Reliability Challenges Research Articles

Regardless of its Merits, Carbon Capture Faces Several Challenges

A theme I have struck multiple times in this column is the reality that energy policy designed to eliminate atmospheric carbon as quickly as possible must—to be genuinely sustainable through political cycles and reliability challenges—recognize that natural gas will be here and play an important role in our energy mix for a very long time to come. Even the Biden Administration, with its stated objective of phasing out all fossil fuels, has recognized this reality, through the projections of its own U.S. Energy Information Administration (EIA). In the 2023 Annual Energy Outlook prepared by EIA, one of its cases in assessing the future energy mix includes one in which non‐hydro resources, primarily solar and wind, become low‐cost enough to be competitive leaders in the marketplace. Figure 1 depicts EIA’s projection, which does see renewables surpassing natural gas in market share by the late 2040s but still leaves natural gas with a 25 percent share of domestic energy consumption, down only 8 percent from current levels. Of course, when exports of liquefied natural gas (LNG) are added to this, as the largest exporter of LNG in the world, the natural gas line is much flatter.

Design of soft error correction flip-flop cells for highly reliable applications

With the development of technology, the vulnerability of integrated circuits to Single Event Effect (SEE) increases, and the sensitivity of flip-flops to soft errors induced by Single Event Transient (SET) and Single Event Upset (SEU) increases. Fault tolerant design is necessary for high reliability applications. In this work, a Soft Error Correction Flip-Flop 1(SEC FF1) is proposed, which can correct soft errors induced by SET and SEU at low clock phase without additional clock cycle. It has the lowest area and power consumption overhead and the lowest Power-Delay Product (PDP) compared with other radiation hardened flip-flops. In order to correct soft errors earlier, a variant of SEC FF1, SECFF2, is proposed at the cost of power consumption and switching delay. As the feature size of transistors shrinks, Double Node Upset (DNU) induced by single-event charge sharing is becoming an emerging reliability challenge. In order tolerate DNU, variant of SEC FFs, DNU Correction (DNUC) FFs, are proposed. In addition, the delay element used in the radiation-hardened flip-flops is also introduced.

Service Reliability Based on Fault Prediction and Container Migration in Edge Computing

With improvements in the computing capability of edge devices and the emergence of edge computing, an increasing number of services are being deployed on the edge side, and container-based virtualization is used to deploy services to improve resource utilization. This has led to challenges in reliability because services deployed on edge nodes are pruned owing to hardware failures and a lack of technical support. To solve this reliability problem, we propose a solution based on fault prediction combined with container migration to address the service failure problem caused by node failure. This approach comprises two major steps: fault prediction and container migration. Fault prediction collects the log of services on edge nodes and uses these data to conduct time-sequence modeling. Machine-learning algorithms are chosen to predict faults on the edge. Container migration is modeled as an optimization problem. A migration node selection approach based on a genetic algorithm is proposed to determine the most suitable migration target to migrate container services on the device and ensure the reliability of the services. Simulation results show that the proposed approach can effectively predict device faults and migrate services based on the optimal container migration strategy to avoid service failures deployed on edge devices and ensure service reliability.

Optimizing Data Processing and Security in Edge Computing For Iot Applications

Edge computing has emerged as a pivotal paradigm in the realm of Internet of Things (IoT) applications, addressing challenges of latency, bandwidth, and reliability by decentralizing computation and storage closer to the data source. This paper explores the critical aspects of optimizing data processing and enhancing security in edge computing environments to foster efficient and secure IoT deployments. In recent years, the proliferation of IoT devices has led to an exponential growth in data generation at the edge. Traditional cloud-centric approaches struggle to cope with the sheer volume and velocity of data produced by these devices, often resulting in latency issues and increased operational costs. Edge computing offers a promising solution by enabling data processing and analytics closer to where data is generated, thereby reducing latency and bandwidth consumption. The first part of this paper reviews the foundational concepts of edge computing, emphasizing its architectural principles and distinguishing characteristics in contrast to centralized cloud computing. Key optimization strategies for data processing at the edge are discussed, including edge analytics, data filtering, and compression techniques. These strategies aim to minimize data transmission to centralized cloud servers, thereby enhancing operational efficiency and real-time responsiveness in IoT applications.

