Fault diagnosis and its prediction in wireless sensor networks using regressional learning to achieve fault tolerance

Due to the wide range of critical applications and resource constraints, sensor node gives unexpected responses, which leads to various kind of faults in sensor node and failure in wireless sensor networks. Many research studies focus only on fault diagnosis, and comparatively limited studies have been conducted on fault diagnosis along with fault tolerance in sensor networks. This paper reports a complete study on both 2 aspects and presents a fault tolerance approach using regressional learning with fault diagnosis in wireless sensor networks. The proposed method diagnose the different types of faulty nodes such as hard permanent, soft permanent, intermittent, and transient faults with better detection accuracy. The proposed method follows a fault tolerance phase where faulty sensor node values would be predicted by using the data sensed by the fault free neighbors. The experimental evaluation of the fault tolerance module shows promising results with R² of more than 0.99. For the periodic fault such as intermittent fault, the proposed method also predict the possible occurrence time and its duration of the faulty node, so that fault tolerance can be achieved at that particular time period for better performance of the network.     

Composite Fault Diagnosis in Wireless Sensor Networks Using Neural Networks

Wireless sensor networks (WSNs) are spatially distributed devices to support various applications. The undesirable behavior of the sensor node affects the computational efficiency and quality of service. Fault detection, identification, and isolation in WSNs will increase assurance of quality, reliability, and safety. In this paper, a novel neural network based fault diagnosis algorithm is proposed for WSNs to handle the composite fault environment. Composite fault includes hard, soft, intermittent, and transient faults. The proposed fault diagnosis protocol is based on gradient descent and evolutionary approach. It detects, diagnose, and isolate the faulty nodes in the network. The proposed protocol works in four phases such as clustering phase, communication phase, fault detection and classification phase, and isolation phase. Simulation results show that the proposed protocol performs better than the existing protocols in terms of detection accuracy, false alarm rate, false positive rate, and detection latency.

     

A heterogeneous fault diagnosis approach to enhance performance of connected vehicles

Evolution of wireless access technology, availability of smart sensors, and reduction in the size of the set up of the communication system have engrossed many researchers toward vehicular ad hoc network (VANET). Vehicle-to-vehicle and vehicle-to-access-point communication in a vehicular environment facilitates the deployment of VANET for many different purposes. The success of any application implemented in a VANET relies on timely and accurate data dissemination across the nodes of the network. Implementation of any application is not going to be fruitful if the communication unit transmits incorrect sensor data due to the presence of a fault. This article focuses on the automatic detection of hard and soft faults for vehicular sensors and the classification of faults into permanent, intermittent, and transient faults using cloud-based VANET. For the cloud service, ThingSpeak cloud is used. At the RSU of the VANET, hard fault detection is performed, and for this purpose, a time-out strategy is proposed. The observation center, after receiving sensor status data over a vehicular cloud, does soft failure detection. The soft fault is identified by utilizing a comparative-based technique during soft fault diagnosis. Soft faults are categorized using two machine learning algorithms: Support vector machine and logistic regression. The effectiveness of the suggested work is assessed using performance metrics like fault detection accuracy, false alarm rate, false positive rate, precision, accuracy, recall, and F1 score.     

Fault diagnosis in wireless sensor network using clonal selection principle and probabilistic neural network approach

The fault diagnosis in wireless sensor networks is one of the most important topics in the recent years of research work. The problem of fault diagnosis in wireless sensor network can be resembled with artificial immune system in many different ways. In this paper, a detection algorithm has been proposed to identify faulty sensor nodes using clonal selection principle of artificial immune system, and then the faults are classified into permanent, intermittent, and transient fault using the probabilistic neural network approach. After the actual fault status is detected, the faulty nodes are isolated in the isolation phase. The performance metrics such as detection accuracy, false alarm rate, false‐positive rate, fault classification accuracy, false classification rate, diagnosis latency, and energy consumption are used to evaluate the performance of the proposed algorithm. The simulation results show that the proposed algorithm gives superior results as compared with existing algorithms in terms of the performance metrics. The fault classification performance is measured by fault classification accuracy and false classification rate. It has also seen that the proposed algorithm provides less diagnosis latency and consumes less energy than that of the existing algorithms proposed by Mohapatra et al, Panda et al, and Elhadef et al for wireless sensor network.     

