João Onofre Pereira Pinto

A neural-network-based space-vector PWM controller for a three-level voltage-fed inverter induction motor drive

A neural network based implementation of space vector modulation (SVM) of a three-level voltage-fed inverter has been proposed in this paper that fully covers the linear undermodulation region. The neural network has the advantage of very fast implementation of SVM algorithm, particularly when a dedicated ASIC chip is used instead of a digital signal processor. A three-level inverter has a large number of switching states compared to a two-level inverter and therefore the SVM algorithm to be implemented in a neural network is considerably more complex. In the proposed scheme, a three-layer feedforward neural network receives the command voltage and angle information at the input and generates symmetrical PWM waves for the three phases with the help of a single timer and simple logic circuits. The ANN based modulator distributes switching states such that neutral point voltage is balanced in open loop manner. The frequency and voltage can be varied from zero to full value in the whole undermodulation range. A DSP based modulator generates the data which are used to train the network by backpropagation algorithm in MATLAB/Neural Network Toolbox. The performance of an open loop volts/Hz speed controlled induction motor drive has been evaluated with the ANN based modulator and compared with that of conventional DSP based modulator, and shows excellent performance. The modulator can be easily applied to a vector-controlled drive, and its performance can be extended to overmodulation region.

Space vector pulse width modulation of three-level inverter extending operation into overmodulation region

Multilevel voltage-fed inverters with space vector pulse width modulation have established their importance in high power high performance industrial drive applications. The paper proposes an overmodulation strategy of space vector PWM of a three-level inverter with linear transfer characteristic that easily extends from the undermodulation strategy previously developed by the authors for neural network implementation. The overmodulation strategy is very complex because of large number of inverter switching states, and hybrid in nature, that incorporates both undermodulation and overmodulation algorithms. The paper describes systematically the algorithm development, system analysis, DSP based implementation, and extensive evaluation study to validate the modulator performance. The modulator takes the command voltage and angle information at the input and generates symmetrical PWM waves for the three phases of an IGBT inverter that operates at 1.0 kHz switching frequency. The switching states are distributed such that the neutral point voltage always remains balanced. An open loop volts/Hz controlled induction motor drive has been evaluated extensively by smoothly varying the voltage and frequency in the whole speed range that covers both undermodulation and overmodulation (nearest to square-wave) regions, and performance was found to be excellent. The PWM algorithm can be easily extended to vector-controlled drive. The algorithm development is again fully compatible for implementation by a neural network.

A neural network based space vector PWM controller for voltage-fed inverter induction motor drive

A neural network based implementation of space vector modulation of a voltage-fed inverter has been proposed in this paper that fully covers the undermodulation and overmodulation regions linearly extending operation smoothly up to square wave. The neural network has the advantage of very fast implementation of SVM algorithm that can increase the converter switching frequency, particularly when a dedicated ASIC chip is used in the modulator. Two ANN-based SVM techniques have been validated: an indirect method with the help of a timer that generates the PWM waveforms from the command voltage vector at the input, and a direct method that synthesizes waveforms directly without any timer. The indirect method has been fully implemented and extensively evaluated in a volts/Hz controlled 5 hp, 60 Hz, 230 V induction motor drive. The performances of the drive with ANN-based SVM are excellent. The scheme can be easily extended to vector-controlled drive. The direct method, although has a simpler topology, needs very large training data and training time.

Adaptive Selective Harmonic Minimization Based on ANNs for Cascade Multilevel Inverters With Varying DC Sources

A new approach for modulation of an 11-level cascade multilevel inverter using selective harmonic elimination is presented in this paper. The dc sources feeding the multilevel inverter are considered to be varying in time, and the switching angles are adapted to the dc source variation. This method uses genetic algorithms to obtain switching angles offline for different dc source values. Then, artificial neural networks are used to determine the switching angles that correspond to the real-time values of the dc sources for each phase. This implies that each one of the dc sources of this topology can have different values at any time, but the output fundamental voltage will stay constant and the harmonic content will still meet the specifications. The modulating switching angles are updated at each cycle of the output fundamental voltage. This paper gives details on the method in addition to simulation and experimental results.

