Search

This paper presents constrained state feedback speed control of a permanent-magnet synchronous motor (PMSM). Based on classical control theory, nonlinear state-space model of PMSM is developed. A simple linearization procedure is employed to design a linear state feedback controller (SFC). Digital redesign of a SFC is carried out to achieve discrete form suitable for implementation in a DSP. Model predictive approach is used to a posteriori constraint introduction into control system. It overcomes limitations of motion control system with nonconstrained SFC resulting in low dynamic properties. The novel concept utilizes machine voltage equation model to calculate the boundary values of control signals which provide permissible values of the future state variables. Secondary control objectives such as zero $d$ -axis current are included. Simulation and experimental results are presented to validate the proposed constrained state feedback control algorithm in comparison to nonconstrained state feedback control and cascade control structure, respectively.

The paper describes a modification to the recently developed plug-in direct particle swarm repetitive controller (PDPSRC) for the sine-wave constant-amplitude constant-frequency (CACF) voltage-source inverter (VSI). The original PDPSRC algorithm assumes that the particle swarm optimizer (PSO) takes into account a performance index defined over the whole reference signal period. Each particle stores all the samples of the control signal, e.g. α = 200 samples for a controller working at 10 kHz and the reference frequency equal to 50 Hz. Therefore, the fitness landscape (i.e. the performance index) is -dimensional ( D), which makes optimization challenging. That solution can be categorized as the single-swarm one. It has been previously shown that the swarm controller does not suffer from long-term stability issues encountered in the classic iterative learning controllers (ILC). However, the convergence of the swarm has to be kept at a relatively low rate to enable successful exploitation in the D search space, which in turn results in slow responsiveness of the PDPSRC. Here a multi-swarm approach is proposed in which we divide a dynamic optimization problem (DOP) among less dimensional swarms. The reference signal period is segmented into shorter intervals and the control signal is optimized in each interval independently by separate swarms. The effectiveness of the proposed approach is illustrated with the help of numerical experiments on the CACF VSI with an output LC filter operating under nonlinear loads.

Purpose – The purpose of this paper is to provide high-efficiency and high-power hybrid energy source for an urban electric vehicle. A power management strategy based on fuzzy logic has been introduced for battery-ultracapacitor (UC) energy storage. Design/methodology/approach – The paper describes the design and construction of on-board hybrid source. The proposed energy storage system consists of battery, UCs and two DC/DC interleaved converters interfacing both storages. A fuzzy-logic controller (FLC) for the hybrid energy source is developed and discussed. Control structure has been tested using a non-mobile experimental setup. Findings – The hybrid energy storage ensures high-power ability. Flexibility and robustness offered by the FLC give an easy accessible method to provide a power management algorithm extended with additional input information from road infrastructure or other vehicles. In the presented research, it was examined that using information related to the topography of the road in the control structure helps to improve hybrid storage performance. Research limitations/implications – The proposed control algorithm is about to be validated also in an experimental car. Originality/value – Exploratory studies have been provided to investigate the benefits of energy storage hybridization for electric vehicle. Simulation and experimental results confirm that the combination of lithium batteries and UCs improves performance and reliability of the energy source. To reduce power impulses drawn from the battery, power management algorithm takes into consideration information on slope of a terrain.

In this paper two different update schemes for the recently developed plug-in direct particle swarm repetitive controller (PDPSRC) are investigated and compared. The proposed approach employs the particle swarm optimizer (PSO) to solve in on-line mode a dynamic optimization problem (DOP) related to the control task in the constant-amplitude constant-frequency voltage-source inverter (CACF VSI) with an LC output filter. The effectiveness of synchronous and asynchronous update rules, both commonly used in static optimization problems (SOPs), is assessed and compared in the case of PDPSRC. The performance of the controller, when synthesized using each of the update schemes, is studied numerically.

