Ahlam Mallak
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ahlam Mallak has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
Madjid Fathi
University of Siegen, Germany
5 shared publications- Christian Weber
University of Siegen
4 shared publications - Madjid Fathi
University of Siegen
4 shared publications Roman Obermaisser
Embedded Systems, University of Siegen, Siegen, Germany
3 shared publicationsAli Behravan
University of Siegen, Institute of Embedded Systems, Siegen, Germany
2 shared publicationsAkash Sonnad
University of Siegen, Germany
1 shared publication
Latest Publications
- Open Access
Sensor and Component Fault Detection and Diagnosis for Hydraulic Machinery Integrating LSTM Autoencoder Detector and Diagnostic Classifiers
Anomaly occurrences in hydraulic machinery might lead to massive system shut down, jeopardizing the safety of the machinery and its surrounding human operator(s) and environment, and the severe economic implications following the faults and their associated damage. Hydraulics are mostly placed in ruthless environments, where they are consistently vulnerable to many faults. Hence, not only are the machines and their components prone to anomalies, but also the sensors attached to them, which monitor and report their health and behavioral changes. In this work, a comprehensive applicational analysis of anomalies in hydraulic systems extracted from a hydraulic test rig was thoroughly achieved. First, we provided a combination of a new architecture of LSTM autoencoders and supervised machine and deep learning methodologies, to perform two separate stages of fault detection and diagnosis. The two phases were condensed by—the detection phase using the LSTM autoencoder. Followed by the fault diagnosis phase represented by the classification schema. The previously mentioned framework was applied to both component and sensor faults in hydraulic systems, deployed in the form of two in-depth applicational experiments. Moreover, a thorough literature review of related work from the past decade, for autoencoders related fault detection and diagnosis in hydraulic systems, was successfully conducted in this study.
Open AccessSensor and Component Fault Detection and Diagnosis for Hydraulic Machinery Integrating LSTM Autoencoder Detector and Diagnostic Classifiers
Anomaly occurrences in hydraulic machinery might lead to massive system shut down, jeopardizing the safety of the machinery and its surrounding human operator(s) and environment, and the severe economic implications following the faults and their associated damage. Hydraulics are mostly placed in ruthless environments, where they are consistently vulnerable to many faults. Hence, not only are the machines and their components prone to anomalies, but also the sensors attached to them, which monitor and report their health and behavioral changes. In this work, a comprehensive applicational analysis of anomalies in hydraulic systems extracted from a hydraulic test rig was thoroughly achieved. First, we provided a combination of a new architecture of LSTM autoencoders and supervised machine and deep learning methodologies, to perform two separate stages of fault detection and diagnosis. The two phases were condensed by—the detection phase using the LSTM autoencoder. Followed by the fault diagnosis phase represented by the classification schema. The previously mentioned framework was applied to both component and sensor faults in hydraulic systems, deployed in the form of two in-depth applicational experiments. Moreover, a thorough literature review of related work from the past decade, for autoencoders related fault detection and diagnosis in hydraulic systems, was successfully conducted in this study.
- Open Access
A Hybrid Approach: Dynamic Diagnostic Rules for Sensor Systems in Industry 4.0 Generated by Online Hyperparameter Tuned Random Forest
In this work, a hybrid component Fault Detection and Diagnosis (FDD) approach for industrial sensor systems is established and analyzed, to provide a hybrid schema that combines the advantages and eliminates the drawbacks of both model-based and data-driven methods of diagnosis. Moreover, it shines light on a new utilization of Random Forest (RF) together with model-based diagnosis, beyond its ordinary data-driven application. RF is trained and hyperparameter tuned using three-fold cross validation over a random grid of parameters using random search, to finally generate diagnostic graphs as the dynamic, data-driven part of this system. This is followed by translating those graphs into model-based rules in the form of if-else statements, SQL queries or semantic queries such as SPARQL, in order to feed the dynamic rules into a structured model essential for further diagnosis. The RF hyperparameters are consistently updated online using the newly generated sensor data to maintain the dynamicity and accuracy of the generated graphs and rules thereafter. The architecture of the proposed method is demonstrated in a comprehensive manner, and the dynamic rules extraction phase is applied using a case study on condition monitoring of a hydraulic test rig using time-series multivariate sensor readings.
