|
|
|
|
|
Now, the program uses the smallest of the beam span and the user-input lateral support spacing as the lu in all cases, and the drawing better reflects a typical configuration.
The components of this problem are discussed in more detail below, along with some smaller issues that have been addressed:
The program had been using the point of zero moment in multi-span beams as if it were a point of lateral support when determining the unsupported length lu for the CL factor in NDS 3.3.3. This is because 3.3.3.3 refers to lateral support of the compressive edge, and the point of zero moment marks the start of the compressive region of the beam edge. For the following reasons, the points along the beam where the moment transitions between positive and negative values are no longer considered start and end points in the calculation of unsupported length:
Because the program was also evaluating lu in the vicinity of the maximum moment (see bug 2708), this frequently resulted in smaller than expected lu values and higher than expected CL.
For a multi-span beam, when you entered a value for intermediate lateral support, the program did not include interior supports as points of lateral support. However, NDS 3.3.3 and 3.3.3.4 say that the beam is to be laterally supported at the points of bearing in the cases of full lateral support and of lateral support by spaced purlins, respectively. This has been corrected, and the program always includes bearing support points as points of lateral support.
The program was determining the lateral support interval to be used in calculating the CL factor for NDS 3.3.3.4 as being the one in the vicinity of the point of maximum moment in a span, so that if lateral supports divided unevenly into span length, and the maximum moment lay within the small remainder distance, that distance would be used as the unsupported length lu rather than the full user input span length. As the exact spacing of the lateral support in this situation is arbitrary, and the aforementioned steel design research indicates that the entire span should be considered as a system, the program now uses the user-input lateral support spacing at all times.
Note that the program would rarely if ever use the small remainder distance, which is positioned at the right end of a span, as the lu , because the maximum positive moment usually occurs at mid-span. The maximum negative moment is usually at a support, in which case the full support distance in the span to the right of the support governs.
However, the use of this method was exacerbating the problems created by bug 2695 (above), in that small lateral support distances created by including the points of zero moments were often in the vicinity of maximum negative moment.
The drawing of the supports was such that the first support was always placed at one half the input lateral support spacing from the end of the beam.. This created a confusing situation when resulting support is the only one drawn, and at other times resulted in more supports than are necessary. Also the positioning of the support used to calculate the unsupported length lu was based on starting the lateral support spacing at the start of the beam, so was not reflected by what was shown on the screen.
The program now draws the first lateral support at the support point for the top of the beam, to reflect the fact that it is to be laterally restrained at the top. At the bottom, it draws the first support at a distance equal to the lateral support away from the bearing support point. It does so independently for each span.
The label indicating whether the input for column lateral support spacing was metric or imperial always showed "in", meaning inches, even if metric units were selected. This has been corrected.
When a known bearing length is specified that is larger than 6" (150mm) but less than 6.375" the CB factor was being applied when it should not. According to NDS Table 3.10.4, bearing lengths of 150mm (O86) or 6" (NDS) or larger should have a CB factor of 1.0.
The value stored in the database for I-joist density was converted from the original 13 lb/ cu ft to an extremely high value. This resulted in unrealistically high self weights, which cause extremely high stresses and nonsensical design. It could be avoided by either changing the value using Database Editor back to 13, or by turning off the automatic self-weight and adding the loads manually.
This has been corrected and the original 13 lb/cu ft density has been restored to the database.
Sometimes, the program was not resetting the auto-eccentricity to the correct value when examining a new section in the design search. As a result, the program was could fail to show a passing section show include a section that actually fails in the list of passing sections. Furthermore the design ratios shown in the Design Summary sheet did not match those calculated when a section was selected for the detailed design check.
The calculation of volume factor CV for fire design now uses the reduced cross section as per the note to NDS Table 16.2.2. Previously it was using the full cross section.