Two novel hybridization schemes for the particle swarm repetitive controller (PSRC) are developed and tested herein. They are designed to be used in power electronic converters, such as grid-tie converters or constant-amplitude constant-frequency (CACF) true sine wave inverters, as well as in motion control systems, especially in assembly line robots. The proposed control algorithms combine the good responsiveness of the classic deterministic iterative learning controller (ILC) with the robustness offered by the evolutionary population-based swarm controller. Properties of the control algorithms are illustrated in the CACF inverter using computer simulations. One out of the schemes is then employed in the three-phase grid-tie converter to facilitate sinusoidal currents under the distorted grid voltage conditions. A weakness not manifested in the case of the CACF inverter, however clearly apparent in the case of the grid-tie converter if envisaged to interface a superconducting magnetic energy storage (SMES), is identified. The concept of an adaptive evaporation rate for the PSRC is then investigated to enhance the performance of the original hybrid repetitive controller. This is to facilitate the controller's future deployment in the power electronic interface for an SMES system where learning should last at most a few seconds instead of tens of seconds.

This paper presents an effective pulse width modulation algorithm for four-leg three-level Neutral-Point-Clamped Voltage Source Converters. The proposed method is based on natural coordinates abc and can be used in three-or four-leg converters. The algorithm is characterized by simplicity and the lack of a coordinate system transformation. The method was verified by simulations and experiments using a laboratory setup.

The paper presents a state-feedback current controller with damped oscillatory terms designed for three-phase grid-connected voltage source converters under distorted grid voltage conditions. The paper focuses on a synthesis of an anti-windup strategy for the LQ current controller. This strategy is based on an active clamping method for the proportional-integral part and the adjusting mechanism of the damping factor (AMDF) for the oscillatory terms. The adjustment uses a simple moving average method (SMA) for smoothing changes of the damping factor. The paper shows the first experimental study of the proposed anti-windup strategy for the LQ current controller with oscillatory terms.

This paper presents an estimation procedure of equivalent electric circuit parameters of a lithium-ion cell using data from a pulse charge/discharge test. An extraction process of the model parameters is based on a transient voltage response of relaxation process following a current pulse. The entire procedure is straightforward and there is no restriction in a number of parallel RC-branches in the electrical model. Convergence problem of the parameter estimation process, existing in classic solutions, is solved by application of a particle swarm optimization method (PSO). The entire parameter identification procedure is performed in automated way in order to reduce time required to obtain a sufficient set of parameter values for different states of charge. The model is verified based on experimental data.

A 15 kVA power electronic system with a battery energy storage system is presented in the paper. The system is designed for areas where a problem with the quality of the electrical power arises. The system is intended for households and small industrial facilities and provides a variety of functionalities such as: ability of providing high quality voltage for the customers load, uninterruptible power supply in the case of power outage, control of the reactive power and low harmonic content of the grid current. An experimental study, that illustrates the functionality of the system, was performed using a full-scale physical model.

In this paper, constrained non-stationary state feedback speed controller for PMSM drive fed by VSI with SiC MOSFETs is presented and experimentally verified. The novelty of proposed approach lies in direct design of non-stationary state feedback controller for non-linear and cross-coupled model of the plant. An artificial bee colony (ABC) optimization algorithm is applied to automatic selection of non-stationary state feedback speed controller coefficients. In order to assure safety operation of the drive, constraints handling method is a posteriori added to non-stationary state feedback control algorithm. Simulation and experimental results prove that proposed control algorithm guarantees efficient operation of PMSM.

