Parameter and state estimation in power electronic circuits

Techniques for reconstructing or estimating unmeasured circuit variables and parameters in power electronic circuits are investigated. In numerous applications in power electronic systems, measurements of circuit variables are required for closed-loop control purposes or for diagnostic purposes, but are expensive or impossible to obtain directly. The method for state estimation relies on observer theory and is based on an intrinsic Lyapunov function, termed the energy in the increment. For cases where circuit parameters are not completely known, an adaptive estimation algorithm is developed for estimating the unknown circuit parameters along with the unmeasured circuit variables. The exponential stability of the combined parameter and state estimation scheme is established using averaging and singular perturbation analysis. Averaging analysis is used in off-line calculations to determine the feasibility of estimating certain parameters and to determine a basis from which to choose parameter update gains. Examples with an up-down converter illustrate the results. A hardware implementation of the state estimation scheme is also documented along with a description of a hardware implementation for parameter estimation.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

This paper introduces the concepts of two basic switching cells, P-cell and N-cell, along with their implications in power electronic circuits. The basic switching cells exist in almost every power electronic circuit. To take advantage of these structures, this paper proposes a novel packaging method for power electronics modules. The proposed packaging method uses the basic switching cells as the unit in a module, instead of traditional anti-parallel connection of active switch and diode. This rearrangement can reduce the stray inductance in the current commutation pass; therefore, the performance and reliability of the power device module and the power electronic system can be improved. A conventional phase leg module and a proposed module are modeled. Electromagnetic simulation is carried out to extract the stray inductance from the two modules. Switching behavior under different package parasitics is studied based on Saber simulation.

The usage of magnetic sensors is increasing steadily in the field of electric drives. Above all magnetoresistive (MR) sensors are experiencing a significant increase in applications in motion control and power electronics systems, not only in the industrial and automotive fields, but also in aerospace applications. MR sensors are not only used for measuring rotational and linear motion, but also for non-contact switching applications and furthermore for highly dynamic current measurement. This is largely the result of increasingly complex demands on the sensors for high performance electric drives [1]. Sensors must not only be accurate and dynamic, but must also be robust under difficult operating conditions and exhibit very high reliability. Recent developments, such as the trend to electro-mobility, are generating additional demands, with respect to compact dimensions and energy efficient operation. This combination of demands is leading to the more intensive use of magnetic sensor principles, compared to optical, inductive or capacitive principles.

The main objective of this paper is to provide a thorough analysis of currently used modulation control schemes for single-phase bidirectional dual active bridge DC/DC converters. In this article, it will be shown that single phase shift, extended phase shift and dual phase shift modulation schemes are special cases of the triple phase shift (TPS) modulation scheme. The article aims to highlight six TPS switching modes and their complements with operational constraints. Unlike previous studies that regarded TPS as a complex scheme, this paper simplifies the analysis of each mode and aims to standardize the understanding of TPS modulation for dual active bridge (DAB) converters. Power equations, range of power transferred and zero-voltage switching (ZVS) are derived for all the modes under TPS without assuming fundamental component analysis. Additionally, a generic optimization algorithm is developed to show the advantages of TPS modulation, and thus, the analysis in this paper offers a valuable insight for single-phase DAB converter designers in identifying a wide range of optimization algorithms to achieve higher efficiency under TPS modulation. This analysis contributes to the advancement of bidirectional DAB converter technology and facilitates its application in various power electronic systems.

Artificial neural network techniques are indicating a lot of promise for application in power electronic systems. So far, these applications are mainly confined to control, identification, and diagnostic problems, but the application in estimation is fairly new. The paper explores the application of neural networks for estimation of power electronic waveforms. The distorted line current waveforms in a single-phase thyristor AC controller and a three-phase diode rectifier that feeds an inverter-machine load have been taken into consideration, and neural networks have been trained to estimate the total RMS current, fundamental RMS current, displacement factor, and power factor. The performance of the neural network-based estimators has been compared with the actual values, and excellent performance is indicated. Neural network-based estimation has the usual advantages of very fast and simultaneous response of all the outputs, noise, and fault-tolerant performance and can be easily implemented in dedicated analog or digital hardware chips, which can coexist with digital signal processor (DSP) and/or application-specific integrated circuit (ASIC) chips. The estimation techniques can be extended to more complex waveforms in power electronics.