Analog circuit diagnosis based on the nullor concept and multiport description of the circuit

The subject of this paper is the fault diagnosis of analog circuits based on the use of nullor concept. The fault location technique presented in the paper can be implemented in the general-purpose analysis program which provides many advantages, of which the most important is the automation of the diagnosis process. A simulation based diagnosis model can be obtained by introducing the norators across the potentially faulty elements and the fixators at the accessible nodes. A practical problem that arises when using this nullor diagnosis model is a lack of an efficient procedure for localization of multiple faults. In the proposed diagnosis technique, the online computational requirements are reduced by introducing a diagnosis model that contains accessible nodes only. The diagnosis model is obtained from the original circuit using relationships among the measured voltages and compensated currents of the faulty elements. The proposed faulty location technique is validated on a benchmark example.     

一种新的容差模拟电路故障屏蔽字典法

提出了一种新的可以诊断容差模拟电路软硬两类故障的故障屏蔽字典法。该法基于统计学原理，计算了容差对故障特征量的影响，在此基础上，通过构造合适的故障隶属函数，将故障定位在一个模糊故障集中，并从中进一步确定出最可能发生的故障。     

Soft Fault Feature Extraction in Nonlinear Analog Circuit Fault Diagnosis

Aiming at the problem to diagnose soft faults in nonlinear analog circuits, a novel approach to extract fault features is proposed. The approach is based on the Wigner–Ville distribution (WVD) of the subband Volterra model. First, the subband Volterra kernels of the circuit under test are cleared. Then, the subband Volterra kernels are used to obtain the WVD functions. The fault features are extracted from the WVD functions and taken as input data into the hidden Markov model (HMM). Finally, with classification of features using HMMs, the soft fault diagnosis of the nonlinear analog circuit is achieved. The simulations and experiments show that the method proposed in this paper can extract the fault features effectively and improve the fault diagnosis.     

Diagnosis of sensor faults in active magnetic bearing systemequipped with built-in force transducers

In this paper, an online diagnosis scheme for sensor faults in an active magnetic bearing system equipped with built-in force transducers is proposed. The scheme, utilizing redundant signals, i.e., displacement, current and force, is based on the causality between faults and symptoms. The main advantage of the scheme is that it is simple and efficient, because monitoring the symptom only is enough to diagnose a sensor fault with no elaborate signal processing. Experiments were also performed to demonstrate the effectiveness of the proposed method in the detection and diagnosis of sensor faults     

Choice of Detection Parameters on Fault Detection in Wireless Sensor Networks: A Multiobjective Optimization Approach

In this paper, the intermittent fault detection in wireless sensor networks is formulated as an optimization problem and a recently introduced multiobjective swarm optimization (2LB-MOPSO) algorithm is used to find an optimum trade-off between detection accuracy and detection latency. Faulty sensor nodes are identified based on comparisons of sensed data between one-hop neighboring nodes. Time redundancy is used to detect intermittent faults since an intermittent fault does not occur consistently. Simulation and analytical results show that sensor nodes with permanent faults are identified with high accuracy and by properly choosing the inter-test interval most of the intermittent faults are isolated with negligible performance degradation.     

