Total Solar Irradiance Forecasting with Keras Recurrent Neural Networks

Muralikrishna, Amita; Vieira, L. E. A.; Santos, Rafael; Almeida, Adriano P.

doi:10.1007/978-3-030-58814-4_18

Cited by 4 publications

(3 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Keras [4] is one of the most popular high-level DL libraries that has been widely used in various domains [15,32,40].…”

Section: Deep Learning Librarymentioning

confidence: 99%

Muffin: Testing Deep Learning Libraries via Neural Architecture Fuzzing

Gu¹,

Luo²,

Zhou³

et al. 2022

Preprint

View full text Add to dashboard Cite

Deep learning (DL) techniques are proven effective in many challenging tasks, and become widely-adopted in practice. However, previous work has shown that DL libraries, the basis of building and executing DL models, contain bugs and can cause severe consequences. Unfortunately, existing testing approaches still cannot comprehensively exercise DL libraries. They utilize existing trained models and only detect bugs in model inference phase. In this work we propose Muffin to address these issues. To this end, Muffin applies a specifically-designed model fuzzing approach, which allows it to generate diverse DL models to explore the target library, instead of relying only on existing trained models. Muffin makes differential testing feasible in the model training phase by tailoring a set of metrics to measure the inconsistencies between different DL libraries. In this way, Muffin can best exercise the library code to detect more bugs. To evaluate the effectiveness of Muffin, we conduct experiments on three widely-used DL libraries. The results demonstrate that Muffin can detect 39 new bugs in the latest release versions of popular DL libraries, including Tensorflow, CNTK, and Theano. CCS Concepts: • Software and its engineering → Software testing and debugging; Software libraries and repositories.

show abstract

“…Keras [4] is one of the most popular high-level DL libraries that has been widely used in various domains [15,32,40].…”

Section: Deep Learning Librarymentioning

confidence: 99%

Muffin: Testing Deep Learning Libraries via Neural Architecture Fuzzing

Gu¹,

Luo²,

Zhou³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Although the BTT can store inputs already processed in the past to positively influence the processing of current inputs towards the desired output, the back-propagated error begins not to interfere significantly in the weights' adjustment, bringing the learning problem of vanishing gradient (Hochreiter, 1998), which the simplest recurring networks cannot deal with, as confirmed in Muralikrishna et al (2020).…”

Section: Recurrent Neural Networkmentioning

confidence: 99%

Exploring possibilities for solar irradiance prediction from solar photosphere images using recurrent neural networks

Muralikrishna¹,

Santos²,

Vieira³

2022

J. Space Weather Space Clim.

Self Cite

View full text Add to dashboard Cite

Studies of the Sun and the Earth's atmosphere and climate consider solar variability and its constant monitoring an important driver in climate models. Solar irradiance is one of the main parameters that allow monitoring this variation, which can be studied in spectrum ranges or in its version that integrates all those ranges. Some physical and semi-empirical models were developed and made available and are very relevant for the reconstruction of irradiance in periods of data failure or absence in the collection. However, the solar irradiance prediction could benefit ionospheric and climate models through prior knowledge of irradiance values hours or days ahead, without the need to know or have available other parameters that would be necessary for their reconstruction. This paper presents a neural network based approach, which uses images of the solar photosphere to extract sunspot and active region information and thus generate inputs for recurrent neural networks to perform the irradiance prediction. Experiments were performed with two recurrent neural network architectures for short- and long-term predictions of total and spectral solar irradiance along three wavelengths. The results show good quality of prediction results for TSI and motivate individual effort in improving the prediction of each type of irradiance predicted in this work. The results obtained for SSI point out that photosphere images do not have the same influence on the prediction of all wavelengths tested, but encourage the bet on new spectral lines prediction. The quality closeness in neural networks and physical models results raise the possibility that prediction is an option to be considered in studies for which only reconstructed data are considered.

show abstract

“…A inteligência artificial tem sido muito utilizada para predições de dados do tipo sequências temporais. Nas pesquisas de Muralikrishna et al (2020) e Santos (2019), as Redes Neurais Recorrentes (RNN -"Recurrent neural networks") são utilizadas e apresentam resultados satisfatórios. As RNN "lembram de informações" isso acontece em virtude delas utilizarem loops em sua estrutura.…”

Section: Rede Neural Long Short-term Memoryunclassified

Avaliação da vida útil de um transformador da subestação de uma usina eólica: Um estudo de caso no nordeste brasileiro

Marini

Baroni

2022

Proceedings IX Simpósio Brasileiro De Sistemas Elétricos

View full text Add to dashboard Cite

This article performs correlation analyzes between the aging of a transformer with forced ventilation of a northeastern wind farm in Brazil with the variation of wind speed. The main objective is to verify if the power oscillation existing in the wind farms accelerates the degradation of the transformer, thus identifying a factor of correct dimensioning of the transformers, preventing high maintenance and operation costs of the companies that generate this type of energy. Such verification's are carried out using international standards with equivalence in the country, that consider apparent power and ambient temperature as the main factors for calculating aging. The results show that the temperature of the hottest point of the winding can reach approximately 130 degrees celsius, which reduces the lifespan of the transformer. In order for the transformer to operate for 35 years, this article concludes that the power of the substation transformer must be about 0.8 of the rated power of the wind farm. Resumo: O presente artigo realiza análises de correlacionamento entre o envelhecimento de um transformador com ventilação forçada de uma usina eólica do nordeste brasileiro com a variação da velocidade do vento. O objetivo principal é verificar se a oscilação de potência existente nos complexos eólicos acelera a degradação do transformador, identificando assim um fator de correto dimensionamento dos transformadores, prevenindo, altos custos de manutenção e operação das empresas geradoras deste tipo de energia. Tais verificações são realizadas utilizando normas internacionais com equivalência no país, que consideram a potência aparente e a temperatura ambiente como principais fatores para o calcula do envelhecimento. Os resultados mostram que a temperatura do ponto mais quente do enrolamento pode chegar a aproximadamente 130 °C o que reduz a vida útil do transformador. Para que o transformador opere por 35 anos este artigo conclui que a potência do transformador da subestação deve ser aproximadamente 0,8 da potência nominal do parque eólico.

show abstract

Total Solar Irradiance Forecasting with Keras Recurrent Neural Networks

Cited by 4 publications

References 18 publications

Muffin: Testing Deep Learning Libraries via Neural Architecture Fuzzing

Muffin: Testing Deep Learning Libraries via Neural Architecture Fuzzing

Exploring possibilities for solar irradiance prediction from solar photosphere images using recurrent neural networks

Avaliação da vida útil de um transformador da subestação de uma usina eólica: Um estudo de caso no nordeste brasileiro

Contact Info

Product

Resources

About