Fixed frequency sliding mode-based robust inversion with a full-bridge current DC-link buck-boost

https://doi.org/10.1016/j.jfranklin.2013.08.009Get rights and content

Abstract

The substitution of the original switches by a full bridge in a Non-Inverting Buck-Boost converter results in an inverter capable of carrying out step-down and step-up tasks as well under sliding mode control. The control law is implemented by means of the Zero Average Dynamics algorithm, which provides a fixed frequency operation and guarantees null error in each switching period, thus achieving a highly accurate tracking of periodic reference profiles. Furthermore, semi-infinite programming techniques are used to reduce power losses and, at the same time, prevent undesirable effects of control action saturation. The performance of the inverter is ensured to be robust in the face of bounded nonlinear and resistive loads. Realistic simulation results obtained with PSIMR software validate the proposed schemes.

Introduction

Full-Bridge Buck-based power supply equipments require a transformer in order to perform step-up tasks (see, for example, [1]), thus resulting in a significant increase of weight and size of the source inverter. This fact has opened a research branch that assesses the possibility of replacing the former structure by a transformerless DC–AC converter, derived from a nonlinear switched DC–DC converter and capable of carrying out step-down and step-up inversion as well.

One of the main challenges in this field deals with the non-minimum phase character exhibited by nonlinear switch-mode DC–DC converters. Direct control of the output voltage by means of passivity-based schemes [2] and PID-type sliding mode controllers [3] has been successfully used for regulation purposes but, as far as these authors know, no results on tracking tasks have been reported. Alternatively, serious attempts to overcome the problem trying current-based indirect control strategies in Boost and Buck-Boost DC–AC inverters [4], [5] cannot avoid sensitivity to external perturbations and parameter uncertainties. PI controllers also offer interesting performance in full-bridge nonlinear inverters [6], [7], [8], [9]; however, it is well known that PI control designs are based on small signal models obtained for nominal parameters, which include a nominal resistive output load. This leads to output waveforms sensitive to power stage parameter variations, such as the output load. The reader has to bear in mind the high variability of the load in case of connection to nonlinear elements as, for example, a full-wave rectifier.

The Full-Bridge Current DC-Link Buck-Boost (FBBB) inverter, which is essentially constituted by a full-bridge inverter in series with a Boost converter, has two control inputs. Hence, it is possible to design control strategies that yield robust tracking of periodic signals by the output voltage and, at the same time, keep the input current regulated at a prescribed level [10].

Sliding mode control [11], [12] is here shown to render an efficient performance of the FBBB in inversion tasks. The proposed scheme uses an error-based switching surface that does not depend on the plant parameters. This, combined with the fact that all disturbances are matched, guarantees robustness in the face of bounded nonlinear and resistive loads. Moreover, a decoupling of the control actions, which allows a fixed-frequency implementation of the sliding mode control law by means of the Zero Average Dynamics (ZAD) algorithm, maintains the robustness properties inherent to sliding mode control and renders null mean tracking error for the output voltage. As a result, highly accurate robust tracking performance of periodic references is achieved. This technique has already shown efficient experimental results in a power inverter. A complete survey about the ZAD algorithm including both theoretical and experimental considerations, together with a comparative study between ZAD and other control implementation techniques, may be found in [13]. Finally, restrictions for candidate signals to be tracked are derived demanding non-saturation of the control action – which has fixed control gains – in the steady state.

It is already known that proper energy transfer constitutes the main goal of power converters, this meaning good efficiency and high output signal quality. Maximizing power efficiency requires minimization of the Root Mean Square (RMS) of the current flowing in the switching converter, this leading to two different effects, namely, optimization of the losses due to the power switching and minimization of the resistive losses in the inductors. Hence, the design incorporates a procedure to reduce power losses that has been successfully applied to the robust tracking of offset sinusoidal signals in a Non-Inverting Buck-Boost under bounded resistive load variation [14]. This technique, which is based on semi-infinite programming theory [15], is here adapted to deal with a family of nonlinear currents that is assumed to be bounded by a known resistive load current. Eventually, the resulting method generalizes that of [14] by relaxing a continuous differentiability condition to Lipschitz continuity and it is also applicable to linear loads.

The proposal is validated by means of illustrative simulations with the realistic power electronics software package PSIMR.

This paper is structured as follows. The mathematical model of the FBBB inverter is established in Section 2. A sliding mode control strategy to achieve the output voltage tracking target is developed in Section 3. The selection of a current reference profile that reduces power losses is studied in Section 4. Simulation results are presented in Section 5, while Conclusions are in Section 6. Appendix A contains the proofs of some of the results stated in Section 4.

Section snippets

The full-bridge current DC-link buck-boost inverter

The FBBB inverter, depicted in Fig. 1, is derived by replacing each one of the original switches of a Non-Inverting Buck-Boost converter [16], [17] by a full bridge of switches allowing bi-directional current flows.

