Output-Voltage Enhancement for A Grid-Connected Photovoltaic System Using a LUO Converter and A Nine-Level Inverter

Abstract

This paper reports a grid-connected photovoltaic conversion chain in which a DC-DC Luo converter raises and regulates the array voltage and a nine-level inverter synthesises the AC output, with a proportional-integral regulated DC link and pulse-width modulation. This study produced quantitative results, and they are treated as the substance of the paper: a simulated boost from 250 V input to 350 V output, a grid voltage oscillating between minus 200 V and plus 200 V, and a simulated total harmonic distortion of 4.9 %. Each is verified arithmetically rather than restated. The voltage ratio is shown to be 1.40, implying a duty ratio of 0.286 for a boost-type transfer function or 0.583 for the buck-boost form the Luo topology provides, so the reported pair is self-consistent with the topology but does not by itself identify which transfer function was simulated. The distortion figure is examined against the harmonic standard rather than against a remembered five per cent, and this revision develops that examination quantitatively. Evaluating the Fourier series of an unfiltered nine-level staircase shows that the same waveform yields 4.1 % when distortion is summed to the fifteenth harmonic and 8.3 % when summed to the fiftieth, which is the range the standard actually prescribes; the reported 4.9 % therefore lies within the band a limited harmonic range would produce, and cannot be interpreted as compliant or non-compliant until the range, the presence of an output filter and the switching-angle scheme are stated. The applicable limit is also shown to depend on which quantity and which voltage class is meant, being 8 % for voltage distortion at or below 1 kV and 5 % for current total demand distortion, so the reported margin of 0.1 percentage points against a remembered five per cent is not a well-defined margin at all. The principal limitation is stated plainly: all reported values are outputs of a simulation model, no hardware measurement was recorded, and the hardware figures in the record are photographs of an oscilloscope rather than logged data. Two internal inconsistencies are reported rather than silently reconciled.