Dynamic vs Kinematic Viscosity

Imagine adjacent fluid layers moving at different speeds. Molecular transport and microscopic interactions resist that velocity difference. For a Newtonian fluid, shear stress is proportional to the velocity gradient, and dynamic viscosity is the proportionality factor. A larger dynamic viscosity means more shear stress is required to maintain the same rate of deformation. Honey resists shearing much more strongly than water, while air has a much smaller dynamic viscosity than either liquid.

Kinematic viscosity divides that shear resistance by density. It can be interpreted as a momentum diffusivity: the rate at which velocity differences spread through the fluid relative to the amount of mass occupying a volume. This normalization is why kinematic viscosity appears naturally in the simplified Reynolds-number expression and in some transient viscous-flow relations. It does not mean density has disappeared physically; density has already been included in the property.

For a simple Newtonian shear flow, tau = mu du/dy. The velocity gradient du/dy describes how quickly speed changes across the fluid thickness, and tau is the resulting shear stress. Dynamic viscosity therefore connects a kinematic description of deformation to a mechanical stress. It is the appropriate property in many bearing, lubrication, pumping, pressure-loss, and wall-shear calculations where forces or stresses appear explicitly.

The Newtonian qualifier matters. Water, air, light oils, and many simple fluids can be approximated as Newtonian over useful ranges, so one viscosity value at a stated temperature and pressure is meaningful. Polymer solutions, slurries, paints, blood, and other complex fluids may be shear-thinning, shear-thickening, yield-stress, or time-dependent. For them, an apparent viscosity belongs to a specified shear rate and constitutive model; a single table value can be misleading.

Common non-SI units include poise and centipoise. One centipoise equals one millipascal-second, so water near room temperature has a dynamic viscosity close to 1 cP or 0.001 Pa s. This convenient numerical relationship is easy to remember, but it should not encourage carrying cP directly into an equation written for SI units. Convert explicitly and show the unit cancellation.

Kinematic viscosity is widely used in fluid dynamics because the momentum equation for a constant-density Newtonian fluid contains mu/rho. It appears in Reynolds number as Re = V L / nu and in viscous time-scale estimates such as L squared divided by nu. A higher nu means velocity gradients diffuse across a given length more quickly relative to inertial transport, even though the underlying dynamic viscosity and density should still be considered when selecting actual forces and equipment.

Stokes and centistokes are common kinematic units. One stoke equals 1 square centimetre per second, or 1e-4 m2/s. One centistoke equals 1e-6 m2/s. Water near room temperature is therefore near 1 cSt as well as near 1 cP, but that numerical coincidence occurs because water density is near 1 g/cm3. It does not hold for most other fluids. A light oil can have 50 cSt while its dynamic viscosity depends on density and is not 50 cP unless density happens to be 1 g/cm3.

Kinematic viscosity is common in petroleum product grading, gravity-driven flow comparisons, and Reynolds-number calculations based on property tables that already report nu. It remains temperature dependent. A label such as 100 cSt is incomplete without the measurement temperature and test method because many oils thin sharply as temperature rises.

Liquid dynamic viscosity usually decreases strongly with temperature because thermal motion makes molecular rearrangement easier. Gas dynamic viscosity generally increases with temperature because faster molecules transport momentum more effectively between layers. Density usually changes more modestly for liquids, but gas density can vary greatly with pressure and temperature. Since nu = mu/rho, kinematic viscosity reflects both trends and may change differently from dynamic viscosity.

A property should therefore be tied to a state. If a pipeline calculation uses oil viscosity measured at 40 C while the actual operating temperature is 80 C, pressure loss and Reynolds number can be badly distorted. For gases, a standard-condition density combined with operating-temperature viscosity creates an inconsistent pair. Use values from the same composition, temperature, and pressure basis, and interpolate only within a range supported by the source.

Density also explains why equal dynamic viscosities do not imply equal kinematic behavior. Two fluids may resist shear similarly per unit velocity gradient but contain different mass per unit volume. The lower-density fluid has larger kinematic viscosity and therefore a smaller Reynolds number at the same velocity and length. Conversely, two fluids can share kinematic viscosity yet exert different shear stress because their dynamic viscosities differ.