Open AccessA Hybrid Approach: Dynamic Diagnostic Rules for Sensor Systems in Industry 4.0 Generated by Online Hyperparameter Tuned Random Forest
In this work, a hybrid component Fault Detection and Diagnosis (FDD) approach for industrial sensor systems is established and analyzed, to provide a hybrid schema that combines the advantages and eliminates the drawbacks of both model-based and data-driven methods of diagnosis. Moreover, it shines light on a new utilization of Random Forest (RF) together with model-based diagnosis, beyond its ordinary data-driven application. RF is trained and hyperparameter tuned using three-fold cross validation over a random grid of parameters using random search, to finally generate diagnostic graphs as the dynamic, data-driven part of this system. This is followed by translating those graphs into model-based rules in the form of if-else statements, SQL queries or semantic queries such as SPARQL, in order to feed the dynamic rules into a structured model essential for further diagnosis. The RF hyperparameters are consistently updated online using the newly generated sensor data to maintain the dynamicity and accuracy of the generated graphs and rules thereafter. The architecture of the proposed method is demonstrated in a comprehensive manner, and the dynamic rules extraction phase is applied using a case study on condition monitoring of a hydraulic test rig using time-series multivariate sensor readings.
- Open Access
- 11 August 2020
- Published by MDPI AG in Preprints.org
Unsupervised Feature Selection Using Recursive k-Means Silhouette Elimination (RkSE): A Two-Scenario Case Study for Fault Classification of High-Dimensional Sensor Data
Feature selection is a crucial step to overcome the curse of dimensionality problem in data mining. This work proposes Recursive k-means Silhouette Elimination (RkSE) as a new unsupervised feature selection algorithm to reduce dimensionality in univariate and multivariate time-series datasets. Where k-means clustering is applied recursively to select the cluster representative features, following a unique application of silhouette measure for each cluster and a user-defined threshold as the feature selection or elimination criteria. The proposed method is evaluated on a hydraulic test rig, multi sensor readings in two different fashions: (1) Reduce the dimensionality in a multivariate classification problem using various classifiers of different functionalities. (2) Classification of univariate data in a sliding window scenario, where RkSE is used as a window compression method, to reduce the window dimensionality by selecting the best time points in a sliding window. Moreover, the results are validated using 10-fold cross validation technique. As well as, compared to the results when the classification is pulled directly with no feature selection applied. Additionally, a new taxonomy for k-means based feature selection methods is proposed. The experimental results and observations in the two comprehensive experiments demonstrated in this work reveal the capabilities and accuracy of the proposed method.
- 11 August 2020
- Published by MDPI AG in Preprints.org
Open AccessUnsupervised Feature Selection Using Recursive k-Means Silhouette Elimination (RkSE): A Two-Scenario Case Study for Fault Classification of High-Dimensional Sensor Data
Feature selection is a crucial step to overcome the curse of dimensionality problem in data mining. This work proposes Recursive k-means Silhouette Elimination (RkSE) as a new unsupervised feature selection algorithm to reduce dimensionality in univariate and multivariate time-series datasets. Where k-means clustering is applied recursively to select the cluster representative features, following a unique application of silhouette measure for each cluster and a user-defined threshold as the feature selection or elimination criteria. The proposed method is evaluated on a hydraulic test rig, multi sensor readings in two different fashions: (1) Reduce the dimensionality in a multivariate classification problem using various classifiers of different functionalities. (2) Classification of univariate data in a sliding window scenario, where RkSE is used as a window compression method, to reduce the window dimensionality by selecting the best time points in a sliding window. Moreover, the results are validated using 10-fold cross validation technique. As well as, compared to the results when the classification is pulled directly with no feature selection applied. Additionally, a new taxonomy for k-means based feature selection methods is proposed. The experimental results and observations in the two comprehensive experiments demonstrated in this work reveal the capabilities and accuracy of the proposed method.
