CERES :: Browsing by Author "Skaf, Zakwan"

Browsing by Author "Skaf, Zakwan"

Now showing 1 - 20 of 27

Open Access
Accommodating repair actions into gas turbine prognostics
(PHM Society, 2013-10-08) Skaf, Zakwan; Zaidan, Martha A.; Harrison, Robert F.; Mills, Andrew R.
Elements of gas turbine degradation, such as compressor fouling, are recoverable through maintenance actions like compressor washing. These actions increase the usable engine life and optimise the performance of the gas turbine. However, these maintenance actions are performed by a separate organization to those undertaking fleet management operations, leading to significant uncertainty in the maintenance state of the asset. The uncertainty surrounding maintenance actions impacts prognostic efficacy. In this paper, we adopt Bayesian on-line change point detection to detect the compressor washing events. Then, the event detection information is used as an input to a prognostic algorithm, advising an update to the estimation of remaining useful life. To illustrate the capability of the approach, we demonstrated our on-line Bayesian change detection algorithms on synthetic and real aircraft engine service data, in order to identify the compressor washing events for a gas turbine and thus provide demonstrably improved prognosis.
Open Access
Aircraft predictive maintenance modeling using a hybrid imbalance learning approach
(SSRN, 2020-10-26) Dangut, Maren David; Skaf, Zakwan; Jennions, Ian
The continued development of the industrial internet of things (IIoT) has caused an increase in the availability of industrial datasets. The massive availability of assets operational dataset has prompted more research interest in the area of condition-based maintenance, towards the API-lead integration for assets predictive maintenance modelling. The large data generated by industrial processes inherently comes along with different analytical challenges. Data imbalance is one of such problems that exist in datasets. It affects the performance of machine learning algorithms, which yields imprecise prediction. In this paper, we propose an advanced approach to handling imbalance classification problems in equipment heterogeneous datasets. The technique is based on a hybrid of soft mixed Gaussian processes with the EM method to improves the prediction of the minority class during learning. The algorithm is then used to develop a prognostic model for predicting aircraft component replacement. We validate the feasibility and effectiveness of our approach using real-time aircraft operation and maintenance datasets. The dataset spans over seven years. Our approach shows better performance compared to other similar methods.
Open Access
Anomaly detection of aircraft engine in FDR (flight data recorder) data
(IET, 2018-05-21) Lee, Chang-Hun; Shin, Hyosang; Tsourdos, Antonios; Skaf, Zakwan
This paper deals with detection of anomalous behaviour of aircraft engines in FDR (flight data recorder) data to improve airline maintenance operations. To this end, each FDR data that records different flight patterns is first sampled at a fixed time interval starting at the take-off phase, in order to map each FDR data into comparable data space. Next, the parameters related to the aircraft engine are only selected from the sampled FDR data. In this analysis, the feature points are chosen as the mean value of each parameter within the sampling interval. For each FDR data, the feature vector is then formed by arranging all feature points. The proposed method compares the feature vectors of all FDR data and detects an FDR data in which the abnormal behaviour of the aircraft engine is recorded. The clustering algorithm called DBSCAN (density-based spatial clustering of applications with noise) is applied for this purpose. In this paper, the proposed method is tested using realistic FDR data provided by NASA's open database. The results indicate that the proposed method can be used to automatically identify an FDR data in which the abnormal behaviour of the aircraft engine is recorded from a large amount of FDR data. Accordingly, it can be utilized for a high-level diagnosis of engine failure in airline maintenance operations.
Open Access
The application of Bayesian Change Point Detection in UAV fuel systems
(Elsevier, 2014-10-31) Niculita, Octavian; Skaf, Zakwan; Jennions, Ian K.
A significant amount of research has been undertaken in statistics to develop and implement various change point detection techniques for different industrial applications. One of the successful change point detection techniques is Bayesian approach because of its strength to cope with uncertainties in the recorded data. The Bayesian Change Point (BCP) detection technique has the ability to overcome the uncertainty in estimating the number and location of change point due to its probabilistic theory. In this paper we implement the BCP detection technique to a laboratory based fuel rig system to detect the change in the pre-valve pressure signal due to a failure in the valve. The laboratory test-bed represents a Unmanned Aerial Vehicle (UAV) fuel system and its associated electrical power supply, control system and sensing capabilities. It is specifically designed in order to replicate a number of component degradation faults with high accuracy and repeatability so that it can produce benchmark datasets to demonstrate and assess the efficiency of the BCP algorithm. Simulation shows satisfactory results of implementing the proposed BCP approach. However, the computational complexity, and the high sensitivity due to the prior distribution on the number and location of the change points are the main disadvantages of the BCP approach.
