No description available.

No description available.

Bayesian networks have been applied to many different domains to perform prognostics, reduce risk and ultimately improve decision making. However, these methods have not been applied to military field and human performance data sets in an industrial environment. Methods frequently rely on a clear understanding of causal connections leading to an undesirable event and detailed understanding of the system behavior. Methods may also require large amount of analyst teams and domain experts, coupled with manual data cleansing and classification. The research performed utilized machine learning algorithms (such as Bayesian networks) and two existing data sets. The primary objective of the research was to develop a diagnostic and prognostic tool utilizing Bayesian networks that does not require the need for detailed causal understanding of the underlying system. The research yielded a predictive method with substantial benefits over reactive methods. The research indicated Bayesian networks can be trained and utilized to predict failure of several important components to include potential malfunction codes and downtime on a real-world Navy data set. The research also considered potential error within the training data set. The results provided credence to utilization of Bayesian networks in real field data – which will always contain error that is not easily quantified. Research should be replicated with additional field data sets from other aircraft. Future research should be conducted to solicit and incorporate domain expertise into subsequent models. Research should also consider incorporation of text based analytics for text fields, which was considered out of scope for this research project.

The performance of the Acoustic Emission (AE) technique was investigated to establish its reliability in detecting and locating fatigue crack damage as well as distinguishing between different AE sources in potential SHM applications. Experiments were conducted to monitor the AE signals generated during fatigue crack growth in coupon 2014 T6 aluminium. The influence of stress ratio, stress range, sample geometry and whether or not the load spectrum was of constant or variable amplitude were all investigated. Timing filters were incorporated to eliminate extraneous AE signals produced from sources other than the fatigue crack. AE signals detected were correlated with values of applied cyclic load throughout the tests. Measurements of Time difference of arrival were taken for assessment of errors in location estimates obtained using time of flight algorithms with a 1D location setup. It was found that there was significant variability in AE Hit rates in otherwise identical samples and test conditions. However common trends characteristic of all samples could be observed. At the onset of crack growth high AE Hit rates were observed for the first few millimetres after which they rapidly declined to minimal values for an extended period of crack growth. Another peak and then decline in AE Hit rates was observed for subsequent crack growth before yet another increase as the sample approached final failure. The changes in AE signals with applied cyclic load provided great insights into the different AE processes occurring during crack growth. AE signals were seen to occur in the lower two-thirds of the maximum load in the first few millimetres of crack growth before occurring at progressively smaller values as the crack length increased. These emissions could be associated with crack closure. A separate set of AE signals were observed close to the maximum cyclic stress throughout the entire crack growth process. At the failure crack length AE signals were generated across the entire loading range. Novel metrics were developed to statistically characterise variability of AE generation with crack growth and at particular crack lengths across different samples. A novel approach for fatigue crack length estimation was developed based on monitoring applied loads to the sample corresponding with generated AE signals which extends the functionality of the AE technique in an area which was previously deficient. It is however limited by its sensitivity to changes in sample geometry. Experiments were also performed to validate the performance of the AE technique in detecting and locating fatigue crack in a representative wing-box structure. An acousto-ultrasonic method was used to calibrate the AE wave velocity in the structure which was used to successfully locate the 'hidden' fatigue crack. A novel observation was made in the series of tests conducted where the complex propagation paths in the structure could be exploited to perform wide area sensing coverage in certain regions using sensors mounted on different components of the structure. This also extends current knowledge on the capability of the AE technique.

The ultimate goal of most prognostic systems is accurate prediction of the remaining useful life (RUL) of individual systems or components based on their use and performance. This class of prognostic algorithms is termed Degradation-Based, or Type III Prognostics. As equipment degrades, measured parameters of the system tend to change; these sensed measurements, or appropriate transformations thereof, may be used to characterize degradation. Traditionally, individual-based prognostic methods use a measure of degradation to make RUL estimates. Degradation measures may include sensed measurements, such as temperature or vibration level, or inferred measurements, such as model residuals or physics-based model predictions. Often, it is beneficial to combine several measures of degradation into a single parameter. Selection of an appropriate parameter is key for making useful individual-based RUL estimates, but methods to aid in this selection are absent in the literature. This dissertation introduces a set of metrics which characterize the suitability of a prognostic parameter. Parameter features such as trendability, monotonicity, and prognosability can be used to compare candidate prognostic parameters to determine which is most useful for individual-based prognosis. Trendability indicates the degree to which the parameters of a population of systems have the same underlying shape. Monotonicity characterizes the underlying positive or negative trend of the parameter. Finally, prognosability gives a measure of the variance in the critical failure value of a population of systems. By quantifying these features for a given parameter, the metrics can be used with any traditional optimization technique, such as Genetic Algorithms, to identify the optimal parameter for a given system. An appropriate parameter may be used with a General Path Model (GPM) approach to make RUL estimates for specific systems or components. A dynamic Bayesian updating methodology is introduced to incorporate prior information in the GPM methodology. The proposed methods are illustrated with two applications: first, to the simulated turbofan engine data provided in the 2008 Prognostics and Health Management Conference Prognostics Challenge and, second, to data collected in a laboratory milling equipment wear experiment. The automated system was shown to identify appropriate parameters in both situations and facilitate Type III prognostic model development.