- Open Access
- 23 July 2020
- Published by MDPI AG in Preprints.org
A Hybrid Approach: Dynamic Diagnostic Rules for Sensor Systems in Industry 4.0 Generated by Online Hyperparameter Tuned Random Forest
In this work, A hybrid component Fault Detection and Diagnosis (FDD) approach for industrial sensor systems is established and analyzed, to provide a hybrid schema that combines the advantages and eliminates the drawbacks of both model-based and data-driven methods of diagnosis. Moreover, spotting the light on a new utilization of Random Forest (RF) together with model-based diagnosis, beyond its ordinary data-driven application. RF is trained and hyperparameter tuned using 3-fold cross-validation over a random grid of parameters using random search, to finally generate diagnostic graphs as the dynamic, data-driven part of this system. Followed by translating those graphs into model-based rules in the form of if-else statements, SQL queries or semantic queries such as SPARQL, in order to feed the dynamic rules into a structured model essential for further diagnosis. The RF hyperparameters are consistently updated online using the newly generated sensor data, in order to maintain the dynamicity and accuracy of the generated graphs and rules thereafter. The architecture of the proposed method is demonstrated in a comprehensive manner, as well as the dynamic rules extraction phase is applied using a case study on condition monitoring of a hydraulic test rig using time series multivariate sensor readings.
- 23 July 2020
- Published by MDPI AG in Preprints.org
Open AccessA Hybrid Approach: Dynamic Diagnostic Rules for Sensor Systems in Industry 4.0 Generated by Online Hyperparameter Tuned Random Forest
In this work, A hybrid component Fault Detection and Diagnosis (FDD) approach for industrial sensor systems is established and analyzed, to provide a hybrid schema that combines the advantages and eliminates the drawbacks of both model-based and data-driven methods of diagnosis. Moreover, spotting the light on a new utilization of Random Forest (RF) together with model-based diagnosis, beyond its ordinary data-driven application. RF is trained and hyperparameter tuned using 3-fold cross-validation over a random grid of parameters using random search, to finally generate diagnostic graphs as the dynamic, data-driven part of this system. Followed by translating those graphs into model-based rules in the form of if-else statements, SQL queries or semantic queries such as SPARQL, in order to feed the dynamic rules into a structured model essential for further diagnosis. The RF hyperparameters are consistently updated online using the newly generated sensor data, in order to maintain the dynamicity and accuracy of the generated graphs and rules thereafter. The architecture of the proposed method is demonstrated in a comprehensive manner, as well as the dynamic rules extraction phase is applied using a case study on condition monitoring of a hydraulic test rig using time series multivariate sensor readings.
- 1 December 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1190-1195
SenGen: A Two-Phase Dynamic Simulation and Toolbox of an Indoor Mobile Wireless Sensor Network for Sensor Monitoring and Dataset Generation
A Mobile Wireless Sensor Network (MWSN) is a network of mobile sensor nodes that are spatially separated in an open or closed space, which work altogether to sense various system environmental and physical parameters. The state-of-the art is full of approaches for modelling WSNs and MWSNs using different simulation tools and programming languages. Such models require the system expert interference to change the simulated model itself whenever any change is required. Without this interference or having knowledge of the simulated system, these models tend to generate fixed-case sensor data and lack the dynamicity and the ability for further user-specific changes at run-time. In this paper, a two-phase dynamic simulation toolbox -so-called 'SenGen'- is presented and tested, where a full simulation of an indoor MWSN system is established using Simulink. Then a Graphical User Interface (GUI) is created with MATLAB, to overall perform as a dynamic toolbox for sensor data generation in MWSNs.
- 1 December 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1190-1195
SenGen: A Two-Phase Dynamic Simulation and Toolbox of an Indoor Mobile Wireless Sensor Network for Sensor Monitoring and Dataset Generation
A Mobile Wireless Sensor Network (MWSN) is a network of mobile sensor nodes that are spatially separated in an open or closed space, which work altogether to sense various system environmental and physical parameters. The state-of-the art is full of approaches for modelling WSNs and MWSNs using different simulation tools and programming languages. Such models require the system expert interference to change the simulated model itself whenever any change is required. Without this interference or having knowledge of the simulated system, these models tend to generate fixed-case sensor data and lack the dynamicity and the ability for further user-specific changes at run-time. In this paper, a two-phase dynamic simulation toolbox -so-called 'SenGen'- is presented and tested, where a full simulation of an indoor MWSN system is established using Simulink. Then a Graphical User Interface (GUI) is created with MATLAB, to overall perform as a dynamic toolbox for sensor data generation in MWSNs.
