What is the difference between active instability control and passive methods?
Except for infrequent special-purpose applications (such as so-called pulse combustors), combustion instabilities are undesirable effects detrimentally influencing life, performance and environmental characteristics of any combustion system. Accordingly, solutions are actively sought that allow the avoidance of this kind of instability or - if oscillatory problems are already being encountered - to suppress them or at least reduce them to acceptable levels. Measures taken in this regard can be subdivided in principle into two distinct groups: passive and active ones.
In general, this term refers to systems devoid of "intelligence", i.e. of closed-loop control; this means that they do not influence or control oscillations arising within a combustion system by specifically interpreting the course of certain system parameters. To simplify even more, passive methods may be said to be characterized by requiring no external energy supply.
Below, some examples for passive methods as well as highlights on modes of action, benefits and disadvantages:
- Damping elements and silencers: acoustic energy is dissipated as heat. Will not prevent oscillations but capable of reducing sound pressure amplitudes occurring. Low-price standard components but frequently bulky; entail additional pressure losses.
- Helmholtz and lambda/4 resonators: disturb the resonant sound field and thus the evolution of combustion instabilities. Effectiveness limited to certain frequencies; low frequencies need large dimensions.
- Baffles: disturb the propagation of sound waves and thus any resonant sound field that might form. Uncomplicated construction but subject to burning off; frequently detrimental to combustion itself; additional pressure losses.
- "Detuning" - geometrical modifications used to modify eigenfrequencies within a combustion system. No significant drawbacks for combustion and efficiency but often expensive and complicated to build. Moreover, effectiveness frequently limited to specific frequencies; thus, there is some danger that oscillations will merely shift to some other frequency.
Above all due to these drawbacks and the limited effectiveness of passive methods, active methods constitute an interesting alternative specifically for systems operating under highly variable conditions.
With "Active Instability Control (AIC)", i.e. active suppression of self-excited combustion instabilities, system parameters are measured and fed into a control system that - depending upon the input signal - uses an actuator to influence the corresponding system so that no oscillations are generated. A simple example: a pressure sensor detects pressure fluctuations arising within a combustion system for a closed-loop control system to determine a signal acting upon a loudspeaker. As a function of signals received, the loudspeaker generates an anti-sound field to be superimposed upon the initial sound field in the system so as to cancel both fields out.
In practice, AIC unfortunately is a little more complicated than that. Various aspects have to be taken into account, and the mechanism triggering self-excited combustion instabilities has to be carefully researched before a decision can be made regarding a suitable AIC strategy. Sensors, actuators and control system architecture depend to a large extent on the combustion system concerned (power, fuel, geometry), on frequencies arising and on sound fields generated in the process. However, any extra expense entailed by AIC will be justified by the benefits linked with it:
- Hardly any need to modify combustion systems, small footprint, relatively uncomplicated retrofitting.
- No detrimental effects on heat release and combustion control, no additional losses.
- In certain cases, pollutant emissions and incomplete burnout can even be reduced.
- True closed-loop control reacts flexibly to actual system behavior and its changes. No intervention except if and while necessary.
- Upon system stabilization by AIC, actuating signals are reduced, thus decreasing power consumption by the actuator and increasing life of the actuator.
What are the causes of acoustic feedback?
Acoustic feedback is characterized by eigenfrequencies and eigenmodes. With most self-excited combustion instabilities, feedback is provided by acoustics. To understand the causes leading to self-excited combustion instabilities, it is thus indispensable to first consider in detail the acoustics of a combustion system. In this context, potentially resonant acoustic frequencies (eigenfrequencies) and the associated eigenmodes of sound pressure and particle velocity are of major importance. Eigenmodes are defined as the standing waves of sound pressure and particle velocity produced by resonance.
The following illustration demonstrates, for a simple tube closed at one end and open at the other, the eigenmodes corresponding to the first, second and third harmonic along the longitudinal axis of the tube. The boundary conditions are idealized and assumed to be acoustically closed for the left end and acoustically open for the right one.
