Computational investigation of probe influence on total pressure measurements in a supersonic turbine linear cascade for ORC applications

The present thesis aims to carry out a numerical investigation through Computational Fluid Dynamics (CFD) to assess the influence of a total pressure probe, inserted downstream a supersonic linear cascade operated with organic fluids, on the flow and the measurement accuracy. The fluid domain and operating conditions considered for this analysis are extracted from an experimental campaign performed on the Test Rig for Organic Vapors (TROVA), a blow-down wind tunnel located at the Laboratory of Compressible fluid dynamics for Renewable Energy Applications (CREA Lab) of Politecnico di Milano. The main purpose of this set of experiments was the investigation of the resulting flow field in a linear cascade that aimed to be representative of an Organic Rankine Cycle (ORC) supersonic turbine stator. ORCs, due to their ability to recover heat from low-temperature sources and their minimal greenhouse gas emissions, are a promising technology in the framework of sustainable power generation. The organic working fluid, however, due to its behavior that highly deviates from ideal gas dynamics, poses many challenges to the optimization of the performances of ORC turbines which are the most crucial components of the cycle, heavily affecting the overall efficiency. This led to a considerable growth in research interest regarding ORC turbine expanding flows, in the last two decades. Nevertheless, a lack of data on such flows urged the need of carrying out experimental investigations on purposely designed cascades to better analyze the non-ideal flow behavior. Being total pressure losses a crucial parameter in the evaluation of turbomachinery efficiency, the measurement of this quantity is a critical factor in the assessment of ORC turbine performances. Total pressure measurements are complicated in supersonic flow by the generation of a bowed shock in front of the probe head, making it necessary to retrieve the pre-shock value by exploiting normal shock equations. Since most of the previous numerical studies did not consider the presence of probes, a specific analysis was needed to evaluate the probe effects in non-ideal supersonic flow measurements. Four specific tests from the aforementioned campaign were considered to characterize the flow in presence of the probe under specific operating conditions. The shock pattern around the probe was investigated by means of Schlieren visualizations, and pressures were acquired and post-processed to be later compared to the numerical results. The numerical study was done by employing ICEM CFD and ANSYS Fluent respectively as the meshing tool and the solver. A simplified geometry of converging-diverging nozzle was investigated first to observe the effect of probe presence and probe misalignment on the measured total pressure. It was demonstrated that the probe, even under 5° of misalignment, still generates a normal shock without disturbing the approaching flow. Shifting to the geometry of the actual test section, a comprehensive numerical study was carried out, including sixteen case studies where the probe was analyzed in two different positions (wake and free stream), under different inclination angles, for both nitrogen and MM. In this way the effects of different factors on measurements accuracy were separately studied. The results were, then, compared with the corresponding experiments. It was found that even in presence of probe misalignment, free stream measurements remain almost unchanged for both nitrogen and MM, while high uncertainties were observed in the wake position. These uncertainties tend to rise in case of MM due to the specific wake shape. Moreover, the shock equation is prone to high error propagation when the shock ahead of the probe becomes less normal. This situation can occur in case of misalignment when the probe is placed in wake, both for nitrogen and MM, leading to wrong predictions of total pressure loss.