Data for the coefficient-wise geometric analysis of the ISO 9869-1 dynamic method

Description

This dataset supports the coefficient-wise geometric analysis of the ISO 9869-1 dynamic inverse method presented in the related manuscript. The study examines how the independent geometric support of the conductance-related regressor affects the sensitivity of the estimated building-envelope thermal conductance and equivalent thermal conductivity.

Steps to reproduce

The two datasets were generated using the same transient heat-conduction and regression workflow. A homogeneous one-dimensional wall with prescribed thermophysical properties was subjected either to ideal sinusoidal interior and exterior surface-temperature signals or to measured realistic surface-temperature histories. The corresponding interior heat-flux response was generated using a transient finite-element heat-conduction model. The ISO 9869-1 dynamic inverse method was then applied through successive sliding regression windows with a sampling interval of 1 h. The conductance coefficient was estimated by least squares and converted to equivalent thermal conductivity using the prescribed wall thickness. For each regression window, the dataset reports the imposed temperatures, numerically generated heat flux, estimated conductivity, least-squares residual, ISO 90% confidence-interval half-width, and the coefficient-wise quantities η and ξ. The “Prevalence” sheets group the estimates into ξ intervals of width 0.01 and report the number of estimates and maximum relative conductivity error in each interval. The realistic and sinusoidal workbooks can be used independently. The heat flux in both cases is numerically generated; only the boundary-temperature histories in the realistic case originate from measurements. Full equations, model parameters, and the coefficient-wise projection formulation are provided in the associated manuscript.

Institutions

• Institut National des Sciences Appliquées de Toulouse

  Occitanie, Toulouse

Categories

Licence