Abstract

This papers presents the analysis and evaluation of the FPGA implementation of a linear adaptive model for the estimation of linearized electrical parameters. These parameters may come, for instance, from non linear models required for complex systems’ efficiency monitoring and/or closed control loop, with speed processing needs not feasible for traditional embedded systems (on the order of at least 1 mega-sample per second). The implemented model is composed by a set of equations, that have been derived from an estimation approach based on a typical Euler numeric differential equation solver. In the particular case here presented, the parameters represent the dynamic behavior of a photovoltaic generator panel. The goal of the model is to evaluate the performance of distributed maximum power tracking algorithms, from a single equation that estimates the current-voltage relations. The algorithm is evaluated first using a Register Transfer Level (RTL) Verilog description, and then is tested on a commercial FPGA with data generated from a high level golden model reference. Final validation on integrated circuits Electronic Design Automation (EDA) tools show that the design is not only feasible to be ported to a commercial CMOS technology, but that is efficient in terms of processing speed and power consumption. Such efficiency makes it adequate for the monitoring and control of interconnected solar panels.

Key words: Linear adaptive estimators; FPGA; digital arithmetic; photovoltaic systems; CMOS integrated circuits