Graceful degradation in algorithm-based fault tolerantmultiprocessor systems

Algorithm-based fault tolerance (ABFT) is a technique which improves the reliability of a multiprocessor system by providing concurrent error detection and fault location capability to it. It encodes data at the system level and modifies the algorithm to operate on the encoded data in order to expose both transient and permanent faults in any processor. Work done till now in this area takes care of only the fault detection and location part of the problem. However, if spare processors are not available, then after a faulty processor has been located, the work initially assigned to it has to be mapped to some nonfaulty processors in the system in such a way that the fault tolerance capability of the system is still maintained with as small a degradation in performance as possible. In this paper, we propose an integrated deterministic solution to the above problem which combines concurrent error detection and fault location with graceful degradation. There exists no previous deterministic ABFT method for the design of general t-fault locating systems, even for the case of t=1. We propose a general method for designing one-fault locating/s-fault detecting systems. We use an extended model for representing ABFT systems. This model considers the processors computing the checks to be a part of the ABFT system, so that faults in the check computing processors can also be detected and located using a simple diagnosis algorithm, and the checks can be mapped to other nonfaulty processors in the system     

Partitioned encoding schemes for algorithm-based fault tolerance inmassively parallel systems

Considers the applicability of algorithm based fault tolerance (ABET) to massively parallel scientific computation. Existing ABET schemes can provide effective fault tolerance at a low cost For computation on matrices of moderate size; however, the methods do not scale well to floating-point operations on large systems. This short note proposes the use of a partitioned linear encoding scheme to provide scalability. Matrix algorithms employing this scheme are presented and compared to current ABET schemes. It is shown that the partitioned scheme provides scalable linear codes with improved numerical properties with only a small increase in hardware and time overhead     

Design of algorithm-based fault-tolerant multiprocessor systems forconcurrent error detection and fault diagnosis

Algorithm-based fault tolerance (ABPT) is a low-overhead system-level concurrent error detection and fault location scheme for multiprocessor systems. We present new methods for the design of ABFT systems. Our design procedure is applicable to a wide range of systems in which processors share data elements. A feature of our design approach is that the type of checks to be used in the final system can be controlled by the system designer. We also present some new bounds on the number of checks needed in ABFT system design     

CDS calibration under an extended JDCEV model

ABSTRACT

We propose a new methodology for the calibration of a hybrid credit-equity model to credit default swap (CDS) spreads and survival probabilities. We consider an extended Jump to Default Constant Elasticity of Variance model incorporating stochastic and possibly negative interest rates. Our approach is based on a perturbation technique that provides an explicit asymptotic expansion of the CDS spreads. The robustness and efficiency of the method is confirmed by several calibration tests on real market data.     

An analysis of algorithm-based fault tolerance techniques

We introduce a unified checksum scheme for the LU decomposition, Gaussian elimination with pairwise pivoting, and the QR decomposition. The purpose is to detect and locate a transient error during a systolic array computation. We show how to represent the error as a rank-one perturbation to the original data, so that we need not worry when the error occurred. Finally, we perform a floating point error analysis to determine the effects of rounding errors on the checksum scheme.     

Optimal input design for multibody systems by using an extended adjoint approach

We present a method for optimizing inputs of multibody systems for a subsequently performed parameter identification. Herein, optimality with respect to identifiability is attained by maximizing the information content in measurements described by the Fisher information matrix. For solving the resulting optimization problem, the adjoint system of the sensitivity differential equation system is employed. The proposed approach combines these two well-established methods and can be applied to multibody systems in a systematic, automated manner. Furthermore, additional optimization goals can be added and used to find inputs satisfying, for example, end conditions or state constraints.     

A hybrid fault prediction method for control systems based on extended state observer and hidden Markov model

Fault diagnosis and prediction for complex control systems rely either on the collection of rich data for training neural networks or on the system models and prior knowledge of faults. These methods are difficult to apply directly in complex integrated systems due to the large uncertainties in practical scenarios. A new fault diagnosis and prediction technique that is based on extended state observer (ESO) and a hidden Markov model (HMM) for control systems is proposed in this paper. Real-time and predictive information that is obtained by ESO of the active disturbance rejection control (ADRC) is utilized to improve the HMM method for the fault prediction of control systems with large uncertain disturbances. The proposed approach realizes a high recognition rate with a small demand for data, and the dependence on the system model is weak without prior knowledge of faults. Fault prediction of the control system output can be realized without additional sensors. The proposed solution is evaluated in simulations of an asynchronous servo motor control system against the traditional control method and the ADRC control. The results indicate that the proposed method performs well in fault prediction and outperforms the traditional method in terms of control when disturbances and failures occur.     

