The impact that rescheduling aircraft maintenance may have on both the network and maintenance roster is hard to grasp due to the inherent complexity of the problem. Various works have been dedicated to developing methods for recovering a disrupted airline schedule, with the primary focus being the recovery of the lines-of-flight in the days after the disruption. More recently, attention has been paid to the specific recovery of aircraft maintenance, constrained by the resources required. However, no comprehensive method exists that combines both network and maintenance scheduling in order to assess the effects of rescheduling maintenance in favor of flight execution. This paper presents a novel framework that provides decision-makers with the opportunity to assess the outcome of a choice they are about to make: what are the short- to medium-term effects of delaying airline maintenance in favor of operating a flight instead of canceling it? At its core lies a mixed-integer programming formulation, which considers a double space-time network, allowing the integral scheduling of network rotations and maintenance tasks in ground slots. A case study involving a heterogeneous fleet of 54 aircraft from a major European airline demonstrated the framework's effectiveness in scheduling 1,886 maintenance tasks and 471 flight rotations across 58 ground slots. The framework shows that preventing a rotation from cancellation by rescheduling maintenance results in approximately 13 schedule changes and 10 hours of additional delay within a two-week period, compared to the same schedule without the maintenance rescheduled. Thus, while the airline has prevented a rotation from cancellation, this came at the cost of multiple scheduling changes and additional network delays. It can be concluded that the initial formulation proves to be beneficial in assisting decision-makers in assessing the implications of delaying aircraft maintenance.

Efficient maintenance management requires an integrated approach that balances downtime (maintenance scheduling, MS) and uptime (tail assignment, TA). Current methods often use sequential decision-making, which neglects the interdependencies between MS and TA, resulting in sub optimal outcomes. Recent research has started to integrate MS and TA modelling, and separately also to integrate MS with predictive maintenance (PM) strategies. These trends highlight the need for a comprehensive approach that optimises these interdependencies and considers the uncertainties of prognostic models.

This research presents a new scheduling framework for optimal maintenance management that integrates PM, MS and TA. Historical sensor data is used with Gaussian Process Regression to predict the end-of-life of aircraft components and quantify prediction uncertainties. These uncertainties are incorporated into the decision-making process using Monte Carlo simulations. The resulting model assigns aircraft and tasks to maintenance slots and flight legs under various predictor scenarios, enabling a more adaptive scheduling strategy.

A case study with Swiss International Air Lines over a three-day period for five aircraft demonstrates the model's effectiveness, showing improvements in operational efficiency, reduced costs, fewer cancellations, and higher fleet availability. The model's predictions and schedules were validated by expert planners, suggesting that a holistic, adaptive scheduling approach has significant potential for operational improvements in the airline industry.

The scheduling of engine shop visits quickly becomes a complex problem to solve as the number of aircraft and engines increases. In recent times, different approaches have been used to tackle this problem and optimize schedules, reducing costs and increasing revenue. This paper formulates the ESV scheduling problem as a Markov Decision Process and develops a reinforcement learning model that includes parameters such as engine performance and life limited parts status, maintenance constraints, and temporal factors. A prioritization algorithm is presented to optimize the learning process and allow for the scheduling of larger fleets. The results show a slight better performance in comparison to a greedy policy when evaluating aircraft availability, flexibility to the initial parameters and reduction in use of spare engines. On the other hand, the reinforcement learning provided lower scores and higher number of removals of aircraft from operations. In conclusion, the methodology proved that reinforcement learning is a viably way to optimize the ESV scheduling process, however a fine tuning of parameters might be necessary to approximate scores to a real revenue and cost relation, and reduce the number of aircraft interventions.

