Heidar A. Malki

New design and stability analysis of fuzzy proportional-derivative control systems

This paper describes the design principle, tracking performance, and stability analysis of a fuzzy proportional-derivative (PD) controller. First, the fuzzy PD controller is derived from the conventional continuous-time linear PD controller. Then, the fuzzification, control-rule base, and defuzzification in the design of the fuzzy PD controller are discussed in detail. The resulting controller is a discrete-time fuzzy version of the conventional PD controller, which has the same linear structure in the proportional and the derivative parts but has nonconstant gains: both the proportional and derivative gains are nonlinear functions of the input signals. The new fuzzy PD controller thus preserves the simple linear structure of the conventional PD controller yet enhances its self-tuning control capability. Computer simulation results have demonstrated this advantage of the fuzzy PD controller, particularly when the process to be controlled is nonlinear. After a detailed stability analysis, where a simple and realistic sufficient condition for the bounded-input/bounded-output stability of the overall feedback control system was derived, several computer simulation results are compared with the conventional PD controller. Although the conventional and fuzzy PD controllers are not exactly comparable, the authors compare them in order to have a sense of how well the fuzzy PD controller performs. For this reason, in the simulations several first-order and second-order linear systems, with or without time-delays, are first used to test the performance of the fuzzy PD controller for step reference inputs: the fuzzy PD control systems show remarkable performance, as well as (if not better than) the conventional PD control systems. Moreover, the fuzzy PD controller is compared to the conventional PD controller for a particular second-order linear system, showing the advantage of the fuzzy PD controller over the conventional one in the sense that in order to obtain the same control performance the conventional PD controller has to employ an extremely large gain while the fuzzy controller uses a reasonably small gain. Finally, in the case of nonlinear systems, the authors provide some examples to show that the fuzzy PD controller can track the set-points satisfactorily but the conventional PD controller cannot. >

Design and analysis of a fuzzy proportional-integral-derivative controller

Abstract This paper describes the design principle, tracking performance and stability analysis of a fuzzy proportional-integral (PI) plus a derivative (D) controller. First, the fuzzy PI+D controller is derived from the conventional continuous-time linear PI+D controller. Then, the fuzzification, control-rule base, and defuzzification in the design of the fuzzy controller are discussed in detail. The resulting controller is a discrete-time fuzzy version of the conventional PI+D controller, which has the same linear structure in the proportional, integral and derivative parts but has nonconstant gains: the proportional, integral and derivative gains are nonlinear functions of the input signals. The new fuzzy PI+D controller thus preserves the simple linear structure of its conventional counterpart yet enhances the self-tuning control capability. Computer simulation results have demonstrated the advantages of the fuzzy controller, particularly when the process to be controlled is nonlinear. After a brief stability analysis, where a simple and realistic sufficient condition for the bounded-input/bounded-output stability of the overall feedback control system was derived, several computer simulation results are shown to compare with the conventional PI+D controller. Computer simulation results have shown the new fuzzy controller indeed has satisfactory tracking performance.

Fuzzy PID control of a flexible-joint robot arm with uncertainties from time-varying loads

This paper presents the design and experiment of a fuzzy proportional integral derivative (PID) controller for a flexible-joint robot arm with uncertainties from time-varying loads. Experimental results have shown remarkable tracking performance of this fuzzy PID controller, and have convincingly demonstrated that fuzzy logic control can be used for flexible-joint robot arms with uncertainties and it is quite robust. In this paper, the fuzzy PID controller is first described briefly, using a simple and practical PD+I controller configuration. This configuration preserves the linear structure of the conventional PD+I controller, but has nonconstant gains: the proportional, integral, and derivative gains are nonlinear functions of their input signals, which have self-tuning (adaptive) capabilities in set-point tracking performance. Moreover, these variable gains make the fuzzy PID controller robust with faster response time and less overshoot than its conventional counterpart. The proposed design was tested using a flexible-joint robot arm driven by a DC motor in a laboratory, where the arm was experienced with time-varying loads. Control performance by the conventional and fuzzy PID controllers for such a laboratory robotic system are both included in this paper for comparison.