Passive filter design is traditionally considered to be very important for many power electronic circuits and power systems. It is is very critical in the design of these power electronic circuits and power systems to know how to obtain accurate corner frequencies. It is known that when the load is large, the inductance of an inductor is known to always change. This phenomenon is due to the fact that when the current load is high, the inductance always suffers from saturation or hystersis. It is extremely difficult to estimate the actual value for the nonlinear inductance in large currents because the inductance is nonlinear. The paper investigates the saturation phenomenon for a nonlinear inductance. The describing function method can be used to linearise the nonlinear inductor and then estimate the inductance in large current situations. Hence, the corner frequency for the low-pass filter can also be calculated accurately. It is shown that, when the current is very large, the corner frequency drifts to a larger value in the low-pass filter. The drift value of the corner frequency can be easily calculated by the describing function. Simulation and experimental results verify this phenomenon. However, it should be stressed that the method used in the paper is restricted to low frequencies. The higher frequency effects are neglected throughout the paper.

Tremendous surges in the application of power electronic converters and systems have led to the utilisation of many analytical techniques by which the time history of the significant variables can be predicted. This paper reviews the existing methods of analysis along with their critical appraisals in a nonmathematical format. A new method of analysis in sequency domain using Walsh function has also been indicated highlighting its salient features and advantages.

Flexible dielectrics with large dielectric constant (ε') coupled with low loss are highly pursued in many applications. To bolster the ε' of raw Zn (zinc)/poly(vinylidene fluoride, PVDF) while maintaining pimping dielectric loss, in this study, the core@double-shell structured Zn@zinc carbonate (ZnCH)@polystyrene (PS) particles are first synthesized through a suspension polymerization of styrene, and then composited with PVDF to elevate the ε' and keep low loss of the composites. By optimizing the PS shells' thickness and tailoring the electrical resistivity of Zn@ZnCH@PS particles, both the slow inter-particle polarization and fast intra-particle polarization in the composites can be decoupled and synergistically tuned, thus, the Zn@ZnCH@PS/PVDF achieves a much higher ε' and lower dielectric loss, simultaneously, which far exceed the unmodified Zn/PVDF. Both experiment and theoretic calculation reveal that the double-shell ZnCH@PS not only induces and promotes multiple polarizations enhancing the composites' ε', especially at the optimized PS's thickness, but also maintains suppressed loss and conductivity thanks to their obvious barrier effect on long-range charge migration. The core@double-shell filler design strategy facilitates the development of polymer composites with desirable dielectric properties for applications in electronic and electrical power systems.

Accurate estimation of submodule capacitance in modular multilevel converters (MMCs) is essential for optimal performance and reliability, particularly in motor drive applications such as permanent magnet synchronous motor (PMSM) drives. This paper presents a novel approach utilizing recurrent neural networks with long short-term memory (RNN–LSTM) to precisely estimate capacitance in MMC-based PMSM drives. By leveraging simulation data from MATLAB, the LSTM neural network is trained to predict capacitance based on voltage, current, and their temporal variations. The proposed LSTM architecture effectively captures the dynamic behavior of MMCs in PMSM drives, providing high-precision capacitance estimates. The results demonstrate significant improvements in estimation accuracy, validated through mean squared error (MSE) metrics and comparative analysis of actual versus estimated capacitance. The method’s robustness is further confirmed under varying operating conditions, highlighting its practical utility for real-time applications in power electronic systems.

