SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
The role of the human operator as a problem solver in man-machine systems such as vehicles, process plants, transportation networks, etc. is considered. Problem solving is discussed in terms of detection, diagnosis, and compensation. A wide variety of models of these phases of problem solving are reviewed and specifications for an overall model outlined.
As part of the Light Water Reactor Sustainability program, an ongoing research effort is being conducted on technical specifications surveillance interval extension of digital equipment in nuclear power plants. The research team is led by Idaho National Laboratory and includes Pacific Northwest National Laboratory, Technology Resources, and Oak Ridge National Laboratory. This research focuses on developing methods for applying the U.S. Nuclear Regulatory Commission (NRC)–approved guidance to implement a licensee-controlled, risk-informed surveillance frequency change program in digital instrumentation and control (I&C) systems that include self-diagnostics and online monitoring (OLM) capabilities. Although approved methods exist for extending technical specifications (TS) surveillance test intervals (STIs) for general equipment, including analog I&C equipment, gaps remain in technology and guidance on crediting newer digital equipment’s internal self-diagnostics and OLM characteristics. Previous research described a general methodology for crediting internal self-diagnostics for extending surveillance test intervals. The methodology used self-diagnostics to detect—and credited recovery from—failure. Self-diagnostics were also applied for performance monitoring during the extended surveillance interval. This report discusses the status of recent activities to evaluate the previously developed methodology using a pilot study. Although both a utility partner for a pilot study and a specific digital asset were identified in FY2020, delays in obtaining proprietary information resulted in a limited ability to fully evaluate the methodology, and further interactions were complicated by the COVID pandemic. Therefore, at that time, the use of public-domain information—along with current processes for surveillance interval extension through a surveillance frequency control program—identified the need to fully assess diagnostic coverage as part of the pilot study. Furthermore, self-diagnostics were also identified as a potential option to replace the drift analyses conducted as part of current STI extension procedures. In FY2022, the project was reconstituted with the industry partner, and information and data were made available by the industry partner to the research team for review. The shared information included failure event descriptions and data for a digital I&C system since its implementation, as well as recent STI extension interval reports developed by the utility partner on that digital I&C system. This report presents an evaluation of this information and data and describes an application of the proposed methodology cited above. The methodology seeks to take advantage of the self-diagnostics and OLM capabilities to reduce risk or reduce the level of qualitative monitoring assessment needed to perform a risk-informed STI extension using existing NRC approved guidance or both. Addressing these issues of STI extension by crediting self-diagnostics is likely to result in benefits for current and future nuclear power plant (NPP) operations, including lowering the barriers to adoption of digital I&C systems and increasing cost savings by deferring or eliminating unneeded preventive maintenance (tasks or checks or activities). Specifically, self-diagnostic and OLM capabilities of newer digital equipment being installed in non-safety and safety applications are designed to detect failures, provide early warning of potential failures, and notify plant operators to take appropriate action to reduce out of service (OOS) time thus protecting safety margins. Moreover, the equipment is expected to provide information that time-related operational degradation is identified early to ensure timely and planned corrective actions instead of a reactive and unplanned approach ahead of an extended-surveillance interval.
In this paper, we will assess the performance of a data-driven anomaly detection algorithm, the Inductive Monitoring System (IMS), which can be used to detect simulated Thrust Vector Control (TVC) system failures. However, the ability of IMS to detect these failures in a true operational setting may be related to the realistic nature of how they are simulated. As such, we will investigate both a low fidelity and high fidelity approach to simulating such failures, with the latter based upon the underlying physics. Furthermore, the ability of IMS to detect anomalies that were previously unknown and not previously simulated will be studied in earnest, as well as apparent deficiencies or misapplications that result from using the data-driven paradigm. Our conclusions indicate that robust detection performance of simulated failures using IMS is not appreciably affected by the use of a high fidelity simulation. However, we have found that the inclusion of a data-driven algorithm such as IMS into a suite of deployable health management technologies does add significant value.