This study presents a new approach for integrated tail assignment and maintenance task scheduling decision support. Wasted maintenance resources, restricted fleet availability for schedule flexibility, inconsistent planning, and no emphasis on schedule resilience are all inefficiencies that now arise from airlines' separate and manual tail assignment and maintenance task scheduling. The framework addresses these issues, something not yet discussed in the literature. The framework's objective is to deliver feasible plans that increase schedule efficiency (no cancellations, high fleet availability, high fleet health, and optimal use of maintenance resources) and schedule stability (limited number of late arrival disruptions during operations) the day before operation. At its core, a mixed integer linear programming optimization model abides by airline-specific requirements and priorities. The schedule is modeled using an innovative two-space time-space network (TSN), with one space dedicated to maintenance and the other network activities. Stability is introduced with additional ground arcs positioned after flights that, if assigned, allow the creation of flight-specific buffers based on flights' historical arrival delay distributions. Task scheduling is done in registration specific slots, based on aircraft's individual maintenance needs, to improve schedule efficiency. The performance of the framework is tested using a case study provided by a major European single hub-to-spoke airline, with a heterogeneous fleet of over 50 wide-body aircraft. Comparison against the airline's plans over seven days shows that the proposed approach can provide real-time decision assistance and produce more efficient and stable plans. A 17% reduction in maintenance time was achieved, consequently resulting in a 10% increase in fleet availability on the day of operations. This was possible through higher labor and task interval utilization, suggesting that the framework is more efficient at scheduling maintenance tasks. Lastly, the framework's plans were more resilient to arrival delays, reducing the number of disruptions and propagated delay on the day of operations by over 40%.

As the volume of telemetry data generated by satellites and other complex systems continues to grow, there is a pressing need for more efficient and accurate anomaly detection methods. Current techniques often rely on human analysis and preset criteria, presenting several challenges including the necessity for expert interpretation and continual updates to match the dynamic mission environment.
This paper critically examines the use of Deep Anomaly Detection (DAD) methods in addressing these challenges, evaluating their efficacy on real-world spacecraft telemetry data. It exposes multiple flaws in current DAD research, highlighting the tendency for performance results to be overestimated and suggesting that simpler methods can sometimes outperform more complex DAD algorithms.
By comparing established metrics for anomaly detection with newly proposed ones, this paper aims to improve the evaluation of DAD algorithms. It underscores the importance of using less accuracy-inflating metrics and offers a comprehensive comparison of DAD methods on popular benchmark datasets and real-life satellite telemetry data.
Among the DAD methods examined, the LSTM algorithm demonstrates considerable promise. However, the paper also reveals the potential limitations of this approach, particularly in complex systems that lack a single, clear predictive failure channel.
The paper concludes with a series of recommendations for future research, including the adoption of best practices, the need for high-quality, pre-split datasets, and the investigation of other prediction error methods. Through these insights, this paper contributes to the improved understanding and application of DAD methods, ultimately enhancing the reliability and effectiveness of anomaly detection in real-world scenarios.

Increasing runway capacity is a key objective for many airports worldwide, including Schiphol Airport. One of the restricting factors that has become more prevalent throughout the years lies in the task load of the Air Traffic Ground Controllers. To overcome this limiting factor, it becomes essential to mitigate this high task load. This study proposes a methodology for estimating ground controller task load at Schiphol Airport, which is used for analyzing the effectiveness of a mitigation method employed at Schiphol Airport, namely the splitting of the combined North-Centre sector. To achieve this, the research adopts the concept of task load estimation via dynamic density modelling, which has primarily been explored when applying it to the airborne segments of Air Traffic Management. The task load estimation models for the different ground control sectors are developed through a multi-phase approach encompassing data collection, candidate independent variable selection, a double correlation analysis, and a sequential search regression for model building. The research successfully constructed task load estimation models for four out of the five analysed sectors. Furthermore, two of these models are employed in a case study to evaluate the task load reduction strategy. The findings highlight that splitting the combined North-Centre sector to alleviate the task load experienced by the North sector ground controller can be an effective mitigation strategy, particularly when the Centre sector experiences a significant level of traffic load.