High Proportion of Distributed PV Reliability Planning Method Based on Big Data

The higher proportion of distributed photovoltaic and lower fossil energy integrated into the power network brings huge challenges in power supply reliability and planning. The distributed photovoltaic planning method based on big data is proposed. According to the impact of stochastic photovoltaics and loads on reliability planning, the probability model of distributed photovoltaic and load is analyzed, and the dynamic capacity–load ratios are presented based on big data. The multi-scenario generation and reduction algorithm of stochastic distributed photovoltaic and load planning is studied, and a source–load scenario matching model is proposed based on big data. According to the big data scenario of source–load, the reliability indexes and dynamic capacity–load ratio may be obtained. Finally, the IEEE 33-bus system is used as an example, and the results show that distributed photovoltaic planning methods based on big data can improve photovoltaic utilization and power supply reliability.

Advanced Test Technologies for Heterogeneous & 2.5D/3D Packaging

Automotive, AI/ML and 5G markets are demanding increasing levels of integration with respect to memory and compute density. To meet the demand for high-performance products in mission-critical, high-bandwidth applications, the semiconductor industry must look for innovative technologies to continue scaling beyond the silicon manufacturing processes. This is often referred to as “More than Moore”, and one of the most promising methods is heterogenous integration of multiple dies inside the same package (System in Package, or SiP). These dies are typically referred to as chiplets and their interfaces are optimized to operate within a multi-die package for both power, drive strength, interface mechanics and ESD protection. Heterogeneous packages at manufacturing test are faced with quality and reliability challenges due to the accessibility and visibility of traditional test approaches. Dies are assembled over the silicon interposer using very small micro-bumps which may suffer from latent defects such as voids or cracks, potentially posing a reliability risk to the assembled product. On organic substrates, rare resistive shorts, known as bridging shorts, can cause signal integrity and performance degradation. These types of latent defects, although rare, could cause a system failure over time if not screened out during final test or detected during in-field operation. To address these challenges, proteanTecs and Advantest have deployed a joint solution at manufacturing test. proteanTecs provides a monitoring solution for heterogeneous packaging based on chip telemetry, combining deep data with machine learning algorithms. proteanTecs’ patent-protected solution is comprised of small-footprint, digital-only agents for monitoring the performance of the parallel interface. These agents (monitoring IP) are designed for analytics. The fact that they have a small footprint, allows them to be placed next to each pin inside the parallel die-to-die (D2D) PHY, to achieve 100% coverage and with no impact on the signal behavior, quality, or timing. proteanTecs’ data analytics platform then trains models that are deployed at test or in the field. To achieve the highest levels of performance, proteanTecs has partnered with Advantest on the ACS Edge™ for real-time decision-making. ACS Edge™ is a high-performance, highly secure edge compute and analytics solution that enables ultra-fast algorithmic AI decision-making with millisecond latencies during test execution. It connects to the user’s test equipment via a private, high-speed encrypted link and uses the advanced container hub to run the user’s protected applications while protecting and keeping secure the user’s data and analytics. Another key piece of the ACS ecosystem is the ACS Solution Store which provides ease of access to ACS real-time data infrastructure solutions. In this paper, we will provide details for the integrated solution and silicon results. We will highlight how the solution provides new levels of visibility for real-time adaptive testing.

AD 2 VNCS: A dversarial D efense and D evice V ariation-tolerance in Memristive Crossbar-based N euromorphic C omputing S ystems

In recent years, memristive crossbar-based neuromorphic computing systems (NCS) have obtained extremely high performance in neural network acceleration. However, adversarial attacks and conductance variations of memristors bring reliability challenges to NCS design. First, adversarial attacks can fool the neural network and pose a serious threat to security critical applications. However, device variations lead to degradation of the network accuracy. In this article, we propose DFS (Deep neural network Feature importance Sampling) and BFS (Bayesian neural network Feature importance Sampling) training strategies, which consist of Bayesian Neural Network (BNN) prior setting, clustering-based loss function, and feature importance sampling techniques, to simultaneously combat device variation, white-box attack, and black-box attack challenges. Experimental results clearly demonstrate that the proposed training framework can improve the NCS reliability.

