Control Of Nonlinear Processes Research Articles

Time domain modeling and control of complex non-linear chemical processes using relay feedback test

This study analyzes with the technology of auto tuning using relay feedback test for non-square MIMO system through process modeling, identification, input-output pairing and control strategies. However, the control configuration selection based on conventional steady-state Relative Gain Array (RGA) matrix sometimes degrades the loop performance and it needs attention. This study also deals with the real challenges in time domain modeling and appropriate pairing of loops for non-linear chemical processes. The choice of input-output pairing for square system has been extended to non-linear chemical processes. This ensures the system’s stability by selecting an appropriate manipulated-controlled variable pairing in non-linear chemical processes and the same is also tested for benchmark of non-linear chemical processes. In this study, two types of non-square systems are considered: one is systems with excess input than output variables and the other is systems with excess output than input variables. Some benchmark non-linear chemical processes, such as processes with mild nonlinearity, moderate to high nonlinearity and highly nonlinearity, are also taken for this study. Three standard benchmark processes are used in analysis of nonlinear chemical processes, namely: (1) Continuous Stirred Tank Reactor (CSTR) process; (2) hydrogen-ion- concentration (pH) process and (3) distillation column and this procedure is illustrated via simulation of 3-by-2 and 2-by-3 non-linear chemical processes.

Process structure-based recurrent neural network modeling for model predictive control of nonlinear processes

In this work, physics-based recurrent neural network (RNN) modeling approaches are proposed for a general class of nonlinear dynamic process systems to improve prediction accuracy by incorporating a priori process knowledge. Specifically, a hybrid modeling method is first introduced to integrate first-principles models and RNN models. Subsequently, a partially-connected RNN modeling method that designs the RNN structure based on a priori structural process knowledge, and a weight-constrained RNN modeling method that employs weight constraints in the optimization problem of the RNN training process are developed. The proposed physics-based RNN models are utilized in model predictive controllers and applied to a chemical process network example to demonstrate their improved approximation performance compared to the fully-connected RNN model that is developed as a black box model.

Diagonal-structure adaptive bilinear filters for multichannel active noise control of nonlinear noise processes

To accommodate multichannel nonlinear active noise control, the second-order Volterra filtered-x LMS (VFXLMS) and filtered-s least mean square (FSLMS) algorithms are usually applied in practical noise control systems. These two algorithms are effective in dealing with polynomial nonlinearity. Recently, a simplified bilinear leaky filtered-x least mean square (SBLFXLMS) algorithm is proposed based on a bilinear model in order to further reduce the computational load but it may have performance degradation. In this paper, we propose a novel nonlinear adaptive algorithm named as the diagonal-structure bilinear filtered-x least mean square (DBFXLMS) algorithm for multichannel nonlinear active noise control. In addition, the developed multichannel DBFXLMS algorithm is equipped with a stability control scheme to maintain a stable condition of the adaptive bilinear filter during the adaptive process. The performance of the proposed algorithm is validated through computer simulations and the computational complexity of the developed algorithm is analyzed. Computer simulations demonstrate that the proposed method has an improvement in terms of control performance in comparison with the conventional first-order FSLMS, second-order VFXLMS, and SBLFXLMS algorithms.

A Novel Fuzzy PI Control Approach for Nonlinear Processes

Classical fuzzy logic controllers fail to control nonlinear systems, which have strong nonlinearity around a certain operating point. In this case, if it is possible, the block-oriented modeling approach could be adopted to separate the highly nonlinear system into two subsystems as a linear time-invariant subsystem and a memoryless nonlinear subsystem. Thereby, a control operation could be carried out on the linear time-invariant block to control the whole system. Respecting this approach, here, a highly nonlinear system (neutralization process) was modeled by using the ARX–polynomial cascade connection. By this means, an opportunity is obtained that the inverse polynomial is utilized to be able to control the ARX subsystem by a classical fuzzy PI controller. And therefore, the fuzzy PI controller performance in controlling nonlinear processes was enhanced. The success of the new approach was confirmed by the results obtained from the control operation carried out in different pH regions for a chemical neutralization process. And also, the results show that the proposed control strategy is as successful as the multiregion fuzzy PI controller in the highly nonlinear process control.

