1.College of Automation Engineering, Qingdao University of Science & Technology , Qingdao 266061, China; 2.Research Institute of Physical & Chemical Engineering of Nuclear Industry,Tianjin 300180,China
Clc Number:
TP29
Fund Project:
Article
|
Figures
|
Metrics
|
Reference
|
Related
|
Cited by
|
Materials
|
Comments
Abstract:
According to the nonlinear relationship between influencing factors including extracting agent concentration, temperature, pH and time In the extraction process of grape skin pigment and the outputs pigment extraction yield, the improved BP neural network prediction model is established. The traditional BP learning algorithm exists the deficiencies of a slow convergence speed and local extremum,momentum term is introduced to improve learning algorithm. The weights and thresholds of the network are trained、simulated and verified by MATLAB continually using iterative optimization control algorithm on the actual datas. The improved BP neural network has the advantages of high precision, strong generalization ability and strong practicability for pigment extraction yield prediction control, providing a good theoretical basis and prediction method for pigment yield.