Extended Lifetime Testing of SiC CMOS Electronics at 500°C

Silicon Carbide (SiC) CMOS electronics is a promising approach for realizing high level of integration, especially for implementing digital functions, while significantly reducing the power dissipation on chip compared to NMOS-only, or JFET approaches. Gate oxide in MOSFET-based electronics pose a potential reliability challenge for operating at extreme temperatures. GE has previously developed gate oxide reliability models that predicted reliable operation at 500°C, but experimental validation had not been conducted previously. Recent testing results of testing of 500°C operation of SiC CMOS based integrated circuit over 60 days of continuous operation are presented. The results show the gate oxide surviving extended operating time of both NMOS and PMOS devices at 500°C. and thereby enabling a roadmap towards realizing more complex digital function at extreme temperature >=500°C.

Meeting the Reliability Challenges of the Clean Energy Transition

Policies intended to meet climate targets must be informed by grid reliability needs. The energy transition requires planning and coordination for the types of generation resources that maintain reliable grid operations throughout the transition to a cleaner grid, and for the specific types of resources that could enable a fully decarbonized electric system. Reliable grid operations require both resource adequacy and operating reliability (NERC 2007).

The 6G Ecosystem as Support for IoE and Private Networks: Vision, Requirements, and Challenges

The emergence of the sixth generation of cellular systems (6G) signals a transformative era and ecosystem for mobile communications, driven by demands from technologies like the internet of everything (IoE), V2X communications, and factory automation. To support this connectivity, mission-critical applications are emerging with challenging network requirements. The primary goals of 6G include providing sophisticated and high-quality services, extremely reliable and further-enhanced mobile broadband (feMBB), low-latency communication (ERLLC), long-distance and high-mobility communications (LDHMC), ultra-massive machine-type communications (umMTC), extremely low-power communications (ELPC), holographic communications, and quality of experience (QoE), grounded in incorporating massive broad-bandwidth machine-type (mBBMT), mobile broad-bandwidth and low-latency (MBBLL), and massive low-latency machine-type (mLLMT) communications. In attaining its objectives, 6G faces challenges that demand inventive solutions, incorporating AI, softwarization, cloudification, virtualization, and slicing features. Technologies like network function virtualization (NFV), network slicing, and software-defined networking (SDN) play pivotal roles in this integration, which facilitates efficient resource utilization, responsive service provisioning, expanded coverage, enhanced network reliability, increased capacity, densification, heightened availability, safety, security, and reduced energy consumption. It presents innovative network infrastructure concepts, such as resource-as-a-service (RaaS) and infrastructure-as-a-service (IaaS), featuring management and service orchestration mechanisms. This includes nomadic networks, AI-aware networking strategies, and dynamic management of diverse network resources. This paper provides an in-depth survey of the wireless evolution leading to 6G networks, addressing future issues and challenges associated with 6G technology to support V2X environments considering presenting +challenges in architecture, spectrum, air interface, reliability, availability, density, flexibility, mobility, and security.

What influences industrial enterprises’ willingness of demand response: A survey in Qinghai, China

Although demand response (DR) is an important means for solving the stability and reliability challenges of renewable energy (RE) power systems, it remains unclear how DR can be widely applied in industrial enterprises, particularly in areas with a high proportional RE system and industrial energy consumption that do not apply any DR programs. The objective of this study is to understand the influencing factors of industrial enterprises' willingness to participate (WTP) in DR in these areas, including one economic factor and three noneconomic factors. This study performs a questionnaire survey in Qinghai, and a total of 556 industrial enterprises completed valid questionnaires, these data are empirically analyzed through Structural Equation Model. The results demonstrate enterprises' DR awareness degree is the most significant predictors of the WTP in DR. Meanwhile, enterprises' electricity consumption level also had a significant impact on the WTP in DR rather than enterprises' potential DR capabilities and the expected minimum compensation. Finally, the study suggests feasible recommendations, such as increasing enterprises' ability of analyzing potential DR, directly encouraging enterprises to participate in DR, designing compensation standards and adopting differentiated DR according to enterprises’ actual DR potential.

