Model measurement based identification of Francis turbine vortex rope parameters for prototype part load pressure and power pulsation prediction

28th IAHR symposium on Hydraulic Machinery and Systems

Manderla, M.^1 ; Weber, W.^1 ; Koutnik, J.^1

Voith Hydro Holding GmbH and Co. KG, Alexanderstr. 11, Heidenheim

89522, Germany^1

关键词: Dynamic measurement;    Future applications;    Non-dimensional parameters;    Operating condition;    Part load operation;    Power oscillations;    Pressure excitation;    System identification methods;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/49/8/082008/pdf
DOI  :  10.1088/1755-1315/49/8/082008

来源: IOP

【 摘 要 】

Pressure and power fluctuations of hydro-electric power plants in part-load operation are an important measure for the quality of the power which is delivered to the electrical grid. It is well known that the unsteadiness is driven by the flow patterns in the draft tube where a vortex rope is present. However, until today the equivalent vortex rope parameters for common numerical 1D-models are a major source of uncertainty. In this work, a new optimization-based grey box method for experimental vortex rope modelling and parameter identification is presented. The combination of analytical vortex rope and test rig modelling and the usage of dynamic measurements allow the identification of the unknown vortex rope parameters. Upscaling from model to prototype size is achieved via existing nondimensional parameters. In this work, a new experimental setup and system identification method is proposed which are suitable for the determination of the full set of part load vortex rope parameters in the lab. For the vortex rope, a symmetric model with cavity compliance, bulk viscosity and two pressure excitation sources is developed and implemented which shows the best correspondence with available measurement data. Due to the non-dimensional parameter definition, scaling is possible. This finally provides a complete method for the prediction of prototype part-load pressure and power oscillations. Since the proposed method is based on a simple limited control domain, limited modelling effort and also small modelling uncertainties are some major advantages. Due to the generality of the approach, a future application to other operating conditions such as full load will be straightforward.