Two cost‐effective fault‐tolerant multistage interconnection networks

Abstract

This paper introduces two concepts for building fault‐tolerant multistage interconnection networks (MIN). The first one is to add multiplexers/demultiplexers at the front/back end of a pre‐constructed network which may be a uniquepath MIN or a fault‐tolerant one. The second one is to integrate multiple fault‐tolerant MINs to form a larger and even more reliable one. Based upon the first concept, we propose a new fault‐tolerant MIN, the Path Sharing Network, which is derived by adding multiplexers/demultiplexers at the front/back end of the Banyan MIN. The Recursive Path Sharing Network is also proposed by integrating two fault‐tolerant MINs based upon the second concept. The reliability and cost‐effectiveness of the proposed networks are analyzed and compared with other fault‐tolerant MINs. Surprisingly, the proposed networks are better than EGN‐4 which was shown to be better than ESC, 3‐Rep and INDRA.     

Reliability analysis of shuffle-exchange network systems

Shuffle-exchange networks (SENs) have been widely considered as practical interconnection systems due to their size of its switching elements (SEs) and uncomplicated configuration. SEN is a network among a large class of topologically equivalent multistage interconnection networks (MINs) that includes omega, indirect binary n-cube, baseline, and generalized cube. In this paper, SEN with additional stages that provide more redundant paths are analyzed. A common network topology with a 2×2 basic building block in a SEN and its variants in terms of extra-stages is investigated. As an illustration, three types of SENs are compared: SEN, SEN with an additional stage (SEN+), and SEN with two additional stages (SEN+2). Finally, three measures of reliability: terminal, broadcast, and network reliability for the three SEN systems are analyzed.     

Hardware implementation of fault‐tolerance in dual computer systems

In this paper, we propose an architectural design for a dual computer system (DCS) that operates in real‐time with the fault‐tolerance implemented purely by hardware. We have a novel design allowing the implementation of hardware that performs the following key services: the determination of fault type (temporary or permanent) and the localization of the faulty computer without using self‐testing techniques and diagnosis routines. We also propose a non‐trivial sequence of services for fault‐tolerance in which the determination of the fault type and the recovery of computational processes after a temporary fault are realized before fault localization. Our design has several benefits: the designed hardware shortens the recovery point time period; the proposed non‐trivial sequence of fault‐tolerant services reduces (to two) the number of logical segments that should be re‐run to recover the computational processes; and the determination of the fault type allows eliminating only the computer with a permanent fault. These contributions bring both an increase in system performance and an increase in the degree of system reliability. Copyright © 2009 John Wiley & Sons, Ltd.     

A closed form formula for the normalized throughput of unbuffered multistage interconnection networks

Abstract

The normalized throughput of an unbuffered multistage interconnection network (MIN) under the uniform traffic model can be computed iteratively [6] or recursively [7]. However, the asymptotic performance of such networks as the number of stages increases cannot be determined using these procedures. A closed form formula for the normalized throughput of unbuffered MINs is presented in this paper. The new formula is proved by mathematical induction on the number of stages. The major advantage of the closed form formula is that the asymptotic performance of unbuffered square MINs can be easily determined. The formula is generalized to unbuffered MINs in the presence of nonidentical input rates and non‐square switch size.     

Development of a compliant legged quadruped robot

This paper presents the detailed steps for design and development of a compliant legged fault tolerant quadruped robot where each leg has two links and two motorized revolute joints for locomotion. The body and upper links of legs are rigid whereas the lower link of each leg is compliant. Amble gait is demonstrated on the developed robot. Safety and reliability are the most critical issues for the quadruped robot. During the failure of any joint, performance of quadruped robot is affected. In this paper, locked joint failure is also discussed. Strategies for fault tolerant control of the quadruped are developed and experimentally validated. The developed robot can be used for various hardware-in-the-loop controller prototyping such as reconfiguration, fault tolerant control, and posture control, etc. pertaining to quadruped robots.     