Open Access
Application of deep reinforcement learning for extremely rare failure prediction in aircraft maintenance
(Elsevier, 2022-02-08) Dangut, Maren David; Jennions, Ian K.; King, Steve; Skaf, Zakwan
The use of aircraft operational logs to predict potential failure that may lead to disruption poses many challenges and has yet to be fully explored. Given that aircraft are high-integrity assets, failures are extremely rare, and hence the distribution of relevant log data containing prior indicators will be highly skewed to the normal (healthy) case. This will present a significant challenge in using data-driven techniques because the model will be biased to the heavily weighted no-fault outcomes. This paper presents a novel approach for predicting unscheduled aircraft maintenance action based on deep reinforcement learning techniques using aircraft central maintenance system logs. The algorithm transforms the rare failure prediction problem into a sequential decision-making process that is optimised using a reward system that penalises proposed predictions that result in a false diagnosis and preferentially favours predictions that result in the right diagnosis. The validation data is directly associated with the physical health aspects of the aircraft components. The influence of extremely rare failure prediction on the proposed method is analysed. The effectiveness of the new approach is verified by comparison with previous studies, cost-sensitive and oversampling methods. Performance was evaluated based on G-mean and false-positives rates. The proposed approach shows the superior performance of 20.3% improvement in G-mean and 97% reduction in false-positive rate.
Open Access
The application of reasoning to aerospace Integrated Vehicle Health Management (IVHM): Challenges and opportunities
(Elsevier, 2019-01-11) Ezhilarasu, Cordelia Mattuvarkuzhali; Skaf, Zakwan; Jennions, Ian K.
This paper aims to discuss the importance and the necessity of reasoning applications in the field of Aerospace Integrated Vehicle Health Management (IVHM). A fully functional IVHM system is required to optimize Condition Based Maintenance (CBM), avoid unplanned maintenance activities and reduce the costs inflicted thereupon. This IVHM system should be able to utilize the information from multiple subsystems of the vehicle to assess the health of those subsystems, their effect on the other subsystems, and on the vehicle as a whole. Such a system can only be realized when the supporting technologies like sensor technology, control and systems engineering, communications technology and Artificial Intelligence (AI) are equally advanced. This paper focuses on the field of AI, especially reasoning technology and explores how it has helped the growth of IVHM in the past. The paper reviews various reasoning strategies, different reasoning systems, their architectures, components and finally their numerous applications. The paper discusses the shortcomings found in the IVHM field, particularly in the area of vehicle level health monitoring and how reasoning can be applied to address some of them. It also highlights the challenges faced when the reasoning system is developed to monitor the health at the vehicle level and how a few of these challenges can be mitigated.
Open Access
A clustering approach to detect faults with multi-component degradations in aircraft fuel systems
(Elsevier, 2020-12-18) Zaporowska, Anna; Liu, Haochen; Skaf, Zakwan; Zhao, Yifan
Accurate fault diagnosis and prognosis can significantly increase the safety and reliability of engineering systems and also reduce the maintenance costs. There is very limited relative research reported on the fault diagnosis of a complex system with multi-component degradation. The Complex Systems (CS) problem, which features multiple components simultaneously and nonlinearly interacting with each other and corresponding environment on multiple levels, has become an essential challenge in system engineering. In CS, even a single component degradation could cause misidentification of the fault severity level and lead to serious consequences. This paper introduces a new test rig to simulate multi-component degradations of the aircraft fuel system. A data analysis approach based on machine learning classification of both the time and frequency domain features is then proposed to detect and identify the fault severity level of CS with multi-component degradation. Results show that a) the fault can be sensitively detected with an accuracy > 99%; b) the severity of fault can be identified with an accuracy of 100%.
