How sonic horns work: the physics of acoustic fouling control

How do sonic horns work?

A sonic horn converts a short supply of compressed air into a high-intensity pressure wave in the gas space of industrial equipment. The wave makes the gas and exposed particulate oscillate. Where a deposit is dry, friable and only weakly attached, that repeated motion can overcome the small adhesive and cohesive forces holding particles to a surface or to one another. Gravity or the process gas then carries the released material away.

That description matters because a sonic horn does not clean by blasting a surface with a narrow air jet. Nor does it depend on every ash or dust particle having a resonant frequency that happens to match the horn. Its useful action is a distributed, oscillating pressure field, applied repeatedly before a deposit develops a strong bond.

The physical chain has four stages. Compressed air supplies the energy. A driver converts that pneumatic energy into rapid mechanical motion. A flared bell couples the motion into the process gas as sound. The resulting pressure and gas-velocity oscillations disturb weak particle contacts throughout the area reached by a sufficient acoustic field.

Understanding each stage explains both the reach of acoustic cleaning and its firm limits.

From compressed air to a pressure wave

Most industrial sonic horns use a diaphragm driver. During a cleaning pulse, an actuated valve admits air from a receiver or branch line into the driver. Pressure acts across a thin metal diaphragm. The diaphragm flexes, uncovers or changes the discharge path, and releases air towards the horn throat. The pressure then falls, the diaphragm returns, and the cycle repeats rapidly.

This feedback between air flow, diaphragm motion and the connected horn produces a self-sustained oscillation for as long as the valve remains open. A diaphragm horn is therefore more than a valve making repeated puffs. Its driver and bell form a coupled acoustic system whose geometry, stiffness, mass and operating air pressure determine whether the oscillation starts cleanly and remains stable.

The air pressure that matters is the pressure at the horn while it is firing. A healthy compressor header does not prove that condition. Undersized pipework, long runs, restrictive valves, wet filters or an inadequate receiver can allow pressure to collapse during the pulse. The horn may still make noise while producing less acoustic output or an unstable waveform. Supply pressure, flow capacity, valve opening time and receiver recovery must be considered together.

The bell horn does not create energy or multiply it without cost. It is a tapered acoustic guide between a small, high-pressure throat and a much larger body of process gas. The change in area acts as an acoustic impedance transformer, improving the transfer of energy from the driver into the gas and influencing directivity and spectrum. Bell length, flare profile, mouth area and the mounting boundary all affect that coupling.

Not every pneumatic acoustic cleaner uses a diaphragm. A piston-whistle horn, for example, interrupts an air stream through another moving or whistle-like element. The source mechanism and output spectrum differ. The rest of this article concentrates on the diaphragm and bell arrangement because that is the common meaning of sonic horn in boiler, air heater, baghouse, ESP, SCR, hopper and silo service.

Frequency, wavelength and horn size

Frequency is the number of pressure cycles per second, measured in hertz. A horn described as 125 Hz completes about 125 oscillation cycles each second while firing. Its fundamental frequency is the lowest dominant mode of the coupled driver and horn assembly, not an arbitrary controller setting.

Changing supply pressure can alter acoustic amplitude, starting behaviour and stability, but it does not normally turn one horn geometry into another frequency class. If a unit produces an unexpected spectrum, the investigation should consider diaphragm condition, assembly, air delivery and mounting rather than treating pressure adjustment as a tuning control. Selecting another fundamental generally requires a different coupled driver and bell design.

Real output is not a perfectly pure tone. It normally contains the fundamental and one or more harmonics, which are components at integer multiples of that fundamental. The balance between them depends on the driver, bell, air supply, mounting and surrounding load. This is why a nameplate frequency is useful but not a complete description of what enters the vessel.

Frequency also sets wavelength, according to the relationship wavelength = speed of sound / frequency. At about 20 degrees Celsius in still air, using a sound speed near 343 metres per second, the approximate values are:

These figures are illustrations, not boiler design values. Sound speed rises with gas temperature, so the wavelength at one frequency becomes longer in a hot pass. Gas composition and flow also affect propagation. A designer should use operating conditions rather than copying the ambient values into a layout.

Long wavelength is one reason a low-frequency acoustic cleaner can establish useful pressure variation around large structures and through open passages. A wave several metres long is not stopped in the same way as a narrow mechanical jet when it meets a tube or plate. It can diffract, reflect and enter connected spaces. That does not mean it passes every obstruction without loss, or that the pressure behind every tube bank is equal.