Fault impact and fault tolerance in multiprocessor interconnection networks

Growing complexity of parallel machines coupled with increasing chip densities escalates the need for fault tolerance and recovery in these systems. In pursuit of the goal of fault-tolerant multiprocessors, many techniques have been proposed. Since methods for designing fault-tolerant processors and memories are relatively mature, the techniques considered in this paper focus on the interconnection network (ICN) linking the processors. The impact of faults on non-fault-tolerant ICNs is contrasted with that in fault-tolerant networks. Fault tolerance in ICNs is addressed at two levels, inter-node or switch level and system level. Inter-node or switch level pertains to data and control integrity and system level deals with maintaining network connectivity and adequate performance levels in the presence of faults. Fault-tolerant schemes at the switching element level warrant some form of concurrent error detection such as the use of codes usually combined with a full handshake protocol. Space–time trade-offs involved in the use of various codes and protocols are investigated. At the system level, several augmented multi-stage switching ICNs, tree and ring networks are studied. The combined provision for fault tolerance together with improved performance in the non-fault condition is emphasized. Finally, strategies for network reconfiguration and rerouting after system failure are presented.     

Fault‐tolerant procedures for redundant computer systems

Real‐time computer systems deployed in life‐critical control applications must be designed to meet stringent reliability specifications. The minimum acceptable degree of reliability for systems of this type is ‘7 nines’, which is not generally achieved. This paper aims at contributing to the achievement of that degree of reliability. To this end, this paper proposes a classification scheme of the fault‐tolerant procedures for redundant computer systems (RCSs). The proposed classification scheme is developed on the basis of the number of counteracted fault types. Table I is created to relate the characteristics of the RCSs to the characteristics of the fault‐tolerant procedures. A selection algorithm is proposed, which allows designers to select the optimal type of fault‐tolerant procedures according to the system characteristics and capabilities. The fault‐tolerant procedure, which is selected by this algorithm, provides the required degree of reliability for a given RCS. According to the proposed graphical model only a part of the fault‐tolerant procedure is executed depending on the absence or presence (type and sort) of faults. The proposed methods allow designers to counteract Byzantine and non‐Byzantine fault types during degradation of RCSs from N to 3, and only the non‐Byzantine fault type during degradation from 3 to 1 with optimal checkpoint time period. Copyright © 2008 John Wiley & Sons, Ltd.     

Sparse convolutional autoencoder-based fault location for drive circuits in nuclear reactors

Drive circuit is a critical part of instrumentation and control systems in nuclear reactors, and its performance directly influences the operation of nuclear reactors. However, comparing with the open circuit IGBT faults, soft faults caused by the degradation of electronic components present much slighter fluctuations to the performance of drive circuits. If the two fault modes co-exist, traditional fault diagnosis models are prone to misclassify soft faults as the normal condition. To improve the accuracy of fault diagnosis of drive circuits, it necessitates to accurately locate the faults of drive circuits, while effectively extracting the distinguishable fault features is one of the critical factors for fault location. In this article, a fault location method combining the empirical modal decomposition (EMD) algorithm and sparse convolutional autoencoder (SCAE) is proposed. The EMD algorithm is applied to decompose the three-phase current signals of drive circuits. An SCAE-based feature extractor is constructed to capture high-dimensional and sparse fault feature data with the aid of the powerful feature autonomic extraction capability of deep learning. A deep classifier is designed to locate faults in the driver circuit. A fault simulation model of the drive circuit is developed and the monitor data is collected. The effectiveness of the proposed method is validated via a real case of drive circuit in nuclear reactors.     