Open Access
Comparison of different classification algorithms for fault detection and fault isolation in complex systems
(Elsevier, 2018-02-08) Jung, Marcel; Niculita, Octavian; Skaf, Zakwan
Due to the lack of sufficient results seen in literature, feature extraction and classification methods of hydraulic systems appears to be somewhat challenging. This paper compares the performance of three classifiers (namely linear support vector machine (SVM), distance-weighted k-nearest neighbor (WKNN), and decision tree (DT) using data from optimized and non-optimized sensor set solutions. The algorithms are trained with known data and then tested with unknown data for different scenarios characterizing faults with different degrees of severity. This investigation is based solely on a data-driven approach and relies on data sets that are taken from experiments on the fuel system. The system that is used throughout this study is a typical fuel delivery system consisting of standard components such as a filter, pump, valve, nozzle, pipes, and two tanks. Running representative tests on a fuel system are problematic because of the time, cost, and reproduction constraints involved in capturing any significant degradation. Simulating significant degradation requires running over a considerable period; this cannot be reproduced quickly and is costly.
Open Access
Data analytics development of FDR (Flight Data Recorder) data for airline maintenance operations
(IEEE, 2017-12-11) Lee, Chang-Hun; Shin, Hyosang; Tsourdos, Antonios; Skaf, Zakwan
In this article, we propose a data analytics development to detect unusual patterns of flights from a vast amounts of FDR (flight data recorder) data for supporting airline maintenance operations. A fundamental rationale behind this development is that if there are potential issues on mechanical parts of an aircraft during a flight, evidences for these issues are most likely included in the FDR data. Therefore, the data analysis of FDR data enables us to detect the potential issues in the aircraft before they occur. To this end, in a data pre-processing step, a data filtering, a data sampling, and a data transformation are sequentially performed. And then, in this analysis, all time series data in the FDR are classified into three types: a continuous signal, a discrete signal, and a warning signal. For each type of signal, a high-dimensional vector by arranging the time series data is chosen as features. In the feature section process, a correlation analysis, a correlation relaxation, and a dimension reduction are sequentially conducted. Finally, a type of k-nearest neighbor approach is applied to automatically identify the FDR data in which the unusual flight patterns are recorded from a large amount of FDR data. The proposed method is tested with using a realistic FDR data from the NASA's open database.
Open Access
Developing a data and knowledge management approach for Integrated vehicle health management.
(Cranfield University, 2021-10) Alexslis Nyuyfoghan Maindze, Xxx Alexslis; Jennions, Ian K.; Skaf, Zakwan
In Integrated Vehicle Management (IVHM), research and engineering activities are conducted that generate large amounts of data and content. These activities include simulations, observations, derivation, experiments and referencing. However, IVHM still faces a range of data- and Knowledge Management (KM) challenges ranging from data accuracy to long-term availability for prognostic and diagnostic health management. IHVM is data-centric and therefore requires a robust data life cycle management to supports its data- and Knowledge Management activities. An understanding of the concept of KM is fundamental to addressing the IVHM data and knowledge management issues. In this regard, this thesis contextualises ‘Knowledge Management’ for IVHM by attempting to resolve the intellectual paradox that has characterised it over the years. It discusses the origins of Knowledge Management as a discipline and addresses its historical inconsistencies. This review of KM and its origins serves as a scoping study guiding a systematic review of data life cycle models. It reviews relevant standards and their role in the data life cycle. Guided by the V-Model, a Data Life Cycle Model is developed as a result and validated using a multi-technique approach combining peer review and expert insights obtained through a purposive survey. The model is then applied to IVHM centre Knowledge Management System development (KMS). The outcome includes an improved requirements gathering process and a solid foundation for resolving IVHM data and Knowledge Management challenges.