This paper presents the design and analysis process of state feedback controllers for NPC type 3-level sine wave inverter. In order to achieve good dynamic features in a state feedback control an internal model of reference input and a feedforward path are introduced. During synthesis process of the state feedback controllers maximum permissible dynamics of voltage control in the linear range of modulation is taken into account. It is shown that initially gains of the controller and feedforward path are non-stationary and depend on the angular velocity. Stationary approximation of a non-stationary state feedback controller is presented in details. Proposed state feedback control structures for continuous voltage shaping NPC type 3-level voltage source inverter are examined in FOC PMSM drive. The novelty of the presented controller lays in a stationary approximated state feedback control structure designed in terms of maximum permissible dynamics of a voltage control system. Simulation and experimental results (at the level of 3 kW) of the designed control algorithms are included. A sinusoidal output voltage waveform of the 3-level in- verter with an output LC filter can be formed in a wide range of the drive operation with the help of the additional voltage control system. Deadbeat controllers (6) or a cascade con- trol structure with classical PI controllers (11) can be applied to shape a sinusoidal output voltage waveform of the invert- er. The mathematical model of an inverter with an output LC filter is non-linear and non-stationary (12). Properties of the model described above cause that, thanks to full vector component decoupling ability, state feedback control is an at- tractive approach to control the NPC inverter with an output LC filter. Non-stationary gain matrices of the state feedback voltage controller can be obtained using pole placement tech- nique (13) or linear-quadratic optimization method (12). In both cases coefficients of the designed state feedback con- troller will be non-stationary (i.e. it depend on the angular velocity). In order to control space vector components of the filter output voltages without steady state error (in a case of step variations of the reference voltages), an internal model of the reference input (14) is added. Further improvement of the dynamic features in state feedback control is possible by introducing a feedforward path into control algorithm (13). Our primary goal is to design and examine a state-space control system for continuous voltage shaping NPC type 3- level voltage source inverter in terms of minimization PMSM torque ripple caused by a non-sinusoidal output voltage of the inverter. There are two case studies presented here: state feed- back with internal model control (SFC1), state feedback with: internal model control, an additional feedforward path and feedback from measurable disturbance (SFC2). During the synthesis process of the state feedback controllers maximum permissible dynamics of voltage control in the linear range of modulation is taken into account. Stationary approxima- tion of a non-stationary state feedback controller is presented in details. Finally, simulation and experimental results of the designed control algorithms are shown.

This paper presents the application of a particle swarm optimization (PSO) to determine iterative learning control (ILC) law gains for an inverter with an LC output filter. Available analytical tuning methods derived for a given type of ILC law are not very straightforward if additional performance requirements of the closed-loop system have to be met. These requirements usually concern the dynamics of a response to a reference signal, the dynamics of a disturbance rejection, the immunity against expected level of system and measurement noise, the robustness to anticipated variations of parameters, etc. An evolutionary optimization approach based on the swarm intelligence is proposed here. It is shown that in the case of the ILC applied to the LC filter, a cost function based on mean squares can produce satisfactory tuning effects. The efficacy of the procedure is illustrated by performing the optimization for various noise levels and various requested dynamics.

In computerized numerical control (CNC) systems, the communication bus between the controller and axis servo drives must offer high bandwidth, noise immunity, and time determinism. More and more CNC systems use real-time Ethernet protocols such as Ethernet Powerlink (EPL). Many modern controllers are closed costly hardware-based solutions. In this paper, the implementation of EPL communication bus in a PC-based CNC system is presented. The CNC system includes a PC, a software CNC controller running under Linux Real-Time Application Interface real-time operating system and servo drives communicating via EPL. The EPL stack was implemented as a real-time kernel module. Due to software-only implementation, this system is a cost-effective solution for a broad range of applications in machine control. All software systems are based on GNU General Public License or Berkeley Software Distribution licenses. Necessary modifications to the EPL stack, Linux configurations, computer basic input/output system, and motherboard configurations were presented. Experimental results of EPL communication cycle jitter on three different PCs were presented. The results confirm good performance of the presented system.

Purpose – The purpose of this paper is to discuss the design and verification of a non‐classical structure of servo‐drive controller with the state feedback and a load torque feedforward compensation.Design/methodology/approach – First a well known nonlinear mathematical model of a PMSM is transformed into a linear form by introducing new variables. The state space new model presented in rotated orthogonal reference frame is decoupled by means of equation in d and q axis. To achieve correct dynamic performance of the servo‐drive system the state feedback with an internal input model and load torque feedforward compensation is proposed. The observed load torque has been used as an input signal for the feedforward compensator. The design of the control system and simulation analysis were performed in Matlab/Simulink. The proposed control algorithm was implemented in a DSP controller (TMS320F2812). The experiments were carried out by using a 0.6 kW PMSM drive system.Findings – It is shown that the proposed c...

In this paper, auto-tuning procedure of state feedback speed controller for permanent magnet synchronous motor is presented. The novelty of the proposed approach lies in applying nature-inspired artificial bee colony optimization algorithm to obtain weighting matrices that are used for calculation of controller coefficients. In order to maintain selected state and control variables of PMSM in an acceptable level, the optimization algorithm is extended to solve constrained optimization problems. Both the main control objective (i.e. satisfactory dynamic of speed response) and secondary control goals (i.e. zero d-axis current and chattering-free q-axis control signal) are taken into account during auto-tuning process. Proper operation of proposed optimization procedure and employed performance index is proven by experimental tests of the drive.

