Direct-sequence maximum power-point tracker for photovoltaic sources

A digital gradient-ascent algorithm is presented to identify and maintain the peak power operating point of a photovoltaic (PV) source with variable I-V characteristics. The approach employs a low-level dither, realized by a one-bit, 64-sample pseudorandom noise (PN) sequence, to perturb the duty cycle of a boost converter that extracts energy from a PV source for battery charging. The digital-domain optimization process operates continuously in the background, and is robust against measurement noise, offset, jitter, and the inherent large-signal nonlinear dynamics of the boost converter. Acquiring a single sample in each switching cycle, i.e., no oversampling for the analog-to-digital converter (ADC), the digital processor consists of a few adders, flip-flops, and one multiplier, which, in conjunction with the ADC, can be integrated with the driver integrated circuit (IC) of the boost converter for very low cost. Simulation verifies the tracking effectiveness of the proposed technique, and demonstrates a stable operation in presence of large power-on and load transients, with an average output power of ¿98.5% of the peak value consistently achieved in steady state.