Design of Adaptive Fuzzy PID Controller for Electropneumatic System

The purpose of the study is to develop an adaptive electropneumatic automatic system based on the classical PID controller using modern technology for controlling like fuzzy controlling. The need of an adaptive setting occurred because of the fact that the classic controller has high sensitivity towards interferences with the main signal and the need for the parameters to be regulated often and be done by hand. Another reason is the inability to provide optimal parameters for the controller for nonlinear processes and changes of the parameters in time. In this article the work of electropneumatic system with linear actuator, is examined. A mathematical model describing the dynamics of the processes, is developed. An adaptive fuzzy PID controller with variable coefficients is synthesized. The system is studied in the environment of Matlab Simulink. The transient processes in the system are compared to processes where the classical PID controller. The results are presented in graphical form.

Control Lyapunov-Barrier function-based predictive control of nonlinear processes using machine learning modeling

Control Lyapunov-Barrier functions (CLBF) have been adopted to design model predictive controllers (MPC) for input-constrained nonlinear systems to ensure closed-loop stability and process operational safety simultaneously. In this work, a CLBF-MPC using an ensemble of recurrent neural network (RNN) models is proposed with guaranteed closed-loop stability and process operational safety for two types of unsafe regions, i.e., bounded and unbounded sets, for nonlinear processes. The application of the proposed RNN-based CLBF-MPC method is demonstrated through a chemical process example.

Control of nonlinear technological objects using fuzzy PDC-controllers

The features of constructing PDC algorithms - control of nonlinear objects using discrete local TS-models are investigated. The schemes of PDC - regulators with compensation of statistical errors are proposed due to the introduction of digital integrators into the structure of the control system and the subsequent calculation of the regulator coefficients for each of the affine submodels. The results of modeling the considered scheme for control of nonlinear processes are presented. The prospects for using the proposed approach are determined.

DCS devices based non-linear process control system design for plants with distributed time-delay using particle filter

Remote control process system with distributed time-delay has attracted much attention in different fields. In this paper, non-linear remote control of a single tank process system with wireless network is considered. To deal with the distributed time-delay in a large-scale plant, the time-delay compensation controller based on DCS devices is designed by using operator theory and particle filter. Distributed control system (DCS) device is developed to monitor and control from the central monitoring room to each process. The particle filter is a probabilistic method to estimate unobservable information from observable information. First, remote control system and experimental equipment are introduced. Second, control system based on an operator theory is designed. Then, process system with distributed time-delay using particle filter is carried out. Finally, the actual experiment is conducted by using the proposed time-delay compensation controller. When estimating with the proposed method, the result is close to the case in which the distributed time-delay does not exist. The effectiveness of the proposed control system is confirmed by experiment results.

Fractional Order PI Controller for Nonlinear Processes using Binomial Theorem Approximation

In Most of the process industries, Integer order Proportional Integral controller (IOPI) has been designed for different processes due to its simplicity but it will damage the final control element due to the sudden change in controller output. So, this research work proposed to eliminate the ringing phenomena by applying large pure time delay in the controller output by proposing Fractional Order Proportional Integral (FOPI) controller using Binomial theorem approximation for the nonlinear conical and spherical tank level processes. The control of liquid level in the nonlinear tank is arduous in nature due to an imbalance in the area of the transverse section of the tank. The robustness of the IOPI and FOPI controllers has demonstrated in Time Domain Specifications (TDS). The FOPI controller eliminates the peak overshoot in the process variable and ringing effect in controller output to protect the pneumatic positioned control valve.

EXPERIMENTAL VERIFICATION STUDY ON TRAJECTORY TRACKING OF A NOVEL REHABILITATION EXOSKELETON SHOULDER JOINT

The trajectory linearization control (TLC) was applied to design an autonomous nonlinear trajectory tracking controller for a novel rehabilitation exoskeleton shoulder joint in this paper. TLC is a relatively new control method which was applied in aircraft and mobile robot, which had good performance on real-time trajectory tracking, anti-jamming and universality. As a new application in the exoskeleton shoulder joint controller design, the controller in this research contained two loops that separately based on the inverse kinematics and pseudo-inverse dynamics models of the exoskeleton shoulder. Two PI controllers as the error regulator can reduce the tracking error. The position and angular velocity error feedback were employed to constitute the closed-loops. Since the controller was based on model and linearization, it can adapt to both linear and nonlinear control processes. The simulation of three different trajectories for single degree of freedom movement of shoulder joint (shoulder flexion and extension, abduction and extension), and the movement of both the two degrees were given. The simulation results showed that the TLC controller can follow the exoskeleton shoulder trajectory steadily and accurately.

Iterative Learning of Optimal Control for Nonlinear Processes With Applications to Laser Additive Manufacturing

Our target application is a laser additive manufacturing (LAM) process that suffers from too wide end points. To overcome this drawback, we propose to extend an algorithm of iterative learning of optimal control (ILOC) to nonlinear systems. We demonstrate its application for a LAM process, not only by simulations but also by comparing the results of 3-D metallic printing with and without learning. Extending previously published results, we provide proof of convergence of an ILOC algorithm for a wide class of nonlinear systems.