In contrast to this simple example, eigenmodes in a real combustion system look significantly different owing to changes in cross-section, to the influence exerted by flow conditions, to temperature fluctuations within the combustion zone and to the complex boundary conditions at the inlets and outlets of the system.
Premix flames are favored in modern combustion systems to reduce emissions, because they have low NOx emissions due to their low flame temperatures. Unfortunately, with this type of combustion, and in conjunction with combustion-chamber acoustics, so-called self-excited combustion dynamics emerge very quickly. These are often also referred to as combustion instabilities, thermoacoustic instabilities, or, in accordance with the audible frequencies, also known as rumbling, humming, screeching or pulsation.
Strong pressure fluctuations at one or more frequencies are the underlying cause of such vibrations. Pressure fluctuations can achieve such high amplitudes that the combustion system itself or components connected upstream and downstream are destroyed. These vibrations can be observed in a range of situations from household heating systems to large-scale firings and from stationary gas turbines to rocket engines.
The IfTA PreCursor is a method for early detection of combustion dynamics- even before high amplitudes occur which can lead to damage.
How does experimental modal analysis work?
Owing to wide fluctuations of all sound field parameters within systems subject to self-excited combustion instabilities, merely determining eigenmodes by measuring local sound-pressure amplitudes and phases does not lead to useful results
Therefore, we developed a special analytical method based on correlation measuring techniques. In simplified terms, a microphone is placed at a reference point while another one is used to determine sound-pressure amplitude and phase relationships in the frequency domain. Complemented by averaging in the frequency domain, this method yields the necessary insights.
How does the numerical modal analysis work?
For computing purposes, any geometry to be analyzed is rendered discrete by means of cylindrical and conical elements that permit the modeling of actual cross-sections to any degree of precision required. This discretizing procedure is backed by comprehensive experimental know-how; after all, the quality of any results obtained is determined to a large extent by appropriate partitioning of the domain to be analyzed. Discretized geometry, together with further parameters - temperature, flow velocity and acoustic boundary conditions - constitute the basic data for a special numerical code. Acoustic boundary conditions, including frequency-dependent ones, can be imposed at will. In the event that complicated boundary conditions cannot be sufficiently formalized by theoretical approaches, we offer experimental studies performed according to the n microphone method. The program code used was written at IfTA; new features are being added continuously.
This version makes it possible to easily vary the length or cross-section of any given tube element to target research on specific shifts in mode or frequency.
What defines the Rayleigh criterion?
The Rayleigh criterion is an important stability criterion for self-excited combustion oscillations
If any resonant frequency is to be excited within a combustion chamber, a source of sound is required that constantly feeds energy into the system at the correct oscillatory frequency. With any combustion system, energy is supplied by heat release rates oscillating around their mean value.
However, any such thermal power oscillation must not only have the correct frequency to trigger self-excited combustion instabilities but also the correct phase angle suitable for reinforcing the sound pressure fluctuations arising within the system. In 1878, Lord Rayleigh first determined the criterion that now bears his name. According to it, combustion instabilities are excited if the oscillation of sound-pressure and a flame's heat release rate are in-phase, or damped down, if they are anti-phase. Putnam and Dennis took up this idea and provided a formal mathematical representation of Rayleigh's criterion, the so-called "Rayleigh integral":
This integral means that the product of thermal power Q(t) and sound pressure oscillation p(t), integrated over the period of oscillation T (=1/f [s]), must be positive for any combustion oscillation to be excited. Whenever it is negative, oscillations are damped.
Applying this equation to any harmonic oscillation, the condition is met if the absolute value for the phase difference between the oscillations of heat release and of sound pressure is less than or equal to 90°.
The research tools now available from us for the one and two-dimensional Rayleigh indexes constitute a logically consistent application of this criterion. The Rayleigh index is a local version of the globally expressed inequality above allowing to identify areas subject to excitation and damping within any combustion zone. With this information, the interactions between thermal and acoustic oscillations can be influenced deliberately.