Open Access
A framework for aerospace vehicle reasoning (FAVER)
(Cranfield University, 2019-10) Ezhilarasu, Cordelia Mattuvarkuzhali; Jennions, Ian K.; Skaf, Zakwan
Airliners spend over 9% of their total revenue in Maintenance, Repair, and Overhaul (MRO) and working to bring down the cost and time involved. The prime focus is on unexpected downtime and extended maintenance leading to delays in the flights, which also reduces the trustworthiness of the airliners among the customers. One of the effective solutions to address this issue is Condition based Maintenance (CBM), in which the aircraft systems are monitored frequently, and maintenance plans are customized to suit the health of these systems. Integrated Vehicle Health Management (IVHM) is a capability enabling CBM by assessing the current condition of the aircraft at component/ Line Replaceable Unit/ system levels and providing diagnosis and remaining useful life calculations required for CBM. However, there is a lack of focus on vehicle level health monitoring in IVHM, which is vital to identify fault propagation between the systems, owing to their part in the complicated troubleshooting process resulting in prolonged maintenance. This research addresses this issue by proposing a Framework for Aerospace Vehicle Reasoning, shortly called FAVER. FAVER is developed to enable isolation and root cause identification of faults propagating between multiple systems at the aircraft level. This is done by involving Digital Twins (DTs) of aircraft systems in order to emulate interactions between these systems and Reasoning to assess health information to isolate cascading faults. FAVER currently uses four aircraft systems: i) the Electrical Power System, ii) the Fuel System, iii) the Engine, and iv) the Environmental Control System, to demonstrate its ability to provide high level reasoning, which can be used for troubleshooting in practice. FAVER is also demonstrated for its ability to expand, update, and scale for accommodating new aircraft systems into the framework along with its flexibility. FAVER’s reasoning ability is also evaluated by testing various use cases.
Open Access
A generalised methodology for the diagnosis of aircraft systems
(IEEE, 2021-01-11) Ezhilarasu, Cordelia Mattuvarkuzhali; Skaf, Zakwan; Jennions, Ian K.
An aircraft is made up of a number of complicated systems which work in harmony to ensure safe and trouble-free flight. In order to maintain such a platform, many diagnostic and prognostic techniques have been suggested, mostly aimed at components but some at the system level. Together these form a patchwork approach to the overall problem of efficiently informing aircraft maintenance to the Original Equipment Manufacturers, the operators /airlines, and the Maintenance, Repair, and Overhaul organisations. It involves these organisations having to support several different approaches to aircraft health management, and is therefore inefficient and costly. In the current work, a streamlined methodology is put forward. This is based on OSA-CBM (Open System Architecture for Condition Based Maintenance) and can be applied to any aircraft system. Integral with this is the use of mRMR (minimum redundancy maximum relevance) for feature selection, the resulting symptom vector being used for fault diagnosis. This approach is demonstrated on three test cases: the engine, the environmental control system, and the fuel system. In each case, the digital twin setup, simulation conditions for healthy and faulty scenarios, a methodology based on OSA-CBM up to diagnostics are detailed. Diagnostics is carried out for each system in turn, using four machine learning supervised algorithms. The best performing algorithm for each system will then subsequently be used in a vehicle level reasoner called FAVER (A Framework for Aerospace Vehicle Reasoning), which requires these system diagnoses as a starting point for vehicle reasoning and fault ambiguity resolution
Open Access
Handling imbalanced data for aircraft predictive maintenance using the BACHE algorithm
(Elsevier, 2022-05-14) Dangut, Maren David; Skaf, Zakwan; Jennions, Ian K.
Developing a prognostic model to predict an asset’s health condition is a maintenance strategy that increases asset availability and reliability through better maintenance scheduling. Therefore, developing reliable vehicle health predictive models is vital in the aerospace industry, especially considering a safety–critical system such as aircraft. However, one of the significant challenges faced in building reliable data-driven prognostic models is the imbalance dataset. Training machine-learning models using an imbalanced dataset causes classifiers to be biased towards the class with majority samples, resulting in poor predictive accuracy in data-driven models. This problem can become more challenging if the imbalance ratio is extreme and classes overlap. In this paper, a novel approach called Balanced Calibrated Hybrid Ensemble Technique (BACHE) is developed to tackle the severe imbalanced classification problem. The proposed method involves the combination of hybrid data sampling and ensemble-based learning. It uses a cascading balanced approach to transfer a class imbalance problem into a sub-problem by decomposing the original problem into a set of subproblems, each characterized by a reduced imbalance ratio. Then uses a calibrated boosting with a cost-sensitive decision tree to enhance recognition of hard-to-learn patterns, which improves the prediction of the extreme minority class. BACHE is evaluated using a real-world aircraft dataset with rare component replacement instances. Also, a comparative experiment of the proposed approach with other similar existing methods is conducted. The performance metrics used are precision, recall, G-mean, and an area under the curve. The final results show that the proposed model outperforms other similar methods. Also, it can attain an excellent performance on large, extremely imbalanced datasets.
