采用实验和计算的方法研究了取向硅钢二次再结晶织构的演变过程。发现取向硅钢通过二次再结晶过程中连续的取向选择，最终获得单一Goss ({110}<001>)织构。在二次再结晶动力学模型中引入依赖取向的相对晶界能系数，可定量描述不同偏差角Goss及非Goss取向晶粒的长大速率差异。通过分析初次再结晶晶粒尺寸分布、晶界特征和抑制力水平等因素对二次再结晶取向选择行为的耦合影响，提出增强Goss晶粒取向选择优势的多参数匹配方法。

The key index of grain-oriented silicon steel is the sharpness of secondary recrystallization Goss ({110}<001>) texture, which is determined by the matrix grain size distribution, texture environment and inhibitor level. In the widely used low-temperature slab heating process in virtue of high efficiency and low-cost manufacturing, the instability of inhibitor and the enlarged matrix grain size distribution seriously restrict the occurrence of secondary recrystallization and the sharpness of Goss texture. The investigation on orientation selection behavior during abnormal grain growth can explore the potential routines to solve the problem. In this work, the evolution process of secondary recrystallization texture in grain-oriented silicon steel has been studied by both experiment and calculation. It is found that single Goss texture is finally obtained by means of continuous orientation selection during secondary recrystallization. The kinetic model for secondary recrystallization, introduced with orientation-dependent relative grain boundary energy coefficient, can describe quantitatively the difference in growth rate between Goss grains with various deviation angles and non-Goss grains. The combined effects of grain size distribution, grain boundary characteristic between Goss and matrix grains, together with inhibition force level on orientation selection behavior are analyzed. Accordingly, a multi-parameter matching method for promoting the advantage of Goss grains in orientation selection is proposed.