Please enter the words you want to search for:

EPE 1993 - 40 - Dialogue Session DS2.5: CONTROL: FUZZY / NEURAL / ADAPTIVE CONTROL AND MODELLING
You are here: EPE Documents > 01 - EPE & EPE ECCE Conference Proceedings > EPE 1993 - Conference > EPE 1993 - 40 - Dialogue Session DS2.5: CONTROL: FUZZY / NEURAL / ADAPTIVE CONTROL AND MODELLING
	[return to parent folder]
	SELF-TUNING SPEED CONTROLLERS FOR AUTOMATIC TUNING OF TORQUE REGULATED AC-DRIVES   By M. Tonnes; H. Rasmussen	[View]   [Download]
Abstract: This paper describes a speed-controlled AC-drive with internal torque regulation. Further, a self-tuner for the speed controller is described. A new method for torque regulation of an induction motor drive is described. The torque regulation method is based on steady-state solutions of the well-known dynamic dq-model. Test results show good torque regulation performance of the drive. A new self-tuning method is implemented in order to simplify installation and tuning of the drive. Pole placement is used as controller design method and robustness against unknown system dynamics is achieved by proper selection of filters in connection with Least Square system identification. The methods are verified through practical tests on a laboratory testrig.
	REAL-TIME SELF-ORGANISING FUZZY LOGIC CONTROLLER FOR DC SERVO   By Hung T. Nguyen; Derek Sands	[View]   [Download]
Abstract: Fuzzy logic controllers (FLC) have been used to provide solutions to control systems which are either ill-defined or too complex to model. For better control quality, self-organising fuzzy logic controllers (SOFLC) have been developed recently to include a self-learning process within the fuzzy logic framework. A major disadvantage of a self-organising fuzzy logic controller is that it is highly computational intensive. In this paper, we concentrate on the development of a real-time implementation of a self-organising fuzzy logic controller for a DC servo. In particular, we propose an efficient algorithm for implementation of the essential rule reinforcer module of the self-organising process. Experimental results show that even with limited knowledge on the system plant, the controller is able to yield good performance with strong robustness.
	SINGLE-PHASE, THREE-LEVEL, CONSTANT FREQUENCY CURRENT HYSTERESIS CONTROL FOR UPS APPLICATIONS   By C. Chiarelli; L. Malesani; S. Pirondini; P. Tomasin	[View]   [Download]
Abstract: The characteristics of hysteresis current control of voltage source inverters make this solution particularly suitable for high precision, high performance UPS, and for similar high demanding applications, like fast response drives. The extension of hysteresis techniques to three-level modulation, which is common in high-performance inverters, is usually quite critical and complex to implement. In this paper a new hysteresis control system is presented, which is capable of producing constant frequency, three-level modulation. The technique is based on the concept of decoupling the contributions to the current error of the inverter phase legs and of the load midpoint voltage. A robust, simple control is obtained. In the paper the decoupling theory is developed and the control implementation is described in detail. Experimental results are reported which confirm the excellent performance of the proposed control.
	NEURAL NETWORK BASED CURRENT REGULATOR FOR PWM INVERTERS   By M. P. Kazmierkowski; P. H. Viet; M. A. Dzieniakowski	[View]   [Download]
Abstract: This paper presents a neural network (NN) based current regulator for transistor PWM inverters. The neural network operate as a digital filter and tunes the controller gain. This NN regulator has the potential of replacing the linear PI regulator which is traditionnaly applied in the system with PWM modulators. The results of comparative study which illustrate performance of the NN regulators in comparison to the systems with PI regulator are presented. Also, learning possibilities and robustness to parameter change are discussed.
	A SERIES OF PWM METHODS OF A MULTIPLE INVERTER FOR ADJUSTABLE FREQUENCY DRIVE   By S. lto; K. Imaie; K. Nakata; S. Ueda; K. Nakamura	[View]   [Download]
Abstract: PWM methods of a multiple inverter for adjustable frequency drive have been investigated. In a multiple inverter, of which configuration inevitably includes transformers connecting output terminals of unit inverters, voltage-time product should be reduced especially in lower frequency region, otherwise magnetic saturation of the transformer cores can occur. Suppression of harmonic components is another important consideration to prevent torque ripple and electromagnetic interference. It is, however, very difficult for a single PWM method to cover the whole operational frequency region, keeping both the voltage-time product and the harmonics reduced. Here, a series of PWM methods are proposed; low, medium and high speed region PWMs. They are changed depending on the output frequency of the inverter, and reduces voltage-time product and harmonics in each operational frequency region. Simulations and experimental results using small capacity test circuit show the effectiveness of the methods.
