Abstract Near-wall modelling is one of the most challenging aspects of CFD computations. In fact, integration-to-the-wall with low-Reynolds approach strongly affects accuracy of results, but strongly increases the computational resources required by the simulation. A compromise between accuracy and speed to solution is usually obtained through the use of wall functions, especially in RANS computations, which normally require that the first cell of the grid to fall inside the log-layer (50 < y + < 200) [1]. This approach can be generally considered as robust, however the derivation of wall functions from attached flow boundary layers can mislead to non-physical results in presence of specific flow topologies, e.g. recirculation, or whenever a detailed boundary layer representation is required (e.g. aeroacoustics studies) [2]. In this work, a preliminary attempt to create an alternative data-driven wall function is performed, exploiting artificial neural networks (ANNs). Whenever enough training examples are provided, ANNs have proven to be extremely powerful in solving complex non-linear problems [3]. The learner that is derived from the multi-layer perceptron ANN, is here used to obtain two-dimensional, turbulent production and dissipation values near the walls. Training examples of the dataset have been initially collected either from LES simulations of significant 2D test cases or have been found in open databases. Assessments on the morphology and the ANN training can be found in the paper. The ANN has been implemented in a Python environment, using scikit-learn and tensorflow libraries [4][5]. The derived wall function is implemented in OpenFOAM v-17.12 [6], embedding the forwarding algorithm in run-time computations exploiting Python3.6m C_Api library. The data-driven wall function is here applied to k-epsilon simulations of a 2D periodic hill with different computational grids and to a modified compressor cascade NACA aerofoil with sinusoidal leading edge. A comparison between ANN enhanced simulations, available data and standard modelization is here performed and reported.

Abstract Additive manufacturing represents a new frontier in the design and production of rotor machines. This technology drives the engineering research framework to new possibilities of design and testing of new prototypes, reducing costs and time. On the other hand, the fast additive manufacturing implies the use of plastic and light materials (as PLA or ABS), often including a certain level of anisotropy due to the layered deposition. These two aspects are critical, because the aero-elastic coupling and flow induced vibrations are not negligible for high aspect ratio rotors. In this work, we investigate the aeroelastic response of a small sample fan blade, printed using PLA material. Scope of the work is to study both the structure and flow field dynamics, where strong coupling is considered on the simulation. We test the blade in two operating points, to see the aero-mechanical dynamics of the system in stall and normal operating condition. The computational fluid-structure interaction (FSI) technique is applied to simulate the coupled dynamics. The FSI solver is developed on the base of the finite element stabilized formulations proposed by Tezduyar et al. We use the ALE formulation of RBVMS-SUPS equations for the aerodynamics, the non-linear elasticity is solved with the Updated Lagrangian formulation of the equations of motion for the elastic solid. The strong coupling is made with a block-iterative algorithm, including the Jacobian based stiffness method for the mesh motion.

Abstract Hybrid marine ship propulsion systems are currently being used, and many more are being considered worldwide, including for commercial ships as well as for navies worldwide. These hybrid propulsion systems offer extended ship range and potentially quiet operation with propulsion on propeller shaft line electric motors. When high power and / or speed may be required, the main engines or turbines can be quickly started, and propeller shaft line clutches provide quick and reliable transition from or back to the electric motors. In the naval marine sector, several ship programs currently use “Combined Diesel Electric or Gas Turbine” (CODELOG) or “Combined Diesel Electric and Gas Turbine” (CODELAG) propulsion. Some utilize Synchronous Self-Shifting Clutches, and some utilize other types of non-automatic friction clutches. Many navies worldwide are considering this type of hybrid propulsion, but variations in machinery arrangements and components can result in significant differences in fuel efficiency, weight, space, first time cost, ship performance, and propulsion machinery lifecycle maintenance with associated cost. Therefore, serious investigation of the alternatives early in the ship design phase is important and accessing the experience of these existing systems with these components is vital, before settling on a machinery arrangement and its components for a future ship program. As the old cliché says, “The Devil is in the Details!” Like any system, the success, the reliability, and the ultimate capability often come down to design details and the components chosen. Naval ships are designed to operate, even when damaged or when control or oil supply systems fail. Simplicity and the ability to improvise is often the key to mission success or failure. Synchronous Self-Shifting Clutches, or SSS Clutches, have proven to be “fit and forget” in these propulsion systems, have performed without fail in battle conditions, and this paper will describe the systems, explore the alternatives for propulsion clutches, provide advantages and disadvantages, and will provide experience of SSS Clutches in hybrid propulsion systems.

