A maximum entropy based approach to fault diagnosis using discrete and continuous features

Xiaodong Zhang, David Jonathan Miller, Roger Xu, Chiman Kwan, Hongda Chen

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    This paper presents a new maximum entropy (ME) based hybrid inference engine to improve the accuracy of diagnostic decisions using mixed continuous-discrete variables. By fusing the complementary fault information provided by discrete and continuous fault features, false alarms due to misclassification and modeling uncertainty can be significantly reduced. Simulation results using a three-tank benchmark system have clearly illustrated the advantages of diagnostics based on mixed continuous-discrete variables. Moreover, in the presence of significant measurement noise, simulation results show that the proposed ME method achieves better performance than the support vector machine classifier.

    Original languageEnglish (US)
    Title of host publication6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006
    Pages438-443
    Number of pages6
    EditionPART 1
    StatePublished - Dec 1 2006
    Event6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006 - Beijing, China
    Duration: Aug 29 2006Sep 1 2006

    Publication series

    NameIFAC Proceedings Volumes (IFAC-PapersOnline)
    NumberPART 1
    Volume6
    ISSN (Print)1474-6670

    Other

    Other6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006
    CountryChina
    CityBeijing
    Period8/29/069/1/06

    Fingerprint

    Maximum entropy methods
    Inference engines
    Failure analysis
    Support vector machines
    Classifiers
    Entropy
    Uncertainty

    All Science Journal Classification (ASJC) codes

    • Control and Systems Engineering

    Cite this

    Zhang, X., Miller, D. J., Xu, R., Kwan, C., & Chen, H. (2006). A maximum entropy based approach to fault diagnosis using discrete and continuous features. In 6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006 (PART 1 ed., pp. 438-443). (IFAC Proceedings Volumes (IFAC-PapersOnline); Vol. 6, No. PART 1).
    Zhang, Xiaodong ; Miller, David Jonathan ; Xu, Roger ; Kwan, Chiman ; Chen, Hongda. / A maximum entropy based approach to fault diagnosis using discrete and continuous features. 6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006. PART 1. ed. 2006. pp. 438-443 (IFAC Proceedings Volumes (IFAC-PapersOnline); PART 1).
    @inproceedings{1f30312e297746eabd766a3dfaf80ea9,
    title = "A maximum entropy based approach to fault diagnosis using discrete and continuous features",
    abstract = "This paper presents a new maximum entropy (ME) based hybrid inference engine to improve the accuracy of diagnostic decisions using mixed continuous-discrete variables. By fusing the complementary fault information provided by discrete and continuous fault features, false alarms due to misclassification and modeling uncertainty can be significantly reduced. Simulation results using a three-tank benchmark system have clearly illustrated the advantages of diagnostics based on mixed continuous-discrete variables. Moreover, in the presence of significant measurement noise, simulation results show that the proposed ME method achieves better performance than the support vector machine classifier.",
    author = "Xiaodong Zhang and Miller, {David Jonathan} and Roger Xu and Chiman Kwan and Hongda Chen",
    year = "2006",
    month = "12",
    day = "1",
    language = "English (US)",
    isbn = "9783902661142",
    series = "IFAC Proceedings Volumes (IFAC-PapersOnline)",
    number = "PART 1",
    pages = "438--443",
    booktitle = "6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006",
    edition = "PART 1",

    }

    Zhang, X, Miller, DJ, Xu, R, Kwan, C & Chen, H 2006, A maximum entropy based approach to fault diagnosis using discrete and continuous features. in 6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006. PART 1 edn, IFAC Proceedings Volumes (IFAC-PapersOnline), no. PART 1, vol. 6, pp. 438-443, 6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006, Beijing, China, 8/29/06.

    A maximum entropy based approach to fault diagnosis using discrete and continuous features. / Zhang, Xiaodong; Miller, David Jonathan; Xu, Roger; Kwan, Chiman; Chen, Hongda.

    6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006. PART 1. ed. 2006. p. 438-443 (IFAC Proceedings Volumes (IFAC-PapersOnline); Vol. 6, No. PART 1).

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    TY - GEN

    T1 - A maximum entropy based approach to fault diagnosis using discrete and continuous features

    AU - Zhang, Xiaodong

    AU - Miller, David Jonathan

    AU - Xu, Roger

    AU - Kwan, Chiman

    AU - Chen, Hongda

    PY - 2006/12/1

    Y1 - 2006/12/1

    N2 - This paper presents a new maximum entropy (ME) based hybrid inference engine to improve the accuracy of diagnostic decisions using mixed continuous-discrete variables. By fusing the complementary fault information provided by discrete and continuous fault features, false alarms due to misclassification and modeling uncertainty can be significantly reduced. Simulation results using a three-tank benchmark system have clearly illustrated the advantages of diagnostics based on mixed continuous-discrete variables. Moreover, in the presence of significant measurement noise, simulation results show that the proposed ME method achieves better performance than the support vector machine classifier.

    AB - This paper presents a new maximum entropy (ME) based hybrid inference engine to improve the accuracy of diagnostic decisions using mixed continuous-discrete variables. By fusing the complementary fault information provided by discrete and continuous fault features, false alarms due to misclassification and modeling uncertainty can be significantly reduced. Simulation results using a three-tank benchmark system have clearly illustrated the advantages of diagnostics based on mixed continuous-discrete variables. Moreover, in the presence of significant measurement noise, simulation results show that the proposed ME method achieves better performance than the support vector machine classifier.

    UR - http://www.scopus.com/inward/record.url?scp=79961144071&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=79961144071&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:79961144071

    SN - 9783902661142

    T3 - IFAC Proceedings Volumes (IFAC-PapersOnline)

    SP - 438

    EP - 443

    BT - 6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006

    ER -

    Zhang X, Miller DJ, Xu R, Kwan C, Chen H. A maximum entropy based approach to fault diagnosis using discrete and continuous features. In 6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes, SAFEPROCESS 2006. PART 1 ed. 2006. p. 438-443. (IFAC Proceedings Volumes (IFAC-PapersOnline); PART 1).