Weighting Factor Design in Model Predictive Control of Power Electronic Converters: An Artificial Neural Network Approach

Abstract

This paper proposes the use of an artificial neural network (ANN) for solving one of the ongoing research challenges in finite-set model predictive control (FS-MPC) of power electronics converters, i.e. the automated selection of weighting factors in cost function. The first step in this approach is to simulate a detailed converter circuit model or run experiments numerous times using different weighting factor combinations. The key performance metrics (e.g. average switching frequency (fsw) of the converter, total harmonic distortion (THD), etc.) are extracted from each simulation. This data is then used to train the ANN, which serves as a surrogate model of the converter that can provide fast and accurate estimates of the performance metrics for any weighting factor combination. Consequently, any arbitrary user-defined fitness function that combines the output metrics can be defined and the weighting factor combinations that optimize the given function can be explicitly found. The proposed methodology was verified on a practical weighting factor design problem in FS-MPC regulated voltage source converter (VSC) for uninterruptible power supply (UPS) system. Designed weighting factors for two exemplary fitness functions turned out to be robust to load variations and to yield close to expected performance when applied both to detailed simulation model (less than 3% error) and to experimental test bed (less than 10% error). Index Terms-Voltage source converter (VSC), finite set model predictive control (FS-MPC), weighing factor design, artificial neural network (ANN).

Model Predictive Control of Power Electronic Systems: Methods, Results, and Challenges

“…To avoid heuristics, there have been some methods that choose the weighting factors based on analytical expressions, thus avoiding tuning altogether [115], [151], [152]. Moreover, emerging methods for the weighting factors tuning utilize techniques from artificial intelligence, such as neural networks and genetic algorithms [143], [153]- [155]. In this way the tuning process is automated and the weighting factors can be adjusted in real time.…”

Optimal Filter Design for Power Converters Regulated by FCS-MPC in the MEA

“…derivative of the voltage reference, switching penalization) are not considered here thus there is only one WF in g t . In [33], two WFs are searched and optimized in a design space using neural network approach. Differently, λ dc is set as a design variable in the optimization.…”

High-Bandwidth Secondary Voltage and Frequency Control of VSC-Based AC Microgrid

“…where v e (i) is the prediction error, ξ lim (i) exposes the current constraint, which is shown in (21) and switching effort (SW ) with a weighting factor (ζ w ) can be expressed as given in (22). An artificial neural network (ANN) approach is presented in [41] and employed in this paper for the selection of weighting factors in the CF.…”