In the dynamic and complex world of air transportation, airports can serve as vital hubs that connect numerous of destinations and facilitate the smooth flow of passengers and aircraft. However, the efficient operation of hub airports can be significantly disrupted by various factors, one of which is the reduction in runway capacity. Runway capacity reductions can be caused by several factors such as maintenance activities, adverse weather conditions, or airspace congestion, resulting in flight cancellations and delays that impact carefully planned schedules of airlines. By evaluating the expected disruption leading to a runway capacity reduction a day before operations, proactive measures could be taken with effective consideration of aircraft, passenger and crew schedules. To address this challenge, a stochastic, multiple scenario, decision support tool was created with the aim of providing the airline with required proactive cancellations for an expected disruption. A case study is performed on the framework in collaboration with a major European airline providing data. Validation has been performed by benchmarking the results of the framework with actual decisions and results of a deterministic tool including only a single scenario. Finally, a new disruption scenario was used to demonstrate the capabilities of the framework and show the sensitivity of some main input parameters. Benchmarking results of the framework with actual implemented cancellations and results of a deterministic tool show that the framework is capable of providing cheaper cancellation options, and overall, a cheaper cost of solving the disruption while complying with crew reserves and maintaining valuable passenger connections.

Bachelor Thesis: Design Synthesis Exercise 2023
Group 13: Electric General Aircraft
in collaboration with
Group 10: Mid-Flight Recharging System

Addressing the pressing need for sustainability amid climate change, the aviation sector is under pressure to mitigate its significant carbon emissions. As aviation accounts for over 2% of global emissions, achieving zero-emissions aviation by 2050 is crucial. The challenge lies in finding alternatives to jet fuel. One option is battery-powered electric aviation, but limitations such as range constraints and emissions during production and disposal must be considered.

This report proposes a solution to extend the range of all-electric passenger aircraft. The ElectriFly aircraft can currently fly 600km at 110m/s before requiring recharging. To achieve perpetual flight capability, a mid-flight recharging system is introduced using a drogue and probe system, similar to military refueling technologies. This approach not only increases range but also reduces turn around time at airports.

Powered by four parallel-connected solid-state batteries with an energy density of 600kWh/kg, the ElectriFly aircraft features 14 propellers using distributed propulsion and potential plasma actuators. This setup improves aerodynamic performance, reduces noise, and allows for regenerative braking, easing the burden on the recharging drone. Real-time data sensors enable predictive maintenance, minimizing downtime and providing pilots with crucial flight anomaly information.

This research presents a promising solution for electric aircraft limitations through mid-flight recharging. The proposed ElectriFly aircraft offers extended range capabilities and improved operational efficiency, supporting the goal of sustainable aviation.

Air cargo is vital for efficient global supply chains, enabling the rapid transport of valuable goods. Combination airlines, a significant player in the industry, use both dedicated full freighters and the belly space of scheduled passenger flights to transport cargo. This research develops a decision tool that combination airlines can use to optimize the routing and scheduling of full freighters. By incorporating anticipated cargo demand data as an input, the tool enables the most cost-effective routing and scheduling of full freighters, also taking into account the available cargo carrying capacity offered by passenger flights. The tool combines and leverages the intrinsic advantages of passenger flights (frequencies) and full-freighters (capacity). The decision tool relies on a Mixed Integer Linear Program (MILP) to minimize operational costs while adhering to operational constraints faced by a combination airline. Although this method performs well for small instances, it suffers from long computational times when applied to larger problem instances. To address this, modifications are introduced to shorten the runtime by reducing the problem size. Specifically, clustering is used to group cargo requests together, and column generation is used to efficiently select optimal routing decisions. These modifications lead to a shorter runtime, albeit with an acceptable increase in operating costs. The decision tool serves multiple purposes. In the short term, it can be used to route full freighters and identify non-profitable full-freighter origin-destination pairs, among other applications. Additionally, it can be used to analyze, simulate, and provide recommendations for future long term full freighter fleet planning decisions.

