Prognostic Health Monitoring Research Articles

A Review of DC Fast Chargers with BESS for Electric Vehicles: Topology, Battery, Reliability Oriented Control and Cooling Perspectives

The global promotion of electric vehicles (EVs) through various incentives has led to a significant increase in their sales. However, the prolonged charging duration remains a significant hindrance to the widespread adoption of these vehicles and the broader electrification of transportation. While DC-fast chargers have the potential to significantly reduce charging time, they also result in high power demands on the grid, which can lead to power quality issues and congestion. One solution to this problem is the integration of a battery energy storage system (BESS) to decrease peak power demand on the grid. This paper presents a review of the state-of-the-art use of DC-fast chargers coupled with a BESS. The focus of the paper is on industrial charger architectures and topologies. Additionally, this paper presents various reliability-oriented design methods, prognostic health monitoring techniques, and low-level/system-level control methods. Special emphasis is placed on strategies that can increase the lifetime of these systems. Finally, the paper concludes by discussing various cooling methods for power electronics and stationary/EV batteries.

A new tool for model examination: Estimation of the mediator transfer function between the model and measured signals

Understanding the phenomenon underlying the behavior of rotating machinery by models can foster the development of algorithms for prognostic health monitoring. Model development is based on several assumptions and results in differences between the simulated and measured signals. In every system, there is always a transfer function between the rotating component and the sensor. The transfer function is one of the main possible sources of differences between the simulated and measured signals, making interpretation of physical results or further model development difficult. In this paper, a new paradigm is proposed to mitigate this problem. The new paradigm facilitates a quantitative comparison between models and elucidates the components that need further development or meticulous re-examination. We examine the performance of the paradigm using measured signals and transfer functions, and demonstrate it on a physical model of a gearbox with its corresponding measured signals.

IIoT Based Anomaly Detection and Maintenance Management of an Industrial Rotary System

Numerous challenges are being faced by implementing the Industrial Internet of Things (IIoT), which enables anomaly detection and optimal maintenance management of industrial production systems and their critical machines. Industries have started adopting this revolutionary technology along with other allied technologies to reap the full benefits out of Industry 4.0 environment. The objectives of this research work were to use the IIoT to develop a continuous monitoring system of the behavior of bottleneck facilities in a production system, to predict and avoid all possible failures, and to improve the overall productivity of the manufacturing system. The proposed prognostic health monitoring system employs IIoT sensors to measure the current values of the operating parameters of a machine, using built-in intelligent decision support mechanism to compare with optimal ranges of values, and to message the appropriate alarming signals as per the severity of the deviation. The system developed was tested with a prototype model comprising the Internet of Things, internet communication technology, and a machine learning algorithm, MEMS with standard input, output, storage and display of components, which was developed in a laboratory but implemented in a case study in a real industrial production plant. After the successful implementation of the developed system, the performance of the critical machine was evaluated in terms of metrics such as the average number of failures, average downtime and average service time spent. It was found that after the implementation, the downtime has decreased by nearly 22% and for the performance in terms of its output, the flow rate exhibited a steady increase with the passage of time.

Health monitoring of a conveyor belt system using machine vision and real-time sensor data

To truly achieve prognostic health monitoring of complex manufacturing systems, real-time data from multiple sensors must be continuously monitored in order to have a system-wide view of the health of the machine and the effect of different operating conditions on system performance. Belt misalignment is a critical factor in the conveyor belt system traditionally assessed through mechanical means. This research successfully implements multiple sensors and communication protocols along with machine vision for monitoring the health of a conveyor belt system. Data is captured using industrial-grade sensors and conditioned in a PLC. This information is sent to a wireless radio frequency module that transmits wirelessly to a remote receiver paired to an IoT gateway. Images taken by a camera are simultaneously processed by a local computer running a machine vision algorithm used to establish if the conveyor belt is operating normally. Both the sensor data and the machine vision information are sent to the cloud for remote users to monitor the system's operating conditions and to detect potential failure prior to its occurrence. The system communicates using a set of Radio Frequency modules, Zumlink™, and the range of this system was tested and determined to be greater than 2.5 miles with obstructions typical in a manufacturing environment.

