Early detection and identification of dangerous states in chemical plants using neural networks

Neumann, Joachim; Deerberg, Görge; Schlüter, Stefan

doi:10.1016/s0950-4230(98)00059-x

Cited by 5 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, because modern industrial processes often present a large number of highly correlated process (Harrou et al (2015)), principal component analysis (PCA) (Harrou et al (2013)), canonical variate analysis, independent component analysis (Chiang et al (2001)), neural networks (Neumann and G. Deerberg (1999)), and support vector machine based methods (Dehestani et al (2011))). Data-based monitoring methods, especially those that utilize PCA or its extensions, have been applied across a wide range of industries, for example in the chemical industry (Simoglou et al (1997)), for water treatment (George et al (2009)), and in ecological studies (Janzekovic and Novak (2012)).…”

Section: Accepted Manuscriptmentioning

confidence: 99%

Amalgamation of anomaly-detection indices for enhanced process monitoring

Harrou

Sun

Khadraoui

2016

Journal of Loss Prevention in the Process Industries

View full text Add to dashboard Cite

Accurate and effective anomaly detection and diagnosis of modern industrial systems are crucial for ensuring reliability and safety and for maintaining desired product quality. Anomaly detection based on principal component analysis (PCA) has been studied intensively and largely applied to multivariate processes with highly cross-correlated process variables; howver conventional PCA-based methods often fail to detect small or moderate anomalies. In this paper, the proposed approach integrates two popular process-monitoring detection tools, the conventional PCA-based monitoring indices Hotelling's T 2 and Q and the exponentially weighted moving average (EWMA). We develop two EWMA tools based on the Q and T 2 statistics, T 2-EWMA and Q-EWMA, to detect anomalies in the process mean. The performances of the proposed methods were compared with that of conventional PCA-based anomaly-detection methods by applying each method to two examples: a synthetic data set and experimental data collected from a flow heating system. The results clearly show the benefits and effectiveness of the proposed methods over conventional PCA-based methods.

show abstract

Section: Accepted Manuscriptmentioning

confidence: 99%

Amalgamation of anomaly-detection indices for enhanced process monitoring

Harrou

Sun

Khadraoui

2016

Journal of Loss Prevention in the Process Industries

View full text Add to dashboard Cite

show abstract

“…Because of this, engineers must keep tweaking and improving the reliability of their processes, watching carefully for signs of anomalies that could lead to disaster. Therefore, it is crucial to be able to detect and identify any possible faults or failures in the system as early as possible [2,4,5].…”

Section: Introductionmentioning

confidence: 99%

An Improved Wavelet‐Based Multivariable Fault Detection Scheme

Harrou¹,

Sun²,

Madakyaru³

2017

Uncertainty Quantification and Model Calibration

View full text Add to dashboard Cite

Data observed from environmental and engineering processes are usually noisy and correlated in time, which makes the fault detection more difficult as the presence of noise degrades fault detection quality. Multiscale representation of data using wavelets is a powerful feature extraction tool that is well suited to denoising and decorrelating time series data. In this chapter, we combine the advantages of multiscale partial least squares (MSPLSs) modeling with those of the univariate EWMA (exponentially weighted moving average) monitoring chart, which results in an improved fault detection system, especially for detecting small faults in highly correlated, multivariate data. Toward this end, we applied EWMA chart to the output residuals obtained from MSPLS model. It is shown through simulated distillation column data the significant improvement in fault detection can be obtained by using the proposed methods as compared to the use of the conventional partial least square (PLS)-based Q and EWMA methods and MSPLS-based Q method.

show abstract