This paper aims to solve the problem of tracking optimal profiles for a nonlinear multivariable fed-batch bioprocess by a simple but efficient closed-loop control technique based on a linear algebra approach. In the proposed methodology, the control actions are obtained by solving a system of linear equations without the need for state transformations. The optimal profiles to follow are directly those corresponding to output desired variables, therefore, estimation of states for nonmeasurable variables is considered by employing a neural networks method. The efficiency of the proposed controller is tested through several simulations, including process disturbances and operation under parametric uncertainty. The optimal controller parameters are selected through the Montecarlo Randomized Algorithm. In addition, proof of convergence to zero of tracking errors is analyzed and included in this article.

No abstract

The problem of optimal profiles tracking control under uncertainties for a nonlinear fed-batch bioprocess is addressed in this paper. Based on the results reported by Pantano et al. [Ind. Eng. Chem. Res. 2017, 56, 6043], this work aims to improve the control system response against parametric uncertainty and process disturbances. The methodology is simple and easy to implement, but with excellent results. The design parameters are optimized by a randomized Monte Carlo algorithm. Besides, demonstration of the tracking error convergence to zero when the system is subjected to uncertainties is included in the article. The control system performance is tested through simulations, showing the improvement achieved.

No abstract

No abstract

Bioethanol is one of the most studied alternative fuels nowadays. Due to its production process complexity and the low quality of the mathematical models that describe it, a reliable controller is needed to maximize the fuel production and minimize its environmental impact, even in the presence of uncertainty. Here, a controller for tracking optimal profiles considering model errors and external perturbations is proposed. This work is an improvement of a previously presented technique. To reduce the earlier mentioned uncertainties’ effect during the fermentation, some tracking error integrators are added in the control action calculation. This simple modification ensures the tracking error convergence to zero, even in the presence of uncertainties (demonstration available). Different tests are carried out and a performance comparison with the original controller is shown to highlight improvements in the tracking error of up to 98% when integrators are incorporated. Furthermore, a classical PI controller is contrasted with the proposed techniques.

No abstract

No abstract