Abhinav Saxena

Evolving an artificial neural network classifier for condition monitoring of rotating mechanical systems

We present the results of our investigation into the use of genetic algorithms (GAs) for identifying near optimal design parameters of diagnostic systems that are based on artificial neural networks (ANNs) for condition monitoring of mechanical systems. ANNs have been widely used for health diagnosis of mechanical bearing using features extracted from vibration and acoustic emission signals. However, different sensors and the corresponding features exhibit varied response to different faults. Moreover, a number of different features can be used as inputs to a classifier ANN. Identification of the most useful features is important for an efficient classification as opposed to using all features from all channels, leading to very high computational cost and is, consequently, not desirable. Furthermore, determining the ANN structure is a fundamental design issue and can be critical for the classification performance. We show that a GA can be used to select a smaller subset of features that together form a genetically fit family for successful fault identification and classification tasks. At the same time, an appropriate structure of the ANN, in terms of the number of nodes in the hidden layer, can be determined, resulting in improved performance.

Integrated diagnosis and prognosis architecture for fleet vehicles using dynamic case-based reasoning

This paper presents a hybrid reasoning architecture for integrated fault diagnosis and health maintenance of fleet vehicles. The aim of this architecture is to research, develop and test advanced diagnostic and decision support tools for maintenance of complex machinery. Artificial intelligence based diagnostic approach has been proposed with particular reference to dynamic case-based reasoning (DCBR). This system refines an asynchronous stream of symptom and repair actions into a compound case structure and efficiently organizes the relevant information into the case memory. Diagnosis is carried out into two steps for fast and efficient solution generation. First the situation is analyzed based on observed symptoms (textual descriptions) to propose initial diagnosis and generate corresponding explanation hypothesis. Next, based on the generated hypothesis relevant sensor data is collected and corresponding data analysis modules are activated for data-driven diagnosis. This approach reduces the computational demands to enable fast experience transfer and more reliable and informed testing. This system also tracks the success rates of all possible hypotheses for a given diagnosis and ranks them based on statistical evaluation criteria to improve the efficiency of future situations. Since the system can interact with multiple vehicles it learns about several operating environments resulting in a rich accumulation of experiences in relatively very short time. A distributed and generic architecture of this system is outlined from technical implementation point of view which can be used for widespread applications where both qualitative and quantitative observations can be gathered. Further, a concept of expanding this architecture for carrying out prognostic tasks is introduced.

Distributed prognostic health management with gaussian process regression

Distributed prognostics architecture design is an enabling step for efficient implementation of health management systems. 12A major challenge encountered in such design is formulation of optimal distributed prognostics algorithms. In this paper, we present a distributed GPR based prognostics algorithm whose target platform is a wireless sensor network. In addition to challenges encountered in a distributed implementation, a wireless network poses constraints on communication patterns, thereby making the problem more challenging. The prognostics application that was used to demonstrate our new algorithms is battery prognostics. In order to present trade-offs within different prognostic approaches, we present comparison with the distributed implementation of a particle filter based prognostics for the same battery data.

A novel blind deconvolution de-noising scheme in failure prognosis

With increased system complexity, condition-based maintenance (CBM) becomes a promising solution for system safety by detecting faults and scheduling maintenance procedures before faults become severe failures resulting in catastrophic events. For CBM of many mechanical systems, fault diagnosis and failure prognosis based on vibration signal analysis are essential techniques. Noise originating from various sources, however, often corrupts vibration signals and degrades the performance of diagnostic and prognostic routines, and consequently, the performance of CBM. In this paper, a new de-noising structure is proposed and applied to vibration signals collected from a testbed of the main gearbox of a helicopter subjected to a seeded fault. The proposed structure integrates a blind deconvolution algorithm, feature extraction, failure prognosis and vibration modelling into a synergistic system, in which the blind deconvolution algorithm attempts to arrive at the true vibration signal through an iterative optimization process. Performance indexes associated with quality of the extracted features and failure prognosis are addressed, before and after de-noising, for validation purposes.

