Dataset for Validation of Two-Box Spacer Modeling Across Spacer Widths
Description
This dataset contains the underlying simulation models, raw data exports, and comparative analysis worksheets for the research paper "Validation of Two-Box Spacer Modeling Across Spacer Widths for Fenestration Thermal Analysis," submitted to Energy and Buildings. The study evaluates the accuracy and limitations of two simplified two-box spacer modeling approaches across a range of spacer widths (8-20 mm) and three spacer families, using detailed THERM simulations as the baseline. The data was gathered using an automated simulation framework that interfaced with LBNL THERM 7.8.71.0 and WINDOW 7.8.71.0. 25 discrete widths were modelled for three generic spacer families using detailed geometries. Corresponding two-box models were generated, one using width-specific calibration and one using nominal-width calibration (calibrated at 8 mm only). A total of 900 full-frame simulations were performed to evaluate the total window thermal U-factor and Condensation Index. The data demonstrates that width-calibrated two-box models are stable across all widths and spacer families. However, for the intermediate conductivity spacer family (metal, double seal), the data shows a linear increase in error as the width deviates from the calibration point. Specifically using an 8 mm calibration for a 20 mm metal double seal spacer resulted in errors exceeding industry-accepted tolerances for U-factor (0.01 W/m²·K). Files are organized into folders mapping to the figures in the associated manuscript. The dataset includes: .xlsx: Calculation spreadsheets for U-factor aggregation and error plotting. .thm: THERM simulation files (serialized binary) .o THERM simulation output file (node data, used for CI Tool) .thmx THERM simulation output file (XML format .thm file, used for CI Tool) .txt Text log file .csv Result CSV file (result data exports) .mdb Access database files (used by WINDOW) Researchers can use this data to: Benchmark the accuracy of other FEA-based thermal modeling tools. Validate alternative simplified spacer modeling approaches. Test how different calibration tolerances affect full-window results. Development of standards and acceptance criteria for two-box spacer models in certification workflows.
Files
Steps to reproduce
Detailed CAD-based geometries for three spacer families (composite single seal, metal single seal, metal double seal) were prepared. Two-box models were derived from these geometries, with two rectangles stacked vertically, the bottom representing the continuous sealant layer of the same conductivity, and the top representing the effective conductivity of the detailed spacer model. All simulations were performed using LBNL THERM 7.8.71.0 and WINDOW 7.8.71.0. Condensation Index results were calculated using the NFRC CI Tool. Two distinct calibration workflows were used. First, two-box models were generated, and calibrated to their detailed counterparts for each width of spacer. Second, a two-box model was generated and calibrated at a nominal width (8 mm) and then simply scaled for the subsequent widths. The resulting spacer models were integrated into head, jamb, and sill cross-sections. Boundary conditions for the frames were set to standard NFRC 100-2023 environmental conditions. Total window U-factors were calculated in LBNL Window 7.8.71.0. Final results were exported to CSV and processed in the provided Excel calculation worksheets to determine the variance across the width range.