Titulo Estágio
Fleet Learning applied to Aircraft Maintenance
Áreas de especialidade
Sistemas Inteligentes
Engenharia de Software
Local do Estágio
Laboratory of Artificial Neural Networks (LARN)
Enquadramento
Aircraft maintenance is currently carried out following one of two strategies: reactive, (also known as ‘fix when it fails’) and preventive (also known as pre-scheduled or interval-maintenance). The fusion of the emerging concepts of IoT, Big Data and Machine Learning have introduced a new promising alternative based on Integrated Fleet Health Monitoring.
Integrated Fleet Health Monitoring will enable a strategy to only repair aircraft parts that are actually damaged or to replace parts that are close to failure. A reliable network of sensors will cover most aircraft systems and primary structural elements, monitoring their health on a permanent basis. Efficient Machine Learning data analytics algorithms and physics models will process the terabytes of data generated per day by these sensors, providing operators with situational awareness and status of the entire fleet. On-board, lean algorithms will be able to diagnose the existence of systems’ faults that can be immediately communicated to the maintenance control centre (i.e., using an edge computing approach). On the ground, powerful algorithms will fuse the coming data generated by the aircraft with historical data to precisely diagnose the existence, location, and severity of structural damages or systems’ faults. Other algorithms will use the same type of data to predict the remaining useful life of systems and structures, anticipating the need for maintenance.
This proposal will focus on the impact of Federated Learning principles on Integrated Fleet Health Monitoring. Federated Learning consists in a distributed machine learning approach which enables training on a large corpus of decentralized data. This will require the development of lean versions of previous models with the use of a more efficient prediction strategy and with a judicious selection of features. This methodology, critical to assure the scalability and resilience of the proposed model, will contribute to decrease the volumes of data that need to be moved, improving overall QoS.
Objetivo
The main objective of this proposal is to develop lean algorithms for on-board data processing and pre-diagnoses for critical aircraft systems, enabling Aircraft Fleet Health Monitoring.
The main goals of this proposal are:
(i) Define a testbench scenario and for evaluation;
(ii) Perform Data Pre-Processing, Normalization and Scaling;
(iii) Select appropriate ML algorithms for building the Predictive Model;
(iv) Perform Sampling and Model Evaluation and validate the overall Model with real data.
(v) Integrate the implemented components reusable platform.
Plano de Trabalhos - Semestre 1
• Overview of project context and importance of Condition Based Maintenance for Fleet Health Management
• Overview of available sensor data and typical data pipeline
• Overview of Machine Learning techniques, namely ones based on Federated Learning
• Propose initial predictive model workflow and prepare a first case study
• Prepare the intermediate report.
Plano de Trabalhos - Semestre 2
• Implement the simulated environment and the final version of the predictive model workflow.
• Evaluate experimental performance of proposed workflow: e.g., study parameter values; compare performance of the reduced datasets vs. previous results, etc.;
• Stable version of the workflow with the deployed algorithms and demonstrable case study.
• Prepare a research paper and the final version of the thesis.
Condições
This work will be carried out in the Laboratory of Neural Networks (LARN) of CISUC, where there will be a regular supervision and feedback on the behalf of the supervisors.
Familiarity with machine learning and data mining algorithms and software tools are essential. Participating students will acquire valuable knowledge and experience with model building and data science by mining massive datasets, which skills are currently in high demand for various technology employers due to the relevance to various applications.
Observações
Grant under REMAP Project will be available during the second semester internship depending on the 1st semester candidate evaluation. A 3-month scholarship of 745 euros per month is foreseen for this work, renewable for another 3 months. Logistics @Laboratory of Neural Networks (LARN)
DEI-FCTUC
Supervisors:
Joel P. Arrais (jpa@dei.uc.pt)
Bernardete Ribeiro (bribeiro@dei.uc.pt)
Orientador
Joel Perdiz Arrais
jpa@dei.uc.pt 📩