• cavitation-induced flow instabilities
• backflow vortex cavitation
• analytical modeling
• system identification
• propulsion
• turbopump inducer

In space rockets, turbopumps used to power the propellant are crucial components of all primary propulsion concepts powered by liquid propellant engines. The crucial nature is due to the severe limitations associated with the design of extremely demanding suction, high power density, and dynamically stable machines capable of meeting pumping and reliability requirements. In typical turbopumps used in space rocket engines, an axial inducer is placed upstream of the centrifugal stages in order to improve suction performance and reduce the propellant reservoir. In particular, inducers currently employed in rocket propellant feed turbopumps almost invariably operate with cavitation. The occurrence of cavitation introduces further complications to the already complex internal flow that characterizes the non cavitating regime and often leads to the development of flow instabilities that can seriously degrade the performance of the machine or even cause its rapid failure.

To date, an analytical model capable of describing perfectly cavitation-induced flow instabilities and rotor dynamics problems in axial rocket inducers and turbopumps has not yet been developed. All research activities are currently based on experimental methods applied to scale models or on the use of numerical methods. As a matter of fact, some aspects concerning unsteady flow phenomena in rocket cavitating inducers and turbopumps are still partially understood and present some challenging aspects with regard to the identification of similarity criteria capable of effectively predicting the actual performance of the machine.
This thesis presents a novel approach introducing an analytical predicting model based on system identification techniques for the effective identification and characterization of one specific form of cavitation: Backflow Vortex Cavitation. The phenomenon concerns the cavitating bubbles and vortices that occur in the annular region of backflow upstream of the inlet plane when the pump is required to operate in a loaded condition below the design flow rate or near the so-called breakdown condition. The system identification techniques are based on the experimental data related to a past test campaign carried out at the ex-ALTA’s test facility named Cavitating Pump Test Facility (CPTF) and its upgraded versions.
The transducers readings used came from the family of DAPAMITO inducer, designed and manufactured by ALTA S.p.A (now SITAEL S.p.A).
After a brief review of the state of the art of modern rocket inducers and fundamentals of cavitation, the experimental apparatus is presented and discussed. The experimental apparatus represents the starting point for the definition of the data analysis techniques and methods used to process the transducers readings and that will be used as input by the identifier. Then, the analytical model, which is the second essential input of the identifier, is proposed and discussed. Afterward, the designed identifier is presented: the method uses the principle of maximum likelihood estimation (MLE) to identify the unknown parameters after a Bayesian estimator found the most correlated frequency between the theoretical model and the experimental data. The identifier implements an error evaluator that is the basis on which the figure of merit is built up for the uncertainty quantification of the estimated parameters. Finally, after simulating multiple relevant cases, both for design and off-design flow conditions, the results are presented.
The obtained features of the instability are in line with the experimental-based literature demonstrating the usability of the predicting model for the identification of the backflow vortex cavitation instability.