A model-based failure diagnosis system has been proposed for real-time diagnosis of SSME failures. Actuation, sensor, and system degradation failure modes are all considered by the proposed system. In the case of SSME actuation failures, it was shown that real-time identification can effectively be used for failure diagnosis purposes. It is a direct approach since it reduces the detection, isolation, and the estimation of the extent of the failures to the comparison of parameter values before and after the failure. As with any model-based failure detection system, the proposed approach requires a fault model that embodies the essential characteristics of the failure process. The proposed diagnosis approach has the added advantage that it can be used as part of an intelligent control system for failure accommodation purposes.
Composite structures are gaining importance for use in the aerospace industry. Compared to metallic structures their behavior is less well understood. This lack of understanding may pose constraints on their use. One possible way to deal with some of the risks associated with potential failure is to perform in-situ monitoring to detect precursors of failures. Prognostic algorithms can be used to predict impending failures. They require large amounts of training data to build and tune damage model for making useful predictions. One of the key aspects is to get confirmatory feedback from data as damage progresses. These kinds of data are rarely available from actual systems. The next possible resource to collect such data is an accelerated aging platform. To that end this paper describes a fatigue cycling experiment with the goal to stress carbon-carbon composite coupons with various layups. Piezoelectric disc sensors were used to periodically interrogate the system. Analysis showed distinct differences in the signatures of growing failures between data collected at conditions. Periodic X-radiographs were taken to assess the damage ground truth. Results after signal processing showed clear trends of damage growth that were correlated to damage assessed from the X-ray images.
The effect of variations in the reliability of an automated monitoring system on human operator detection of automation failures was examined in two experiments. For four 30-min sessions, 40 subjects performed an IBM PC-based flight simulation that included manual tracking and fuel-management tasks, as well as a system-monitoring task that was under automation control. Automation reliability - the percentage of system malfunctions detected by the automation routine - either remained constant at a low or high level over time or alternated every 10 min from low to high. Operator detection of automation failures was substantially worse for constant-reliability than for variable-reliability automation after about 20 min under automation control, indicating that the former condition induced 'complacency'. When system monitoring was the only task, detection was very efficient and was unaffected by variations in automation reliability. The results provide the first empirical evidence of the performance consequences of automation-induced 'complacency'. We relate findings to operator attitudes toward automation and discuss implications for cockpit automation design.
Nuclear security is one of the defining challenges of our time. Nuclear threats range from deliberate dispersal of radioactive material to contaminate the vital infrastructure to diversion and smuggling of special nuclear material for clandestine nuclear programs and nuclear terrorism, respectively. There is an associated need to develop and sustain the nuclear forensics capabilities, which requires the understanding of complex processes that occur in plasmas of nuclear materials. The area of nuclear safety has seen a resurgence of public interest, and there is a concomitant need to safely store used nuclear fuel and detect structural material failure in nuclear power systems, especially in innovative reactor designs envisioned for future adoption. Laser-produced plasmas are complicated extreme environments that can generate intense and rich, highly specific signatures of nuclear and radiological materials, which can then be explored in a wide range of applications. They include interdiction and rapid detection of nuclear materials, including their isotopic composition, detection over long distances, laboratory simulation of weapons effects, monitoring the condition of structural materials in dry cask storage containers, and novel instrumentation for nuclear power systems. We present a compilation of recent representative examples of the application of laser spectroscopy, and laser-induced breakdown spectroscopy in particular, to nuclear safety and security problems. A case is made that spectroscopic techniques based on laser-produced plasmas offer complementary, and sometimes unique, capabilities that motivate the continued exploration of their efficient production and understanding of the signatures they produce
Classification machine learning models require high-quality labeled datasets for training. Among the most useful datasets for photovoltaic array fault detection and diagnosis are module or string current-voltage (IV) curves. Unfortunately, such datasets are rarely collected due to the cost of high fidelity monitoring, and the data that is available is generally not ideal, often consisting of unbalanced classes, noisy data due to environmental conditions, and few samples. In this paper, we propose an alternate approach that utilizes physics-based simulations of string-level IV curves as a fully synthetic training corpus that is independent of the test dataset. In our example, the training corpus consists of baseline (no fault), partial soiling, and cell crack system modes. The training corpus is used to train a 1D convolutional neural network (CNN) for failure classification. The approach is validated by comparing the model’s ability to classify failures detected on a real, measured IV curve testing corpus obtained from laboratory and field experiments. Results obtained using a fully synthetic training dataset achieve identical accuracy to those obtained with use of a measured training dataset. When evaluating the measured data’s test split, a 100% accuracy was found both when using simulations or measured data as the training corpus. When evaluating all of the measured data, a 96% accuracy was found when using a fully synthetic training dataset. The use of physics-based modeling results as a training corpus for failure detection and classification has many advantages for implementation as each PV system is configured differently, and it would be nearly impossible to train using labeled measured data.
