SimOps Simulator

Global air traffic is steadily increasing. The major airports are already working at their capacity limits. Also the available space on the air traffic routes decreases constantly by the rising traffic already causing considerable delays during flights. Under these conditions, it is becoming increasingly difficult to establish an economic schedule, i.e. a flight plan with the highest possible productivity of the aircraft but with as little as possible delays under as any outside influences.

The flight plan and network optimization of a major airline with many connections is a complex problem with many stochastic variables. The flight times depend on the weather conditions on the route, the selected runways depends on the wind direction and delays during takeoff and landing depend on the load on the airports. In addition, the flights are interconnected by transfering passengers and crews.

Using event-driven stochastic simulation allows to model and simulate the entire route network of airlines, including all relevant properties and operational processes. In this way, critical points of a flight plan can be identified and possible changes can be evaluated.

SimOps is used at Lufthansa AG for analysis of the continental traffic.

Simulation model

SimOps is an event-driven simulation system for modeling flight plans incl. all flights, airports, cabin crew and passengers. In the simulation model the following characteristics are considered for each flight:

  • Timely availability of the aircraft (possibly using a spare aircraft)
  • Availability of crews (incl. Into account the different attendance requirements of cabin and cockpit crews and consideration of Dead-Head crews)
  • Availability of the passengers (incl. waiting rules)
  • ATFM delays of starts by routes, times and starting and destination airports
  • Stochastic taxi-in and taxi-out times (depending on the selected parking position and the selected runway and the aircraft type)
  • Selection of the runway by wind direction and approach and departure direction
  • Capacity model to simulate the runway capacity
  • Stochastic flight times, their distribution may also depend on the weather conditions (clustering)
  • Stochastic transfer times for crews and passengers depending on start and destination gate
  • Necessary ground times of aircraft (depending on the airport and aircraft type)

With the help of the simulation model one can examine in a simple manner, which reticulated consequences individual delays will have and what impact load shifts at individual airports will have.

Features

Using the simulator SiimOps a complete air traffic network can be simulated on the computer.

The SimOps Simulator provides appropriate interfaces in order to load and process simulation models based on historical data. Since the simulation models are stored in xml format, at any time a good automatable access to the simulation process is possible.

Input data

For the simulation data about all flights and airports are needed. In addition, information for linking of the flights (transfering passengers and crews) are used. Depending on the specific questions the simulation model can get more details at various points. Simulations can be performed with or without crews and passengers. (Also only partially present crew data may be used.) Similarly, the airports can be designed differently in detail, this the capacity model for each airport can be turned on or off separately and fallback rules for taxi and ground times and models can be simulated, if no information are present for all airports for all aircraft types separately.

Simulation process

SimOps reproduces the individual movements of the aircraft, crews and passengers using event-driven simulation. All relevant parameters of the system (that is, all the flights, all airports, passengers and of information related to routes and time slots) are recorded continuously. All data is stored in form of mean and standard deviation. Moreover, the the delay distributions and the values various thresholds are recorded.

Performance

The simulator is designed for the simulation of complex flight networks (week schedules with more than 10,000 flights, more than 100 airports and more than 10,000 crew and passenger records). Through efficient memory utilization of the simulator is even in this case, only about 200 MB. With multi-core support SimOps could uses modern CPUs optimal.

In the simulation of a complex weekly timetable (about 250 planes with 10,000 flights) on a modern CPU (Intel Core 2 Duo, 2.66 GHz) can be expected statistically stable results after about one minute.

Output of the results

The simulation results are output in the form of tables and graphs. The results can thereby be displayed directly in the simulator or can be provided as HTML reports or as XML files for further processing by other programs.

The delay information is available on a per flight, per route, per airport and per half-hour interval basis. In addition to the mean and standard deviation the distribution of the delays in the form of a density and in the form of histograms are provided. In addition, the respective time performance at the usual threshold (0 min., 3 min., 5 min. and 15 min.) is outputed.

Interfaces

SimOps can load simulation models from XML files. The structure of the files is detailed documented in the online help for the program and there are dtd files available for automatic syntax checking of scenario files. Moreover, the names of the tags within the XML files can be configured to your needs accordingly.

When editing a simulation scenario in the scenario editor data for individual properties (as base load values for a runway or empirically flight time distributions) can be loaded from csv files.

The simulation results can be viewed not only in the program or be issued in the form of html reports and CSV tables, but also be stored in a xml file for further processing by other programs. The file format is documented here as well as the format of the scenario files and can also be configured in the same way.

Screenshots

The following screenshots show how model editor, simulation and statistical output of SimOps look like in practice.

The individual images can be enlarged by clicking on them.

Model editor

Scenario editor (editing the taxi times)
Scenario editor (editing the flights list)
Scenario editor (editing the day-type depending flight times)

Simulation

Running a simulation
SimOps server (using remote network computers for simulation)

Statistic output

Statistic output
Statistic output
Report generator
 

Sitemap  Contact  Data Privacy  Imprint
© TU Clausthal 2019