Control valves are major sources of noise in power plants and petrochemical installations. Valve noise is not amenable to theoretical models or numerical flow simulation; model tests are commonly used instead.
Scaling rules are introduced based on the thermodynamic power dissipated in the valve. Thus the reference power is only a function of flow conditions before and after the valve and is independent of valve geometry. Valves with fundamentally different design can therefore be compared directly. The scaling rules are verified with tests in flow models of different size and with compressed air and superheated steam.
The results demonstrate the importance of considering the spectral distribution of the control valve noise. Part of the noise may be generated above the audible frequency range in the model and shift to a sensitive frequency range in large scale installations. Examples are given to illustrate the importance of the spectral distribution.