Short-term solar power forecasting: Investigating the ability of deep learning models to capture low-level utility-scale Photovoltaic system behaviour

Short-Term Deterministic Solar Irradiance Forecasting Considering a Heuristics-Based, Operational Approach

“…Clusters for different patterns for inputs were constructed based on data similarity using either a shape similarity approach [38] or a pattern discovery method [39]. While these approaches proved successful, the method is highly dependent on the characteristics of the data, as the resulting clusters are defined based on mathematical abstractions and no straightforward relations allow for simple understanding of the data characteristics and cluster boundaries, but more importantly, the interoperability or how the resulting construct can be applied to a different context, a shortcoming shared by [13][14][15] and already expressed in the introductory section. Examples of this approach are presented in Bhardwaj et al [38], where Markov chains were used to model transitions for atmospheric variables (temperature, humidity, sunshine hours, atmospheric pressure and wind speed) and a generalized fuzzy model was developed to take these variables as inputs to produce solar irradiance as the output.…”

“…Clusters for different patterns for inputs were constructed based on data similarity using either a shape similarity approach [38] or a pattern discovery method [39]. While these approaches proved successful, the method is highly dependent on the characteristics of the data, as the resulting clusters are defined based on mathematical abstractions and no straightforward relations allow for simple understanding of the data characteristics and cluster boundaries, but more importantly, the interoperability or how the resulting construct can be applied to a different context, a shortcoming shared by [13][14][15] and already expressed in the introductory section.…”

“…Du Plessis et al [15] Several neural networks were developed for meteorological data to forecast a photovoltaic plant output at the subunit level and then scaled up to plant-wide production.…”

“…Despite the fact that several well-known metrics and procedures for forecasting performance assessment are presented in the literature, most studies are limited to assessing performance of the entire evaluation period, while some further detail at how performance varies with month or season [16,[29][30][31], but comparatively few studies provide details at the type-of-day [15] level or for specific values or ranges for K t [14,32]. This analysis is of interest, because as Figure 1 shows, the same method can yield very different results as a function of the variability of the location under analysis.…”

Deep Learning Based on Wireless Remote Sensing Model for Monitoring the Solar System Inverter

“…Any system including the PV system when managed by deep learning should be understood first in order to get utmost benefit. Training such system is not easy and needs more experience and costs time under fixed conditions [9]. Due to the proposed system which is managed remotely, we have to train the remote-sensing system to be able to adapt with the deep learning system.…”