Robust recurrent neural network modeling for software fault detection and correction prediction

Software fault detection and correction processes are related although different, and they should be studied together. A practical approach is to apply software reliability growth models to model fault detection, and fault correction process is assumed to be a delayed process. On the other hand, the artificial neural networks model, as a data-driven approach, tries to model these two processes together with no assumptions. Specifically, feedforward backpropagation networks have shown their advantages over analytical models in fault number predictions. In this paper, the following approach is explored. First, recurrent neural networks are applied to model these two processes together. Within this framework, a systematic networks configuration approach is developed with genetic algorithm according to the prediction performance. In order to provide robust predictions, an extra factor characterizing the dispersion of prediction repetitions is incorporated into the performance function. Comparisons with feedforward neural networks and analytical models are developed with respect to a real data set.     

Diagnosis Modelling for Dependability Assessment of Fault‐Tolerant Systems Based on Stochastic Activity Networks

The growing demand for safety, reliability, availability and maintainability in modern technological systems has led these systems to become more and more complex. To improve their dependability, many features and subsystems are employed like the diagnosis system, control system, backup systems, and so on. These subsystems have all their own dynamic, reliability and performances and interact with each other in order to provide a dependable and fault‐tolerant system. This makes the dependability analysis and assessment very difficult. This paper proposes a method to completely model the diagnosis procedure in fault‐tolerant systems using stochastic activity networks. Combined with Monte Carlo simulation, this will allow the dependability assessment by including the diagnosis parameters and performances explicitly. Copyright © 2014 John Wiley & Sons, Ltd.     

Design of fault simulator

Fault simulator is proposed to understand and evaluate all possible fault propagation scenarios, which is an essential part of safety design and operation design and support of chemical/production processes. Process models are constructed and integrated with fault models, which are formulated in qualitative manner using fault semantic networks (FSN). Trend analysis techniques are used to map real time and simulation quantitative data into qualitative fault models for better decision support and tuning of FSN. The design of the proposed fault simulator is described and applied on experimental plant (G-Plant) to diagnose several fault scenarios. The proposed fault simulator will enable industrial plants to specify and validate safety requirements as part of safety system design as well as to support recovery and shutdown operation and disaster management.     

Optical binary de Bruijn networks for massively parallel computing: design methodology and feasibility study

The interconnection network structure can be the deciding and limiting factor in the cost and the performance of parallel computers. One of the most popular point-to-point interconnection networks for parallel computers today is the hypercube. The regularity, logarithmic diameter, symmetry, high connectivity, fault tolerance, simple routing, and reconfigurability (easy embedding of other network topologies) of the hypercube make it a very attractive choice for parallel computers. Unfortunately the hypercube possesses a major drawback, which is the complexity of its node structure: the number of links per node increases as the network grows in size. As an alternative to the hypercube, the binary de Bruijn (BdB) network has recently received much attention. The BdB not only provides a logarithmic diameter, fault tolerance, and simple routing but also requires fewer links than the hypercube for the same network size. Additionally, a major advantage of the BdB network is a constant node degree: the number of edges per node is independent of the network size. This makes it very desirable for large-scale parallel systems. However, because of its asymmetrical nature and global connectivity, it poses a major challenge for VLSI technology. Optics, owing to its three-dimensional and globalconnectivity nature, seems to be very suitable for implementing BdB networks. We present an implementation methodology for optical BdB networks. The distinctive feature of the proposed implementation methodology is partitionability of the network into a few primitive operations that can be implemented efficiently. We further show feasibility of the presented design methodology by proposing an optical implementation of the BdB network.     

FAILURE ANALYSIS FOR AN AIRBAG INFLATOR BY PETRI NETS

Petri nets are useful for modelling a variety of asynchronous and concurrent systems, such as automated manufacturing, computer fault tolerant systems, and communication networks. This study employs an airbag inflator system as an example to demonstrate a Petri net approach to failure analysis. This paper uses Petri nets to study minimum cut sets finding, marking transfer, and dynamic behaviour of system failure. For Petri net models incorporating sensors, fault detection and higher-level fault avoidance is dealt with. Compared with fault trees that present only static logic relations between events, Petri nets indeed offer more capabilities in the scope of failure analysis. © 1997 John Wiley & Sons, Ltd.     

