Accuracy of Energy Prediction Methodologies

2010

During the four years of the PERFORMANCE project enormous contributions to pre-normative research related to photovoltaic modules and systems have been achieved. The contributions ranged from instantaneous device characterisation and system measurements to their life-time performance prediction and assessment. This paper presents the results of the 3rd and final modelling Round Robin (RR3) for PV module energy prediction. It describes the inter-comparison of various energy prediction methods with a direct link to the currently discussed IEC61853 energy rating standard.

Energy Conversion and Management, 2017

Accurate and reliable modelling of photovoltaic (PV) modules is necessary for design and performance estimation of PV systems. Manufacturers of PV modules usually provide some basic electrical parameters specified at only one single operating condition, which is commonly known as standard test conditions (STCs). However, PV systems operate over a wide range of outdoor conditions, and the manufacturers' reports do not cover sufficient information about how the PV modules react with respect to changes in the two most important environment-oriented conditions, which are solar intensity and operating cell temperature. In this respect, an accurate and reliable tool is required by designers to predict electrical characteristics and thermodynamic performance parameters of PV modules. Therefore in this research, a novel mathematical model is developed to determine solar intensity and cell temperature dependency of PV module parameters and thermodynamic efficiency figures. A simple one-diode model is proposed considering series resistance and shunt conductance. Mitsubishi PV-UJ225GA6 225 W polycrystalline silicon PV module and Kyocera KD205GH-2P 205 W multicrystal PV module are utilised for model assessment. Model results are compared with environmental chamber tests and manufacturers' performance reports, and a very good agreement is achieved.

Energy Conversion and Management, 2018

The recent boom in distributed generation together with decreases in the cost of technology has resulted in photovoltaic energy becoming the most widely used means of producing electricity in buildings. Analysing this technology requires models that use the data available in the manufacturers' catalogues to study energetic and economic impact of the different commercial solutions available. Furthermore, if the model is to be integrated into a complex simultaneous simulation algorithm for different solutions, it should have a low computational cost. The electricity output of photovoltaic modules is characterized by the I-V response according to the ambient temperature. A review of the literature leads to the conclusion that existing models are difficult to parameterize using the data available in the technical product datasheets. This study presents an explicit, simplified model to determine the characteristic I-V curve of photovoltaic modules by means of a direct link with the data available in the catalogues of manufacturers. The model's validity and quality of results it provides have been tested on a sample of 259 modules of different brands, sizes and technologies, as well as under varying operating conditions. The explicit model proposed is highly accurate, easy to reproduce, and can be adapted to the parameters found in most manufacturers' catalogues.

Energy Conversion and Management, 2017

Electrical behavior predicting of photovoltaic modules, at different operating climatic conditions, remains a crucial issue for the estimation of output power from photovoltaic (PV) plants. In this paper, a simplified explicit model to describe the behavior of PV modules is introduced; the model is based on a simple mathematical equation relating the current to the voltage (I-V). The model requires the estimation of three parameters which are: open circuit voltage (V oc), the short circuit current (I sc) and a shape parameter (S). The model validation has been performed through experimental measurements for four different PV modules technologies (mono-crystalline Silicon, multi-crystalline Silicon, Copper Indium Gallium Selenide and Cadmium Telluride) at two different locations. To show its effectiveness, the model is then compared with four explicit models. Results showed that the model accurately predicts the I-V characteristics for the four examined PV modules at different range of solar irradiance levels and cell temperatures. Moreover, the developed model performs better than other investigated models in terms of accuracy and simplicity.

2017

Several factors impact the power output from solar photovoltaic (PV) modules. Some are deterministic and controllable and others are uncontrollable (e.g., meteorological conditions), yet critical to performance. e objective of this paper is to assess the relationship between the meteorological variables and the power output of a mono silicon PV module using Multiple Linear Regression modelling. e approach involved exposing one mono-silicon photovoltaic module to the open atmosphere for a period of time and measuring the electrical energy output as a function of natural variation in the meteorological factors. A regression model for the power output was developed and significant variables identified. e model derived for this material grade of silicon tended to match the validation data more closely for clear skies, but not as accurate for times of cloud cover. Results of this study will provide useful design and application insight on critical factors that impact the energy capabi...

