Sequential adaptive networks: An ensemble of neural networks for feed forward control of L-methionine production

Generalized hybrid control synthesis for affine systems using sequential adaptive networks

“…Figure 4(a) and (b) present a comparison of performance of the hybrid control strategy with the feedforward control strategy presented in previous work 41. The performance of these control strategies, where the maximum concentration of methionine was achieved, has been compared.…”

“…(A) Hybrid control strategy developed in this study. (B) Feedforward control strategy developed in previous work 41…”

“…In our previous work,41 a SAN‐based simple feedforward control strategy was developed to track the methionine concentration along a predefined linear trajectory. The implementation procedure was very straightforward and only a one‐step manipulation of the product equation was required to calculate the amount of substrate needed to achieve the control objective.…”

“…Substituting, Equations (25) and (26) into Equation (24) we obtain Finally, the SAN–hybrid controller is obtained as where the constants k 1 and k 2 assume the role of the feedforward tuning parameters subject to the physical constraints of the process. This control law has a different structure compared with the simple feedforward control law previously implemented for this system 41. Here the objective function is imbedded in the control law.…”

Fed-Batch Production of Bacterial Ghosts Using Dielectric Spectroscopy for Dynamic Process Control

“…Different types of soft-sensors have been reviewed [ 24 , 28 ]. Data driven soft-sensors that use statistical regressions derived from multivariate data analysis but need huge training data sets to build process models were successfully applied [ 29 , 30 ]. During bioprocess development, often those training data sets are not available and elemental balancing approaches are the primary choice for the design of a soft sensing strategy.…”