Logistic regression is usually used to model
probabilities of categorical responses as functions of covariates. However, the
link connecting the probabilities to the covariates is non-linear. We show in
this paper that when the cross-classification of all the covariates and the
dependent variable have no empty cells, then the probabilities of responses can be
expressed as linear functions of the covariates. We demonstrate this for
both the dichotmous and polytomous dependent variables.
A linear three dimensional finite element (FE) study has been carried out to examine the structural response of a prestressed concrete (PSC) inner containment (IC) dome of reactor building (double containment system) of a typical Indian Nuclear Power Plant, having large steam generator (SG) openings with due emphasis on the local behaviour of the steel-concrete interfaces at the SG openings, due to initial prestress transfer. The predominant thrust of the study has been placed on the objective of predicting the possibilities of separation at the steel-concrete interface zones adjacent to the embedded plates (EPs) of the SG openings. Two types of modeling and analysis have been made to study the overall and local behaviour of the structure. Prestressing ducts, passive reinforcements and EPs have been included in the models in certain ways. For the FE analysis, the interface zone has been modeled using interface elements, the properties of which were derived from the results of past experiments conducted on steel plate-concrete inter-face specimens. The FE analysis results have been compared with the results of the past two FE analytical studies on the linear behaviour of the same PSC IC dome. Important observations have been made regarding dome deformation and stresses throughout the structure with special emphasis on the local behaviour of steel-concrete interfaces at and around the SG openings.