A Multiagent Framework for a Diagnostic and Prognostic System Irtaza Barlas 124 Pages Directed By: Dr. George Vactsevanos The shortcomings of the current diagnostic and prognostic systems stem from the limitations of their frameworks. The framework is typically designed on the passive, open loop, static, and isolated notions of diagnostics, in that the framework does not observe its diagnostic results (open-looped), hence can not improve its performance (static). Its passivity is attributed to the fact that an external event triggers the diagnostic or prognostic action. There is also no effort in place to team-up the diagnostic systems for a collective learning, hence the implementation is isolated. In this research we extend the current approaches of the design and implementation of diagnostic and prognostic systems by presenting a framework based upon Multiagent systems. This research created novel architectures by providing such unique features to the framework, as learning, reasoning, and coordination. As the primary focus of the research the concept of Case-Based Reasoning was exploited to reason in the temporal domain to generate better prognosis, and improve the accuracy of detection as well as prediction. It was shown that the dynamic behavior of the intelligent agent helps it to learn over time, resulting in improved performance. An analysis is presented to show that a coordinated effort to diagnose also makes sense in uncertain situations when there are certain number of systems attempting to communicate certain number of failures, since there can be high probability of finding a shareable experience.

In shipboard power systems (SPS), it is important to provide continuous power to vital loads so that their desired missions can be completed successfully. Several components exist between the primary source and the vital load such as transformers, cables, or switching devices. These components can fail due to mechanical stresses, electrical stresses, and overloading which could lead to a system failure. If the normal path to a vital load cannot supply power to it, then it should be powered through its alternate path. The process of restoring, balancing, and minimizing power losses to loads is called network reconfiguration. Prognostics is the ability to predict precisely and accurately the remaining useful life of a failing component. In this work, the prognostic information of the power system components is used to determine if reconfiguration should be performed if the system is unable to accomplish its mission. Each component will be analyzed using the Weibull Distribution to compute the conditional reliability from present time to the end of the mission. To determine if reconfiguration is needed, all components to a given load will be utilized in structure functions to determine if a load will be able to survive during a time period. Structure functions are used to show how components are interconnected, and also provide a mathematical means for computing the total probability of a system. This work will provide a method to compute the conditional survivability to a given load, and the results indicate the top five loads that have the lowest conditional survivability during a mission in known configuration. The results show the computed conditional survivability of loads on an all electric navy ship. The loads conditional survivability is computed on high/medium voltage level and a low voltage level to show how loads are affected by failing components along their path.

