点击次数：

论文类型：期刊论文

发表时间：2021-01-10

发表刊物：INTERNATIONAL JOURNAL OF MECHANICAL SCIENCES

卷号：187

ISSN号：0020-7403

关键字：Ceramic-centered annular electrode; Resistance welding; Artificial neural network; Simulation

摘要：A resistance welding method that uses a composite ceramic-centered annular electrode (RAW) was adopted to improve the weldability of high-strength steel. The design idea was to increase the strength of welded joints by the formation of a large nugget with low indentation depth. The strength and the nugget formation of joints produced by the RAW, and the traditional spot welding (RSW) were comparatively investigated by the microstructural, artificial neural network, and finite element model analyses. The highest failure load reached 22.3 kN in the RAW, and the strongest influence on the strength was welding time, followed by welding current, and electrode force at the end. The interaction between these welding parameters significantly influences the strength of joints in the RAW but not in the RSW. Based on the ANN models, a detailed relationship between the parameters and the strength in RAW were obtained, which also determined that the optimal window of welding parameters in the RAW was larger than that in the RSW. The FEM model analysis showed that the annular nugget was generated first, then expanded outward and inward in RAW, and finally completely fused in the nugget center. The mode of heat transfer and the nugget formation could contribute to better dissipation of the resistance heat generated in the RSW. This "soft" heat distribution yields a large nugget with low indentation depth.