Adel A. Ghandakly

A Neural Network Parallel Adaptive Controller for Dynamic System Control

A neural network (NN)-based intelligent adaptive controller that introduces a new concept of intelligent supervisory loop is proposed. The scheme consists of an online radial basis-function NN (RBFNN) in parallel with a model reference adaptive controller (MRAC) and uses a growing dynamic RBFNN to augment MRAC. Updating of the RBFNN width, center, and weight characteristics is performed such that error reduction and improved tracking accuracy are accomplished. The RBFNN architecture adapts its centers and radii and tunes the relevant parameters dynamically. These adaptations effectively address the issues that are related to initial error and dimensional growth that are inherent in static NN design. The strength of the proposed scheme is in its ability to perform effectively, even when the plant mode swings and functional changes occur. Theoretical results are validated by simulation studies based on a nonlinear single-link flexible robotic manipulator position tracking of changing reference pattern. Compared to single and multiple fuzzy reference adaptive control approaches, the proposed intelligent controller produced better tracking with reduced tracking error in the event of functional changes and is capable of delivering plant output to track the reference precisely.

Design of a model reference adaptive stabilizer for the exciter and governor loops of power generators

The authors present a decentralized model reference adaptive control (DMRAC) scheme for the design of power system stabilizers (PSS) and a means for coordinating the generating unit excitation and governor control loops. In the excitation and governor control loops. In the proposed scheme, the state variables of the generating unit (GU) are to track those of an explicitly specified reference system which is designed to have desirable performance characteristics. The adaptive control law for coordinating the exciter-governor stabilizer signals is derived from a Lyapunov energy function, and thus assures system stability. Decentralized regulation and tracking tests on simulated one-machine infinite bus system show significant improvement in system performance. >

Digital Controller Design Method for Synchronous Generater Excitation and Stabilizer Systems Part I: Methodology and Computer Simulation

An effective straightforward method to design digital excitation and stabilizer controls for generating units is presented. The excitation controller is designed in the S-domain with a proportional plus integral (PI) effect and then transformed to the Z-domain using a step invariance technique. The stabilizer is also designed in the S-domain by pole-zero placing techniques and mapped to the Z-domain to obtain the digital equivalent. Both controllers are feasible for implementation by simple software and therefore suitable for on-line computer control applications. Simulation study results on a laboratory size generator are presented to show the merits of the controllers obtained. The actual hardware design and implementation study results are presented in the companion part II of this paper.

A parametrically optimized self-tuning regulator for power system stabilizers

A parametrically optimized self-tuning regulator (POSTR) is proposed for the design of digital power system stabilizers. The proposed POSTR consists of an identification scheme which identifies the nonlinear power system with a predictive model and a control technique based on parameter optimization to derive the control law. The control design consists of choosing a suitable regulator structure then turning the associated parameters. The proposed stabilizer design technique has flexibility in specifying the order and the structure of the regulator which offers advantages in selecting and autotuning well-known effective controller structures. The computational requirements of the regulator are suitable for real-time implementation. Simulation study results are presented which show that the proposed technique outperforms the fixed parameters conventional stabilizer and the adaptive minimum variance self tuner. >

An adaptive synchronous generator stabilizer design by generalized multivariable pole shifting (GMPS) technique

A robust adaptive power system stabilizer algorithm using a generalized multivariable pole shifting (GMPS) technique is presented. The method is capable of handling multiple-input-multiple-output (MIMO) systems with arbitrary numbers of inputs and outputs. The pole shifting factor is self-adjusted online such that the control constraints will not be violated. The algorithm was applied to a single-machine infinite-bus system excitation control. Both shaft speed and accelerating power were used as control inputs. A set of multifault simulation studies are presented to compare the system performance obtained by a conventional power system stabilizer. In all cases, the proposed controller demonstrates consistent superiority, and most importantly consistent reliability and robustness. >