Open Access
An integrated machine learning model for aircraft components rare failure prognostics with log-based dataset
(Elsevier, 2020-05-11) Dangut, Maren David; Skaf, Zakwan; Jennions, Ian K.
Predictive maintenance is increasingly advancing into the aerospace industry, and it comes with diverse prognostic health management solutions. This type of maintenance can unlock several benefits for aerospace organizations. Such as preventing unexpected equipment downtime and improving service quality. In developing data-driven predictive modelling, one of the challenges that cause model performance degradation is the data-imbalanced distribution. The extreme data imbalanced problem arises when the distribution of the classes present in the datasets is not uniform. Such that the total number of instances in a class far outnumber those of the other classes. Extremely skew data distribution can lead to irregular patterns and trends, which affects the learning of temporal features. This paper proposes a hybrid machine learning approach that blends natural language processing techniques and ensemble learning for predicting extremely rare aircraft component failure. The proposed approach is tested using a real aircraft central maintenance system log-based dataset. The dataset is characterized by extremely rare occurrences of known unscheduled component replacements. The results suggest that the proposed approach outperformed the existing imbalanced and ensemble learning methods in terms of precision, recall, and f1-score. The proposed approach is approximately 10% better than the synthetic minority oversampling technique. It was also found that by searching for patterns in the minority class exclusively, the class imbalance problem could be overcome. Hence, the model classification performance is improved
Open Access
Knowledge management yesterday and tomorrow: exploring an ‘Intellectual Paradox’
(IOS Press, 2017-09-07) Maindze, Alexslis; Jennions, Ian K.; Skaf, Zakwan
Knowledge management continues to be characterized by strong contextual application with diversity of techniques, tools and applications which practitioners far and wide seem to agree and adopt. However, when it comes to its philosophical distinctness, it is yet to achieve something as seemingly easy as a common definition. There is significant agreement on fluidity and methods of application but limited consensus on philosophical interpretation. Furthermore, that we know what it is, acknowledge its impact, functional relevance and yet cannot articulate a common methodology points to what this paper terms an ‘intellectual paradox’. An intellectual paradox is the phenomenon whereby professionals and academics acknowledge a concept, practice it, write about it, and promote its relevance individually but as a collective lack a consensus on exactly what it is. This paper seeks to explore this phenomenon in detail and to propose a philosophical framework. It further explores the role of the traditional composition; people, process and technology in sustaining this suggested conundrum. This phenomenon seems to tie neatly with the tacit form of knowledge on the basis of the difficulty in articulating a common definitional framework of perception, though it could be argued that it is merely exhibiting characteristics of ‘Tacit’ knowledge management; thereby justifying the status quo. Some authors point to “descriptive frameworks” and insufficient addressing of learning including structural differences in organisations. This difficulty per some writers, results from the use of multiple and variable methods, tools techniques and strategies. Their alternative proposition views for a both ‘descriptive and prescriptive’ framework still did not yield a consensus either. This paper seeks to explore the problem and to propose a new definition.
Open Access
A machine learning-based clustering approach to diagnose multi-component degradation of aircraft fuel systems
(Springer, 2021-10-07) Liu, Haochen; Zhao, Yifan; Zaporowska, Anna; Skaf, Zakwan
Accurate fault diagnosis and prognosis can significantly reduce maintenance costs, increase the safety and availability of engineering systems that have become increasingly complex. It has been observed that very limited researches have been reported on fault diagnosis where multi-component degradation are presented. This is essentially a challenging Complex Systems problem where features multiple components interacting simultaneously and nonlinearly with each other and its environment on multiple levels. Even the degradation of a single component can lead to a misidentification of the fault severity level. This paper introduces a new test rig to simulate the multi-component degradation of the aircraft fuel system. A machine learning-based data analytical approach based on the classification of clustering features from both time and frequency domains is proposed. The scope of this framework is the identification of the location and severity of not only the system fault but also the multi-component degradation. The results illustrate that (a) the fault can be detected with accuracy > 99%; (b) the severity of fault can be identified with an accuracy of almost 100%; (c) the degradation level can be successfully identified with the R-square value > 0.9.
