HAMSTAD Benchmark Test Report

Software Version: Better Building with EnergyPlus v25.1 I Date of Testing: xxxx - 1xxxx I Report Version: 1.0 I Prepared by Darren 'Dea

1. Purpose and Scope

This report documents the results of running all five HAMSTAD benchmark cases against [Software Name]. Unlike the EN 15026 benchmark, which provides a single analytical reference solution, HAMSTAD benchmarks use a reference solution band derived from several established simulation tools. Your software's results need to fall within that scatter band, which is a somewhat more forgiving criterion but also means you can't pin a precise deviation figure on a single ground truth. Welcome to the real world.

The five benchmarks test progressively more complex behaviour:

Benchmark
Primary Phenomenon
Layers
Heat Transfer

BM1

Isothermal capillary redistribution

1

No

BM2

Vapour diffusion, hygroscopic range

1

No

BM3

Capillary suction, one-sided wetting

1

No

BM4

Coupled vapour and liquid transport, interface

2

No

BM5

Full hygrothermal, realistic climate

Multi

Yes

Pass criterion: for all benchmarks, simulated profiles and time courses must lie within the envelope of results produced by the reference codes (DELPHIN, WUFI, and others) as defined in the HAMSTAD WP2 final report.


2. Software Configuration — General Notes

Before getting into individual benchmarks, document any global settings in [Software Name] that affect all five cases.

Setting
Value
Notes

Transport model for vapour

[e.g. vapour pressure gradient / RH gradient]

Transport model for liquid

[e.g. capillary pressure / water content gradient]

Coupled heat and moisture

[Yes / No / Per benchmark]

Temperature dependence of liquid transport

[Yes / No / Configurable]

Note if viscosity correction is always active

Hysteresis in moisture storage

[Yes / No / Configurable]

HAMSTAD benchmarks assume no hysteresis

Surface transfer model

[description]

If any of the above cannot be changed and conflicts with HAMSTAD assumptions, document it here and describe the workaround used. There is almost always at least one.


3. Benchmark 1: Isothermal Moisture Redistribution

3.1 Scenario Description

A single-layer homogeneous brick wall with an initial non-uniform moisture distribution is sealed on both sides. No external boundary fluxes are applied. The system evolves purely by internal capillary redistribution until equilibrium is reached. Because there is no boundary forcing and no heat transfer, this is a clean test of the liquid transport implementation in isolation.

Parameter
Value

Wall thickness

0.2 m

Material

Brick (HAMSTAD specification)

Initial condition

Non-uniform moisture profile (see BM1 specification)

Left boundary

No flux

Right boundary

No flux

Temperature

Isothermal, 20°C

Simulation period

[value] days

Profile output times

[e.g. 1, 7, 30 days]

3.2 Material Properties

Property
EN Specification
Value Used
Notes

Bulk density ρ₀

kg/m³

Specific heat capacity c₀

J/(kg·K)

Not relevant here; include for completeness

Moisture storage function

Analytical / tabulated

Liquid transport coefficient D_w (suction)

Analytical / tabulated

Liquid transport coefficient D_w (redistribution)

Analytical / tabulated

HAMSTAD distinguishes suction from redistribution

Vapour permeability δ_p / µ

Note that HAMSTAD explicitly separates the liquid transport coefficient for suction (wetting front advancing) from redistribution (moisture levelling out). If your software uses a single D_w for both, document which value was used and why.

[Insert table: Moisture storage function tabulation used as input]

[Insert table: Liquid transport coefficient tabulation used as input]

Tabulation error (moisture storage function): [value] kg/m³ max Tabulation error (liquid transport coefficient): [value] % max (log scale)

3.3 Numerical Setup

Parameter
Value
Notes

Grid elements

Grid refinement strategy

Fine grid at interfaces if any; uniform acceptable here

Time step

[value] h

3.4 Results

[Insert Figure: Moisture profiles at t = 1, 7, 30 days. Plot simulated results against the HAMSTAD reference band.]

Output Time
Within Reference Band?
Notes

[t₁]

[Yes / No / Marginal]

[t₂]

[Yes / No / Marginal]

[t₃]

[Yes / No / Marginal]

BM1 Verdict: [Pass / Fail / Marginal]


4. Benchmark 2: Vapour Diffusion, Hygroscopic Range Only

4.1 Scenario Description

A single-layer wall is exposed to a step change in relative humidity on one side. Moisture transport occurs purely by vapour diffusion within the hygroscopic range (no capillary liquid transport). This isolates the vapour transport model. Temperature is held constant throughout, so any temperature-dependent δ adjustments the software makes are operating on a fixed value.

Parameter
Value

Wall thickness

[value] m

Material

[HAMSTAD BM2 material]

Initial RH

[value] % (uniform)

Left boundary RH

[value] %

Right boundary

[sealed / fixed RH — specify]

Temperature

Isothermal, [value]°C

Simulation period

[value] days

Profile output times

[list]

4.2 Material Properties

Property
Specification
Value Used
Notes

Moisture storage function

Hygroscopic range only

Vapour permeability δ_p / µ

Constant or moisture-dependent

Liquid transport

N/A

Not active

Must be confirmed as disabled or negligible

If the software does not allow liquid transport to be explicitly disabled, confirm that the moisture range in this benchmark stays entirely within the hygroscopic domain so capillary transport would not be triggered regardless.