An extended model of design process of lean production systems by means of process variables

In this paper, we present an axiomatic modeling of lean production system design, using process variables (PVs). So far, we had developed a model for conceptual design of lean production systems by means of FR–DP relationships, the key characteristics of axiomatic design (AD) methodology, appeared in the proceedings of Second International Conference of Axiomatic Design. Albeit the model in question was thorough enough to be applied in various cases, its embedded abstract principles hamper straightforward applications and the required resources, tools, and techniques are not clarified. In AD terms, it lacks PVs created by mapping the design parameters (DPs) to the process domain to clarify the means that produce the specified DPs. Owing to the difficulties involved in the definition of PVs for manufacturing systems, there is few works in this area. This paper is an attempt to introduce PVs in production system design. When we are developing a product (i.e. a part), we can simply interpret the set of PVs as the process design. In the case of a production system we interpret PVs as the tools, methods, and resources, required for implementing a lean production system. In this paper, according to AD methodology, we have developed a hierarchical structure to model the design process of a lean production system, composed of FRs, DPs, and PVs. Serving as an efficient guideline for the design process and clarifying the required tools, methods, and resources, this structure is general enough to be applied for different cases.     

Conceptual design of remote monitoring and fault diagnosis systems

Condition monitoring and fault diagnosis are of fundamental importance for many industrial systems. In the last decade, substantial research efforts have been made on the surveillance and diagnosis systems for different types of equipment, with the approach of integrating information technologies and intelligent computing methods. This paper presents the conceptual design of a distributed information system of condition monitoring and fault diagnosis for a growing number of gas turbine-based power generation systems. Each individual information system that monitors a specific gas turbine system, locally deployed in a power plant, is linked to another information system, deployed at the manufacturer's site, which oversees all the gas turbine systems in parallel. The systems are constructed on the basis of COM components, which are conceptually separated into three tiers. Subsequently, this paper proceeds to present a generic business domain model with components encapsulating physical entities of interest. Finally, this paper addresses the interactions among components, which considerably affect the performance of the system including efficiency and effectiveness. It has been identified that both asynchronous and synchronous communication mechanisms are required for exchanging information for different scenarios. COM+ services are also required for supporting object pooling, transaction coordination, and security control.     

Integrated design of fault detection systems in time-frequencydomain

Problems related to the integrated design of robust fault detection (FD) systems are studied. First, it is revealed that due to the time window introduced to realize the 2-norm based evaluation function, an optimal design of a FD system with the 2-norm based evaluation function may not ensure the expected optimal performance when the system is realized in real applications. To solve this problem, an integrated design method of FD systems using the absolute value of residual signal as evaluation function is then proposed. It leads to a residual generator which is much easier to be realized. Different from the usual 2-norm based approaches whose mathematical basis is the relationship between the energy of the output and input signals of a dynamic system, a relationship between the instant power of the output signal and the energy of the past input signal of a dynamic system is established and further used for FD system design. Another new kind of evaluation function based on the absolute value of wavelet transform of residual signal and the corresponding integrated design approach for FD systems are further proposed     

EkmEx - an extended framework for labeling an unlabeled fault dataset

Multimedia Tools and Applications - Software fault prediction (SFP) is a quality assurance process that identifies if certain modules are fault-prone (FP) or not-fault-prone (NFP). Hence, it...     

Recursive identification algorithms to design fault detection systems

This paper deals with the recursive identification of the parity space based fault detection systems. We propose two such algorithms that update the eigenstructure after every new measurement with significantly less computational cost. Its immediate application is in the design of adaptive parity space based residual generators. The method improves the fault detection performance against uncertain changes, especially the frequent shifts in operating points, or parameter variations. The algorithms are compared with the existing techniques and applied to the hybrid simulation platform of continuously stirred tank heater.     

Synthesis of algorithm-based fault-tolerant systems from dependencegraphs

Algorithm-based fault tolerance (ABFT) is a method for improving the reliability of parallel architectures used for computation-intensive tasks. A two-stage approach to the synthesis of ABFT systems is proposed. In the first stage, a system-level code is chosen to encode the data used in the algorithm. In the second stage, the optimal architecture to implement the scheme is chosen using dependence graphs. Dependence graphs are a graph-theoretic form of algorithm representation. The authors demonstrate that not all architectures are ideal for the implementation of a particular ABFT scheme. They propose new measures to characterize the fault tolerance capability of a system to better exploit the proposed synthesis method. Dependence graphs can also be used for the synthesis of ABFT schemes for non-linear problems. An example of a fault-tolerant median filter is provided to illustrate their utility for such problems     

