The detection of filling-to-packing switchover point during injection molding plays a crucial role in ensuring the quality of the molded *** detection methods,such as using predetermined screw cushioning or checking i...
详细信息
The detection of filling-to-packing switchover point during injection molding plays a crucial role in ensuring the quality of the molded *** detection methods,such as using predetermined screw cushioning or checking injection time have the disadvantages of depending on human experience and failure of adaptation to process parameter *** study presents a filling-to-packing detection method using the theory of singularity detection based on wavelet transform modulus maxima,according to the jump feature at filling-to-packing *** adopted the pressure profile of a previous batch to obtain wavelet decomposition scale and switching threshold etc.,and then used sliding window to conduct on-line detection of switch *** experimental results indicated that the innovative switchover method yielded a more uniform product mass than any traditional methods.
基于残差思想的异常数据检测方法的关键之处在于对数据的准确预测.针对这一问题,提出基于改进径向基网络(radial basis function network,RBFN)的过程工业时间序列预测方法,该方法通过改变RBF网络的输入形式,使改进后的RBF网络能够更方...
详细信息
基于残差思想的异常数据检测方法的关键之处在于对数据的准确预测.针对这一问题,提出基于改进径向基网络(radial basis function network,RBFN)的过程工业时间序列预测方法,该方法通过改变RBF网络的输入形式,使改进后的RBF网络能够更方便地引入遗忘因子以及惩罚因子,以适应于基于残差思想的异常数据检测方法要求的动态性能和鲁棒性.通过理论证明改进的RBF网络与传统RBF网络的等效性,并通过实验比较证明改进后的RBF网络较传统的网络结构更简单,参数意义更明确.
暂无评论