Periodic Reporting for period 3 - ReMAP (Real-time Condition-based Maintenance for Adaptive Aircraft Maintenance Planning)
Periodo di rendicontazione: 2021-06-01 al 2022-08-31
ReMAP reinforced the European leadership in aeronautics by developing an open-source Integrated Fleet Health Management (IFHM) solution for aircraft maintenance. The goal was to modernize aircraft maintenance by following adaptive condition-based interventions, according to which maintenance is only performed when needed, following continuous monitoring of the components’ health conditions. The project addressed the specific challenge of replacing fixed-interval aircraft maintenance inspections with adaptive condition-based interventions. A data-driven approach was implemented to achieve this, developing probabilistic health algorithms for systems and structures. A similar approach was followed to develop a maintenance management optimisation solution capable of adapting to the real-time health conditions of the aircraft fleet.
It is expected that ReMAP will have an estimated benefit to European aviation of more than 700 million euros per year due to a direct decrease in maintenance costs, reduced unscheduled aircraft maintenance events, and increased aircraft availability.
The main objective for ReMAP for the final period was to integrate multiple technologies into the IFHM solution and to demonstrate this solution in a 6-month demonstration. The demonstration and complementary studies allowed the assessment of the cost and reliability benefits and proposed standards, paving the way for implementing CBM in the European air transport system.
It is expected that ReMAP will have an estimated benefit to European aviation of more than 700 million euros per year due to a direct decrease in maintenance costs, reduced unscheduled aircraft maintenance events, and increased aircraft availability.
The main objective for ReMAP for the final period was to integrate multiple technologies into the IFHM solution and to demonstrate this solution in a 6-month demonstration. The demonstration and complementary studies allowed the assessment of the cost and reliability benefits and proposed standards, paving the way for implementing CBM in the European air transport system.
To fulfil the main ReMAP objective, six main technical goals were defined. These technical goals are listed below, with a summary of the work carried out towards their achievement.
Objective 1 – develop an integrated approach for CBM
The integration work has been developed in the last period. Eight PHM solutions and a couple of maintenance planning models were integrated into the IT platform. This integration was tested and validated in the successful 6-month Demonstration Exercise.
Objective 2 − explore and optimize the use of different sensing technologies for structural health management (SHM).
The work performed involved the design, the specifications definition, the manufacture, and preliminary tests of four sensing technologies to be used for SHM. In particular, the equipment of 40 L1 (level 1) coupons and 8 L2 (level 2) coupons with piezoelectric elements, optical fibre sensors, and acoustic emissions sensors, together with the associated software for laboratory tests.
Objective 3 − develop data-driven probabilistic algorithms for aerostructures damage monitoring (diagnosis) and remaining useful life (RUL) estimation (prognosis)
An unprecedented lab test campaign was performed during the project. Several diagnostic methodologies targeting the various levels of SHM (i.e. detect, locate, quantify damage) were developed and successfully tested in lab-scale scenarios (details in several ReMAP publications). Anomalies are detected with success rates from 95.2% to 100%, based on methodologies leveraging strain readings with optical fibre or distributed sensing, acoustic emissions, or lamb waves. Results from Lamb wave testing show up to 10% of the Mean Absolute Percentage Error in the localization and up to 15% in the sizing of impact damage or disbond in the skin/stringer interface.
Objective 4 − develop a hybrid approach, combining machine-learning-based data analytics algorithms and physics-based models for diagnostics, prognostics and health management (PHM) of dissimilar aircraft systems.
The consortium developed data-driven PHM algorithms that have been trained and tested in 10 aircraft systems from four aircraft types. Together, one conceptual algorithm using the Federated Learning concept takes advantage of the platform architecture. An initial physics-based model concept was developed to improve route cause analysis of failure prediction from data-driven methods. Develop user interfaces to interpret the results obtained by some of the developed algorithms.
Objective 5 − develop an efficient maintenance packaging and schedule optimisation algorithm for real-time adaptive fleet maintenance management.
The consortium has developed a set of machine-learning algorithms, and scheduling models were developed to produce fast schedules. In addition, a prototype graphical user interface (GUI) was developed to help maintenance schedulers to interpret the prognostics and the schedule solutions. Results show that the ground time for maintenance can be reduced by 20%.
Objective 6 − develop a quantitative safety risk assessment methodology for CBM.
An agent-based model of the maintenance process was produced. Using stochastic simulations, the produced model allows the simulation of maintenance strategies and computes safety indicators, such as the probability of undesired events. Furthermore, a case study considering the break wear was performed.
Objective 1 – develop an integrated approach for CBM
The integration work has been developed in the last period. Eight PHM solutions and a couple of maintenance planning models were integrated into the IT platform. This integration was tested and validated in the successful 6-month Demonstration Exercise.