Prognostics-enhanced Automated Contingency Management for advanced autonomous systems

This paper introduces a novel prognostics-enhanced automated contingency management (or ACM+P) paradigm based on both current health state (diagnosis) and future health state estimates (prognosis) for advanced autonomous systems. Including prognostics in ACM system allows not only fault accommodation, but also fault mitigation via proper control actions based on short term prognosis, and moreover, the establishment of a long term operational plan that optimizes the utility of the entire system based on long term prognostics. Technical challenges are identified and addressed by a hierarchical ACM+P architecture that allows fault accommodation and mitigation at various levels in the system ranging from component level control reconfiguration, system level control reconfiguration, to high level mission re-planning and resource redistribution. The ACM+P paradigm was developed and evaluated in a high fidelity unmanned aerial vehicle (UAV) simulation environment with flight-proven baseline flight controller and simulated diagnostics and prognostics of flight control actuators. Simulation results are presented. The ACM+P concept, architecture and the generic methodologies presented in this paper are applicable to many advanced autonomous systems such as deep space probes, unmanned autonomous vehicles, and military and commercial aircrafts.

Rolling element bearing feature extraction and anomaly detection based on vibration monitoring

In this paper, an anomaly detection structure, in which different types of anomaly detection routines can be applied, is proposed. Bearing fault modes and their effects on the bearing vibration are discussed. Based on this, a feature extraction method is developed to overcome the limitation of time domain features. Experimental data from bearings under different operating conditions are used to verify the proposed method. The results show that the extracted feature has a monotonic decrease trend as the dimension of fault increases. The feature also has the ability to compensate the variation of rotating speed. The proposed structure are verified with three different detection routines, pdf-based, k-nearest neighbor, and particle-filter-based approaches.

A methodology for analyzing vibration data from planetary gear systems using complex Morlet wavelets

Planetary gear trains are complex flight critical components of helicopters and other aircraft. Failure modes on such components may lead to loss of life and/or aircraft. It is essential, therefore, that incipient failures or faults be detected and isolated as early as possible and corrective action be taken in order to avoid catastrophic events. Research thus far has focused on gear teeth faults and available methods could not detect a crack in the planetary gear plate under all operating conditions. A wavelet domain methodology is suggested for the analysis and feature extraction of the vibration data from the planetary gear system of military helicopters. Complex Morlet wavelets are employed and the time domain knowledge, preserved by the wavelet decomposition, is used to extract useful features that distinguish between faulted and healthy gear plates from experimental data made available from both on-aircraft and test cell experiments. A statistical method based on the z-test is also suggested to evaluate the relative performance of these features.

A hybrid reasoning architecture for fleet vehicle maintenance

This article has described a novel approach for integrated diagnosis/prognosis of systems. The suggested architecture enables encoding of analytical techniques from a systems point of view and its expansion for prognosis tasks under the same structure. The performance of such a knowledge-based system depends on the degree of completeness of its enables encoding of analytical techniques from a systems point of view and its expansion for prognosis tasks under the same structure. The performance of such a knowledge-based system depends on the degree of completeness of its knowledge base. Since the system can interact with multiple vehicles, it learns about several operating environments, resulting in a rich accumulation of experiences in relatively very short time. At the same time, it also serves multiple systems. A natural language processing technique has been developed to extract information from the textual descriptions that is less computationally expensive than the usual NLP techniques and still preserves the meaning of the text. The experimental test data are currently being gathered for the experiments from the domain of automobiles to demonstrate the capability of the system

Simulation-based Design and Validation of Automated Contingency Management for Propulsion Systems

Automated contingency management (ACM), or the ability to confidently and autonomously adapt to fault and/or contingency conditions with the goal of still achieving mission objectives, can be considered the ultimate technological goal of a health management system. To establish confidence on the ACM system, objective performance evaluations should be executed. The need for verification and validation (V&V) techniques for ACM has also been specifically identified by DOD agencies and within the NASA community recently. This paper presents a general process and related techniques for developing and validating ACM systems for advanced propulsion systems. A novel ACM modeling paradigm, optimization-based ACM strategies, V&V approaches and performance metrics are developed. While some well-established formal methods such as model checking techniques are applicable to some sub-problems, this research has been more focused on innovative informal methods that attempt to address ACM performance requirements, optimality, robustness, etc. A pressure fed, monopropellant propulsion system for a small space flight vehicle is utilized as initial proof-of-concept implementation for the proposed techniques and preliminary simulation results are presented.

Genetic Algorithms for Artificial Neural Net-based Condition Monitoring System Design for Rotating Mechanical Systems

Stable single film bipolar membranes of prolonged life and improved performance particularly for use in electrodialysis water splitting process, are prepared by introducing a more stable interface in the membrane structure. After the cationic exchange groups are preformed on an insoluble cross-linked aromatic polymeric matrix, the dissociable anionic exchange groups may be introduced more intimately chemically bonded in position by using multi-functional compounds containing mixed tertiary, secondary and primary amine groups, so that the resulting interface is comparatively more stable, and is less likely to be neutralized, therefore, attaining longer life-time and a higher level of performance.