Exponential function LMS and fractional order pid based voltage power quality enhancement in distribution network

Abstract

The complexity of distribution power system becomes a menace for utility person and causing severe power quality issues. The polluted voltage power quality seriously affects the performance of critical equipment's and causing load disruptions. Dynamic voltage restorer (DVR) is recommended to address the voltage Power Quality (PQ) in distribution network using advance control scheme of Least Mean Square such as Exponential Function Least Mean Square control (EFLMS). It employs an exponential function for step size update to extract the fundamental quantity and this excels the speed of convergence, more adaptable in environments of noise and reduces the error coefficients associated with non-ideal grid. The proposed techniques derive the fundamental component using learning factor which is dependent on exponential error function and minimizes the error to estimate the reference load voltage. Secondly, the voltage ripples and fluctuation of DC and AC link are stabilized with a controller called Fractional-Order Proportional-Integral-Derivative (FOPID). The proposed FOPID have five gain values to adjust the voltage oscillations and integrated with Tyrannosaurus optimization to reduce the computation effort and achieves the superior performance over classical PI with a settle time, peak overshoot, undershoot of 0.15 s, 2.45 % and 5.7 % respectively. This auto-tuning scheme restrict the load harmonic voltage as per IEEE-519 standard and results of Tyrannosaurus proves its efficacy during fine-tuning of FOPID variables. Nonetheless, the proposed methodology was tested and validated with the MATLAB and experimental captured results and highlights the DVR compensation effectiveness under different grid voltage issues.