Pipes

Description

The Pipe asset models a physical pipe segment in a thermal network. It represents the hydraulic and thermal transport of fluid between two nodes, accounting for pressure losses, heat losses, and flow characteristics. The asset is mapped from ESDL (Energy System Description Language) objects and is parameterised by its geometric and material properties.

Parameters

Parameter	Description	Unit	ESDL Asset Property
length	Pipe length	m	length
inner_diameter	Inner diameter	m	see remarks on diameter [1]
roughness	Wall roughness	m	roughness
alpha_value	Heat transfer coefficient	W/(m²·K)	see remarks on alpha value [2]
minor_loss_coefficient	Minor loss coefficient	—	minor_loss_coefficient
external_temperature	Surrounding temperature	K	external_temperature
qheat_external	External heat flow	W	qheat_external

[1]

The inner diameter can be specified in multiple ways. If the innerDiameter property is not null, it is used directly. Otherwise, if diameter is supplied, the inner diameter is derived from the EDR database using default schedule S1. If neither is available, a default diameter of 1.2 m is used.

[2]

The alpha value is derived from the thermal conductivity table of the ESDL asset. It uses a routine (get_thermal_conductivity_table) to extract the diameters and corresponding alpha values from the ESDL asset. If the table is empty, the alpha value will be set to 0 W/(m²·K) by default, meaning no heat loss.

Additional simulation outputs

In addition to the default per-port outputs for mass flow, pressure, temperature, and volume flow, the pipe asset provides the following additional outputs:

Signal	Description	Unit
velocity	Fluid velocity at each port, see remark [3]	m/s
pressure_loss	Pressure loss	Pa
pressure_loss_per_length	Pressure loss per unit length	Pa/m
heat_loss	Heat loss	W

[3]

The fluid velocity is multiplied by a negative sign for odd ports to ensure that the flow direction is consistent with the sign convention used in the equations.

Physics and Assumptions

The pipe asset computes hydraulic and thermal behaviour using the relationships below.

Temperature dependence of fluid properties (density, viscosity, specific heat) is accounted for using interpolation functions of the property tables for water at constant pressure.

Pressure loss

Pressure loss \(\Delta p\) is calculated using the Darcy-Weisbach equation, which accounts for both frictional losses along the pipe length and minor losses due to fittings, bends, etc.:

\[\Delta p = f \frac{L}{D} \frac{\rho v^2}{2} + K_{minor} \frac{\rho v^2}{2}\]

where:

\(\Delta p\)	Pressure loss [Pa]
\(f\)	Darcy friction factor [—]
\(L\)	Pipe length [m]
\(D\)	Inner diameter [m]
\(\rho\)	Fluid density [kg/m³]
\(v\)	Fluid velocity [m/s]
\(K_{minor}\)	Minor loss coefficient [—]

Estimation of Darcy friction factor

The Darcy friction factor \(f\) depends on the Reynolds number \(Re=\left(\rho v D\right)/\mu\) and is estimated as follows:

• For \(Re < 100\):

  \[f = 0.64\]

• For \(100 \leq Re < 2000\):

  \[f = \frac{64}{Re}\]

• For \(2000 \leq Re < 4000\), the friction factor is interpolated linearly between the laminar and turbulent values.

  \[f = f_{laminar} + \left( \frac{Re - 2000}{4000 - 2000} \right) (f_{turbulent} - f_{laminar})\]

  where \(f_{laminar} = 64/2000\) and \(f_{turbulent}\) is calculated using the turbulent correlation at \(Re = 4000\).

• For \(Re \geq 4000\) (turbulent), the Mileikovskyi and Tkachenko (2020) explicit correlation is used:

  \[f = \left( \frac{8.128943 + A_1}{8.128943 A_0 - 0.86859209 A_1 \ln \left( \frac{A_1}{3.7099535 Re} \right)} \right)^2\]

  with

  \[A_0 = -0.79638 \ln \left( \frac{\epsilon/D}{8.208} + \frac{7.3357}{Re} \right)\]
  \[A_1 = Re \frac{\epsilon}{D} + 9.3120665 A_0\]

  where \(\epsilon\) is the pipe roughness.

