Research of Ellipsoid Bounded Algorithm in Hybrid Modeling
-
摘要: 并行结构混合建模主要由机理模型与误差补偿模型组成.一般地,误差补偿模型不宜过于复杂,且模型应具有校正功能,以免精度随时间不断下降.针对这个问题,本文选择单层神经网络作为误差补偿模型,并将椭球定界算法应用于单层神经网络的参数更新,不仅能够保证建模误差稳定有界,同时能够提高网络的收敛速度.将提出的方法应用于氧化铝生产过程,改进了原有的苛性碱和氧化铝组 分浓度软测量方法.实验研究结果表明,椭球定界算法的应用提高了模型的精度和网络的收敛速度.除此之外,在存在噪声干扰下,改进 的方法比原有方法更稳定,进一步证明了方法的有效性和优越性.Abstract: A parallel structure hybrid model is mainly composed of a mechanism model and an error compensation model. In general, the error compensation model should be simple and have correction functions to prevent the accuracy from varying with time. Aiming at this problem, a single layer neural network with an ellipsoid bounded algorithm is proposed, and it is taken to enhance the convergence speed and ensure the boundedness of modeling error. The proposed method is applied to an alumina production process to improve the soft sensing method of caustic soda and alumina component concentration. The experimental results show that the ellipsoid bound algorithm can improve the model accuracy and the speed of network convergence. In addition, the improved method is more stable than the original method the presence of noise.
-
Key words:
- Hybrid modeling /
- soft sensing /
- ellipsoid bounded /
- neural network
-
[1] Chai Tian-You. Operational optimization and feedback control for complex industrial processes. Acta Automatica Sinica, 2013, 39(11): 1744-1757(柴天佑. 复杂工业过程运行优化与反馈控制. 自动化学报, 2013, 39(11): 1744-1757) [2] Tang J, Chai T Y, Zhao L J, Yu W, Yue H. Soft sensor for parameters of mill load based on multi-spectral segments PLS sub-models and on-line adaptive weighted fusion algorithm. Neurocomputing, 2012, 78(1): 38-47 [3] Gui Wei-Hua, Yang Chun-Hua, Chen Xiao-Fang, Wang Ya-Lin. Modeling and optimization problems and challenges arising in nonferrous metallurgical processes. Acta Automatica Sinica, 2013, 39(3): 197-207(桂卫华, 阳春华, 陈晓方, 王雅琳. 有色冶金过程建模与优化的若干问题及挑战. 自动化学报, 2013, 39(3): 197-207) [4] Tang Jian, Chai Tian-You, Yu Wen, Zhao Li-Jie. On-line KPLS algorithm with application to ensemble modeling parameters of mill load. Acta Automatica Sinica, 2013, 39(5): 471-486(汤健, 柴天佑, 余文, 赵立杰. 在线KPLS 建模方法及在磨机负荷参数集成建模中的应用. 自动化学报, 2013, 39(5): 471-486) [5] Ding J L, Chai T Y, Wang H. Offline modeling for product quality prediction of mineral processing using modeling error PDF shaping and entropy minimization. IEEE Transactions on Neural Networks, 2011, 22(3): 408-419 [6] Wu F H, Chai T Y. Soft sensing method for magnetic tube recovery ratio via fuzzy systems and neural networks. Neurocomputing, 2010, 73(13-15): 2489-2497 [7] Cong Q M, Yu W, Chai T Y. Cascade process modeling with mechanism-based hierarchical neural networks. International Journal of Neural Systems, 2010, 20(1): 1-11 [8] Qi H Y, Zhou X G, Liu L H, Yuan W K. A hybrid neural network-first principles model for fixed-bed reactor. Chemical Engineering Science, 1999, 54(13-14): 2521-2526 [9] Kramer M A, Thompson M L, Bhagat P M. Embedding theoretical models in neural networks. In: Proceedings of the 1992 American Control Conference. Chicago, USA: IEEE, 1992. 475-479 [10] Wang W, Yu W, Zhao L J, Chai T Y. PCA and neural networks-based soft sensing strategy with application in sodium aluminate solution. Journal of Experimental and Theoretical Artificial Intelligence, 2011, 23(1): 127-136 [11] Sun Xian-Fang, Zhang Zhi-Fang, Ning Wen-Ru, Fan Yue-Zu. Robustness of OBE algorithms to underestimation of error bounds. Acta Automatica Sinica, 1998, 24(6): 784-788 (孙先仿, 张志方, 宁文如, 范跃祖. OBE算法对误差界低估的鲁棒性. 自动化学报, 1998, 24(6): 784-788) [12] Rubio J, Yu W, Ferreyra A. Neural network training with optimal bounded ellipsoid algorithm. Neural Computing and Applications, 2009, 18(6): 623-631 [13] Yu W, Rubio J. Recurrent neural networks training with stable bounding ellipsoid algorithm. IEEE Transactions on Neural Networks, 2009, 20(6): 983-991 [14] Kosmatopoulos E B, Polycarpou M M, Christodoulou M A, Ioannou P A. High-order neural network structures for identification of dynamical systems. IEEE Transactions on Neural Networks, 1995, 6(2): 422-431 [15] Wang W, Chai T Y, Yu W, Wang H, Su C Y. Modeling component concentrations of sodium aluminate solution via hammerstein recurrent neural networks. IEEE Transactions on Control Systems Technology, 2012, 20(4): 971-982
点击查看大图
计量
- 文章访问数: 1526
- HTML全文浏览量: 127
- PDF下载量: 1508
- 被引次数: 0