Control valves have long been recognised as major sources of noise in fluid process and transmission systems common to the petrochemical and refining industries. Notable advancements are the ability to predict noise and the development of quiet valves. This article deals with techniques for abatement of control valve noise. Use of vibration measurement to calculate level of noise radiated by piping systems is also discussed.
Quiet Valves:
Parameters that determine the level of noise generated by compressible flow through a valve are: geometry of flow path, differential pressure (ΔP) across the restricting elements, ratio of differential pressure to absolute inlet pressure (ΔP/P1) and the number of ports or restrictions exposed to the flow stream.
ADVERTISEMENTS:
Fig. 19.1 shows four basic approaches to the design of quiet control valves. Approach (a) utilises a small tortuous path (diameter < 1/32″) providing a flow path with a high fℓ/D ratio designed to maximise the percentage of total pressure drop obtained by viscous stresses induced by the shearing action of the fluid and boundary layer turbulence.
The noise characteristic or noise potential of a control valve increases as a function of the square of the differential pressure (ΔP2) and the ratio of differential pressure to absolute static pressure at the inlet (ΔP/P1). Thus for high pressure ratio applications (ΔP/P1 > 0.7) an appreciable reduction in noise can be effected by staging the pressure loss through a series of restrictions to produce the total pressure head loss required.
Concept (b) uses multiple series restrictions to limit the ΔP/P1 ratio across individual restrictions to optimum operating points and provides a favourable velocity distribution in the expansion area. Approaches (a) and (b) can easily be incorporated into cage-style trim fabricated from stacks of discs machined to stage the total pressure drop through a series of circumferential restrictions.
It should be noted that both (a) and (b) function as strainers and are more susceptible to plugging as a result of solid particles in the flow stream than single-stage orifical-type restrictions. Multi-staging the pressure drop of hydrocarbons can result in the formation of hydrates in the intermediate stages of pressure reduction, resulting in the plugging or blocking of the final stages.
The third approach (c) uses a multiplicity of narrow parallel slots specifically design to both minimise the turbulence level and provide a favourable velocity distribution in the expansion area. This is an economical approach to quiet valve design and can provide substantial noise reduction (15-20 dB) with little or no decrease in the total flow capacity.
ADVERTISEMENTS:
It can be shown that the acoustic power of a single flow restriction increases as a function of area squared. Changing the area by a factor of 2 results in a corresponding 6 dB change of power level; in reality, the power level is changed only 3 dB when the number of equal noise sources acting independently is changed by a factor of 2.
Critical to the total noise reduction that can be derived from utilisation of many small restrictions versus a single or few large restrictions is the proper size and spacing of restrictions so that the noise generated by jet interaction is not greater than the summation of the noise generated by the jets individually.
It has been found that the optimum size and spacing are very sensitive to pressure ratio, ΔP/P1. Recently a company has developed the technology to utilise the above approach for the design of quiet valve trim using single-stage pressure reduction. This approach, depicted by (d), provides equivalent or better noise performance than (a), (b) or (c).
For control valve applications operating at high pressure ratios (ΔP/P1 > 0.8), splitting the total pressure drop between the control valve and a fixed restriction (diffuser) downstream of the valve can be very effective in minimising the noise. To optimise the effectiveness of a diffuser, it must be designed (size and spacing of flow holes in the diffuser) for each given installation so that the noise levels generated by the valve and diffuser are equal. Fig. 19.2 depicts a typical valve-plus-diffuser installation.
Pertaining to the design of quiet valves for liquid application, the problem resolves itself to one of designing to eliminate cavitation.
Service conditions that will produce cavitation can be calculated readily. The use of staged or series reductions provides a very viable solution to cavitation and hence hydrodynamic noise.
Path Treatment:
A second approach to noise control is that of path treatment. Sound is transmitted via longitudinal waves through the elastic medium or media that separate the source from the receiver. The speed and efficiency of sound transmission depends on the properties of the medium through which it is propagated. Path treatment consists of regulating the impedance of the transmission path to reduce the acoustic energy that is communicated to the receiver.
The fluid stream is an excellent noise transmission path. When critical flow exists (fluid velocity at the vena contracta is at least at the sonic level), the vena contracta acts as a barrier to the propagation of sound upstream via the fluid. At subcritical flow, however, valve noise can be propagated in the upstream direction almost as efficiently as it is downstream.
ADVERTISEMENTS:
The impedance to the transmission of noise upstream at subcritical flow is primarily a function of valve geometry. The valve geometry that provides a direct line of sight through the valve (i.e., ball valves and butterfly valves) offers little resistance to noise propagation. Globe-style valves provide approximately 10 dB attenuation.
