Determination of Poisson's ratio by spherical indentation using neural networks - Part II: Identification method

In a previous paper it has been shown that the load and the unloading stiff ness are, among others, explicit functions of the Poisson's ratio, Efa sphe rical indenter is pressed into a metal material. These functions can be inv erted by using neural networks in order to determine the Poisson 's ratio a s a function of the load and the unloading stiffness measured at different depths. Also, the inverse function possesses as an argument the ratio of th e penetration depth and that depth, at which plastic yield occurs for the f irst time. The latter quantity cannot be measured easily. In the present pa per same neural networks are developed in order to identify the value of Po isson 's ratio, After preparing the input darn appropriately, two neural ne tworks are trained, the first one being related to Set 2 of the previous pa per: In order to avoid an explicit measurement of the yield depth, the seco nd neural network has to be trained in such a way, that its solution inters ects with that of Set 2 at the correct value of Poisson 's ratio. This allo ws to identify, Poisson's ratio with high accuracy within the domain of fin ite element data.