Abstract As a consequence of the increasing share of renewable energy sources in present-day electrical grid systems, time variations of the power demand for fossil fuel plants can become more sudden. Therefore, an ability to respond to sudden load changes becomes an important issue for power generation gas turbines. This paper describes a real-time model for predicting the transient performance of gas turbines. The method includes basic transient phenomena, such as volume packing and the heat transfer between the working fluid and the structural elements. The dynamics of components are quantified by solving ordinary differential equations with appropriate initial and boundary conditions. Compressor and turbine operating points are determined from corresponding performance maps previously calculated using sophisticated aerodynamic, through-flow codes. This includes a sufficient number of such characteristics to account for the variations in speed and machine geometry. The developed dynamic model was verified by comparison of simulation results with experimentally recorded operating parameters for a real engine. This includes the start-up sequence and the change of load. Additional simulation covers the system response to a step increase in fuel flow. The simulation is carried out faster than the real process.

Abstract Despite significant advancements in computational power and various numerical modeling in past decades, flow simulation of a multi-stage axial-flow compressor is still one of the most active areas of research, for it is the critical component in engine performance and operability, and there are so many elements that need to be looked into to predicting correct matching of the stages and accurate flow distribution inside the machine. Modeling unsteadiness, both deterministic and random types, and real geometries are among the most important features to be considered in such prediction. The authors have conducted in their previous studies a series of unsteady RANS (URANS) simulations of a 6.5-stage high-speed highly-loaded axial-flow compressor, and explored many unsteady effects as well as effects of real geometries such as Variable Stator Vane (VSV) clearance and inter-stage seal leakage flow on the compressor performance. However, all the analyses failed to predict correct stage matching, total pressure and temperature radial profiles, or mass-flow with adequate accuracies. In the present study, an Improved Delayed Detached Eddy Simulation (IDDES) with SST k-omega model is applied to the simulation of the same compressor configuration at aerodynamic design point. Fifth-order WENO scheme is employed for improved spatial accuracy to suppress significant increase in mesh size. Total number of mesh points are over 400 million for 1/10th sector model. Computations are ensemble averaged for 20 sector passage. Computed overall performance and flow field are compared with the compressor rig test data. The predictions of inter-stage total temperature radial profiles are noticeably improved over the URANS with the same mesh, discretization scheme and eddy turbulence model. Good comparison with the rig data indicates the current simulation is properly capturing the span-wise mixing phenomena. Unsteady flow field are compared between IDDES and URANS to locate the cause for the enhanced mixing. It is shown that components of Reynolds stress responsible for radial diffusion and anisotropic features are intensified in the tip leakage vortex at the rotor exit for the IDDES.

Abstract In a compressor used for power generation, a rotating stall due to an operating point mismatch between the front and rear stages occurs at front stages in low-speed condition during startup. IGV, VSV opening schedule and bleed flow rates are determined in order to obtain stable operating conditions, with the occurrence of pressure fluctuations or blade excitation force derived from the rotating stall. The setting of the number of stages for variable stator vane, the vane opening and the bleed flow rate are largely based on the rig test and experience of the actual machine, however we focused on the capability study for direct simulation with unsteady CFD for the rotating stall. If this can be predicted in advance, it makes it possible to reduce the number of stages of the variable stator vanes, the capacity of the thyristor, and to make the bleed chambers and piping compact, leading to cost reductions. The method of rotating stall prediction during startup condition, and the comparison between predicted and actual measurements on how this number changes with different starting conditions and different machines, are shown in this paper. CFD shows the different stall cell position, the number of stall cells and the pressure fluctuation level in each condition, and these results were consistent with the measurement data. Furthermore, we have found that these phenomena can be controlled by variable stator vane angle during start up.

