Loading Pyodide...

Fan Type Selector

Phase-0 fan architecture triage for axial, mixed-flow, and centrifugal fans. Start from the duty point, expose the specific-speed math, estimate wheel size, and then hand the shortlist into the curve explorer.

Version v1.0
Scope Architecture screening before vendor curves
Method Specific speed + Cordier sizing
Tool Purpose & README

Loading README...

Inputs

Duty Point The required flow and pressure rise. This is the part that should stay simple. 1.00 m³/s at 500 Pa
Advanced Duty Option Pressure basis: static

Most screens should stay on static pressure. Total pressure is carried mainly for downstream curve work and verification.

Machine Limits Use only when RPM or wheel size is already fixed by the motor, package, or retrofit geometry. Unconstrained sizing
Machine Constraint ?
Geometry Override Optional family-aware override for hub or eye geometry when the internal flow section is already known. Family defaults

Leave this closed unless you already know the hub, eye, or annulus geometry. These overrides affect the reference flow-section model only; they do not replace the duty point or the Cordier sizing step.

Axial Geometry ?
Mixed-Flow Geometry ?
BC Centrifugal Geometry ?
FC Centrifugal Geometry ?
Passage & Ducting Optional packaging and acoustics realism checks, not independent aerodynamic duty variables. No passage check
Passage Geometry Check ?
Environment Air density shifts specific speed, diameter, and power. Most comfort-air cases will stay near the defaults. 20 °C, 0 m
Computed density: -- kg/m³
Scoring Weights Adjust how much each factor influences the final ranking. The tool normalises to 100%. Defaults
58%
27%
15%

Normalised: 58 / 27 / 15 = 100%

Using default values until Pyodide finishes loading.
Work:

Loading calculation engine...

Waiting for results...

Viewport Basis

Waiting for results...

Selected Family

Waiting for results...

Duty Status

Waiting for results...

Why The Flow Span Looks Like This

Waiting for results...

Why The Pressure Span Looks Like This

Waiting for results...

Boundary Rule

Waiting for results...

This is a Balje diagram: a 2D map of efficiency on the specific-speed / specific-diameter plane. The heavy black curve threading it is the Cordier line — the empirical locus of peak-efficiency points. The gray dashed curves are iso-efficiency contours showing how efficiency falls off away from the Cordier line. Colored family-peak markers on the ridge mark each architecture's natural best-efficiency Ns. Both axes are dimensionless. The efficiency field comes straight from the Python scoring model, so plotted colors match the efficiencies used by the recommendation.

What The Axes Mean

X-axis: N_s, the dimensionless specific speed. Y-axis: D_s, the dimensionless specific diameter. Both are unitless screening parameters derived from flow, pressure, density, speed, and diameter. This matches the standard Balje / Cordier orientation used in Balje (1981) and Dixon & Hall.

How The Point Is Built

Waiting for results...

How To Read The Marks

  • Waiting for results...

Why Points Can Converge

Waiting for results...

What The Efficiency Field Means

Waiting for results...

Why It Matters

The plot shows where an efficient machine for the current duty tends to live, then compares that location against the usual neighborhoods of axial, mixed-flow, and centrifugal designs.

What It Is

The Cordier line is an empirical turbomachinery correlation showing where efficient machines tend to land in dimensionless size-speed space.

How To Read It

Move right and machines trend toward larger specific diameter (slower, bigger wheels). Move up and machines trend toward higher specific speed (faster, smaller wheels).

How This Tool Uses It

The tool uses your duty point to estimate N_s, reads the corresponding D_s from the Cordier relation, and turns that into a first-pass outer wheel diameter.