Optimizing the routes of firefighting aircraft can reduce the time it takes to contain wildfires and make sure they remain within control. In this paper, a novel formulation of the Vehicle Routing Problem (VRP) is developed to improve aerial firefighting operations by optimizing aircraft routes. The formulation is a capacitated split delivery multi-trip VRP with time windows and hierarchical objectives. The primary objective is to minimize the time of carrying out all requested drops, and the secondary objective is to minimize the total flight time. Two types of aircraft are used: Scoopers and tankers. The main difference is that scoopers can refill their water tank from a water body. By easily adjusting the capacity and speed of the aircraft, most firefighting aircraft can be modelled using these two types, including helicopters. The program allows the user to input the number and types of aircraft available, the locations of airfield, fires, and nearest water body, intensity of each fire, and more. Several random cases and case studies were solved within the expert-recommended time limit of 5 minutes, yielding reasonable optimized routes. The problem is scalable and sizes ranging from one to 80 drops were tested and solved within 22 minutes. Furthermore, given a certain fire situation, the model can be simulated with various aircraft combinations to gain insights into fleet optimization. In one case study, it was demonstrated that replacing a scooper with a tanker can result in halving the total operation time. Strategic fleet planning is also demonstrated in a case study with the use of a Monte Carlo simulation, in order to compare the performance of different fleet options for a given setting. Therefore, the model is not only applicable in live situations, but can also be used as a supportive tool in planning for upcoming fire seasons, or reviewing and learning from past fires

The complexities associated with airline operations require operations planning to be divided into multiple problems solved sequentially by the respective departments: (1) network planning, and (2) maintenance planning.
Furthermore, airline operations take place in an intrinsically uncertain environment, which requires the development of robust plans and the use of effective recovery policies. Despite the close interaction of network and maintenance plans in this dynamic environment, it is current airline practice to evaluate plans from the two domains separately, thus not representing airline operations from an integrated perspective. To this end, a modular, stochastic, discrete event simulation model of airline operations named ANEMOS (Airline Network and Maintenance Operations Simulation) is presented in this paper. The model integrates network and maintenance operations dynamics, allowing the evaluation of plans, policies, and scenarios from both domains. The model is validated using data provided by a major European airline, and it is shown that the simulated results closely resemble the airline's historical operational performance. Finally, the model's capabilities are demonstrated with a case study investigating the effects of adding a second reserve aircraft to a fleet of fifty wide-body aircraft. Results show that the second reserve is capable of reducing cancellations by 55%, but the lost revenue associated with keeping an aircraft non-operational make it a very costly solution, with the avoided costs of disruptions quantified at 6.2% of the lost profit.

With the aviation sector growing each year, the need for a reduced climate impact is becoming increasingly important. Electrification of the propulsion system is believed to offer promising avenues in achieving this reduction. Additionally, airlines operating these aircraft have to adapt their operations and network to optimally utilize these aircraft. This research presents a methodology for the coupled design of a hybrid-electric aircraft fleet with strategic airline planning to optimally serve a specific network. The objective is to maximize the airline profit and minimize the network CO2 emissions. Aircraft design trade-offs in payload, range and runway length will guide the creation of new aircraft until the optimal aircraft fleet is determined. The methodology is tested in a case study for the regional airline network of SATA Air Acores. The study investigates the impact of introducing new hybrid-electric aircraft designs in the fleet on the creation of new aircraft, the aircraft allocation and the network performance. By directly integrating hybrid-electric aircraft design (having a parallel hybrid architecture) with strategic airline planning, it is possible to reduce the network CO2emissions by -11% at the cost of an airline profit decrease of -13%. When including a climate optimization, an additional reduction of network CO2 emissions is achieved of -27% with a small additional decrease in profit of -1%. Network profitability and climate impact are mainly dictated by the fleet diversity and the assumed technology level of the batteries employed in the aircraft. This research highlights the importance of including climate optimization in the design of new aircraft and the need for more advanced hybrid-electric propulsion architectures (such as distributed propulsion systems) to further contribute to climate impact reduction.

