Anomaly detection: A robust approach to detection of unanticipated faults

Zhang, Bin; Sconyers, Chris; Byington, Carl S.; Patrick, Romano; Orchard, Marcos E.; Vachtsevanos, George

doi:10.1109/phm.2008.4711445

Cited by 28 publications

(16 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most common goal of anomaly detection is to raise an alarm when anomalous observations are encountered, such as insider threats [17], cyber attacks [15,23,16,21], machine component failures [27,28,1], sensor failures [7], novel astronomical phenomena [26], or the emergence of cancer cells in normal tissue [22,10]. In all of these cases, the underlying goal is to detect observations that are semantically distinct from normal observations.…”

Section: Requirements For Anomaly De-tection Benchmarksmentioning

confidence: 99%

Systematic construction of anomaly detection benchmarks from real data

Emmott

Das

Dietterich

et al. 2013

Proceedings of the ACM SIGKDD Workshop on Outlier Detection and Description

112

View full text Add to dashboard Cite

Research in anomaly detection suffers from a lack of realistic and publicly-available problem sets. This paper discusses what properties such problem sets should possess. It then introduces a methodology for transforming existing classification data sets into ground-truthed benchmark data sets for anomaly detection. The methodology produces data sets that vary along three important dimensions: (a) point difficulty, (b) relative frequency of anomalies, and (c) clusteredness. We apply our generated datasets to benchmark several popular anomaly detection algorithms under a range of different conditions.

show abstract

Section: Requirements For Anomaly De-tection Benchmarksmentioning

confidence: 99%

Systematic construction of anomaly detection benchmarks from real data

Emmott

Das

Dietterich

et al. 2013

Proceedings of the ACM SIGKDD Workshop on Outlier Detection and Description

112

View full text Add to dashboard Cite

show abstract

“…Both Markov-based and AI-based methods require extensive training before actual detection, and in practice, there might not be sufficient data for training purpose. Fault detection can also be performed with particle filters [27]- [29]. Particle filters are usually computationally expensive and also require extensive trainings beforehand.…”

Section: In Addition Online Diagnosis Methods Allow Automatic Remotementioning

confidence: 99%

“…Proof: The proof is given in Appendix A. Theorem 1: Assume the elements in the feature vector x n are independent. With the multi-candidate procedure defined in 1, the probability of false alarm is upper bounded by PFA ≤θ w A (27) whereθ = ∞ k=0 π k k is the prior mean of the change point θ , and w is the number of candidates.…”

Section: Performance Analysismentioning

confidence: 99%

Low Latency Bearing Fault Detection of Direct-Drive Wind Turbines Using Stator Current

Nath

Zhao

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Low latency change detection aims to minimize the detection delay of an abrupt change in probability distributions of a random process, subject to certain performance constraints such as the probability of false alarm (PFA). In this paper, we study the low latency detection of bearing faults of direct-drive wind turbines (WT), by analyzing the statistical behaviors of stator currents generated by the WT in real-time. It is discovered that the presence of fault will affect the statistical distribution of WT stator current amplitude at certain frequencies. Since the signature of a fault can appear in one of the multiple possible frequencies, we need to monitor the signals on multiple frequencies simultaneously, and each possible frequency is denoted as a candidate. Based on the unique properties of WT bearing faults, we propose a new multi-candidate low latency detection algorithm that can combine the statistics of signals from multiple candidate frequencies. The new algorithm does not require a separate training phase, and it can be directly applied to the stator current data and perform online detection of various possible bearing faults. The theoretical performance of the proposed algorithm is analytically identified in the form of upper bounds of the PFA and average detection delay (ADD). The algorithm allows flexible parametric adjustment of the tradeoff between PFA and ADD. INDEX TERMS Bearing fault, fault detection, quickest change detection, wind turbine.

show abstract

“…The idea is that measurements using multiple sensors in combination with environmental, operational, and performance-related parameters can provide a more accurate system health status. The sensor data can also be used along with statistical pattern recognition and machine-learning techniques to detect changes in machine parameter data, isolate faults, and estimate the remaining useful life (RUL) of the machines [6][7][8][9]. This approach assumes a product's loading and operating conditions, geometry, material properties, and failure mechanisms as the parameters to estimate RUL.…”

Section: Industry Segment Global Basementioning

confidence: 99%

Diode Laser‐Based Sensors for Extreme Harsh Environment Data Acquisition

Mitra¹,

Sharma²

2016

High Energy and Short Pulse Lasers

View full text Add to dashboard Cite

The world has witnessed several step changes in living standards, productivity, growth, and innovation. We are currently witnessing a convergence of intelligent devices, intelligent networks, and intelligent decision making. Obtaining long-term accurate, in situ, and real time data from the machines is necessary for enabling the industrial Internet. This relies heavily on sensor systems. Development of robust sensors that can operate reliably in extreme environments will make it possible to gather data from previously inaccessible locations in the equipment. This will enable machine operators to monitor and optimize the performance of their machines. Diode laser-based diagnostics technology has found applications in a variety of areas and a versatile range of operating conditions. It has proven to be a strong and reliable technique for remote measurements of concentrations and temperatures in harsh environments. Some of the major challenges for implementation of these sensors in real world are machine vibrations, window clogging, cooling, etc. In this chapter, the authors discuss about the application details and specific technologies suitable for the applications. Few case studies are considered, and the theoretical approach, algorithm development, and experimental validation are also discussed.

show abstract

Anomaly detection: A robust approach to detection of unanticipated faults

Cited by 28 publications

References 21 publications

Systematic construction of anomaly detection benchmarks from real data

Systematic construction of anomaly detection benchmarks from real data

Low Latency Bearing Fault Detection of Direct-Drive Wind Turbines Using Stator Current

Diode Laser‐Based Sensors for Extreme Harsh Environment Data Acquisition

Contact Info

Product

Resources

About