Performance analysis of ANN-based multilevel UPQC under faulty and overloading conditions

ABSTRACT

This study presents Levenberg–Marquardt back propagation (LMBP)-based artificial neural network controlled UPQC. The proposed power distribution system includes LMBP, back-to-back connected 5-level diode clamped converter (DCC), three-phase three wire distribution system, sinusoidal PWM, Hysteresis current controller and Power Angle Adjustment (PAA). Various abnormal faults like L–G, L–L–G and L–L–LG faults are considered as the major challenging issues for analysis. ANN-based PAA control is developed in this work to share reactive power between shunt and series converters without increasing the converter rating of UPQC and hence reduces the overall cost of the system. Simulation models of back-to-back connected 5-level DCC-based UPQC, Synchronous Reference Frame (SRF) and proposed ANN-based power angle control schemes are developed using SPS toolbox in Matlab/Simulink. ANN control scheme demonstrates more effective solution compared to SRF-based control, under the system subjected to different states.

KEYWORDS:

• Artificial Neural Network (ANN)/Levenberg–Marquardt back propagation (LMBP)
• diode clamped converter (DCC)
• Power Angle Adjustment (PAA)
• Total harmonic distortion (THD)
• Unified Power Quality Conditioner (UPQC)

Disclosure statement

No potential conflict of interest was reported by the authors.