Interpretation note: the line does not pick a vendor fan by itself. It tells you where an efficient machine for that duty tends to live, then the family comparison and real fan curves decide the practical answer.
Boundary note: the efficiency field is a shared per-family Gaussian envelope (Ns hump × off-ridge decay) that matches the Python scoring, not a certified performance map. Use it to read where efficient machines cluster, not as a procurement boundary.
From Inputs to N_s $$\omega_s = \frac{\omega \sqrt{Q}}{(\Delta P / \rho)^{3/4}}$$
Waiting for results...
From Inputs to D_s $$\delta_s = \frac{D (\Delta P / \rho)^{1/4}}{\sqrt{Q}}$$
Waiting for results...
Mode-Specific Calculation Path
  • Waiting for results...
These are representative family curves, not vendor curves. They are scaled so each family's nominal best-efficiency point passes through the same duty point for shape comparison only.
Contents
  1. Overview
  2. Ranking Logic
  3. Cordier Line
  4. From Inputs to Ns and Ds
  5. Reference Flow Area Model
  6. Key Equations
  7. How to Interpret the Output
  8. References

1. Overview

This tool is meant to answer the earlier question than catalog selection: given the duty point, what fan architecture should you be looking for? It stays at family level, then hands the problem into the curve explorer for real-fan comparison.

What this tool does: specific-speed screening, Cordier-based diameter estimation, family-specific reference flow area estimation from rotor annulus or inlet eye geometry, representative curve comparison, radial subtype branching, optional passage-geometry realism checks, and practical audits for nominal size, drive recommendation, acoustic risk, and architecture margin.

2. Ranking Logic

The ranking is dominated by how well the implied or unconstrained specific speed aligns with each family's natural range. Estimated efficiency and tip speed temper that recommendation, and the optional passage check adds a geometry penalty or boost.

The default recommendation now also exposes practical consequence signals that engineers often use to overrule a pure dimensionless winner: nominal hardware size, drive realism, acoustic proxy risk, and architecture margin. Those signals are meant as sanity checks and audit prompts, not as standards-grade performance predictions.

Important: the score is an architecture screening aid, not a substitute for a real fan curve, stall-margin check, or acoustic review.

3. The Cordier Line

The Cordier diagram is a classic turbomachinery map that relates specific speed and specific diameter, both of which are dimensionless. The line itself is the approximate locus of best-efficiency machines across pumps, compressors, and fans. In this tool, it is used as a first-pass bridge from a duty point to an expected fan size.

For fan screening, the practical reading is: once a duty point implies a specific speed, the Cordier relation suggests the dimensionless diameter an efficient machine would tend to want. Converting that back into a real wheel size gives the user a concrete packaging estimate instead of only a family label.

The efficiency field shown here is not copied from a single canonical standards chart. It is the max-over-families Balje envelope built from per-family Gaussians in ln(Ns) multiplied by an off-ridge exp(-β·ln²(Ds/Ds,opt)) decay. Family peaks sit on the Cordier locus by construction. That makes the plot explanatory rather than normative: it visualizes where efficient screened designs usually cluster, but it is not a stall line, procurement boundary, or catalog map.

Axes: the Cordier plot uses N_s on the x-axis and D_s on the y-axis. Both are dimensionless quantities, so the axis unit is effectively [dimensionless] or [—].
Interpretation: a point near a family peak anchor means the duty is naturally compatible with that architecture. A point between anchors means multiple architectures may be credible and the trade is likely to hinge on packaging, efficiency, contamination tolerance, or noise.
Limitation: the Cordier line is not a stall-margin predictor, acoustic model, or catalog selector. It is a physically grounded screening tool that should be followed by real fan curves.

4. From Inputs to Ns and Ds

The exact path depends on which inputs are fixed. The underlying flow Q, pressure rise \Delta P, and density \rho always come first. After that, the tool needs either a speed to compute N_s, or a diameter to compute D_s, or else it must assume a family-optimal N_s as a screening hypothesis.

Unconstrained mode: the tool assumes each candidate family's preferred N_s, then solves for the RPM and diameter that would satisfy the duty near the Cordier line. This is why unconstrained screening can compare architectures even when the user has not entered speed or wheel size.
Fixed RPM mode: the entered speed provides \omega, so the tool computes N_s directly from Q, \Delta P, and \rho. It then uses the Cordier relation to estimate the corresponding efficient D_s and outer wheel diameter.
Fixed diameter mode: the entered wheel diameter gives D_s directly from Q, \Delta P, and \rho. The tool then inverts the Cordier relation to find the matching N_s, and from that backs out the required RPM.
Interpretation: in unconstrained mode there is not a single duty-only N_s before an architecture assumption is made. The plotted candidate points are architecture hypotheses anchored to the same duty point, not one universal fan-independent point.
Why the points can collapse in constrained mode: once the duty point is paired with an imposed RPM or an imposed wheel diameter, the actual N_s and D_s are already fixed. That means all candidate types share one common actual Cordier point, and the architecture comparison comes from how far that point sits from each type's preferred neighborhood, plus efficiency, tip-speed, and passage-fit checks.

