Joshua D Isom

Abstract: A common approach to process monitoring based on principal component analysis (PCA) assumes that fault-free, noise-free data is sampled from a low-dimensional subspace. Although widely described and applied, process fault detection and isolation using PCA is not robust to outliers in the training data, is hard to properly tune, and is not capable of isolating multiple faults. A newly introduced method called principal component pursuit (PCP) optimally decomposes a data matrix as the sum of a low-rank matrix and a sparse matrix. When applied to the process monitoring problem, PCP simultaneously accomplishes the objectives of model building, fault detection, fault isolation, and process reconstruction with a single convex optimization problem, thereby overcoming the key shortcomings of PCA-based approaches for process monitoring. The use of PCP for process monitoring is described and illustrated using data from a manufacturing process.

Abstract: The time synchronous average (TSA) is a simple but powerful tool for extracting deterministic vibration components that are synchronous with one single shaft and thus diagnosing faults caused by shafts or gears. The task of extracting random vibration components associated with gears, shafts, or bearings is more challenging. The residual obtained by subtracting the single shaft TSA contains components from both bearings and shafts/gears do not synch with that shaft. This paper describes a simple approach for integrating the single shaft TSAs to generate a composite time synchronous average which can be subtracted from the original signal to generate a pure second-order cyclostationary residual. This approach is applied to vibration signals collected from a two-stage gearbox and compared with other techniques including an all-shaft TSA in which angular re-sampling over the minimum common multiple of shaft revolutions is conducted, high/low-pass filtering and self-adaptive noise cancellation (SANC). The results demonstrate that the proposed approach produces an integrated TSA signal is clean and includes only the shaft components, gear mesh components and the sidebands associated with all the shafts, while the bearing vibration signatures and random noise are kept in the residuals. The results produced by three alternative techniques do not separate the components well and have a low signal-to-noise ratio.

Abstract: The Integrated Vehicle Health Management System (IVHMS) fielded on the UH-60M aircraft has an open-system architecture and third-party software capability that has allowed Sikorsky Aircraft to expedite the development and insertion of advanced health-monitoring technologies. Sikorsky has used this capability over the past two years to develop real-time applications for 1) airframe and rotor system load estimation, 2) engine health monitoring, and 3) advanced planetary gearbox health monitoring. The first application was tested in a bench-test facility in 2008 and flight-tested in 2009; the second application was lab-tested and flight-tested in 2009; and the third application was lab-tested in 2009. Each application demonstrated unique features of the third-party software capability. Sikorsky has demonstrated that the entire algorithm implementation and verification process can be completed in three to six months.

Abstract: Probabilistic latent component analysis (PLCA) is applied to the problem of gearbox vibration source separation. A model for the probability distribution of gearbox vibration employs a latent variable intended to correspond to a particular vibration source, with the measured vibration at a particular sensor for each source the product of a marginal distribution of vibration by frequency, a marginal distribution of vibration by shaft rotation, and a sensor weight distribution. An expectation-maximization algorithm is used to approximate a maximum-likelihood parameterization for the model. In contrast to other unsupervised source-separation methods, PLCA allows for separation of vibration sources when there are fewer vibration sensors than vibration sources. Once the vibration components of a healthy gearbox have been identified, the vibration characteristics of damaged gearbox elements can be determined. The efficacy of the technique is demonstrated with an application on a gearbox vibration data set.

Abstract: Change detection problems are ubiquitous, but there are no existing tractable techniques for epsilon-exact solutionof discrete-time, discrete-observation problems. A polynomial-time algorithm for solution of Markov Bayes change detectionproblems is presented. The algorithm combines backward induction on a tree with induction of bounds on the value function at specific points from those of neighboring points.

Abstract: The Integrated Vehicle Health Management System (IVHMS) fielded with the Army UH-60M Black Hawk provides drivetrain diagnostics, rotor track and balance, exceedance, operational usage, engine performance, bearing monitoring, structural regime recognition, cockpit voice and flight data recording functionality. As with the systems predecessor, the Integrated Diagnostics Health and Usage Management System (IMD-HUMS), this vehicle health management system was developed with an open system architecture which supports third party technology insertion. This capability includes insertion of technology developed by other vendors internal and external to the various subsystems within the IVHMS product boundary. This paper presents an overview of the open systems architecture embedded within the current generation of Goodrich Corporation vehicle health products, interfaces available for third party technology insertion in the Integrated Health Management Unit (IVHMU) and experiences of the open systems integrator for the UH-60M, Sikorsky, with the IVHMU opens systems interface.

Abstract: A novel general framework for distributed anomaly detection with theoretical performance guarantees is proposed. Our algorithmic approach combines existing anomaly detection procedures with a novel method for computing global statistics using local sufficient statistics. Under a Gaussian assumption, our distributed algorithm is guaranteed to perform as well as its centralized counterpart, a condition we call ‘zero information loss’. We further report experimental results on synthetic as well as real-world data to demonstrate the viability of our approach.