Recurrent-neural-network-based implementation of a programmable cascaded low-pass filter used in stator flux synthesis of vector-controlled induction motor drive

The concept of programmable cascaded low-pass filter for stator flux vector synthesis by ideal integration of stator voltages at any frequency was introduced by Bose and Patel. A new form of implementation of this filter is proposed that uses a combination of recurrent neural network trained by Kalman filter and a polynomial neural network. The proposed structure is simple, permits faster implementation by digital signal processor, and gives improved performance.

Fraud identification in electricity company customers using decision tree

The objective of this work is to develop a system that pre-select electricity energy company customers who will undergo in-site inspection for frauds or faulty measurement equipments identification. The pre-selection system was built based on the electricity company database. It used attributes such as monthly energy consumption, type of consumers, previous inspection outcome, and others. A decision tree based classification system was used to reach such goal. The identification was designed, trained and tested using MATLAB code. The fraud/faulty equipments identification per number of in-site inspection rate was 40% of the total of pre-selected customers, which was above the expectation.

Extending the constant power speed range of the brushless DC motor through dual-mode inverter control

An inverter topology and control scheme has been developed and tested to demonstrate that it can drive low-inductance, surface mounted permanent magnet motors over the wide constant power speed range (CPSR) required in electric vehicle applications. This new controller, called the dual-mode inverter controller (DMIC) , can drive both the Permanent Magnet Synchronous Machine with sinusoidal back emf, and the brushless dc machine (BDCM) with trapezoidal emf as a motor or generator. Here we concentrate on the application of the DMIC to the operation of the BDCM in the motoring mode. Simulation results, supported by closed form analytical expressions, show that the CPSR of the DMIC driven BDCM is infinite when all of the motor and inverter loss mechanisms are neglected. The expressions further show that the ratio of high-to-low motor inductances accommodated by the DMIC is 11 making the DMIC compatible with both low- and high-inductance BDCMs. Classical hysteresis-band motor current control used below base speed is integrated with DMICs phase advance above base speed. The power performance of the DMIC is then simulated across the entire speed range. Laboratory testing of a low-inductance, 7.5-hp BDCM driven by the DMIC demonstrated a CPSR above 6:1. Current peak and rms values remained controlled below rated values at all speeds. A computer simulation accurately reproduced the results of lab testing showing that the limiting CPSR of the test motor is 8:1.

Differential-Evolution-Based Optimization of the Dynamic Response for Parallel Operation of Inverters With No Controller Interconnection

In this paper, the use of differential evolution (DE), a global search technique inspired by evolutionary theory, to find the parameters that are required to achieve optimum dynamic response of parallel operation of inverters with no interconnection among the controllers is proposed. Basically, in order to reach such a goal, the system is modeled in a certain way that the slopes of P- ω and Q -V curves are the parameters to be tuned. Such parameters, when properly tuned, result in systems eigenvalues located in positions that assure the systems stability and oscillation-free dynamic response with minimum settling time. This paper describes the modeling approach and provides an overview of the motivation for the optimization and a description of the DE technique. Simulation and experimental results are also presented, and they show the viability of the proposed method.

Pulse-width optimization in a pulse density modulated high frequency AC-AC converter using genetic algorithms

As the size and the cost of power semiconductor switches are decreasing, converter topologies with high device count are starting to draw more attention. One such type of converter is the high frequency AC (HFAC) link converters. A popular control method for these converters is pulse density modulation (PDM). The HFAC link voltage of the converter in this paper is a high frequency, three-step, variable pulse-width (PW) square wave voltage waveform. A genetic algorithm approach is used to determine the PW to optimize the output voltage harmonic content.

Rainflow Algorithm-Based Lifetime Estimation of Power Semiconductors in Utility Applications

Rainflow algorithms are one of the popular counting methods used in fatigue and failure analysis in conjunction with semiconductor lifetime estimation models. However, the rainflow algorithm used in power semiconductor reliability does not consider the time-dependent mean temperature calculation. The equivalent temperature calculation proposed by Nagode et al. is applied to semiconductor lifetime estimation in this paper. A month-long arc furnace load profile is used as a test profile to estimate temperatures in insulated-gate bipolar transistors (IGBTs) in a STATCOM for reactive compensation of load. The degradation in the life of the IGBT power device is predicted based on time-dependent temperature calculation.