Here is a summary of the technical program as of 13 May 2019, subject to change and final approval.
Tuesday 4 June 2019
Keynote Address
Ekbel Bouzgarrou
Vice President Information Technology Distributed Services
Air France - KLM
Tuesday 4 June 2019 - Friday 7 June 2019
Airline Operations Presentations
TENTATIVE
(Scheduling and order not yet determined -- soliciting additional presentations. Abstracts in order of submission)
Blockchain-Inspired Integrated Airline Operations Recovery
Zachary Bluestein, Dr. John-Paul Clarke, Dr. David Goldsman
Georgia Institute of Technology
A novel algorithm for combinatorial, multi-agent optimization has been developed for integrated airline operations recovery. When inclement weather or other delays occur, the re-assignment of crew, passenger, and equipment to flights is optimized within banks of departing flights. This optimization is done in accordance with FAA, union, and airline regulations. A “master problem” oversees the combinatorial optimization between flight banks and distributes relief crew and available aircraft as necessary to alleviate systemic delays. Using a blockchain-inspired record keeping system, all decisions made during the optimization process are recorded in a decentralized, secure, and verifiable ledger. This increases system awareness, simplifies data storage and analysis, and maintains information security.
Modeling Activity at Charlotte Douglas Airport for Strategic Planning of Airline Operations
L. Doug Smith, Canser Bilir
University of Missouri-St. Louis
We demonstrate the power of coupling statistical models based on detailed Aerobahn flight data with a discrete-event simulation model (structured as a network of staged queues) to support planning for airline operations at a major U.S. hub airport. With the Aerobahn data and other information (e.g., compass bearing and distance to the airport of origin or destination), we construct statistical models for arrival delays, turnaround times, and departure delays. and embed them into the simulation model, which may be used to experiment with various strategies for reducing delays and optimizing use of transportation assets. Our presentation illustrates the statistical modeling involved, features of the simulation model, the model-validation process, and insight that may be derived from the simulation model.
Centralized Deicing Facilities During Winter Operations
Ryan Leick, Aaron Organ, Jordan Stacy
Utah Valley University
Several airports in North America and Europe have constructed centralized deicing facilities (CDF) to improve winter operations. Reduced taxiout times during adverse conditions are a supposed benefit to CDF’s but airlines have reported an increase following construction. Data was provided by masFlight for over 5 million flights at 15 large and medium hub airports; 5 airports with true centralized deicing facilities and 10 airports with various of deicing practices. Traditional operating metrics at CDF airports were compared to peer airports. Contrary to expectations, CDF's do not improve taxi-out times during adverse winter operations. However, analyzing time to takeoff (sch out – act off) showed negligible differences at CDF and peer airports. Arrival variance at CDF airports was less relative to peer airports. CDFs free up gates for arriving aircraft and relocate queues away from concourses reducing apron congestion.
Digital Operations Transformation
Fernando Bosch
BCG
The BCG - KLM partnership has advanced since last year with optimization and AI tools now integrating advanced machine learning to have decision support tools interacting with each other across silos (OCC, Crew, Ground Services) and across time horizons (long/ medium term planing and day of execution. We would like to discuss how the models interact with each other and how that enables changes in the way of working, required skill sets, planning processes, etc.
Fast-Time Gate to Gate Simulation for Today & Tomorrow
Anthony Weatherington
Jeppesen Sanderson, Inc.
TAAM is a comprehensive, commercially available gate to gate simulation tool,capable of simulating an airline network including aircraft operations at airports and within airspace and is used worldwide to support complex decision making, planning, and analysis.
Recent sim projects have focused on airport planning and air navigation services. Key projects will be presented with methodology and outcomes to showcase the means by which simulation successfully assists in creating efficiency gains and decision making for airports and ANSPs. Potential airline use cases will then be suggested.
Goal is to familiarize attendees with the impact fast-time simulation has on the aviation domain & solicit input & opinion on usefulness to an airline context. A free-form discussion will be encouraged to learn how attendees might apply such a tool in research as well as identify shortcomings or areas of improvement to make the process or tool more utilitarian to air carriers.
