论文部分内容阅读
提出了一种基于ANFIS的烧结终点预测模型,该模型采用减法聚类来确定隶属度函数的中心,接着采用正交最小二乘参数辨识对神经网络进行训练。由于影响烧结终点的因素较多,若要全部考虑到可能会引起神经网络输入的维数灾,文中首次采用了主成分分析法来减小ANFIS输入维数,避免出现维数灾。通过现场采集的数据,对该模型进行了仿真。实验证明,该模型有较好的学习能力和自适应能力,为烧结终点预测提供了一种新的算法。
An ANFIS-based sintering end-point prediction model is proposed. The model uses subtractive clustering to determine the center of the membership function, and then the neural network is trained using orthogonal least squares parameter identification. Because of the many factors that affect the sintering end point, if we consider the dimensionality disaster which may cause the input of neural network, the principal component analysis is used for the first time to reduce the input dimension of ANFIS and avoid the dimensionality disaster. Through the data collected in the field, the model was simulated. Experiments show that the model has good learning ability and self-adaptive ability, which provides a new algorithm for the prediction of sintering end point.