Rolling Mill Looper Mass-Flow Detection for Imminent Strip Breakout

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for predicting impending breakout situations in rolling processes lack accuracy and transparency, requiring operators to manually assess complex mass flow issues across multiple stand areas, leading to potential delays and increased downtime due to the 'black box' nature of artificial intelligence-based systems.

Innovation Solution

A method that uses a physical model incorporating process data and kinematics of the loop lifter to calculate strip storage lengths from two sources, comparing them to detect mass flow problems, and provides visual diagrams and key figures to support operator monitoring, with the application of bandpass filtering to reduce noise and ensure accurate assessments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If AI-based black box methods are used for breakout prediction, then prediction capability is provided, but transparency and reliability are reduced

Engineering Contradiction:
Improveautomated breakout predictionVSAvoidtransparency of prediction process
Core Design Contradiction:
Extent of automationVSLoss of information

Solution Approach 1:

The patent introduces correlation diagrams as an intermediary visual tool between the complex AI prediction system and the operator. These diagrams display relationships between multiple stand areas and loop storage, translating black box AI outputs into transparent, interpretable visual information that maintains automation while restoring understanding.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent segments the complex mass flow system into discrete visualizable components including individual stand areas, loop storage sections, and correlation relationships. This segmentation allows operators to understand specific local conditions rather than facing an undifferentiated black box prediction, maintaining automation transparency.

Inventive Principle:
Principle #1Segmentation

2Reliability

If operators manually assess mass flow across multiple stand areas, then comprehensive monitoring is achieved, but operator workload and response time are negatively impacted

Engineering Contradiction:
Improvemass flow monitoring accuracyVSAvoidmonitoring system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent merges information from multiple stand areas and loop storage measurements into unified correlation diagrams. This consolidation presents comprehensive mass flow status across all areas in a single integrated visual display, maintaining comprehensive monitoring while dramatically reducing the cognitive complexity operators must manage.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent transforms complex multi-dimensional mass flow data from multiple stands and loop areas into a two-dimensional correlation diagram visualization. This dimensional transformation preserves all relevant information while presenting it in an intuitively understandable format that reduces operational complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Productivity

If AI-based prediction methods are implemented, then breakout detection is automated, but error reconstruction and verification become difficult

Engineering Contradiction:
Improvebreakout detection speedVSAvoiderror reconstruction capability
Core Design Contradiction:
ProductivityVSEase of repair

Solution Approach 1:

The patent implements feedback through correlation diagrams that continuously display the relationship between predicted breakouts and actual mass flow conditions. This visual feedback loop allows operators to verify AI predictions in real-time, reconstruct errors by observing correlation patterns, and understand the reasoning behind predictions, thereby maintaining high detection speed while enabling error analysis.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4176985B1Method for predicting an imminent breakout situation of a rolled strip in a rolling mill
Publication Date: 2024.02.14 SMS GROUP GMBH
  • EP4176985B1 patent drawingFigure 1
  • EP4176985B1 patent drawingFigure 2
  • EP4176985B1 patent drawingFigure 3

AI summary

The invention relates to a method and a computer program for predicting an impending breakout situation of a rolled strip, in particular a metal strip, between a first and a subsequent second rolling stand of a rolling mill with a looper installed between the two rolling stands. To achieve this, the present invention proposes to compare the strip storage lengths between two successive rolling stands of a rolling mill, wherein the strip storage lengths are derived from two different sources/calculations: - The strip storage length resulting from the looper angle and looper kinematics; and - The strip storage length resulting from the integrated speed difference of the strip upon leaving the first rolling stand minus the strip speed upon entering the subsequent second rolling stand.It is assumed that discrepancies between these two calculated tape storage lengths indicate mass flow problems, i.e., impending outbreak situations.