Lower frequency is not automatically better. A larger bell is usually needed to couple a low frequency efficiently, and the useful result still depends on source output, vessel dimensions, absorption, reflections and the deposit. At the other extreme, an infrasonic cleaner operates below conventional sonic-horn frequencies and is a distinct equipment class. Frequency selection is a matching problem, not a contest for the lowest number.

The acoustic field inside real equipment

The pressure wave leaving the bell spreads through the process gas. At any measurement point, the alternating pressure can be expressed as sound pressure level, or SPL. It is a logarithmic ratio relative to a reference pressure. A decibel value therefore does not behave like a linear percentage, and a stated figure is incomplete unless it identifies what was measured and under what conditions.

Sound power and sound pressure are not interchangeable. Sound power describes the total acoustic output of a source. Sound pressure describes the local field at a particular point. One horn can have the same sound power while producing very different SPL readings beside the bell, across an open yard and inside a reflective steel casing.

Frequency reporting matters too. An overall number can conceal whether energy sits at the intended fundamental or in less useful components. An octave-band or narrow-band spectrum, together with the weighting method and microphone position, gives an engineer more information than a single unqualified decibel figure. A measurement taken one metre in front of an unmounted horn is not evidence that the far corner of a vessel receives the same pressure.

In an ideal open free field, acoustic intensity from a point source follows the inverse-square law. Doubling distance reduces intensity to one quarter and corresponds to an SPL reduction of about 6 dB. That rule is a useful reference, but it is not a vessel model. Acoustic attenuation also arises from absorption in the gas and dust, scattering by internals, leakage and wall losses. Local SPL varies further through constructive and destructive interference.

Close to the bell, pressure and particle velocity have complex spatial and phase relationships. Farther away, the field may behave more like a propagating wave. The distinction between near field and far field depends on source size, wavelength and geometry, so there is no universal distance at which a rating becomes representative.

Boilers, air heaters, baghouses and silos add reflecting walls, openings, tube bundles, plates, catalyst layers and moving gas. Incident and reflected waves combine. Where their phases reinforce, local pressure increases. Where they oppose, local pressure falls. If a dimension and boundary condition support resonance, the gas space can amplify a mode. A standing wave can then form pressure antinodes and nodes, creating strong and weak zones at fixed locations.

This is not an argument for trying to make an entire boiler resonate uniformly. Source resonance, gas-space modes and structural resonance are different phenomena. A useful gas pressure maximum in one bay does not prove adequate coverage elsewhere, and exciting a casing mode is not a cleaning objective. Layout work should treat reflections as part of a complex field and should check critical surfaces, rather than rely on a nominal frequency alone.

How pressure oscillation releases particulate

Sound in a gas involves alternating pressure and alternating gas velocity. As the wave passes a deposit, exposed particles and the gas around them are accelerated back and forth. A particle does not necessarily follow the gas exactly. Its response depends on size, density, shape and the frequency and amplitude of the field. The resulting relative motion creates oscillating drag and inertial forces at particle contacts and at the interface with the equipment surface.

For a loose ash or dust layer, those contacts are numerous but individually weak. Van der Waals attraction, electrostatic effects, small mechanical interlocks and light compaction can hold the layer together. The acoustic cycle repeatedly changes the load direction. When the imposed force, assisted by gravity and process flow, exceeds the local adhesive or cohesive resistance, a particle or small agglomerate separates.

The word fluidisation is sometimes used for the visible result. It can be a helpful practical description when a surface layer becomes mobile, but it should not be confused with a deliberately engineered fluidised bed. A horn can agitate and disperse loose particulate without establishing the uniform gas distribution or operating regime required for true bed fluidisation.

Acoustic streaming is another real but frequently overstated effect. It is a steady, time-averaged gas motion generated when sound is absorbed or interacts with viscous boundary layers. This mean flow is superimposed on the much faster back-and-forth acoustic motion. It may help circulate fine material locally or change particle transport, particularly near boundaries and in complex gas-particle systems.

Streaming is not the main explanation for every successful installation. The primary useful action is the oscillatory pressure and velocity field repeatedly stressing weak contacts. Nor can streaming replace the plant's transport path. Once particles detach, gravity, hopper geometry, a conveying system or sufficient process-gas flow still has to move them to a legitimate exit. Otherwise the horn can redistribute dust, leave it suspended temporarily or allow it to settle on another surface.

Why prevention is easier than recovery

Freshly arriving dry particles usually have fewer and weaker contacts than an aged deposit. With time, more material accumulates, the lower layers compact and particles find additional contact points. Temperature cycling, chemical reaction, moisture and pressure can increase bond strength further. A layer that was initially friable can become a dense cake or a sintered mass.

