A GMDH neural network-based approach to robust fault diagnosis: Application to the DAMADICS benchmark problem

Witczak, Marcin; Korbicz, Józef; Mrugalski, Marcin; Patton, Ron J.

doi:10.1016/j.conengprac.2005.04.007

Cited by 111 publications

(52 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Results obtained by the authors of [34] show that the small type fault f 16 and incipient type fault f 18 are not detectable, while the authors of [33] stated that using their method the faults f 9 and f 16 are not detectable. The authors of [5] did not detected some types of faults: f 5, f 8, f 9, f 12 and f 14 using the GMDH neural network-based approach. In the presented study all of the faulty cases were considered in the analysis.…”

Section: Data Preparationmentioning

confidence: 99%

“…The authors of [30] focused on the abrupt large faults available in the DAMADICS benchmark for faults detection using a spectral estimation approach. Other approaches were to consider GMDH neural networks [5], interval observers [31] fuzzy classifiers for fault detection and isolation [32], structural analysis [33] in order to evaluate fault isolability, etc. The only study, which considered the whole set of faults possible to simulate in the DAMADICS benchmark, was performed by the authors of [34].…”

Section: B Previous Studies Related To Diagnosis Using Damadics Bencmentioning

confidence: 99%

“…In [5] the authors applied the group method of data handling (GMDH) neural networks, while in [6], [7] the authors used a neurofuzzy approach and ANN with decision trees, in [8] the hidden Markov model for the fault diagnosis was applied.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Faults diagnosis using self-organizing maps: A case study on the DAMADICS benchmark problem

Katunin

Amarowicz

Chrzanowski

2015

Annals of Computer Science and Information Systems

View full text Add to dashboard Cite

Abstract-This paper deals with a method of faults detection and identification based on the clusterization of the multiple diagnostic signals. Various types of faults and character of their occurrence were simulated using DAMADICS Benchmark Process Control System. A great advantage of the applied approach based on self-organizing (Kohonen) maps is that even the smallest differences in signals allow for detection, isolation and identification of type of occurred faults with respect to the healthy condition of the investigated system based on the unsupervised learning. It was shown that in some cases the faults, which are undetectable during monitoring of simple heuristic and statistical parameters and other previously applied methods, are recognizable when the approach based on self-organizing maps is applied. The case studies presented in this paper show the faults detection procedure as well as clusterization of types and successful classification of almost all the unique faulty states of the investigated system.

show abstract

Section: Data Preparationmentioning

confidence: 99%

Section: B Previous Studies Related To Diagnosis Using Damadics Bencmentioning

confidence: 99%

See 1 more Smart Citation

Faults diagnosis using self-organizing maps: A case study on the DAMADICS benchmark problem

Katunin

Amarowicz

Chrzanowski

2015

Annals of Computer Science and Information Systems

View full text Add to dashboard Cite

show abstract

“…Thus, it seems desirable to have a tool which can be employed for automatic selection of the ANN structure, based only on the measured data. To overcome this problem, GMDH neural networks (Ivakhnenko and Mueller, 1995;Witczak et al, 2006) have been proposed. The synthesis process of a GMDH model is based on iterative processing of a sequence of operations.…”

Section: Robust Gmdh Neural Networkmentioning

confidence: 99%

Towards Robustness in Neural Network Based Fault Diagnosis

Patan¹,

Witczak²,

Korbicz³

2008

International Journal of Applied Mathematics and Computer Science

Self Cite

View full text Add to dashboard Cite

Challenging design problems arise regularly in modern fault diagnosis systems. Unfortunately, classical analytical techniques often cannot provide acceptable solutions to such difficult tasks. This explains why soft computing techniques such as neural networks become more and more popular in industrial applications of fault diagnosis. Taking into account the two crucial aspects, i.e., the nonlinear behaviour of the system being diagnosed as well as the robustness of a fault diagnosis scheme with respect to modelling uncertainty, two different neural network based schemes are described and carefully discussed. The final part of the paper presents an illustrative example regarding the modelling and fault diagnosis of a DC motor, which shows the performance of the proposed strategy.

show abstract

“…Another difficulty follows from the fact that the model obtained during system identification is usually uncertain (Witczak et al, 2006). Model uncertainty can appear during both stages of system identification, i.e., model structure selection and parameter estimation.…”

mentioning

confidence: 99%

An unscented Kalman filter in designing dynamic GMDH neural networks for robust fault detection

Mrugalski¹

2013

International Journal of Applied Mathematics and Computer Science

Self Cite

View full text Add to dashboard Cite

This paper presents an identification method of dynamic systems based on a group method of data handling approach. In particular, a new structure of the dynamic multi-input multi-output neuron in a state-space representation is proposed. Moreover, a new training algorithm of the neural network based on the unscented Kalman filter is presented. The final part of the work contains an illustrative example regarding the application of the proposed approach to robust fault detection of a tunnel furnace.

show abstract

A GMDH neural network-based approach to robust fault diagnosis: Application to the DAMADICS benchmark problem

Cited by 111 publications

References 12 publications

Faults diagnosis using self-organizing maps: A case study on the DAMADICS benchmark problem

Faults diagnosis using self-organizing maps: A case study on the DAMADICS benchmark problem

Towards Robustness in Neural Network Based Fault Diagnosis

An unscented Kalman filter in designing dynamic GMDH neural networks for robust fault detection

Contact Info

Product

Resources

About