Fault Detection and Prognostic Health monitoring of Towed array sonars

Sonars are used to detect underwater targets and are especially important in maintaining naval superiority. Towed array sonars can operate at very low frequencies thus giving larger ranges and can be deployed to any desired depth of operation. Towed array sonars offer long range surveillance capability and is the sensor of choice for sustained surveillance operations. Reliable operation and maintenance of towed array sonars need effective methods of health monitoring and reliability prediction. For any prognostic health monitoring to be done we need to identify certain parameters which can be observed and will give system health status in the present condition. This paper proposes some metrics which are easily measurable in-situ and which offer insights into the health of the sonar system. These metrics give direct measureable impact for each of the failure modes and offer insights into the current health of an operational towed array sonar. Simulation results are shown to demonstrate the effectiveness of the proposed metrics and detailed trial data results from different towed array trials are analysed to validate them in operational scenarios.

A fiber optic conjugate stress sensor for instantaneous tangent modulus detection targeting prognostic health monitoring applications

In this work, a surface-mounted fiber optic conjugate stress sensor (FOCSS) is developed, fabricated, and experimentally demonstrated to accurately evaluate the instantaneous tangent modulus of various metals, thus allowing measurement of the elastic-plastic stress–strain behavior of the host material. The originally proposed embedded version of the conjugate stress (CS) sensor measured the deformations experienced in two fully embedded, ellipsoidal, collocated sensors of significantly different stiffnesses, and related those sensor deformations to host modulus via Eshelby’s equivalent inclusion method. In contrast, the current experimental study focuses on a surface-mounted implementation of the CS sensor. As Eshelby’s analysis is not applicable in this case, a simplified lumped-element spring model is developed to instantaneously relate the deformations in the CS sensor to the local stress–strain state of the host. Therefore, tangent modulus of the host can be estimated from the output of the FOCSS, which can serve as an important damage precursor in structural and prognostic health monitoring for a prominent set of failure mechanisms, including mechanical overstress, fatigue, and corrosion. In this study, the FOCSS data is used in conjunction with the lumped-element spring model to measure the elastic modulus of three different materials of significantly different stiffness: aluminum, copper, and steel. Additionally, as validation of instantaneous tangent modulus tracking, the elastic-plastic stress–strain curves of copper and steel are reconstructed from the FOCSS outputs during uniaxial tensile tests (with a goodness of fit R 2 > 0.98). The results demonstrate the ability of the FOCSS to detect instantaneous modulus as materials experience plastic deformation.

Transfer Across Different Machines by Transfer Function Estimation.

A digital twin is a promising evolving tool for prognostic health monitoring. However, in rotating machinery, the transfer function between the rotating components and the sensor distorts the vibration signal, hence, complicating the ability to apply a digital twin to new systems. This paper demonstrates the importance of estimating the transfer function for a successful transfer across different machines (TDM). Furthermore, there are few algorithms in the literature for transfer function estimation. The current algorithms can estimate the magnitude of the transfer function without its original phase. In this study, a new approach is presented that enables the estimation of the transfer function with its phase for a gear signal. The performance of the new algorithm is demonstrated by measured signals and by a simulated transfer function.

On integrating prior knowledge into Gaussian processes for prognostic health monitoring

Gaussian process regression is a powerful method for predicting states based on given data. It has been successfully applied for probabilistic predictions of structural systems to quantify, for example, the crack growth in mechanical structures. Typically, predefined mean and covariance functions are employed to construct the Gaussian process model. Then, the model is updated using current data during operation while prior information based on previous data is ignored. However, predefined mean and covariance functions without prior information reduce the potential of Gaussian processes.This paper proposes a method to improve the predictive capabilities of Gaussian processes. We integrate prior knowledge by deriving the mean and covariance functions from previous data. More specifically, we first approximate previous data by a weighted sum of basis functions and then derive the mean and covariance functions directly from the estimated weight coefficients. Basis functions may be either estimated or derived from problem-specific governing equations to incorporate physical information.The applicability and effectiveness of this approach are demonstrated for fatigue crack growth, laser degradation, and milling machine wear data. We show that well-chosen mean and covariance functions, like those based on previous data, significantly increase look-ahead time and accuracy. Using physical basis functions further improves accuracy. In addition, computation effort for training is significantly reduced.