基于BP神经网络的导弹故障自动诊断研究

精确制导空空导弹是一个复杂系统,及时、准确的识别和诊断导弹故障是化解风险的重要措施。由于导弹故障模式复杂,对其故障进行识别和诊断的难度很大。文中以某型空空导弹测试问题为例,提出一种基于BP神经网络算法的导弹故障自动识别与诊断技术。通过对导弹测试数据的采集和整理形成数据样本,利用神经网络系统的学习和判断能力自动识别及诊断导弹故障,并使用Matlab神经网络工具箱进行仿真验证。验证结果证明,该技术能够快速、准确的识别和诊断导弹故障。     

基于MIV和BRBP神经网络的电路板红外诊断方法

针对BP神经网络对于海量数据训练及多维数据训练收敛困难的问题,在使用增加动力项、自适应学习速率等方法的基础上,引入均值影响度算法(MIV)构造了贝叶斯正则化反向传播(BRBP)神经网络,以此提高电子线路板红外故障诊断算法的效率。利用红外测温方式,获取了不同室温及运行状态下电路板中21个元器件温度数据。将此21个参数作为故障诊断模型的初始输入变量,经过MIV算法简约为12个参数输入至BRBP神经网络,进行故障评估和诊断。结果表明:相对于传统的BRBP神经网络,本文设计的基于MIV和BRBP神经网络模型诊断方法极大简化了数据训练的数据量并解决了数据收敛的困难,因此效率更高,用时更省。     

Synergistic use of soft computing technologies for fault detection in gas turbine engines

In this paper, we present a synergistic approach to startup fault detection and diagnosis (FDD) in gas turbine engines. The method employs statistics, signal processing, and soft computing techniques in a complementary manner to address fault detection at transient conditions. Traditional turbine engine FDD methods are based on engine data collected at steady-state conditions. However, incipient faults are difficult to diagnose using steady-state engine data; only engine faults that are fairly developed can be detected using conventional methods. Because incipient engine component faults are often manifest in the engine startup characteristics, we present a method to characterize the engine transient startup. Engine sensor data during engine startup are recorded in time series format. The sensor profiles corresponding to "good" and "bad" engine startups are sampled using the bootstrap technique. A feature vector is extracted in two steps, and signal processing is followed by the feature vector selection. In the signal processing step, principal component analysis (PCA) is applied to reduce the samples consisting of sensor profiles into a smaller set. In the feature vector selection step, a cost function is defined, and important discriminating features for fault diagnosis are distilled from the PCA output vector. The features obtained from this step are then classified using neural-network-based methods. The "leave-one-out" approach to cross validation is applied to obtain an objective evaluation of the neural network training. The proposed FDD method is evaluated using actual engine startup data, and the results are presented.     

基于闭环集成运放的模拟电路模块级软故障诊断

基于斜率故障模型,提出了一种诊断模拟电路中基于闭环集成运算放大器的模块级软故障的字典法.在由闭环运放组成的模拟电路中,通过对电路以闭环运放及与其输入直接相连的元件看作一个整体划分模块,对各个模块中的任一元件或进行宏模型替代之后的运放等效电路,利用电路中的两节点电压增量计算出的斜率作为统一故障特征,建立故障字典,实现电路中相应模块包含的运放和所有元件的软故障诊断.给出了运放的等效宏模型和模块级软故障的诊断步骤,并用仿真实例证明了该诊断方法的有效性.     

基于深度动态贝叶斯网络的服务功能链故障诊断算法

针对5G端到端网络切片场景下底层物理节点出现故障会导致运行在其上的多条服务功能链出现性能异常的问题,该文提出一种基于深度动态贝叶斯网络(DDBN)的服务功能链故障诊断算法。首先根据网络虚拟化环境下故障的多层传播关系,构建故障与症状的依赖图模型,并采用在物理节点监测其上多个虚拟网络功能相关性能数据的方式收集症状。其次,考虑到基于软件定义网络(SDN)和网络功能虚拟化(NFV)的架构下网络症状观测数据的多样性以及物理节点和虚拟网络功能的空间相关性,引入深度信念网络对观测数据特征进行提取,使用加入动量项的自适应学习率算法对模型进行微调以加快收敛速度。最后,利用故障传播的时间相关性,引入动态贝叶斯网络对故障根源进行实时诊断。仿真结果表明,该算法能够有效地诊断故障根源且具有良好的诊断准确度。     

Diagnosis of Wireless Sensor Networks in Presence of Permanent and Intermittent Faults