In any path treatment approach to control valve noise abatement, consideration must be given to the amplitude of noise radiated by both the upstream and downstream piping. Dissipation of acoustic energy by use of acoustical absorbent materials is one of the most effective methods of path treatment. Whenever possible the acoustical material should be located as close to the source as possible.
This approach to abatement of valve noise is accommodated by absorption-type inline silencers. Inline silencers effectively dissipate the noise within the fluid stream and attenuate the noise level propagated within the fluid stream. Where high mass flow rates and/or high pressure ratios across the valve exist, inline silencers are often the most realistic and economical approach to noise control.
Use of absorption-type inline silencers can provide almost any degree of attenuation desired. However, economic considerations generally limit the insertion loss to approximately 25 dB.
Noise that cannot be eliminated within the flow stream must be eliminated by external treatment or isolation. This approach to the abatement of control valve noise includes the use of heavy-walled piping, acoustical insulation of the exposed solid boundaries of the fluid stream and use of insulated boxes, rooms and buildings to isolate the noise source.
In closed system (not vented to atmosphere) any noise produced in the process becomes airborne only by transmission through the solid boundaries that contain the flow stream. The sound field in the contained flow stream forces the solid boundaries to vibrate, in turn causing pressure disturbances in the ambient atmosphere that are propagated as sound to the receiver.
Because of the relative mass of most valve bodies the primary surface of noise radiation to the atmosphere is the piping adjacent to the valve. An understanding of the relative noise transmission loss as a function of pipe size and schedule is essential to the development of the most economical approach to noise control of fluid transmission systems.
However, it should be recognised that the spectral density of the noise radiated by the pipe has been shaped by the transmission loss characteristic of the pipe and is not that of the noise field within the confined flow stream.
Acoustical insulation of the exposed solid boundaries of the fluid stream is an effective means of noise abatement for localised areas. Test results indicate that ambient noise levels can be attenuated as much as 10 dB per inch of insulation thickness.
Path treatment such as heavy-walled pipe or external acoustical insulation can be a very economical and effective technique for localised noise abatement. However, it should be pointed out that noise is propagated for long distances via the fluid stream and that the effectiveness of the heavy-wall pipe or external insulation terminates where the treatment is terminated.
Measurement of Noise Radiated by Piping Systems:
Control valve noise can be accurately predicted; however, to the best knowledge, reliable techniques for predicting noise resulting from other sources in the system (abrupt expansions, pumps, compressors, pipe fittings) are not available. Thus, actual measurement of the sound radiated to the atmosphere may be required to determine the total contribution made by the piping system to the overall noise level.
A simple sound survey of a given area will establish compliance or non-compliance to the governing noise criterion, but it will not necessarily either identify the primary source of noise or quantify the contribution of individual sources. Frequently piping systems are installed in environments where the background noise due to highly reflective surfaces and other sources of noise in the area makes it impossible to use a sound survey to measure the contribution the piping system makes to the overall ambient noise level.
Vibration measurements provide a viable technique for determining the noise level radiated by a piping system in an environment that precludes the use of sound pressure level measurements. The basic theory employed is the physical relationship between the velocity of a vibrating surface and the acoustic power radiated.
Ideally, the pressure of an acoustic wave is proportional to the particle velocity of the medium, through which the wave passes, with the constant of proportionality being the acoustic impedance of that medium. At the surface of a pipe, particle velocity is assumed equal to the velocity at which the pipe wall is vibrating. From this, acoustic wave pressure (ps) at the wall can be related ideally to wall velocity (v) by –
ps = ρocov
Where poco = acoustic impedance of the atmosphere.
Vibration measurements of the pipe wall can thus be used to calculate sound pressure levels radiated to the atmosphere. The mean squared acoustic pressure (ps2) at a point in space a distance r from the axial centreline of the pipe is related to the acoustic impedance of the atmosphere (poco), rms of the pipe surface velocity (v), pipe diameter (d) and radiation efficiency (ζ) as shown below –
The radiation efficiency (ζ) is defined as the ratio of acoustic power at the source to the acoustic power transmitted (W/Wa). The radiation efficiency has been found to equal unity above coincident frequency (fc) and is directly proportional to frequency below coincidence.
The theory is appropriate for the shell modes only. This type of response would be present if a flat vibrating plate were rolled up into a cylinder. At low frequencies, however, the shell modes are not present and the response of the pipe is due to the entire length of pipe acting as a beam. Beam modes can vibrate with very high amplitudes; however, the efficiency of their coupling to the acoustic field on the outside of the pipe is extremely low.