Machine Learning Application for Automatic Analysis of Aircraft Operations Interruption Reports
Wlamir Olivares Loesch Vianna, Juliano Elias Cardoso Cruz
Empresa Brasileira de Aeronáutica S.A (EMBRAER)
Events such as aircraft failures or operation interruptions are recorded to provide reliability data to authorities, manage operations performance and provide directives for airframers and equipment manufacturers to engage in engineering campaigns for product improvements. In general, these events are reported as free text and must be interpreted and structured for further statistical analyses and relevant information creation. This process is repetitive and time consuming. This work proposes a novel method to automate this process by the usage of text mining and supervised machine learning techniques for several indications, such as: fault class; contingency action performed; and technical accountability reporting the potential responsible for the interruption (e.g. weather). A numerical implementation using historical field data is used to demonstrate the performance and potential benefits and drawbacks of the proposed method.
Using data science to optimize ground staff planning under delay uncertainty
Benoit Patacq
Air France
The creation of ground staff shifts is usually done as a succession of tasks without considering delays. However, airlines’ ground tasks are highly dependent on its flights and thus on their unexpected delays, which results in numerous and costly staff planning adjustments. In this context, the project aims at building stochastic shifts for Air France ground staff in order to anticipate delays and reduce planning disruptions on the day of operations.
Why is your flight late? Mining airline data to assess root-causes and impact of delay propagation
Goran Stojkovic
Boeing Global Services
Air transportation systems are exposed to daily disruptions, which have significant impact on operations. Airlines operate tight schedules to maximize resource utilization. However, the lack of sufficient buffers often result in propagating delays. In this paper, we propose a framework for automatic detection of root-causes of delays and their propagation effects. We test our framework on empirical data of several airlines. Presentation on one case study where airline is prone to delay propagation through passenger connections. Additionally, majority of those delays are related to airport capacity, resource allocation, and passengers, and mainly originate from the hub. Obtained results and future research focus areas are discussed.
Data Blind Spots Preventing Optimal IROPS Recovery Solutions
Rex Bull (SlickOR) and David White (Cirium)
Data without a solution is just data, and solutions without the best data are sub-optimal at best. This presentation will explore various data blind spots that prevent airlines from optimal IROPS recovery solutions. We will also outline a path to bring more relevant data to the solution earlier in the recovery process that enhances the solution and streamlines the downline impacts. For example, bringing more robust passenger data into the aircraft recovery optimization model.
You will learn how better and more relevant data can facilitate better solutions on which flights to cancel, delay, swap, and divert that minimizes the impact on schedules, passengers, and crew. This approach will move airlines closer to generating holistic aircraft and passenger recovery solutions that provide airline operations with the cost metrics and passenger impact necessary to make better decisions in the heat of the moment.
The need for Airspace Users-driven flight prioritisation in case of delay
Nadine Pilon
EUROCONTROL
With resumed air traffic growth for a few years now, the European Air Traffic Management Network is about to reach its capacity limits. This growth will continue to generate increasing delays, to flights and for passengers. There are two ways to address such increases in delay: strive to augment the capacity; or reduce the impact of the delay on airlines and passengers.
Our research aims to reduce the impact of delays on Airspace Users and passengers whilst not reducing the performance of the airports, flow managers and the network manager. The User Driven Prioritization Process (UDPP) provides additional flexibility for airlines within constrained airspaces and airports when delays occur. The concept allows a complete prioritization over a number of flights, beyond the existing slot swapping. Current validation aims to demonstrate a reduced impact for Airspace Users is possible and is not detrimental to the other ATM stakeholders.
Refining the decisions in the Operations Centre
Anders Bohlin
Jeppesen
In this presentation we will look at some comparisons between controlling parts of a process industry and controlling parts of an airline operations centre. In the process industry, in particular in refineries, automation, multi variable control and optimization of throughput on a profit basis has been around for decades. In most airline operations centres the majority of decisions are taken by and implemented by individuals. We will cover an approach where automation of optimization can be a way forward to control the operations centre with profit in mind.