Day-ahead energy-mix proportion for the secure operation of renewable energy-dominated power system

Advancements in various scientific fields have encouraged the development of novel tools, techniques, components, methodologies, and innovations aimed at addressing the challenges encountered in modern power systems dominated by inverter-based resources (IBRs). This paper focuses on a concept that leverages historical time-series data obtained from transmission system operators (TSOs) to enhance the secure management and operation of power systems. By employing a data-driven model, the day-ahead values of power generation and load consumption are estimated and integrated with a dynamic model of the power system for further analysis. To optimize energy generation and ensure grid stability, an energy-mix operation and reserve scheduling model is utilized. This model optimally combines different power-generating technologies, including synchronous generators (SGs), grid-following converters (GFLs), and grid-forming converters (GFMs), to meet the energy demands of the day while enhancing the overall system strength. The findings are supported by quantitative analysis utilizing variables such as frequency, power production, terminal voltages, and system non-synchronous penetration (SNSP). Simulation results demonstrate that implementing the proposed concept enables the power system under consideration to operate securely, even in the face of a 38% increase in immediate load, with a maximum SNSP ratio of 59%. These findings highlight the effectiveness of the proposed approach in addressing the reliability, system dynamics, stability, control efficiency, and security challenges posed by IBR-dominated power systems. Furthermore, it is believed that this research contributes to the ongoing efforts in decarbonization, renewable energy integration, and combating global warming by facilitating the secure and optimized operation of renewable energy-dominated power systems.

Experimental and numerical study of solder bumps impact on the underfill fluid flow under electrohydrodynamic effect

Solder bumps can increase filling time, which is one of the main challenges in electronic packaging reliability. Here, we compare the capillary flow behavior between parallel plates with and without solder bumps to examine how solder bumps affect the length of the underfill fluid flow under the effects of an electric potential. We found that the solder bumps restrained the flow length, while the electric field enhanced it. By enhancing the voltage from 0 to 1000 V, in the case without solder bumps, the flow length increased by up to 30%, and it increased by up to 25% in the case of solder bumps. To determine the optimum bump design, we selected the diameter and pitch size of the solder bumps as the independent variables. The results revealed that larger pitch sizes and smaller diameters show longer fluid flow lengths. The effect of the electric field on varying nozzle positions was also investigated. We found that the fluid flow length increased when the nozzle was between the solder bumps compared to the top of the solder bumps. According to our observations, the nozzle position is also the main factor in determining the fluid flow length compared with the bump diameter and pitch sizes for the design of the underfill packaging process. Numerical simulations were also performed to compare the experimental results, and the average discrepancy between the experimental and numerical results at various time steps for different solder bump parameters was between 5 to 10%. Our findings demonstrate the potential of using electric potential in conjunction with solder bumps to control underfill flow, which can benefit flip-chip packaging applications. Numerical methods can accurately predict underfill fluid flow with solder bumps under the electric field effect.

XGBoost and Random Forest Algorithms: An in Depth Analysis

Machine learning is playing an increasingly important role in many facets of our lives as technology develops, including forecasting weather, figuring out social media trends, and predicting prices on the world market. This significance invoked the demand of some efficient predicting models that can easily handle complex data and provide maximum accurate results. XGBoost and Random Forest are upgradable ensemble techniques used to solve regression and classification problems that have evolved and proved to be dependable and reliable machine learning challenge solvers. In this research paper, we undertake a comprehensive analysis and comparison of these two prominent machine learning algorithms. The first half of the research includes relevant overview on both of the techniques, significance and evolution of both algorithms. The latter part of this study involves a meticulous comparative analysis between Random Forest and XGBoost, scrutinizing facets such as time complexity, precision, and reliability. We examine their distinctive approaches to handling regression and classification problems while closely examining their subtle handling of training and testing datasets. A thorough quantitative evaluation using a variety of performance metrics, such as the F1-score, Recall, Precision, Mean Squared Erro