Lifetime and degradation analysis of AgPt alloy thick film/AlN heater for semiconductor wafer annealing

In this study, a heater for 300-mm-diameter semiconductor wafer annealing was fabricated by high-temperature firing after screen-printing a silver–platinum (AgPt) paste on an aluminium nitride (AlN) substrate. For life testing, the changes in the electrical characteristics and thermodynamic properties of the Ag65Pt20-alloy thick film on AlN were analysed according to aging by varying the temperature. Ag65Pt20/AlN heater specimens were fabricated to have a resistance of 10 Ω for thermal life tests. The input power was adjusted in the range of 10%–50%, and the temperature was set between 200 °C–600 °C. The Ag65Pt20/AlN heater was aged for approximately 3000 h per unit input power to measure the changes in electrical characteristics. The resistance increased with aging at each temperature and decreased subsequently. The activation energy was 88.7 kJ/(mol·K) [0.92 eV/K]. Moreover, the resistance decreased from the electrode with AgPt metallization, where the bias is applied, to the inner part. Physical analysis of the Ag65Pt20/AlN specimen revealed that the modified Ag65Pt20 alloy exhibited Pt-rich dendrite phases. The separated Pt-rich dendrite phases increased the capacitance. Further, electrical characteristics and material thermodynamic characteristics were analysed through high-temperature aging of the Ag65Pt20/AlN heater. Through such analyses, prognostic health monitoring factors of the Ag-Pt/AlN heater were deduced, thus enabling service lifetime period prediction for predictive maintenance.

Principal Component Analysis and Independent Component Analysis-Based Prognostic Health Monitoring of Electronic Assemblies Subjected to Simultaneous Temperature-Vibration Loads

Abstract This paper discusses methods for estimating different feature vectors from strain signals of an electronic assembly under combined temperature and vibration load. A vibrational load of 14 G acceleration-level with an ambient temperature of 55 °C is selected as the operating conditions for this experiment. Strain signals were measured at different time intervals during the vibration of the printed circuit board, and resistance values of the packages on the printed circuit board are monitored to identify the failure. The frequency response was measured by taking the fast Fourier transform of the signal and quantized by frequency quantization techniques. These techniques were able to identify the increase in the number of higher frequency components in the strain signal before failure with increase vibration time. The time-frequency response was also compared by employing different time–frequency analysis, joint time–frequency analysis, and statistical techniques such as principal component analysis (PCA), and independent component analysis (ICA). Statistical techniques like PCA and ICA were used to identify the different patterns of the original strain and filtered signals. These techniques discretely separated the before and after failure strain signals but were unable to predict the progression of failure in the packages. The instantaneous frequency of the strain signal displayed an interesting behavior, in which the variance of the PCA components of the instantaneous frequency had an increasing trend and reached a maximum value before continuously decreasing and reaching a lower value just before failure, indicating a progression of the before failure strain components.

A Dynamic Bayesian Network Structure for Joint Diagnostics and Prognostics of Complex Engineering Systems

Dynamic Bayesian networks (DBNs) represent complex time-dependent causal relationships through the use of conditional probabilities and directed acyclic graph models. DBNs enable the forward and backward inference of system states, diagnosing current system health, and forecasting future system prognosis within the same modeling framework. As a result, there has been growing interest in using DBNs for reliability engineering problems and applications in risk assessment. However, there are open questions about how they can be used to support diagnostics and prognostic health monitoring of a complex engineering system (CES), e.g., power plants, processing facilities and maritime vessels. These systems’ tightly integrated human, hardware, and software components and dynamic operational environments have previously been difficult to model. As part of the growing literature advancing the understanding of how DBNs can be used to improve the risk assessments and health monitoring of CESs, this paper shows the prognostic and diagnostic inference capabilities that are possible to encapsulate within a single DBN model. Using simulated accident sequence data from a model sodium fast nuclear reactor as a case study, a DBN is designed, quantified, and verified based on evidence associated with a transient overpower. The results indicate that a joint prognostic and diagnostic model that is responsive to new system evidence can be generated from operating data to represent CES health. Such a model can therefore serve as another training tool for CES operators to better prepare for accident scenarios.

A Data-Driven Machine Learning Approach for Corrosion Risk Assessment—A Comparative Study

Understanding the corrosion risk of a pipeline is vital for maintaining health, safety and the environment. This study implemented a data-driven machine learning approach that relied on Principal Component Analysis (PCA), Particle Swarm Optimization (PSO), Feed-Forward Artificial Neural Network (FFANN), Gradient Boosting Machine (GBM), Random Forest (RF) and Deep Neural Network (DNN) to estimate the corrosion defect depth growth of aged pipelines. By modifying the hyperparameters of the FFANN algorithm with PSO and using PCA to transform the operating variables of the pipelines, different Machine Learning (ML) models were developed and tested for the X52 grade of pipeline. A comparative analysis of the computational accuracy of the corrosion defect growth was estimated for the PCA transformed and non-transformed parametric values of the training data to know the influence of the PCA transformation on the accuracy of the models. The result of the analysis showed that the ML modelling with PCA transformed data has an accuracy that is 3.52 to 5.32 times better than those carried out without PCA transformation. Again, the PCA transformed GBM model was found to have the best modeling accuracy amongst the tested algorithms; hence, it was used for computing the future corrosion defect depth growth of the pipelines. This helped to compute the corrosion risks using the failure probabilities at different lifecycle phases of the asset. The excerpts from the results of this study indicate that my technique is vital for the prognostic health monitoring of pipelines because it will provide information for maintenance and inspection planning.

