Omitir los comandos de cinta
Saltar al contenido principal
Inicio de sesión
Universidad EAFIT
Carrera 49 # 7 sur -50 Medellín Antioquia Colombia
Carrera 12 # 96-23, oficina 304 Bogotá Cundinamarca Colombia
(57)(4) 2619500
EAFITProgramas académicosPosgradosMaestría en IngenieríaMaestría en Ingeniería / EstudiantesNon Intrusive Detection of Rotating Stall Occurring in the Turbine Quadrant of Francis Type Pump Turbines

Non Intrusive Detection of Rotating Stall Occurring in the Turbine Quadrant of Francis Type Pump Turbines

Francisco Javier Botero Herrera


The so called rotating stall (RS) is an undesired hydrodynamic phenomenon affecting the regular functionof Francis type  pump turbines. Stronger noise and vibration increase the  risk of catastrophic failures and economic loses. In the past,  such phenomenon was considered negligible and the  search  for  a  definitive  solution  was  not  a  priority.
Today, since pump turbines have been recognized as important agents in the  global ecological stage, the  interest in RS is renewed. Scientist search  for  answers,  mainly  in  two fields: computational  models  and  laboratory  testing;  but neither has been  successfully  transferred to  prototypes. 
In  the  first  case, prediction methods have not reached the accuracy required to be considered in practical implementations; and in  the  second  case,  experimental approaches remain limited to reduced  scale  models  and laboratory testing conditions. To compensate for the lack of reliable prediction methods, to simplify the conventional experimental arrangements, and of course, to extend the application range from scale models to prototypes, three non-intrusive RS detection approaches have been developed. In fact, they allow detection  of  RS  by  examining  shaft  dynamics,  structure and  airborne  noise  with  a  reduced number of instruments, and specific signal processing. For validation purposes, a reduced scale pump turbine model featuring RS is instrumented with accelerometers, proximity probes, microphones and pressure sensors; additionally, high speed visualization and specific image processing techniques are used as well. The model is tested in turbine rotation.
It is initially set to a high efficiency and RS free operating point. Then, rotating speed is increased up to runaway and turbine brake when RS develops. Pressure  sensors  and tuft visualization are  employed  to confirm the occurrence of the phenomenon whilst accelerometers, microphones and proximity probes are used to non intrusively detect RS phenomena.