Aron Dobos

This work presents a novel analytical approximation of the effect of inter-row shading on large photovoltaic (PV) arrays. Computation time is reduced orders of magnitude relative to full numerical simulations, allowing this method to be used in annual production estimation software – for instance, it is currently implemented in the National Renewable Energy Laboratory’s System Advisor Model program. A further advantage of the analytical approach is that, unlike numerical simulations… Show more

This work presents a novel analytical approximation of the effect of inter-row shading on large photovoltaic (PV) arrays. Computation time is reduced orders of magnitude relative to full numerical simulations, allowing this method to be used in annual production estimation software – for instance, it is currently implemented in the National Renewable Energy Laboratory’s System Advisor Model program. A further advantage of the analytical approach is that, unlike numerical simulations, computation time does not increase with the size of the PV installation. Comparisons with full I–V curve simulations indicate that this simplified approach has typical error of 1% over multiple module fill factor, shade extent and shade opacity assumptions. Maximum error of 2-6% was found for simulations of crystalline silicon modules. Comparisons with experimental results show good agreement between the experiment and the model over a range of operating conditions, and intercomparison with prior modeling methods indicates a spread of possible model results, depending on model assumptions. Show less

Presented at Solar 2013
Baltimore, Maryland
April 16-20, 2013

This paper describes a straightforward methodology for modeling photovoltaic arrays comprised of variously configured sub-arrays connected to a single inverter. Particularly in rooftop applications, PV arrays must be installed within the constraints of various roof slopes and geometries. This reality calls into question the typical modeling assumption that each panel operates at its maximum power point, even when shading effects are ignored. A series of scenarios are presented with a variety of… Show more

This paper describes a straightforward methodology for modeling photovoltaic arrays comprised of variously configured sub-arrays connected to a single inverter. Particularly in rooftop applications, PV arrays must be installed within the constraints of various roof slopes and geometries. This reality calls into question the typical modeling assumption that each panel operates at its maximum power point, even when shading effects are ignored. A series of scenarios are presented with a variety of array orientations, string configurations, and temperature effects. Each scenario is modeled in detail using industry standard modeling tools, and the operation characteristics and DC losses due to sub-array layout mismatch are presented. Typical losses resulting from sub-optimal relative alignment of fixed array layouts are on the order of a one percent or less on an annual basis, suggesting that sub-array orientation in the absence of shading is not a major factor in small to medium scale system energy yield. Show less

Presented at the 2012 World Renewable Energy Forum
Denver, Colorado
May 13-17, 2012

Presented at the 2012 World Renewable Energy Forum
Denver, Colorado
May 13-17, 2012

This paper describes an improved algorithm for calculating the six parameters required by the California Energy Commission (CEC) photovoltaic (PV) Calculator module model. Rebate applications in California require results from the CEC PV model, and thus depend on an up-to-date database of module characteristics. Currently, adding new modules to the database requires calculating operational coefficients using a general purpose equation solver—a cumbersome process for the 300+ modules added on… Show more

This paper describes an improved algorithm for calculating the six parameters required by the California Energy Commission (CEC) photovoltaic (PV) Calculator module model. Rebate applications in California require results from the CEC PV model, and thus depend on an up-to-date database of module characteristics. Currently, adding new modules to the database requires calculating operational coefficients using a general purpose equation solver—a cumbersome process for the 300+ modules added on average every month. The combination of empirical regressions and heuristic methods presented herein achieve automated convergence for 99.87% of the 5487 modules in the CEC database and greatly enhance the accuracy and efficiency by which new modules can be characterized and approved for use. The added robustness also permits general purpose use of the CEC/6 parameter module model by modelers and system analysts when standard module specifications are known, even if the module does not exist in a preprocessed database. Show less