Information theoretic fault detection

A novel statistical model based on the higher-order statistics projection to the latent structure (HPLS) is proposed, which uses a combination of higher-order statistics (mutual information and differential entropy) instead of covariance to extract latent variables. This model not only has an explicit representation similar to the projection to latent structure model but can also capture the non-Gaussian process features. Furthermore, in order to reduce the redundant components contained in the latent variables independent of the quality variables, a novel strategy called independent signal correction is proposed. Finally, the novel independent signal correction HPLS (ISC-HPLS) model and the corresponding process monitoring strategy are developed. This model considers higher-order statistics that may involve certain key features of non-Gaussian processes and eliminates redundant components. Experimental results from synthetic numerical simulations and the Tennessee-Eastman process benchmark test verified the effectiveness of the proposed method.

show abstract

“…The ISC algorithm with mutual information as a criterion is proposed. The basic steps of the ISC algorithm are introduced in Algorithm 1 .…”

Section: Isc-hpls Model and Fault Detection Based On Isc-hplsmentioning

confidence: 99%

Non-Gaussian Quality Relevant Process Monitoring Based on Higher-Order Statistics Projection to the Latent Structure and Independent Signal Correction

Zhou

Yang

2023

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

show abstract

“…The linear relationship between the virtual loop controller and the actual controller can be obtained according to (26), as follows:…”

Section: Performance Assessment and Controller Tuning Resultsmentioning

confidence: 99%

“…X and Y are independent of each other when I(X; Y ) = 0. Mutual information extended to multidimensional inputs and outputs with [26]:…”

Section: Basic Theorymentioning

confidence: 99%

“…The CARMA-LADI algorithm is used to calculate the noise process function N i for different virtual loops, then the optimal controller G Copti for each virtual loop can be obtained by (33). Finally, the linear relationship obtained by (26) combined with G Copti can give the actual controller tuning direction to make the system work better.…”

Section: Parameter Estimationmentioning

confidence: 99%

See 1 more Smart Citation

Performance Assessment of MIMO System With Non-Gaussian Disturbances

Jia

Zhou

2022

IEEE Access

View full text Add to dashboard Cite

Performance assessment of control loops is of great importance for industrial production. This paper proposes a novel performance assessment and controller tuning method for non-Gaussian MIMO feedback control systems. First, an algorithm based on minimum entropy and mutual information projection to latent structure (ME-PLS) is proposed to replace the canonical correlation analysis algorithm (CCA). The ME-PLS algorithm decomposes the system data into independent components related to inputs and outputs, and this algorithm applies to both Gaussian and non-Gaussian systems. Each pair of principal components represents a virtual non-Gaussian control loop. Next, the performance of each virtual loop is calculated separately with the non-Gaussian minimum entropy method. Finally, a least absolute deviation iterative algorithm based on the CARMA model (CARMA-LADI) is given to identify the parameters of each virtual loop and to give the actual controller tuning directions based on the relationships obtained by the ME-PLS algorithm.INDEX TERMS control loop performance assessment, independent components, MIMO system, maximum mutual information, minimum entropy

show abstract

“…Many research studies have been conducted in the field of fault detection. To study the relationship among variables, an information‐theoretic measure called MI was suggested in [18–20] in which the MI is used as a measure of correlation among variables to identify faults. Kraemer et al .…”

Section: Literature Reviewmentioning

confidence: 99%