The design of acoustic enclosures for large heat pumps and large refrigeration systems is based on the following parameters. 

1) Physical dimensions of the systems 
2) Required air volume at full load 
3) Static pressure of the fans 
4) Sound insulation to be achieved
5) Access required for service and maintenance work.

The following physical parameters must be set in the correct ratio during planning:

Air volume / free areas (open Areas) / air velocity / pressure loss.

Free air inlet and air outlet areas (net open areas in which the air can enter and exit the bonnet). These open areas must be dimensioned so that an air velocity of 7 metres/sec is not exceeded. This is due to the pressure loss of 28 Pa. at 7 metres/sec. This pressure loss is partially compensated for by the static pressure of the fans, which have to overcome a resistance on the evaporator on the air inlet side and have to expel the air above the fans on the pressure side. 

Up to an air velocity of 7 metres/sec, the ratio between heat transfer and pressure loss is ideal. At higher air speeds, the heat transfer is negatively affected and the pressure loss becomes too high. In addition, a higher air velocity leads to flow noise which results in an increased noise level.

What does this mean for example for the design of a sound bonnet for a system with an air volume of 170,000 m3/h? 

To calculate the required free area at an air volume of 170,000 m³/h and an air velocity of 7 m/s, the air volume flow must first be converted into the appropriate units in order to then use the formula for the air volume flow.

The free area for an air volume of 170,000 m³/h at an air velocity of 7 m/s is around 6.93 m².

As the exhaust air and supply air areas are separated to prevent recirculation of the air, the free area of at least 6.93 m2 must be ensured both on the intake side and at the air outlet.

Heat pumps and refrigeration systems are caught between energy-efficient performance and acoustic compatibility with the neighbourhood. A central phenomenon in this context is sound masking effects, which significantly influence both the technical design and the acoustic perception of these systems. These effects result from the complex interaction of different frequency ranges and operating conditions, which often lead to unexpected noise emissions, even if individual components have been optimised.
​
Sound masking effects describe the phenomenon in which the reduction of certain frequency ranges leads to other frequency components being perceived subjectively louder. This effect is based on psychoacoustic interactions: The human ear is less able to localise low frequencies and perceives them as more dominant when the background noise is reduced.

With heat pumps, such effects are typically caused by:

1
Frequency overlaps between fan noise (usually medium to high frequencies) and compressor noise (low frequencies),

2
Operating state-dependent modulation as typically occurs during de-icing cycles, in L/W heat pumps where fan speed changes shift the frequency spectrum.  

3
Reflection from building structures that amplify certain frequency bands through constructive interference.

Air-to-water heat pumps emit sound in the range of 30-70 dB(A), with the critical frequency bands lying between 63 Hz (low humming) and 4 kHz (high humming). Masking effects occur in particular when high frequencies are attenuated by sound insulation measures, making low frequency components more prominent in relative terms. For example, attenuating a 2 kHz signal by 10 dB(A) can result in a 100 Hz hum being perceived as 6-8 dB(A) louder

One of the main causes of masking effects are thermodynamically induced changes in operating conditions. At air temperatures around 0°C with high humidity, ice forms on the evaporator fins, which leads to the following effects:

4
Pressure losses in the air flow force higher fan speeds (frequency increase of 15-30%).

5
Compressor load changes when switching to defrosting modes generate pulse-like low-frequency oscillations

6
Material expansion on iced components cause additional resonances in the 80-200 Hz range.

These dynamic changes are superimposed on the basic noise spectrum and lead to non-linear masking effects that can hardly be detected by stationary sound measurements.

The interaction of these spectra leads to complex superpositions. For example, the 100 Hz component of the compressor can suppress the perception of 800 Hz fan sounds, while at the same time harmonics at 1600 Hz are amplified by resonances in the housing.