Stochastic simulation of predictive space–time scenarios of wind speed using observations and physical model outputs

Bessac, Julie; Constantinescu, Emil M.; Anitescu, Mihai

doi:10.1214/17-aoas1099

Cited by 16 publications

(17 citation statements)

References 65 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first is a synthetic test problem meant to illustrate the methodological details described in Section 3. The second is our motivating application where G is a space-time hierarchical Gaussian process used for wind speed forecast [1].…”

Section: Extracting Parameter Correlationsmentioning

confidence: 99%

“…In this section we apply the proposed method to analyze the motivating statistical model [1]. The model aims at forecasting wind speed by fusing two heterogeneous datasets: numerical weather prediction model (NWP) outputs and physical observations.…”

Section: Analysis Of a Space-time Gaussian Processmentioning

confidence: 99%

“…This paper provides a framework to use existing GSA tools along with tools from statistics to address these challenges and yield a new GSA approach for analysis of statistical models. This work is motivated by a statistical model that fuses two datasets of atmospheric wind speed in order to provide statistical prediction of wind speed in space and time [1]. The predictions are generated from a Gaussian process whose mean and covariance are parameterized through a large number of parameters, which are determined by numerical optimization.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Global Sensitivity Analysis for Statistical Model Parameters

Hart¹,

Bessac²,

Constantinescu³

2019

SIAM/ASA J. Uncertainty Quantification

Self Cite

View full text Add to dashboard Cite

Global sensitivity analysis (GSA) is frequently used to analyze the influence of uncertain parameters in mathematical models and simulations. In principle, tools from GSA may be extended to analyze the influence of parameters in statistical models. Such analyses may enable reduced or parsimonious modeling and greater predictive capability. However, difficulties such as parameter correlation, model stochasticity, multivariate model output, and unknown parameter distributions prohibit a direct application of GSA tools to statistical models. By leveraging a loss function associated with the statistical model, we introduce a novel framework to address these difficulties and enable efficient GSA for statistical model parameters. Theoretical and computational properties are considered and illustrated on a synthetic example. The framework is applied to a Gaussian process model from the literature, which depends on 95 parameters. Non-influential parameters are discovered through GSA and a reduced model with equal or stronger predictive capability is constructed by using only 79 parameters.

show abstract

Section: Extracting Parameter Correlationsmentioning

confidence: 99%

Section: Analysis Of a Space-time Gaussian Processmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Global Sensitivity Analysis for Statistical Model Parameters

Hart¹,

Bessac²,

Constantinescu³

2019

SIAM/ASA J. Uncertainty Quantification

Self Cite

View full text Add to dashboard Cite

show abstract

“…Standard Kalman filter equations (Kalman and Bucy, 1961;Kalman, 1960) KF.KalmanFilter Perturbed-observation (stochastic) EnKF (Burgers et al, 1998;EnKF.EnKF Houtekamer and Mitchell, 1998) Deterministic EnKF (Sakov and Oke, 2008) EnKF.DEnKF Ensemble transform Kalman filter (ETKF) (Bishop et al, 2001) EnKF.ETKF Local least-squares EnKF (Anderson, 2003) EnKF.LLSEnKF Hybrid Monte Carlo (HMC) sampling filter hmc_filter.HMCFilter Family of cluster sampling filters multi_chain_mcmc_filter.MultiChainMCMC A vanilla implementation of the particle filter (Gordon et al, 1993) PF.PF the observation time, the assimilation time, the observation vector, and the forecast state or ensemble, are also passed to the constructor upon instantiation and can be updated during runtime. Table 3 summarizes the filters implemented in the initial version of the package, which is DATeS v1.0.…”

Section: Filtering Algorithm Dates Implementationmentioning

confidence: 99%

Authors Response to Review Comments

Attia¹

2018

Preprint

View full text Add to dashboard Cite

A flexible and highly extensible data assimilation testing suite, named DATeS, is described in this paper. DATeS aims to offer a unified testing environment that allows researchers to compare different data assimilation methodologies and understand their performance in various settings. The core of DATeS is implemented in Python and takes advantage of its object-oriented capabilities. The main components of the package (the numerical models, the data assimilation algorithms, the linear algebra solvers, and the time discretization routines) are independent of each other, which offers great flexibility to configure data assimilation applications. DATeS can interface easily with large third-party numerical models written in Fortran or in C, and with a plethora of external solvers.

show abstract

“…Therefore, when building ARMA model for wind speed prediction, the time series must be non-stationary by adding trend and periodicity. References [6] and [7] proposed different wind speed prediction models based on the statistical method. 3) Artificial Intelligence (AI) is a good prediction method developed in recent years.…”

Section: Introductionmentioning

confidence: 99%

Untitled

2019

Adv. Electr. Comp. Eng.

View full text Add to dashboard Cite

Wind power is one of the most promising means of power generation. But the time-varying of wind speed is the most fundamental problem for power generation control system. Therefore, accurate wind speed prediction becomes particularly important. However, traditional wind speed predictions often lack consideration of the influence of atmospheric dynamic system. And few papers have introduced VMD method into the field of wind speed prediction. Thus, combined with four neural networks, this paper develops a wind speed prediction method based on Lorenz system and VMD, obtains LD-VMD-Elman wind speed prediction model. Simulation results show that: 1) As for wind speed prediction, Elman neural network has higher prediction accuracy and smaller error. 2) The models which added Lorenz disturbance can describe the actual physical movement of wind more accurately. 3) VMD can abstract the changing rules of different wind speed frequencies to improve the prediction effect. This paper makes up for the lack of consideration of atmospheric dynamic system. The Lorenz equation is used to describe the atmospheric dynamic system, which provides a new thought for wind speed prediction. The LD-VMD-Elman model significantly improves the accuracy of wind speed prediction and contribute to the power dispatch planning.

show abstract

Stochastic simulation of predictive space–time scenarios of wind speed using observations and physical model outputs

Cited by 16 publications

References 65 publications

Global Sensitivity Analysis for Statistical Model Parameters

Global Sensitivity Analysis for Statistical Model Parameters

Authors Response to Review Comments

Untitled

Contact Info

Product

Resources

About