Objective 2 − explore and optimize the use of different sensing technologies for structural health management (SHM).
The work performed involved the design, the specifications definition, the manufacture, and preliminary tests of four sensing technologies to be used for SHM. In particular, the equipment of 40 L1 (level 1) coupons and 8 L2 (level 2) coupons with piezoelectric elements, optical fibre sensors, and acoustic emissions sensors, together with the associated software for laboratory tests.
Objective 3 − develop data-driven probabilistic algorithms for aerostructures damage monitoring (diagnosis) and remaining useful life (RUL) estimation (prognosis)
An unprecedented lab test campaign was performed during the project. Several diagnostic methodologies targeting the various levels of SHM (i.e. detect, locate, quantify damage) were developed and successfully tested in lab-scale scenarios (details in several ReMAP publications). Anomalies are detected with success rates from 95.2% to 100%, based on methodologies leveraging strain readings with optical fibre or distributed sensing, acoustic emissions, or lamb waves. Results from Lamb wave testing show up to 10% of the Mean Absolute Percentage Error in the localization and up to 15% in the sizing of impact damage or disbond in the skin/stringer interface.
Objective 4 − develop a hybrid approach, combining machine-learning-based data analytics algorithms and physics-based models for diagnostics, prognostics and health management (PHM) of dissimilar aircraft systems.
The consortium developed data-driven PHM algorithms that have been trained and tested in 10 aircraft systems from four aircraft types. Together, one conceptual algorithm using the Federated Learning concept takes advantage of the platform architecture. An initial physics-based model concept was developed to improve route cause analysis of failure prediction from data-driven methods. Develop user interfaces to interpret the results obtained by some of the developed algorithms.
Objective 5 − develop an efficient maintenance packaging and schedule optimisation algorithm for real-time adaptive fleet maintenance management.
The consortium has developed a set of machine-learning algorithms, and scheduling models were developed to produce fast schedules. In addition, a prototype graphical user interface (GUI) was developed to help maintenance schedulers to interpret the prognostics and the schedule solutions. Results show that the ground time for maintenance can be reduced by 20%.
Objective 6 − develop a quantitative safety risk assessment methodology for CBM.
An agent-based model of the maintenance process was produced. Using stochastic simulations, the produced model allows the simulation of maintenance strategies and computes safety indicators, such as the probability of undesired events. Furthermore, a case study considering the break wear was performed.
ReMAP extended the current state of the art and practice in many ways. The consortium developed innovative machine-learning models and an IT platform. The IT platform architecture developed is unique, allowing airlines and other partners to share models without the data having to leave the servers of the airlines, guaranteeing cooperation between stakeholders and the confidentiality of the data. Despite the low TRL expectation from the call (TRL 4-5), the platform and models were deployed and tested at KLM, using live operational data from 50 of their aircraft. In the first couple of months of this demonstration, the models allowed KLM to detect and repair air-conditioning pump failures early in two of their aircraft. We have also developed similar innovative models for aircraft structures following unprecedented laboratory tests. The test data from these tests were made publicly available for future research. For the first time, comprehensive maintenance planning models were developed to optimise the maintenance plans for a fleet of aircraft while considering health prognostics from multiple aircraft systems.
ReMAP’s approach and all these models were tested in an unprecedented 6-month operational demonstration involving several systems from two different aircraft fleets to contribute to this goal. Furthermore, the involvement of European aviation stakeholders, particularly primes and suppliers, in the definition and support of a strategic plan towards exploiting CBM in practice and the IFHM solution in particular.
The impacts on society are clear – fewer maintenance results in less waste and costs, components are more efficient when well-maintenance, flight delays and cancellations can be avoided with fewer technical disruptions, and future aircraft can be lighter if prepared for a continuous health monitoring policy.
Together with some partners, I am proposing the creation of a CBM Academy to continue this effort. With this academy, I want to promote the link between education, research, and practice, supporting the development of the technicians that in the future will deploy CBM in practice.
ReMAP’s approach and all these models were tested in an unprecedented 6-month operational demonstration involving several systems from two different aircraft fleets to contribute to this goal. Furthermore, the involvement of European aviation stakeholders, particularly primes and suppliers, in the definition and support of a strategic plan towards exploiting CBM in practice and the IFHM solution in particular.
The impacts on society are clear – fewer maintenance results in less waste and costs, components are more efficient when well-maintenance, flight delays and cancellations can be avoided with fewer technical disruptions, and future aircraft can be lighter if prepared for a continuous health monitoring policy.
Together with some partners, I am proposing the creation of a CBM Academy to continue this effort. With this academy, I want to promote the link between education, research, and practice, supporting the development of the technicians that in the future will deploy CBM in practice.