即時事件導向之航機延誤預測模式

A Real-time Event Oriented Flight Delay Forecasting Model

以往預測航機延誤之文獻，往往著重於探討航機因班表規劃不佳、相關作業時間不確定或機場容量限制所造成之延誤。然而，機場運作發生異常狀況，向來才是航機延誤之主因。即時且確實的預估異常狀況下航機之延誤狀況，對相關單位之運作、設備與人員調度很有幫助。本研究針對天候異常狀況，包括風向轉變造成機場起降方向轉變，與機場關閉造成航機停止起降等情境，討論其運作情形，並分別推導航機延誤模式，計算個別離、進場航機的延誤。在模式驗證部份，利用SIMMOD模擬模式建構研究所需之空域與空邊系統，並以其輸出結果，與航機延誤模式預測數值作比較。驗證結果顯示兩者平均誤差時間僅為0.78分鐘，顯示所推估之延誤模式能有效地預測航機延誤的時間。

Previous studies on the prediction of flight delay focused mainly on delays due to poor flight schedule, operational disruptions and airport capacity. Weather change, however, is often the root cause of flight delay. For the need of smooth operations, it's important for airport operators, airline staffs and air traffic controllers (ATC) to predict flight delays in abnormal weather conditions. In this research, a mathematical real-time delay forecasting model is derived to estimate the delays of each flight, in the cases of runway direction change and airport closing, on the basis of the actual performance of airport operations. To validate the developed delay forecasting model, a SIMMOD-based simulation model was developed, calibrated and validated using ATC data and observed operational data. Compared to the outputs of the SIMMOD model, the average error of the delays derived from the mathematical formulas is only 0.78 minute. The result suggests that the developed formulas can be used to effectively predict flight delays and, with this information, better decisions can be made in advance.