The wait is over: ETABS 2013 was released last Tuesday, April 23rd. CSI was kind enough to upgrade my firm’s licenses that same day, so I’ve spent the better part of two weeks tinkering with the new version. Following more than half a decade after the release of ETABS 9, the new release has gobs of improvements and changes, meaning it will be months before I explore and understand all of the intricacies. What follows is a “first look” at the release along with general impressions.
- Better Graphics – Spinning a model in a 3D view is no longer painful.
- Improved Solvers – ETABS 2013 implements the SAPfire solver used by SAP, leading to vastly improved run times. Using the advanced solver can reduce run times by 80% or more compared to ETABS 9, while the standard solver is at least twice as fast. One caveat: the “percent force and moment errors” given in the log file by ETABS 9 are apparently no longer available, even with the standard solver.
- Rigorous P-Delta – The best ETABS 9 could do was an iterative P-Delta performed after the elastic solution. The results from this analysis could only match the AISC benchmarks if columns were split into at least two elements. In ETABS 2013, the P-Delta analysis is performed prior to running the elastic analysis, which more closely approximates a full “rigorous” approach. ETABS 2013 can match the AISC benchmark results without subdividing elements. Note that this is still not true “rigorous” analysis (where superposition does not apply), but is close enough for all but the most extreme theoretical cases.
- Text File – The $et/e2k files that permit text-based model manipulation are in the same format as version 9. This is not an improvement per se, but the fact that they are still easily readable (unlike SAFE v12 text files) is a win.
- Lateral Point Loads – After importing existing ETABS 9 models, I am consistently seeing numerical errors at points where lateral loads are applied to diaphragms. These nodes are not directly connected to model elements, but are used by the program to deliver, say, wind loads to the structure. ETABS 9 never batted an eye; ETABS 2013 returns an error message for every wind load at every floor. Even though this may be trivial by itself (no load is lost), it could obscure other ‘true’ errors in the long list.
- Frame Subdivision – The very first model I ran in 2013 was the single column AISC benchmark (See figure C-C2.2 in the 2010 Specification, page 16.1-276). When I subdivided this column into two elements, it created an overlapping double element for the top half of the column. After I found this problem and deleted the redundant member, it recreated it when the analysis was run. This error is easy to miss and is not found by the “Check Model” command, and is thus very problematic.
- 64-bit vs. 32-bit – I initially installed the 64-bit version, thinking that bigger is better. Not so. The 64-bit version does not play nicely with Microsoft Office 32-bit, meaning that exporting tables to Excel or Access is not possible. Because Microsoft discourages using 64-bit Office (due to limited compatibility with add-ins), I installed the 32-bit version of ETABS. CSI tells me and I have confirmed that analysis times do not suffer; the sole advantage of 64-bit is that the 2GB memory limitation is removed. Any model larger than this ought to be simplified anyway.
ETABS 2013 represents a major step forward for design software and will be my default software – in a few months. My initial recommendation is to not use it for production; let the chumps like me sort out the bugs. As they say with Windows: wait for the first service pack.
So, what do you like or dislike in the new version? Leave a comment and let me know!