ASHRAE Journal - June 2011 - 40

sure losses across a fitting or series of fittings. The result is a false-positive measurement that indicates that the fittings have nothing to do with deficient performance.

but prevents laminar flow until it reaches a part of the duct system that will allow laminar flow to develop, albeit at a much lower velocity than the design called for. The reason system effect must be the result of energy lossSource of System Effect: V/CV es due to V/CV is best illustrated in Figure 3 by the lack of It was not until we visualized airflow at the molecular level any significant pressure loss (potential energy) in a 90 degree that we could determine the underlying cause of system effect fitting is illustrated even with turning vanes. Velocity losses was turbulence due to V/CV Figures 1, 2 and 3 clarify what can be significantly higher than the very small pressure loss . happens inside fan/duct systems that promote significant system indicated in design guides (SMACNA) and as measured in losses that are directly attributable to duct design and fan con- the field. The velocity losses are even greater if additional 90 nections. These figures will help you gain a better understanding degree fittings or transitions are located in close proximity. of system effect, and why it is imperative to believe it to “see” it. These are noted in an AMCA International publication as sysBy tracking a few air molecules through a system, it is easier to tem effect factors (SEFs).2 In light of our new findings, these see why the underlying problem is the result of velocity losses losses would more appropriately be called velocity loss facand not pressure losses. It is imperative to forget static pres- tors (VLFs). sures. They have nothing to do with system effect. Why? While the velocity losses are to be expected as a result of As we know, static pressure is a measure of potential energy poor duct design and poor fan connections, the exact amount whereas velocity is the conversion of potential energy to kinet- of loss is unpredictable. We have measured losses on the order ic energy. As the figures show, of 60%; but even a 10% loss not all of the potential energy has a dramatic affect on the is properly converted to kinetic amount of brake horsepower energy (velocity) as intended. required to overcome the veFigure 1 shows that when locity losses in the fan/duct air molecules encounter a flat system. surface they ricochet back into Since we can neither “see” the oncoming air molecules. nor measure system effect, it Impact Loss The first velocity loss (shown must be believed to be “seen” in red) is due to the impact and understood. This also Velocity with the flat surface. The secserves notice to designers to V/CV (Kinetic Energy) Velocity ond loss is attributable to the avoid fan and duct connecLoss subsequent impact between the tions that promote velocity ricocheting molecules and onlosses. Based on experience, Velocity Loss (Kinetic Energy) coming molecules (also shown radius fittings (as long as posFan Total SP (Potential Energy) Net Velocity Remaining in red). sible and with turning vanes (Remaining Kinetic Energy) Now consider that thousands in larger size ducts) greatly (or more) air molecules are do- Figure 1: Velocity loss on impact and loss due to counter velocity. reduce velocity losses. This is ing the same thing. The result is likely due to the fact that the a loss in velocity in the total airstream due to counter velocity. radius is less disruptive to laminar flow even in the presence Therefore, a part of the potential energy (static pressure) is con- of a change in direction. While immediate changes in flow sumed at one or more points in the system. This loss can occur direction using close fitted radius turns are to be avoided inside the fan cabinet or in the attached duct system, or both. they are far superior to square/rectangular ducts with turnIn most cases velocity losses are not accurately translated back ing vanes. Nonetheless, large radius turns will benefit from to fan total pressure. The fan’s total pressure may show that the turning vanes as well. fan is operating above design. This is yet another example of a Other velocity losses are also associated with too steep of a false-positive measurement. convergent (Figure 3) or divergent change to duct sizes particEach subsequent encounter with a flat surface within the ularly when preceded by or immediately after a duct transition. duct system will result in additional velocity losses as set out Another issue is the conversion from one duct size to anabove. Importantly, these losses can be system wide but from other one of “equal area”. For example, a 1 ft by 2 ft duct is experience appear to be more devastating the closer they are 2 ft2. A 6 in. by 48 in. duct is also 2 ft2. However the linear to the inlet or discharge of the fan. This is also true at various equivalent of one is 72 in. while the other is 108 in. It is obviterminal connections. ous that the latter will have a much greater friction loss than Since not all molecules encounter a direct impact, glancing the former. The corresponding velocity loss is even greater blows also absorb some energy. As the molecules continue to when the divergent connection at the inlet and convergent conbounce around off other molecules, a wall of turbulence (Fig- nection at the outlet of the reconfigured “equivalent” duct are ure 2) develops. The turbulent wall not only absorbs energy sharply angled.
40 ASHRAE Journal ashrae.org June 2011

ASHRAE Journal - June 2011

Table of Contents for the Digital Edition of ASHRAE Journal - June 2011