Case Study: Severstal achieves 6.5% productivity boost and 100K tonnes/year additional processing with Cloudera

Improving manufacturing processes and productivity through streaming and AI

Severstal, a leading vertically integrated Russian steel and mining company, needed to boost production efficiency, product quality and cost leadership across dispersed mills and mines. The challenge was to ingest and act on massive, real-time telemetry and imagery—millions of data points per minute and ten cameras producing over five million photos a day—while leveraging a 6 PB industrial data lake to enable end-to-end digital solutions.

Severstal migrated to a Cloudera CDH architecture with streaming (Kafka, NiFi, MiNiFi), computer vision and ML models, achieving 0.7–1.5 second end-to-end latency. The solution automated control and analytics, driving >6.5% performance improvement on the continuous pickling line (≈100k tonnes extra metal/year), productivity gains of >5% on selected units, and broad improvements in quality, cost and energy efficiency.

Severstal

Boris Voskresenskii

Chief Digital Officer