2009

This paper presents some of the results obtained within the European project PERFORMANCE in the field of module energy prediction. It describes the second of three modelling inter-comparison campaigns (round robins). The main aim of the study is to further reduce and to better understand the uncertainty in energy predictions and to support the development of the IEC61853 energy rating standard for PV modules. The round robin results showed that an improvement in energy predictions by implementing spectral and angle of incidence effects is very difficult to achieve without increasing the reliability and availability of the required input parameters. Independently of the type of technology, none of the seven models investigated was able to significantly improve the annual energy prediction respect to the basic approach where only in-plane irradiance, as measured by a broad band pyranometer, and measured back of module temperature were used as input. A real improvement could only be achieved for clear sky days. The temperature models were shown to be less critical, leading to no major decrease in annual energy prediction accuracy. However, the translation from horizontal to in-plane irradiance led to larger errors, these were mainly due to the high uncertainty in diffuse irradiance measurements and/or its translation to the tilted plane.

2017 IEEE 44th Photovoltaic Specialist Conference (PVSC), 2017

Knowing the uncertainty in the PV production forecast is crucial in the optimization of smart-grid operations and in its stability. In this paper, we present uncertainty calculation in the PV energy production forecast process, calculated step by step. From horizontal irradiance and ambient temperature, all the steps needed to evaluate the PV production are studied with a special focus on the comparison between calculation and measurements. The uncertainty is evaluated by computing the relative mean bias error and the relative mean absolute error.

Renewable and Sustainable Energy Reviews, 2017

An essential stage in assessing the feasibility of a PV project is the energy yield prediction, which estimates the total energy production of a PV system at a specific site. Photovoltaic (PV) performance models are mathematical representations used to estimate the energy yield of power systems based on PV technology. The PV performance models are subjected to a series of errors derived from the different steps in the modeling chain of the PV system. Although some studies have been conducted to assess the accuracy of these models, limited research have focused on studying the accuracy of the individual submodels that comprise the PV performance model. The main objective of this paper is to assess the performance of different combinations of the most cited models aiming to find a PV performance model with good accuracy. There were studied a total of 20 PV performance models derived from the combination of four plane-of-the-array (POA) irradiance models, five PV module models and two inverter models. All the PV performance models were implemented computationally and their performance was compared with measurements collected by a data acquisition system in a real 2.2 kWp photovoltaic system. The best PV performance model presents an accuracy of −0.201% (rMBE) and 15.099% (rRMSE) with respect to the measured AC power output, which is in line with the values reported in the literature. Several sources of error were identified, which can greatly influence PV system energy yield estimation. Among them, the uncertainty in the derating factors which represent all the non-temperature dependent losses present in the PV system is the most critical.

In the current market, the specific annual energy yield (kWh/kWp) of a PV system is gaining in importance due to its direct link to the financial returns for possible investors who typically demand an accuracy of 5% in this prediction. This paper focuses on the energy prediction of photovoltaic modules themselves, as there have been significant advances achieved with module technologies which affect the device physics in a way that might force the revisiting of device modelling.

Journal of Solar Energy Engineering, 2006

Computer simulation tools used to predict the energy production of photovoltaic systems are needed in order to make informed economic decisions. These tools require input parameters that characterize module performance under various operational and environmental conditions. Depending upon the complexity of the simulation model, the required input parameters can vary from the limited information found on labels affixed to photovoltaic modules to an extensive set of parameters. The required input parameters are normally obtained indoors using a solar simulator or flash tester, or measured outdoors under natural sunlight. This paper compares measured performance parameters for three photovoltaic modules tested outdoors at the National Institute of Standards and Technology (NIST) and Sandia National Laboratories (SNL). Two of the three modules were custom fabricated using monocrystalline and silicon film cells. The third, a commercially available module, utilized triple-junction amorpho...