Automated timestamp generation through machine learning for solid collaboration in turnaround management
Manuel van Esch
Mussie Beian
zeroG
Turnaround processes are a data black box. The lack of reliable turnaround process data makes it impossible to do a root-cause analyses of what caused delays or monitor the process. Solutions like A-CDM from Eurocontrol tried to solve this issue, but as they rely on manual input the data quality is still unreliable.
Our solution is to use an AI technology called object detection to automatically generate reliable turnaround process time stamps in real time. This data serves turnaround managers and OPS controllers on the day of operations to monitor the process and address issues before they cause delays. Beyond the day of OPS the data enables fact-driven SLA negotiations, root-cause analyses of delays, simulate the best turnaround process and resource allocation.
In out talk we share how the solutions works (challenges), how other industries use this technology and why the aviation industry can concretely benefit.
Understanding and overcoming barriers of implementing disruption management systems in airline operations control
Sebastian Heger
M2P Consulting
Continuous growth in commercial air transport in recent decades has made the management of operations increasingly complex. Since today’s highly competitive airlines are more severely affected by disruptive events and their effects than ever before, there is an urgent need for better decision support within the operations control centers in order to minimize the cost induced by these disruptions.
This talk will review methodological approaches to enhance the disruption management systems, which up to now are mainly a theoretical construct and have hardly been used, to be successfully established as practical systems. To deeply understand how decision-makers fulfill their different tasks, assumptions about business practices and processes as well as organizational designs will be further analyzed, limitations of legacy information systems are discussed and the variety of needed models and algorithms from the field of mathematical optimization are reinvestigated.
The die is cast, gear up for the unknown.
Davide Bardelli / Tim Nickel / Björn Bech
Lufthansa Systems
European Regulation 716/2014 (the Pilot Common Project) deadlines are approaching. The SESAR functionalities that are being deployed pose additional requirements on the OCCs in terms of processes, operational decision-making and workflows (including the way OCC interacts with the cockpit crews and the air traffic controllers) and structures. OCCs must have the competence to perform consistently with them, not least to reap the benefits the new environment will disclose. Those who won’t be capable of properly interacting with the other stakeholders could undergo an inefficient access to ATM resources and jeopardize the system-wide benefits expected by the upcoming ATM paradigm.
Building a Shorter Approach Fuel Saving Best Practice through Clustering
Nicolas Pinchemel
OpenAirlines
While Fuel Efficiency programs are spreading across the industry, there are untapped best practice opportunities. The ability for pilots to operate shorter than “standard” approaches depends on their own initiative, but an established best practice would significantly increase its adoption.
Beyond air traffic constraints, this requires knowing where shorter approaches can be performed, their likelihood, operating conditions and fuel savings. This way, the pilot is made aware of the opportunity at briefing time, and the airline can monitor and promote the adoption.
Together with our airline clients, we are developing a generic best practice for any airport and aircraft type, based on historical analysis and segmentation of the shorter approaches through a 2-stage clustering model. This addresses the challenges of flight trajectory comparisons and identification of “standard” approaches.
The results show average savings of 40 kilograms of fuel per approach.
Airline Disruptions and Holistic Recovery
Hocine Bouarab
GE Aviation
Dealing with an airline disruption involves a complex iterative decision process that starts by adjusting the flight schedule and tail plan then recovering the crew and re-accommodating the passengers. Often, flight and tail recovery decisions have to be re-adjusted based on the feedback of crew and passenger’s recovery teams. The existing technologies support the process without breaking its sequential nature. GE Aviation’s Digital Group is building the next generation of recovery solutions to support the airlines in solving the disruptions holistically. In this presentation, we’ll be discussing the challenges and opportunities in building such a solution and the future of airline recovery.