- 1 May 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 572-577
ADISTES Ontology for Active Diagnosis of Sensors and Actuators in Distributed Embedded Systems
Fault Detection, Isolation and Recovery (FDIR) are becoming increasingly indispensable for real-time systems. Active diagnosis systems, therefore, are endeavoring to increase the reliability of real-time systems by computing at run-time the diagnostic information within FDIR systems, to actively decide on recovery actions. For embedded systems that contain various components, especially those including sensors and actuators, it is essential to provide a platform in which these components are able to communicate and also serve for an active diagnosis system. This paper proposes the ADISTES ontology (Active Diagnosis based on Semantic Web Technologies for Distributed Embedded Real-Time Systems) for an active diagnosis system of sensors and actuators. The proposed ADISTES ontology is a general ontology that can be used by different cases and systems to integrate their sensors and actuators; it can be instantiated from heterogeneous sources. As a characteristic of ontologies, they provide information in a machine interpretable format. Hence, the ADISTES ontology can also play a role for different sensors and actuators to share and communicate their status with one another.
- 1 May 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 572-577
ADISTES Ontology for Active Diagnosis of Sensors and Actuators in Distributed Embedded Systems
Fault Detection, Isolation and Recovery (FDIR) are becoming increasingly indispensable for real-time systems. Active diagnosis systems, therefore, are endeavoring to increase the reliability of real-time systems by computing at run-time the diagnostic information within FDIR systems, to actively decide on recovery actions. For embedded systems that contain various components, especially those including sensors and actuators, it is essential to provide a platform in which these components are able to communicate and also serve for an active diagnosis system. This paper proposes the ADISTES ontology (Active Diagnosis based on Semantic Web Technologies for Distributed Embedded Real-Time Systems) for an active diagnosis system of sensors and actuators. The proposed ADISTES ontology is a general ontology that can be used by different cases and systems to integrate their sensors and actuators; it can be instantiated from heterogeneous sources. As a characteristic of ontologies, they provide information in a machine interpretable format. Hence, the ADISTES ontology can also play a role for different sensors and actuators to share and communicate their status with one another.
- 1 June 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 000377-000382
A Graph-Based Sensor Fault Detection and Diagnosis for Demand-Controlled Ventilation Systems Extracted from a Semantic Ontology
Fault detection and diagnosis in HVAC systems such as demand-controlled ventilation systems, is a crucial step to attain optimal user comfort and energy saving, through providing fast and accurate monitoring, control, and recovery solutions in buildings. This can be accomplished by increasing the number of the utilized sensors and actuators in each zone, which leads to an enormous increase in the complexity of the system, due to the need to capture these components' values and the interactions between them. To overcome the complexity issue, it is important to establish an accurate model of the system, which contains the system components like sensors and actuators, and the relationships between them. However, achieving an accurate model design of the building, its technical components and the relationships between them is rarely available. Especially, the lack of precision in technical measurements, which is needed in model-based diagnostic methods to develop precise thresholds and conditions that are required to make the decisions. In this paper a model was developed to overcome the previous challenges, by the following: 1) creating a simulated model for a building, to extract the missing sensors' thresholds, values and relationships between those sensors. 2) A semantic model represented by the building ontology, to model the relationships between sensors and their containing systems, created based on the diagnostic information provided by the simulated model. And 3) A novel diagnostic directed graph is extracted from the ontology to offer more automation to the diagnosis and lessen the complexity of the system, by providing a clear graph of the decision making process.
- 1 June 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 000377-000382
A Graph-Based Sensor Fault Detection and Diagnosis for Demand-Controlled Ventilation Systems Extracted from a Semantic Ontology
Fault detection and diagnosis in HVAC systems such as demand-controlled ventilation systems, is a crucial step to attain optimal user comfort and energy saving, through providing fast and accurate monitoring, control, and recovery solutions in buildings. This can be accomplished by increasing the number of the utilized sensors and actuators in each zone, which leads to an enormous increase in the complexity of the system, due to the need to capture these components' values and the interactions between them. To overcome the complexity issue, it is important to establish an accurate model of the system, which contains the system components like sensors and actuators, and the relationships between them. However, achieving an accurate model design of the building, its technical components and the relationships between them is rarely available. Especially, the lack of precision in technical measurements, which is needed in model-based diagnostic methods to develop precise thresholds and conditions that are required to make the decisions. In this paper a model was developed to overcome the previous challenges, by the following: 1) creating a simulated model for a building, to extract the missing sensors' thresholds, values and relationships between those sensors. 2) A semantic model represented by the building ontology, to model the relationships between sensors and their containing systems, created based on the diagnostic information provided by the simulated model. And 3) A novel diagnostic directed graph is extracted from the ontology to offer more automation to the diagnosis and lessen the complexity of the system, by providing a clear graph of the decision making process.