The paper proposes a significant modification to the plug-in direct multi-swarm repetitive controller (PDMSRC). The recently developed PDMSRC for the sine wave constant-amplitude constant-frequency (CACF) voltage-source inverter (VSI) with an output LC filter serves as a case study. Initially, a purely stochastic control law was employed. Due to the nature of the particle swarm optimization algorithm used to solve the dynamic optimization problem posed by the repetitive control task in real time, the responsiveness of the controller was relatively sluggish. On the other hand, the classic P-type iterative learning control (ILC) law offers fast response. Unfortunately, the classic non-filter-augmented ILC does not offer a practical level of robustness, e.g. a repetitive disturbance can easily destabilize the system, which is not the case in PDMSRC thanks to the appropriate cost function definition. Here a hybridization is proposed to combine the advantages of both approaches. The stochastic control law is replaced with the original part-stochastic part-deterministic swarm movements by incorporating the classic ILC rule into the swarm movements. A convincing improvement in convergence ratio is illustrated with the help of numerical experiments.

The paper discusses selected modifications to a constricted, i.e. canonical, version of the particle swarm optimizer (PSO) enabling effective operation of a repetitive controller. The recently developed plug-in direct multi-swarm repetitive controller (PDMSRC) for the sine wave constant-amplitude constant-frequency (CACF) voltage-source inverter (VSI) with an output LC filter serves as a case study. Online dynamic optimization problems (DOPs), i.e. those solved without switching into some sort of offline/commissioning mode, impose restrictions that are challenging for population-based optimization techniques. In the PDMSRC samples of the control signal are directly stored in the particles. Solutions available for swarms operating in dynamic and noisy environments are briefly discussed and candidates for the PDMSRC are indicated. Geometric versus statistical and cumulative versus dimension-wise diversity measures are tested. The advantages of replacing the frequently used non-dimension-wise repulsion with a dimension-wise mechanism are demonstrated here. The effectiveness of the proposed approach is illustrated with the help of numerical experiments.

This paper presents the adaptive speed control in the permanent synchronous motor (PMSM) drive for a non-stationary repetitive process using particle swarms. The implementation of the particle swarm optimizer (PSO) for the speed controller on-line optimization in PMSM drive system is proposed. The PSO is utilized for the purpose of optimizing in real-time the proportional and integral gains of the speed PI controller for the PMSM drive system with a variable moment of inertia. The variable evaporation factor is introduced, instead of the classical evaporation constant for the PSO. The proposed PSO optimization solution is investigated with the aid of numerical simulation.

The paper presents a recently developed repetitive neurocontroller (RNC) that does not require additional filtering and/or forgetting to robustify it, i.e. to circumvent the long horizon stability issue present in the classic iterative learning control (ILC) scheme. Initially, the Levenberg-Marquardt (L-M) error backpropagation (BP) algorithm was used as a DOP(dynamic optimization problem)-capable search mechanism. At that time the choice of the training algorithm was made based on the frequently reported effectiveness of the L-M method in static optimization problems. However, there is an abundance of neural network training methods characterized, e.g., by different convergence rates, computational burden, noise sensitivity, etc. The performance of a particular optimization method is always problem specific. The case study of a constant-amplitude constant-frequency (CACF) voltage-source inverter (VSI) with an LC output filter is analysed here and some recommendations regarding the trade-off between convergence rate and computational complexity are made. The robustness to a measurement noise is also tested. The comparison is based on the results of numerical experiments. A couple of algorithms is then suggested for real-time implementation.

This paper presents a synthesis of a control system of a three-phase three-level PWM rectifier operated under unbalanced grid voltage conditions. The main goal is to obtain balanced grid currents. The proposed control method contains a PI dc-link voltage controller and an inner state feedback current controller coordinated in synchronous reference frame. The state-space model, which describes the dynamics of an input L-type filter of the rectifier, is augmented by a disturbance model that contains integral and oscillatory terms. Moreover, two approaches of reference tracking are considered. In order to determine the current controller gains, linear-quadratic (LQ) optimization is used. The discussed issues are illustrated by experimental results to validate the accuracy of the proposed control design.

