The Machine Learning methods require a large data sets for model training and it often causes a problem in a real application. Usage of a model based approach to generate the data can be a way to generate data in various degrees of malfunctions, e.g. with different sizes of unbalance. The aim of this paper is to demonstrate the applicability of a 1D rotor-bearing model to reproduce the unbalance conditions during rotors coast-down operation. The model parameters adjustment case study is focused on an application of the model in order to reproduce rotor unbalance conditions of a 200 MW steam turbine. The rotor coast-down operation is considered to reduce external forces related to start-up or steady-state rotor operation. This allows to reduce and in turn clearly isolate unbalance conditions. The developed 1D model consists of first-principle rotor motion equations along the hydrodynamic bearing-support and foundation equations. The gray-box approach was applied to reduce the number of parameters required to be adjusted during system identification process. The rotor geometry and related mass-stiffness parameters were derived from the bearing-rotor assembly drawings while other phenomenological parameters were adjusted based on the measurements to obtain a good correlation with amplitude of vibrations at measurement locations along the shaft line.
CITATION STYLE
Barszcz, T., Czop, P., & Zabaryłło, M. (2020). Development and Tuning of a Simplified 1D Model for Generation of Transient States in Large Turbomachinery. In Lecture Notes in Mechanical Engineering (pp. 541–554). Pleiades Publishing. https://doi.org/10.1007/978-981-13-8331-1_40
Mendeley helps you to discover research relevant for your work.