5. Reference Flow Area Model

The wheel diameter from the Cordier step is only an outer impeller size. It is not the usable flow area by itself. To estimate a real flow section, the tool now applies a family-specific annulus model and subtracts the center blockage.

Axial and mixed-flow: use the rotor annulus. The outer diameter is the predicted wheel tip diameter, the inner diameter is a family-default hub or cone diameter, and the gross annulus area is then reduced by a blockage factor.
Centrifugal: use the inlet eye annulus. The outer wheel diameter is first converted to an inlet eye diameter with a family-default eye-to-wheel ratio, then the eye hub is subtracted and the result is reduced by a blockage factor.
Limitation: this is still a first-pass geometry model. It does not solve blade width, scroll cutoff, rectangular outlet area, bellmouth details, spinner shape, or detailed blockage from blade count and thickness. It is a real area model, but not a CAD model.

6. Key Equations

Equation (1): Specific Speed $$\omega_s = \frac{\omega \sqrt{Q}}{(\Delta P / \rho)^{3/4}}$$
  • \(\omega\): angular speed [rad/s]
  • \(Q\): volume flow [m^3/s]
  • \(\Delta P\): pressure rise [Pa]
  • \(\rho\): air density [kg/m^3]
Waiting for results...
Equation (2): Specific Diameter $$\delta_s = \frac{D (\Delta P / \rho)^{1/4}}{\sqrt{Q}}$$
  • \(D\): outer wheel diameter [m]
  • \(\delta_s\): dimensionless specific diameter
Waiting for results...
Equation (3): Cordier Line $$\ln(\delta_s) = 0.833 - 0.524 \ln(\omega_s) + 0.008 [\ln(\omega_s)]^2$$

This empirical locus gives the approximate best-efficiency diameter for a given specific speed.

Equation (4): Shaft Power $$P_{shaft} = \frac{Q \Delta P}{\eta}$$
  • \(\eta\): estimated peak efficiency for the screened family
Waiting for results...
Equation (5): Tip Speed $$U_{tip} = \frac{\pi D N}{60}$$
  • \(N\): rotational speed [RPM]
Waiting for results...
Equation (6): Gross Annulus Area $$A_{gross} = \frac{\pi}{4}(D_o^2 - D_i^2)$$
  • \(D_o\): outer diameter of the modeled reference section [m]
  • \(D_i\): inner blocked diameter of the hub or eye centerbody [m]
Waiting for results...
Equation (7): Effective Area and Reference Velocity $$A_{eff} = \phi_b A_{gross} \qquad V_{ref} = \frac{Q}{A_{eff}}$$
  • \(\phi_b\): blockage factor for blades, struts, and centerbody details [—]
Waiting for results...
Equation (8): Passage Check $$A = \frac{Q}{V} \qquad \text{or} \qquad V = \frac{Q}{A}$$
  • \(A\): flow passage area [m^2]
  • \(V\): bulk passage velocity [m/s]
Inactive until a passage constraint is enabled.

7. How to Interpret the Output

Use the recommendation banner first: it gives the lead family and the first-pass physical sizing that most users actually need to act on.
Use the decision trace second: it unwraps the equations, substituted values, score logic, and practical consequence checks without forcing all of that detail onto the default screen.
Use the comparison table and Cordier tab third: they help interpret how the candidate types trade physical size and where the duty sits relative to the family neighborhoods.
Use the radial branch only when radial wins: backward-curved is usually the safer efficiency-first answer, while forward-curved is mainly justified by compactness and lower-speed packaging.

8. References

  • Balje, O.E. Turbomachines: A Guide to Design, Selection, and Theory
  • Dixon, S.L. & Hall, C.A. Fluid Mechanics and Thermodynamics of Turbomachinery
  • Eck, B. Fans
  • ANSI/AMCA 210-25 and AMCA Publication 201-23
  • DOE Improving Fan System Performance Sourcebook