China
Africa
Explore chapters and articles related to this topic
Fault Tolerant Industrial Wireless Sensor Networks
Published in V. Çağri Güngör, Gerhard P. Hancke, Industrial Wireless Sensor Networks, 2017
Ataul Bari, Jin Jiang, Arunita Jaekel
The goal of fault detection is to detect “abnormalities” by verifying that the services being provided are functioning properly and to predict, if possible, whether they will continue to function in the near future [43]. In many cases, it is possible for a node to perform a self-diagnosis to check for any possible anomalies. For example, the work in [20] has introduced fault tolerance into WSNs by monitoring the status of each wireless sensor node. The focus has been on the detection of physical malfunctions of sensor nodes caused by impacts or incorrect orientation. Similarly, battery depletion can be detected if nodes can measure the battery output [34]. Incorrectly generated values from a faulty sensor node can also be detected using group detection. Typically, it is expected that sensors from the same region will generate similar values as additional/redundant sensors can usually be deployed in the regions in order to obtain finer-grained information [24]. Hence, a fault probability can be calculated based on the information from the neighbors of a node [24]. Another approach, based on quartile method, has been proposed in [47] for fault tolerant sensing. The approach can select correct data based on data discreteness so that actors can perform appropriate actions.
Analytics for Operations and Equipment Maintenance in Buildings and on Campuses
Published in John J. “Jack” Mc Gowan, Energy and Analytics, 2020
In the past, predictive maintenance was limited to high-value assets. What’s different today is that highly skilled professionals can deploy analytics technology to automatically collect and analyze enough data so that predictive maintenance can even be applied to small end point devices. Some of these devices might include variable air volume boxes and process utility connection points, such as water for injection and compressed air connection points. The computer does the work, so automatic fault detection and diagnostics can be scaled down to the low-cost ubiquitous devices and sensors in a system.
Introduction
Published in Sunan Huang, Kok Kiong Tan, Poi Voon Er, Tong Heng Lee, Intelligent Fault Diagnosis and Accommodation Control, 2020
Sunan Huang, Kok Kiong Tan, Poi Voon Er, Tong Heng Lee
Figure 1.4 depicts fault diagnosis research. Fault diagnosis involves fault detection, fault isolation and fault identification. Fault detection is to check whether there is a malfunction or fault in the system and make the decision that a fault has occurred. Furthermore, for a practical problem, even when we have detected the occurrence of a fault, it is necessary to isolate the faulty element and find out the fault types. This is called fault isolation. When the detected fault cannot be isolated, fault identification is triggered to find out the fault characteristics.
A Review of machine learning techniques for wind turbine’s fault detection, diagnosis, and prognosis
Published in International Journal of Green Energy, 2023
Prince Waqas Khan, Yung-Cheol Byun
Machine learning is widely used in various fields, including the wind energy sector. One of the major uses of machine learning is to detect and diagnose faults in wind turbines. Fault detection is the process of identifying the presence of faults or abnormalities in the system, while fault diagnosis is the process of identifying the specific type and location of the fault. In the context of the wind turbine, fault detection can be achieved through the analysis of SCADA data and vibration signatures, while fault diagnosis typically requires more detailed analysis and inspection, such as visual inspection of the turbine components or analysis of oil samples. Fault prognosis refers to the process of predicting the future behavior or condition of a system or component, based on its current state and past performance data. In the context of wind turbines, fault prognosis involves predicting the Remaining Useful Life (RUL) of a component or system.
Hybrid feature adaptive fusion network for multivariate time series classification with application in AUV fault detection
Published in Ships and Offshore Structures, 2023
Shaoxuan Xia, Xiaofeng Zhou, Haibo Shi, Shuai Li
Security is a key consideration in the operation of AUVs. Fault detection technologies form the basis of ensuring security, and thus, preventing or reducing potential accidents and economic losses. This paper focuses on the security of AUV, exploring an intelligent fault detection method. In recent years, research into AUV fault detection has seen considerable advances, with many researchers contributing to the development of this technology. However, most of the current studies have largely focused on the detection of a single type of fault, such as thruster fault (Yu D et al. 2020). In this study, we expand the scope of fault types to include light damage to the propeller, severe damage to the propeller, failure of the depth sensor, and load increase. It should be noted that these faults are deliberately designed for the experiments. The proposed algorithm has been proven to possess the capacity to acquire knowledge of other faults and effectively detect them.
Active fault diagnosis of 2 DoF helicopter using particle filter-based log-likelihood ratio
Published in International Journal of Control, 2022
S. Kanthalakshmi, M. Raghappriya
Fault detection and diagnosis uses model-based method to detect the presence of faults in the system. System without fault and with faults are modeled as a normal model and fault models respectively. The normal model represented by (1) and (2) is chosen as .The faults in the system are characterised by the changes in the parameters of the system. The system with faults can be represented by modifying (1) and (2) by including the fault bias vector. γ denotes the scaling factor of effectiveness representing the faults.The fault model is represented as where f is the number of faults in the system. Thus a total of F fault models available are represented as state space model . Hypothesis testing methods are the most popular to monitor parametric changes in the system (Alrowaie et al., 2012). To detect the changes, one can consider two hypotheses, one for the normal model and the other for the faulty model. Null hypothesis represents the normal model and alternate hypothesis represents fault model respectively.