Deep Neural Networks-Based Fault Diagnosis Model For Process Systems

Created by W.Langdon from gp-bibliography.bib Revision:1.8414

@InCollection{Shahab:2024:ESCAPEISPSE,

• author = "Mohammad Shahab and Zoltan Nagy and Gintaras Reklaitis",
• title = "Deep Neural Networks-Based Fault Diagnosis Model For Process Systems",
• booktitle = "34th European Symposium on Computer Aided Process Engineering / 15th International Symposium on Process Systems Engineering",
• publisher = "Elsevier",
• year = "2024",
• editor = "Flavio Manenti and Gintaras V. Reklaitis",
• volume = "53",
• series = "Computer Aided Chemical Engineering",
• pages = "1963--1968",
• keywords = "genetic algorithms, genetic programming, fault diagnosis, deep neural networks, ANN, feature engineering, pharmaceutical process",
• ISSN = "1570-7946",
• URL = "https://www.sciencedirect.com/science/article/pii/B9780443288241503288",
• DOI = "doi:10.1016/B978-0-443-28824-1.50328-8",
• abstract = "The diagnosis of faults is crucial to ensure process safety and increased product quality. Faults can emerge with time across different unit operations due to changes in the process that the process controllers are unable to handle appropriately. This undesirable divergence in the variables of the system is found to adversely affect the product quality in process industries. In this work, a deep neural network (DNN) model driven by feature engineering on the process dataset using genetic programming is developed to classify faults in a process system. Feature extraction and construction using process data is carried out before the transformed features are used for fault diagnosis in a DNN. The DNN model performs fault diagnostics on the process data that contains the normal operating conditions and the abnormal operating conditions which arise due to variations in the characteristic quality of the system. The genetic programming driven DNN methodology is illustrated on a benchmark chemical process, where its effectiveness is evaluated by classifying faults in the Tennessee Eastman Process (TEP) and is compared against existing methodologies",

}

Genetic Programming entries for Mohammad Shahab Zoltan Nagy Gintaras Reklaitis

Citations