Abstract The effects of blade row interactions on stator-mounted instrumentation in axial compressors are investigated using unsteady numerical calculations. The test compressor is an 8-stage machine representative of an aero-engine core compressor. For the unsteady calculations, a 180deg sector (half-annulus) model of the compressor is used. It is shown that the time-mean flow field in the stator leading edge planes is circumferentially non-uniform. The circumferential variations in stagnation pressure and stagnation temperature respectively reach 4.2% and 1.1% of the local mean. Using spatial wave number analysis, the incoming wakes from the upstream stator rows are identified as the dominant source of the circumferential variations in the front and middle of the compressor, while towards the rear of the compressor, the upstream influence of the eight struts in the exit duct becomes dominant. Based on three circumferential probes, the sampling errors for stagnation pressure and stagnation temperature are calculated as a function of the probe locations. Optimization of the probe locations shows that the sampling error can be reduced by up to 77% by circumferentially redistributing the individual probes. The reductions in the sampling errors translate to reductions in the uncertainties of the overall compressor efficiency and inlet flow capacity by up to 50%. Recognizing that data from large-scale unsteady calculations is rarely available in the instrumentation phase for a new test rig or engine, a method for approximating the circumferential variations with single harmonics is presented. The construction of the harmonics is based solely on the knowledge of the number of stators in each row and a small number of equi-spaced probes. It is shown how excursions in the sampling error are reduced by increasing the number of circumferential probes.

Abstract Counter-rotating turbomachines have the potential to be high efficiency, high power density devices. Comparisons between conventional and counter-rotating turbomachines in the literature make multiple and often contradicting conclusions about their relative performance. By adopting appropriate non-dimensional parameters, based on relative blade speed, the design space of conventional machines can be extended to include those with counter-rotation. This allows engineers familiar with conventional turbomachinery to transfer their experience to counter-rotating machines. By matching appropriate non-dimensional parameters the loss mechanisms directly affected by counter-rotation can be determined. A series of computational studies are performed to investigate the relative performance of conventional and counter-rotating turbines with the same non-dimensional design parameters. Each study targets a specific loss source, highlighting which phenomena are directly due to counter-rotation and which are solely due to blade design. The studies range from two-dimensional blade sections to three-dimensional finite radius stages. It is shown that, at hub-to-tip ratios approaching unity, with matched non-dimensional design parameters, the stage efficiency and work output are identical for both types of machine. However, a counter-rotating turbine in the study is shown to have an efficiency advantage over a conventional machine of up to 0.35 percentage points for a hub-to-tip ratio of 0.65. This is due to differences in absolute velocity producing different spanwise blade designs.

Abstract The adoption of very high bypass ratios is an effective strategy to improve the performance of turbofans for civil aviation. Very high-bypass ratio turbofans are characterized by large-diameter fans and large intake diameters, whereas the overall length of the engine and its installation have to be contained to prevent excessive weight growth. As a result, the main components of the low-pressure compression system (LPC) are closer to each other in modern and future engines than they were in less recent designs, if distances are measured in terms of engine diameter. As the axial decay of potential distortions takes place over distances comparable to their tangential wavelength, modern and future engines also display stronger interactions between the intake, the fan, the OGVs and the pylons, causing performance and integrity issues that need to be addressed at the design stage. The direct computation of whole LPC systems through CFD is expensive on account of the large extent of the domain and the wide range of wavelengths/frequencies that need to be resolved. This makes CFD unsuitable for bypass design. However, as the interaction between intake, fan, OGV and bypass takes place mainly through flow potential over long wavelengths, drastic simplifications can be introduced. Correspondingly, considerable computational savings are possible by resorting to semi-analytical models. In this Part I of a two-part paper, a model for the intake/fan/OGV bypass interaction problem based on potential flow is presented. The model can describe the potential component of the LPC flow field, as well as vorticity distortions transmitted through the fan and those generated by the fan as a result of the non-uniform work done around the circumference. The method is validated against CFD results on a configuration typical of modern machines for civil aviation service and is demonstrated as a useful preliminary design tool.