Circuits for redundant space-borne systems - detection and location of failure
As offshore wind systems become more complex, the risk of human error or equipment malfunction increases during experimental testing. This study investigates a lab-scale incident involving a 1 : 50 scale 5 MW wind turbine, where a generator failure led to rotor overspeed and a blade–tower strike. To improve early fault detection, we propose a data-driven method based on multivariate long short-term memory (LSTM) models. High-frequency measurements are projected onto principal components, and anomalies are identified using reconstruction error and its time derivative. Two models are trained on different healthy datasets and tested using single- and multi-principal component (1PC and MPC) variations. Results show that combining both error and error derivative improves detection accuracy. The 1PC model detects faults faster, has a higher recall rate, and achieves a 43 % improvement in anomaly detection accuracy, while the MPC model yields higher precision. This approach provides a simple and effective tool for early anomaly detection in lab-scale experiments, helping to reduce the risk of future failures during the testing of new technologies.
Electric Power Group, LLC (EPG) was awarded DOE OE0000850 to design, develop, and demonstrate a real-time software application for substation secondary equipment health monitoring and management at host utility American Electric Power (AEP), a cost share partner. This project addresses the need for monitoring substation equipment health and providing operators with tools to identify and take pre-emptive action to avoid catastrophic equipment failure. By monitoring synchrophasor data in real time, data anomalies that indicate potential asset failure can be detected and alerts can be sent to operators in time to take corrective actions. EPG developed two data driven algorithms to identify abnormal equipment signature patterns as well as a method using substation linear state estimation (SLSE). The software has been deployed on hardened PC’s and tested and validated for cost share partner AEP’s substations - 138kV and 765 kV. The SSHM software has been accepted by AEP and final demonstration of DOE - OE0000850 for AEP and DOE was successfully completed on March 17th, 2020. EPG developed the Substation Secondary Asset Health Monitoring (SSHM) Platform to analyze equipment failure signatures in PMU data and alert substation personnel when pre-emptive inspection, repairs and other actions are warranted to prevent potential catastrophic failures. This is the first system that automatically monitors the health of substation secondary assets and alerts users to emerging failures using synchrophasor measurements. SSHM uses high resolution PMU data from PTs, CTs, and CCVTs to analyze equipment signatures and identify anomalies that are precursor indicators of potential equipment failure. When an anomaly is detected, there is a likelihood of potential failure of monitored equipment, the system alerts the user with visual alarms including alarm trend charts and indicator lights on a oneline diagram of the substation.
Transistor second breakdown location, size and temperature detected and analyzed before failure by high resolution IR microscope
Owing to the frequency of occurrence and high risk associated with bearings, identification and characterization of bearing faults in motors via nondestructive evaluation (NDE) methods have been studied extensively, amongst which vibration analysis has been found to be a promising technique for early diagnosis of anomalies. However, a majority of the existing techniques rely on vibration sensors attached onto or in close proximity to the motor in order to collect signals with a relatively high SNR. Due to weight and space restrictions, these techniques cannot be used in unmanned aerial vehicles (UAVs), especially during flight operations since accelerometers cannot be attached onto motors in small UAVs. Small UAVs are often subjected to vibrational disturbances caused by multiple factors such as weather turbulence, propeller imbalance or bearing faults. Such anomalies may not only pose risks to UAV's internal circuitry, components or payload, they may also generate undesirable noise level particularly for UAVs expected to fly in low-altitudes or urban canyon. This paper presents a detailed discussion of challenges in in-flight detection of bearing failure in UAVs using existing approaches and offers potential solutions to detect overall vibration anomalies in small UAV operations based on IMU data.