This means that a vibration measurement may indicate high energy content at low frequencies with very little contribution to the observed sound pressure level. Frequencies associated with the lowest shell mode should be considered as a low frequency cut-off for the direct applicability of the theory. This is generally not restrictive in evaluating control valve noise or other broadband and high-frequency noise.
Obviously the accuracy of Sound Pressure Levels (SPL) calculated from vibration measurements is limited to the accuracy of the vibration measurements. The technique is restricted to measurements taken a minimum of two diameters from an end connection of a straight run of pipe. Critical to accurate measurement of pipe vibration is the attachment of the accelerometer to the pipe wall.
Ideally, the accelerometer should be rigidly attached to a small metal pad that is welded to the pipe. Also, some device should be used to electrically isolate the accelerometer from the pipe, such as an insulated stud or washer. This attachment method will yield valid information over the entire frequency range for which the particular probe is specified.
An alternative method is to attach pads or studs to the pipe wall, using an adhesive. As long as a stiff, thin-layered adhesive is used, this method can be effective over the specified range of probe. Different adhesives are necessary depending on the temperature of the application. Magnetic attachments should be of special design to give a firm bond to a cylindrical surface.
Even with a good magnetic attachment, the high-frequency response is limited. If a magnetic base is used, then the surface should be clean of paint and dirt to ensure maximum contact. Hand-held accelerometers generally are limited to very low- frequency measurements.
It should be recognised that vibration measurements are not a panacea for all of the problems associated with noise analysis of a fluid transmission system. Sound and/or vibration surveys can quantify the noise levels radiated by a fluid transmission system; however, because the noise generated within the flow stream is both structure-borne and fluid-borne, the primary source of noise generation is not necessarily obvious, thereby making it difficult to know which sources (if any) are controlling the sound field.
Closure:
A substantial amount of progress in the area of control valve noise technology has been achieved in the last very few years. Among the most notable advancements are the ability to predict accurately the level and spectral density of control valve noise radiated to the atmosphere via the adjacent piping and the development of quiet valve trim.
From a purely noise consideration, there are few control valve installations that can be considered truly standard. They will be unique from the standpoint of installation geometry, service conditions, noise attenuation requirements or some combination of these. With so many possible installation variables and the numerous pieces of control valve noise abatement equipment, it becomes extremely important that knowledgeable persons are consulted in the application of this equipment.
For example – several approaches may be taken to the same problem. One approach might produce the very quietest installation but at a prohibitive cost, whereas another approach could meet the required noise specification at a substantial saving. Without the ability to predict noise levels and without the choices of equipment, optimising a given installation from the noise and cost standpoints would not be possible.
Where do we go from here? Comparing current noise technology with other important control valve technologies, such as systems analysis and valve sizing, indicates that the noise technology is in an emerging status. If this is the case, then dramatic progress can be expected in the immediate future.
Studies in progress are intended to increase the understanding of the noise generation mechanisms and identify parameters not presently being considered. These studies and others should generate new and more efficient items of equipment and result in more precise techniques for the prediction of control valve noise.
Hydrodynamic Control of Valve Noise:
Liquid flow through a control valve (called hydrodynamic flow) can and often does create noise. There are three categories of hydrodynamic noise: noise from non-cavitating liquid flow, cavitating liquid flow and flashing liquid flow. Of the three, cavitating flow is the major noise problem. Laboratory testing and field experience show that non-cavitating and flashing liquid flow noise levels are quite low and generally not a problem.
Cavitation is a two-stage phenomenon. The first stage involves the formation of vapour bubbles in the fluid stream. As liquid flow passes through the orifice of a control valve, its velocity causes pressure at the vena contracta to drop below the vapour pressure of the liquid and vapour bubbles are formed.
The vena contracta is the point beyond an orifice where the flow cross-section is smallest, pressure is lowest and velocity is highest. However, since clearance between the valve plug and seat ring is the primary restriction in most conventional valves, the vena contracta is formed near the valve seat line.
The second stage is the implosion of these vapours bubbles. As the fluid moves downstream from the vena contracta into a large flow area, velocity decreases with a resulting pressure recovery. When static pressure exceeds the vapour pressure of the liquid, the vapour bubbles implode, generating extremely high-pressure shock waves that hammer against the valve outlet and piping. (Pressures in the collapsing cavities reportedly can approach 1,00,000 psi in magnitude). Noise and damage result.
