Adaptive fault detection and diagnosis using parsimonious Gaussian mixture models trained with distributed computing techniques

Nakamura, Thiago Akio; Palhares, Reinaldo M.; Caminhas, Walmir M.; Menezes, B.R.; Campos, Mário Cesar Mello Massa de; Fumega, Ubirajara; Bomfim, Carlos H. de M.; Lemos, André Paim

doi:10.1016/j.jfranklin.2016.11.024

Cited by 12 publications

(3 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This method achieved a high FDR; however, the parameter division requirement was difficult to implement and involved a complicated algorithm. Recently, many fault detection methods have chosen Gaussian distribution functions to analyze feature parameters [21][22][23][24]. A Gaussian distribution is also called the normal distribution function.…”

Section: Related Workmentioning

confidence: 99%

Fault Detection Method for Wi-Fi-Based Smart Home Devices

Cheng

Zhang

et al. 2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

At present, the dynamic nature and unstable network connections in the deployment environments of Wi-Fi-based smart home devices make them susceptible to component damage, crashes, network disconnections, etc. To solve these problems, researchers have used various fault detection methods, such as alarming when monitored fault parameters exceed the preset values, model-based mathematical methods, device signal processing-based methods, and artificial intelligence-based methods. However, these methods require large numbers of fault parameters, the model are complex, and their fault detection accuracy is relatively poor. To more quickly and accurately detect faults in smart home devices and ensure the continuity of people’s daily work and lives, this paper analyzes both the Wi-Fi traffic characteristics of smart home devices and the complexity and difficulty of traditional fault detection methods and proposes a fault detection method based on TDD (Throughput and Delay Distribution). This method obtains throughput and data packet delay distribution by capturing Wi-Fi communication and sending test data. By dividing the throughput into heartbeat data and command information, we can calculate the real-time throughput and further calculate the similarity between the real-time throughput and the throughput in database. Also, the resulting delay distribution is compared with the probability distribution of delay in the database. When the throughput values are sufficiently similar and the delays are all in the normal range, the smart home secure devices are functioning properly. The experimental results show that the proposed TDD method can detect faults in household devices in real time and that it achieves high recall and good detection accuracy in Wi-Fi communication environment.

show abstract

Section: Related Workmentioning

confidence: 99%

Fault Detection Method for Wi-Fi-Based Smart Home Devices

Cheng

Zhang

et al. 2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

show abstract

“…Todo componente da planta e do sistema de controle está passível de falhas. O sistema de Detecção e Diagnóstico de Falhas (FDD)(do inglês, Fault Detection and Diagnosis) deve ser capaz de detectar problemas independente de onde ocorre (Figura 1), mesmo não indicando a localização e a causa da falha (Nakamura et al, 2017).…”

Section: Sistemas De Detecção E Diagnóstico De Falhas (Fdd)unclassified

Proposta de Framework com Foco em Métodos Evolutivos para Detecção, Diagnóstico e Causa Raiz de Falhas em Processos Industriais

Almeida

Caminhas

2020

Anais Do Congresso Brasileiro De Automática 2020

View full text Add to dashboard Cite

Nas últimas décadas houve um crescimento expressivo do grau de complexidadee dimensionalidade dos processos nas indústrias, em parte, devido à revolução tecnológica que vem ocorrendo em igual período. Por consequência, surgiu a necessidade de ferramentas computacionais capazes de lidar com essa nova dinâmica e contribuir positivamente para a supervisão. Neste trabalho, propõe-se um framework escalável para uso no monitoramento de processos e equipamentos que utiliza métodos evolutivos, independente de sua abordagem, para modelar a dinâmica do processo, e permite, ainda, acoplar a saída, um método que buscarexplicar eventuais anormalidades detectadas pelo modelador, por meio da indicação de um conjunto de variáveis que desviam do comportamento esperado. Um experimento utilizando o benchmark DAMADICS é realizado para demonstrar o uso da ferramenta proposta.

show abstract

“…Nakamura et al proposed a statistical model-based fault detection and diagnosis system, which is based on a series of parsimonious GMMs (PGMMs) distributions, and has good performance in correctly detecting and diagnosing faults. [5] For chemical process fault detection and diagnosis, Yu designed a nonlinear kernel GMM (NKGMM)-based inference monitoring method. [6] To detect the faults in semiconductor manufacturing, Yu proposed a principal component-based GMM (PCGMM), which is superior to PCA-based monitoring models and enables accurate early detection of various types of faults in complex manufacturing processes.…”

Section: Introductionmentioning

confidence: 99%

Dynamic stationary subspace analysis based on Gaussian mixture models for ironmaking process monitoring

Zhang,

Fan,

Guo

et al. 2023

Can J Chem Eng

View full text Add to dashboard Cite

Blast furnace ironmaking process monitoring is an important and challenging task. Due to the influence of hot blast stove switching and large fluctuations in the quality of raw materials, the measurements of ironmaking processes show obvious non‐stationary characteristics, and in addition, the observed data are also characterized by time‐series dynamic and non‐Gaussian characteristics. In this paper, a dynamic stationary subspace analysis method based on the Gaussian mixture model (DSSA–GMM) is proposed to address the difficulties in blast furnace ironmaking process monitoring. The time‐series dynamic relationship of the data is conducted by introducing a sliding time window. The Gaussian mixture model (GMM) is used to deal with the non‐Gaussian characteristics of the data, and the parameters of the GMMs are estimated using the expectation–maximization algorithm. The stationary projection matrix is obtained by optimizing the Kullback–Leibler (K–L) divergence between GMMs of different periods to realize the stationary subspace separation. Finally, the convex hull of the stationary subspace is established for fault detection, thus realizing the monitoring for non‐stationary and non‐Gaussian dynamic processes. The effectiveness of the DSSA–GMM method is verified by a numerical simulation and a dataset collected from an actual blast furnace ironmaking process.

show abstract

Adaptive fault detection and diagnosis using parsimonious Gaussian mixture models trained with distributed computing techniques

Cited by 12 publications

References 55 publications

Fault Detection Method for Wi-Fi-Based Smart Home Devices

Fault Detection Method for Wi-Fi-Based Smart Home Devices

Proposta de Framework com Foco em Métodos Evolutivos para Detecção, Diagnóstico e Causa Raiz de Falhas em Processos Industriais

Dynamic stationary subspace analysis based on Gaussian mixture models for ironmaking process monitoring

Contact Info

Product

Resources

About