On Monday 25th June 2001 (MNM Week 26) we reported the Annual Seminar and Summer Management Meeting of the EuroFlam consortium at Cardiff - Wales, on Thursday 21st and Friday 22nd June. Sixteen Graduates/Visiting Investigators who presented their completed reports or progress reports attended the meeting. In this and the following editions of EuroFlam News,  we are publishing abstracts of the presentations made.

Turbulent Characteristics of Swirl Flow with and without a PVC
by Marzia Taddei - Cardiff

Best Presentation Award (Certificate plus a cheque for €150), was given by Air Products Plc.

Summary of the presentation

The general aim of this EuroFlam project is to highlights the importance of the PVC-Precessing Vortex Core and to determine whether this effect is beneficial or not on the flow’s energy. Swirl burners are widely used in industry because their high degree of swirl generates Vortex Breakdown phenomena including the Reverse Flow Zone (RFZ) and the PVC. These phenomena induce very strong mixing of the flame therefore improving combustion efficiency by ensuring that all the unburned fuel molecules have plenty of oxygen molecules floating around them. The three-dimensional instability called PVC occurs when the central vortex core of a swirl flow moves off-centre and starts to precess around the axis of symmetry.

To study the interaction of the PVC and the RFZ in the exhaust of a swirl burner and to analyse the differences between a flow with and without a PVC, all experiments were performed using a 150 kW small swirl burner, a scale model of a 2 MW swirl burner. The method of swirl generation was via tangential inlets. Under isothermal conditions pressure measurements were carried out to set up the rig and to characterize the occurrence of the PVC. For an airflow rate of 2500 l/min and a swirl geometric number of 0.76 the typical PVC main frequency was 87.5 Hz. Using the same configuration of the rig but with an insert of 50% in the tangential air inlets, different airflow rate, premixed and diffusion gas were used to create combustion conditions.

Laser Doppler Anemometry (LDA) measurements were used to investigate the fluid flow structure. Through laser measurements, four profiles have been chosen with different distances from the exit of the burner to understand the aerodynamics for an open flame with and without the effect of the PVC. These four profiles were taken for a partially premixed flame, with the following axial distances: 20, 40, 70 and 195 mm.

The mean axial velocities measured show that with the PVC, in the second and third profile, the RFZ is wider and more intense, with higher forward axial velocities to compensate with the peak axial velocity occurring at wider radius. For both types of flame (with and without the PVC) the peak of axial velocity is at the boundary of the RFZ. The corresponding tangential velocities show that higher tangential velocities occur throughout the flow field in the third profile. Peak tangential velocities with the PVC are nearly twice that without the PVC. For the third profile, the PVC appears to act as a coherent structure considerably reducing decay of tangential velocity in the axial direction. As to be expected the passage of the large coherent structure, the PVC, through the flow increases the level of axial and tangential velocity fluctuation throughout the flow.

Regarding energy spectra it is possible to notice the presence of peaks around 90 Hz for the flow with a PVC. Also the range of the frequency is very similar for both kind of flames, but the value of energy increased by a factor of 10 with the presence of the PVC at the small eddy sizes. This implies that the smaller eddies (associated with high frequency) have increased their influence in the flow, by means of bringing up the larger eddies.

To study the turbulent scales of the flow considered, normalised time correlation series were carried out to characterise the velocity field and calculate the integral length scale of the large scale eddies. The average length of the large scale eddies, in the tangential direction is larger with the presence of the PVC than without this effect. Instead in the axial direction it is not possible notice a clear effect of the PVC on the value of the integral length scale. This means that in the axial direction the RFZ creates large instabilities in the flame.

The future work associated with this project could consist of carrying out other LDA measurements with higher data rates and different axial distances above the burner exit to study in more detail the flame. It will be interesting also to analyse the dissipation length scale to have more information about the size of eddies with a PVC presence in the flow.

Active Control System of combustion instabilities in Gasturbines
by Alain Guillot - ENEL

Abstract

Instabilities in combustion systems in general occur when the combustion process excites one or more natural acoustics modes of the system.

To date, these instabilities have been damped by using passive methods which consist of one or more of the following:

• Modification of the combustion process to reduce the magnitude and/or to change the frequency of its driving;
• Increasing the combustion chamber damping;
• Shifting the frequencies of the combustor unstable modes away from the range where the combustion process driving is maximized.

However, due to limited understanding of the processes that control combustion instabilities, these passive control approaches are generally costly and require considerable time, and often, produce solutions which are failing to damp correctly instabilities.

That’s why, recently, the interest went in the development of Active Control System (ACS) for preventing instabilities. In contrast to a passive solution, ACS controls the instability by continous sensing and evaluating the state of the combustor and forcing it to perform in a desired manner.

A typical control system may consist of a pressure sensor, an observer, a controller and an actuator.

The pressure sensor measures the combustor pressure and sends the measured signal to an observer.