The above illustration shows the one and two-dimensional Rayleigh indexes computed from experimental data obtained for a liquid fueled combustor. The one-dimensional Rayleigh index is plotted versus burner length (bottom), the two-dimensional one as an intensity plot together with the burner geometry (top). It is obvious that flame sections subject to excitation and damping alternate along the combustion chamber and that exciting domains predominate, a fact which leads to the observed combustion instabilities.
How do self-excited or self-induced combustion oscillations develop?
Given certain operating conditions, industrial-type combustion and propulsion systems may be subject to self-excited combustion instabilities, sometimes designated flame pulsations or oscillations, which are characterized by pressure oscillations arising at discrete frequencies. With low-power systems, such as domestic or auxiliary vehicle heating systems, combustion instabilities normally only produce high levels of noise.
With higher-power systems, such as regenerating air heating units, process gas heaters, gas turbines or rocket engines, the pressure amplitudes reached are sometimes so high that the alternating loads entailed by them may result in mechanical system failure of the combustion chambers as well as the upstream and downstream components. For instance, pressure amplitudes measured for process gas heaters reached 0.5 bar, and even 5 bars for a rocket combustion chamber.
In addition to pressure fluctuations, self-excited combustion instabilities invariably induce, via sound particle velocities, equivalent fluctuations of flow velocities that lead, in their turn, to substantially increased amounts of heat being transferred to combustion chamber walls. Thus, thermal loads will increase, in addition to mechanical ones, so that there are definite thermal hazards that might even destroy the combustion chamber. When amplitudes reach substantial values, burner flames can often no longer be stabilized so that they may flash back or be extinguished.
Basically, any self-excited combustion instability is due to several quantities interacting physically and leading, given appropriate preconditions, to resonance. In order to allow any such oscillation to arise in a self-excited manner, sound-pressure and heat-release oscillations have to reinforce each other. For this to happen, there must be a feedback mechanism inducing heat release fluctuations strong enough to reinforce sound pressure oscillations. In most cases, feedback is caused by the acoustics of the combustion system concerned. However, other mechanisms have likewise been observed.
For instance, structural vibrations in rocket drives may modulate fuel flows supplying the combustion chamber and thus produce oscillations of the heat release rate of the flame. The vibrations caused by these periodic changes of the heat release rate in turn reinforce the structural oscillations, thus closing the loop.
Self-excited combustion dynamics are also often referred to as combustion instabilities, thermoacoustic instabilities, or, in accordance with the audible frequencies, also known as rumbling, humming, screeching or pulsation.
What defines thermal heat release rates of flames?
Unsteady heat release rates of flames, i.e. fluctuations of the energy introduced into the flow of air by combustion per unit of time, constitute an important parameter for self-excited combustion instabilities. Frequently, modulated flows of air-fuel mixtures into the combustion zone cause this type of fluctuating heat release. Quantifying these fluctuations constitutes an important precondition for analyzing any unsteady combustion and its classification as self-excited or externally excited.
The combustion of hydrocarbons in oxygen or air is not a single bulk reaction but proceeds over a sequence of numerous (sometimes several hundred) intermediate steps involving a corresponding number of chemical species. Many of those intermediate reactions proceed extremely rapidly, resulting in correspondingly short-lived species, such as OH radicals whose lifetime is less than 600 ns. OH radicals emit radiation at 306.4 nm that can be measured using suitable probes; this parameter provides a sufficiently precise figure for the burning rate and thus for the heat release rate.
In addition to spatially integrated measurements of the radiation emitted by OH radicals (total heat released), IfTA offers 1D and 2D measurements for certain points or segments of a flame front. We use special photo multipliers for integral measurement purposes. 1D and 2D measurements are performed by means of special fiber optics. The measured signals are analyzed in the frequency domain in order to determine the amplitude and phase relationship between fluctuating heat release rates and oscillations of the sound field parameters measured at the same time. This type of data may be used not only for direct analytical purposes but also to compute the Rayleigh-Index for example.