Soldering Parameter-Setting Approach Based on Improved Adaptive Neuro Fuzzy Inference Model

Abstract

Abstract: To solve the problem in setting parameters for nonlinear dynamic system, an adaptive neuro fuzzy inference system was adopted to model the dynamic I/O system. A neural network predictor was introduced to enhance the adaptive ability of fuzzy neural networks and nonlinear approximation ability. By virtue of the nearest neighbor clustering algorithm, the parameters and the structure of the fuzzy rules were adjusted and updated, which improves the accuracy of fuzzy reasoning and enhances the system’s fault tolerance and robustness. The simulation experiment’s result shows that the proposed method can effectively achieve the quick and accurate setting of the nonlinear I/O system parameters.

Key words: adaptive neuro fuzzy inference system(ANFIS)； neural network predictor(NNP), nearest neighbor clustering algorithm(NNCA), soldering, online parameter setting

CLC Number:

F253.3

SHA Jian-Jun, PAN erShun. Soldering Parameter-Setting Approach Based on Improved Adaptive Neuro Fuzzy Inference Model[J]. Journal of Shanghai Jiaotong University, 2011, 45(12): 1741-1746.

References

［1］Jin J H, Yu D. Online automatic process control using observable noise factors for discrete part manufacturing ［J］. IIE Transactions，2004，36：899 - 911.

［2］周德俭. SMT 组装质量检测与控制［M］.北京：国防工业出版社，2007.

［3］Tsai T, Yang T. A neurocomputing approach to the thermal profile control of the secondside reflow process in surface mount assembly［J］. Management, 2005, 16(3): 343359.

［4］Kosko. Neural networks and fuzzy systems［M］, Englewood Cliffs, USA: Taylor & Francis, 1991: 13112.

［5］Shooreh Deli, Mahdi Aliyari, Teshnehlab, et al. Identification using ANFIS with intelligent hybrid stable learning algorithm approaches and stability analysis of training methods［J］. Applied Soft Computing，2009, 9(2): 833850.

［6］Sadeghi B H M. A BPneural network predictor model for plastic injection molding process［J］. Journal of Materials Processing Technology，2000，103(3): 411416.

［7］Zahid N, Abouelala O, Limouri M, et al. Fuzzy clustering based on K-nearest-neighbours rule［J］. Fuzzy Sets and Systems，2001，120(2): 239247.

［8］贾忠中. SMT工艺质量控制［M］. 北京：电子工业出版社，2007：135138.

［9］潘开林，周德俭，覃匡宇.SMT再流焊接工艺预测与仿真技术研究现状［J］.电子工艺技术， 2000，45(5)：6570.

PAN Kailin, ZHOU Dejian, QIN Kuangyu. Study on simulation of reffow soldering process of SMT［J］. Journal of Electronics Process Technology,2000,45(5):6570.

［10］Liang Y C，Feng D P，Lee H P，et al. Successive approximation training algorithm for feedforward neural network ［J］. Neurocomputing，2002，42(14)：311322.

［11］Takagi T，Sugeno M. Fuzzy identification of systems and its application to modeling and control ［J ］.IEEE Trans on Systems Man and Cybern，1985，15 (1)：116132.

［12］Parlos A G，Rais O T，Atiya A F. Multistepahead prediction using dynamic recurrent neural networks［J］. Neural Networks，2000，13(7)：765786.

［13］卫东，曹广益，朱新坚. 基于一种改进自适应模糊神经技术的PEMFC 系统建模和控制［J］. 上海交通大学学报，2005，38(9)：15811586.

WEI Dong, CAO Guangyi, ZHU Xinjian. Modeling and novel adaptive fuzzy neural network control of proton exchange membrane fuel cell［J］. Journal of Shanghai Jiaotong University,2005,38(9):15811586.

［14］Philippe T, De Geuser F, Duguay S, et al. Clustering and nearest neighbour distances in atomprobe tomography［J］. Ultramicroscopy， 2009，109(10)：13041309.

［15］Yang T，Tsai T N，Yeh J. A neural networkbased prediction model for fine pitch stencilprinting quality in surface mount assembly［J］. Engineering Applications of Artificial Intelligence，2005，18(3)：335341.