Nonlinear Predictive Functional Control Research Articles

Speed Control for Variable Speed PMSM Drive System Using Nonlinear Variable-Horizon Predictive Functional Control

Electrohydrostatic actuators (EHAs) adopted variable speed permanent magnet synchronous motor (PMSM) drive systems are widely used by surface actuation systems in more electrical aircraft. However, the dynamic performance and the antidisturbance performance of EHA need to be improved due to the limited dynamic response of variable speed PMSM drive systems and uncertain aerodynamic disturbance. To improve the transient-state operation, nonlinear variable-horizon predictive functional control (NVHPFC) is proposed for the speed controller of the variable speed PMSM drive system. By analyzing the relationship between the prediction horizon and the electromagnetic torque, a nonlinear variable predictive horizon strategy is presented, which replaces the fixed predictive horizon in classical PFC. Aiming to improve the prediction accuracy and antidisturbance performance of NVHPFC, a discrete prediction error compensation strategy is incorporated with the NVHPFC. Meanwhile, the stability of the proposed method is verified by the Lyapunov function in the discrete domain. Finally, simulation and experimental results of the proposed method show stronger antidisturbance, smaller overshoot, and shorter settling time performance compared with the classical PFC method based on a field-programmable gate array (FPGA) test bench.

Design and Implementation of Hybrid Modeling and PFC for Oxygen Content Regulation in a Coke Furnace

This paper proposes the implementation of hybrid data driven modeling and predictive functional control (PFC) strategy for regulation of oxygen in an industrial coke furnace. A comprehensive model that incorporates simple step-response test and nonlinear optimization using neural network is first developed through process operation data. Then, a nonlinear PFC is designed to improve the dynamic response and steady operation. The proposed PFC overcomes the disadvantages of proportional-integral-derivative or linear advance control strategies because the developed process model yields better process dynamics prediction and facilitates the subsequent PFC controller to improve process operation. In addition, the linear iterative form of the controller design is implemented such that engineers can easily apply it to industrial processes. Results are shown by way of simulations and experimental tests to demonstrate the effectiveness of the proposed strategy.

Nonlinear modeling and predictive functional control of Hammerstein system with application to the turntable servo system

This article deals with the identification of nonlinear model and Nonlinear Predictive Functional Controller (NPFC) design based on the Hammerstein structure for the turntable servo system. As a mechanism with multi-mass rotational system, nonlinearities significantly influence the system operation, especially when the turntable is in the states of zero-crossing distortion or rapid acceleration/deceleration, etc. The field data from identification experiments are processed by Comprehensive Learning Particle Swarm Optimization (CLPSO). As a result, Hammerstein model can be derived to describe the input–output relationship globally, considering all the linear and nonlinear factors of the turntable system. Cross validation results demonstrate good correspondence between the real equipment and the identified model. In the second part of this manuscript, a nonlinear control strategy based on the genetic algorithm and predictive control is developed. The global nonlinear predictive controller is carried out by two steps: (i) build the linear predictive functional controller with state space equations for the linear subsystem of Hammerstein model, and (ii) optimize the global control variable by minimizing the cost function through genetic algorithm. On the basis of distinguish model for turntable and the effectiveness of NPFC, the good performance of tracking ability is achieved in the simulation results.

Nonlinear Adaptive Predictive Functional Control Based on the Takagi–Sugeno Model for Average Cracking Outlet Temperature of the Ethylene Cracking Furnace

The conventional PID control has been proven insufficient and incapable for this particular petro-chemical process. This paper proposes a nonlinear adaptive predictive functional control (NAPFC) algorithm based on the Takagi–Sugeno (T-S) model for average cracking outlet temperature (ACOT) of the ethylene cracking furnace. In this algorithm, in order to overcome the effect on system performance under model mismatch, the structure parameters of the T-S fuzzy model are confirmed, and the model consequent parameters are identified online using the forgetting factor least-square method. Prediction output is calculated according to the identified parameters instead of computing the Diophantine equation, thereby obtaining directly the predictive control law and avoiding the complex computation of the inverse of the matrix. Application results on ACOT of the ethylene cracking furnace show the proposed control strategy has strong tracking ability and robustness.

Vehicle Stability Control Study Based on Neural Network Predictive Method

A control method is proposed to improve vehicle yaw stability based on neural network predictive. A vehicle steering model using neural network control strategy is set up, at the same time, the nonlinear predictive functional control using the neural network model is developed for control of high-nonlinear system. New structure of neural network multi-step prediction that is different from cascade or parallel is given. The results illustrate that the nonlinear predictive functional control using neural network model is more effective for control nonlinear system than PID control. A simulation is performed with it during two different conditions: step input and sinusoidal input, the results showed that compared with uncontrolled, the presented controller achieve good steady response of side slip angle and yaw rate, and lighten the burden of the driver and improve vehicle yaw stability.

Modeling and nonlinear predictive functional control of liquid level in a coke fractionation tower

The level control of the fractionation tower in industrial coke unit is not very easy due to its complex characteristics and nonlinearity. Most are controlled by PID or linear predictive control methods without considering the complexity. This paper shows that a more comprehensive nonlinear-model-based predictive control method can further improve control performance. A verification of a nonlinear process model with plant data is first shown. Then the design of nonlinear predictive functional control is discussed and results are shown by way of simulations and application to demonstrate the effectiveness and feasibility of the proposed control strategy.

A simplified linear iterative predictive functional control approach for chamber pressure of industrial coke furnace

This paper describes the application of a linear iterative design of nonlinear predictive functional control for chamber pressure of industrial coke furnace. The control target is to maintain the chamber pressure of a coke furnace within a set range while rejecting the effect of disturbances (mainly the coke removing process and the switch of coke towers).The proposed control is easy to implement, as it selects a simple structure and a linear iterative approach instead of nonlinear control schemes to improve control performance in such a way that results compared with traditional PID controllers are improved. Application of industrial plant demonstrates the efficiency of the method.

Support vector machine based predictive functional control design for output temperature of coking furnace

A new support vector machine based nonlinear predictive functional control design method has been developed and applied to an industrial coking furnace, which leads to the improvement of regulatory capacity for both reference input tracking and load disturbance rejection compared with traditional PFC and PID control strategies. The nonlinear process is first treated into a linear part plus a nonlinear part, then a convergent overall linear predictive functional control law is designed. The method gives a direct and effective multi-step predicting method and uses linear methods to get the control law which avoids the complicated nonlinear optimization. Comparison results and application to the temperature control of the industrial heavy oil coking furnace are presented in the article showing the efficiency of the method.

Nonlinear Predictive Functional Control Based on Artificial Neural Network

An Artificial Neural Network (ANN) is an adequate tool for modeling nonlinear systems and can be applied straightforward in the predictive functional control. New structure of ANN multi-step prediction that is different from cascade or parallel is presented, at the same time, the nonlinear predictive functional control using this ANN model has been developed in this paper. The useful of this control strategy is evaluated by applying it to a Continuous Stirred Tank Reactor (CSTR). The simulation results indicate that it is more effective than PID control.