The paper presents an evolutionary optimization of a novel neurocontroller for the single phase sine-wave inverter. The controller is trained in online mode. The adaptation algorithm takes into account repetitiveness of the process to be controlled. The cost function evaluates performance of the controller over the whole period of the reference signal and the weights are updated only once per period of this signal. A model-free concept is implemented and hence no prior identification of the plant is needed. The controller employs the backpropagation algorithm for updating its weights in order to adapt to changing load conditions. The controller is nonlinear and the tuning procedure involves determining a good set of values for at least three parameters: a number of neurons, an error gain and an output gain. In our recent publication devoted to the development of the control algorithm we relied on guessing and checking at the tuning stage. Here the particle swarm optimizer (PSO) is used to find the optimal set of values for these parameters. The gradientless search algorithm enables a designer to work with any user-defined performance index that reflects desired system behavior. The effectiveness of the proposed approach is illustrated with the help of numerical experiments. The controller tuned by the PSO is capable to maintain a high-quality output voltage waveform in the presence of the periodic disturbance caused by nonlinear loads.

In this paper new discrete non-stationary linear-quadratic speed regulator for PMSM fed by 3-level NPC sinusoidal inverter is presented. Controlled sinus wave inverter with an output LC filter was used in order to reduce electromagnetic torque ripple of the motor. The novelty of the proposed control system consists in one state feedback controller using in order to control the angular velocity of the motor as well as to provide true sine wave of the motor voltages. It was found that some gains of the designed discrete controller are non-stationary and depends on the angular velocity of the motor. Use of non-stationary controller causes, that linearization and decoupling process of the motor with LC filter is not needed. Simulation test results of the designed control system are compared with results obtained for PMSM fed by classical 2-level inverter. Proposed control system can be used in industrial applications where torque ripple minimization is needed instead high dynamic performance.

This paper presents simulation studies and preliminary experimental results related to a hybrid battery-ultracapacitor energy source for an urban electric car. A proposal of energy management strategy based on fuzzy logic is introduced. Control strategy for the hybrid source is designed to achieve high-efficiency and high-power source for the vehicle with a relatively small battery storage. Simulation studies show the possibility to use information from a navigation system to improve performance of the source.

In this paper we develop a new approach to combined full-state and disturbance feedforward control system synthesis for a three-phase four-leg true sine wave inverter. Proposed scheme is an extension of the augmented state feedback controller with multi-oscillatory terms. Multi-resonant controllers are well-established solution for this kind of converters if high quality output voltage is needed in the presence of nonlinear loads. They take advantage of repetitiveness of the control process under the steady load conditions (this includes nonlinear and unbalanced loads). No presumption on load type entails introduction of oscillatory terms for all three voltage components with resonant frequencies up to available bandwidth dictated by technically justified switching/sampling frequency. This makes the practical implementation challenging. Our idea is to reduce number of needed oscillatory terms by adding disturbance feedforward. In the discussed system the disturbance signal is easily accessible/measurable and hence this feedback can be very effective because of lack of estimation errors. The control system is designed in dq0 reference frame rotating with the fundamental output frequency. Extensive numerical experiments indicate that the number of oscillatory terms can be reduced significantly and high quality of output voltage can be maintained.

The paper investigates possible advantages from using nonlinear adaptive ANN (Artificial Neural Network)-based controller in a control system of autonomous variable speed electrical energy source with internal combustion engine. The speed is adjusted automatically as a function of load power demand. When the system is in a light or no load condition, the Main Voltage Controller automatically reduces the engine speed in order to reduce the fuel consumption, environmental noise and mechanical wear of engine parts. Optimization of the controller is difficult due to the non-linearity and non-stationarity of the plant.

The paper presents control system synthesis for a true sine wave four-leg inverter with an LC output filter. Our main goal is to ensure high quality voltage waveform for unbalanced and nonlinear loads. An LQR approach was chosen. The LQR design approach has proven effective in many industrial applications. Although weights in a cost function are usually set by guess and check method, the LQR in proposed system gives satisfactory results as far as some general and very intuitive rules are applied, e.g. the penalty coefficient for an integral of control error is a few orders of magnitude away from penalty coefficients related to plant state variables. In order to immunise system against nonlinear loads oscillatory terms are included in the control scheme. They can be tuned for typical harmonics like 5th, 7th, 11th and 13th. A novelty of the proposed solution lies in the fact that all controller gains (including all integral and oscillatory terms) are determined altogether in one LQR procedure call. Controls are calculated in rotating reference frame. Simulation (continuous and discrete) and experimental results (at the level of 10kW) are included. They establish consistent set of tests that validates effectiveness of the design approach.