In this paper we consider the problem of distributed fault diagnosis in Wireless Sensor Networks (WSNs). The proposed Fault Diagnosis Algorithm (FDA) aims to handle both permanent and intermittent faults. The sensor nodes with permanent communication faults can be diagnosed by using the conventional time-out mechanism. In contrast, it is difficult to detect intermittent faults due to their inherent unpredictable behavior. The FDA is based on the comparison of sensor measurements and residual energy values of neighboring sensor nodes, exploiting their spatial correlations. To handle intermittent faults, the comparisons are made for \(r\) rounds. Two special cases of intermittent faults are considered: one, when an intermittently faulty node sends similar sensor measurement and similar residual energy value to some of its neighbors in all \(r\) rounds; another, when it sends these values, either or both of which deviates significantly from that of some neighbors in all \(r\) rounds. Through extensive simulation and analysis, the proposed scheme is proved to be correct, complete, and efficient to handle intermittent faults and hence, well suited for WSNs.     

On-line test for fault-secure fault identification

In an increasing number of applications, reliability is essential. On-line resistance to permanent faults is a difficult and important aspect of providing reliability. Particularly vexing is the problem of fault identification. Current methods are either domain specific or expensive. We have developed a fault-secure methodology for permanent fault identification through algorithmic duplication without necessitating complete functional unit replication. Fault identification is achieved through a unique binding methodology during high-level synthesis based on an extension of parity-like error correction equations in the domain of functional units. The result is an automated chip-level approach with extremely low area and cost overhead     

Artificial neural-network model-based observers

Describes a pseudorandom testing scheme for fault diagnosis of analog integrated circuits. The goal is to implement a BIST technique with both a built-in pattern generator and a response analyzer for fault diagnosis. We have chosen a diagnostic framework for the analog ICs using a pseudorandom noise generator as the test-pattern generator and a model-based observer to detect and diagnose faults. The observer is implemented through a multilayer feedforward ANN trained with a back-error propagation (BEP) algorithm. Both the test-pattern generator and the model-based observer proposed in this article can be implemented either on- or offline depending on the need of the application and silicon area overhead.     

子网络级模拟故障诊断的神经网络实现

利用容差模拟电路节点电压灵敏度序列守恒定理,得到了模拟电路元件的软、硬故障统一样本。然后利用统一样本集训练BP神经网络,并将神经网络用于子网络级模拟故障诊断。实例验证表明,软、硬故障统一样本集使得用于神经网络训练所需样本数目大大减少,但经过训练的神经网络可以诊断容差模拟电路的全部软、硬故障,而且诊断正确率较高。     

Fault Diagnosis of Analog Circuits Using Systematic Tests Based on Data Fusion

An analog fault diagnosis approach using a systematic step-by-step test is proposed for fault detection and location in analog circuits with component tolerance and limited accessible nodes. First, by considering soft faults and component tolerance, statistics-based fault detection criteria are established to determine whether a circuit is faulty by measuring accessible node voltages. For a faulty circuit, fuzzy fault verification is performed using the accessible node voltages. Furthermore, using an approximation technique, the most likely faulty elements are identified with a limited number of circuit gain measurements at selected frequencies. Finally, employing the D-S evidence theory, synthetic decision is made to locate faults according to the results of fault verification and estimation. Unlike other methods which use a single diagnosis method or a particular type of measurement information, the proposed approach makes use of the redundancy of different types of measurement information and the combined use of different diagnosis methods so as to improve diagnosis accuracy.     

基于小波包分析和Elman神经网的故障诊断方法研究

由非线性电力电子装置组成的电路发生故障时,故障特征信息不易提取和识别。对此提出一种基于小波包分析和Elman神经网的电力电子装置故障诊断的方法,先运用小波包分析法提取电路在不同故障状态下电压及电流信号的特征信息,然后对数据进行归一化处理并作为Elman神经网的输入,由具有智能学习功能的神经元故障分类器完成故障识别和定位。以12脉冲整流电路为例,在Matlab软件下建立电路模型进行仿真实验,结果表明该方法能快速、准确的完成故障诊断。