Feasibility study of a scalable optical interconnection network for massively parallel processing systems

The theoretical modeling of a novel topology for scalable optical interconnection networks, called optical multimesh hypercube (OMMH), is developed to predict size, bit rate, bit-error rate, power budget, noise, efficiency, interconnect distance, pixel density, and misalignment sensitivity. The numerical predictions are validated with experimental data from commercially available products to assess the effects of various thermal, system, and geometric parameters on the behavior of the sample model. OMMH is a scalable network architecture that combines positive features of the hypercube (small diameter, regular, symmetric, and fault tolerant) and the mesh (constant node degree and size scalability). The OMMH is implemented by a free-space imaging system incorporated with a space-invariant hologram for the hypercube links and fiber optics to provide the mesh connectivity. The results of this work show that the free-space links can operate at 368 Mbits/s and the fiber-based links at 228 Mbits/s for a bit-error rate of 10(-17) per channel. The predicted system size for 32 nodes in the OMMH is 4.16 mm × 4.16 mm × 3.38 cm. Using 16-bit, bit-parallel transmission per node, the system can operate at a bit rate of up to 5.88 Gbits/s for a size of 1.04 cm × 1.04 cm × 3.38 cm.     

An enhanced chiller FDD strategy based on the combination of the LSSVR-DE model and EWMA control charts

Development of the fault detection and diagnosis (FDD) for chiller systems is very important for improving the equipment reliability and saving energy consumption. The results of FDD performance are strongly dependent on the accuracy of chiller models. Since the accuracy of the chiller models depends on the indefinite model parameters which are normally chosen by experiments or experiences, an accurate chiller model is difficult to build. Therefore, optimization of model parameters is very useful to increase the accuracy of chiller models. This paper presents a new FDD strategy for centrifugal chillers of building air-conditioning systems, which is the combination between the nonlinear least squares support vector regression (LSSVR) based on the differential evolution (DE) algorithm and the exponentially weighted moving average (EWMA) control charts. In this strategy, the nonlinear LSSVR, which is a reformulation of SVR model with better generalization performances, is adopted to develop the reference feature parameter models in a typical non-linear chiller system. The DE algorithm which is a real-coding optimal algorithm with powerful global searching capacity is employed to enhance the accuracy of LSSVR models. The exponentially weighted moving average (EWMA) control charts are introduced to improve the fault detection capability as well as to reduce the Type II errors in a t-statistics-based way. Six typical faults of the chiller from the real-time experimental data of ASHRAE RP-1043 project are chosen to validate proposed FDD methods. Comprehensive comparisons between the proposed method and two similarly previous studies are performed. The comparison results show that the proposed method has achieved significant improvement in accuracy and reliability, especially at low severity levels. The proposed DE-LSSVR-EWMA strategy is robust for fault detection and diagnosis in centrifugal chiller systems.     

A study on group decision-making based fault multi-symptom-domain consensus diagnosis

In the field of fault diagnosis for rotating machines, the conventional methods or the neural network based methods are mainly single symptom domain based methods, and the diagnosis accuracy of which is not always satisfactory. In this paper, in order to utilize multiple symptom domains to improve the diagnosis accuracy, an idea of fault multi-symptom-domain consensus diagnosis is developed. From the point of view of the group decision-making, two particular multi-symptom-domain diagnosis strategies are proposed. The proposed strategies use BP (Back-Propagation) neural networks as diagnosis models in various symptom domains, and then combine the outputs of these networks by two combination schemes, which are based on Dempster–Shafer evidence theory and fuzzy integral theory, respectively. Finally, a case study pertaining to the fault diagnosis for rotor-bearing systems is given in detail, and the results show that the proposed diagnosis strategies are feasible and more efficient than conventional stacked-vector methods.     

