No / Optimum multi-layered feed-forward neural network (NN) models using a resilient back-propagation algorithm and early stopping technique are built to predict the shear capacity of reinforced concrete deep and slender beams. The input layer neurons represent geometrical and material properties of reinforced concrete beams and the output layer produces the beam shear capacity. Training, validation and testing of the developed neural network have been achieved using 50%, 25%, and 25%, respectively, of a comprehensive database compiled from 631 deep and 549 slender beam specimens. The predictions obtained from the developed neural network models are in much better agreement with test results than those determined from shear provisions of different codes, such as KBCS, ACI 318-05, and EC2. The mean and standard deviation of the ratio between predicted using the neural network models and measured shear capacities are 1.02 and 0.18, respectively, for deep beams, and 1.04 and 0.17, respectively, for slender beams. In addition, the influence of different parameters on the shear capacity of reinforced concrete beams predicted by the developed neural network shows consistent agreement with those experimentally observed.

Dans la nature les plantes sont sans cesse soumises à des sollicitations mécaniques qui affectent et modifient leur croissance. Un aspect remarquable de cette réponse est qu’elle n’est pas seulement locale mais non-locale : la flexion d’une tige ou d’une branche inhibe rapidement la croissance loin de la zone sollicitée. Cette observation suggère l'existence d'un signal pouvant se propager à travers toute la plante. Parmi les différentes hypothèses, il a été suggéré que ce signal pouvait être purement mécanique, et provenir d’un couplage hydro/mécanique entre la déformation du tissu et la pression de l’eau contenue dans le système vasculaire de la plante. L’objectif de cette thèse est de comprendre l’origine physique de ce couplage par une approche biomimétique. Pour cela, nous avons développé des branches artificielles micro-fluidiques possédant des caractéristiques mécaniques et hydrauliques similaires à celles d'une branche d'arbre. Nous avons montré que la flexion de ces branches génère une surpression globale non-nulle dans le système, qui varie comme le carré de la déformation longitudinale. Un modèle simple basé sur un mécanisme analogue à l’ovalisation des tubes permet de prédire cette réponse poroélastique non-linéaire et d’identifier le paramètre physique clé pilotant cette réponse en pression : le module de compressibilité de la branche. A la lumière de ces résultats, des expériences sur des branches d'arbre ont ensuite été conduites et des signaux similaires sont obtenus et comparés au modèle théorique. La similitude suggère le caractère générique du mécanisme physique identifié pour la génération de signaux hydraulique dans les plantes. / Plants are constantly subjected to external mechanical loads such as wind or touch and respond to these stimuli by modifying their growth and development. A fascinating feature of this mechanical-induced-growth response is that it is not only local, but also non-local: bending locally a stem or a branch can induce a very rapid modification of the growth far away from the stimulated area, suggesting the existence of a signal that propagates across the whole plant. The nature and origin of this signal is still not understood, but it has been suggested recently that it could be purely mechanical and originate from the coupling between the local deformation of the tissues and the water pressure in the vascular system. The objective of this work is to understand the origin of this hydro/mechanical coupling using a biomimetic approach. Artificial microfluidic branches have been developed, that incorporate the mechanical and hydraulic key features of natural ones. We show that the bending of these branches generates a steady overpressure in the whole system, which varies quadratically with the bending deformation. A simple model based on a mechanism analogue to tube ovalization enables us to predict this non-linear poroelastic response, and identify the key physical parameter at play, namely the elastic bulk modulus of the branch. Further experiments conducted on natural tree branches reveal the same phenomenology. Once rescaled by the model prediction, both the biomimetic and natural branches falls on the same master curve, showing the universality of the identified mechanism for the generation of hydraulic signals in plants.