In recent years, the aviation industry has faced escalating environmental scrutiny, prompted by the global shift towards heightened environmental awareness. This study explores a collaboration framework set up by airlines which envision to partially or totally shoulder the environmental impact costs generated by their airborne emissions. The methodology developed includes the production of detailed emissions data to be employed during the airline’s planning strategy. The model conceived introduces a taxation approach exploited by airlines to include and incentive more environmentally sustainable solutions and, a collaboration mechanism based on the possibility to share passenger demand between collaborating airlines. The environmental incentive and collaboration are implemented in a Linear Programming Network and Frequency Planning model with the goal of maximising airlines profitability. A case study is analysed focusing on two airlines, KLM and TAP, collaborating to achieve improved environmental and economic performance. The airlines engage in a scenario where the environmental impact externalities generated by airborne operations are shouldered by the airlines themselves. Employing the model developed, the airlines explore solutions that either reduce the environmental impact or that solely maximise profit of the two airlines. The results underscore the potential benefits of including collaboration in the airline’s planning strategy, with a total 1.33% reduction in environmental impact when prioritising environmental cost abatement. Collaboration also demonstrated to be a consistent tool to achieve a profit recovery, showing up to a total anticipated profit increase of 2.21% when prioritising profitability or 0.71% when the priority is given to environmental impact reduction. In conclusion, this study highlights the potential efficacy of collaboration in addressing aviation’s environmental and economic challenges.

Efficient maintenance scheduling is a critical objective for aircraft carriers and Maintenance Repair and Overhaul Organizations (MROs) to minimize aircraft downtime. While predictive maintenance models have improved, accurately identifying materials, especially spare parts, for specific maintenance events remains challenging. This paper combines the challenges of identifying spare parts by MROs and aftermarket distributor demand models by developing a robust prediction model (SPSO-CM) to forecast subsequent customer-specific purchases of maintenance planning document (MPD) related spare parts, considering technical documentation and previous sales records. The proposed architecture employs a gradient-boosting algorithm with numerous data mining improvement techniques to predict the likelihood of a subsequent spare part purchase from a customer. A k-means clustering algorithm is used to group spare parts with similar characteristics, as certain specific spare part properties significantly influence demand prediction models. A unique feature selector and nested group K-fold TimeSeriesplit cross-validation method were developed and incorporated into the Bayesian search space to optimize hyperparameters and improve performance. Two test cases were simulated, and the results demonstrated that SPSO-CM is more effective in forecasting proprietary parts and frequently purchased spare parts than those with extremely lumpy demand patterns. Two potential applications for an aftermarket distributor are discussed, one from a customer-level perspective and another at a larger supply-chain level, highlighting its promising capabilities.

Airlines frequently deal with unexpected disruptions, which have to be resolved in order to resume their operations again. Decision-support tools help airlines with disruption management. However, long computational times are associated with integrated recovery models that compute globally optimal solutions, which makes these tools unfit for large airlines. Faster sequential approaches have been proposed, which recover one resource at a time, but often only provide local optima. This paper presents a real-time decision-support tool to solve the integrated airline recovery problem. The model simultaneously recovers the schedule and allocates aircraft and pilot pairs to the flights, while minimizing additional incurred costs. Passengers are implicitly recovered by considering missed connections. Recovery actions include delaying and cancelling flights, swapping aircraft, and swapping, deadheading or using reserve crew. A machine-learned ranking algorithm reduces the computational complexity of the problem by selecting a subset of the resources that are likely to be involved in the recovery plan, such that only part of the network is considered. The decision-support tool was evaluated on disruption scenarios of one of the largest airlines in the world: Delta Airlines. The results show that the machine learning selection reduces the average computational time 15- fold compared to the integrated recovery model that uses the complete network, increasing the percentage of solutions computed under two minutes from 13% to 96%. The proposed model is able to find globally optimal solutions in 58% of the cases and yields similar results in terms of delays, but more flight cancellations in comparison with the globally optimal solutions. The proposed integrated model found a feasible solution to all disruption instances, while a benchmarked sequential model returned infeasible solutions in 4.4% of the cases. Besides, the sequential model produces 68% more flight cancellations, 4 times more aircraft sink node violations and twice as many crew sink node violations in comparison with the proposed integrated airline recovery model.