For HPC systems to date, application resilience to faults and failures has been accomplished by the brute- force method of checkpoint/restart, which allows an application to make forward progress in the face of system and application faults, errors, and failures independent of root cause or end result. It has remained the primary resilience mechanism because we lack a way to identify faults and anticipate consequences early enough to take meaningful mitigating action. However, checkpoint/restart implementations put a tremendous burden on system resources and on the applications themselves and is becoming less feasible at scale. Because we have not yet operated at scales at which checkpoint/restart fails to provide forward progress, despite increasing costs, vendors have had little motivation to provide the instrumentation necessary for early identification of faults and failures. However, as we move from petascale to exascale, component mean time to failure (MTTF) will render the existing techniques ineffectual and/or too expensive. Furthermore, fault recovery mechanisms such as failover and/or error correction introduce performance inconsistency. Instrumentation allowing early indication of problems and tools to enable use of such information by systems, operating systems, and applications offer an alternative, more scalable and less costly solution. In the HMDR project, we built on our experience and expertise developed and accumulated over years of research on design, monitoring, measurement, and assessment of resilient computing systems. Analysis of field data on the current and past generations of extreme-scale systems revealed several challenges that, if not addressed in increasingly larger and more complex systems, may hinder the effectiveness of future exascale computing systems. Specifically, i) file systems and interconnects in current-generation large-scale systems already operate at the margins of resiliency, including consistent performance, and may not scale to larger deployments; ii) automated, software-based failover mechanisms are frequently inadequate and can introduce wider failures, such that failures during recovery may lead to system/application failures, including system-wide outages; and iii) silent data corruption represents a critical fault mode and will require efficient detection mechanisms if next-generation applications are to take full advantage of exascale hardware. To address the above challenges, we assembled a team of world-renowned experts in resilient extreme- scale computing from the University of Illinois (Electrical and Computer Engineering, Computer Science, and NCSA), SNL, LANL, NERSC, and Cray. Our team includes representatives from centers that house many of the largest HPC resources in the world, both today and over the coming years. The team has a unique track record of research in i) system and application failure characterization based on the analysis of field data, ii) data-driven design of fault/error detection mechanisms, and iii) experimental characterization of system/application resiliency. The team includes system owners/operators who provide continuous data collection and access and ensure installation of appropriate analysis tools.
Analytic techniques have been developed for detecting and identifying abrupt changes in dynamic systems. The GLR technique monitors the output of the Kalman filter and searches for the time that the failure occured, thus allowing it to be sensitive to new data and consequently increasing the chances for fast system recovery following detection of a failure. All failure detections are based on functional redundancy. Performance tests of the F-8 aircraft flight control system and computerized modelling of the technique are presented.
Commercial airplanes are becoming increasingly more sophisticated, placing an increasing burden on pilots to detect and resolve the exhaustive set of possible control effector failures. Automatic techniques are needed to either reconfigure an existing control law or restructure a new control law after failure. A discrete control law has been designed for the longitudinal channel of a mildly statically unstable commercial airplane, to track the glideslope during the approach to landing phase of flight. Single effector failures with time delays for failure detection and identification are analyzed for both the reconfigured and restructured control laws, and results are compared with those from previous research using a statically stable airplane. Strategies considered include reconfiguration and restructuring with new flight conditions. Validation of all cases is made using a 6 DOF nonlinear airplane simulation.
Failure mechanisms, test techniques, and quality control procedures for difficult-to-detect class of failures have been devised. Test circuit provides electrostatic discharge similar to those encountered in production situation and assists in evaluation of susceptibility of specific circuits and devices
The monitoring of control system sensors and actuators for failures is presently undertaken by means of an exceptionally simple form of generalized parity relations based on a discrete-time model of the dynamics of linear, time-invariant systems. These generalized parity relations are constructed by recourse to a transfer matrix-description of the system that is additionally useful in the interpretation of their properties. Attention is given to a novel method for constructing the parity relation of minimum length that depends on the output of only a single sensor.