Nowadays, the estimation of the PhotoVoltaic (PV) power production systems is crucial for ensuring their economic feasibility as well as their proper sizing in order to avoid outages and guarantee quality and continuity of supply. The working temperature of a PV module or system is a key parameter for the assessment of the actual performance of photovoltaic modules. PV modules are usually rated at Standard Test Conditions (STC = 1000W/m2, AM1.5, 25°C), but their operating temperatures are typically considerably higher. Power production can be highly influenced by cell working (module) temperature whose increase respect to standard one could gradually deteriorate system's energy performance. Correlations to evaluate performances referring to STC and/or applying some theoretical simplifications/assumptions are available in literature. However, it be noticed that the use of these correlations, under the same operative conditions, does not produce univocal results. In this paper, an Artificial Neural Network (ANN)-based model to forecast the working temperature of a Concentrator PhotoVoltaic (CPV) module (back - plate temperature) is proposed. A dataset consisting of meteorological data (i.e. solar direct normal irradiance, ambient temperature, wind speed and wind direction), measured every 1-minute and recorded from November 2012 to April 2014, concerning a 50 kWp CPV plant installed in the southwestern Europe, is used for the training and testing of the ANN developed. The main advantage of the proposed approach is that it acts as black box tool, making easy to model an arbitrary complex non-linear relationship between inputs and outputs. In fact, in this case no in-depth knowledge of the system and its components is required contrary to what needed by deterministic techniques. In order to verify the effectiveness and the accuracy of the approach here proposed, measured and estimated data were analysed and compared considering different error metrics. © 2016, American Institute of Physics Inc. All rights reserved.
Forecasting the working temperature of a concentrator photovoltaic module by using artificial neural network-based model
Graditi, G.;Ferlito, S.;Cancro, C.
2016-01-01
Abstract
Nowadays, the estimation of the PhotoVoltaic (PV) power production systems is crucial for ensuring their economic feasibility as well as their proper sizing in order to avoid outages and guarantee quality and continuity of supply. The working temperature of a PV module or system is a key parameter for the assessment of the actual performance of photovoltaic modules. PV modules are usually rated at Standard Test Conditions (STC = 1000W/m2, AM1.5, 25°C), but their operating temperatures are typically considerably higher. Power production can be highly influenced by cell working (module) temperature whose increase respect to standard one could gradually deteriorate system's energy performance. Correlations to evaluate performances referring to STC and/or applying some theoretical simplifications/assumptions are available in literature. However, it be noticed that the use of these correlations, under the same operative conditions, does not produce univocal results. In this paper, an Artificial Neural Network (ANN)-based model to forecast the working temperature of a Concentrator PhotoVoltaic (CPV) module (back - plate temperature) is proposed. A dataset consisting of meteorological data (i.e. solar direct normal irradiance, ambient temperature, wind speed and wind direction), measured every 1-minute and recorded from November 2012 to April 2014, concerning a 50 kWp CPV plant installed in the southwestern Europe, is used for the training and testing of the ANN developed. The main advantage of the proposed approach is that it acts as black box tool, making easy to model an arbitrary complex non-linear relationship between inputs and outputs. In fact, in this case no in-depth knowledge of the system and its components is required contrary to what needed by deterministic techniques. In order to verify the effectiveness and the accuracy of the approach here proposed, measured and estimated data were analysed and compared considering different error metrics. © 2016, American Institute of Physics Inc. All rights reserved.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