Abstract: We describe an exact dynamic programming update for constrained partially observable Markov decision processes (CPOMDPs). State-of-the-art exact solution of unconstrained POMDPs relies on implicit enumeration of the vectors in the piecewise linear value function, and pruning operations to obtain a minimal representation of the updated value function. In dynamic programming for CPOMDPs, each vector takes two valuations, one with respect to the objective function and another with respect to the constraint function. The dynamic programming update consists of finding, for each belief state,the vector that has the best objective function valuation while still satisfying the constraint function. Whereas the pruning operation in an unconstrained POMDP requires solution of a linear program, the pruning operation for CPOMDPs requires solution of a mixed integer linear program.

Abstract: A system for reconstructing sensor data in a rotor system that comprises a rotating component of the rotor system, a plurality of sensors in the rotating component to sense at least one of loads and motion characteristics in the rotating component and to generate sensor data, and an analysis unit to generate reconstructed sensor data from the sensor data using numerical analysis for low-rank matrices.

Abstract: A system for monitoring for pushrod faults in an aircraft includes a pushrod; a sensor mounted on the pushrod, the sensor wirelessly transmitting measured pushrod load data; a receiver receiving the measured pushrod load data; a model receiving flight data and generating a modeled pushrod load profile; and a fault detector comparing measured pushrod load data to the modeled pushrod load profile to detect a pushrod fault.

Abstract: A computer-implemented method, system, and computer program product for virtual monitoring of aircraft fleet loads are provided. The method includes calculating virtual load data associated with an aircraft from a set of orthogonal waveforms. The method also includes calculating a set of coefficients as a function of parametric data and high frequency data associated with an aircraft. The method further includes storing the set of coefficients on the aircraft and transmitting the set of coefficients to a ground-based system configured to reproduce the virtual load data based on a copy of the set of orthogonal waveforms and the received set of coefficients in order to perform aircraft fleet management.

Abstract: A system and method for operating fuel cell power plant includes enclosing fuel bearing components, such as fuel cell stack and reformer, into a fuel compartment separate from motorized components in a motor compartment, and consuming leaked fuel in the fuel compartment using a fuel bearing component such as cell stack and/or burner, thereby reducing fuel emissions from the plant.

Abstract: A method of real-time rotor fault detection includes measuring a set of loads to obtain measured signals and virtually monitoring the set of loads to obtain estimated signals. The estimated signals are subtracted from the measured signals to obtain residuals and the residuals are compared to a categorical model. A categorical output representative of a rotor fault is identified within the categorical model.

Abstract: A fuel cell assembly includes a cell stack assembly having a flow field configured to provide a fluid flow. The fluid flow has an acid and a water content. A condenser heat exchanger is arranged downstream from and fluidly connected to the flow field by a fluid flow passage. The condenser heat exchanger is configured to receive the fluid flow from the flow field through the fluid flow passage. A water supply system including a water source in fluid communication with the flow passage is arranged downstream from the flow field. The water source is configured to provide additional water to the fluid flow at a water inlet and increase the water content. The increased water content is within the condenser heat exchanger.

Abstract: A fluidized contaminant separator and water-control loop decontaminates a fuel reactant stream of a fuel cell. Water passes over surfaces of an ammonia dissolving means within a fluidized bed while the fuel reactant stream simultaneously passes over the surfaces to dissolve contaminants from the fuel reactant stream into a separated contaminant and water stream. A fuel-control heat exchanger upstream from the scrubber removes heat from the fuel stream. A water-control loop directs flow of the separated contaminants and water stream from an accumulator through an ion exchange bed which removes contaminants from the stream. Decontaminated water is directed back into the scrubber to flow through the fluidized bed. Separating contaminants from the fuel reactant stream and then isolating and concentrating the separated contaminants within the ion exchange material minimizes costs and maintenance requirements.

Abstract: A phosphoric acid fuel cell (PAFC) system includes a cell stack assembly having an anode, a cathode and a coolant portion. At least one heat exchanger is fluidly interconnected with at least one of the anode, the cathode and the coolant portion and provides a fluid path for receiving a fluid from the anode, the cathode and/or the coolant portion. An absorption cycle refrigerant system includes an absorber having a solution of refrigerant and absorbent, and an absorbent loop and a refrigerant loop communicating with the absorber and respectively carrying absorbent and refrigerant. The at least one heat exchanger is arranged in the absorbent loop and is configured to transfer heat from the fuel cell system to the absorption chiller.

Abstract: Fuel processing by a reformer and a shift reactor converts hydrocarbon feedstock and steam to hydrogen-rich reformate, such as for use in a fuel cell power plant. Some of the reformate is recycled through a restriction to the inlet of a feedstock pump, thereby increasing its pressure sufficiently to cause recycle flow through a hydrodesulfurizer and the secondary inlet of an ejector driven by the steam. Recycle pressure is maintained by steam pressure through a valve regulated by a controller.

Abstract: The control system utilizes an oxygen sensor to sense an oxygen concentration within a burner exhaust of a fuel processing system wherein the burner device utilizes an anode exhaust stream from a fuel cell to supply heat to a reformer. If the anode utilization by the fuel cell anode exceeds an acceptable range, less hydrogen is available for the burner device and more oxygen will therefore be sensed by the oxygen sensor. An oxygen sensor controller, in response to the increase in sensed oxygen, increases flow of a fuel feedstock into the reformer to provide more hydrogen fuel to the anode to thereby return anode utilization to an acceptable anode utilization range. An opposite control sequence occurs if anode utilization falls below the acceptable range.

