Jaap van Kampen

§

Master’s project in co-operation with the NLR and the Dutch Air Force

Heat transfer in the Pratt&Whitney 220 Turbofan Engine

At the National Aerospace Laboratory (NLR) simulation software is used to predict the lifetime of the F100-PW-220 turbofan jet engine (see figure 1), which is used in the F-16 jet fighters of the Dutch Air Force. The simulation software, called GSP (Gasturbine Simulation Package), uses mission information, such as altitude and throttle handle position to monitor the wear of the turbofan engine. For this purpose heat transfer models and failure models are used.

Figure 1: The F100-PW-220 Turbofan Engine

Up to now, the heat transfer to the PW-220 combustor walls was estimated by assuming a constant heat transfer coefficient and a constant temperature inside the combustion chamber. Effects of wall film cooling, a non-constant temperature distribution and radiation are not yet included.

Preliminary Computational Fluid Dynamics (CFD) calculations showed the possibility of using CFD to predict the wall heat flux in the F100-PW-220 combustor. However, many simplifications were made in these calculations. For example: only one burner is modeled. The annular combustor of the real engine contains 12 burners. A more sophisticated model should include the interaction between the burners. Moreover, the heat transfer was predicted with a wall function. A further analysis is to be made with a low Reynolds turbulence model, showing the development of the boundary layer. Because the focus of research will be on heat transfer, a gaseous combustion model will be used, although the real engine is fired with liquid fuel.

The master’s project will focus on the following aspects:

1)

Study of the current model and specify its shortcomings;

2)

Refine the model of the F100 combustor;

3)

Implement an improved low Reynolds model;

4)

Perform k-low Reynolds simulations;

5)

Create a database with the wall heat flux as a function of the operating conditions. This database can subsequently be used in NLR’s GSP-software.

More information:

B. de Jager

b.dejager@ctw.utwente.nl, tel: 053-489 1091, Room N-209

Jaap van Kampen

j.f.vankampen@ctw.utwente.nl, tel: 053-489 2417, Room N-215

§

Master’s project within the framework of ‘DESIRE’

(Design and demonstration of highly reliable low NOx combustion systems for gas turbines):

Numerical thermoacoustic modeling

In modern gas turbine designs, low NOx emissions are achieved by introducing lean premixed techniques in combination with annular combustion chambers. The price that has to be paid is a higher risk for thermoacoustic instabilities in the combustion chamber of the gas turbine (see figure 1). These instabilities arise from a feedback mechanism between oscillatory flow and heat release perturbations and often lead to large amplitude pressure and velocity perturbations in the combustor. These acoustic perturbations significantly reduce lifetime and regions of operability of the combustor.

Figure 1: Siemens V94.3A 265 MW gas turbine

At the laboratory of Thermal Engineering an acoustic model is developed with which the unstable regions in a combustor can be predicted. An important parameter in this model is the so-called “flame transfer function”, which indicates how the flame responses to perturbations in its inlet. Inlet perturbations are for example fluctuations in the fuel/air ratio and in the mixture mass flow. The heat release of the flame in response to these perturbations is directly related to the acoustic source of the flame.

One way to obtain the flame transfer function is by means of Computational Fluid Dynamics (CFD). The laboratory has its own combustion model, which is being implemented into the CFD-code, CFX 5.X. Within this framework, this combustion model has to be formulated in an optimized way to give the heat release of the flame. Subsequently, the inlet conditions of the flame can be perturbed in a transient CFD run, and the response of the flame can be monitored. This flame response will cause acoustic waves in the numerical domain. Wave propagation in the numerical model is also an aspect which should be studied.

The master’s project will focus on the following aspects:

1)

Study the combustion model and set up a thermochemical database for partially premixed combustion and heat release.

2)

Optimize the implementation in a transient turbulent flame code to give the heat release of the flame; The CFD-code to be used is CFX 5.X, which is available at the laboratory;

3)

Obtain flame transfer functions for different frequencies and various mean conditions;

4)

Study the modeling of acoustic waves in CFX 5.X.

More information:

Jaap van Kampen

j.f.vankampen@ctw.utwente.nl, tel: 053-4892417, Room N-215

§

Master’s project within the framework of ‘DESIRE’

(Design and demonstration of highly reliable low NOx combustion systems for gas turbines):

Measurements and simulations on a swirl burner

The laboratory of Thermal Engineering participates in a European project called ‘DESIRE’. This project aims to reduce emissions and to increase the efficiency of modern large scale gas turbines (power: ~250MW). One of the tasks within this project is to build a test-setup of a combustor. This test-setup will be built in the summer of 2003. In the test-rig, a swirl burner will be used. The swirling flow generated by this burner causes a recirculation zone which stabilizes the flame. When modeling the combustion process using Computational Fluid Dynamics (CFD), it is extremely important that the recirculation zone can be accurately predicted. Preliminary CFD-calculations were already performed. Figure 1 shows the velocity in the DESIRE combustion chamber. One can easily identify the recirculation zones.

Figure 1: Velocity in the DESIRE-combustion chamber

To study the swirling flow through the DESIRE burner a so-called water tunnel will be built. The water tunnel is an exact copy of the future DESIRE combustion chamber, except for two aspects: 1) water is used instead of air, and 2) the water tunnel is constructed from Perspex. When the flow-velocities in the water tunnel are such that the Reynolds-number is the same as in the DESIRE-combustion chamber, both flows will behave similarly. When using water, which has a 10-fold lower kinematic viscosity than air, the flow velocities in the water tunnel are 10 times lower than in the original combustion chamber. With the help of a laser doppler measurement system, the velocity profiles can be determined in the water tunnel. These profiles can be used to validate CFD-calculations.

The master’s project will focus on the following aspects:

1)

Perform measurements on the flow profile in the water tunnel using laser doppler velocimetry.

2)

Use CFD to simulate the flow through the water tunnel, and compare the results with the measurements.

3)

Study the effect of different turbulence models and boundary conditions on the CFD-results.

4)

To create a combustible mixture in the DESIRE-combustor, natural gas will be mixed with the swirling air. This mixing can also be visualized in the water tunnel by injecting ink in the swirling flow. Results must be compared with mixing simulations in CFD.

More information:

Jaap van Kampen

j.f.vankampen@ctw.utwente.nl, tel: 053-4892417, Room N-215