Tuesday 4 June 2019 - Wednesday 5 June 2019
MRO Special Session Presentations
TENTATIVE
Blockchain Enabled Resource Optimizer for Emergency Maintenance
Craig Fisher, Dr. John-Paul Clarke, Dr. David Goldsman
Georgia Institute of Technology
An innovative approach to applying combinatorial optimization within a blockchain data structure to streamline decision making in aircraft maintenance. This system is primarily designed to address siloing within emergency maintenance by providing real time guidance on where to position aircrafts, inventory, labor, and repair station resources. The use of a helper optimizer along with the transparent blockchain will greatly reduce the turn time from the time a request for work is sent, to the time all assets have been allocated and are moving to the worksite. Blockchain provides the added benefit of data logging since it’s an immutable service that will provide insight to regulators when needed. Due to the highly distributed nature of aviation, blockchain provides much needed real time transparency to decision makers and has many opportunities to event based optimization schemes.
Line Maintenance Check Optimization
Patrice Yapo
American Airlines
At American Airlines, Line Maintenance Planning is the department within the Technical Operations Division that is responsible for the scheduling of planned maintenance work on a fleet of 900 plus aircraft. Prior to 2018, the process was manual for all checks, but was especially onerous for certain scheduled overnight checks, that required large dedicated crews at specific maintenance locations. Line Maintenance Planning partnered with the Operations Research & Advanced Analytics team to develop the Optimizer; the Optimizer is an automated scheduling tool to assist the planning Supervisors in developing a plan for all overnight checks and provide visibility to that plan in real-time for all parties responsible for executing it, i.e., workload planners, aircraft routers, and line stations. In this presentation we will introduce the tool, including the model and its implementation as well as share some lessons learned during the development process.
A Case Study in Augmenting Engineering Knowledge with Analytics
Greg Jackson
Boeing
As data science and technology grow and evolve, they can be used to augment existing engineering expertise to optimize maintenance operations. Boeing has a long history of predictive maintenance products and services that have traditionally been powered by engineering expertise. Augmenting this core strength with analytics expertise improves both the quality and time-to-market for services.
This talk will review lessons learned from collaboration of engineers and data scientists on a prescriptive maintenance solution. The team leveraged insights from all stages of the component lifecycle to build a fully-deployed prescriptive maintenance solution based on engineering knowledge and powered by machine learning. This successful collaboration has not only turned hours of unscheduled maintenance into a single non-intrusive scheduled maintenance task, but also provided the blueprint for how engineers and data scientists can collaborate to improve asset reliability.
Use cases on machine learning and airplane health management
Wannes Meert
KU Leuven
Over the past few years KU Leuven, TUI fly and Boeing have looked at applications of machine learning for airplane health management. Since airplanes collect a plethora of flight parameters there is ample opportunity to apply advanced data analytics methods. There are, however, a number of challenges which make that a naive machine learning approach is not adequate. For instance, the majority of the data is unlabeled, maintenance actions can change patterns present in the data, and each airplane might show unique behavior. In this talk we will detail the challenges we encountered and results we obtained in a number of use cases focussing on the cabin pressure controller, the auxiliary power unit, and the airplane spoilers.
Kick starting your analytics program, key tools and data required for maintenance optimization
Helmuth Naumer
IBM
Predictive maintenance, knowing parts will fail before they fail, is the Holy Grail for airline technical operations. Many airlines struggle with getting cognitive solutions up and running. Cognitive tools have matured, and offer a great foundation for analyzing task cards, log books, and other maintenance data. By providing an analysis of historical data to line maintenance or maintenance control, technical operations can resolve faults faster, which will reduce their maintenance costs, aircraft downtime, and AOG’s. Predictive maintenance combines this analysis with reliability and operational data/analysis/modeling. Those airlines that take advantage of these new technologies will greatly outperform their competitors and show much higher operational margins. This presentation will summarize the steps technical operations can take to initiate the journey to predictive maintenance while gaining many benefits along the way.