Multivariate Analysis for Solar Resource Assessment Using Unsupervised Learning on Images from the GOES-13 Satellite

Abstract

Solar resource assessment is of paramount importance in the planning of solar energy applications. Solar resources are abundant and characterization is essential for the optimal design of a system. Solar energy is estimated, indirectly, by the processing of satellite images. Several analyses with satellite images have considered a single variable—cloudiness. Other variables, such as albedo, have been recognized as critical for estimating solar irradiance. In this work, a multivariate analysis was carried out, taking into account four variables: cloudy sky index, albedo, linke turbidity factor (TL2), and altitude in satellite image channels. To reduce the dimensionality of the database (satellite images), a principal component analysis (PCA) was done. To determine regions with a degree of homogeneity of solar irradiance, a cluster analysis with unsupervised learning was performed, and two clustering techniques were compared: k-means and Gaussian mixture models (GMMs). With respect to k-means, the GMM method obtained a smaller number of regions with a similar degree of homogeneity. The multivariate analysis was performed in Mexico, a country with an extended territory with multiple geographical conditions and great climatic complexity. The optimal number of regions was 17. These regions were compared for annual average values of daily irradiation data from ground stations using multiple linear regression. A comparison between the mean of each region and the ground station measurement showed a linear relationship with a R2 score of 0.87. The multiple linear regression showed that the regions were strongly related to solar irradiance. The optimal sites found are shown on a map of Mexico.

Annual Daily Irradiance Analysis of Clusters in Mexico by Machine Learning Algorithms

“…The creation of the dataset used is discussed in [1], wherein detailed explanations are provided regarding the preprocessing, modeling and evaluation of satellite images. This process resulted in the generation of datasets and maps that cluster Mexico into distinct regions based on annual values of albedo, Linke, cloudiness index and altitude for the year 2015.…”

“…Planning the strategic locations for the installation of a comprehensive measurement network proves to be indispensable for the assessment of solar resources, especially in vast territories characterized by a multitude of climatic variations [1,2]. The substantial financial commitment entailed in establishing solar radiation measurement stations, coupled with the inherent challenges associated with their sustained maintenance, underscores the critical need for methodological approaches capable of precisely determining both the optimal number of stations and the most suitable deployment sites [3].…”

“…The substantial financial commitment entailed in establishing solar radiation measurement stations, coupled with the inherent challenges associated with their sustained maintenance, underscores the critical need for methodological approaches capable of precisely determining both the optimal number of stations and the most suitable deployment sites [3]. To address this imperative, recent years have witnessed the publication of several works that delve into the identification of potential sites for solarimetric station installations, leveraging advanced techniques such as data mining and cluster analysis [1,2,4]. These methodological advancements not only contribute to the refinement of the selection process, but also pave the way for enhanced resource assessment methodologies, ensuring a more robust foundation for solar energy planning and utilization.…”

“…A work was recently published that performs regionalization based on solar irradiation, measuring different climatic characteristics such as albedo, cloud cover, altitude and atmospheric turbidity [1]. In this research, regionalization is accomplished through the application of the unsupervised K-means classification method and Gaussian models.…”

“…These same techniques allow for the calculation of the significance of these parameters in relation to measurements from surface solar radiation stations. Notably, the study delves into the application of diverse methodologies, including Random Forest (RF), Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Multiple Linear Regression (MLR), derived from the bases established in [1], and secondly works to identify the machine learning algorithm that presents the best model to forecast and estimate solar irradiance. To achieve the above, the annual daily irradiation from 26 meteorological stations were localized in each cluster, averaged and then related to the cluster's centers of each k cluster.…”