Browsing by Keyword "Industrial prognosis"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item Adaptive Dendritic Cell-Deep Learning Approach for Industrial Prognosis under Changing Conditions(2021-11) Diez-Olivan, Alberto; Ortego, Patxi; Ser, Javier Del; Landa-Torres, Itziar; Galar, Diego; Camacho, David; Sierra, Basilio; Tecnalia Research & Innovation; IAIndustrial prognosis refers to the prediction of failures of an industrial asset based on data collected by Internet of Things sensors. Prognostic models can experience the undesired effects of concept drift, namely, the presence of nonstationary phenomena that affects the data collected over time. Consequently, fault patterns learned from data become obsolete. To overcome this issue, contextual and operational changes must be detected and managed, triggering rapid model adaptation mechanisms. This article presents an adaptive learning approach based on a dendritic cell algorithm for drift detection and a deep neural network model that dynamically adapts to new operational conditions. A kernel density estimator with drift-based bandwidth is used to generate synthetic data for a faster adaptation, focusing on fine-tuning the lowest neural layers. Experimental results over a real-world industrial problem shed light on the outperforming behavior of the proposed approach when compared to other drift detectors and classification models.Item Advanced Prognostics for a Centrifugal Fan and Multistage Centrifugal Pump Using a Hybrid Model(Springer Science and Business Media Deutschland GmbH, 2023) Vila-Forteza, Marc; Jimenez-Cortadi, Alberto; Diez-Olivan, Alberto; Seneviratne, Dammika; Galar-Pascual, Diego; Juuso, Esko; Galar, Diego; Tecnalia Research & Innovation; IAPredictive maintenance is fully implemented in the oil and gas industry, and the impressive development of field sensors, big data, and digital twins offers a wide field for the ongoing experimentation and development of diagnostic and prognostic tools for machinery. Although a wide range of technologies and sensors is available, vibration analysis remains the preferred predictive technique for rotating machinery diagnostics. It is well-known, widely used, and has proven efficacious in evaluating the health of rotating machinery and preventing failures. Taking advantage of vibration analysis development and computing capabilities, this study develops three digital twins of one multistage centrifugal pump and two centrifugal fans using real vibration data and synthetic data. This hybrid model approach permits the use of failure data which are not usually found in the normal operation of these machines. The study improves and tunes the accuracy of those models using real operating data obtained from a distributed control system (DCS), thus obtaining results in accordance with process conditions. Maintenance decisions can be supported by these models. They are based on online vibration and process data; they diagnose the health of a machine and give its remaining useful life (RUL). The models may also be used for other API plant assets (multistage centrifugal pumps or centrifugal fans) by changing the configuration parameters and process DCS tags.