Neural Networks for Modelling and Predicting the Chlorella Protothecoides Cultivation Processes

摘要: The availability of estimators in the algal cultivation processes can lead to improving prediction and optimization. In this study a new simulation method is introduced into field cultivation. Two network models that describe process are developed. based on neural Chlorella protothecoides with highly nonlinear characteristics. types feed-forward networks, trained Levenberg-Marquardt algorithms (LMNN) radial basis function (RBFNN), considered paper. Modelling effort was focused selection structures, verification used. Data sets input-output patterns were obtained from thesis by computer simulation. Neural networks tested for their predictive abilities, an agreement between predicted values test data set shown. Follow-up studies indicate these be used as establishment controllers. Possible developments context modelling also discussed.