Open Access
New application of data analysis using aircraft fault record data
(AIAA, 2018-04-06) Lee, Chang-Hun; Shin, Hyosang; Tsourdos, Antonios; Skaf, Zakwan
Open Access
Prognostics: Design, Implementation, and Challenges
(Curran Associates, 2015-09-30) Skaf, Zakwan
Prognostics is an essential part of condition-based maintenance (CBM), described as predicting the remaining useful life (RUL) of a system. It is also a key technology for an integrated vehicle health management (IVHM) system that leads to improved safety and reliability. A vast amount of research has been presented in the literature to develop prognostics models that are able to predict a system’s RUL. These models can be broadly categorised into experience-based models, data-driven models and physics-based models. Therefore, careful consideration needs to be given to selecting which prognostics model to take forward and apply for each real application. Currently, developing reliable prognostics models in real life is challenging for various reasons, such as the design complexity associated with a system, the high uncertainty and its propagation in the degradation, system level prognostics, the evaluation framework and a lack of prognostics standards. This paper is written with the aim to bring forth the challenges and opportunities for developing prognostics models for complex systems and making researchers aware of these challenges and opportunities.
Open Access
Progress towards a Framework for Aerospace Vehicle Reasoning (FAVER)
(PHM Society, 2019-09-23) Ezhilarasu, Cordelia Mattuvarkuzhali; Skaf, Zakwan; Jennions, Ian
This paper proposes a reasoning framework to diagnose faults at the vehicle level in a complex machine like an aircraft. The current focus of Integrated Vehicle Health Management (IVHM) is on diagnosing and prognosing faults at the component and subsystem levels; only a few IVHM systems consider the interaction between the systems. To diagnose faults at the vehicle level, an IVHM System needs a framework that recognizes the causal relationships between systems and the likelihood of fault propagation between them. The framework should also possess an element of reasoning to assess data from all systems, to assign priorities, and to resolve ambiguities. The Framework for Aerospace VEhicle Reasoning (FAVER) that is proposed in this paper uses a digital twin of the aircraft systems to emulate functioning of the aircraft and to simulate the effect of fault propagation due to systems interactions. FAVER applies reasoning that can handle fault signatures from multiple systems in the form of symptom vectors, to detect and isolate cascading faults and their root causes. The blending of a digital twin and reasoning in this framework will enable FAVER to: i) isolate faults that have both local and cascading effects on the concerned systems, ii) identify faults that were previously unknown, and iii) resolve ambiguous faults. This paper explains the different steps involved in developing FAVER and how this framework can be demonstrated in the aforementioned scenarios with the help of different use cases. This paper also talks about the challenges to be faced while developing this framework and ways to overcome them.
Open Access
A rare failure detection model for aircraft predictive maintenance using a deep hybrid learning approach
(Springer, 2022-03-26) Dangut, Maren David; Jennions, Ian K.; King, Steve; Skaf, Zakwan
The use of aircraft operation logs to develop a data-driven model to predict probable failures that could cause interruption poses many challenges and has yet to be fully explored. Given that aircraft is high-integrity assets, failures are exceedingly rare. Hence, the distribution of relevant log data containing prior signs will be heavily skewed towards the typical (healthy) scenario. Thus, this study presents a novel deep learning technique based on the auto-encoder and bidirectional gated recurrent unit networks to handle extremely rare failure predictions in aircraft predictive maintenance modelling. The auto-encoder is modified and trained to detect rare failures, and the result from the auto-encoder is fed into the convolutional bidirectional gated recurrent unit network to predict the next occurrence of failure. The proposed network architecture with the rescaled focal loss addresses the imbalance problem during model training. The effectiveness of the proposed method is evaluated using real-world test cases of log-based warning and failure messages obtained from the fleet database of aircraft central maintenance system records. The proposed model is compared to other similar deep learning approaches. The results indicated an 18% increase in precision, a 5% increase in recall, and a 10% increase in G-mean values. It also demonstrates reliability in anticipating rare failures within a predetermined, meaningful time frame.