Classification of daily solar radiation distributions using a mixture of Dirichlet distributions

Soubdhan, Ted; Emilion, Richard; Calif, Rudy

doi:10.1016/j.solener.2009.01.010

Cited by 98 publications

(65 citation statements)

References 16 publications

(16 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The available data set was divided, as usual in good practice, into a training and a testing data set, in order to avoid polarization of the identified models. From a visual standpoint, the performances of two independent 1-step ahead neural networks based NAR models, to predict W r and H daily values, respectively, are shown in Figure (8). As it is possible to see, the prediction models behave quite similar to a persistent model, that is, the predicted time series exhibits peaks delayed by 1-step with respect to the true time series.…”

Section: A Identify Nar Models For W R (D) and H(d)mentioning

confidence: 93%

“…Decision trees based on if − then rules, which are one of the most popular methods used in machine learning for classification, have been proposed by [7] with the aim of implementing short term wind speed prediction models. In related renewable energy source fields, such as solar radiation time series, classification of daily time series has been proposed by [8]. Data mining techniques and clustering approaches to classify wind speed data in different cities of Turkey have been adopted by [9].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

One day ahead prediction of wind speed class

Fortuna

Nunnari

Guariso

2015

2015 International Conference on Renewable Energy Research and Applications (ICRERA)

View full text Add to dashboard Cite

Abstract-This paper deals with the problem of clustering daily wind speed time series based on two features referred to as Wr and H, representing a measure of the relative daily average wind speed and the Hurst exponent, respectively. Daily values of the pairs (Wr, H) are first classified by means of the fuzzy c-means unsupervised clustering algorithm and then results are used to train a supervised MLP neural network classifier. It is shown that associating to a true wind speed time series a time series of classes, allows performing some useful statistics. Further, the problem of predicting 1-step ahead the class of daily wind speed is addressed by introducing NAR sigmoidal neural models into the classification process. The performance of the prediction model is finally assessed.

show abstract

Section: A Identify Nar Models For W R (D) and H(d)mentioning

confidence: 93%

Section: Introductionmentioning

confidence: 99%

One day ahead prediction of wind speed class

Fortuna

Nunnari

Guariso

2015

2015 International Conference on Renewable Energy Research and Applications (ICRERA)

View full text Add to dashboard Cite

show abstract

“…In work of Soubdhan et al (2009) a method to classify daily clearness index t k distributions using mixture Dirichlet distributions. Solar radiation data sampled with frequency of 1 Hz was used for this work.…”

Section: Cloud Layer Simulationmentioning

confidence: 99%

“…In a similar manner as in (Soubdhan et al 2009) and (Chia, Hutchinson 1991) the peak of PDF is chosen in known average for a given month.…”

Section: Cloud Layer Simulationmentioning

confidence: 99%

“…2. Probability density functions, defining the probability of cloud coverage ratio in the sky during a day in one of four months (February, May, August and November) in Lithuania The albedo of low clouds can be as high as 0.6 (Zekai 2008), so the light reflected from the clouds also reach surface of PV module and in some cases total instantaneous radiation can become higher than of clear sky (Soubdhan et al 2009). …”

Section: Cloud Layer Simulationmentioning

confidence: 99%

See 1 more Smart Citation

Solar Radiation Model for Development and Control of Solar Energy Sources / Saulės Spinduliavimo Modelis Fotovoltinių Elektrinių Projektavimui Ir Valdymui

Vasarevičius

Pikutis

2016

Mokslas – Lietuvos Ateitis

View full text Add to dashboard Cite

ELEKTroNIKa Ir ELEKTroTEChNIKa ELECTroNICS aNd ELECTrICaL ENGINEErINGIn previous work a model for a clear sky direct solar irradiance was presented (Vasarevičius, Martavičius 2011). Estimating only direct component is usually enough when analysing solar thermal collectors, but PV modules also gain benefit from diffused and especially reflected radiation components when the surrounding environment has high albedo (e. g. snow, water ponds in evening and morning). Presented model is implemented in Matlab/Simulink environment and generates total solar irradiation values at a given location, described by geographical coordinates and altitude for any time of the year. Model also outputs Zenith and azimuth angles, which can be used for sun tracking applications. Due to this many analytic expressions are used, to meet computational possibilities of embedded devices and increase modelling speed in Matlab.Some control systems like MPPT tracking are very sensitive to rapidly varying solar radiation due to clouds and usually have reduced efficiency in such conditions. The main aim of presented model is development of MPPT algorithms, so the model of cloud layer simulation is included. Abstract. The model of solar radiation, which takes into account direct, diffused, and reflected components of solar energy, has been presented. Model is associated with geographical coordinates and local time of every day of the year. It is shown that using analytic equations for modelling the direct component, it is possible to adopt it for embedded systems with low computational power and use in solar tracking applications. Reflected and diffused components are especially useful in determining the performance of photovoltaic modules in certain location and surroundings. The statistical method for cloud layer simulation based on local meteorological data is offered. The presented method can't be used for prediction of weather conditions but it provides patterns of solar radiation in time comparable to those measured with pyranometer. Cloud layer simulation together with total solar radiation model is a useful tool for development and analysis of maximum power point tracking controllers for PV modules.

show abstract

Harmonic modelling of PV systems for probabilistic harmonic load flow studies

Ruíz-Rodríguez

Hernández

Jurado

2014

Circuit Theory & Apps

View full text Add to dashboard Cite

Summary Probabilistic harmonic load flow (HLF) is currently enjoying renewed popularity. This is not surprising since in the near future, thousands of photovoltaic (PV) systems will be integrated into distribution systems. However, as yet, there is no model capable of explaining PV harmonic current behaviour in probabilistic HLF studies. To fill this gap, the harmonic modelling of PV systems proposed in this paper has three key points. First, it provides an effective model of the relationship between PV harmonic current emission and background harmonic voltages. Second, it statistically characterises PV harmonic currents (relative magnitude and phase angle) at different fundamental‐frequency current output intervals using historical time‐series data. In this statistical characterisation, the first fifth moments of each PV harmonic current are used to accurately approximate the raw probability density function (PDF) by means of the Legendre series. Finally, the third key point of this harmonic modelling is a method capable of determining the distribution functions of PV harmonic currents (absolute magnitude and phase angle), based on the statistical characterisation and a fundamental‐frequency probabilistic PV model. The numerical results obtained confirm the effectiveness of this PV model. Copyright © 2014 John Wiley & Sons, Ltd.

show abstract

Classification of daily solar radiation distributions using a mixture of Dirichlet distributions

Cited by 98 publications

References 16 publications

One day ahead prediction of wind speed class

One day ahead prediction of wind speed class

Solar Radiation Model for Development and Control of Solar Energy Sources / Saulės Spinduliavimo Modelis Fotovoltinių Elektrinių Projektavimui Ir Valdymui

Harmonic modelling of PV systems for probabilistic harmonic load flow studies

Contact Info

Product

Resources

About