	AN ADAPTIVE CONTROL FOR INDUCTION MOTOR DRIVES BASED ON A FULLY LINEARIZED MODEL   By A. Bellini; G. Figalli	[View]   [Download]
Abstract: The development of microprocessor-based control systems, of power semiconductors and of advanced control techniques has made it possible to obtain induction motor drives which are able to give a maximum power nearly constant in a large variation range of the speed and are characterized by a dynamic behavior much better than that of corresponding d.c. motor drives. To this end, field oriented control techniques have been used. The direct implementation of the field oriented control requires the determination of the rotor flux components, these later are generally obtained by a state observer. The availability of an observer, characterized by a reduced sensitivity to the parametric variations, has aroused interest to use also more sophisticated control strategies based on exact linearization of the machine model by a non-linear feedback. After a short description of an approach that makes it possible to obtain a fully linearized model of the induction machine, the paper inquires into the possibility to identify some drive parameters by the adaptation mechanism.
	A PRACTICAL IMPLEMENTATION OF A DSP BASED INDUCTION MOTOR DRIVE   By D. J. Atkinson; P. P. Acarnley	[View]   [Download]
Abstract: The practical implementation of a high performance induction motor drive is described. This system uses a 20kHz switching, 2kW IGBT inverter controlled by a TMS320C31 digital signal processor. In order to achieve rapid well damped torque response the drive is based on vector control principles. A digital space vector delta current controller is used to position the instantaneous stator current vector. The success of this method is due to the relatively high inverter switching rates. Both simulation and experimental results are included. Also a new current control strategy is proposed which operates with dc link current measurement only. This method uses a modified form of the above mentioned delta modulation scheme in which a redundant switching state is used to yield current vector position. Simulation results are included for this scheme.
	TECHNIQUE OF OPTMIZED PWM FOR MINIMIZING TORQUE HARMONICS   By D. Kulka; G. Klug; W. Gens; G. Berger	[View]   [Download]
Abstract: The paper presents a technique for the off-line optimization of pulse patterns applicated in controlling high performance drive systems with an induction motor fed by low switching voltage inverters. The amplitudes of torque harmonics are used as criteria for optimizing pulse patterns besides the value of stator winding losses and the amplitudes of current harmonics. Furthermore synchronous pulse patterns without 90°-symmetry in their courses are offered allowing the variation of amplitudes and phase angles of current harmonics. This is advantageous for eliminating several torque harmonics.
	NEURAL NETWORKS APPROACH TO ELECTRIC MACHINE ON-LINE DIAGNOSTICS   By F. Filippetti; G. Franceschini; C. Tassoni	[View]   [Download]
Abstract: A neural network approach to diagnosis of induction machine rotor faults is presented. The neural network substitutes the diagnostic indexes obtained through a complete model of the investigated machine. This model requires a large number of simplifying assumption in conjunction with many design parameters not easy obtainable. In this work the machine model is used to the aim of evidencing the input and the output patterns to train the neural network. The attention is focused on the conditioning of machine variables and parameters in order to select the most suitable inputs and outputs to realize a diagnostic system applicable to machine of a definite power range. The training process convergence and the evolution of the network weights put in evidence the effectiveness of the chosen inputs and outputs and the feasibility of this diagnostic procedure.
	AN APPLICATION OF GENERAL REGRESSION NEURAL NETWORK TO NONLINEAR ADAPTIVE CONTROL   By Clemens Schäffner; Dierk Schröder	[View]   [Download]
Abstract: Neural Networks have the potential to learn multivariable static mappings via the adjustment of internal weights. Therefore they are able to form a self-organizing control structure in order to handle unknown or slowly varying plant parameters and nonlinearities. In this paper it is shown that the General Regression Neural Network can be applied to a broad class of such systems. The feasibility of the approach is demonstrated with a second order plant with unknown non-linearity and unknown PT1-parameters in order to perform input-output linearization. The neural network interacts with the plant to estimate and compensate the nonlinearity and PT1-parameters. The learning scheme is fed by error signals generated by a comparison between the states of a reference model and the actual plant. It can be demonstrated that the overall system is stable in the sense of Ljapunov. The attractive features of this approach are the high speed of the GRNN implemented in parallel hardware and the ability for constant learning in a changing environment.