A neural network based intelligent model reference adaptive controller

This paper presents a novel neural network based intelligent model reference adaptive controller. In this scheme the intelligent supervisory loop (ISL) is incorporated into the traditional model reference adaptive controller (MRAC) framework by utilizing an online growing dynamic radial basis function neural network (RBFNN) structure in parallel with it. The idea is to control the plant by a direct MRAC with a suitable single reference model, and at the same time respond to plant multimodal dynamics by on line tuning of an RBFNN controller. This parallel RBFNN controller is designed in order to precisely track the system output to the desired command signal trajectory, regardless of system multimodality and/or unmodeled dynamics. The updating details of the RBFNN width, centers and weights are derived to ensure error reduction and for improved tracking accuracy. The importance of the proposed scheme is in its ability to perform effectively even when the plant mode swings without using multiple model concept or a multiple reference model adaptive controller if a suitable reference model structure can be established. Further, the parallel controller will be able to precisely track the reference trajectory even with system showing unmodeled dynamics. The performance ability of the scheme is confirmed by applying to control the angular position of the robotic manipulator under tip load variations.

A Neural Network Parallel Adaptive Controller for Fighter Aircraft Pitch-Rate Tracking

A fighter aircraft pitch-rate command-tracking controller based on a neural network parallel controller is proposed. The scheme consists of an online radial basis function neural network (RBFNN) in parallel with a model reference adaptive controller (MRAC) and uses a growing dynamic RBFNN to augment MRAC. Updating the RBFNN width, the center and weight characteristics are performed such that the error reduction and improved tracking accuracy are accomplished. The RBFNN architecture adapts its centers and radii and tunes the relevant parameters, dynamically addressing the issues related to initial error and dimensional growth inherent in static neural network design. The total control signal is used to change the elevator deflection, keeping the other control surface deflections at random values, even when the aircraft operates at different maneuvers. Moreover, a suitable reference model structure is used for all aircraft operating modes, and the system is then fully tuned by the parallel controller. The strength of the proposed scheme is in its ability to effectively perform, even when plant mode swings and functional changes occur. Theoretical results are validated by conducting simulation studies on a nonlinear F16 fighter aircraft model operating at different modes created by a randomly changing parameter set.

Multiple Fuzzy Reference Model Adaptive Controller Design for Pitch-Rate Tracking

A multiple fuzzy reference model adaptive controller (MFRMAC) is proposed for nonlinear aircraft pitch-rate tracking. The controller is developed using a direct model reference adaptive control (MRAC) scheme with variable fuzzy logic reference model. The proposed controller provides a soft reference model switching for the multiple modes of operation of the aircraft without any prohibitive computation or explicit system identification. The effectiveness of the technique is assessed by simulation studies based on a 6-DOF high-performance fighter aircraft model undergoing pitch-rate control at a wide range of operating contingencies. It is demonstrated that the scheme performs extremely well when controlling multimodal dynamic systems, as opposed to conventional MRAC designs.

An intelligent paradigm for electric generator control based on supervisory loops

In this paper a new approach to a neural network based intelligent adaptive controller, which consists of an online growing dynamic Radial Basis Function Neural Network (RBFNN) structure along with a Model Reference Adaptive Control (MRAC), is proposed. RBFNN control is used to approximate the nonlinear function and the MRAC control adapts when plant parametric set changes. The adaptive laws, including neural network approximation error, are derived based on a Lyapunov function. The update details of the RBFNN width, centers, and weights are derived in order to ensure the error reduction and for improved tracking accuracy. Main advantage and uniqueness of the proposed scheme is the controllers ability to complement each other in case of parametric and functional uncertainty. Moreover, the online neural network produces a plant functional approximation control with growing and pruning nodes. The theoretical results are validated by conducting simulation studies on a single machine infinite bus (SMIB) system for electric generator control.

A model to predict current distributions in bundled cables for electric glass melters

The authors present a model for predicting the current distribution in high current cables that are constructed from bundles of parallel conductors (strands). These cables are typically used in electric glass melters to interconnect the power transformers and the melter secondary bus installations. Due to the skin effect, the magnetic field forces inside the bundled cable will tend to drive the current toward the outermost strands. The model is developed by starting from the fundamentals of the skin effect phenomenon for solid conductors and extending the concept to bundled cables. It can be easily coded in a simple computer program that can be used for ampacity prediction and sizing of bundled cables used to supply large electric glass melters or other similar heavy-current applications. Results obtained using the model for a variety of cable-bundling configurations are presented for demonstration purposes. The results show that currents in the outermost layers of bundled cables can exceed those of the innermost layers by factors of four or more, depending on the total cable size.<<ETX>>