Multi-step wind speed prediction based on turbulence intensity and hybrid deep neural networks

Li, Fei; Ren, Guorui; Lee, Jay

doi:10.1016/j.enconman.2019.02.045

Cited by 72 publications

(31 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, from just a year's data, we can reach the average trend and isolate the stochastic part of the variables, i.e., the Fourier-statistical analysis is not considered for not being periodic and for the following events: the environmental disturbances induced by men, the natural disasters, the climate change, the ecological processes (the transformation of natural habitats into agricultural and urban land), and biodiversity loss, among others. is study shows that meteorological historical sequences may be employed for the training of neural networks with some predictability capabilities [7][8][9][10]. Furthermore, the key features of meteorological temperature and humidity behavior are investigated by this technique.…”

Section: Discussionmentioning

confidence: 99%

“…It is the solutions C to equation (8), obtained through Cramer's rule, that will be tabulated in this work along with the corresponding value of the mean square deviation σ given by equation (5). ere is an important difference between the least-squares approach, based on equations (6) and (7) and the first one, which employs the Fourier coefficients given by equations (2)- (4). Suppose that we have adjusted a trigonometric polynomial of degree m using the Fourier coefficients of the equations (2)-(4).…”

Section: (7)mentioning

confidence: 99%

“…Consequently, new and efficient methodologies are required, to be able to achieve the inference for nontrivial models within an affordable time [5], including embedded, distributed, and parallel computing of inferential processes and anomalies from global or local distributions [6]. Recently, the wind speed prediction using Wind Speed and Turbulence Intensity-based Recursive uses artificial neural network [7] but requires that the maximum time-resolution of turbulence intensity should be longer than the prediction interval [8]. Likewise, the temperature prediction uses the artificial neural network model and it was able to predict the indoor temperature [9], but the majority of previous studies applying either ambient or ground have tended to emphasize the structural improvement of individual forecasting models without considering the periodicity of data [10].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Meteorological Temperature and Humidity Prediction from Fourier-Statistical Analysis of Hourly Data

Castañeda-Miranda

Icaza-Herrera

Castaño

2019

Advances in Meteorology

View full text Add to dashboard Cite

The temperature readings for all the 365 days and the 24 hours may be fitted through a 3 × 3 matrix (the so-called T-matrix). The mean square deviation between this fit and the actual meteorological measurements is smaller than three degrees Celsius. Four entries of this (nonsymmetric) matrix may be fixed by other means, leaving only five independent components. However, the same method applied to the humidity measurements produces a larger mean square deviation. A strong stochastical connection is found between the T-temperature matrix and the U-humidity matrix. The computer program, in C, may be used to adjust a (2M + 1) × (2m + 1) matrix simply by changing the arguments at the command line and has been tested with m and M ranging from zero to 11 (eleven) (more than 24 readings per day are necessary for larger values of m). The physical meaning of these constants is given only in the case m = M = 1. Our results have also been connected to fundamental cosmological properties: Earth’s orbit, the ecliptic angle, and the latitude of Querétaro (or whatever geographical location is chosen). A separate program calculates the angular position of the Sun as measured in the sky of Querétaro, to determine the length of the day or the mean value of the solar cosine. This work introduces several new variables which happen to be stochastically connected.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: (7)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Meteorological Temperature and Humidity Prediction from Fourier-Statistical Analysis of Hourly Data

Castañeda-Miranda

Icaza-Herrera

Castaño

2019

Advances in Meteorology

View full text Add to dashboard Cite

show abstract

“…Knowledge of meteorological conditions and site characteristics could be essential for optimal extrapolation accuracy. In the same vein, Li et al (2019) found that adding turbulence intensity as an input greatly improves wind speed forecasting accuracy, showing that the input feature set may be highly influential for machine learning tools applied to meteorological problems. Following such developments, the present study focuses on proper extraction and selection of meteorological features, across multiple sites, for a neural network designed for vertical extrapolation of wind speed.…”

Section: Introductionmentioning

confidence: 95%

Decreasing Wind Speed Extrapolation Error via Domain-Specific Feature Extraction and Selection

Vassallo¹,

Krishnamurthy²,

Fernando³

2019

Preprint

View full text Add to dashboard Cite

Abstract. Model uncertainty is a significant challenge in the wind energy industry and can lead to mischaracterization of millions of dollars' worth of wind resource. Machine learning methods, notably deep artificial neural networks (ANNs), are capable of modeling turbulent and chaotic systems and offer a promising tool that can be used to produce high-accuracy wind speed forecasts and extrapolations. This paper quantifies the role of domain knowledge on ANN wind speed extrapolation accuracy using data collected using profiling lidars over three field campaigns. A series of 11 meteorological features are used as ANN inputs and the resulting output accuracy is compared with that of a simple power law extrapolation. It is found that normalized inputs, namely turbulence intensity, normalized current wind speed, and normalized previous wind speed, are the features that most reliably improve ANN accuracy, providing up to a 52 % increase in extrapolation accuracy over the power law predictions. The volume of input data is also deemed important for achieving robust results. One test case is analyzed in-depth using dimensional and non-dimensional features, showing that feature normalization drastically improves network accuracy and robustness for uncommon atmospheric cases.

show abstract

“…Based on the territory of use, some studies use a global area [8] while other studies are predictive of specific regions [9]. Meanwhile, climate prediction has a time frame, such as daily [8] of solar forecasting, hours of wind speed [9], hours of temperature [10], multi-step time of wind speed [11], extreme climate every day [5]. Other research estimated rainfall in a short time or hours [12], days [13], weeks [14].…”

Section: Introductionmentioning

confidence: 99%

Precipitation prediction using recurrent neural networks and long short-term memory

Priatna

Djamal

2020

TELKOMNIKA

View full text Add to dashboard Cite

Prediction of meteorological variables such as precipitation, temperature, wind speed, and solar radiation is beneficial for human life. The variable observations data is available from time to time for more than thirty years, scattered each observation station makes the opportunity to map patterns into predictions. However, the complexity of weather variables is very high, one of which is influenced by Decadal phenomena such as El-Nino Southern Oscillation and IOD. Weather predictions can be reviewed for the duration, prediction variables, and observation stations. This research proposed precipitation prediction using recurrent neural networks and long short-term memory. Experiments were carried out using the prediction duration factor, the period as a feature and the amount of data set used, and the optimization model. The results showed that the time-lapse as a shorter feature gives good accuracy. Also, the duration of weekly predictions provides more accuracy than monthly, which is 85.71% compared to 83.33% of the validation data.

show abstract

Multi-step wind speed prediction based on turbulence intensity and hybrid deep neural networks

Cited by 72 publications

References 46 publications

Meteorological Temperature and Humidity Prediction from Fourier-Statistical Analysis of Hourly Data

Meteorological Temperature and Humidity Prediction from Fourier-Statistical Analysis of Hourly Data

Decreasing Wind Speed Extrapolation Error via Domain-Specific Feature Extraction and Selection

Precipitation prediction using recurrent neural networks and long short-term memory

Contact Info

Product

Resources

About