This paper addresses the passivity of continuous bioreactors around their non-zero equilibrium point by representing the system into a port-Hamiltonian system. In case of continuous bioreactors, the steady-state values of substrate(s) and biomass are not zero, thus passivity at shifted equilibrium point becomes interesting. A general representation of bioreactors is written in a form that allows to derive very simple conditions to prove their shifted passivity, using a quadratic energy function. In addition to that, the case of enforcing shifted passivity using feedback proportional control for general representation of bioreactors is also considered. Practical conditions are given for general single microbial reactions, and an example of anaerobic digestion of volatile fatty acids illustrates the multiple microbial reactions case.

Fuel cells are multi-physics systems which exhibit nonlinear phenomena that are difficult to capture. Model Predictive Control allows to design an optimal controller based on a prediction of the fuel cell dynamical behavior. However, the nonlinear optimal problem involves a large number of variables and is difficult to solve. A new algorithm is presented which is based on model inversion by deep neural networks (Neural Network Inversion-based Model Predictive Control). The feedforward neural network is able to retrieve a sequence of controls from a string of past and future outputs. Next, a specified trajectory is supplied which can be parametrized by a convergence rate. The algorithm finds the convergence rate that optimizes an optimal criterion such as the efficiency of the fuel cell; the sequence of optimal control is obtained by feeding the neural network with the reference trajectory. Using the neural network based model inversion, the sequence of optimal controls over a receding horizon which optimizes a given criterion can be obtained from the parametrized reference trajectory. As usual in Model Predictive Control, the first control input is applied, and the procedure is repeated at the next time instant. This method allows to work out a predictive controller even for nonlinear systems, with very little optimization variables. The simulation results are displayed in the last part.

This paper proposes a two-layer microgrid supervisor based on Model Predictive Control (MPC). The supervisor in the upper layer relies on an economical optimization that considers the cost of energy and the load and production forecasts to define the State of Charge (SoC) targets for each storage device on a timescale of 15 minutes. The lower layer displays a shorter timescale and aims to control the equipment to ensure the stability of the overall system and SoC tracking while satisfying the economic constraints specified by the upper layer. These two layers require an uniformization of the timestep and of the references in order to behave properly. The main contributions of this paper are the microgrid network modelling embedded in the optimization routine of the lower layer and a discretization for integrating upper-layer references.

This paper deals with a new approach for complex systems modeling and control based on neural and fuzzy clustering algorithms. It aims to derive a base of local models describing the system in the whole operating domain. The implementation of this approach requires three main steps: 1) determination of the structure of the model-base, the number of models are found out by using Rival Penalized Competitive Learning (RPCL), and the operating clusters are selected referring to the fuzzy K-means algorithm, 2) parametric model identification using the clustering results 3) determination of the global system control parameters obtained by a fusion of local control parameters. The case of a second order nonlinear system is studied to illustrate the efficiency of the proposed approach.

An indirect iterative learning algorithm has shown to be able to update the parameters of an acceleration feedback controller for flexible manipulators. The fine estimation of the masses of a chain of mass-spring-dampers units allows the simultaneous tuning of both the feedback controller and the trajectory generation. This algorithm has been validated on an industrial robot arm.

The design of stable fuzzy controllers can be achieved by determining a series of embedded sets for which the attractor of any set is included into the next set of the series. The framework of the comparison principle, along with overvaluing systems and vector norms, allows the computation of the attractors and to handle the robustness with respect to uncertainties. Considering the inverted pendulum control problem, this procedure is applied to the parameter tuning of a fuzzy controller with singleton defuzzification, whose regionwise structure leads to a rational partition of the state space.