Process plant noise control: preventive versus remedial noise control
The noise emission from process plant equipment is now routinely considered at the design stage when the overall specification for the plant is drawn up. Preventive noise control is the most cost-effective approach to ensuring the final installation does not give rise to complaints. However, noise specifications are frequently taken from other contracts without suitable modifications to allow for the specific requirements of the project of concern. Sometimes this works, but if it does not, then the remedial noise control can be very expensive.
For a noise specification to be effective the engineer responsible should ask himself a number of questions. The answers to these questions will lead to a noise specification that is meaningful in the context of the project and should be enforceable at completion of the project. The questions are:
- Why is the noise specification required? For example, is it for hearing conservation or is it for environmental constraints?
- Where does the noise specification have to be met? By the side of the equipment, at the boundary of the plant or at a more remote location?
- Is the target to be met in terms of overall dB(A) levels or frequency band levels?
- Should the specification be described as sound pressure levels or sound power levels? The engineer needs to understand the difference between the two parameters.
- At what equipment operating conditions should the noise specification be met? Normal operations, maximum conditions, etc.
- Should suppliers quote unsilenced or silenced equipment noise levels, or both, in tenders?
- How are intermittency and tonality of the noise to be defined and treated?
- Are there any recognised guidelines that can be referred to in the specification to reduce ambiguity, such as ISO or BS standards, Industry guidelines (such as EEMUA 140)?
While these questions do not cover every aspect of a noise specification they will assist the engineer to ensure the equipment specification is meaningful in the context of the project. Unfortunately even with a carefully prepared noise specification there are times when noise problems arise that could not have been foreseen.
ISVR Consulting has been involved in two such cases recently, one in a power station and the other in a refinery. In each case sophisticated methods were employed to identify the cause of the noise and vibration. Some of the issues addressed have wide applicability so it is useful to consider these cases further.
Gas turbine power station
The power station was new and housed two 48 megawatt gas turbines. During commissioning unusually high noise and vibration levels arose during certain operating conditions. They were so severe that concerns were expressed about potential structural damage to the installation.
An extensive series of tests was carried out at a range of operating conditions to explore the problem. The most striking aspect of the measurements was that the offending noise only occurred when water injection took place. Without water injection, to reduce the particulate emissions, the noise and vibration levels were acceptable and did not display the offending tones irrespective of the power setting.
Using correlation techniques it was demonstrated that there was a strong coincidence between the vibration of the water injection feed pipes to the combustor and the external sound pressure level during water injection.
Based on this, and the other test results, it was concluded that the problem was a pulsation in the water feed line that induced a pulsation in the combustion process which, in turn, caused the strong tonal noise in the combustion exhaust.
Reducing noise from a catalytic distillation column
This project concerned the noise from a Cat Cracker exhaust stack, used for the purpose of converting heavy oil into gasoline products. Following the upgrading of the Cat Cracker, there were persistent community complaints of an irregularly varying noise that sounded like an “overflying jet aircraft”. Since the exhaust from the plant ran at about 300°C with a high, in-duct exhaust velocity it was not possible, at the time, to measure the noise inside the exhaust stack.
Noise measurements were carried on-plant, at the stack tip and in the community. Unlike many process plant noises this one was irregular in occurrence and short in duration so short sections of the recordings of the noise events were analysed in order to characterise the offending noise. From the recordings, and from theoretical considerations, it was surmised that the problem was the interaction of turbulence that was shed from a process valve striking a downstream Multi-Holed Orifice plate (MHO) just before the flue gas entered the exhaust stack.
A scale model of the process was constructed in the laboratory and tests carried out to explore qualitatively the nature of the potential interaction between the valve and the MHO in the stack. The tests concluded that:
- The close proximity of the MHO to the valve caused an increase in the low frequency noise, which was consistent with that found at the stack tip.
- The noise level increased for reduced separation between the valve and the MHO.
- The noise level increased for increased outlet velocity for a fixed arrangement of valve and MHO.
- Substantial reductions in noise arose when more than one MHO was installed in the duct.
No single test of the many tests that were carried out proved conclusively the cause of the excessive noise levels and the reason for the fluctuations in levels. However, the accumulated evidence strongly supported the theory that the source of noise was due to turbulent interaction between the valve and the MHO. At the stack tip the noise level was steady. A separate study, using computer modelling, confirmed that the effects of the wind and turbulence were the probable cause of the irregular fluctuations in the measured noise in the community.
Following the investigation the existing MHO and valve were replaced by three MHO’s in series. This arrangement was chosen so that the valve in question could be removed and the required pressure drop maintained with the three MHO’s spaced as far apart as possible.
Following these modifications a repeat survey was carried out in the community. The results showed a reduction in the community noise of between 10 dB and 14 dB at the critical frequencies.
The purpose of this short article is to stress that noise must be considered at the design stage of a project and the noise specification must be specific to the project of concern if noise control is to be cost-effective. However, not all noise problems are foreseeable. When these arise then sophisticated and systematic investigations may be needed to identify the cause if effective remedial treatment is to be successful.
Further details of the catalytic converter project have been published in these technical papers:
- Reducing noise from an oil refinery cat cracker R D Rawlinson, J Alberola and P Joseph. Paper No 2103. Proceedings of Internoise 2005 Rio de Janeiro, Brazil. (pdf file, 270 KB)
- Reducing noise from an oil refinery catalytic distillation column R. D. Rawlinson, J. Alberola, P. F. Joseph, and M. G. Smith. Noise Control Engineering Journal, 54 (6), Nov-Dec 2006. reprinted with permission (large pdf, 5281 KB)
Sound power measurements for hot exhaust stacks
The continuation of the Cat Cracker project involved noise measurements being taken inside the high temperature stack using a twin microphone technique developed at the ISVR. Further details are presented in the following paper:
- Sound Power Measurements for Hot Exhaust Stacks R. David Rawlinson and Philip F. Joseph. Paper 448 Proceedings of Internoise 2006 Honolulu, Hawaii. (pdf file, 560 KB)