Submodule Voltage Estimation Scheme in Modular Multilevel Converters with Reduced Voltage Sensors Based on Kalman Filter Approach

Abstract

On the other hand, as with most conventional converters, the MMC has specific control requirements. For instance, the voltage-balancing of the submodule (SM) capacitors in the converter is critical. However, due to the series of cascaded SMs which are used to construct the converter, the reliability of the MMC is an important challenge [5]. In one example of a real application based on the MMC is presented in [6], where hundreds of SMs per leg were used. This also requires a great many voltage sensors in order to achieve the voltage-balancing of the system. This problem has been extensively investigated in recent research [3,7]; but many voltage sensors are always used in such studies. However, it would be preferable if this issue could be resolved with lower cost and complexity.Several proposals have been attempted to minimise the number of voltage and current sensors required. For instance, successful simulation and practical results have been achieved with fewer current sensors [8][9][10]. However, no reductions in voltage sensor number were discussed in those studies. Another attempt based on an open-loop scheme has been suggested, where fixed PWM signals were applied to the converter [11]. Although, this proposed scheme does not use any sensors since it does not require any form of feedback control, the wellknown disadvantages of open-loop control schemes may threaten the performance of the system. In some recent studies, online observers have been introduced to estimate individual SM voltages [12,13]. For example, fewer voltage sensors could be achieved based on a sliding mode where the proposed method merely monitors the total input voltage and arm current of the converter [12]. This technique has been proposed for fault detection scheme. However, variations in SM capacitance were not considered in this study. In an attempt to ameliorate this problem, the estimation of capacitance values for each individual SM was considered in [13]. This improvement provides an important solution against concerning capacitance; however, in all observer-based methods the effect of variations associated with arm inductor values was not included in the design. In very recent research [14,15], the voltage-balancing of a seven-level MMC has been achieved with important reductions in the number of voltage sensors required, where the lowest number of voltage sensors needed is two when sevenlevel MMC is used. Two sensors are contributed to measure total arm voltage of the converter. On the other hand, the main UK

Improved model predictive control for modular multilevel converter with reduced voltage sensors based on Kalman filter

“…Noise v (t i ) is also added to the measurement equation with variance matrix R (t i ) that is a positive real number R (t i ) > 0. 20 So, Equation ( 19) is rewritten as follows 15 :…”

“…By comparing Equations ( 17) and (22), the similarities between the two systems can be seen intuitively. So, Equation ( 17) is rewritten as follows 15 :…”

Denoising and Voltage Estimation in Modular Multilevel Converters Using Deep Neural-Networks

“…Several efforts have been made to reduce the number of voltage sensors in MMCs, in order to simplify the hardware requirements and increase reliability and noise immunity [6]. The Kalman filter has been proposed as an estimator for capacitor voltage in [7], but the formulation uses an extra voltage sensor per stack and its slow dynamic response may affect control performance during transients and faults. In [8], an adaptive linear neural network is proposed for SM voltage estimation, adding sensors to the stack inductors and output voltage of each phase.…”

Master–slave structure‐based capacitor voltage measuring technique for hybrid modular multilevel converters

“…In the second category, various algorithms for estimating SM capacitor voltages based on the SMs' states and the measured AC voltage are proposed [30][31][32][33]. In [30], the arm AC-side voltage is measured to estimate the SMs' voltages combined with switching signals.…”

“…To improve the accuracy, some advanced algorithms are introduced into the voltage estimation. For example, a Kalman filter (KF) [31] and an exponentially weighted recursive least square (ERLS) [32] are typically used. While using these proposed schemes, any conventional voltage balancing technique can be easily incorporated.…”