Hybrid Block Storage for Efficient Cloud Volume Service

The migration of traditional desktop and server applications to the cloud brings challenge of high performance, high reliability, and low cost to the underlying cloud storage. To satisfy the requirement, this article proposes a hybrid cloud-scale block storage system called Ursa . Trace analysis shows that the I/O patterns served by block storage have only limited locality to exploit. Therefore, instead of using solid state drives (SSDs) as a cache layer, Ursa proposes hybrid storage structure that directly stores primary replicas on SSDs and replicates backup replicas on hard disk drives (HDDs) . At the core of Ursa ’s hybrid storage design is an adaptive journal that can bridge the performance gap between primary SSDs and backup HDDs for random writes by transforming small backup writes into journal appends, which are then asynchronously replayed and merged to backup HDDs. To efficiently index the journal, we design a novel range-optimized merge-tree structure that combines a continuous range of keys into a single composite key {offset,length} . Ursa integrates the hybrid structure with designs for high reliability, scalability, and availability. Experiments show that Ursa in its hybrid mode achieves almost the same performance as in its SSD-only mode (storing all replicas on SSDs), and outperforms other block stores (Ceph and Sheepdog) even in their SSD-only mode while achieving much higher CPU efficiency (IOPS and throughput per core).

A novel fault location strategy based on Bi‐LSTM for MMC‐HVDC systems

AbstractThe integration of modular multilevel converters (MMCs) with high voltage direct current (HVDC) transmission systems is an efficient method for transporting electricity from distant renewable energy sources to demand centres. However, MMC‐HVDC systems face reliability challenges during DC overcurrent faults, often caused by component failures that can lead to HVDC network shutdowns. Consequently, a reliable fault location approach is crucial for grid protection and restoration, aiding in fault isolation and alternate power flow identification. Conventional fault location methods struggle with manual protective threshold setting, susceptibility to fault resistance and noise, and the need for communication channels, resulting in signal delays. In multi‐terminal HVDC networks, fault location becomes even more complex due to poor selectivity and sensitivity in traditional schemes. This study proposes a robust fault location approach based on bidirectional long short‐term memory (bi‐LSTM). The method offers a simplified decision‐making model with low computational requirements, utilizing fault features from one end of the network, eliminating the need for a communication channel. Remarkably, this approach achieves high fault location accuracy, even with varying fault types, resistances, and noise levels, as demonstrated by an MSE of 0.006 and a percentage error below 1% in simulations conducted using a real‐time simulator with MATLAB/Simulink.

A Machine Learning Based Reliability Analysis of Negative Bias Temperature Instability (NBTI) Compliant Design for Ultra Large Scale Digital Integrated Circuit

NBTI is a key reliability challenge in nanoscale digital design, and it is vital to address it throughout the exploration of design space at high levels of abstraction in order to improve reliability. A prediction model of aging that is adequate for these levels ought to be faster. In addition to this, the model should be able to forecast the recently discovered stochastic consequences of growing older. The purpose of this study is to offer a model that is based on machine learning (ML) and can predict aging effects. After obtaining a training set of sufficient size using Synopsis HSPICE (MOSFET Reliability, MOSRA) in the beginning, the machine-learning-based model is then trained and built in order to forecast the aging statistical features. Evaluation is done on a number of machine learning techniques, including Adaptive Neuro-Fuzzy Inference System (ANFIS), K-Nearest Neighbor (KNN), Support Vector Machine (SVM) and Random Forest (RF). The findings indicate that ANFIS algorithms are very effective in the process of age prediction. The proposed technique shows that the aging prediction runtime is reduced by more than 99% when compared to the MOSRA-based approach, and accurate predictions of the statistical properties of aging are obtained with an accuracy of more than 99% on complementary metal oxide semiconductor (CMOS) and metal gate/high-K (MGK) circuits at the 22nm technology node.

Artificial Intelligence in Digital Pathology for Bladder Cancer: Hype or Hope? A Systematic Review.

