3.4 Internal Losses with Radial Fans
129
et al. [1] and Kim and Seo [7]. These studies reveal that the size of the recirculation
zone is not constant over the periphery of the rotor and that the size of the recirculation zone gets larger at lower delivered flow rate. For zero net flow, the recirculation
flow is the strongest and creates significant energy transfer to the fluid in the volute.
For some choices of the blade angles, it is even higher than for the design flow rate.
3.4.10 Applicability of the Loss Models
Loss models, in the style as described above, are used with hand calculations of the
performance of centrifugal fans, pumps and compressors. Modern loss correlations
are mostly somewhat more complex than the formulae given above, but are inspired
by the same principles [2, 4–6]. Moreover, more detailed expressions for the loss
coefficients are employed, while the constant loss coefficients used here, give only
an approximate value, like μ i = 0.7 or c f = 0.005. Here, we only aim at a principal
discussion. For more details, we refer to the literature cited and to a recent overview
paper on loss correlations by Kim et al. [8]. In particular, this paper shows that the
dump loss model and the incidence loss model for the volute entrance are still used
nowadays, giving very reasonable loss estimates.
Performance evaluation with a one-dimensional flow representation, using
the slip formulae of Stodola or Pfleiderer and with the loss formulae is only well
justified if the flow through the rotor channels is sufficiently homogeneous in a
cross section. This requirement is satisfied for centrifugal machines with backward
curved blades operating not extremely far away from the design conditions. For
very low flow rate, the flow in the rotor becomes a recirculation flow as shown on
Fig. 3.9 (left). Such a flow cannot be analysed as a one-dimensional flow.
3.4.11 Optimisation of the Rotor Inlet of a Centrifugal Fan
In the design of a power receiving turbomachine for a constant density fluid, the
target performance parameters are the flow rate (volume flow rate Q) and the mechanical energy rise (
/ )
∆p 0 ρ . The principal machine parameters to be determined
in a first design phase are the rotor diameter ( d 2 ) and the rotational speed ( Ω). These
follow from similitude considerations (see Chap. 7), implying a global optimisation, and considerations about the application of the machine. In essence, the application type determines the degree of reaction, thus the repartition of the rotor work
into enthalpy increase and kinetic energy increase in the rotor (see Sect. 3.6). Once
the main parameters of the machine are determined ( d 2 and Ω), other parameters
typically follow from local optimisation considerations. For a radial fan, without
pre-swirl inlet vanes, it is generally assumed in turbomachinery theory that optimum efficiency is obtained by the minimum of the relative velocity at the entrance
of the rotor. The most important loss mechanism is the mixing loss at the outlet
of the rotor (dump into the volute). The second largest, but already much lower,
129
et al. [1] and Kim and Seo [7]. These studies reveal that the size of the recirculation
zone is not constant over the periphery of the rotor and that the size of the recirculation zone gets larger at lower delivered flow rate. For zero net flow, the recirculation
flow is the strongest and creates significant energy transfer to the fluid in the volute.
For some choices of the blade angles, it is even higher than for the design flow rate.
3.4.10 Applicability of the Loss Models
Loss models, in the style as described above, are used with hand calculations of the
performance of centrifugal fans, pumps and compressors. Modern loss correlations
are mostly somewhat more complex than the formulae given above, but are inspired
by the same principles [2, 4–6]. Moreover, more detailed expressions for the loss
coefficients are employed, while the constant loss coefficients used here, give only
an approximate value, like μ i = 0.7 or c f = 0.005. Here, we only aim at a principal
discussion. For more details, we refer to the literature cited and to a recent overview
paper on loss correlations by Kim et al. [8]. In particular, this paper shows that the
dump loss model and the incidence loss model for the volute entrance are still used
nowadays, giving very reasonable loss estimates.
Performance evaluation with a one-dimensional flow representation, using
the slip formulae of Stodola or Pfleiderer and with the loss formulae is only well
justified if the flow through the rotor channels is sufficiently homogeneous in a
cross section. This requirement is satisfied for centrifugal machines with backward
curved blades operating not extremely far away from the design conditions. For
very low flow rate, the flow in the rotor becomes a recirculation flow as shown on
Fig. 3.9 (left). Such a flow cannot be analysed as a one-dimensional flow.
3.4.11 Optimisation of the Rotor Inlet of a Centrifugal Fan
In the design of a power receiving turbomachine for a constant density fluid, the
target performance parameters are the flow rate (volume flow rate Q) and the mechanical energy rise (
/ )
∆p 0 ρ . The principal machine parameters to be determined
in a first design phase are the rotor diameter ( d 2 ) and the rotational speed ( Ω). These
follow from similitude considerations (see Chap. 7), implying a global optimisation, and considerations about the application of the machine. In essence, the application type determines the degree of reaction, thus the repartition of the rotor work
into enthalpy increase and kinetic energy increase in the rotor (see Sect. 3.6). Once
the main parameters of the machine are determined ( d 2 and Ω), other parameters
typically follow from local optimisation considerations. For a radial fan, without
pre-swirl inlet vanes, it is generally assumed in turbomachinery theory that optimum efficiency is obtained by the minimum of the relative velocity at the entrance
of the rotor. The most important loss mechanism is the mixing loss at the outlet
of the rotor (dump into the volute). The second largest, but already much lower,