Fan-energy savings with vav

Variable air volume (vav) is often cited as the most energy efficient method of air conditioning. While this may be true, particularly for part-load operation, recent research has shown that such systems’ energy consumption will be strongly influenced by ductwork configuration, the degree of asymmetry in the cooling load, and the use of efficient fan speed control¹.

Previous work by UMIST reported significant additional energy savings when a vav system is configured as a looped duct rather than a conventional radial duct². It was also suggested that the location of the static pressure sensor influenced the system operation, with the potential to offer further savings.

Recent tests at the BRE have compared the performance of a duct loop system with a conventional radial vav system. The pressure loss in each branch leg, and the loop section, can be varied in the BRE test facility to represent different section lengths.

The system has two vav terminal units which are controlled from a computer according to a representative building model. The model was developed using the CIBSE Admittance method, with a cooling load profile based on a simplified two-office building.

The building was assumed to have a north/south axis with cellular offices having glazing on the east and west facades only. This produces a significant degree of asymmetry between the cooling load for the representative east and west offices (figure 2).

The system test rig was set up to represent two 30 m radial branch legs and a 10 m loop section (figure 1). Pressure losses in all ducts were based on a typical design value of 1 Pa/m at maximum flow rates.

The static pressure controller was set to control the fan speed via a Danfoss frequency inverter to maintain a static pressure at the sensor points of 300 Pa in all tests.

The configuration tests

The tests involved placing the static pressure sensor (used by the fan speed controller) in different locations to see what effect it had on fan speed. The normal industry rule of thumb for positioning a static pressure sensor is two-thirds down the index duct run from the supply fan. This was taken as the position for the sensor when it was placed in either of the two branch legs of the system, east or west.

For the loop system, the sensor was additionally placed in a position representing the end of the east or west branch legs before the loop section (figure 1).

Table 1 shows the maximum instantaneous fan energy savings for the six tests, and the overall measured fan energy consumption and percentage savings for the hours 07.00 h – 19.00 h. These hours of operation were considered typical for vav systems in office buildings.

Percentage fan energy consumption savings are calculated using the radial duct system with the static pressure sensor located in the west leg as the base case.

Figures 3 and 4 compare the influence on fan energy consumption for the two duct designs of the positioning of the static pressure sensor and show the total cooling load profile for the building.

Fan energy savings

From Figures 3 and 4 it is evident that, at the point of maximum instantaneous fan energy consumption, the looped duct system offers significant energy savings over the radial duct system. However, when the fan energy consumption is considered over a 12 h period the savings are reduced to between 5% and 7%.

The reduction in fan energy consumption of the loop duct design compared to that of the radial duct design is directly related to the level of asymmetry of the cooling load. It is also directly linked to the volume of air required by each vav unit (figure 2).

The level of asymmetry varies from zero, when both the east and west side cooling loads (volumetric air flow rates) are identical, to a maximum asymmetry when the difference in the cooling loads between the two sides of the building is at its maximum.

When the effect of this asymmetry is considered in relation to the two types of duct system, the reason for differing fan energy consumption becomes apparent.

First, in a loop system which requires an asymmetric cooling load, the air is supplied through both legs, with a proportion of the air for the high demand side passing through the loop. In this configuration the static pressure sensor is subject to only small variations in air volume flow rate and the fan energy consumption profile closely follows the profile of the total building load regardless of the level of asymmetry in cooling load (figure 4).

In the radial system, all of the air required to offset the cooling load in each side of the building is supplied down the duct leg on that side of the building. In this configuration, as the cooling load asymmetry changes, the static pressure sensor is subjected to a highly varying volumetric airflow rate.

When the required flow rate is high in the leg where the static pressure is being sensed, the fan speed increases to maintain the control static pressure at the setpoint. This results in increased fan energy consumption for this period of the day (figure 3).

It follows that for an asymmetric cooling load, the maximum difference in instantaneous fan energy consumption between the two duct systems will occur when the load is at its maximum level of asymmetry and the static pressure sensor detects pressure in the leg where the volumetric airflow rate is highest in the radial duct system.

It follows that as the level of asymmetry varies during the day, the total difference in fan energy consumption will be significantly reduced from the maximum instantaneous value and is governed by the building cooling load profile.

The position of the static pressure sensor within the loop duct system had no discernible influence upon instantaneous or total fan energy consumption. In the radial system, the instantaneous fan energy consumption was directly influenced by the volumetric airflow rate down the leg in which the sensor was sited.

When the flow rate was high in the leg in which the sensor was sited, with a high level of cooling load asymmetry, the instantaneous fan energy consumption was at a maximum. The slight difference in the level of asymmetry between the loads in the morning and afternoon resulted in a small difference in maximum instantaneous and total fan energy consumption.

Conclusions from the tests

The work has shown that previous theoretically predicted2 benefits of a looped duct compared to a conventional radial duct design may have overestimated the potential energy savings.

It is evident from this investigation that the degree of asymmetry in the load profile used on the test rig influenced the potential savings to be gained from moving from one duct configuration to another. The overall potential for energy saving is therefore not straightforward.

Further study is needed to assess system sensitivity to a broader range of parameters and values. More realistic building load profiles exhibiting a range of real asymmetry are also needed.

This work has been undertaken by the BRE for the DETR Construction Directorate.

The project was supported by Danfoss who loaned the BRE an HV-AC VLT 3000 inverter drive for the duration of the work.