Bladder cancer (BC) diagnosis and prediction of prognosis are hindered by subjective pathological evaluation, which may cause misdiagnosis and under-/over-treatment. Computational pathology (CPATH) can identify clinical outcome predictors, offering an objective approach to improve prognosis. However, a systematic review of CPATH in BC literature is lacking. Therefore, we present a comprehensive overview of studies that used CPATH in BC, analyzing 33 out of 2285 identified studies. Most studies analyzed regions of interest to distinguish normal versus tumor tissue and identify tumor grade/stage and tissue types (e.g., urothelium, stroma, and muscle). The cell's nuclear area, shape irregularity, and roundness were the most promising markers to predict recurrence and survival based on selected regions of interest, with >80% accuracy. CPATH identified molecular subtypes by detecting features, e.g., papillary structures, hyperchromatic, and pleomorphic nuclei. Combining clinicopathological and image-derived features improved recurrence and survival prediction. However, due to the lack of outcome interpretability and independent test datasets, robustness and clinical applicability could not be ensured. The current literature demonstrates that CPATH holds the potential to improve BC diagnosis and prediction of prognosis. However, more robust, interpretable, accurate models and larger datasets-representative of clinical scenarios-are needed to address artificial intelligence's reliability, robustness, and black box challenge.

Reliability oriented techno- economic assessment of fast charging stations with photovoltaic and battery systems in paired distribution & urban network

Installing fast charging electric vehicle stations (FCEVS) is crucial for increasing public acceptance of electric vehicle (EV) adoption. The enormous energy demands of FCEVS, as well as the inclusion of renewable energy resources (RES) into utility grids, may have a significant influence on system reliability and pose challenges owing to uncertainty. Therefore, a new reliability-oriented analysis-based two-stage planning framework for optimal deployment of solar photovoltaic (SPV) powered FCEVS and battery energy storage systems (BESS) in coupled transportation and distribution networks is proposed in this article, while encapsulating the complex interactions between EVs, SPVs, and BESS. A novel composite metric, the average voltage deviation reliability index (AVDRI), has been developed for the deployment of FCEVS infrastructure considering the goals and limitations of both distribution and transportation networks. Models such as the demand for charging electric vehicles, the reliability evaluation system, and the economic analysis model are offered to provide essential inputs for the study of distribution network reliability and profitability. In the first stage, AVDRI, active power loss, captured EV flow, and capital, along with operating and maintenance costs of the FEVCS and SPV system, are all taken into consideration as objectives when optimizing the placement and sizing of SPV assisted FCEVS using Multi Objective Particle Swarm Optimization (MOPSO) and a fuzzy satisfaction-based hybrid technique. Then, the optimal resource sizing, such as the count of charging spaces, the capacity of the SPV system at each optimal FCEVS, and captured EV flow, are estimated. In the second step, a bi-section approach is used to compute the capacity of the BESS and the extra SPV rating necessary to power it at each optimal FCEVS location, taking into account the irradiance profile for all conceivable seasons within a planning year. A 25-node transportation network coupled to an IEEE 33 bus distribution system is used to validate the proposed methodology under a range of seasonal scenarios.

Matroidal connectivity and conditional matroidal connectivity of star graphs

As modern parallel computer systems continue to grow in scale, the occurrence of failures becomes inevitable, which places severe reliability challenges on the systems' underlying interconnection network. Recently, two novel indicators, named the matroidal connectivity and conditional matroidal connectivity, were proposed to better evaluate the reliability of interconnection networks. These indicators enable flexible restrictions on faulty edges in each dimension and have proven to be promising in improving the edge fault-tolerance of interconnection networks. In this paper, we aim to apply these indicators to the n-dimensional star graph Sn, which is a famous interconnection network with high symmetry and recursive hierarchical structure. We partition the faulty edge set F⊆E(Sn) into n−1 parts according to each faulty edge's dimension. By sorting the cardinality of these parts and exerting the restriction that the (n−i)-th largest part's cardinality does not exceed i!−1 for 1≤i≤n−1, we prove that Sn−F is always connected where |F|≤∑i=1n−1(i!−1). Then we determine the exact values of matroidal connectivity and conditional matroidal connectivity of star graphs. Moreover, we make comparisons to demonstrate our results outperform other similar works in terms of edge fault-tolerance.