Searching an optimal experiment observation sequence to estimate the thermal properties of a multilayer wall under real climate conditions

TL;DR

This paper introduces a D-optimum criterion-based methodology to efficiently estimate the thermal properties of multilayer walls with reduced measurement durations, significantly decreasing computational costs while maintaining accuracy.

Contribution

It proposes a novel optimal experiment duration selection method using a D-optimum criterion for thermal property estimation in buildings.

Findings

Reduced measurement duration leads to lower computational costs.

The methodology accurately estimates thermal conductivity with fewer data.

Reliable physical simulation results are achieved with the estimated properties.

Abstract

The \emph{in situ} estimation of the thermal properties of existing building wall materials is a computationally expensive procedure. Its cost is highly proportional to the duration of measurements. To decrease the computational cost a methodology using a D-optimum criterion to select an optimal experiment duration is proposed. This criterion allows to accurately estimate the thermal properties of the wall using a reduced measurement plan. The methodology is applied to estimate the thermal conductivity of the three-layer wall of a historical building in France. Three different experiment sequences (one, three and seven days) and three spatial distributions of the thermal conductivity are investigated. Then using the optimal duration of observations the thermal conductivity is estimated using the hybrid optimization method. Results show a significant reduction of computational time; and reliable simulation of physical phenomena using the estimated values.