The observer uses a measured combustor pressure to determine the state of the process (Amplitude, Frequency, Phase of the modes). These informations are used by the controller to generate a control signal for the actuator. The latter permits to inject an oscillating fuel stream into the combustor in order to damp the instability.

The aim of this work is to realize a numerical simulation of pressure oscillations occuring in the combustion chamber of a gasturbine, and to develop an active control system which is able to damp oscillations.

The first step is to implement in Simulink, Dynamic System Simulation for Matlab, a control algorithm of the combustion instability, based on Van der Pol equation.

The second step is to use Neural Network, to develop and to implement an Adaptive Control for identification and damping of instabilities.

The last step consists in to test the developed control system by means of experimental data.

Experimental Studies of Oil Spray Combustion in Highly Preheated Air
by Maddalena Allegorico - IFRF

Abstract

To reduce energy consumption and to minimize pollutant emissions, like Kyoto protocol requires, are the key factors to encourage European industries to develop new efficient and competitive technologies.

Good results towards these goals were achieved by using High Temperature Air Combustion (HTAC) technology. I focused my attention on the applicability of this new technique on oil spray firing.

This new technology requires highly preheated combustion air, obtained using regenerative burners, and fuel injectors positioned away from the comburent inlet. This is the main difference with Conventional Combustion where comburent and fuel were injected into the same channel and swirled. The HTAC leads to:

1. Substantial increase in thermal efficiency
   or reduction in CO2 emission by more than 30% compared with the conventional combustion system
2. Drastic reduction in NOx emission
   because combustion takes place in the whole furnace with less O2, which produce a slow combustion and suppress the peak flame temperature.

The objectives of this work are:

• Characterize four different oil atomizers using a laser technique, Phase Doppler Particle Analyzer (PDPA);
• Investigate their combustion performances with HTAC technology;
• Examine the applicability of HTAC to light (LFO) and heavy (HFO) fuel oil firing.

In particular I was interested on last two points, carrying out in furnace experiments. The furnace facility has a square cross-section of 2x2m and a length of 6.2m. The internal wall is refractory lined. The input/output parameters, the wall temperature and the cooling waterflow are continuously monitored. A precombustor uses NG as fuel to preheat the comburent up to 9000C. The precombustor simulates the regenerator. Before coming into the furnace O2 is added to the flue gas to keep the O2 level to 21% vol. wet. The oil mass flow rate used is 50kg/h giving a thermal input of 1 MW. The nitrogen content on light fuel oil is < 0.05 (wt%) and on heavy fuel oil is 0.44(wt%). The two oil guns was positioned at 250mm and 355mm from the centerline of combustion air inlet. The oil atomizers used take part of twin fluid category and are: Y-jet, Air Assist Internal Mixing (AAIM), Air Ring (AR) and Flat Flame (FF).

PDPA technique was used to characterize spray. Droplet size and velocity distributions were measured at several locations within the spray. As expected different atomizers design give different velocity profiles and different droplet size. Y-jet and AR produce very small droplets. The Sauter mean Diameter is 40 mm. Otherwise the other two atomizers give particle size of roughly 65 mm.

Soot measurements was done at 2% of O2 in flue gas using a soot-sampling probe. We did not find substantial differences between atomizers. Soot concentration is around 8 mg/Nm3 on LFO and 12 mg/Nm3 on HFO. NOx concentration is 10ppm on LFO and 120ppm on HFO.

The O2 content in the flue gas, below that is not possible to work because CO concentration became too high, is roughly 1% on both fuels. The NOx level increases by increasing excess air level due to higher conversion efficiency of fuel nitrogen because much more O2 is available. NOx concentration value measured on HTAC is 100ppm-vol.dry lower than the value obtained in previous works made with conventional combustion.

Also in Total Radiative Heat Flux there is a substantial difference between the two combustion systems. A peak in the first meter of the furnace was observed in previous works on conventional combustion. Ellipsoidal measurements carried out on HTAC show a very flat profile along the furnace and high radiative fluxes (450 kW/m2) on both oils. The flat profile is due to a very high degree of mixing because oil jets entrain the recirculated flue gas before mixing with air giving a low O2 concentration and a slow combustion.

In flame measurements were performed firing heavy fuel oil, atomized using FF atomizer.

Temperature field is uniform in the whole furnace and the highest value measured is 14500C. This value is considerably low considering that 17000C were measured with conventional combustion technique. High O2 concentration is measured only at burner inlet, the rest of the furnace is filled with combustion products containing 1% O2. The highest NOx concentration value measured was 200ppm-vol. dry. Low NOx concentration is correlated with the low flue gas temperature. NOx concentration measured at chimney was roughly 400 mg/Nm3 at 3% O2.

The furnace efficiency is roughly 50% firing both oils.

We can conclude saying:

• The investigated atomizers are fully applicable to HTAC:
  • Combustion performance is not affected by tested atomizer design;
  • No peak flame temperature distributions;
  • Uniform radiative heat flux.
• HTAC technique is applicable to LFO and HFO combustion.