URALCHEM* has embraced machine learning to accurately forecast output amount and quality, build predictive maintenance models, and test business hypotheses, while taking advantage of Data Lake and sensor-based equipment monitoring system implementation by IT company CROC.
Creating a single information environment to handle sensor-generated data was the first step towards implementation of a single Big Data and Business Intelligence (BI) system across all URALCHEM production facilities.
The system has already been integrated with a BI platform to allow for real-time reporting and implementation of a deviation analysis algorithm capable of monitoring and detecting inoperable equipment sensors in near-real time.
"URALCHEM is undergoing massive digital transformation. Digital teams that used to perform service functions have become a driving force behind this project, creating technological solutions to optimize our corporate processes and increase business value. Although we’ve only just got started, the Data Lake, built by CROC, already provides ample opportunities for leveraging Big Data and Machine Learning across the entire holding".Vladimir Rudenko
Director of IT Department at Unified Service Center Branch and Head of Machine Learning Practice, URALCHEM Holding PLC.
"Until a few years ago, most manufacturing companies could not even imagine that data from equipment sensors would be a valuable asset to improve production performance.Igor Zeldets
Now it's different: manufacturers have embraced data mining, storage and advanced analytics to allow for more effective decision making. This joint project will significantly help URALCHEM obtain tangible business results through more accurate and high-quality predictions, thus offering the possibility of a single manageable environment comprising all manufacturing and business data. Comprehensive monitoring of the production process is another initiative with strong potential to look forward to".
Business Development Director for Oil, Gas, and Chemicals, CROC
CROC specialists designed a scalable platform to analyze and store data streams from machinery sensors, even including data generated 5+ years ago. The platform can become the basis for manufacturing, finance, shipment and stock data consolidation, while also performing data analysis and KPI monitoring jobs. Machine Learning and Big Data technologies allow for the creation of new tools to solve practical business tasks faced by the holding, while analysis of collected data provides valuable insights to improve product quality and production efficiency. The Data Lake also has a feature to develop and implement predictive MRO based on machine learning.
"Digital enterprise is no longer a whim or option, but a must-do, given current market conditions and competition. Business processes, production, and logistics cannot be separated from IT and digital culture like in the past. We are happy to partner with URALCHEM and leverage our knowledge to assist its digital journey towards Industry 4.0. Our experienced and passionate specialists managed to deploy a complex system within the shortest period possible, thus delivering a solid foundation to handle Big Data across the entire holding".Egor Osipov
Head of Big Data Practice, CROC
1 About URALCHEM
URALCHEM Holding PLC is one of the largest producers of nitrogen and phosphate fertilizers in Russia and the CIS. The company has facilities capable of producing over 3 million tons of ammonia, 3 million tons of ammonium nitrate, 1.2 million tons of urea, and 1 million tons of phosphate and compound fertilizers per year. The holding is the leading producer of ammonium nitrate and one of the top three ammonia, urea and nitrogen fertilizer producers in Russia. URALCHEM’s key assets include: Azot Branch in Berezniki, Perm Territory; PMU Branch in Perm, Perm Territory; KCKK Branch in Kirovo-Chepetsk, Kirov Region; and Voskresensk Mineral Fertilizers in Voskresensk, Moscow Region.