Abstract The flow losses in a U-bend and return channel system of an inter stage has an important influence on the stage performance. To study the flow characteristics in such system, a pseudo-stage which consists of pseudo guide vanes, U-bend and return channel, combined with a normal stage, was designed and its performance curves were measured at different machine Mach numbers. Simultaneously, the numerical simulations, in which different data processing methods of CFD post process were adopted, were performed and the prediction accuracy was verified by comparison with the experimental results. The results indicated that the design of pseudo guide vanes provides an approximately constant flow direction at the inlet of U-bend with the variation of inlet Reynolds number, which matches the blade metal angle of the return channel well. The further analyses on the measured and numerical results showed that the loss characteristics vary with the inlet Reynolds number. When the Reynolds number is greater than a critical Reynolds number, the loss coefficient in the U-bend and return channel keeps nearly constant. The non-uniformity of flow angle and total pressure at the outlet of return channel were also illustrated and discussed.

Abstract The performance of radial inflow turbines, and specifically of turboexpanders for oil & gas applications, has been traditionally described in terms of efficiency versus velocity speed ratio (U/C) and discharge flow coefficient (Q/N). Especially in the testing phase, this latter parameter has been often preferred to the angle setting of moveable inlet guide vanes (IGV), which are standard equipment for most turboexpanders. In practice, the expander U/C has been often considered to give the performance backbone, while the Q/N ratio has been used for secondary corrections. Moreover, although the role of pressure ratio (PR) is recognized, its impact has been experimentally unexplored in those cases where testing facilities had capacity limitations. Eventually, in case of variable nozzles, the inlet flow capacity curve has been rarely included among the output performance variables, being the attention mainly focused on efficiency. In the present paper, beside an overview and an explanation of the physical meaning of traditional performance parameters, an alternative approach based on torque mapping versus U/C is introduced and discussed in detail. As a matter of fact, numerical and experimental data show smooth and regular trends when torque coefficient is used instead of adiabatic efficiency. Moreover, performance based on torque coefficient can be more conveniently extrapolated at extreme off-design conditions such as start-up (locked rotor condition) or full speed no load. The ease of extrapolation is particularly important for machine operability, which often requires accurate modeling of transient missions at very partial loads (as for instance during start-up or shut-down). Examples will be offered to show the advantages of torque coefficient representation and how sensitive this is to IGV setting and pressure.

Abstract An additive manufactured (AM) vaned diffuser for use in a centrifugal compressor research facility was designed and implemented. Utilizing an AM process to manufacture the diffuser reduces the long lead time that is associated with conventionally manufactured diffusers, and it increases the instrumentation capabilities within the flow path. Several AM techniques and a variety of plastic and metal materials were evaluated for this application. A high-temperature, stereolithography (SL) resin was chosen because of the tight dimensional tolerances maintained by the SL process. Utilizing a high-temperature plastic also results in manufacturing costs that are significantly less than using a metal material. Samples of the chosen material were subjected to mechanical testing to investigate the effects of build direction (BD) and to verify its properties in the high-temperature compressor environment. To fit within the manufacturing space of an SL machine, the AM diffuser consists of seven radially symmetric sections that are assembled to form a complete flow path. Considerations for modifying the research facility to allow for this unique installation are presented. Precision measurements of the AM components were obtained to compare printed and modeled geometry, and they demonstrate close alignment of flow path dimensions.

Abstract Rotary compressors such as screw compressors, roots blowers, and turbo compressors are used in industry to compress process gases, or as vacuum or backing pumps to evacuate vessels. Gas is sucked in at low-pressure side, transported and compressed by size-changing chambers (PD machines) or energy transmission from rotor to fluid (turbo machinery), and released at high-pressure side. In expanders or turbines, flow direction is from high to low pressure side to gain energy from pressurized gases. The 3D CFD simulation of such compressors/expanders is complex and time-consuming due to its transient nature and fine meshes to ensure a proper representation of radial and axial gaps in the range of some microns with machine dimensions up to meters. Due to this complexity, 3D CFD simulation should focus on the component, i.e. the compressor, and the attached overall system with vessels, valves, pipes, and consumers should be simulated in a 1D network or system simulation. Due to oscillations in the gas flow and interaction with the connected system a transient coupling is necessary. In this paper we show a 3D CFD simulation of a screw compressor using ANSYS CFX in a co-simulation with the 1D Flownex simulation environment of a network modelling the pressurized gas distribution. Whereas the 3D solver works on meshes with up to several million nodes in parallel on HPC systems, the 1D solver typically works serially on several thousand nodes that discretize the flow direction. The transient coupling is based on the exchange of variables at the boundaries of each simulation for every time step allowing for detailed analysis. The impact of the acoustic propagation of pressure fluctuations and the pulsating fluid flow provided by the compressor on the distribution system, and in return the effects of the system response on the compressor are evaluated. Furthermore transient scenarios such as start-up procedures or component failure will be shown.