Heat loss

Heat loss \(Q_{loss}\) is calculated using a linear heat transfer model based on the overall heat transfer coefficient \(U\):

\[Q_{loss} = U A (T_{fluid} - T_{ext})\]

where:

\(Q_{loss}\)	Heat loss [W]
\(U\)	Overall heat transfer coefficient [W/(m²·K)]
\(A\)	Pipe surface area [m²]
\(T_{fluid}\)	Fluid temperature [K]
\(T_{ext}\)	External (surrounding) temperature [K]

Estimation of U (overall heat transfer coefficient)

The total heat transfer coefficient \(U\) is calculated as:

\[\frac{1}{U} = \frac{1}{\alpha_{fluid}} + \frac{1}{\alpha_{wall}}\]

where:

• \(\alpha_{fluid}\) is the convective heat transfer coefficient (from Nusselt number correlations)

• \(\alpha_{wall}\) is the wall (external) heat transfer coefficient

The temperature dependent properties in the overall heat transfer coefficient are derived from the inflow temperature of the fluid. For \(\dot{m} < 0\), the inflow temperature is determined by the fluid properties at the odd port, while for \(\dot{m} \geq 0\), it is determined by the fluid properties at the even port. If there is no flow (\(\dot{m} = 0\)), the inflow temperature is set to be the ambient temperature (\(T_{ext}\)).

For laminar flow (\(Re \leq 10^4\)), the Nusselt number is determined using the Graetz number. The Graetz number is defined as:

\[Gz = \frac{D}{L} Re Pr\]

and \(k\) is the thermal conductivity of the fluid, \(D\) is the pipe diameter, \(L\) is the pipe length, \(Re\) is the Reynolds number, and \(Pr\) is the Prandtl number.

For small Reynolds numbers (\(Re < 1\times 10^{-6}\)), the Graetz number is set to 10 to ensure a finite, but low, Nusselt number. This prevents numerical issues while still reflecting the low convective heat transfer in this regime.

The Nusselt number for laminar flow is then calculated as follows:

\[\begin{split}Nu = \begin{cases} 1.62 Gz^{-1/3} & \text{if } Gz \leq 0.1 \\ 3.66 & \text{if } Gz > 0.1 \end{cases}\end{split}\]

For turbulent flow (\(Re > 10^4\)):

\[Nu = 0.023 Re^{0.8} Pr^{0.33}\]

For both laminar and turbulent flow, the convective heat transfer coefficient is calculated as:

\[\alpha_{fluid} = \frac{Nu k}{D}\]

Assumptions

• Steady-state, one-dimensional flow.

• Fluid properties (density, viscosity, specific heat) are assumed constant along the pipe length but can vary with temperature.

• Heat loss is linear with temperature difference.

• No phase change or chemical reactions occur.

• Minor losses are lumped into a single coefficient.

Limitations

• No modelling of transient (dynamic) effects.

• No explicit modelling of pipe ageing, fouling, or insulation degradation.

• Assumes fully developed flow and uniform fluid properties along the pipe.

• The current documentation describes the intended physical meaning of the configured pipe parameters; if minor losses, ambient conditions, or external heat input are critical to a study, verify the active model configuration against the implementation used in the simulation setup.

See Also

• Junction — Network connectivity and flow conservation

• ProductionCluster — Heat source

• DemandCluster — Heat sink

References

References

• Mileikovskyi, V. & Tkachenko, T. (2020). Precise explicit approximations of the Colebrook-White equation for engineering systems.

• Incropera, F.P., DeWitt, D.P., Bergman, T.L., & Lavine, A.S. (2007). Fundamentals of Heat and Mass Transfer, 6th Edition.

• White, F.M. (2011). Fluid Mechanics, 7th Edition.