Design and Real-Time Implementation of SHEPWM in Single-Phase Inverter Using Generalized Hopfield Neural Network

A Universal Formulation for Multilevel Selective-Harmonic-Eliminated PWM With Half-Wave Symmetry

“…As stated before, this feature is important in MVSCs where there are many different possible waveforms and many of them could not be valid to find a solution for a particular m a . The proposed system of equations can be solved with different existing search methods, such as optimization algorithms (offline technique) [6] or generalized hopfield neural networks (online technique) [39]. GAs have been selected to solve the proposed system of equations (the search algorithm is described in detail in [22]) due to their ability to find a high number of firing angles.…”

Solving nonlinear ordinary differential equations using neural networks

“…As w related to weight and bias of corresponding l initial conditions, due to usually estimation with zero until the end of neural network lear nd bias matrices of ther words, we set By integrating the output of the graph the neural network uitable number of olution for solving to zero. ws: (8) are ( ) y a A = and error for equations (9) al equations of the ons, except that in on line that has a A = . As shown in calculated by the equation as shown th two successive s and taking into ut of each of the will be zero at the and velocity of this al network will be well as parameters layers are equal to error value equal rning.…”

“…Some of the approaches are a solution in the form of an array that contains the value of the solution at a selected group of points. In other hands, other methods use basis-functions to represent the solution in analytic form and transform the original problem usually in a linear equations [8], [9].…”

Low-order harmonics control in staircase waveform useful in high-power application by a novel technique

“…Depending upon the MI, a particular set of switching angles is applied to the power electronics switches in order to obtain the desired waveform at the output. Some literature also tried to implement it in real times by adopting techniques such as artificial neural network (ANN), model predictive control, and criteria‐based control techniques . Real‐time implementation is useful especially in unequal direct current (DC) voltages such as application in solar photovoltaic as the size of lookup table is huge and will consume a lot of memory of the processor board.…”

“…Some literature also tried to implement it in real times by adopting techniques such as artificial neural network (ANN), model predictive control, and criteria-based control techniques. [28][29][30][31] Real-time implementation is useful especially in unequal direct current (DC) voltages such as application in solar photovoltaic as the size of lookup table is huge and will consume a lot of memory of the processor board. A criteria-based modulation (CBM) strategy is also proposed, which selects the optimum switching angles from a given objective function while considering all the practical constraints such as minimum ON and OFF times from the set of solution stored in the lookup table.…”