[Insert table: µ or δ_p tabulation used]

Tabulation error: [value] %

4.3 Numerical Setup

Parameter
Value

Grid elements

Time step

4.4 Results

[Insert Figure: Moisture profiles at required output times, compared to HAMSTAD reference band.]

Output Time
Within Reference Band?
Notes

[t₁]

[t₂]

BM2 Verdict: [Pass / Fail / Marginal]


5. Benchmark 3: Capillary Suction — One-Sided Wetting

5.1 Scenario Description

A single-layer brick specimen, initially dry (or at a low reference moisture content), is exposed to liquid water contact on one face. Moisture is drawn in by capillary suction. This is the classic sorptivity test scenario and is where most software tools show the sharpest differences, because the wetting front is steep and numerically demanding. If your grid is too coarse, you will know immediately.

Parameter
Value

Specimen thickness

[value] m

Material

Brick (HAMSTAD BM3 specification)

Initial moisture content

[value] kg/m³

Left boundary

Water contact (capillary saturation condition)

Right boundary

Sealed

Temperature

Isothermal, [value]°C

Simulation period

[value] hours / days

Profile output times

[list]

5.2 Material Properties

Property
Specification
Value Used
Notes

Moisture storage function

Must extend to free saturation

Liquid transport coefficient D_w (suction)

Critical for this benchmark

Liquid transport coefficient D_w (redistribution)

Less active here; document if combined

Vapour permeability

Secondary role

The suction-phase liquid transport coefficient is the dominant parameter in this benchmark. Tabulation resolution near free saturation matters more here than in any of the other cases.

[Insert tabulations]

Tabulation error (D_w, suction): [value] % max (log scale)

5.3 Numerical Setup

Parameter
Value
Notes

Grid elements

Fine grid near wetted surface is essentially mandatory

Grid refinement strategy

Time step

Adaptive or fixed; note if adaptive

5.4 Results

[Insert Figure: Moisture profiles at required output times, with HAMSTAD reference band.]

[Insert Figure (if available): Cumulative water uptake vs. √t — linear relationship expected in early capillary regime.]

Output Time
Within Reference Band?
Notes

[t₁]

[t₂]

BM3 Verdict: [Pass / Fail / Marginal]


6. Benchmark 4: Two-Layer System with Interface

6.1 Scenario Description

A two-layer construction (typically insulation and concrete, or mineral wool and brick, depending on specification version) is subjected to boundary conditions that drive moisture across the material interface. The interface behaviour is the point of interest: how the software handles continuity of moisture potential and flux across a discontinuity in material properties. This is where implementation details that software documentation glosses over tend to surface.

Parameter
Value

Layer 1 (left)

[Material, thickness]

Layer 2 (right)

[Material, thickness]

Initial condition

[Uniform RH or moisture content]

Left boundary

[T, RH or moisture condition]

Right boundary

[T, RH or moisture condition]

Temperature

Isothermal or coupled — [specify]

Simulation period

[value] days

Profile output times

[list]

6.2 Material Properties

Layer 1: [Material Name]

Property
Specification
Value Used

Bulk density

Moisture storage function

D_w (suction)

D_w (redistribution)

µ or δ_p

Layer 2: [Material Name]

Property
Specification
Value Used

Bulk density

Moisture storage function

D_w (suction)

D_w (redistribution)

µ or δ_p

[Insert tabulations for both layers]

6.3 Interface Treatment

Document how [Software Name] handles the material interface:

Aspect
Software Behaviour
Notes

Moisture potential continuity at interface

[Continuous RH / p_suc / w — specify]

Flux continuity

[Enforced / interpolated / describe]

Grid element at interface

[Shared node / separate elements]

Any special interface resistance applied

[Yes / No]

Must be zero for this benchmark

6.4 Numerical Setup

Parameter
Value
Notes

Grid elements (Layer 1)

Grid elements (Layer 2)

Grid refinement at interface

Finer elements on both sides recommended

Time step

6.5 Results

[Insert Figure: Moisture profiles at required output times, with HAMSTAD reference band.]

Output Time
Within Reference Band?
Notes

[t₁]

[t₂]

BM4 Verdict: [Pass / Fail / Marginal]


7. Benchmark 5: Full Hygrothermal Simulation with Realistic Climate

7.1 Scenario Description

This is the only HAMSTAD benchmark that includes coupled heat and moisture transfer under realistic (time-varying) boundary conditions. It is also the most representative of actual building physics practice, which is either reassuring or alarming depending on how the previous four went. A multi-layer wall assembly is exposed to outdoor and indoor climate data over a simulation period of one or more years.

