Efficient use of three-dimensional CFD (Computational Fluid Dynamics) and CHT (Conjugate Heat Transfer) analyses is becoming increasingly critical, in order to compress the development process required for either on-going development, or the design of new large-bore engines for power generation applications.
Engine performance and reliability targets force engine developers to progressively refine and screen design iterations, from the conceptual stage up to the design-freeze, by means of CAE (Computer Aided Engineering) methods, which have to be accurate, robust and cost-effective, in order for them to effectively contribute to the product design. An efficient deployment of these tools usually requires extensive efforts to consolidate the analysis procedures and allow loosening of particular accuracy requirements, in favor of a shorter overall turn-around time. Finally, validation of the models against measurements enables the definition of best-practice guidelines for future programs.
The aim of this paper is to summarize the three-dimensional thermal-fluid simulation methodologies developed in GE’s Distributed Power business, for supporting the design of reciprocating engine cylinder-heads. The work reviews the two main types of analyses which are carried out during the defined development process.
Isothermal simulations are performed in order to estimate the flow-field velocities in the cylinder-head water jacket, without modeling the wall heat-transfer. If backed up by consolidated guidelines, they can be highly efficient for down-selecting design variants, simply by looking at bulk results, with minimum requirements in terms of turbulence modeling.
On the other end of the complexity spectrum, CHT simulations are used to model the thermal behavior of the cylinder-head assembly, by coupling and solving at run-time the RANS (Reynolds Averaged Navier-Stokes) equation set in the fluid domain and Fourier’s equation for the heat-transfer in the solid domain.
The challenging timeline associated to a new engine development program induced the authors to carefully review and adapt, on a case-by-case basis, general CFD best-practice guidelines for near-wall turbulence modeling, well-established in the CFD community. This mitigation was driven by the high complexity of a typical water-jacket geometry and by a number of uncertainties in the real-world application, related to manufacturing tolerances, material properties and operating conditions, which should be considered in order to find the optimal trade-off between absolute accuracy and computational costs.
Verification and validation CHT test-cases were carried out in order to support this approach. In particular, a comparison between the predicted CHT temperature solution and thermocouple measurements, performed on a GE Jenbacher engine, is described, in order to check the effectiveness of the proposed methodology and identifying opportunities for future developments.