An analytical method for reliability analysis of hardware‐software co‐design system

Hardware‐software co‐design systems abound in diverse modern application areas such as automobile control, telecommunications, big data processing, and cloud computing. Existing works on reliability modeling of the co‐design systems have mostly assumed that hardware and software subsystems behave independently of each other. However, these two subsystems may have significant interactions in practice. In this paper, an analytical approach based on paths and integrals is proposed to analyze reliability of nonrepairable hardware‐software co‐design systems considering interactions between hardware and software during the system performance degradation and failure process. The proposed approach is verified using the Markov‐based method. As demonstrated by case studies on systems without and with warm standby sparing, the proposed approach is applicable to arbitrary types of time‐to‐failure or degradation distributions. Effects of different transition and fault detection/recovery parameters on system performance are also investigated through examples.     

Application of a max–min ant system to joint layout and size optimization of pipe networks

The application of a max–min ant algorithm to the layout and size optimization of pipe networks is described in this paper. The formulation conventionally used for the pipe size optimization of networks with fixed layout is extended to account for the layout determination of the networks. This is achieved by including new constraints regarding the reliability of the network and modifying some of the constraints of the optimization problem. A deterministic concept of reliability is used in which the number of independent paths from source nodes to each of the demand nodes is considered as a measure of reliability. The method starts with a predefined layout which includes all possible links. The method is capable of designing the layout and pipe sizes of water distribution networks of predefined reliability including tree-like and looped networks. It is also shown that a layout optimization of a network followed by size optimization does not lead to an optimal or a near-optimal solution. This emphasizes the need for simultaneous layout and size optimization of networks if an optimal or near-optimal solution is desired. The performance of the method for layout and pipe size optimization of pipe networks is tested against two benchmark examples in the literature and the results are presented. The first example is considered to show the necessity of joint layout and size optimization even for the simple tree networks while the second example is considered to illustrate the efficiency of the proposed method for layout and size optimization of real-world networks with different levels of reliability.     

Fault Diagnosis of Power Electronic Circuits Based on Adaptive Simulated Annealing Particle Swarm Optimization

In the field of energy conversion, the increasing attention on power electronic equipment is fault detection and diagnosis. A power electronic circuit is an essential part of a power electronic system. The state of its internal components affects the performance of the system. The stability and reliability of an energy system can be improved by studying the fault diagnosis of power electronic circuits. Therefore, an algorithm based on adaptive simulated annealing particle swarm optimization (ASAPSO) was used in the present study to optimize a backpropagation (BP) neural network employed for the online fault diagnosis of a power electronic circuit. We built a circuit simulation model in MATLAB to obtain its DC output voltage. Using Fourier analysis, we extracted fault features. These were normalized as training samples and input to an unoptimized BP neural network and BP neural networks optimized by particle swarm optimization (PSO) and the ASAPSO algorithm. The accuracy of fault diagnosis was compared for the three networks. The simulation results demonstrate that a BP neural network optimized with the ASAPSO algorithm has higher fault diagnosis accuracy, better reliability, and adaptability and can more effectively diagnose and locate faults in power electronic circuits.     

基于深度置信网络与信息融合的齿轮故障诊断方法

针对齿轮在复杂运行工况下故障特征提取困难,传统故障诊断方法的识别精度易受人工提取特征的影响,以及单传感器获取信息不全面等问题,提出基于深度置信网络(DBN)与信息融合的齿轮故障诊断方法。通过多传感器信息融合技术对每个传感器采集的振动信号进行数据层融合;利用DBN进行自适应特征提取从而实现故障分类。为了避免因人为选择DBN结构参数,导致模型识别精度下降的问题,利用改进的混合蛙跳算法(ISFLA)对DBN结构参数进行优化。试验表明,与BP神经网络、未经优化的DBN以及单传感器故障诊断相比,该研究提出的信息融合及优化方法具有更高的故障识别精度。