Solving cavitation problems begins first with either controlling the cavitation process or ideally, eliminating cavitation altogether. In controlling cavitation, the techniques employed are often defensive in nature. For example – valve parts subject to damage are furnished in hardened materials in an attempt to extend valve life against the erosion and shock generated by the imploding vapour bubbles. Another technique simply lets cavitation exhaust itself by destroying some sacrificial part of the piping system (an elbow downstream of the valve orifice plate).
A third technique, developed within the past few years, involves special cage-type valve trim for globe valves that moves the primary fluid restriction away from the valve plug seat line. In new trim design, a number of pairs of small, diametrically opposed flow holes are located in the wall of the cage. As the valve plug moves away from the seat, increasing numbers of these holes (always in pairs) are opened to the inside of the cage.
Each hole admits a jet of cavitating liquid that ensures substantial pressure recovery in the centre of the cage. Collision of this jet with that of the opposing hole creates a continuous fluid cushion. The cushion in turn prevents cavitating liquid from contacting the valve plug and seat line of the valve. Under certain conditions, this trim can reduce valve noise as much as 6 dB.
Ball valves are often used in installations that require wide ranges of flow control. A ball valve, however, is a device having a greater pressure recovery from the vena contracta pressure than that given by a globe valve? This pressure recovery advantage has its price in that a cavitating condition is more easily reached.
An answer to this problem is a special trim consisting of a bundle of parallel flow tubes that, under certain conditions, create a back pressure within the ball valve to keep static pressure above the vapour pressure of the fluid. However, when cavitation does occur, these flow tubes restrict the size and number of vapour bubbles. Damage tests with soft aluminium rods indicated that damage downstream of the flow tubes was insignificant as compared to damage without them.
When cavitation control methods are used, noise reduction is gained by applying acoustical insulation on the valve and associated piping, by using heavy-walled piping, by installing the valve in an enclosure or by burying the pipeline. These techniques commonly referred to as path treatment, do not reduce the level of noise carried in the fluid stream; they only shroud it. Therefore, it is important to note that where path treatment stops, fluid noise may annoyingly reappear.
If the cavitation is eliminated, cavitation-created noise is also eliminated. Several techniques can be applied to eliminate cavitation. One involves placing the control valve within the piping system at a point where pressure drop and fluid temperature conditions will not create cavitation. If this proves impossible, then two or more valves in series, each taking a portion of the total desired pressure reduction (called staging), can be used to prevent dropping pressure within the valve below the vapour pressure of the fluid.
The technique of staging has been designed into special valve trim for globe bodies. An example of this style trim is a cage consisting of one or more concentric, cylindrical sections that contain specially drilled orifices. In operation, each section stages the pressure drop, the number of stages required depending on the inlet pressure and the total pressure drop across the valve.
This type of trim would be applied normally when the pressure drop is in the 1000 to 3000 psi range and it can be characterised when the pressure drop decreases with increasing valve travel. The characterisation provides pressure staging and cavitation prevention at low valve plug lift, with the staging effect becoming progressively less as the pressure drop becomes smaller at greater valve travel.
A problem is encountered in this trim design as the pressure drop exceeds 3000 psi. With the valve plug off its seat, the pressure drop-staging devices below the plug seat line are open to flow and perform as designed. However, pressure drop staging does not occur in those devices still blocked to flow; as a result, full inlet pressure registers against the valve plug through them.
Due to plug-cage clearances, flow at nearly full inlet pressure moves between the plug and cage into the downstream pressure area. This high-pressure, high-velocity flow stream exiting into the downstream flow passage may cavitate near the valve plug seat line. The higher the inlet pressure, the greater the possibility of damage.
This clearance flow problem is solved by a newly-developed trim that stages both normal flowing pressure drop and all clearance flow between the valve plug and cage.
To optimise the noise control effort, it is essential first to determine existing or potential operating noise levels. Through research, a hydrodynamic noise prediction technique has been developed that involves the factors of valve style; size and type of trim, size and schedule of adjacent piping; inlet pressure, pressure drop and liquid vapour pressure; and valve capacity.
There are no control valve industry measurement standards with regard to distance from the noise source; however, since noise attenuates with distance, some reference point always should be selected and recorded. Use of this prediction technique provides a valid and ready determination of valve’s noise level. The noise level then becomes the basis for evaluating available noise treatment methods.
Should the predicted noise level exceed established noise standards, a choice must be made whether to:
(i) Control cavitation and use path treatment methods; or
(ii) Prevent cavitation.
At first glance, the simple answer is to prevent cavitation, but in actuality, the decision is based on economic trade-offs. Table 19.1 provides a general comparison of noise control effectiveness and economics for the various techniques.
