Hourly Solar Irradiance Prediction Based on Support Vector Machine and Its Error Analysis

Bae, Kuk Yeol; Jang, Han Seung; Sung, Dan Keun

doi:10.1109/tpwrs.2016.2569608

Cited by 107 publications

(62 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We convert the irradiance into PV output power based on the prediction data in [15]. In this paper, we assume that the PV output power is linearly proportional to the irradiance without considering the module temperature.…”

Section: Prediction Results For Pv Output Powermentioning

confidence: 99%

See 1 more Smart Citation

Effect of Prediction Error of Machine Learning Schemes on Photovoltaic Power Trading Based on Energy Storage Systems

et al. 2019

View full text Add to dashboard Cite

Photovoltaic (PV) output power inherently exhibits an intermittent property depending on the variation of weather conditions. Since PV power producers may be charged to large penalties in forthcoming energy markets due to the uncertainty of PV power generation, they need a more accurate PV power prediction scheme in energy market operation. In this paper, we characterize the effect of PV power prediction errors on energy storage system (ESS)-based PV power trading in energy markets. First, we analyze the prediction accuracy of two machine learning (ML) schemes for the PV output power and estimate their error distributions. We propose an efficient ESS management scheme for charging and discharging operation of ESS in order to reduce the deviations between the day-ahead (DA) and real-time (RT) dispatch in energy markets. In addition, we estimate the capacity of ESSs, which can absorb the prediction errors and then compare the PV power producer’s profit according to ML-based prediction schemes with/without ESS. In case of ML-based prediction schemes with ESS, the ANN and SVM schemes yield a decrease in the deviation penalty by up to 87% and 74%, respectively, compared with the profit of those schemes without ESS.

show abstract

Section: Prediction Results For Pv Output Powermentioning

confidence: 99%

“…In addition, a number of prediction studies have developed nonlinear solar or PV output power prediction schemes based on machine learning (ML). In specific, artificial neural networks (ANN) [14], support vector machine (SVM) [15,16], and hybrid schemes [17] have been utilized in prediction models.…”

Section: Introductionmentioning

confidence: 99%

Effect of Prediction Error of Machine Learning Schemes on Photovoltaic Power Trading Based on Energy Storage Systems

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Even though the forecasting error is inevitable, it is assumed that the forecasting module provides accurate forecasts of load demand and renewable generation output power. This is not a radical assumption, and already there have been many studies that show high accuracy with many forecasting algorithms such as linear regression, clustering, and support vector machine (SVM) [24,25]. When the amounts of accumulated data in the database increase over time, the accuracy of forecasts will improve.…”

Section: Modularized Programming Architecturementioning

confidence: 99%

Comparative Study on Optimization Solvers for Implementation of a Two-Stage Economic Dispatch Strategy in a Microgrid Energy Management System

et al. 2020

View full text Add to dashboard Cite

A microgrid energy management system (MEMS) optimally schedules the operation of dispatchable distributed energy resources to minimize the operation costs of microgrids (MGs) via an economic dispatch (ED). Actual ED implementation in the MEMS relies on an optimization software package called an optimization solver. This paper presents a comparative study of optimization solvers to investigate their suitability for ED implementation in the MEMS. Four optimization solvers, including commercial as well as open-source-based ones, were compared in terms of their computational capability and optimization results for ED. Two-stage scheduling was applied for the ED strategy, whereby a mixed-integer programming problem was solved to yield the optimal operation schedule of battery-based energy storage systems. In the first stage, the optimal schedule is identified one day before the operating day; in the second stage, the optimal schedule is updated every 5 min during actual operation to compensate for operational uncertainties. A modularized programming strategy was also introduced to allow for a comparison between the optimization solvers and efficient writing of codes. Comparative simulation case studies were conducted on three test-bed MGs to evaluate the optimization results and computation times of the compared optimization solvers.Commercial optimization solvers generally have superior computational capability compared with their open-source counterparts. The ED implementation environment is then programmed either in general-purpose programming languages (GPLs), e.g., Python and Julia, or in algebraic modeling languages (AMLs). AML is a high-level programming language dedicated to optimization applications. Popular AMLs include GAMS and AMPL.Numerous studies on ED have been conducted from the perspectives of its optimization algorithms [4][5][6][7][8][9][10] and actual implementation methods [11][12][13][14][15]. Table 1 summarizes the optimization algorithms, optimal solvers, programming languages, and demonstration methods used in these studies. The optimization algorithms to the ED problem have been well studied. Particularly, in references [4-10], various analytic-based optimization methods, such as linear programming (LP), mixed-integer programming (MIP), and convex optimization, have been investigated for the application to ED. In reference [4], day-ahead ED was formulated as an LP problem to determine the hourly planned operation set points (i.e., charging and discharging power references) of battery-based energy storage systems (ESSs) and electric vehicles (EVs). In reference [5], convex optimization method was used to determine the optimal charging and discharging schedules of an ESS. In reference [6], an online battery power control method based on an MIP formulation was used over a rolling horizon window. The ED strategies proposed in [4][5][6] planned the operation schedule only once and the schedule is followed without any modifications during operation. However, one-time scheduling strategies are vulne...

show abstract

“…3) Statistical and Artificial intelligence (AI) techniques are recently presented for a number of solar irradiance and PV power estimation/regression problems. As discussed in [13], the non-stationary and highly nonlinear characteristics of solar radiation time series lead to the superiority of AI approaches over the traditional statistical models. Machine learning algorithms are employed as target function approximators, to estimate future solar irradiance or PV power.…”

Section: Introductionmentioning

confidence: 99%

Convolutional Graph Autoencoder: A Generative Deep Neural Network for Probabilistic Spatio-Temporal Solar Irradiance Forecasting

Khodayar

Mohammadi

Khodayar

et al. 2020

IEEE Trans. Sustain. Energy

127

View full text Add to dashboard Cite

Machine Learning on graph-structured data is an important and omnipresent task for a vast variety of applications including anomaly detection and dynamic network analysis. In this paper, a deep generative model is introduced to capture continuous probability densities corresponding to the nodes of an arbitrary graph. In contrast to all learning formulations in the area of discriminative pattern recognition, we propose a scalable generative optimization/algorithm theoretically proved to capture distributions at the nodes of a graph. Our model is able to generate samples from the probability densities learned at each node. This probabilistic data generation model, i.e. convolutional graph autoencoder (CGAE), is devised based on the localized first-order approximation of spectral graph convolutions, deep learning, and the variational Bayesian inference. We apply our CGAE to a new problem, the spatio-temporal probabilistic solar irradiance prediction. Multiple solar radiation measurement sites in a wide area in northern states of the US are modeled as an undirected graph. Using our proposed model, the distribution of future irradiance given historical radiation observations is estimated for every site/node. Numerical results on the National Solar Radiation Database show state-of-the-art performance for probabilistic radiation prediction on geographically distributed irradiance data in terms of reliability, sharpness, and continuous ranked probability score.

show abstract

Hourly Solar Irradiance Prediction Based on Support Vector Machine and Its Error Analysis

Cited by 107 publications

References 29 publications

Effect of Prediction Error of Machine Learning Schemes on Photovoltaic Power Trading Based on Energy Storage Systems

Effect of Prediction Error of Machine Learning Schemes on Photovoltaic Power Trading Based on Energy Storage Systems

Comparative Study on Optimization Solvers for Implementation of a Two-Stage Economic Dispatch Strategy in a Microgrid Energy Management System

Convolutional Graph Autoencoder: A Generative Deep Neural Network for Probabilistic Spatio-Temporal Solar Irradiance Forecasting

Contact Info

Product

Resources

About