Due to rapid advancements in sensing and computation technology, multiple types of sensors have been embedded in various applications, on-line automatically collecting massive production information. Although this data-rich environment provides great opportunity for more effective process control, it also raises new research challenges on data analysis and decision making due to the complex data structures, such as heterogeneous data dependency, and large-volume and high-dimensional characteristics. This thesis contributes to the area of System Informatics and Control (SIAC) to develop systematic data fusion methodologies for effective quality control and performance improvement in complex systems. These advanced methodologies enable (1) a better handling of the rich data environment communicated by complex engineering systems, (2) a closer monitoring of the system status, and (3) a more accurate forecasting of future trends and behaviors. The research bridges the gaps in methodologies among advanced statistics, engineering domain knowledge and operation research. It also forms close linkage to various application areas such as manufacturing, health care, energy and service systems. This thesis started from investigating the optimal sensor system design and conducting multiple sensor data fusion analysis for process monitoring and diagnosis in different applications. In Chapter 2, we first studied the couplings or interactions between the optimal design of a sensor system in a Bayesian Network and quality management of a manufacturing system, which can improve cost-effectiveness and production yield by considering sensor cost, process change detection speed, and fault diagnosis accuracy in an integrated manner. An algorithm named “Best Allocation Subsets by Intelligent Search” (BASIS) with optimality proof is developed to obtain the optimal sensor allocation design at minimum cost under different user specified detection requirements. Chapter 3 extended this line of research by proposing a novel adaptive sensor allocation framework, which can greatly improve the monitoring and diagnosis capabilities of the previous method. A max-min criterion is developed to manage sensor reallocation and process change detection in an integrated manner. The methodology was tested and validated based on a hot forming process and a cap alignment process. Next in Chapter 4, we proposed a Scalable-Robust-Efficient Adaptive (SERA) sensor allocation strategy for online high-dimensional process monitoring in a general network. A monitoring scheme of using the sum of top-r local detection statistics is developed, which is scalable, effective and robust in detecting a wide range of possible shifts in all directions. This research provides a generic guideline for practitioners on determining (1) the appropriate sensor layout; (2) the “ON” and “OFF” states of different sensors; and (3) which part of the acquired data should be transmitted to and analyzed at the fusion center, when only limited resources are available. To improve the accuracy of remaining lifetime prediction, Chapter 5 proposed a data-level fusion methodology for degradation modeling and prognostics. When multiple sensors are available to measure the degradation mechanism of a same system, it becomes a high dimensional and challenging problem to determine which sensors to use and how to combine them together for better data analysis. To address this issue, we first defined two essential properties if present in a degradation signal, can enhance the effectiveness for prognostics. Then, we proposed a generic data-level fusion algorithm to construct a composite health index to achieve those two identified properties. The methodology was tested using the degradation signals of aircraft gas turbine engine, which demonstrated a much better prognostic result compared to relying solely on the data from an individual sensor. In summary, this thesis is the research drawing attention to the area of data fusion for effective employment of the underlying data gathering capabilities for system modeling, performance assessment and improvement. The fundamental data fusion methodologies are developed and further applied to various applications, which can facilitate resources planning, real-time monitoring, diagnosis and prognostics.

abstract: Prognostics and health management (PHM) is a method that permits the reliability of a system to be evaluated in its actual application conditions. This work involved developing a robust system to determine the advent of failure. Using the data from the PHM experiment, a model was developed to estimate the prognostic features and build a condition based system based on measured prognostics. To enable prognostics, a framework was developed to extract load parameters required for damage assessment from irregular time-load data. As a part of the methodology, a database engine was built to maintain and monitor the experimental data. This framework helps in significant reduction of the time-load data without compromising features that are essential for damage estimation. A failure precursor based approach was used for remaining life prognostics. The developed system has a throughput of 4MB/sec with 90% latency within 100msec. This work hence provides an overview on Prognostic framework survey, Prognostics Framework architecture and design approach with a robust system implementation. / Dissertation/Thesis / M.S. Computer Science 2010

abstract: In-situ fatigue damage diagnosis and prognosis is a challenging problem for both metallic and composite materials and structures. There are various uncertainties arising from material properties, component geometries, measurement noise, feature extraction techniques, and modeling errors. It is essential to manage and incorporate these uncertainties in order to achieve accurate damage detection and remaining useful life (RUL) prediction. The aim of this study is to develop an integrated fatigue damage diagnosis and prognosis framework for both metallic and composite materials. First, Lamb waves are used as the in-situ damage detection technique to interrogate the damaged structures. Both experimental and numerical analysis for the Lamb wave propagation within aluminum are conducted. The RUL of lap joints under variable and constant fatigue loading is predicted using the Bayesian updating by incorporating damage detection information and various sources of uncertainties. Following this, the effect of matrix cracking and delamination in composite laminates on the Lamb wave propagation is investigated and a generalized probabilistic delamination size and location detection framework using Bayesian imaging method (BIM) is proposed and validated using the composite fatigue testing data. The RUL of the open-hole specimen is predicted using the overall stiffness degradation under fatigue loading. Next, the adjoint method-based damage detection framework is proposed considering the physics of heat conduction or elastic wave propagation. Different from the classical wave propagation-based method, the received signal under pristine condition is not necessary for estimating the damage information. This method can be successfully used for arbitrary damage location and shape profiling for any materials with higher accuracy and resolution. Finally, some conclusions and future work are generated based on the current investigation. / Dissertation/Thesis / Doctoral Dissertation Mechanical Engineering 2016