Towards Automated Condition Monitoring of Blowout Preventer Wellbore Packers

According to Bureau of Safety and Environmental Enforcement (BSEE), in 2017 a total of 45 out of 59 rigs operating in the Gulf of Mexico reported component failures of well control related equipment. The aftermath of the oil spill from the Deepwater Horizon rig provides stark illustration that acceptance of the equipment failure status quo is untenable. In this work, the authors propose novel automated strategies to monitor the health of blowout preventers (BOPs) on offshore drilling rigs. Using physicsbased models, we demonstrate computational detection of pressure tests by identifying characteristic features in timeseries pressure data. After detection, we present a methodology for extracting features relevant for prognostic health monitoring, including pressure decay and hold durations. Augmenting these computational models with domain knowledge from BOP experts, we produce health indices (HIs) for the respective equipment as the output. In addition, we demonstrate the optimized enumeration and identification of possible pressure test plans for BOPs in different combinatorial configurations. By combining our novel detection approach with health indices and automated test planning, this work contributes to prediction and amelioration of well control equipment failures on offshore drilling rigs.

A combined framework for automated diagnosis and prognosis based on surrogate modelling and a particle filter

Bayesian approaches have proven to be successful in prognostic health monitoring, especially for inverse problem solutions, such as the estimation of damage evolution model parameters from some damage dependent observations. In practice, they stem from the evaluation of the posterior distribution of a vector of parameters of interest conditioned on the observation of some signal features, which relies on the direct calculation of the observation likelihood, i.e. a measure of the probability that the observations are associated to some realizations of the system parameter vector. However, for realistic structures, a numerical simulation might be required for the evaluation of each sample likelihood, which can make the whole procedure for posterior pdf estimation computationally unfeasible. In this work, this problem is addressed by leveraging on surrogate modelling. Particle filter is used as a general framework for the combined health state estimation and prognosis of a skin panel subject to fatigue crack growth, while observing the strain field pattern acquired at some specific locations. A surrogate model consisting of an artificial neural network, trained on a set of analytical simulations, is used to predict the strain as a function of the crack position and length, thus allowing a fast calculation of the strain observation likelihood. The algorithm is tested with an analytic case study of a crack propagating in an infinite plate, allowing for a simultaneous diagnosis of the crack position and length, as well as a real-time updating of the evolution model parameters and system prognosis.

Prognostic Health Monitoring Method for Fatigue Failure of Solder Joints on Printed Circuit Boards Based on a Canary Circuit

Devices mounted on printed circuit boards (PCBs) are subject to temperature variations resulting from power switching and ambient temperature changes, and may be subject to random dynamic load histories from sources such as vibration. Since solder material is mechanically the most ductile part, fatigue failure may occur in solder joints. Health monitoring for fatigue life under field conditions is a key issue for improving availability and serviceability for maintenance. We have developed a failure precursor detection technology and a fatigue life estimation method for ball grid array (BGA) solder joints, based on a canary circuit. This method estimates fatigue failure life of an actual circuit by detecting failure connections in a canary circuit (a dummy circuit of daisy-chained solder joints). The canary circuit is designed to fail before the actual circuit under the same failure mode by using accelerated reliability testing and inelastic stress simulation. A feasibility study of the failure probability estimation method is conducted by applying the method to a PCB on which a BGA component is mounted. It is confirmed that the fatigue life under a thermal cyclic load can be estimated from a canary circuit, that estimation of fatigue life under a random dynamic load is feasible, and that the estimation results are consistent with results from actual random vibration tests. The proposed method is found to be useful for prognostic health monitoring of solder joint fatigue failure.