基于扩展滤波器的非线性系统迭代学习故障诊断算法

针对一类存在执行器和传感器故障的非线性系统,提出基于滤波器的故障重构方法。为了使算法同时适用于状态和输出端,通过由系统方程构造新状态方程对系统作扩展变换,将原系统输出端非线性和故障转换到扩展系统的状态方程,在此基础上设计故障诊断滤波器,采用迭代学习调节算法更新虚拟故障使之逼近实际故障。该算法可以检测和估计系统故障,并且对不同类型故障具有一定的适应性。在单关节机器人模型上进行仿真实验,实验结果验证了所提出算法的可行性和有效性。     

片上网络故障模型及容错设计方法合理性分析

结合片上网络的硬件实现结构和先进工艺集成电路物理效应对片上网络存在的故障行为进行分析。进而对现的多种故障模型合理性进行了深入探讨,指出了其存在的缺陷及问题。在此基础上从故障产生的成因出发,对面向片上的容错多种设计方法进行了检讨和反思。通过本文的分析可以看出,需要在结合其自身硬件结构的基础上深入分析物理效应所引起的故障行为,才能有针对性的设计出真正高效、可靠的片上网络。     

Relaxed observer-based controller design method of discrete-time multiplicative noised LPV systems via an extended projective lemma

ABSTRACT

The aim of this paper is to discuss a relaxed observer-based control problem of multiplicative noised Linear Parameter Varying (LPV) systems with unmeasurable states. For discussing the problem, positive definite matrices without neglecting any element are employed to construct a novel parameter-dependent Lyapunov function. Based on the Lyapunov function, some relaxed sufficient conditions are derived to hold the freedom in searching feasible solutions. Furthermore, an extended projective lemma is developed to convert those conditions into Linear Matrix Inequality (LMI) form. Solving these LMI conditions, controller gain and observer gain can be simultaneously obtained in one-step procedure via using convex optimization algorithm. Therefore, an observer-based controller can be established to guarantee the asymptotical stability of the multiplicative noised LPV systems in the sense of mean square. At last, two numerical examples are used to show effectiveness and applicability of the proposed design method.     

Subspace aided data-driven design of robust fault detection and isolation systems

This paper deals with subspace method aided data-driven design of robust fault detection and isolation systems. The basic idea is to identify a primary form of residual generators directly from test data and then make use of performance indices to make uniform the design of different type robust residuals. Four algorithms are proposed to design fault detection, isolation and identification residual generators. Each of them can achieve robustness to unknown inputs and sensitivity to sensor or actuator faults. Their existence conditions and multi-fault identification problem are briefly analyzed as well and the application of the method proposed is illustrated by a simulation study on the vehicle lateral dynamic system.     

Subspace method aided data-driven design of fault detection and isolation systems

This paper deals with data-driven design of fault detection and isolation (FDI) systems. The basic idea is to identify a primary form of residual generators, instead of the process model, directly from test data and, based on it, to design advanced FDI systems. The proposed method can be applied for the parity space and observer based detection and isolation of sensor and actuator faults as well as the construction of soft-sensors. The application of the proposed method is illustrated by a simulation study on the Tennessee Eastman process.     

Performance of an extended certainty-weighted detection model

This paper introduces a certainty-weighted detection system (CWDS) based on distributed decision makers that can classify a binary phenomenon as true, false, or uncertain. The CWDS is composed of two main blocks: the definite decision block (DDB), which provides a decision regarding the presence or absence of the phenomenon, and the uncertainty measure block (UMB) that provides a measure of uncertainty. The final decision, which may be definite (true or false) or uncertain, depends on characteristic parameters that define the region of uncertainty (RU/sub i/ and /spl alpha/) used by piecewise linear certainty functions in the DDB and in the UMB. The Bayes cost analysis is extended to include the cost of uncertain classifications and cost of errors. A cost function is used to compare the CWDS to decision structures based on the Dempster-Shafer theory and fuzzy logic that also provide uncertain decisions. The CWDS performs similarly to a classical Bayes detection system when no uncertain classifications are provided. By changing the parameters RU/sub i/ and /spl alpha/, the CWDS can also be adjusted to perform similarly to the Dempster-Shafer and fuzzy structures. The differences between these approaches are mainly in their characterization of uncertainty, and they can reduce the total costs below that of the Bayesian model if the cost of uncertain classifications is sufficiently smaller than the cost of errors. The performance of the CWDS was less sensitive to changes in the ratio of the costs of uncertain decisions to the cost of incorrect certain decisions, showing the CWDS to be more robust to system parameters than the fuzzy and Dempster-Shafer systems.