	INFERENTIAL PROCESSOR BASED FUZZY CONTROL FOR AN ELECTRIC DRIVE   By S. Bolognani; M. Zigliotto	[View]   [Download]
Abstract: The paper is aimed to illustrate the design and the implementation of Fuzzy Logic Controllers (FLCs) for electrical drives based on the inferential processor OMRON FP3000. The fundamentals of the FLCs are preliminarily illustrated. A description of the inferential processor and a discussion about its capabilities and limitations for this field of applications is first given. Then, the design and the implementation of two different FLCs for the speed and current control of a DC motor drive is given. Experimental results are also included in the paper.
	AN ADJUSTABLE SINUSOIDAL PWM SAMPLING PROCESS FOR A.C. DRIVES   By P. H. Mellor	[View]   [Download]
Abstract: The paper presents a new sinusoidal PWM samping process which can be adjusted to give some control over the generated waveform spectrum. It is shown how this feature allows the generation of waveforms with low harmonic distortion or selected low order harmonics to be suppressed. However, like established regular sampling techniques, the process is suitable for direct implementation in microprocessor based digital controllers. The process is compared against optimal and sub-optimal regular sampling methods and it is demonstrated that at low frequency ratios the process has superior performance to regular sampling and compares favourably with the published optimal techniques.
	A NOVEL SLIDING MODE CONTROL STRATEGY FOR A VARIABLE SPEED PERMANENT MAGNET DRIVE SYSTEM   By J. Wang; D. W. Shimmin	[View]   [Download]
Abstract: A new sliding mode controller is proposed for variable speed drive systems which incorporated high field permanent magnet motors. The hardware uses a commercially available three-phase pulse-width modulator to drive the power bridge. The controller has two conventional current-tracking loops with d-q axis decoupling, and an optimal sliding mode speed control algorithm. The sliding mode parameters are directly related to the required system performance in the time domain, simplifying the design procedure. The simulation and experimental results show that the overall system performance is entirely satisfactory.
	STATE FEEDBACK ANALOG CONTROL FOR VARIABLE FREQUENCY CONVERTERS   By A. Oukaour; B. Le Pioufle; J.-P. Louis; F. Forest	[View]   [Download]
Abstract: The presented work treats the modelling and the control of variable frequency converters. The modelling of these converters is rather delicate because of their variable framework. The variation of their states implies a global non-linear behaviour with respect to the state (output current and voltage) and to the entry (frequency). Thus we linearise the discrete model of the converter. For the analog control, a V.C.O. is needed. The behaviour of this one being non-linear at low frequencies, we propose a precise modelling of this component. To make nil the static error and to do the poles placement, we suggest an analog state feedback control with integrator. Associating to the model of the V.C.O., the converter one and the controller one we can build the closed loop transfer function and study the evolution of the poles locus, parametered by the parameters of the controller. In our case we have three degrees of freedom and we can impose the desired behaviour by mean of the resolution of a non-linear equations system.
	SOFTWARE CONTROLLERS FOR PERMANENT MAGNET SYNCHRONOUS MOTOR DRIVE   By M. F. Rahman; K. S. Low; K. W. Lim	[View]   [Download]
Abstract: High performance brushless dc (BLDC) drives based on permanent magnet synchronous motor require fast acting current/torque controllers. The key to good torque control is the ability of the current control to track the current reference signals adequately. This paper describes approaches to achieve this objective using fully digital speed and current control structures in software. It will be shown that current regulated PWM (CRPWM) inverter using the asynchronous PWM/PI regulators has a phase lag problem. An improved controller which eliminates this problem using state feedback of the back emfs in the current loop regulators is then demonstrated. Both these approaches do not regulate the d-axis dynamics when they are viewed in the synchronous reference frame. Better performance can be expected by performing the current regulation in the synchronous reference frame at the expense of more computation time. In this paper, the design and implementation issues of such a controller on a fully software platform are discussed. Experimental results of these implementations are also presented.
	OPTIMAL PWM FOR THE CONTROL OF MULTILEVEL VOLTAGE SOURCE INVERTER   By G. Carrara; D. Casini; S. Gardella; R. Salutari	[View]   [Download]
Abstract: The use of multilevel Voltage Source Inverters allows improvements in the quality of the AC side voltage waveforms as well as an increase in the handled power and DC side voltage. In such high power applications only a limited switching frequency is allowed for the PWM control of the converter, giving importance to the optimal exploitation of the few available commutations. In this paper optimal precalculated switching patterns for the control of a generalized multilevel VSI are presented. The problems addressed are the minimization of the global harmonic distortion of the output voltage waveforms as well as the elimination of selected harmonics. Possible sub-optimal simplifications to avoid discontinuities in the switching patterns as the amplitude of the fundamental varies are discussed. Quantization errors and memory requirements for the implementation of a microprocessor controller are briefly analyzed.