Its dynamical behavior can be modelled by means of a two-dimensional, bilinear system with the inductor current iL and the capacitor voltage vC as state variablesLdiLdτ=VABriL(1q)vCCdvCdτ=iR+(1q)iL.The variable iR stands for the load current. The control actions envisage the

Sliding mode control of the full-bridge current DC-link buck-boost inverter

The sliding surface used in [10] for system (5), (6) requires a somehow complicated variable switching frequency implementation due to the coupling suffered by the control actions. However, a fixed frequency implementation is possible with a selection of switching surfaces that result in a decoupling of the control actions.

With this aim, consider the switching surfaces candidatesσ1(x1,x2,t)12(x12+x22)12[x1d2(t)+x2d2(t)],σ2(x1,x2,t)x2x2d(t).

Proposition 1

Let Assumption A be fulfilled. Then, the control law

Power loss reduction

According to Remark 3 and to the definition of the unsaturation region in Eqs. (10), (11), the selection of a constant and high enough inductor current reference signal x1d may suffice to guarantee the desired performance. However, due to both technical and economical reasons, it is extremely convenient to reduce as much as possible power losses in the converters. These losses are of two types, namely switching losses and conduction losses; in this article we focus on the second ones.

The power

Simulation results

The power electronics software PSIMR is used to carry out the simulations, the numerical integration time step being 2.5×106. The reader is referred to the PSIMR User's Manual for further details.

The parameters of the FBBB inverter simulation prototype are: a DC source of Vg=50V, an inductance of L=100μH with an internal resistance of 0.1Ω and a capacitor of C=220μF. The inverter also has two switches: the first one is a unidirectional voltage and bidirectional current full-bridge, and has

Conclusions and suggestions for further research

This article presents an inverter obtained from the replacement of the original switches by a full bridge in a Non-Inverting Buck-Boost converter. Operating under a sliding regime that yields a control action decoupling, the converter performance involves step-down and step-up tasks which are robust in the face of nonlinear and resistive loads with known bounds. The control law is implemented by means of the ZAD algorithm, which provides a fixed operation frequency and guarantees null error in

Acknowledgment

This work has been partially supported by the spanish Ministerio de Educación under projects DPI2009-14713-C03-03 and DPI2010-15110.

References (19)

There are more references available in the full text version of this article.

Cited by (6)

  • Affine discretization methods for the digital resonant control of uninterruptible power supplies

    2019, Journal of the Franklin Institute
    Citation Excerpt :

    In this context, usual performance parameters are individual harmonic distortion (IHD), total harmonic distortion (THD) and recovering time, with the admissible values imposed by standards such as the IEC 62040-3 [1]. In the last decades, several strategies have been developed to the pulse width modulation (PWM) control of DC-DC and DC-AC converters, e.g. proportional-integral-derivative (PID) controllers [2] or nonlinear techniques, such as sliding-mode [3,4] and model predictive control [5]. More recently, authors considered active disturbance rejection control (ADRC) and type-2 fuzzy logic techniques [6,7] to improve the overall system performance under time-varying parameter or uncertainties.

  • Optimized sliding mode current controller for power converters with non-minimum phase nature

    2019, Journal of the Franklin Institute
    Citation Excerpt :

    In addition, the boost and buck-boost converters cannot be directly controlled by using a single loop compensator processing the output voltage error, because of the presence of a right half-plane zero (RHPZ) in their model. So far, classical PI controllers [15,16], current mode regulators [17], intelligent adaptive control scheme [18], passivity-based control strategy [19], adaptive backstepping technique [20], fuzzy logic controllers [21], output-feedback model predictive controller [22], neural network controller [23], repetitive method with phase-lead compensation [24], robust H-infinity controller [25], fixed frequency sliding mode controller [26–28] have been reported for DC/DC converters. In the case of the continuous control approaches, the performance of the relevant controllers is generally degraded, due to use of the pulse width modulation (PWM) technique.

  • Robust sliding mode control of a DC/DC Boost converter with switching frequency regulation

    2018, Journal of the Franklin Institute
    Citation Excerpt :

    In this case, such variable switching frequency implies an inconvenient from the power converter point of view, as it was early discussed. Different approaches have been proposed to overcome this problem ([8–13], among others). The reader is referred to [14] for an analysis of the main benefits and drawbacks of several SMC implementation proposals.

  • Adjustable PID control based on adaptive internal model and application in current shared control of multi inverters

    2017, Journal of the Franklin Institute
    Citation Excerpt :

    With the depletion of traditional energy and its serious pollution to the environment, the micro-grid composed of renewable energy gets more and more attention for its large reserves, and the economic and environmental advantages [1–4]. Distributed generation is the main way in micro-grid [5–10]. The inverters connected to the grid are the interface between the distributed micro power sources and the utility grid.

  • Model Predictive and Sliding Mode Controllers for the DC-DC Boost Converter

    2022, 2022 IEEE 2nd International Conference on Sustainable Energy and Future Electric Transportation, SeFeT 2022
  • Study of Constraint SMC Low-Frequency Full-Bridge Inverter

    2018, Tien Tzu Hsueh Pao/Acta Electronica Sinica
View full text