Numerous computer analysis techniques for use in the seismic design of reinforced concrete structures are available to the design engineer and are finding general use. Unfortunately, these techniques are not "exact". Rather, they are forced to make a large number of questionable assumptions about earthquake characteristics and building behavior. To the practicing engineer, whose complex structures and structural elements defy symmetry, regularity and simplicity, the valid use of such technique depends on a complete understanding of the analysis limitations and inaccuracies and requires constant review of the results for analysis generated errors. This paper, while presenting a practical analysis application, addresses the serious difficulties and the inherent inaccuracies encountered in applying the most commonly used computer analysis techniques to concrete shear wall buildings. It is based on the actual computer analyses of a variety of middle-rise concrete shear wall buildings performed over the past few years at H. J. Degenkolb & Associates. This paper, while it addresses and identifies the invalid results that can be easily produced, believed and designed for, in concrete shear wall building analysis, also provides usable techniques for identifying, adjusting and correcting the problems that are encountered. As such, it provides the practicing engineer with additional insight and understanding of his computer analysis techniques.

Results of an experimental investigation to determine inelastic load-deformation character istics of reinforced concrete structural walls are reported. Sixteen large structural walls have been tested. These tests show that structural walls possess signif icant rotational ductility when subjected to reversing loads. In addition, it was found that shear distortions within the hinging region of a wall are coupled to flexural rotations. Therefore, inelastic shear distortions should be considered in structures designed to utilize the inelastic capacity of struc-tural walls for earthquake resistant construction.

The concepts of capacity design philosophy, which aims to ensure a desirable sequence in the plastification of ductile frames during severe seismic excitation, are introduced, In the establishment of desirable strength hierarchies the relationship between beam load inputs and ideal column strengths are examined in detail. Simple procedures, that recognise relative strength values and general characteristics of dynamic behavior, are proposed for the evaluation of moments, axial and shear forces for columns of frames. The proposals intend to ensure a high degree of protection against premature damage to columns, and to eliminate the possibility of concentrated energy dissipation in storey mechanisms in frarnes that may be subjected to unidirectional earthquake attack or to concurrent orthogonal seismic excitations. Because hinging is not expected in upper storey columns, the re-quirements for confining reinforcement in the end regions of these columns can be greatly relaxed.

Results are reported of the correlation between finite element predictions of the strength and behavior of flat slab to interior column connections transferring moment and response observed in tests on such connections. Predictions of the elastic response of the connections were made using a general elasticity program and a plate bending program. Although the general elasticity program considered shear deformations, it did not provide a markedly better correlation with the test data than the plate bending program. The plate bending analysis was extended into the inelastic range using an incremental procedure that recognized variations in stiffness and yielding of the slab with the directions of the reinforcement and principal moments. While the measured strains in the reinforcement and concrete were in reasonable agreement with the predictions, the measured deflections were about twice the predicted deflections. That discrepancy was found to be caused by bond slip of the reinforcement passing through the column and the inability of the program to correctly assess the torsional cracked section stiffness of reinforced concrete.

A finite element model which separately describes the linear and the nonlinear propertiesof steel and concrete materials is developed to analyze the cyclic inelastic behavior of a reinforced concrete plane stress shear wall. The reinforced concrete plane stress element is divided into subregions. The linear behavior of each subregion is defined by an elastic plane stress element and the nonlinear behavior is provided by a joint element of zero initial width connected to a boundary of the subregion. The stiffness matrix of the combined plane stress and joint elements is obtained from the individual material properties of concrete and orthogonally placed reinforcing steel. Four hysteretic stress-strain relationships are developed for the concrete and reinforcing steel. One model defines the cyclicbehavior of concrete under normal stresses, a second model defines the cyclic concrete shear behavior in the uncracked state and after cracking, a third model defines the cyclic behavior of reinforcing steel by a simple bi-linear model, and a fourth model defines reinforcing steel buckling in compression after crushing of concrete. The applicability of the model is demonstrated by comparing analytical solutions with experimental results obtained by PCA on two slender shear walls.