Nozzle design for an organic vapor wind tunnel

Andrea Spinelli, V. Vandecauter, Alberto Guardone, Vincenzo DossenaProceedings of ... European Conference on Turbomachinery Fluid Dynamics & Thermodynamics 2011 · DOI

TL;DR

This paper presents the design of specialized nozzles for a new "dense gas" wind tunnel aimed at studying the behavior of organic fluids near their critical points. These nozzles are crucial for accurately simulating the complex, non-ideal gas flows found in Organic Rankine Cycle (ORC) engines, which are vital for efficient energy recovery. The design methodology, based on the Method of Characteristics and advanced real-gas thermodynamic models, yields nozzles significantly different from those predicted by simpler ideal-gas assumptions.

What Problem Does This Paper Address?

Organic Rankine Cycles (ORCs) are increasingly important for converting low-to-medium temperature heat sources (like waste heat or renewables) into electricity. However, the efficiency of ORC turbines is limited by the complex "dense gas" behavior of organic working fluids, especially when they operate near their critical points. This behavior deviates significantly from ideal gas assumptions, leading to challenges in accurately predicting and designing turbine components. Specifically, the flow within turbine blades can become highly supersonic, and the real-gas effects can drastically alter flow properties like the speed of sound and the fundamental derivative of gasdynamics. This makes it difficult to design efficient turbomachinery. To address this, the authors are building a novel test rig, the "Test Rig for Organic Vapors" (TROVA), to experimentally investigate these dense gas flows. A key component of this rig is a converging-diverging Laval nozzle designed to create controlled subsonic-to-supersonic expansion of these organic fluids. This paper focuses on the design of these critical nozzles.

Methods

The core of the paper is the design of the supersonic section of the Laval nozzle. The authors employ the standard Method of Characteristics (MOC), a well-established technique for designing supersonic nozzles. What's novel here is the integration of state-of-the-art multi-parameter thermodynamic models, specifically the Span-Wagner models for fluids MDM and R245fa. These models accurately capture the "real-gas" behavior of fluids near their critical points, which is crucial for dense gas flows.

The design process involves:

  1. Solving the transonic potential equation using an approximate method (Sauer's method) to obtain initial conditions at the nozzle throat.
  2. Applying the MOC to design the divergent section. This involves iteratively determining the nozzle wall geometry by imposing uniform Mach number and flow direction at the exit, while ensuring mass flow conservation.
  3. Crucially, the fluid properties (like speed of sound and the fundamental derivative of gasdynamics) are calculated using the real-gas models at each step, rather than assuming constant specific heats as in ideal gas models.

The authors compare their real-gas nozzle designs with those obtained using a simplified ideal-gas model to highlight the differences.

Datasets

The paper does not use pre-existing datasets for training or evaluation in the machine learning sense. Instead, it relies on:

  • Thermodynamic property data generated by the Span-Wagner models for MDM and R245fa. These models are based on extensive experimental data and theoretical understanding of these fluids.
  • Experimental operating conditions (pressure, temperature, fluid type) defined for four specific cases (A, B, C, D) representing typical ORC scenarios, including both sub-critical and super-critical operations. These conditions are summarized in Table 1.

Key Insights and Ideas

  • Dense gas effects are significant: The paper strongly emphasizes that organic fluids near their critical points behave very differently from ideal gases. This non-ideal behavior leads to substantial differences in nozzle geometry and flow characteristics.
  • Nozzle length is dramatically increased: A key finding is that real-gas nozzles designed for dense fluids are significantly longer (up to three times) than their ideal-gas counterparts for the same expansion ratio. This is a critical consideration for practical turbomachinery design.
  • Speed of sound behavior is non-monotonic: Unlike ideal gases where the speed of sound generally decreases with expansion, in these dense gas flows, the speed of sound can increase along much of the expansion process. This counter-intuitive behavior profoundly impacts the flow dynamics and nozzle design.
  • Fundamental derivative of gasdynamics is key: The paper highlights the importance of the fundamental derivative of gasdynamics (Γ) in characterizing real-gas flows. Its deviation from the ideal-gas value (typically (γ+1)/2) is directly linked to the observed differences in flow behavior and nozzle shape.
  • Ideal gas models are insufficient: The direct comparison between real-gas and ideal-gas designs clearly shows that using ideal-gas assumptions would lead to inaccurate and potentially non-functional nozzle geometries for these applications.

Key Results

  • The paper successfully designs four converging-diverging Laval nozzles for the TROVA test rig, catering to specific operating conditions with MDM and R245fa fluids.
  • For design case A (MDM), the real-gas nozzle is approximately 31 Ht long (where Ht is the throat semi-height), whereas an ideal-gas model would predict a much shorter nozzle.
  • The pressure, temperature, and Mach number profiles along the nozzle axis are presented for the designed cases, showing distinct differences between real-gas and ideal-gas predictions (e.g., Figure 5 and 7).
  • The fundamental derivative Γ is shown to be significantly lower than the ideal-gas value in the dense gas region, impacting the expansion rate.

Limitations and Future Work

  • Negligible viscous and thermal boundary layers: The design assumes that viscous and thermal boundary layer thicknesses are negligible compared to the nozzle height. This simplification might not hold true for all operating conditions or very small nozzles.
  • Quasi-1D flow assumption: While the Method of Characteristics is used, the initial transonic flow solution and the overall design are based on quasi-one-dimensional (1D) flow assumptions.
  • Unsteady flow effects: The TROVA facility operates in unsteady conditions due to the emptying of the high-pressure reservoir. While the nozzle design is based on initial steady-state conditions, the paper acknowledges that the effects of flow unsteadiness, including the start-up process and reservoir emptying, will be investigated in future work.
  • Off-design conditions: The current design focuses on specific operating points. Future work will investigate off-design performance and the impact of flow unsteadiness.

Summary updated

Abstract

A novel test rig for investigating the real-gas behavior of organic fluids operating at subsonic-supersonic speed in the proximity of the liquid-vapor critical point and the saturation curve is currently under construction at the Politecnico di Milano, Italy. This is a blow-down facility in which an organic vapor is expanded from a high-pressure reservoir kept at controlled super-heated or super-critical conditions into a low-pressure reservoir, where the vapor is condensed and pumped back into the high-pressure reservoir. Expansion to supersonic speeds occurs through a converging-diverging Laval nozzle. The present paper documents the design of the supersonic portion of the nozzle by means of the standard method of characteristics, that is coupled to the state-of-the-art multi-parameter Span-Wagner thermodynamic models of fluids MDM and R245fa. Four different operating conditions are considered and the resulting nozzle designs are thoroughly discussed. Keywords: Organic Rankine Cycle, Real-gas wind tunnel, Gasdynamics nozzles
BibTeX
@inproceedings{spinelli2011nozzle,
  author = {Andrea Spinelli and V. Vandecauter and Alberto Guardone and Vincenzo Dossena},
  title = {{Nozzle design for an organic vapor wind tunnel}},
  booktitle = {Proceedings of ... European Conference on Turbomachinery Fluid Dynamics \& Thermodynamics},
  year = {2011},
  doi = {10.29008/etc2011-107},
}
Figure