Acoustic cleaning works best by interrupting that progression. Short, repeated horn cycles disturb particles before a continuous, load-bearing layer develops. The objective is not to wait for a passage to choke and then apply more sound. It is to keep deposition below the point at which bond strength and thickness move outside the acoustic field's capability.

This preventive role affects commissioning. A horn installed against an established hard deposit may appear ineffective even if its field is suitable for keeping the same surface cleaner after an outage wash or mechanical clean. Starting condition, firing interval and the route for released material are therefore part of the physics, not merely operating preferences.

The firing interval cannot be selected from frequency alone. Deposition rate changes with load, fuel, gas temperature, dust loading and local flow. Excessive firing wastes compressed air and adds diaphragm cycles without necessarily improving cleanliness. Infrequent firing allows contacts to strengthen. The useful sequence is the least intensive one that interrupts deposit growth under the actual process conditions.

Why sticky, molten and sintered deposits stay put

The same physics that explains success on dry, friable dust explains failure on other deposits. A sonic horn distributes acoustic energy through a gas volume. The alternating stress available at any one bond is modest compared with the concentrated momentum of a sootblower jet, the direct impulse of a rapper or the thermal and hydraulic action of water washing.

Wet or sticky material forms liquid bridges between particles and between the deposit and the wall. Capillary forces increase adhesion, while viscosity dissipates oscillatory motion as heat. A tacky layer can deform with the wave and then relax without breaking away. If the surface remains sticky, any particles that are released can attach again nearby.

Molten or partly molten ash has the same fundamental problem at higher temperature. It flows or deforms instead of fracturing as a dry, brittle network, and new arriving particles can become embedded. As it cools, mineral phases can join through solid necks. Sintering changes a collection of weak contacts into a stronger solid structure. Hard-bonded slag, clinker, cement build-up and chemically bonded scale likewise demand stresses well beyond those produced by a distributed airborne field.

For those reasons, acoustic cleaning is weak or useless on sticky, wet, molten, sintered or hard-bonded deposits. A sonic horn never replaces sootblowers, ESP rappers or water washing. It may complement them by reducing the rate at which suitable dry particulate accumulates between stronger cleaning events, but each method acts through a different force and must be assigned to the deposit state it can actually change. A steam sootblower applies a concentrated jet to deposits that need more local energy, so the two methods are complementary rather than interchangeable.

What a useful horn specification must describe

A meaningful specification connects the source, acoustic field, deposit and removal path. The source description should identify the fundamental frequency and measured spectrum, not simply call the unit low frequency. It should state the air pressure and flow required at the horn during firing, the pulse duration, receiver demand and recovery time. Materials and mounting must suit process temperature, corrosion, dust ingress and pressure boundary requirements.

Acoustic output should be reported with measurement distance, direction, environment, weighting and frequency band. Coverage should then be considered at the surfaces that matter, allowing for internals, open dampers, gas conditions and likely standing-wave patterns. A near-bell SPL value cannot substitute for that work.

The process side is equally important. Engineers should identify whether the target is loose fly ash, dry filter cake, catalyst dust, powder on a hopper wall or another friable particulate. They should check temperature and moisture excursions, existing deposit condition, deposition rate and what carries released material away. If the deposit is already sticky or sintered, changing horn frequency is not a remedy.

Finally, a horn is only one component. A complete acoustic cleaning system includes the air receiver, valves, pipework, controls, sequencing, mountings and isolation needed to reproduce the intended pulse reliably. Air demand, maintenance access and avoided cleaning or outage costs should be assessed only after the physical fit has been established.

The bottom line

Sonic horns work by converting compressed-air energy into repeated, low-frequency pressure and gas-velocity oscillations. A diaphragm generates the motion, the bell couples it into the process gas, and the field applies cyclic drag and inertial forces to exposed particulate. Reflections, attenuation, wavelength and vessel geometry determine where that field is strong enough to matter.

The method is most effective as online prevention on dry, friable and loosely bonded deposits, with gravity or process flow available to remove what is released. Acoustic streaming can assist local transport, but it is a secondary effect, not a universal cleaning mechanism.

The limit is physical rather than commercial. Distributed sound can break weak particle contacts; it cannot provide the concentrated mechanical, hydraulic or thermal action needed to remove sticky, wet, molten, sintered or hard-bonded material. A sound design begins by identifying which side of that boundary the deposit occupies.