|RER Renewable Energy Research||français|
Research conducted on behalf of the Photovoltaics and Hybrid Systems Program at the CETC-Varennes (Natural Resources Canada) with partial funding from the Panel on Energy Research and Development (PERD).
Ross, Michael M. D. Hybrid Cycling with PV and Genset: Simulation Results. Report to CETC-Varennes (Natural Resources Canada). Montreal, Qc: RER Renewable Energy Research, 2005.
This document examines the simulated operation of a specific PV hybrid system, that being the hybrid system test bench of the Photovoltaic and Hybrid Systems Group at the CANMET Energy Technology Centre—Varennes. During a five-week test run on this system in September and October, 2004, a photovoltaic array and a genset provided power to a DC load, with energy surpluses and deficits being accommodated by a battery. Monitored data from this test were compared with simulations, run on the PV Toolbox tool developed by the CANMET Energy Technology Centre—Varennes. Following this validation, several interesting aspects of system operation, noticed in analysis of the monitored data from the test, were investigated using PVToolbox. This permitted the system operation over the whole year to be examined, and demonstrated the complementary nature of simulation and the hybrid test bench.
The accuracy of the PV Toolbox system simulations far exceeded expectations. The design team had aimed to estimate total energy flows within 10% accuracy; in this test, when using plane-of-array irradiance input data, energy flows were often within 1or 2%, and there was a very close correspondence between the time of genset operation in simulation and reality. The mean bias error in the battery voltage was around 10 mV per cell, and the root mean squared error was only 40 mV per cell. Coulombic efficiency and round-trip battery efficiency were accurate to within 1 to 4 percentage points. Admittedly, these simulations assumed a 1152 Wpeak array (instead of the 1200 Wpeak measured with an IV curve tracer after the study had been completed) and an array wiring resistance too low by a factor of 60%, but this is nonetheless impressive given the presence of a charge controller and the non-linearity of genset dispatch. Accuracy deteriorated only slightly (and improved for certain key variables) when horizontal ir
radiance data were used in place of plane-of-array data. It is argued that the exemplary accuracy of the simulation was due in part to the constant (as opposed to varying) load, the use of a time criterion (as opposed to a voltage or current threshold) for the termination of genset operation, and the high solar fraction, which avoided extended periods of cycling between partial states-of-charge. This suggests that further comparisons should be conducted with test bench runs using varying loads, voltage or current thresholds, and lower solar fractions.
The PV Toolbox was used to estimate the plane-of-array irradiance based on the measured global irradiance on the horizontal, and this was compared with the measured plane-of-array irradiance. The agreement was excellent: the error in PV Toolbox’s estimate (about 6%) was comparable to the uncertainty in the measurement. This excellent result masked inaccuracy in the algorithm for splitting global irradiation into beam and diffuse components, however. In other locations and situations, such as at high latitudes, this underlying problem might seriously affect the accuracy of the plane-of-array irradiance estimate.
Simulation studies suggest that 1) maximum power point tracking offers few advantages over clamped operation for hybrid power systems, 2) charge control losses are small when solar fractions are considerably below unity, and therefore dispatch should not focus on minimizing them, and 3) battery charge currents higher than recommended by most manufacturers could significantly reduce genset fuel consumption in many hybrid systems. These conclusions are not firm, however, and require further investigation.
Created 2007/03/06 Updated 2007/03/06 ©2007 RER Renewable Energy Research