一般化学习网络及其对化学反应器的建模应用(英文)

Universal learning network and its application in modeling chemical reactor

本文研究了一般化学习网络(Universal Learning Network)在多变量连续釜式反应器(CSTR)系统的建模应用.一般化学习网络具有节点之间有多重分支、任意2个节点互连且节点之间可具有任意的时间延迟的特点,因此能够应用在高度非线性复杂系统的辨识中.分别用一般化学习网络和常规的递归神经网络对多变量连续釜式反应器(CSTR)进行系统辨识比较,仿真结果验证了一般化学习网络结构比递归神经网络Elman的辨识精度更高,且网络结构更简洁紧凑的特点.

Universal Learning Network for modeling chemical reactor is discussed in this paper. Universal learning network consists of a number of interconnected nodes and each pair of nodes can be connected by multiple branches with arbitrary time delays. With all these structural characteristics, it provides a generalized framework to model and control highly complicated nonlinear system. Both the universal learning network and the conventional recurrent network have been used to identify the CSTR system. The simulation results verify the capability and effectiveness of universal learning network in process identification. The architecture of multi-branch with time-delay and the learning algorithm independent of the initial parameter values make it more accuracy than the recurrent network Elman in identification, and furthermore, the network structure is more simple and eompact.