Mechanism‐Guided and Data‐Driven Head Wedge Precontrol With Adjacent‐Coil Feature Inheritance in Hot Strip Finishing Rolling

Head wedge in hot strip finishing rolling is difficult to control because the strip head passes the mill before downstream wedge feedback becomes available. To address this head‐end blind zone, a mechanism‐guided and data‐driven precontrol method is developed. First, a baseline roll‐gap leveling model is established from historical production records under unchanged work‐roll and step‐pad conditions to determine the initial head presetting of the seven finishing stands. Second, a dynamic finite element model of the roll‐strip system is used to quantify the effects of strip width, thickness, and reduction ratio on wedge response, from which wedge control coefficients and interstand wedge inheritance coefficients are identified. Third, adjacent‐coil consistency is recognized from rough‐rolling deviation curves and local head asymmetry features, and the measured head wedge of the previous coil is transferred to the current coil to calculate additional roll gap adjustments for stands F1–F3. Industrial application on a 2250 mm hot strip finishing mill shows that the head wedge hit rate increases by 10.57 percentage points and the whole‐coil wedge hit rate of silicon steel increases by 5.21 percentage points. The proposed method provides a practical and physically interpretable route for feedforward head wedge precontrol in finishing rolling.