Gianpaolo Vitale

Current Harmonic Compensation by a Single-Phase Shunt Active Power Filter Controlled by Adaptive Neural Filtering

This paper presents a single-phase shunt active power filter (APF) for current harmonic compensation based on neural filtering. The shunt active filter, realized by a current-controlled inverter, has been used to compensate a nonlinear current load by receiving its reference from a neural adaptive notch filter. This is a recursive notch filter for the fundamental grid frequency (50 Hz) and is based on the use of a linear adaptive neuron (ADALINE). The filters parameters are made adaptive with respect to the grid frequency fluctuations. A phase-locked loop system is used to extract the fundamental component from the coupling point voltage and to estimate the actual grid frequency. The current control of the inverter has been performed by a multiresonant controller. The estimated grid frequency is fed to the neural adaptive filter and to the multiresonant controller. In this way, the inverter creates a current equal in amplitude and opposite in sign to the load harmonic current, thus producing an almost sinusoidal grid current. An automatic tuning of the multiresonant controller is implemented, which recognizes the largest three harmonics of the load current to be compensated by the APF. The stability analysis of the proposed control system is shown. The methodology has been applied in numerical simulations and experimentally to a properly devised test setup, also in comparison with the classic sinusoidal current control based on the P-Q theory.

Sensorless Control of PMSM Fractional Horsepower Drives by Signal Injection and Neural Adaptive-Band Filtering

This paper presents a sensorless technique for permanent-magnet synchronous motors (PMSMs) based on high-frequency pulsating voltage injection. Starting from a speed estimation scheme well known in the literature, this paper proposes the adoption of a neural network (NN) based adaptive variable-band filter instead of a fixed-bandwidth filter, needed for catching the speed information from the sidebands of the stator current. The proposed NN filter is based on a linear NN adaptive linear neuron (ADALINE), trained with a classic least mean squares (LMS) algorithm, and is twice adaptive. From one side, it is adaptive in the sense that its weights are adapted online recursively. From another side, its bandwidth is made adaptive during the running of the drive, acting directly on the learning rate of the NN filter itself. The immediate consequence of adopting a variable-band structure is the possibility to enlarge significantly the working speed range of the sensorless drive, which can be increased by a factor of five. The proposed observer has been tested experimentally on a fractional horsepower PMSM drive and has been compared also with a fixed-bandwidth structure.

An Improved Active Common-Mode Voltage Compensation Device for Induction Motor Drives

This paper presents a new device for the compensation of common-mode (CM) disturbance in induction motor drives, based on the active cancellation approach. The detailed design and the experimental implementation issues of the new active compensation device for a 380-V/50-Hz pulsewidth-modulation (PWM) induction motor drive are discussed. Starting from the idea of the active common-noise canceller, the proposed active compensation device is suitably improved in order to overcome the limitations of similar previously proposed circuits. In fact, it can be successfully used within a drive system with a rated voltage of 380 V or higher by employing an improved active circuit with a dedicated dc power supply derived from the ac power supply line. In addition, the design follows the criteria of compactness and minimum cost. The performance of the realized active compensation device is verified through experimental measurements of the CM voltage, the CM current, and the motor shaft voltage. The effectiveness of the proposed solution is demonstrated by experimental results.

MRAS Speed Observer for High-Performance Linear Induction Motor Drives Based on Linear Neural Networks

This paper proposes a neural network (NN) model reference adaptive system (MRAS) speed observer suited for linear induction motor (LIM) drives. The voltage and current flux models of the LIM in the stationary reference frame, taking into consideration the end effects, have been first deduced. Then, the induced part equations have been discretized and rearranged so as to be represented by a linear NN (ADALINE). On this basis, the transport layer security EXIN neuron has been used to compute online, in recursive form, the machine linear speed. The proposed NN MRAS observer has been tested experimentally on suitably developed test set-up. Its performance has been further compared to the classic MRAS and the sliding-mode MRAS speed observers developed for the rotating machines.

Real time simulation of renewable sources by model-based control of DC/DC converters

This paper presents a DC/DC buck converter circuit for real-time laboratory simulation of renewable sources. The DC/DC converter, suitably driven, can accurately describe the current-voltage characteristic of a photovoltaic (PV) array and of a fuel cell (FC). In perspective its hardware structure, if the source modelling is correctly known and implemented, can reproduce any renewable source with rated data compatible with those of the DC/DC converter. The I-V laws of the PV and the FC have been obtained by an appropriate modelling of the considered renewable sources. Particular care has been given to the design of the converter control, to ensure the desired stability and dynamics. It has been verified in numerical simulation and experimentally that the designed DC/DC buck converter is able to reproduce the electrical characteristics of the experimental generator both in steady state and transient conditions, due to either load or parameters variations. The effectiveness of the proposed circuit is verified by laboratory experiments, obtained by implementing the converter control on a low cost DSP board.