In the airline industry revenue and inventory is managed tightly, with the intent to sell each seat at the right price to the right customer at the right time. This does not always go to plan and seats remain unsold, or demand can be greater than what is available. The question then becomes how can capacity and revenue be managed more efficiently? How can capacity and revenue be maximized to ensure that seat spoilage is minimized. This thesis takes a look at this problem and aims to solve this problem by dynamically allocating capacity as well as becoming paired with revenue management processes. By being able to move capacity around to different flights, it is believed that revenue can be maximized and that seat spoilage can be minimized, creating a more effective use of a perishable product. The goal of this thesis is to show that increases are possible, even if by small margins. This thesis looks to use real-world data to help support its findings along with a random distribution to pair alongside a Monte Carlo simulation to help ensure accuracy. In this data driven methodological approach revenue management and a capacity swapping algorithm are integrated. The revenue management module creates a series of seat inventory allocations as well as demand forecasting. This information is then fed into a simulation which will then simulate the entire booking life- cycle of a fight. Here each flight will build its passenger load and will swap aircraft with other aircraft of varying capacities when the algorithm has determined it is necessary and would create a positive increase in revenue and passenger counts. To assist in this approach a further random forest model is developed to help make informed decisions. These changes are further evaluated by the algorithm to ensure that every change that was made was beneficial and if so will accept the swap and continue to build loads as this new flight. Different case studies were developed that introduced different timing scenarios as well as different decision making tools. The two different decision making tools would use either the random forest model that was developed or a simple logic model that would look at estimated final revenue and estimated final passenger loads. The timing scenarios would look to see when swapping would be needed, whether it is a true daily swap or if swaps can be performed once a month. Ultimately there were positive results, with each scenario tested yielding positive increases in the three KPIs examined (Load Factor, Passengers, Revenue). Each scenario involved testing either the simple logic, or random forest model with varying swapping intervals. The best performing scenario was one where flights were allowed to be swapped daily using a random forest decision model. In this best scenario load factor saw an increase of 1%, the number of passengers increased 1.4%, and revenue increased 0.9%; all against the baseline scenario where zero flights were swapped. This translates into a 36,000 USD weekly increase for a small sample in a route network. Across a full route network, even greater increases can be achieved.

To reduce the environmental impact of the aviation sector in Europe, climate goals have been set as stated in the Paris Agreement and the European Green Deal. This reduction can be achieved by fleet replacement with electrified aircraft. Challenging is that electrified aircraft are yet in a conceptual phase and most likely suitable for smaller airline networks. This research presents a study on how regional airlines can strategically adapt to fleet replacement by not-yet-designed electrified aircraft. Earlier research shows interest in integrating strategic airline planning and aircraft design, but the design and operation of electrified aircraft are not yet considered.
A methodology is presented to couple strategic airline planning and electrified aircraft design. In the strategic network and fleet planning model, the operation of electrified aircraft is considered. The electrified aircraft design is provided by the Initiator and aircraft configuration options by the Breguet Range equation. Based on current network information, this approach considers the adaptation of an airline on not-yet-designed electrified aircraft. Impacts on network performance, such as profit, CO2 emission and operating routes, as well as chosen aircraft are studied.
In this thesis, the methodology is tested with a case study of the airline SATA Air Açores. The impact of replacing their current fleet with electrified aircraft on the network performance, development, and aircraft design choices is considered. Furthermore, the impact of airlines’ choices and external factors are evaluated. Under the assumptions, the adaption of electrified aircraft seems profitable for the airline and the CO2 emission is highly reduced. A choice for smaller fully-electric and medium-sized hybrid-electric aircraft is presented.

This research aims to investigate and experiment on a state-of-the-art problem to treat the non-monotonic behavior of fault progression trends in predictive maintenance. Well-established algorithms and literature are researched for fault progression prognostics, however, not considerable attention has been given to
monotonic constraints at a preprocessing stage. A non-monotonic trend carries complex information which has outliers and nonessential signal values. The goal of the project is to motivate the usage of monotonic constraints to treat non-monotonic signals of a degrading component. The problem is presented as follows: Determining if the monotonic constrained method at a preprocessing step shall assist prognostics to estimate the remaining useful life of a component accurately. To explore this research, a monotonic constraint - Average Conditional Displacement (ACD) is used at the preprocessing step of a model, in comparison with regular preprocessing methods. The model is experimented on the NASAs simulated C-MAPSS datasets of turbofan engine and modelled with two prognostics algorithms. The model performance is measured
with performance metrics. The results showcase that by treating non-monotonic trends with monotonic constraints does improve the prognostics. However, they are not significantly advanced compared to other preprocessing steps.