Arcelor Mittal

Artificial-intelligence-to-reduce-maintenance-time-sets-in-50

Through Artificial Intelligence (AI) and digital transformation tools, companies in the steel sector could reduce their maintenance-time by 50%. Arcelor Mittal implemented the KNet tool, reducing 50% of its units’ maintenance-time in a year, the company’s representatives declared.

Artificial Intelligence in Arcelor Mittal

At the “Emerson Exchange Virtual Series” event, Arcelor Mittal’s and Laurentide Controls’ members shared their perspectives regarding the artificial intelligence and digital transformation relevance for decreasing maintenance-time.

According to the panelists, Arcelor Mittal reduced its maintenance times by 50% after implementing the artificial intelligence and digital transformation “KNet” tool for a year.

The KNet tool is a system that measures and adapts itself to machine behavior in a production plant. Such a course identifies the equipment’s ideal conditions, sending data and corrective recommendations for the assets corrects operation, maintenance, and management.

In that sense, the event’s participants said that to implement a strategy that considers the use of artificial intelligence, companies must take into account the following steps: the project’s definition, planning, scheduling, and execution.

Following the implementation line for artificial intelligence, users will collect, co-relation, and analyze data; build models to identify different operation modes; and revisit adopted models with new information in real-time.  

KNet goals

In the beginning, Arcelor Mittal’s objective was to save between 40 and 60 hours in maintenance-time per year, the panelists declared. Arcelor Mittal is a company with an international presence devoted to steel manufacture, shipping, analysis, research, and project development.

The participants also said the company would address such a goal by developing and implementing an intelligent app based on machine learning and prescriptive analytics.

The reason behind using these artificial intelligence tools was to predict failures way before they happened. In that sense, the maintenance-time would be reduced due to the anticipated analyses that KNet would provide to the company, the panelists shared.

After a year of implementing the KNet system, Arcelor Mittal is currently collecting the obtained data through machine learning and looking to identify the root causes for failure to optimize its maintenance and operational processes.

Arcelor Mittal: results and future actions

Along this year, Arcelor Mittal added 30 different sensors with a connection to controllers. The information generated by those sensors was stored in an existing port and a remote terminal for the open-access of operators to the data.

Therefore, the panelists shared the KNet system is also connected to different databases. Through artificial intelligence, the system can generate other prescriptive analytics models to identify existing and potential threats.  

In the first phase, the company found out the existing sensors could only identify past failures in 30%. This way, KNet responded, giving data and building recovery algorithms.

Possibly of your interest: MPC and 10-year digitalization programs – Emerson

In a second phase, implemented in June this year, Arcelor Mittal installed more data collection sensors. The goal was to obtain new algorithms with a broader reach in prediction within its diagnoses.

The company has covered approximately 76% of its failures with more robust algorithms after adding those sensors.

According to the panelists, the six following months will be crucial for the impact measurement regarding the last sensors’ installation and maintenance-time. The company will collect data from its sensors, define the algorithms, and monitor the implementation results.

Related posts

MPC and 10-year digitalization programs – Emerson

editor

Cognizant to deliver extended digital and AI support for Aker Solutions

editor

Rare Element Resources advances separation project for DOE funding

editor