Multi-model ensemble wake vortex prediction

Abstract

Tag der mündlichen Prüfung: 25.07.2017 Diese Dissertation ist auf den Internetseiten der Universitätsbibliothek online verfügbar. Für meinen Großvater. v Eidesstattliche Erklärung Ich versichere hiermit an Eides Statt, dass ich die vorliegende Dissertation mit dem Titel "Multi-Model Ensemble Wake Vortex Prediction" selbstständig und ohne unzulässige fremde Hilfe erbracht habe. Ich habe keine anderen als die angegebenen Quellen und Hilfsmittel benutzt. Für den Fall, dass die Arbeit zusätzlich auf einem Datenträger eingereicht wird, erkläre ich, dass die schriftliche und die elektronische Form vollständigübereinstimmen. Die Arbeit hat in gleicher oderähnlicher Form noch keiner Prüfungsbehörde vorgelegen. Zudem versichere ich Eides statt, dass ich kein erfolgloses Promotionsverfahren bzw. Promotionsstudium zum Dr.-Ing/ Dr.rer.nat. durchgeführt habe. AbstractAs a response to lift, a complex flow pattern is shed from the wings of an aircraft that evolves into a pair of counter-rotating vortices. Their behavior is strongly influenced by the atmosphere and further underlies a complex interaction with the ground. Due to the rolling momentum that the vortices may induce and the forces that they can exert they pose a potential hazard to following air traffic, especially along the glide path and in ground proximity. To avoid dangerous incidents, separation regulations based on aircraft mass exist that is currently under revision in the context of RECAT. However, air traffic is expected to further increase in the future, making wake vortex encounters more likely. For this reason fast-time wake vortex models have been developed in the past that predict the vortex position and strength dependent on the aircraft parameters and the ambient conditions in order to avoid dangerous situations. Furthermore, they may reduce flight delays at congested airports as they allow adapting overly conservative separations tactically under certain conditions. However, forecasts include uncertainties that originate from inaccurate and highly variable initial and ambient conditions as well as from inadequate understanding and simplification of the underlying physics. As a consequence, deterministic forecasts are to be complemented by probabilistic envelopes.In the field of meteorology such envelopes are frequently computed by combining the forecasts of multiple independent models to quantify the model uncertainty. This approach, also known as Multi-Model Ensemble (MME), has not only shown to increase the reliability of the probabilistic forecast but also to enhance its deterministic skill. For this reason the capability of the MME approach to improve both deterministic and probabilistic wake vortex forecasts is assessed and existing approaches are further developed in this thesis. Three of the ensemble members have been provided in the frame of a NASA-DLR cooperation. The employed models comprise APA 3.2, APA 3.4 and TDP 2.1 (NASA) and are combined with the DLR model D2P. Several MME approaches are examined, comprising the Direct Ensemble Average...

Multimodel Ensemble Methods for Prediction of Wake-Vortex Transport and Decay

“…Please note that, unlike in previous sections, the standard deviation used for defining 2σ bounds is not based on any specific distribution. Sensitivity analysis conducted in the previous studies has shown good convergence property of the method [46]. Figure 9 shows the MCS results for a single WakeMUC landing.…”

“…In this case, the vortices are assumed to be further advected by the wind while staying at constant altitude. Although the models' sensitivities to the ambient conditions differ, it has been found in a previous study [39] that no so-called expert models exist. They are defined as models that show significantly better performance than the other models for a distinct meteorological parameter interval.…”

“…Model Overview A brief overview of the distinct features of the fast-time wakevortex models is given in Körner and Holzäpfel [25] and in Gloudemans et al [27]. In general, the fast-time wake-vortex models are empirical algorithms used for fast predictions of wake transport and decay, based on aircraft parameters and ambient weather conditions.…”

“…Körner and Holzäpfel [25] have shown that the ensemble approach is capable of improving the skill of wake-vortex predictions based on results achieved by the reliability ensemble average (REA) method. To further enhance the prediction skill (the deterministic prediction skill for REA could be improved by 1.5%), we employ here other methods including the Bayesian model averaging (BMA) [23] and Monte Carlo simulations (MCSs).…”

Probabilistic Multimodel Ensemble Wake-Vortex Prediction Employing Bayesian Model Averaging

“…In a previous study [15] the evaluation of suitable MME methods demonstrated that the ensemble approach can indeed improve wake vortex forecasts. The Bayesian Model Averaging [16] turned out to achieve better results for our application while being more robust than the Reliability Ensemble Averaging [17,18] and is thus chosen to be further refined. Compared to Monte-Carlo Simulations the probabilistic BMA-forecast could reach a much higher reliability in reproducing the vortices measured by lidar as demonstrated in Körner et al [19].…”