Modelling of Contact Dynamics in Kaplan Turbines
2024 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE credits
Student thesis
Abstract [en]
Hydropower is the single largest renewable way of producing electricity in Sweden, with the Kaplan turbine being the most common type of turbine used for production. It is also the focus of this master’s thesis project which is carried out on behalf of the state-owned Swedish energy company Vattenfall. Between its blades and the surrounding discharge ring, the Kaplan turbine has a gap that is small enough to introduce a risk of contact between the two parties, possibly leading to dire consequences as a result. In addition to hefty repair costs, the consequences of contact can in the worst case scenario be operation disruption and thereby production losses.Contact can be a result of general rotor unbalance and uneven pressure distribution from blockage in the surrounding spiral casing. It is a well-studied phenomenon currently studied using various 2D models that are constrained to in-plane movement. While 3D models have no such constraint and are superior in terms of capturing a complex geometry and its effect on contact behaviour, they are costly to simulate. Hence, it is important to be able to evaluate a course of contact through simulation and finding the optimal way to do so, with respect to accuracy and computational cost. Therefore, simulation results from two 2D FEM models in the form of self-written routines in MATLAB have been compared to those from an idealised real 3D runner geometry FEM analysis of the Porjus U9 runner in LS-DYNA, to determine which way of modelling contact is the most advantageous and which simplifications are justified.The two 2D models used were the overhung rotor model and an expanded rotor model. The latter describes the complete appearance of an actual hydropower unit, while the former can be considered to represent the bottom part of such a unit. They are both comprised of lumped masses supported by elastic beam elements and thus, they could be described by the so-called equation of motion and solved numerically through time integration techniques in MATLAB by utilising state-space modelling. The 3D model was created with the purpose of replicating the expanded rotor model, by first idealising the 3D runner geometry in Siemens NX and then exporting it to LS-DYNA for further modelling.Focus was put on studying the resultant contact force as well as Fast Fourier Transforms, or FFTs, of the obtained resultant displacement and contact force data for information about the general contact behaviour. The simulations were run at a typical operating speed resembling normal operation and a runaway speed resembling a worst case scenario, with and without the presence of friction in the contact interface, for a simulation time of 60 seconds.The results indicated a more dynamic behaviour for the 3D model compared to the 2D models, which was shown, in part, as a contact force with a higher and varying amplitude. The effect of friction in the contact interface was also more noticeable for the 3D model. While the FFT of the displacement data did not reveal much in the way of differences in contact behaviour between the 3D model and the 2D models, the FFT of the contact force data did. An intricate contact behaviour was revealed for the 3D model compared to that of the 2D models, possibly owing to the ability of the runner in the 3D model to move out-of-plane which resulted in a change in its contact geometry. This is one of the key arguments for using a 3D model over a 2D model and it was shown when contact at more than one point on the runner in the 3D model simultaneously was discovered.Model simplification is possible in the context of 3D modelling in order to make it more relevant compared to 2D models with respect to computational cost, with one such example being the discharge ring modelled as a round, cylindrical and rigid analytical surface. Due to the uncertainty in the magnitude of the obtained results, however, it is difficult to gauge the accuracy of a 3D model and how much better it is compared to a 2D model, as well as if the further model simplification that is the 2D models is justifiable or not in that regard. Though further research on 3D modelling practices for hydropower rotor dynamics needs to be conducted, it can be stated that a 3D model does indeed capture a complex geometry and its effect on contact behaviour in a way that a 2D model simply lacks an answer for, which on its own might justify the use of one depending on the application in question.
Place, publisher, year, edition, pages
2024. , p. 41
Keywords [en]
Contact Dynamics, Kaplan Turbines, Modelling, Hydropower, Rotor Dynamics
Keywords [sv]
Kontaktdynamik, Kaplanturbiner, Modellering, Vattenkraft, Rotordynamik
National Category
Mechanical Engineering
Identifiers
URN: urn:nbn:se:ltu:diva-108443OAI: oai:DiVA.org:ltu-108443DiVA, id: diva2:1886385
Educational program
Mechanical Engineering, master's level
Supervisors
Examiners
2024-08-072024-07-312025-10-21Bibliographically approved