针对平整分卷机组的分卷工作模式，采用弹塑性有限元方法对三辊张力装置增张过程进行数值模拟，得到了板厚、板宽、开卷张力、上张力辊压下位移量和结构参数等对张力装置增张能力的影响规律。将有限元模拟结果作为训练样本，结合国内某厂实测数据，建立了三辊张力装置增张能力的BP神经网络预测模型，预测结果与实测数据吻合良好，为张力装置增张定量控制提供了依据。

Aiming at deliver coil mode of skin pass mill, the tension enhanced process of the tension unit was simulated with nonlinear elastic-plasticity finite element method. From the simulation results, the influences of thickness and width of plate, decoil tension, depress of top tension roll on the amount of tension enhanced of the tension unit was achieved. With the simulation results and actual measurement data from some skin pass mill product line, a BP artificial neural network prediction model for tension enhanced ability of 3-roll tension unit was setup. The prediction results are in accordance with the actual measurement data, which provides a valuable method to establish more accurate operating system for the tension unit.