This paper considers iterative learning control law design for bothtrial-to-trial error convergence and along the trial performance in the control system of the single-phase inverter (DC/AC). The inverter with passive LC filter act as a high quality AC power source with sinusoidal output voltage. Such solution is typically used for instance in Uninterruptible Power Supply (UPS) system. Standard feedback control cannot eliminate the periodic distortion caused by nonlinear loads. It is shown how a class of control laws can be designed using thetheory of linear repetitive processes for this problem where thecomputations are in terms of Linear Matrix Inequalities (LMIs).The simulations have been performed with the use of Simulink accompanied by PLECS library, which allows to simulate power electronic converter facilities in their realistic working conditions.

This paper describes a discrete linear quadratic speed regulator (LQR) for permanent magnet synchronous motor design. Linearization and decoupling processes of a motor model were discussed. An internal input model was introduced to control algorithm in order to eliminate the steady state speed error caused by step reference speed as well as load torque variations. Simulation tests of the drive system with the PMSM, were carried out in the Matlab-Simulink environment. The designed discrete control algorithm was implemented in a digital signal processor (TMS320F2812) and tested in a setup system. Simulation and experiment test results confirm properly operation of the designed drive system.

Paper presents a novel DC link balancing scheme for a generic n-level back-to-back system with multi-level diode clamped topologies. The proposed algorithm is generalization of the control strategy formerly introduced with the five-level back-to-back system and it relays on measurement of adjacent capacitor voltages which provide information about the potential variation in consecutive nodes of the n-level DC link network. Subsequently the adequate control signals are used to modify the multi-carrier PWM modulators on the rectifier side as well as on the inverter side respectively. In result, the voltages across all the capacitors in DC link network are effectively maintained balanced. The new balancing scheme has been demonstrated in simulations with 5-level, 7-level and 9-level single phase back-to-back systems and it can be extended to any n-level back-to-back system.

The paper presents a novel autonomous power generating system with a three-phase five-level cascaded inverter (5LCI) topology fed from a diesel engine with multiple-winding permanent magnet synchronous generator set (DE/PMSG) for UPS application. The power electronics system consists of six isolated windings which supply six full-bridge diode rectifiers with boost converters stabilizing the DC link voltages for a particular cell of the 5LCI. The five-level current mode control algorithm provides the high quality PWM waves which mitigate the high harmonic content and satisfy a robust output voltage regulation even under the unbalanced load conditions. The interfaced battery energy storage (BES), which constitutes an integral part of the entire ldquogen-setrdquo system, it supplies the DC link capacitor voltages to particular cells of the 5LCI in a standby emergency mode, when the DE/PMSG are off-line. The topology under discussion can be applied to medium or high power UPS systems.

This paper is devoted to the field of artificial intelligence for drive control. In previous works, we presented possible advantages from using an artificial neural network (ANN) for speed control in a DTC-SVM (direct torque controlled-space vector modulated) drive. Learning of the neural controller was set on-line. Starting from a random configuration of the speed controller, the network adapts its weights according to an error criterion. Although the use of such specialized controller allows potential adaptive and robust control skills, tuning of an ANN for online learning control is a long iterative procedure. Indeed, optimization of the neural controller induces determination of ten parameters acting critically on the control dynamics. However, using optimization algorithms, one can reduce efforts to reveal this set of parameters. Several optimization algorithms are based on description of biological evolutions. We call such algorithms evolutionary algorithms (EA). Genetic algorithm (GA) is a EA inspired by genetic processes leading human race toward optimal individuals capable of controlling their environment. This paper presents GA for optimization of ANN-based speed controller for induction motor drive.

In this paper, a vector-controlled three-phase voltage-source inverter with an output LC filter for AC motor drives is introduced. A control strategy of the sinusoidal inverter is based on a multivariable state feedback controller with a steady-state error compensation, and a feedforward control from disturbance and reference inputs. The controller is designed in the reference frame rotating with the synchronous angular velocity of the reference output-voltage space vector. The performance of the controller is obtained by a pole placement technique. The features of the sinusoidal inverter have been verified by experiments.

