Résumé
In order to prevent damage to painted wooden panels from the cultural heritage sector, curators and conservators must provide the hygroscopic and thermal conditions necessary for their conservation. Furthermore, the design of physical reinforcements such as frames or cradles has to be considered. In this paper, experiments were designed to determine the kinematic response of a painting to hygrometric variations both before and after remedial conservation treatments were undertaken. This study was carried out in the Fabre Museum, Montpellier on the panel painting "La Sainte Trinité couronnant la Vierge" (« The Coronation of the Virgin by the Holy Trinity »), which was placed in a climate controlled showcase in a museum exhibition space. A glass panel on either side allowed visitors to see both sides of the panel during the experiment. At the same time, two stereoscopic devices allowed us to follow the relief, the shape deformation and strain of the panel on its front and back sides. First, the mechanical effect of the frame and cradle was studied. The majority of the deformation was found to be due to the cradle; initially the panel had a complex warp (-5 to 10 mm), decreasing to (-1 to +2 mm) when the cradle was removed. The continuous monitoring (two images per hour) was carrried out to assess the behaviour of the panel during imposed relative humidity (RH) variations of about 13% RH (between 65% RH for the maximum and 52% RH for the minimum) for both the absorption and desorption phases. The global mechanical effect on the curvature was very clear with an opposite direction of deflexion between absorption and desorption phases. A comparison of the behaviour of the panel with and without the cradle showed that the change in curvature was about twice as great (4 mm) without the cradle, and its direction moved from an initial slight negative curvature state to a positive one. In terms of strain on both the surfaces, we observed a horizontal shrinkage of about-0.1% during the desorption phase and a swelling of about +0.1% during the absorption phase for both the front and back sides of the panel. A further local study allowed us to assess the mechanical phenomenon for each board with the local determination of cupping angles and strains. This local study highlighted the change in behaviour of the panel without the cradle. On both sides, the local strains were decomposed into the mean strain and the deviation from the mean strain. The study highlighted the fact that the intensity of the mean strain (local swelling or shrinkage due to the hygroscopic changes) was the same for each configuration and was about 0.1%. Nevertheless, in