Network Intrusion Detection System Using Neural Networks

This paper presents a neural network-based intrusion detection method for the internet-based attacks on a computer network. Intrusion detection systems (IDS) have been created to predict and thwart current and future attacks. Neural networks are used to identify and predict unusual activities in the system. In particular, feedforward neural networks with the back propagation training algorithm were employed in this study. Training and testing data were obtained from the Defense Advanced Research Projects Agency (DARPA) intrusion detection evaluation data sets. The experimental results on real-data showed promising results on detection intrusion systems using neural networks.

Predicting contaminant removal during municipal drinking water nanofiltration using artificial neural networks

Abstract An artificial neural network model for steady-state contaminant removal during nanofiltration of ground and surface waters under conditions typical of drinking water treatment is derived and validated. Operating conditions such as flux, feed water recovery, and element recovery (surrogate for cross-flow velocity), and feed water quality parameters including pH, total dissolved solids concentration (surrogate for ionic strength), target contaminant concentration, and where possible the diffusion coefficient were used as inputs to predict the ratio of permeate to feed concentration of the target contaminant. Contaminants reported herein include dissolved organic carbon, precursors to total organic halide, four trihalomethanes and nine haloacetic acids containing chlorine and bromine, hardness, alkalinity, and total dissolved solids. Additionally, source waters from seven different locations and two commercial thin-film composite membranes operating in a wide range of permeate fluxes and feed water recoveries were considered. Deterministic and pseudostochastic simulations showed that artificial neural networks closely predicted permeate concentrations of each one of these organic and inorganic contaminants. Therefore, neural networks can be used to circumvent difficulties associated with formulating and solving the highly non-linear Nernst–Planck equation to calculate solute removal from multi-component solutions at high recovery. Moreover, neural networks can predict the transport of heterogeneous and difficult to characterize water treatment contaminants such as natural organic matter and disinfection by-product precursors, whose physicochemical properties are unknown. Such models can be used to screen membranes prior to conducting expensive large-scale tests as well as in the better design and interpretation of data obtained from site-specific water treatment nanofiltration studies conducted in support of plant design.

Power flow management of microgrid networks using model predictive control

In this paper, we present a power flow management method for a network of cooperating microgrids within the context of a smart grid by formulating the problem in a model predictive control framework. In order to reliably and economically provide the required power to the costumers, the proposed method enables the network of microgrids to share the power generated from their renewable energy sources and minimize the power needed from the micro gas turbines. To corroborate the viability of the proposed method, we will illustrate simulation results on a model consisting of three microgrids in a network.

Stochastic model predictive control method for microgrid management

This paper presents a stochastic model predictive control method for managing a microgrid. In order to reliably provide the required power for costumers, the proposed method enables the microgrid to use the renewable energy sources as much as possible while keeping the storage device to its maximum state of charge and minimizing the power generated by the micro gas turbine. The performance and effectiveness of the proposed method will be finally illustrated by simulating a microgrid model consisting of three nodes including a renewable generation source and a battery, customers, and a micro gas turbine.

Fuzzy variable structure control for MIMO systems

This paper presents a fuzzy variable structure controller for MIMO systems. Motivated by the principle of variable structure control strategy, a systematic design procedure is proposed. The proposed fuzzy controller is constructed to approximate the unknown system and to reduce the chattering phenomenon. The controller parameters can also be adapted online to utilize the control energy more efficiently. The simulation results show that the proposed control algorithm exhibits strong robustness against model uncertainties and external disturbances.

Hardware Implementation for a Genetic Algorithm

A genetic algorithm (GA) can find an optimal solution in many complex problems. GAs have been widely used in many applications. A flexible-very-large-scale integration intellectual property for the GA has been proposed in this paper. This algorithm can dynamically perform various population sizes, fitness lengths, individual lengths, fitness functions, crossover operations, and mutation-rate settings to meet the real-time requirements of various GA applications. It can be seen from the simulation results that our design works very well for the three examples running at an 83-MHz clock frequency.

Evaluation of cosine radial basis function neural networks on electric power load forecasting

This paper presents the results of a study aimed at the development of a system for short-term electric power load forecasting. This was attempted by training feedforward neural networks (FFNNs) and cosine radial basis function (RBF) neural networks to predict future power demand based on past power load data and weather conditions. This comparison indicates that both neural network models exhibit comparable performance when tested on the training data but cosine RBF neural networks generalize better since they outperform considerably FFNNs on the testing data.