Fuzzy neural networks (FNNs) are quite useful for nonlinear system identification when only the input/output information is available. A new FNN framework first proposed by combining an AutoRegressive with exogenous input (ARX) Tanh function in Takagi-Sugeno (T-S) type fuzzy consequent part. An improved genetic algorithm then designed to optimize structure and parameters of simultaneously under unknown plant dynamics. The hybrid encoding/decoding, neighborhood search operator, maintain...

Iterative learning control (ILC) has been successfully applied to numerous batch processes over the past decades. Monotonic convergence of tracking error is a desired characteristic that attracts much attention in academia. Many factors arising industrial practice, such as strong nonlinearity and parameter uncertainty, have challenged most existing monotonically convergent ILC approaches. This motivates development nonlinear (NMC-ILC) this paper. The proposed NMC-ILC an optimization-based...

This paper proposes an enhanced model predictive control (MPC) using a new state space structure for temperature of industrial coke furnace. The advantage the proposed controller lies in fact that its implementation only requires simple step-response process model, whereas design can be based on formulation to improve regulation. To ensure performance effectiveness under model/process mismatch and uncertainties, predictions cost function optimization are done basis improved model. MPC is...

The chamber pressure modeling of the industrial coke furnace is difficult due to flame instability in fuel burner and various disturbances. To deal with this issue, a new optimization method using radial basis function (RBF) neural network proposed improve accuracy simplify structure. An improved multi-objective evolutionary algorithm (MOEA) optimize input layer, hidden parameters functions RBF network. structure/parameter encoding local search, prolong pruning operators are designed make...

Modeling of distributed parameter systems is difficult because their nonlinearity and infinite-dimensional characteristics. Based on principal component analysis (PCA), a hybrid modeling strategy that consists decoupled linear autoregressive exogenous (ARX) model nonlinear radial basis function (RBF) neural network are proposed. The spatial-temporal output first divided into few dominant spatial functions finite-dimensional temporal series by PCA. Then, ARX designed to the dynamics modes...

The oxygen content modeling of the coke furnace is important for advanced control design but not an easy job because various disturbances and nonlinearity. A novel approach proposed by using improved genetic algorithm (IGA) combined with dynamic autoregressive exogenous input (ARX) Takagi–Sugeno (T-S) fuzzy model. IGA automatically generates variable, appropriate if–then rules, ARX structure to characterize nonlinear feature processing operation data from industrial furnace. And a more...

In this paper, the iterative learning fault-tolerant control problem for multiphase batch processes with uncertainty and actuator faults is studied. First, making full use of characteristics two-time dimension (2D) feature repetitiveness in introducing state error output between adjacent batches, established model transformed into an equivalent 2D-Roesser switched system different dimensions. Under framework 2D theory by means average dwell time method, sufficient conditions ensuring to be...

In this paper, a T-S model-based fuzzy delay-range-dependent iterative learning control (ILC) scheme is developed for highly nonlinear batch processes with interval time-varying delays. The two-dimensional (2D) time-delay model constructed to remedy the disadvantage that overall linear cannot sufficiently describe process. Then, exploiting repetitive nature of processes, 2D designed. stabilization problem and H∞ are studied by using Lyapunov function under system framework. At same time,...

In an industrial coke process, the dynamic relationship between input and output devices is complicated. Since performance closely related to control used, improved alternatives are necessary. This paper deals with oxygen content loop proposes state-space model predictive (MPC) aimed at satisfying steady air supply into process. The details of proposed MPC first described tested on a typical example then implemented furnace. Simulation examples real-time implementation results both show...

The paper presents a combination modeling procedure and the implementation of nonlinear predictive control scheme for optimization industrial chemical processes. model structure is first based on simple step response method. This provides way to use prior knowledge about dynamics, which has general validity, while additional information process behavior derived from measured plant data-model error. data error driven framework applicable wide range operating units under certain policy. same...

Concerning multiphase batch processes with delays, disturbances, and actuator faults, the design of 2D robust hybrid composite iterative learning fault-tolerant guaranteed cost controller is put forward. First, a control law introduced process interval time-varying delays converted to an equivalent 2D-FM switched system faults by introduction output tracking errors state consideration fault effects. Next, sufficient condition for satisfying asymptotical stability that closed-loop has upper...