	DIGITAL CONTROL OF HIGH FREQUENCY PWM CONVERTERS   By P. R. Holme; C. D. Manning	[View]   [Download]
Abstract: A method is presented for the digital control of high frequency pwm converters. Digital current mode control uses samples of inductor current and output voltage to implement a control strategy in software. Average inductor current is calculated each switching cycle and compared to the current program level, providing true current mode control. This has several advantages over traditional methods. Components due to the up and down slopes of the inductor current waveform are not present in the current error signal. Current error amplifier gain is not, therefore, limited at the switching frequency in order to prevent subharmonic oscillation. Slope compensation is not required to stabilise the converter at duty ratios above 0.5 and there is no peak to average error introduced. A modular architecture is put forward for practical circuit implementation. Loop responses using the Texas Instruments TMS320C51 are discussed and features introduced by the use of digital control analysed.
	A NOVEL CURRENT CONTROL STRATEGY IN TRAPEZOIDAL EMF ACTUATORS TO MINIMIZE TORQUE RIPPLES DUE TO PHASES COMMUTATIONS   By J. Cros; J. M. Vinassa; S. Clenet; S. Astier; M. Lajoie-Mazenc	[View]   [Download]
Abstract: Rectangular current fed permanent magnet actuators with trapezoidal emf, supplied by a current controlled PWM voltage source inverter, enable to fulfil the modern large scale distribution applications with regard to cost, performances and electronics integration. However, this kind of variable speed actuator produces an important torque ripple occuring at each commutation between phases which not only depends on an emf waveshape, but also on the current control strategy. In this paper, a novel current control strategy is presented. which enables to minimize this ripple over the whole speed range and to improve the actuator performances. This control strategy is particularly simple to implement, because it only needs three phase current sensors and a rustic rotor position sensor. Simulations of the actuator working with this current control strategy by means of a global simulation software put emphasize on improvements compared to classical control strategies.
	A SINGLE PHASE, SLIDING MODE CONTROLLED INVERTER WITH THREE LEVELS OUTPUT VOLTAGE FOR UPS OR POWER CONDITIONING APPLICATIONS   By M. Carpita; P. Farina; S. Tenconi	[View]   [Download]
Abstract: In the paper, a single phase, sliding mode controlled inverter with three levels output voltage for fixed frequency AC generation is described. After a brief review of the sliding mode control theory, the way to obtain the three levels output voltage will be described from the theoretical point of view. Next a simplified approach will be shown. Simulation results are presented. In the last part of the paper the hardware implementation of the control and some experimental results relevant to a single phase, 5 KVA, 110 Vdc, 220 Vac, 50 Hz, IGBT inverter will be presented.
	THE NEW CONTROL METHODS OF CONVERTERS SYSTEMS   By A. S. Iskhakov; S. G. Obukhov; A .V. Ushakov; N. V. Dedkov	[View]   [Download]
Abstract: The new methods of converters control systems design are discussed. These methods allow to regulate not only one converter coordinate, as in the up-to-date systems, but also two coordinates - voltage and current. It gives a possibility to change an external characteristic of a power supply. The novel control systems structures allow to form an unlimited set of external characteristics of an arbitrary shape.
	A NOVEL EXCITATION SYSTEM FOR POWER PLANT GENERATORS   By L. Bojtor; B. Nagy; Gy. Rakovszky	[View]   [Download]
Abstract: A novel microcomputer controlled excitation system has been developed at Ganz Ansaldo Electric Ltd. primarily for power plant synchronous generators. The paper introduces the main characteristics of the excitation system: the redundant system configuration, the control functions and structure, the hot standby microcomputer system, the high-level man-machine interface and diagnostic tools. The prototype excitation system has been tested under power plant circumstances and the results are presented in the paper.
	DISCRETE LQ-CONTROL OF A BUCK CONVERTER   By J. A. J. M. Geerligs; G. van Rhee	[View]   [Download]
Abstract: This paper deals with the control of a buck converter. The 120 kW converter is fed by a varying 1500 V DC supply. Digital voltage as well as current control is discussed. A discrete linearized state space model is derived. A controller structure is chosen and the linearized model is extended with the controller states. For this extended model an optimal output feedback gain is computed by means of an LQ-algorithm. Digital simulation results are compared with full scale test bench measurements.