Design and Control of Nonlinear Hybrid Tank Process using Conventional Controllers in Petrochemical Industries

In this paper the plan of customary controllers, which can be utilized for the control of non-straight frameworks to give an ideal degree of a nonlinear tank. The non linear process takes up for the examination is Hybrid tanks due to its utilization in the field of Pharmaceutical, petro chemical ect., Its non-linearity is because of the cross sectional conduct of the procedure because of the departure of items without wastage is conceivable. The shut circle execution are resolved to get the ideal level control utilizing customary P , PI, PID controllers for different systems like Ziegler Nicholas and Cohen coon technique. The real preferred position of this strategy is straightforwardness. Reproduction results utilizing MATLAB programming to decide the outperformance of controller strategies.

Sensitivity-based hierarchical distributed model predictive control of nonlinear processes

Hierarchical distributed model predictive control (HDMPC) is a promising control framework for industrial processes. With the development of efficient solvers for large-scale nonlinear programming (NLP), the implementation of hierarchical distributed control with nonlinear models becomes achievable. However, there is a lack of systematic method that handles the online computational delay in HDMPC which may lead to deterioration of performance or stability. To speed up online computation, an NLP sensitivity-based HDMPC algorithm is proposed in this paper. The implementation strategy is divided into background and online stages. All the local MPC controllers solve their sub-optimization problems iteratively in background based on the predicted future state, and a sensitivity update step is performed online to correct the predicted optimal inputs. The system-wide sensitivity equation is formulated in the upper control level by combining the optimality information of local controllers. The optimality and stability analysis for the proposed method is given. Three case studies are presented to demonstrate the controller performance.

Real-Time Optimization and Control of Nonlinear Processes Using Machine Learning

Machine learning has attracted extensive interest in the process engineering field, due to the capability of modeling complex nonlinear process behavior. This work presents a method for combining neural network models with first-principles models in real-time optimization (RTO) and model predictive control (MPC) and demonstrates the application to two chemical process examples. First, the proposed methodology that integrates a neural network model and a first-principles model in the optimization problems of RTO and MPC is discussed. Then, two chemical process examples are presented. In the first example, a continuous stirred tank reactor (CSTR) with a reversible exothermic reaction is studied. A feed-forward neural network model is used to approximate the nonlinear reaction rate and is combined with a first-principles model in RTO and MPC. An RTO is designed to find the optimal reactor operating condition balancing energy cost and reactant conversion, and an MPC is designed to drive the process to the optimal operating condition. A variation in energy price is introduced to demonstrate that the developed RTO scheme is able to minimize operation cost and yields a closed-loop performance that is very close to the one attained by RTO/MPC using the first-principles model. In the second example, a distillation column is used to demonstrate an industrial application of the use of machine learning to model nonlinearities in RTO. A feed-forward neural network is first built to obtain the phase equilibrium properties and then combined with a first-principles model in RTO, which is designed to maximize the operation profit and calculate optimal set-points for the controllers. A variation in feed concentration is introduced to demonstrate that the developed RTO scheme can increase operation profit for all considered conditions.

Neural-network-based iterative learning control of nonlinear systems

This work reports on a novel approach to effective design of iterative learning control of repetitive nonlinear processes based on artificial neural networks. The essential idea discussed here is to enhance the iterative learning scheme with neural networks applied for controller synthesis as well as for system output prediction. Consequently, an iterative control update rule is developed through efficient data-driven scheme of neural network training. The contribution of this work consists of proper characterization of the control design procedure and careful analysis of both convergence and zero error at convergence properties of the proposed nonlinear learning controller. Then, the resulting sufficient conditions can be incorporated into control update for the next process trial. The proposed approach is illustrated by two examples involving control design for pneumatic servomechanism and magnetic levitation system.

Optimal control of nonlinear fed-batch process using direct transcription method

In this paper, we investigate the direct transcription method for dynamic optimization applied to fed-batch processes. The performance index proposed for optimization has a more general form than the classical ones. As a result, more design characteristics such as the minimum time, maximum production, desired production, and minimum energy are provided. As a matter of fact, abrupt changes in the feed profiles can lead to undesirable behaviors in bioprocesses, such as decreasing the productivity. To handle the mentioned drawback, a novel idea is presented that results in achieving smooth control policies by limiting the higher-order derivatives of the feed rates. Owing to the nonlinear model of the fed-batch processes, the presence of constraints and free final time analysis, the problem is reformulated as a nonlinear programming optimization by the direct trapezoidal method. Finally, the proposed method is applied to the invertase and PHB production processes and the effectiveness of the presented approach is demonstrated by comparing with several works in the literature.

