ETNA empirical validation: initial steady-state cases and simulation trials

Aim

This work seeks to include the first empirical validation test cases in ASHRAE Standard 140, Method of Test for Evaluating Building Performance Simulation Software. Such cases compare software results with empirical data that provide a physical truth standard within the uncertainty of measured data. Current Standard-140 test suites only compare software to each other (no truth standard) and/or to analytical solutions (mathematical truth standard based on assumed physics model).

Method

An experimental facility is applied with twin test cells specifically designed for empirical validation of building thermal fabric models. Both test cells are configurable as artificial climate (test cells fully guarded) or natural climate (south guard zone detached). This work begins with artificial-climate steady-state characterization cases, developed as a pre-requisite for more dynamic cases later. The cases compare recent software results with previously collected empirical data, applying a new test specification based on the experiments. The following aspects are featured:

• Steady-state heating energy measurements and their uncertainty, which provide a target range for software results
• Imputation of selected thermal conductivities based on empirical determination of as-built thermal boundary surface UA-values and interior and exterior surface heat transfer coefficients; such imputations provide model inputs that accurately characterize the test facility.

Results and Preliminary Findings

Simulation results from several models run by independent modelers allow conclusions regarding the ability to validate and improve software accuracy, such as:

• Modeling steady-state conduction – we expect this, but difficult to design and implement an experiment to prove it and then apply that as a foundation for more dynamic cases later
• Using the test suite diagnostic output to isolate: 1) input errors, allowing better test models, and 2) software issues, improving accuracy beyond the tests
• Quantifying the importance of imputing selected thermal conductivities based on measured surface UA-values in the context of two 16.5 m2 floor-area test cells, which may have greater 2-D and/or 3-D corner conduction effects than a typical building
• Validating surface heat transfer algorithms appropriate to the test cases, and the importance of comparing interior surface convection algorithms driven by forced convection (mechanical) airflow versus by natural convection (buoyancy) airflow.

Other

The paper will also summarize:

• Improvements to the test specification and software from multiple simulation trials
• Next steps to move from the foundational work to developing more dynamic artificial and natural climate test cases from the existing data
• Selected previous related research.