In-flight cost index optimisation upon weather forecast updates

Abstract

This paper presents an optimisation framework to compute the altitude and speed profiles of a trajectory in the execution phase of the flight, such that the expected total cost (ETC) of the operation is minimised (i.e. modelling the expected cost of delay and fuel -including arrival uncertainties -at the arrival gate). This is achieved with a two-stage optimisation strategy: a trajectory optimiser that minimises a generalised direct operating cost function, for a given cost index; and an upper-level optimiser, which obtains the best cost index that minimises the ETC. Several case studies are presented for different departure delays, while considering the impact of two different weather forecast updates too: a region with relative high head-winds appearing half way across teh flight; and a cold atmosphere scenario, with a tropopause altitude lower than standard conditions. ETC savings with respect to following the operational flight plan increase with departure delay, as expected. Due to the non-linearity of the cost function, however, the benefits of considering the weather update depend on the actual value of the departure delay, showing the convenience of integrating the proposed approach into a crew decision support tool in order to avoid sub-optimal decisions.

Considering Expected TMA Holding into In-flight Trajectory Optimization

“…This is done using a binary-search algorithm. 16) Note that even if CI is used to generate the trajectories, the final optimization minimizes the total cost function (i.e., ETC) and not DOC.…”

“…The impact of temperature, for instance, on the optimization is analyzed in more detail in another publication. 16) As shown in Fig. 4, the OFP for this scenario consists of an initial cruise at FL360 followed by a step-climb to FL380 approximately 850 NM from the destination airport.…”

“…trajectories in cold atmospheric conditions, the reader is referred to. 16) When the delay at FL100 while flying the OFP is expected to be 40-45 min, the Pilot3 solution proposes to immediately descend to FL340 at the triggering point (i.e., the TOC for these case studies) and perform a second descent at about half of the cruise to FL300. The case for an expected delay of 45 min is shown in Fig.…”