A just-in-time-learning based two-dimensional control strategy for nonlinear batch processes

In this paper, we study the problem of two-dimensional (2D) integrated model predictive iterative learning control for nonlinear batch processes. A nonlinear process is presented by using local models together with a just-in-time-learning (JITL) method. In order to deal with the problem of massive computation for identifying JITL models, a three-layer searching strategy and an updating mechanism for databases are proposed. First, a dynamic model whose parameters vary with time and batch is established for batch processes. Then, a JITL-based 2D control strategy is devised to realize comprehensive control by combining iterative learning control in batch-axis with model predictive control in time-axis. As a result, not only the closed-loop control performance can be improved, but also the optimization procedure can be simplified. Finally, performance analysis verifies the effectiveness of the proposed methods.

Construction of software using gain-schedule/auto-adaptive control strategy for automation in drilling fluid production plants

Although many advanced nonlinear process control techniques have been developed over the past decade, classic control based on feedback response still has its place. This is mostly so because feedback empirical control is robust and simple to implement and does not require fancy calculations or high-qualified operational manpower to operate it. This work has developed an application written in LabVIEW® environment capable of doing a fully automated single-input single-output control. Preliminary tests were performed in a drilling fluid production unit, controlling flow rate through manipulation of the pump power engine. In the future, tests in the same plant of pressure control by choke valves manipulation will be performed (as found in rig sites, where wellbore pressure is controlled by manipulation of such valves). The final goal is to implement such software in a real rig site, to help operators in drilling control areas such as flow rate and wellbore pressure. The produced software has embedded three self-developed features: automatic plant identification (API), auto-tuning (ABAP) and controllers auto-switch (CAS). The API determines automatically the linearity of the process determining the empirical parameters according to Sundaresan and Krishnaswamy technique. In sequence, it calculates the parameters for P, PI and PID controllers using Cohen–Coon and Ziegler–Nichols methods. The API method automates the sequence of tests necessary to implement the Sundaresan and Krishnaswamy empirical approach. The ABAP feature based on heuristic rules tunes in real time the controllers’ parameters to optimize its response. The CAS allows automatic switch between controllers and parameters to avoid instability, overshoots and creates a synergy with ABAP feature. The results have shown that the API feature is a good optimizer reducing the invested time to calculate all the parameters, from hours to a few minutes. The CAS results demonstrated an associative property with the ABAP feature to mitigate instabilities and overshoots. Therefore, the preliminary results suggested this software is a unique and important tool to improve performance, profitability and reliability during offshore and onshore drilling operations. Moreover, this application could be used in any industry with an approximate first-order dynamic system due to its robustness and a low human interaction need.

Machine learning‐based predictive control of nonlinear processes. Part I: Theory

AbstractThis article focuses on the design of model predictive control (MPC) systems for nonlinear processes that utilize an ensemble of recurrent neural network (RNN) models to predict nonlinear dynamics. Specifically, RNN models are initially developed based on a data set generated from extensive open‐loop simulations within a desired process operation region to capture process dynamics with a sufficiently small modeling error between the RNN model and the actual nonlinear process model. Subsequently, Lyapunov‐based MPC (LMPC) that utilizes RNN models as the prediction model is developed to achieve closed‐loop state boundedness and convergence to the origin. Additionally, machine learning ensemble regression modeling tools are employed in the formulation of LMPC to improve prediction accuracy of RNN models and overall closed‐loop performance while parallel computing is utilized to reduce computation time. Computational implementation of the method and application to a chemical reactor example is discussed in the second article of this series.

Machine‐learning‐based predictive control of nonlinear processes. Part II: Computational implementation

AbstractMachine learning is receiving more attention in classical engineering fields, and in particular, recurrent neural networks (RNNs) coupled with ensemble regression tools have demonstrated the capability of modeling nonlinear dynamic processes. In Part I of this two‐article series, the Lyapunov‐based model predictive control (LMPC) method using a single RNN model and an ensemble of RNN models, respectively, was rigorously developed for a general class of nonlinear systems. In the present article, computational implementation issues of this new control method ranging from training of the RNN models, ensemble regression of the RNN models, and parallel computation for accelerating the real‐time model calculations are addressed. Furthermore, a chemical reactor example is used to demonstrate the implementation and effectiveness of these machine‐learning tools in LMPC as well as compare them with standard state‐space model identification tools.