Parameter
Value

Wall assembly

[List layers, materials, thicknesses left to right]

Climate file

HAMSTAD BM5 prescribed climate data

Indoor climate

[Constant or prescribed — specify values]

Initial conditions

[Uniform T and RH, or from pre-conditioning run]

Simulation period

[value] years

Output quantities

Temperature, RH, and/or water content profiles; surface fluxes

Output times / positions

[Specify monitor points and dates]

7.2 Material Properties (All Layers)

Repeat the material property table structure from BM4 for each layer. For BM5, thermal properties are also active and must be correctly specified.

[Layer N]: [Material Name]

Property
Specification
Value Used
Notes

Bulk density ρ₀

kg/m³

Specific heat c₀

J/(kg·K)

Thermal conductivity λ(w)

Moisture-dependent if specified

Porosity

Moisture storage function

D_w (suction)

D_w (redistribution)

µ or δ_p

(Repeat for each layer.)

7.3 Boundary Conditions

Outdoor (Left) Surface

Parameter
Source
Notes

Temperature

HAMSTAD BM5 climate file

Relative humidity

HAMSTAD BM5 climate file

Solar radiation

HAMSTAD BM5 climate file (if included)

Wind-driven rain

HAMSTAD BM5 specification

Heat transfer coefficient

[value] W/(m²K)

Vapour transfer coefficient / s_d

[value]

Indoor (Right) Surface

Parameter
Value
Notes

Temperature

[value] °C (constant or schedule)

Relative humidity

[value] % (constant or schedule)

Heat transfer coefficient

[value] W/(m²K)

s_d value

[value] m

7.4 Numerical Setup

Parameter
Value
Notes

Grid elements per layer

[list by layer]

Grid refinement at surfaces

Fine grid at exposed face

Grid refinement at interfaces

Time step

Adaptive preferred for realistic climate

Pre-conditioning run

[Yes / No, duration]

Recommended to establish realistic initial conditions

7.5 Results

Temperature

[Insert Figure: Simulated temperature at monitor positions vs. time, compared to HAMSTAD BM5 reference band.]

Monitor Position / Output Time
Within Reference Band?
Notes

Relative Humidity / Moisture Content

[Insert Figure: Simulated RH or water content at monitor positions vs. time, compared to HAMSTAD BM5 reference band.]

Monitor Position / Output Time
Within Reference Band?
Notes

Surface Fluxes (if evaluated)

[Insert Figure: Heat flux and/or moisture flux at surfaces vs. time, compared to HAMSTAD BM5 reference band.]

BM5 Verdict: [Pass / Fail / Marginal]


8. Summary of Results

Benchmark
Description
Verdict
Notes

BM1

Isothermal redistribution

[Pass / Fail / Marginal]

BM2

Vapour diffusion, hygroscopic range

[Pass / Fail / Marginal]

BM3

Capillary suction, one-sided wetting

[Pass / Fail / Marginal]

BM4

Two-layer system, interface behaviour

[Pass / Fail / Marginal]

BM5

Full hygrothermal, realistic climate

[Pass / Fail / Marginal]

Overall HAMSTAD compliance: [Pass / Fail / Partial — specify which benchmarks failed]


9. Deviations, Limitations, and Workarounds

This is the section that will actually be read most carefully if results are marginal. Be specific about what the software does and what the standard assumes, rather than just writing "minor differences exist."

Issue
HAMSTAD Assumption
Software Behaviour
Impact
Workaround

Suction vs. redistribution D_w

Two separate coefficients

[Single / dual / configurable]

[High / Medium / Low]

[Action taken]

Hysteresis

Not modelled

[Active / Inactive / Configurable]

Disable or confirm inactive

Temperature dependence of liquid transport

Not modelled in isothermal BMs

[Always active / Configurable]

Relevant to BM5 only if active

[Action taken]

Interface continuity condition

Continuity of capillary pressure

[Capillary pressure / RH / w — specify]

BM4 sensitivity

[Action taken]

Vapour transport driving potential

Partial vapour pressure gradient

[p_v / RH / specify]

Equivalent if correctly converted

Confirm conversion

[Other software-specific issue]


10. Notes and Observations

BM3 is almost always the most numerically demanding benchmark for the wetting front resolution. Note here any grid sensitivity runs performed, time-step sensitivity checks, or cases where the software struggled before the configuration was dialled in.

[Free text — include observations on convergence, run times, any unexpected behaviour, and anything that required iteration to resolve]


Appendix A: Complete Material Property Tabulations

A.1 BM1 — [Material Name]

Moisture storage function:

RH (–)
w (kg/m³)

Liquid transport coefficient (suction):

w (kg/m³)
D_w (m²/s)

Liquid transport coefficient (redistribution):

w (kg/m³)
D_w (m²/s)

(Repeat structure for BM2, BM3, BM4, and BM5 materials.)


Appendix B: Climate Data (BM5)

Reference the source of the climate data file used. Confirm it matches the prescribed HAMSTAD BM5 dataset without modification.

Parameter
Source / Value

Climate data file

[Filename / origin]

Time resolution

[value] h

Period

[dates or duration]

Any modifications from HAMSTAD original

[None / describe]


Appendix C: Software Project Files

Attach or reference the project files, input decks, or configuration exports used to run each benchmark case.

Benchmark
File / Reference

BM1

BM2

BM3

BM4

BM5

Last updated

Was this helpful?