Abstract To improve the fuel efficiency demanded by airlines and regulations, the turbomachinery industry is required to steadily enhance engine performances and numerical prediction capabilities. One of the solutions is the lean burn combustor which dramatically reduces NOx levels compared to rich one. However, one drawback of this technology is its impact on the High-pressure turbine due to large swirl and reduced cooling airflow, inducing large spatial and temporal variations in the turbine inlet condition. This can drastically change the operation of the turbine and our ability to model it using standard practice, usually RANS computation. To investigate this combustor-turbine interaction, the European Commission-funded project FACTOR (Full Aerothermal Combustor-Turbine interactiOns Research) was launched several years ago. A test rig of a combustor simulator coupled with a 1.5 stage turbine was built at a DLR facility. An extensive test campaign comprising 5 holes probes and infrared imaging was performed. These produced an array of aerodynamic quantities at different points of interest along the machine axis. With this project reaching its term by the end of 2017, results have been disseminated to the partners. This allows a comparison of measurements with RANS modeling on this configuration. The present paper deals with this analysis using several RANS computations and the results of the test campaign. First, single row computation of the Nozzle Guide Vane and rotor blade were performed. To impose the boundary conditions, the experimental map were azimuthally averaged to obtain profiles of total temperature, total pressure and flow angles. Second, the impact of some geometrical features was investigated. This was done using the recent addition of unstructured mesh capability in the elsA solver. Finally, multi-stage computations, both steady (mixing plane) and unsteady (sliding mesh) give an insight on the relative accuracy of these interstage models. All these computations were then used to investigate the behavior of this particular turbine. In addition to classical analysis using profiles of averaged data, the loss sources were identified by computing the viscous and thermal entropy production. This paves the way for a better understanding of the possibilities and limitations of our simulation capabilities.

Abstract Fouling and erosion are two problems that severely affect gas turbines. The shape of the blade, its roughness, and its structural stability can vary as a consequence of these phenomena. The outcomes of this occurrence can span from the efficiency reduction to the engine shut down according to the nature of the material ingested, to the concentration of contaminants in the air, the cleanliness of fuel and to the particular design of the machine. In this work, an axial turbine airfoil is modified according to the requirement of less sensibility to the phenomena above mentioned, utilizing an automatic optimization algorithm. An artificial neural network surrogate approach is used for searching the optimal shape, minimizing the computational cost of the entire process. The optimum design of the blade is therefore achieved, in order to reduce the effects of deposition on the performance. The methodology here proposed is fully general and it is applied to an HPT nozzle in the present analysis.

Abstract Nowadays a preliminary evaluation of environmental impact of a new product becomes more and more important, especially when the case study refers to an industrial gas turbine both for power generation and mechanical drive applications. The environmental impact evaluation, as well as the preliminary lifecycle cost analysis, will represent a critical driver to develop a competitive product during the conceptual design phase where the engine architecture is an outcome of different alternatives trade-offs. Scope of the following paper is the presentation of a set of Design-for-Environment considerations obtained through gas turbine functional decomposition in modules, identification of the most critical, assessment of their contribution compared to the whole engine in terms of environmental impact as well as the effect on the engine use depending on ambient and operating conditions. The outcome of this study is an approach to preliminarily evaluate the engine life-cycle impact as well as a set of indications to drive machine architecture, material selection and production processes towards the sustainability during manufacturing and operational phases.

