Improved Prediction of Runway Usage for Noise Forecast

Abstract

The research deals with improved prediction of runway usage for noise forecast. Since the accuracy of the noise forecast depends on the robustness of runway usage prediction, improved accuracy of runway usage prediction will result in improved accuracy of noise load prediction. The main motivation behind this research is that the current method for runway usage prediction does not account for certain factors such as anticipating changes in weather forecast, additional meteorological phenomena, operational disturbances, which influence the controllers in the runway configuration selection decision-making process. The main objectives of the research are to develop runway usage models with increased accuracy of runway usage prediction compared to the current models and to investigate the effect of the developed models on the results of the computations of the noise load around the airport. The novelty of this research comes from improving the accuracy of runway usage prediction and noise forecast and identification of the main factors that influence runway usage. Most of the recent research in this area focuses on runway usage prediction for tactical and strategic planning. There has been very few research carried out on runway usage prediction for noise forecast and this research aims to fill that knowledge gap. Based on literature study, it was identified that modeling with the use of historical data (empirical modeling) can be used to predict runway usage more accurately since it includes the controller’s decision-making patterns. Two prediction algorithms were chosen for the development of runway usage models: Nearest Neighbor and Neural Networks. Two approaches were chosen for runway usage prediction: determination of runway usage directly and determination of runway usage from runway combination prediction. The combination of the prediction algorithms along with approaches was used to develop four runway usage models. The main factors that influence runway usage were identified and used as predictors for the models. The developed models were verified by a comparison with the actual runway usage. Various predictors were analysed to see if it improves the runway usage prediction accuracy. The developed runway usage models were compared with each other in terms of noise forecast accuracy. Based on the effect of the developed runway usage models on the results of the noise load computations around the airport, the runway usage model that resulted in the highest noise forecast accuracy was identified to be the model developed using neural networks that determines runway usage from runway combination prediction. The main factors that influence runway usage were identified to be – wind direction, wind speed, visibility, period of the day, required capacity, type of operation (landing/take-off), and origin/destination. The developed runway usage models were validated for Schiphol airport and can be applied for other complex multi-runway airports like Schiphol airport. This will aid in noise load prediction around the airport for transparency with surrounding communities, determining annual usage plan and analyzing noise mitigation measures.