Abstract: Systems and methods for identifying a faulty sensor are provided. One system includes multiple sensors, each generating a first set of parameters based on detected measurands. The system also includes a fusion module for generating a second set of parameters by fusing the first sets of parameters, and a sensor fault module including multiple fault templates defining sensor faults, an estimation module for generating a third set of parameters based on the first and second sets of parameters, and an evaluation module. The evaluation module is configured to compare the third set of parameters and the fault templates, and determine a sensor fault based on the comparison. A method includes receiving a first set of parameters based on measurands from multiple sensors and a second set of parameters based on a measurand from a sensor, and determining a sensor fault based on the first and second set of parameters.

Abstract: A method for benchmarking diagnostic algorithms for a particular application is provided. The diagnostic algorithms are rank ordered based on a specified criterion so as to weed out weak algorithms, selecting more robust algorithms, defined in some sense, for deployment. This is realized by evaluating various parameters subsequently mentioned. A normalized product entropy ratio parameter is obtained. A performance parameter vector is fixed to define a plurality of sensitivity parameters including a plurality of threshold parameters and a plurality of data parameters. The plurality of threshold parameters and the plurality of data parameters are perturbed to obtain a threshold sensitivity parameter and a data sensitivity parameter.

Abstract: A method for performing diagnostics for an engine comprises the steps of identifying an engine component as potentially being related to operational data of an engine, calculating a deviation from a thermodynamic model, and comparing the deviation with root cause deviation measures. The deviation relates the engine component to an adjustment to the thermodynamic model with respect to a variable of the thermodynamic model, based at least in part on the operational data. Each root cause deviation measure relates one of a plurality of potential root causes to the thermodynamic model with respect to the variable of the thermodynamic model.

Abstract: A health monitoring system for a vehicle system includes an operational support system including a plurality of managers and a decision support module. Each manager corresponds to a different sub-system of the vehicle system, and comprises a plurality of reasoners and a fusion block. Each reasoner is configured to obtain data and provide preliminary output regarding a different component of the sub-system based on the data. The fusion block is coupled to the plurality of reasoners, and is configured to receive the preliminary output and generating manager output based on the preliminary output. The decision support module is coupled to the plurality of managers, and is configured to receive the manager output and provide a decision support output based on the manager output.

Abstract: A health monitoring system for a vehicle system includes a plurality of managers and a decision support module. Each manager corresponds to a different sub-system of the vehicle system, and generates manager output pertaining to the sub-system based at least in part on an intermediate output. Each manager comprises a plurality of reasoners. Each reasoner corresponds to a different component of the sub-system, and comprises a plurality of modules and a reasoner fusion block. Each module obtains data regarding a different aspect of the component and generates a preliminary output based on the data. The reasoner fusion block is coupled to the plurality of modules, receives the preliminary output, and generates the intermediate output based on the preliminary output. The decision support module is coupled to the plurality of managers, receives the manager output therefrom, and provides a decision support output based on the manager output.

Abstract: A method for performing diagnostics on a system of a vehicle includes the steps of receiving a request form an external source, obtaining data relevant to the request, processing the request using the data, and generating a report based at least in part on the processing of the request. The report comprises a diagnostic vector and a life usage vector. The diagnostic vector comprises diagnostic information for the system based at least in part on the processing of the request. The life usage vector comprises information as to usage of the system throughout a life history of the system, based at least in part on the processing of the request.

Abstract: At least one statistical output associated with a process model and a rate of change associated with each statistical output is identified, such as by using historical data associated with the process model. The statistical outputs and the rates of change are used to generate a graphical display, such as a phase plane plot. Each point in the display is based on one of the statistical outputs and its associated rate of change. The graphical display could include multiple portions, such as quadrants, and one of the portions can be selected and highlighted. The different portions of the graphical display may represent whether the process model is a poor fit to current conditions and approaching a better fit, a poor fit and approaching a poorer fit, a good fit and approaching a poorer fit, and a good fit and approaching a better fit.

Abstract: Methods and apparatus are provided pertaining to a design of experiments. The method comprises generating a data set from historical data; identifying and removing any fault data points in the data set so as to create a revised data set; supplying the data points from the revised data set into a nonlinear neural network model; and deriving a simulator model characterizing a relationship between the input variables and the output variables. The apparatus comprises means for generating a data set from historical data; means for identifying and removing any fault data points in the data set so as to create a revised data set; means for supplying the data points from the revised data set into a nonlinear neural network model; and means for deriving a simulator model characterizing a relationship between the input variables and the output variables.

Abstract: A system for obtaining diagnostic information, such as evidence about a mechanism, within an algorithmic framework, including filtering and aggregating the information through, for instance, a stochastic process. The output may be an overall belief value relative to a presence of an item such as, for example, a fault in the mechanism.

Abstract: A system and method is provided for detecting anomalies in turbine engines emanating from the main shaft and/or main shaft bearings. The anomaly detection system includes a sensor data processor and a matrix analysis mechanism. The sensor data processor receives engine sensor data, including main engine speed data during spin down, and formats the engine sensor data into an appropriate matrix. The matrix analysis mechanism receives the sensor data matrix and performs a singular value analysis on the sensor data matrix to detect potential anomalies in the turbine engine main shaft and/or bearings. The output of the matrix analysis mechanism is passed to a diagnostic system where further evaluation of the anomaly detection determination can occur.