Analytical Versus Neural Real-Time Simulation of a Photovoltaic Generator Based on a DC–DC Converter

This paper presents a simulator of a PV (photovoltaic) field where the current-voltage characteristic is obtained either with a fully analytical model or with a numerical model based on a Growing Neural Gas (GNG) Network. The power stage is obtained with a DC-DC buck converter driven by the current-voltage-irradiance-temperature relation of the PV array. The improvements introduced here, respect to previous works, are the following: 1) the mathematical model is given as a continuous surface in the irradiance domain, 2) a relation between temperature and irradiance is obtained by a LSR (Leasr Square Regression) method, 3) the thermal constant of the PV field is introduced, 4) a lower number of neurons is used, 5) a better learning of the data is achieved, 6) an experimental prototype of higher rating has been devised and constructed. For both the approaches a more performing control technique of the converter has been used. Finally a PV simulator prototype is experimentally tested.

Neural MPPT of Variable-Pitch Wind Generators With Induction Machines in a Wide Wind Speed Range

This paper proposes a maximum power point tracking (MPPT) technique for variable pitch wind generators with induction machines, which can suitably be adopted in both the maximum power range and the constant power range of the wind speed. To this aim, an MPPT technique based on the Growing Neural Gas (GNG) wind turbine surface identification and corresponding function inversion has been adopted here to cover also the situation of constant rated power region. To cope with the constant power region, the blade pitch angle has been controlled on the basis of the closed-loop control of the electrical active power absorbed by the induction machine. The wind speed is then estimated in the constant power region on the basis of the actual position of the blade pitch angle. The proposed methodology has been verified both in numerical simulation and experimentally on a properly devised test set-up.

Growing Neural Gas (GNG)-Based Maximum Power Point Tracking for High-Performance Wind Generator With an Induction Machine

This paper presents a maximum power point tracking (MPPT) technique for a high-performance wind generator with induction machine based on the growing neural gas (GNG) network. Here, a GNG network has been trained offline to learn the turbine characteristic surface torque versus wind speed and machine speed. It has been implemented online to perform the inversion of this function, obtaining the wind free speed based on the estimated torque and measured machine speed. The machine reference speed is then computed by the optimal tip speed ratio. For the experimental application, a back-to-back configuration with two voltage source converters has been considered: one on the machine side and another on the grid side. Finally, two comparisons have been made. The first approach maintains the same generator structure and compares the GNG MMPT with the classic perturb-and-observe MPPT, whereas the second approach compares the squirrel-cage induction generator with a doubly fed induction generator, both integrated with the GNG MPPT. Both comparisons have been made on real wind speed profiles.

A Single-Phase DG Generation Unit with Shunt Active Power Filter Capability by Adaptive Neural Filtering

This paper deals with a single-phase distributed generation (DG) system with active power filtering (APF) capability, devised for utility current harmonic compensation. The idea is to integrate the DG unit functions with shunt APF capabilities, because the DG is connected in parallel to the grid. With the proposed approach, control of the DG unit is performed by injecting into the grid a current with the same phase and frequency of the grid voltage and with an amplitude depending on the power available from renewable sources. On the other hand, load harmonic current compensation is performed by injecting into the alternating current system harmonic currents like those of the load but with an opposite phase, thus keeping the line current almost sinusoidal. Both detection of the grid voltage fundamental and computation of the load harmonic compensation current have been performed by two neural adaptive filters with the same structure, one in a configuration ldquonotchrdquo and the other in the complementary configuration ldquoband.rdquo The ldquonotchrdquo filter has been used to compute the compensation current by eliminating only the contribution of the fundamental of the load current, whereas the ldquobandrdquo configuration is able to extract the fundamental of the coupling point voltage. Furthermore, because the active power generation and the APF features require current control of components at different frequencies, respectively, a multiresonant current controller has been adopted. The methodology has been tested successfully both in numerical simulation and experimentally on a suitably devised test setup. The stability analysis of the proposed control approach has been performed in the discrete domain.

An Optimized Feedback Common Mode Active Filter for Vehicular Induction Motor Drives

This paper proposes a common mode (CM) electromagnetic interference active filter devised for application in automotive induction motor drives. The active filter is based on a feedback scheme and it is realized using linear amplifiers. It performs the compensation of the CM voltage at the motor input, allowing an increase of the drive reliability and a reduction of the harmonic content of leakage high-frequency CM currents that affect the vehicle electromagnetic compatibility. A size-optimized layout is proposed and the influence of the power linear amplifier performance on the CM voltage compensation is discussed. General guidelines for the active filter design are given and its experimental implementation is presented. Finally, the filter performance is assessed by simulations and experimental tests.