- 1 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 0273-0279
Fault injection framework for fault diagnosis based on machine learning in heating and demand-controlled ventilation systems
The main purpose of a heating, ventilation and air-conditioning system is to provide thermal comfort and acceptable indoor air quality through adequate natural ventilation by exchanging the building zone air with the outside fresh air, as an option. This air exchange during the cold seasons can be considered as a heating load for the heating system which increases the energy consumption of the heating system. The use of the demand-controlled ventilation system can be counted as a potential energy saving method by its automatic adjustment of ventilation, which modifies the amount of fresh air coming from outside and causes less energy consumption by less heating load, as a result. Faults in system components such as sensors and actuators can result in different types of failures and severe implications on the efficiency of the heating and demand-controlled ventilation system. However, tracing the component and system behavior back to the faults is a challenging task. This study demonstrates a successful approach for finding the nature, value, time of occurrence, and locality of these faults using a mapping from failures to faults with knowledge-based diagnostic techniques. We introduce fault models for online diagnosis using machine learning. These fault models are established for a heating and demand-controlled ventilation system with fault injection blocks design using a simulation framework based on MATLAB/Simulink.
- 1 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 0273-0279
Fault injection framework for fault diagnosis based on machine learning in heating and demand-controlled ventilation systems
The main purpose of a heating, ventilation and air-conditioning system is to provide thermal comfort and acceptable indoor air quality through adequate natural ventilation by exchanging the building zone air with the outside fresh air, as an option. This air exchange during the cold seasons can be considered as a heating load for the heating system which increases the energy consumption of the heating system. The use of the demand-controlled ventilation system can be counted as a potential energy saving method by its automatic adjustment of ventilation, which modifies the amount of fresh air coming from outside and causes less energy consumption by less heating load, as a result. Faults in system components such as sensors and actuators can result in different types of failures and severe implications on the efficiency of the heating and demand-controlled ventilation system. However, tracing the component and system behavior back to the faults is a challenging task. This study demonstrates a successful approach for finding the nature, value, time of occurrence, and locality of these faults using a mapping from failures to faults with knowledge-based diagnostic techniques. We introduce fault models for online diagnosis using machine learning. These fault models are established for a heating and demand-controlled ventilation system with fault injection blocks design using a simulation framework based on MATLAB/Simulink.
- 1 October 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Active diagnosis automotive ontology for distributed embedded systems
Active diagnosis is the retrieval and analysis of the diagnostic information at runtime. In order to enhance the system ability for detecting faults, isolating them, and deciding the most suitable recovery action during the shortest possible completion period at runtime. Thus, it is necessary to build a precise and appropriate knowledge-based model that is able to serve the diagnosis and the fixing tasks efficiently and accurately. This paper presents an approach for an active ontology-based automotive diagnostic and repair for distributed embedded systems. This approach helps in creating better understanding and realization of the knowledge-based diagnostic concepts in automotive systems, as well as the integration of these concepts with distributed embedded systems, to achieve better diagnosis and repair for actual automotive systems at runtime when deployed in the future.
- 1 October 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Active diagnosis automotive ontology for distributed embedded systems
Active diagnosis is the retrieval and analysis of the diagnostic information at runtime. In order to enhance the system ability for detecting faults, isolating them, and deciding the most suitable recovery action during the shortest possible completion period at runtime. Thus, it is necessary to build a precise and appropriate knowledge-based model that is able to serve the diagnosis and the fixing tasks efficiently and accurately. This paper presents an approach for an active ontology-based automotive diagnostic and repair for distributed embedded systems. This approach helps in creating better understanding and realization of the knowledge-based diagnostic concepts in automotive systems, as well as the integration of these concepts with distributed embedded systems, to achieve better diagnosis and repair for actual automotive systems at runtime when deployed in the future.
- 26 October 2016
- Published by Springer Nature
- p. 334-341
Computational and Comparative Study on Multiple Kernel Learning Approaches for the Classification Problem of Alzheimer’s Disease
- 26 October 2016
- Published by Springer Nature
- p. 334-341
Computational and Comparative Study on Multiple Kernel Learning Approaches for the Classification Problem of Alzheimer’s Disease