This paper presents selected methods in angular rotor speed estimation for induction motor drives. Modern control schemes for high-performance drives (e.g. Field Oriented Control, Direct Torque Control) are based on precise knowledge of some physical quantities that are hard to measure or are immeasurable in standard machine (e.g. fluxes). Other quantities needed for control require additional transducers that are expensive in comparison to overall costs of the drive or can decrease reliability when operated in hostile environment (e.g speed encoders). Thus, many researchers focus on developing various estimation algorithms to enable construction of low-cost high-performance reliable drives. The favoured estimation model type depends very much on the researcher and the process to be controlled and hence this paper does not attempt to make a value judgement as to which is best. Nevertheless, we will try to take a critical look at some aspects of selected artiflcial-neural-networks-based solutions.

Paper presents a novel topology of autonomous power generating system with a diesel engine and a permanent magnet synchronous generator (DE/PMSG) in configuration with the multi-level converters. The star-connected three-phase back-to-back topology with a five-level diode-clamped rectifier (5LDCR) and a five-level diode-clamped inverter (5LDCI) supplies a three phase output load. The proposed multi-level current mode control technique represented in dq frame and subsequently converted to abc frame provide an effective solution to voltage imbalance problem in a DC link network. The proposed balancing method is applied to "back-to-back" operational mode only. In case of operation from the interfaced battery energy storage (BES), the balancing problem does not exist because individual DC-DC converters (SU/SD) stabilize voltages across four capacitors in the DC link network. Importantly, the gen-set is able to operate with an asymmetrical, a nonlinear and an impact load in three phases and it provides the power factor correction on the generator side.

This work presents a novel method of stator flux estimation for induction motor, using artificial neural networks. Proposed estimation scheme can be employed in any vector controlled drive, e.g. DTC (direct torque control) or SFOC (stator field oriented control) drive. It does not exploit pure integration, therefore there is no problem with drift and initial conditions. Moreover, it does not require any stator resistance adaptation algorithm. A multilayer perceptron is trained off-line to approximate stator flux space vector. The approximation space is spanned by stator voltages and currents preprocessed with different low-pass filters. The performance of the SFOC drive with the above stator flux estimator has been investigated in simulation. The results seem to be very promising. Developed estimator is quite insensitive to the stator resistance variations.

In this paper, an artificial neural network, which realizes a reference model control idea, has been applied to the control system of a variable speed generating system. The speed is adjusted automatically as a function of the load power demand. The energy balance in the system is achieved by control of the DC-link voltage. The control system is based on an artificial neural network trained on-line. A neural emulator is designed (trained) off-line in order to apply the back propagation learning method. The number of neurons in the network structures as well as characteristic design parameters were specified a priori. The computer test results are presented to show performance of proposed neural controller

In this paper, we investigate the possibility to implement a novel ANN (artificial neural network)-based stator flux estimator in the DTC-SVM (direct torque controlled-space vector modulated) drive. Proposed estimation scheme does not exploit pure integration, therefore there is no problem with drift and initial conditions. Moreover, it does not require any stator resistance identification algorithm. A multilayer perceptron (MLP) is trained off-line to approximate stator flux space vector. The approximation space is spanned by stator voltages and currents preprocessed with different low-pass filters. The value of stator resistance does not occur openly in elaborated estimator, thus some level of robustness to its variations can be observed. The effectiveness of the novel approximation scheme has been confirmed in simulation as well as by experimental results. The results indicate its superiority to selected solutions based on back EMF integration (under assumption of no additional stator resistance identification algorithm). The performance of the DTC drive with ANN-based stator flux estimator has been compared to the analogous drives with PLPF (programmable low-pass filter)-based estimator and improved integration scheme with an amplitude limiter in polar coordinates. If the windings' temperature rises, there is observable speed dynamics deterioration during reversals of the latter ones, whereas there is no significant change in performance of the former one

Hygen is an adjustable-speed load-adaptive electricity generating system comprising one or more energy sources, e.g. an engine-generator unit, an energy storage means and a power electronic converter system with microprocessor control. A medium frequency permanent magnet alternator is integrated into a diesel engine to form a compact generator unit. A variable-ratio boost rectifier converts the variable voltage/frequency of the generator unit and regulates the output as required by the converter system. This combination provides a compact and efficient energy delivery sub-system with many benefits.