Abstract: A system and method is provided for fault detection in a turbine engine. The fault detection system and method uses discrete event system modeling to provide improved fault diagnosis and prognosis. The fault detection system and method receives sensor taken at multiple dynamic events occurring in different time windows. An event determination mechanism evaluates the received sensor data to determine if specified events have occurred. Indications of the occurrence of specified events are then passed to a discrete event system model. The discrete event system model analyzes the timing and sequencing of the event occurrences to determine if a fault has occurred. This method does not require a detailed modeling of the system, and thus can be applied to complex systems such as turbine engines.

Abstract: A system and method for combining conclusions from multiple fault detection techniques to isolate likely faults in a turbine engine is provided. The system and method provide the ability to effectively deal with multiple concurrent faults in the engine. Additionally, the embodiments of the invention provide the ability to correctly characterize multiple conclusions generated from evidence having different levels of interdependence. In one embodiment, the conclusions based on device data with high dependency are aggregated using a high dependency aggregation rule, and the resulting high-dependency sets are then further aggregated using a weak dependency rule. Finally, any conclusions based on independent evidence can be aggregated using an independent combination rule. The resulting aggregation determines which fault(s) are most likely indicated by the plurality of conclusions, taken into account the dependency of the device data used to generate the conclusions.

Abstract: A method for characterizing engine wear includes the steps of generating operational data representative of engine operation, comparing the operational data with baseline operational data generated by a baseline operational model of the engine and generating a first deviation vector based on this comparison, generating a plurality of data images of an engine component following engine operation, comparing each of the plurality of data images with a baseline image of the engine component and generating a second deviation vector based on this comparison, and quantifying a relationship between the first deviation vector and the second deviation vector. The first deviation vector represents variation between the operational data and the baseline operational data. The second deviation vector represents variation between the plurality of data images and the baseline images.

Abstract: A system and method for fault detection is provided. The fault detection system provides the ability to detect symptoms of fault in turbine engines and other mechanical systems that have nonlinear relationships. The fault detection system uses a neural network to perform a data representation and feature extraction where the extracted features are analogous to principal components derived in a principal component analysis. This neural network data representation analysis can then be used to determine the likelihood of a fault in the system.

Abstract: A system and method is provided for fault detection in a turbine engine. The fault detection system and method uses discrete event system modeling to provide improved fault diagnosis and prognosis. The fault detection system and method receives sensor taken at multiple dynamic events occurring in different time windows. An event determination mechanism evaluates the received sensor data to determine if specified events have occurred. Indications of the occurrence of specified events are then passed to a discrete event system model. The discrete event system model analyzes the timing and sequencing of the event occurrences to determine if a fault has occurred. This method does not require a detailed modeling of the system, and thus can be applied to complex systems such as turbine engines.

Abstract: Methods and apparatus are provided pertaining to a design of experiments. The method comprises generating a data set from historical data; identifying and removing any fault data points in the data set so as to create a revised data set; supplying the data points from the revised data set into a nonlinear neural network model; and deriving a simulator model characterizing a relationship between the input variables and the output variables. The apparatus comprises means for generating a data set from historical data; means for identifying and removing any fault data points in the data set so as to create a revised data set; means for supplying the data points from the revised data set into a nonlinear neural network model; and means for deriving a simulator model characterizing a relationship between the input variables and the output variables.

Abstract: A system and method for detecting erosion in turbine engine blades is provided. The blade erosion detection system includes a sensor data processor and a cluster analysis mechanism. The sensor data processor receives engine sensor data, including exhaust gas temperature (EGT) data, and augments the sensor data to determine sensor data residual values and the rate of change of the sensor data residual values. The augmented sensor data is passed to the cluster analysis mechanism. The cluster analysis mechanism analyzes the augmented sensor data to determine the likelihood that compressor blade erosion has occurred. Specifically, the cluster analysis mechanism performs a 2-tuple cluster feature analysis using Gaussian density functions that provide approximations of normal and eroded blades in a turbine engine. The 2-tuple cluster feature analysis thus provides the probability that the sensor data indicates erosion has occurred in the turbine engine.

Abstract: A system and method for fault detection is provided. The fault detection system provides the ability to detect symptoms of fault in turbine engines and other mechanical systems that have nonlinear relationships between two or more variables. The fault detection system uses a neural network to perform feature extraction from data for representation of faulty or normal conditions. The values of extracted features, referred to herein as scores, are then used to determine the likelihood of fault in the system. Specifically, the lower order scores, referred to herein as “approximate null space” scores can be classified into one or more clusters, where some clusters represent types of faults in the turbine engine. Classification based on the approximate null space scores provides the ability to classify faulty or nominal conditions that could not be reliably classified using higher order scores.

Abstract: A system and method is provided to detect surface anomalies. The detection system and method provides the ability to automatically detect and characterize anomalies on a surface, such as detecting and characterizing pitting corrosion on metal surfaces. The surface anomaly detection system and method uses an image of the surface under evaluation. The image data is passed to a noise filter. The noise filter uses an estimated characterization of the background noise to remove the background noise from the image data. The smoothed pixel intensity is then compared to a threshold to identify significant departures. The detected anomalies are then passed to a contour detector, which determines the contours of anomalies in the surface to provide more precise characterization of the detected anomalies. The detected closed contours of anomalies may then be passed to a defect severity estimator that provides an anomaly factor metric.

Abstract: A system having a combination including a predictive controller and a correlation model analysis which may include a principal component analysis (PCA) calculator. The system may encompass a discrete sampler of trajectories of variables from the predictive controller for a correlation model analysis calculation. There may be a compensator for providing initialization and estimation values of other system variables for the correlation model analysis. An uncertainty calculator connected to the predictive controller may output an anomaly impact prediction zone based on upper and lower bounds from the predictive controller.