【摘要】 本文采用极致梯度提升机法（XGBOOST）建立了基于船用止裂钢小尺寸试样性能参数的大尺寸厚板止裂温度预测模型，该模型可通过小尺寸试样的侧面无塑性转变温度、心部屈服强度、心部抗拉强度、板厚和主应力等表征参量，快速预测船用止裂钢厚板止裂温度。结果表明，该模型决定系数R2为0.969，平均误差百分比为9.56%，均方根误差为2.4℃，模型预测精度较高。更多还原

【Abstract】 This paper employs the eXtreme Gradient Boosting(XGBoost) method to establish a prediction model for the crack-arrest temperature of large-sized thick plates of shipbuilding crack-arrest steel, based on the performance parameters of small-sized specimens. This model can rapidly predict the crack-arrest temperature of thick plates using characterization parameters such as the Lateral Nil Ductility Transition Temperature, core yield strength, core tensile strength, plate thickness, and principal stress obtained from small-sized specimens. The results indicate that the model achieves a coefficient of determination R2 of 0.969, a mean absolute percentage error of 9.56%, and a root mean square error of 2.4℃, demonstrating high prediction accuracy.更多还原