Abstract: Provision is made in an organic rankine cycle power plant for modulating the flow of hot gases from a thermal source to the evaporator. The modulation device may be a blower on the downstream side of the evaporator and/or a valve on the upstream side thereof. The modulation device is controlled by generation of a digital signal to enable or disenable the modulation device and an analog signal for adjusting the position of the modulation device.

Abstract: A homogeneous burner in a steam generator for fuel processing in a fuel cell power plant consumes unused hydrogen-containing reformate during startup of the fuel cell power plant, without damaging a catalytic burner which is upstream of the homogenous burner, and which consumes anode exhaust and produces steam during normal operation of the fuel cell power plant. During startup, little air is provided to the catalytic burner and a lot of air is provided to the homogenous burner; during normal operation, little air is provided to the homogenous burner but sufficient air is provided to the catalytic burner to consume all of the anode exhaust.

Abstract: Fuel from a source passes through a hydrogen desulfurizer and a proportioning mixing valve to a CPO, the temperature of the output of the CPO being monitored to provide a signal which a controller utilizes to adjust the valve. The output of the CPO may be passed through a water gas shift reactor and a preferential CO oxidizer to provide fuel to a fuel cell system. The air provided to the valve may be humidified, such as by an enthalpy recovery device receiving the oxidant outflow from the fuel cell system.

Abstract: The challenge of detecting a change in the distribution of data is a sequential decision problem that is relevant to many engineering solutions, including quality control and machine and process monitoring. This dissertation develops techniques for exact solution of change-detection problems with discrete time and discrete observations. Change-detection problems are classified as Bayes or minimax based on the availability of information on the change-time distribution. A Bayes optimal solution uses prior information about the distribution of the change time to minimize the expected cost, whereas a minimax optimal solution minimizes the cost under the worst-case change-time distribution. Both types of problems are addressed. The most important result of the dissertation is the development of a polynomial-time algorithm for the solution of important classes of Markov Bayes change-detection problems. Existing techniques for epsilon-exact solution of partially observable Markov decision processes have complexity exponential in the number of observation symbols. A new algorithm, called constellation induction, exploits the concavity and Lipschitz continuity of the value function, and has complexity polynomial in the number of observation symbols. It is shown that change-detection problems with a geometric change-time distribution and identically- and independently-distributed observations before and after the change are solvable in polynomial time. Also, change-detection problems on hidden Markov models with a fixed number of recurrent states are solvable in polynomial time. A detailed implementation and analysis of the constellation-induction algorithm are provided. Exact solution methods are also established for several types of minimax change-detection problems. Finite-horizon problems with arbitrary observation distributions are modeled as extensive-form games and solved using linear programs. Infinite-horizon problems with linear penalty for detection delay and identically- and independently-distributed observations can be solved in polynomial time via epsilon-optimal parameterization of a cumulative-sum procedure. Finally, the properties of policies for change-detection problems are described and analyzed. Simple classes of formal languages are shown to be sufficient for epsilon-exact solution of change-detection problems, and methods for finding minimally sized policy representations are described.

Abstract: A control method and arrangement are provided in a fuel cell power plant for regulating fuel flow to a steam-based fuel processing system (FPS) associated with a low-temperature fuel cell stack assembly. A portion of the fuel provided by the FPS is used to provide steam for the FPS. The fuel flow to the FPS is regulated as a function of the power demand (I) on the fuel cell and at least the enthalpy of the steam (P, T), such that the steam enthalpy is regulated to meet increases and decreases in power demand without exceeding steam pressure limits. In addition to reliance on steam pressure (P) as a fundamental measure of steam enthalpy, the control may additionally use reaction temperature (T) at, or in, a reformer, such as a catalytic steam reformer, to regulate fuel flow and thus, steam enthalpy.

Abstract: A fuel cell system includes a fuel cell for reacting a hydrogen rich gas; a fuel processor system for converting a hydrocarbon fuel-steam mixture into said hydrogen rich gas; and a system for preparing the hydrocarbon fuel-steam mixture which includes (a) structure for superheating a hydrocarbon fuel so as to provide a superheated fuel, and (b) structure for mixing water with the superheated fuel so as to provide the hydrocarbon fuel-steam mixture.

Abstract: A fuel cell power plant having a plurality of functionally integrated components including a fuel cell assembly provided with a fuel stream, an oxidant stream and a coolant stream. The fuel cell power plant functionally integrates a mass and heat recovery device for promoting a transfer of thermal energy and moisture between a first gaseous stream and a second gaseous stream, and a burner for processing the fuel exhausted from the fuel cell assembly during operation thereof. A housing chamber is utilized in which the oxidant stream exhausted from the fuel cell assembly merges with a burner gaseous stream exhausted from the burner. The resultant airflow from the common chamber is directed back to the mass and heat recovery device as the first gaseous stream.