RISABase uses Finite Element Analysis to obtain accurate analysis results for the base plate model.
The base plate itself is modeled using plate/shell finite elements. The base plate to bearing surface contact area is modeled using an assembly of compression only springs. The plate mesh itself is obtained by identifying the smallest distance between any two adjacent control points, otherwise called the minimum critical distance. This distance is then divided by three and the plates are then meshed into squares of this dimension.
If the coarse solution is ran, then the minimum critical distance will be divided by two, not three as in the typical mesh.
Control points are placed at the ends and intersections of the columns shape profile, at anchor bolt locations, at the ends of stiffeners, and at the edges of the base plate. (For the pipe shape, the control points are spaced evenly around the circumference).
The minimum critical distance is the distance between elements on the base plate. The critical distance that typically controls is the distance from the anchor bolt to the edge of the baseplate. RISABase will take this dimension and divide that by three and create a square plate mesh of these dimensions.
The column shape is modeled using an assembly of rigid links. Shapes that do not have all their parts welded can have the non-welded parts modeled using compression only links. (Thus a flange that is not welded will still bear down on the plate if it is in compression, but it will lift free from the plate and not induce plate bending if put into tension ).
If present, the actual stiffness of the anchor bolts is used for the analysis. The stiffness is based on the bolt area, modulus of elasticity and the anchor bolt stretch length.
The clip angle and channel type connections are modeled using link elements to connect the load point with the appropriate anchor bolt locations. In most cases this link will only transfer vertical tension and compression forces, but for the special case where you might have a column that is not directly welded to the base plate, the links will also transfer shear forces directly to the anchor bolts.
Any shear lugs in the model are assumed to be infinitely rigid and will take all the shear forces. The presence of any shear lugs is assumed to prevent any rotation that might occur due to a shear load not being coincident with the center of resistance.
The pedestal is modeled as a series of compression only springs, with a stiffness equal to the modulus of elasticity of the concrete. This value is calculated per ACI 318-11, Section 8.5
See Solution Speed for more information.
The methods used by AISC, Blodgett, et. al. to solve the base plate problem for axial load only, are based almost exclusively on the assumption of a rigid plate. The notable exception to this would be the Murray/Stockwell method used by AISC, which does attempt to model limited flexible plate behavior.
The generally accepted methods currently used to obtain results for base plates subject to uniaxial moment either assume a rigid plate, or they make assumptions about the bearing pressure distribution under the plate. These methods typically assume either an arbitrary linear stress and/or linear strain distribution (which is only valid for a rigid plate!) between the plate and the bearing surface. Anchor bolts (if any) are generally assumed to be infinitely rigid.
RISABase, by default, models a flexible plate and also uses the actual stiffness of the anchor bolts. The stress and strain distribution between the plate and the bearing surface is generally not linear.
To attempt comparisons between the other solution methods and RISABase, the user must do several things to force RISABase to match the assumptions used in the other methods.
When the above 2 items are done, the bearing pressure results and the anchor forces will very closely match the rigid plate solution results given by AISC, Blodgett, and others.
The plate stresses calculated by RISABase will only match the AISC, et al. solutions if the stresses in the plate are primarily “one way”. The AISC method to calculate a plate thickness only accounts for “one way” stress.
RISABase uses the Von Mises stress as the design stress for the plate. This value should probably be directly compared to the yield stress since it is an ultimate stress value and is combining stresses from both directions. RISABase checks this ultimate stress against a calculated allowable stress to comply with the AISC code. Since RISABase is always accounting for the actual 2 way plate stresses (2 way plate stresses occur even in plates only subject to axial load or uniaxial moment ), and RISABase detects subtle effects such as stress concentrations, it may produce slightly larger thicknesses than AISC solutions which are only looking at “one way” stresses. Depending on the engineer's familiarity with the Von Mises stress criteria and the engineer's professional judgement, the engineer could make the decision to base the plate design on the full yield or plastic stress instead of the allowable bending stress.