A gas turbine is usually installed inside a package to reduce the acoustics emissions and protect against adverse environmental conditions. An enclosure ventilation system is keeps temperatures under acceptable limits and dilutes any potentially explosive accumulation of gas due to unexpected leakages. The functional and structural integrity as well as certification needs of the instrumentation and auxiliary systems in the package require that temperatures do not exceed a given threshold. Moreover, accidental fuel gas leakages inside the package must be studied in detail for safety purposes as required by ISO21789. CFD is routinely used in BHGE (Baker Hughes, a GE Company) to assist in the design and verification of the complete enclosure and ventilation system. This may require multiple CFD runs of very complex domains and flow fields in several operating conditions, with a large computational effort. Modeling assumptions and simulation set-up in terms of turbulence and thermal models, and the steady or unsteady nature of the simulations must be carefully assessed. In order to find a good compromise between accuracy and computational effort the present work focuses on the analysis of three different approaches, RANS, URANS and Hybrid-LES. The different computational approaches are first applied to an isothermal scaled-down model for validation purposes where it was possible to determine the impact of the large-scale flow unsteadiness and compare with measurements. Then, the analysis proceeds to a full-scale real aero-derivative gas turbine package. in which the aero and thermal field were investigated by a set of URANS and Hybrid-LES that includes the heat released by the engine. The different approaches are compared by analyzing flow and temperature fields. Finally, an accidental gas leak and the subsequent gas diffusion and/or accumulation inside the package are studied and compared. The outcome of this work highlights how the most suitable approach to be followed for industrial purposes depends on the goal of the CFD study and on the specific scenario, such as NPI Program or RQS Project.

Gas turbine packages require a ventilation system in order to keep temperatures under acceptable limits and to dilute any hazardous accumulation of gas due to unexpected leakages. As part of the design phase, a detailed CFD (Computational Fluid Dynamics) analysis is performed on the complete system to assess the fluid dynamic behavior of the flow in terms of flow path, temperature distributions and velocity field. In this work, as additional approach, a detailed experimental 3D assessment of an entire aero-derivative gas turbine (GT) package was performed by creating a scale model (1:8) of the real configuration. The original package can be as much as 60 m3 in volume where details from pipes to valves can create severe flow distortions and the 3D CFD study might not necessarily include all details up to such level during the design phase. The scale model, built using sintered plastic material through rapid prototyping, was used for a test campaign reproducing the operation of the ventilation system, copying the dynamic similarity of the real scale. The model was equipped with a set of instruments to acquire measurements of pressure and velocity in several locations and at different flow rates. A significant benefit of using a scale model with transparent plexiglass for the external structure of the enclosure and ventilation ducts walls, was that it allowed to carry out a smoke test. This has been done by injecting a visible gas from several locations allowing the visualization of the streak lines of the local flow field. The aim of this approach was to find a fast and reliable way to investigate in detail complex phenomena such as gas leakage dilution and local flow distribution. A good agreement between experimental and computational data was found confirming that the CFD studies currently performed during the standard design phase are accurate and reliable enough to provide a proper prediction of the flow field inside the entire package even when a high level of details is included.

Ventilation systems are used in gas turbine packages to control the air temperature, to protect electrical instrumentation and auxiliary items installed inside the enclosure and to ensure a proper dilution of potentially dangerous gas leakages. These objectives are reached only if the ventilation flow is uniformly distributed in the whole volume of the package, providing a good air flow quality as prescribed by international codes such as ISO 21789. To evaluate the effectiveness of the ventilation design, numerical computations are performed for several purposes, one of which is the identification of poorly ventilated portions of the enclosure. In fact, it is essential to accurately detect the regions which are less ventilated, since they could be prone to the accumulation of an accidental fuel gas leak. There are different approaches to identify these portions, such as decay regression or inlet source analysis, that require unsteady simulations of the flow field inside the package. The present work discusses the implementation of a new methodology using machine learning and artificial neural networks (ANN) to detect the poorly ventilated regions where a gas cloud can accumulate. The concentration of fuel gas is estimated starting from a steady-state computation without running a more expensive unsteady computation. The entire process is built around an accurate training of the ANN using a proper set of simpler test-cases that have been identified to match the characteristics of the gas turbine enclosure. During the training phase accuracy and overfitting of the ANN were monitored to ensure robustness of the method. The procedure is then applied to a real case scenario and the results are presented in this paper highlighting the main advantages of this approach respect to a conventional use of CFD analysis. Computations of the flow fields are carried out using OpenFOAM with RANS and U-RANS approaches, while the ANN is developed and trained in Python.

Gas turbines usually are installed inside an enclosure, which is used as protection from the external environment and to provide an acoustic insulation. A ventilation system is required to control the temperature inside the enclosed volume and to dilute any potential gas leakage that may come from faulty pipes or flanges. The system has to be properly designed to avoid any unexpected explosion which would generate an overpressure not contained by enclosure walls. The most common approach to predict the effectiveness of the ventilation system requires to perform CFD analyses, which are very expensive in computational terms. A new approach has been proposed by authors, using machine learning and artificial neural networks (ANN) to identify the poorly ventilated zones. This methodology has been further developed, optimized and applied to a real gas turbine packages of new generation. In the present paper the authors will show the application of this procedure to the LM9000 package and the comparison with the results predicted using conventional CFD techniques. The tangible improvement introduced by this methodology is that the computational time is reduced from about three weeks with the common CFD approach to few minutes. The artificial neural network is developed in a Python environment that is applied during the CFX post-process phase of a steady state CFD simulation, providing results equivalent to unsteady CFD simulation. Besides the immediate benefits of this particular application, the suggested approach looks to be a great candidate to substitute the conventional and time-consuming CFD simulations with a fast post-processing algorithm that is able to learn and self-optimize as long as it is used.

Design of gas turbine packages is subjected to safety issues and the related guidelines are provided by ISO-21789. According to this code, the ventilation system shall guarantee a good and safe dilution in case of an unexpected gas leakage from components of the fuel gas system inside the enclosure. The evaluation of the dilution is commonly carried out by CFD simulations and the ISO-21789 indicates the criteria to evaluate the danger of a gas leak by estimating the cloud volume of the explosive mixture. To follow this prescription and to properly calculate the exact volume cloud, it is fundamental to accurately reproduce the fuel gas leak, which is always a supersonic jet of fuel gas into an air-ventilated domain. The main criticality is to simulate a supersonic jet into a complex domain such as the gas turbine package, considering the industrial goals in terms of accuracy and time constraints. The complexity is due to the geometry of the package and to the multiple locations where the leakage could occur. In such context, it is preferable to develop an advanced modeling of the phenomenon rather than simply improve the detail of the CFD, that could turn out to be unfeasible for industrial goals. For this reason, the authors present a series of simulations of under-expanded jets at high pressure ratios carried out to investigate the applicability of the sonic source approach to not-round jets.