Development of an Effective Strategy for Prognostic Monitoring of a Large Centrifugal Air Compressor in an Automotive Plant

Prognostic monitoring of health condition of a large centrifugal air compressor that supplies compressed air in an automotive plant is crucial because its failure will seriously impair operation of the entire plant. It was desired todevelop an effective prognostic maintenance methodology of air compressors after the failure of an air compressor in one of major automotive companies in US, which brought a highly undesirable situation to the manufacturing line of the plant. In this work, the shaft motion of the compressor measured at transient and steady-state conditions were used to develop techniques and a strategy for effective prognostic monitoring. The pseudo frequency response function (FRF) obtained from the Campbell diagram and directional Power Spectrum (dPS) were new techniques employed to develop the prognostic health monitoring strategy. The analytic wavelet transform (AWT) is adopted to monitor temporal change of the system characteristics during the start-up period. In addition, AWT was utilized to monitor the steady state condition.

FIXING THE CHAIN: THE DATA ENRICHMENT CYCLE IN PROGNOSTIC HEALTH MONITORING

Only a very modest percentage of maintenance of capital goods is done condition-based, as opposed to plan-based or corrective. Doing much more would be very beneficial from a safety and operations perspective. Moreover, the basic methods have been known for some time. This paper addresses the question that, if condition-based maintenance (CBM) or Prognostic Health Management (PHM), as the terms commonly used to denote the same maintenance policy, is so logical and beneficial, why is it currently not more often applied in practice?This paper suggests that the answer lies in a complex interplay of technical and organizational issues. This interplay is best conceptualized as a data enrichment chain, which starts and ends at the physical asset to be maintained and ends at the actual prognostic maintenance being executed and evaluated. This chain is described in more detail, and illustrated with examples from an open innovation project in the process industry in the Netherlands, the CAMPIONE fieldlab. From this description, it becomes evident that this chain is currently broken in many places. If CBM/PHM is to become more established as a business practice, these breaks will need to be fixed. The paper discusses several of the options of how to fix this data enrichment chain in practice.

A New Feature Extraction Technique Based on 1D Local Binary Pattern for Gear Fault Detection

Gear fault detection is one of the underlying research areas in the field of condition monitoring of rotating machines. Many methods have been proposed as an approach. One of the major tasks to obtain the best fault detection is to examine what type of feature(s) should be taken out to clarify/improve the situation. In this paper, a new method is used to extract features from the vibration signal, called 1D local binary pattern (1D LBP). Vibration signals of a rotating machine with normal, break, and crack gears are processed for feature extraction. The extracted features from the original signals are utilized as inputs to a classifier based onk-Nearest Neighbour (k-NN) and Support Vector Machine (SVM) for three classes (normal, break, or crack). The effectiveness of the proposed approach is evaluated for gear fault detection, on the vibration data obtained from the Prognostic Health Monitoring (PHM’09) Data Challenge. The experiment results show that the 1D LBP method can extract the effective and relevant features for detecting fault in the gear. Moreover, we have adopted the LOSO and LOLO cross-validation approaches to investigate the effects of speed and load in fault detection.

Microelectronics packaging technology roadmaps, assembly reliability, and prognostics

This paper reviews the industry roadmaps on commercial-off-the shelf (COTS) microelectronics packaging technologies covering the current trends toward further reducing size and increasing functionality. Due to the breadth of work being performed in this field, this paper presents only a number of key packaging technologies. The topics for each category were down-selected by reviewing reports of industry roadmaps including the International Technology Roadmap for Semiconductor (ITRS) and by surveying publications of the International Electronics Manufacturing Initiative (iNEMI) and the roadmap of association connecting electronics industry (IPC). The paper also summarizes the findings of numerous articles and websites that allotted to the emerging and trends in microelectronics packaging technologies. A brief discussion was presented on packaging hierarchy from die to package and to system levels. Key elements of reliability for packaging assemblies were presented followed by reliabilty definition from a probablistic failure perspective. An example was present for showing conventional reliability approach using Monte Carlo simulation results for a number of plastic ball grid array (PBGA). The simulation results were compared to experimental thermal cycle test data. Prognostic health monitoring (PHM) methods, a growing field for microelectronics packaging technologies, were briefly discussed. The artificial neural network (ANN), a data-driven PHM, was discussed in details. Finally, it presented inter- and extra-polations using ANN simulation for thermal cycle test data of PBGA and ceramic BGA (CBGA) assemblies.

The Model of Estimating the Loss of Maintenance Potential the High Mobility Wheeled Vehicle

In the paper, the model of estimating the loss of maintenance potential of high mobility wheeled vehicles within proving ground tests is presented. This problem is up-to-date, especially important for military vehicles. The authors have analysed the selected requirements directed to this class of vehicles. In the main body of the paper, some model that could be applied in on-line system of prognostic Health and Usage Monitoring System (pHUMS) was presented. This model can be used to analyse information about the maintenance process of vehicle before some maintenance decision would be made.