Permanent Magnet Synchronous Generators are an emerging technology in wind power applications. A simplified Thevenin equivalent simulation model of a wind generator has been developed in the ATPDraw environment and validated by comparing its dynamics with that of a Matlab/Simulink detailed model. The simplified model includes also power electronic circuits, for interfacing the generator to a MV primary distribution grid, and their controls. A 2,5 MVA wind generator was simulated in ATPDraw in order to assess flicker severity dependence upon different grid and circuit parameters. Simulation results show a strong flicker severity dependence on grid short circuit power and X/R line characteristics, in particular when generator supplies reactive power to the grid, as required by some National standards.

The issues related to integrating these systems into the grids continue to gain importance with the increasing use and importance of renewable energy sources. Therefore, the importance of power distribution transformers is increasing. Besides, these power distribution transformers are connected to the grid with power electronics circuits and inverters. Considering the modular inverter structures, ease of maintenance, and connection, three-level T-type inverters are chosen for this study. The secondary output voltage of the power transformer is estimated by using circuit parameters such as the dead time of the inverter circuit, PWM switching frequency, and modulation rate. Based on the finite element analysis analysis according to the selected parameters, 810 data are obtained with time-dependent parametric analysis. The adaptive neuro-fuzzy inference system model is constructed by considering the simulation data to estimate the secondary output of the power transformer of these parameters. In the training phase of the model, 648 randomly selected data from 810 data obtained by ANSYS-Electronics/Simplorer are used. The remaining 162 data are used in the testing process to measure system performance. As a result of the analysis made by ANFIS, the Root Mean Square Error (RMSE) error is found as 2.475%. Since the values obtained in the estimation process of the study are very close to the simulation values, the ANFIS method can be used as an estimation method that will give accurate results during the design phase.

In this paper, an adaptive nonlinear method is proposed for both Maximum Power Point Traking Control and output voltage regulation of a Single-Ended Primary Inductance Converter (SEPIC). The main objective of this adaptive synergetic controller is to maintain a constant switching frequency of the converter and to stabilize the output voltage under uncertain weather conditions in real time. A hybrid control scheme has been derived and a Radial Basis function neural network used for approximation of unmeasurable or unmeasured variables. The effect of the controller to provide a maximum power transfer from PV side to SEPIC converter under unpredictable weather conditions and load variations has been verified through simulations. The stability of the close-loop system is insured by Lyapunov’s theory and the proposed algorithm gives good results compared to the Sliding Mode Controller used in the same context.

The paper demonstrates how electrical variables, when formulated as vector quaternions, can be manipulated. Circuit parameters can easily be obtained, for analytical purposes. Quaternions, which were discovered 170 years ago, are the basis for four dimensional vectors. Electrical variables fit into this realm, as some circuit variables are in quadrature with one another when viewed across different circuit components. The manipulation of the variables to obtain power, energy and impedance is shown to be simple and visual when vectors are looked at, in various planes of the four dimensional space. Other applications of quaternions and complex numbers in the general field of electrical engineering are also alluded to in the paper for completeness and in particular the famous and initial application of quaternions in electric field theory analysis by James Clerk Maxwell.

This paper presents pseudo-coevolutionary genetic algorithms (GA's) for power electronic circuit (PEC) optimization. Circuit parameters are optimized through two parallel co-adapted GA-based optimization processes for power conversion stage and feedback network, respectively. Each process has tunable and untunable parametric vectors. The best candidate of the tunable vector in one process is migrated into the other process as untunable vector through a migration controller, in which the migration strategy is adaptively controlled by a first-order projection of the maximum and minimum bounds of the fitness value in each generation. Implementation of this method is suitable for systems with parallel computation capacity, resulting in considerable improvement of the training speed. Optimization of a buck regulator for meeting requirements under large-signal changes and at steady state is illustrated. Simulation predictions are verified with experimental results.

A Solid-State Transformer as Enabling Technology for Microgrids for Integration into the Power Distribution System: An Overview