Suppose a hydraulic oil has dynamic viscosity 0.082 Pa s and density 860 kg/m3 at the operating temperature. Divide 0.082 kg/(m s) by 860 kg/m3. The kilograms cancel and metres move to the numerator, giving nu = 9.535e-5 m2/s. Converting to centistokes uses 1 cSt = 1e-6 m2/s, so the result is about 95.35 cSt.

The same oil at a higher temperature might have a much lower dynamic viscosity while density falls only slightly. If mu becomes 0.026 Pa s and rho becomes 835 kg/m3, kinematic viscosity becomes about 3.114e-5 m2/s, or 31.14 cSt. This change would raise Reynolds number by roughly a factor of three for the same geometry and velocity. The example shows why a viscosity conversion without operating temperature is not enough for flow analysis.

When reporting the result, retain sensible precision. If 0.082 Pa s and 860 kg/m3 are rounded property values, 95 cSt may be more defensible than 95.35 cSt. Record the source state and whether density was measured, specified, or taken from a table. That documentation matters more than extra decimals.

Use the form required by the governing equation or source correlation. If the relation contains shear stress, torque, drag force, or pressure loss and identifies mu, use dynamic viscosity. If it contains a Reynolds number written as V L / nu or a viscous diffusion time, use kinematic viscosity. Do not convert merely because one unit feels more familiar; conversion requires a density value and introduces the assumptions attached to that density.

CFD material models commonly request dynamic viscosity because the momentum equations use stress. Some solver interfaces also accept kinematic viscosity for incompressible formulations, so the software documentation must be checked. Experimental reports may provide either property. When reproducing a case, preserve the original property basis and state rather than converting through an unrelated density. In lubrication, equipment specifications may quote kinematic grades, while film and friction calculations may still require dynamic viscosity.

If both are available from a trustworthy table at the same state, verify that mu/rho agrees with nu within expected rounding. A mismatch can expose unit confusion, a temperature mismatch, or a transcription error before it reaches a larger calculation.

A frequent mistake is treating cSt as cP because water has values near one in both units. Centistokes measure kinematic viscosity; centipoise measure dynamic viscosity. Another mistake is converting mu to nu without density, effectively assuming rho = 1 in an unstated unit system. The relationship is dimensional, so density must be present with compatible units.

Temperature mismatch is equally serious. Copying viscosity from one row of a table and density from another creates a property pair that may describe no real fluid state. Composition matters for mixtures, humidity matters for some gases, and pressure can matter for gases and high-pressure liquids. Verify the fluid grade and state before choosing significant digits.

Other errors include using a non-Newtonian apparent viscosity as though it were constant, inserting cP into an SI formula without multiplying by 0.001, and overlooking whether a software field expects dynamic or kinematic viscosity. Label every value with property type and unit; the Greek symbol alone is not enough in plain-text records where fonts and encodings can vary.

Start with a property source and note fluid, composition, temperature, pressure, and test basis. Use the Dynamic Viscosity Converter or Kinematic Viscosity Converter only to change units, not to estimate missing temperature behavior. If conversion between property types is required, obtain density at the same state and perform nu = mu/rho with a documented unit check. The Density Converter can normalize units before the division.

For a Reynolds-number calculation, decide whether the available source supports the dynamic or kinematic form. Enter geometry and velocity on the same coherent basis, then compare the result from both forms if both viscosity values are available. Agreement is a useful data check. For CFD, enter the property type expected by the solver and keep the original source beside the case setup. Repeat the calculation at plausible temperature bounds when viscosity sensitivity could change a regime or pressure-loss conclusion.

Related in this workflow: Dynamic Viscosity Converter, Kinematic Viscosity Converter, Density Converter, Reynolds Number Calculator.

Simple conversions assume one consistent fluid state and a meaningful scalar viscosity. They do not create temperature curves, pressure corrections, mixture models, or non-Newtonian constitutive behavior. Tabulated values may carry measurement uncertainty and batch variation. Published oil grades can specify a range rather than one exact value. Use manufacturer data, recognized property sources, or measurements appropriate to the required accuracy.

A correct viscosity value does not by itself validate a pressure-drop, lubrication, heat-transfer, or CFD model. Geometry, roughness, flow regime, compressibility, boundary conditions, and numerical choices remain important. For consequential equipment selection or safety decisions, check the complete method against applicable standards and validated data, and use qualified engineering judgement.