Comparison of Two Techniques for the Simulation of PV Systems

摘要

For several years, MIT Lincoln Laboratory has conducted computer simulations of the performance of photovoltaic solar energy systems in order to size system components, to define designs of potential economic feasibility, to test various control schemes, and to monitor the performance of working systems in the field. When used as an aid-to-design, these hourly simulations step through a full year's worth of insolation and weather data at a specific geographical site. These data are available on computer tapes in the SOLMET format from the National Climatic Center. More recently, a simulation technique has been developed that does not require marching through time but instead works with probability-density functions of daily values of insolation and load as inputs while still providing estimates of the usual measures of system performance (e.g., auxiliary energy required, surplus energy thrown away, fraction of load displaced). Results obtained compare well with results previously obtained from an hourly simulation of a daytime radio station. This technique may be used to study the effect on system performance of varying degrees of correlation of load with insolation and to test the sensitivity of economic analyses to variations in utility escalation rate (discounted for inflation), PV module and balance-of-system costs. (ERA citation 05:015222)