Abstract Air contamination by solid particles represents a real hazard for compressors for both heavy-duty and aero-propulsion gas turbines. Particles impacting the inner surfaces of the machine can stick to such surfaces or erode them. The presence of deposits entails the reduction in performance of the machinery. As the severity of the problem increases, the performance reduction can become so big to demand engine shut-down and off-line washing. Numerical modeling is one of the techniques employed for tackling the fouling problem. In this work, an innovative procedure is proposed in order to evaluate the losses and the variation in the fluid flow due to the deposits. Specifically, as the deposit grows, it is assumed it forms a porous medium attached to the wall. The porosity of this zone (related to the packing of the particles and to the amount of particles that sticks in to a zone) is responsible for the deposition-induced losses. Different approaches to compute such losses are proposed and discussed. By using this methodology, the two main effects of fouling (variation in roughness and in shape of the airfoil) can be easily included in a comprehensive analysis of the variation of the performance of the compressor over time. Furthermore, this approach overcomes the difficulties that may arise by using a mesh morphing technique. The computational grid is not modified and thus its quality is retained, without remeshing requirements, even for large deposits.

Abstract Micro gas turbine engines in the range of 1–100 kW are playing a key role in distributed generation applications, due to the high reliability and quick load following that favor their integration with intermittent renewable sources. Micro-CHP systems based on gas turbine technology are obtaining a higher share in the market and are aiming at reducing the costs and increasing energy conversion efficiency. An effective control of system operating parameters during the whole engine lifetime is essential to maintain desired performance and at the same time guarantee safe operations. Because of the necessity to reduce the costs, fewer sensors are usually available than in standard industrial gas turbines, limiting the choice of control parameters. This aspect is aggravated by engine aging and deterioration phenomena that change operating performance from the expected one. In this situation, a control architecture designed for healthy operations may not be adequate anymore, because the relationship between measured parameters and unmeasured variables (e.g. turbine inlet temperature or efficiency) varies depending on the level of engine deterioration. In this work, an adaptive control scheme is proposed to compensate the effects of engine degradation over the lifetime. Component degradation level is monitored by a diagnostic tool that estimates performance variations from available measurements; then, the information on the gas turbine health condition is used by an observer-based model predictive controller to maintain the machine in a safe range of operation and limit the reduction in system efficiency.

Abstract The combustion of heavy fuel oil (HFO) in gas turbines (GT) generates some operational constraints and expenses which are too often accepted fatalistically. A first development has addressed the technical challenge caused by the deposition of ash on the hot parts of the turbines. Indeed, the combustion of HFO significantly increases the volume of ash as one must treat the fuel with a “vanadium inhibitor” that acts as an ash modifier preventing hot corrosion by vanadium. This fouling effect is the most serious drawback of HFO operation as it progressively shrinks the performances and reduces the availability of the machines. To tackle this issue, a genuine bimetallic vanadium inhibitor has been developed and field tested step by step between 2015 and 2018. The last step that took place at Yugadanavi in March-April 2018 has allowed validating a ready-to-use version of this bimetallic inhibitor product. Upon the completion of this program, the rate of power degradation during HFO operation has been halved and the GT availability significantly increased while the emission of particulates has been substantially reduced. As a further improvement effort, the teams have tested in the field the effect of changing the temperature of the HFO on its viscosity and monitored the impact induced on the quality of fuel atomization that underlies namely the level of particulate emissions. This second program devoted to the optimization of HFO heating, has enabled defining a rational minimum temperature of the fuel which establishes a fair compromise between atomization effectiveness and thermal energy consumption. This paper summarizes the main outcome of this multi-year collaborative test program.

Abstract Dry, low NO x gas turbines are extremely complex machines that are heavily relied upon in the power industry as baseload, cycling, and/or peaking units. These low-emission gas turbines present potential maintenance and monitoring challenges due to the intrinsically harsh pressure and temperature environments, which make diagnostics and prognostic capabilities extremely difficult. One such challenge involves understanding and interpreting combustion dynamics data. This paper focuses on gas turbine combustion dynamics monitoring (CDM) and describes an algorithm to determine combustor health based upon dynamic pressure. The ongoing CDM and diagnostic work has progressed from taking basic binned FFT data and transforming this data to statistically-based health indicators that can be continuously calculated to determine combustion system anomalies. These anomalies can be detected hours, days, and sometimes even weeks before passive CDM alarm levels are reached, thus, giving additional time to plan for shutdown, inspection, and repair. The paper will discuss real-time observed successes and challenges associated with combustor health monitoring, including sensor health determination and factors associated with the non-linear nature of combustion dynamics. Overall, this work is helping to better alleviate the user’s “black box” perspective of combustion dynamics monitoring systems through automated, real-time interpretation for combustion system health for can annular gas turbines.