PLS-based EWMA fault detection strategy for process monitoring

PLS-based memory control scheme for enhanced process monitoring

“…The method is tested using a distillation column process simulated by Aspen (see [7] for details) with added zero-mean Gaussian noise, where the predictor variables consist of ten temperatures (T c) in different stages of the monitored column, feed flow rates and reflux stream, and the composition of the light component in the distillate stream represents the response variable. The Aspen simulator is used to generate 1024 data samples to be used in constructing the reference PLS model.…”

“…Process monitoring is employed by various process industries [1], [3], [4], [5]. Partial least square (PLS) is among the most widely used multivariate statistical process monitoring method for monitoring multivariate processes [6], [7]. Roughly speaking, PLS aims to extract from the predictor (input) variables latent variables which are the most correlated and relevant to predicted (output) variables [8].…”

“…In this framework, several PLS variants have been proposed to overcome some shortcomings of the classical PLS, such as multiway PLS [10], a modified PLS [11], [12], multi-block PLS [13], a multi-scale PLS [14], dynamic PLS [15], recursive PLS [16] as well as multi-phase PLS [17]. Therefore, PLS-based process monitoring method as well its variants has been largely exploited and used to different engineering applications [18], [7], [19].…”

A statistical-based approach for fault detection and diagnosis in a photovoltaic system

“…This control chart have two chief advantages and which are : the high ability to detect small shift in the process mean, since its time-weighted nature, and its great flexibility, which can be seen in the possibility of obtaining high accurate results by adjusting only few parameters. EWMA's has been originally introduced by Roberts et al in [15], then it has been significantly applied in time series analysis [16,17]. Overall several years, EWMA control chart has been applied by lot of engineers and scientists from various areas [18,19].…”

Early detection of abnormal patient arrivals at hospital emergency department

“…Generally, small values of λ increase the chart's sensitivity to small shifts in the process mean, while large values of λ increase its sensitivity to large shifts [18], [29]. The standard deviation of z t is defined as σ zt = σ 0…”