This paper presents a neural estimator of torque and stator flux, based on knowledge of stator current and speed. The suggested estimator is constructed from a multilayer feedforward neural network. Training sets are based on calculations using an FEM model of an induction machine. The steady-state performance of an AC motor has been assumed and the nonlinearity of the laminated core has been taken into consideration while modelling the problem. The end-region's impedance has been considered by means of modification of the rotor cage conductivity.

Light loading of generator sets has long been the cause of accelerated aging and premature engine failure. This problem is exacerbated by impact and nonlinear loads because generator sets are oversized to cope with this condition whilst maintaining the voltage within the specified limits. A new hybrid engine-friendly load-adaptive variable speed generating system has been developed which decouples the engine from the load by means of a decoupling converter. The engine load torque can be precisely controlled independently of the load so that optimal engine operation is achieved for a wide range of load power demands. Test results show outstanding performance under the worst load conditions, superior to that of comparable systems.

The paper presents the concept of a parallel active filter which can compensate the nonactive power in the supply line and can also operate as a controlled DC source to supply an electric drive. Such a system provides sinusoidal shape of input current and operates with unity power factor. Simulation studies and laboratory tests have been carried out to study the performance of the filter in two cases. In the first case the filter compensates the nonactive power in the line without any DC load connected to the filter and in the second case a DC load is also linked to it. The control diagram of the system is proposed and tested. The filter uses an AC/DC converter with a diode bridge rectifier.

This paper deals with the problem of torque ripple minimization of permanent magnet synchronous motor. The novelty of the presented approach lays in precisely maintain the level of the voltage source inverter DC voltage demanded for proper operation of the motor. An additional voltage matching circuit with state feedback controller is introduced in order to control of the inverter DC voltage. In the proposed solution model of a plant (i.e. permanent magnet synchronous motor fed by voltage source inverter with additional voltage matching circuit) is non-linear and non-stationary. An adaptive state feedback controller is developed by using an artificial neural network, which approximates non-linear control gain surfaces. A simple adaptation algorithm based on 2 low-order low-pass filters is used. Simulation results illustrate the proposed approach in comparison to typical drive with voltage source inverter and stationary state feedback controller.

This paper describes a new linear-quadratic non-stationary position regulator (LQR) for the permanent magnet synchronous motor servo-drive with sinusoidal voltage source inverter. The LC filter is used in order to reduce an electromagnetic torque ripple and to improve the servo-drive performance. Modelling process of a motor and LC filter was discussed. An internal input model was introduced to proposed control algorithm in order to eliminate the steady state position error caused by ramp reference position as well as load torque variations. Zero direct current control strategy was used in the described algorithm. Linear-quadratic optimization method was applied to determine non-stationary controller coefficients. Properties of the proposed control system have been verified in the simulation studies.

The Hygen off-grid adjustable-speed load-adaptive generating system produces standard AC 50 Hz, 230 V RMS voltage for common applications- and for very sensitive loads. A permanent magnet generator generates a variable voltage and variable frequency which is converted by an AC/DC/AC power electronic system. The system provides an uninterruptible high quality output voltage under the most adverse load conditions-large step/impact loads, nonlinear loads and arc welding machines. Experimental results produced by Volt Ampere confirm the high quality of the output voltage.

This paper presents the design of a control system of a three-phase three-level pulse-width modulated rectifier operated under unbalanced and distorted grid voltage conditions. The developed scheme is based on a voltage-oriented control method. Instead of proportional-integral current controllers a state-feedback current controller in rotating reference frame is used. In order to achieve a zero steady-state current control error in response to step commands, integral terms are incorporated into the current control scheme. Similarly, to ensure the rejection of the sinusoidal component of current control error caused by unbalance and distortion in the grid voltage, the oscillatory terms are added to the current control system. To calculate the current controller gains, an augmented plant model is developed and linear-quadratic optimization method is applied. The physical experiments indicate that using state-feedback current control with oscillatory terms allows for notable improvement of the shape of the input currents of the converter and increases the power conversion efficiency.