Member Designer

Get the most Optimized Sections for your designs, in a single click.

The SkyCiv Member Design Designer saves users time, by adding checking every possible section in a given database and choosing the shape with the lowest cross sectional area and best utility ratio. Currently, the Designer supports the SkyCiv steel, cold formed steel and timber member design codes (AISC, BS, Eurocode, AS, NZS, CSA, IS, AISI, AS/NZS 4600, NDS and AS 1720). There are a few criteria for the Designer to work correctly:

• The Section must be loaded from a library in the Section Builder
• The Member Design Check function must be able to run

The Designer exists on all of our steel, wood and cold formed member design software via SkyCiv Structural 3D, SkyCiv Beam and our Standalone Design software. Want a full video walkthrough? Visit our example tutorial on this feature.

Launching the Designer

To launch the Member Design Designer, you’ll need to first open up the Member Design module which you are designing your structure to. In this example we are running a Eurocode 3 (EN 1993-1-1:2005) steel check. Once the module is open, simply click the Run Designer button in the top right corner of the module. In the below model, a steel frame with HE 200 B columns, IPE 360 beams and SHS100x100x5 bracing, all members pass the check with a governing utility ratio of 0.767, so there is room to use lighter sections:

Designer Settings

Users have a number of controls and settings available, including a targeted Utility Ratio, Section Height/Width limits and which sections you wish to design. By clicking the Run Designer button, the Designer Settings will appear. Its Member Design tab holds the following options:

The General tab has the Auto Commit Changes option: when ticked, the Designer automatically commits the optimized sections into the current model.

Filters

By default, the Designer looks at all the sections of the library the section was loaded from. These options will govern what sections the Designer looks at when selecting the most optimized section:

• Section Height Between and Section Width Between: only test sections whose height or width is in the given range (in mm for metric models).
• Section Exclude List: skip any section whose name contains one of these entries. For instance, W6x will exclude any section with W6x in its name. Comma separate multiple entries.
• Section Inclusion List: only test sections whose name contains one of these entries. This overrides the Section Exclude List. Comma separate multiple entries.

Criteria

• Maximum Utility Ratio: any result with a higher UR than specified will not be considered a valid choice. The default is 0.95.
• Minimize Section Area: the Designer always looks for the passing section with the smallest cross sectional area (this option is always on).
• Stop on First Pass: on by default. Skips the remaining sections once the lightest section that meets the criteria is found, saving unnecessary results and solving time.
• Consider Buckling: also uses the buckling factor along with the Maximum Utility Ratio to decide if the criteria is met. To enable it, re-solve your model with Linear Static + Buckling.
• Re-run analysis for passing sections: on by default. Re-runs the analysis whenever a section meeting the criteria is found, whatever the Analysis Interval. This may slow down the Designer, but gives a more accurate model once the changes are committed.

Member settings

In this area of the settings, the Design Code of the Member Design module is shown. Under Design Sections by, users can choose to design by Sections (the default) or by Groups. The Designer will import whichever sections are designable in this module, but the user has the control to add/remove which sections they wish to design. The Designer will design a single section at a time.

The Analysis Interval is an important setting, and will tell the software how often to re-run the FEA solver. Its quickest option is Fast Run (an interval of 0) – meaning we will use the same internal forces when testing each section. This is usually ok, since the bending moment, shear force and axial results won’t change too much when changing a section. However, deflection results will change and can affect the results – as we use a lighter section, the deflection result will need to be updated as the self weight is reduced.

So users can increase the frequency at which the solver is run between testing sections. The only caveat here is that it will slow down the process, since the FEA solver has to run between section changes. The options are:

• Fast Run (low accuracy): does not re-run the analysis between sections. The sections are tested against the current force results.
• Recommended (balanced-quick): re-runs the analysis every 7 sections.
• Recommended (balanced-accurate): re-runs the analysis every 3 sections. This is the default.
• Slow Run (highest accuracy): re-runs the analysis between every single section test.

For large models, the default is lowered to keep the run time reasonable: balanced-quick for models with more than 500 members, and Fast Run for models with more than 2000 members.

The Search Method sets how the Designer searches the section library for a section that meets your criteria:

• Bisection (fast): the default. Jumps through the ranked list of sections, halving the range left to search after each test, and skips sections that are unlikely to meet the criteria. This can save up to 70% of the design checks. In rare cases, it might not find the absolute best section.
• Sequential (slow): tests the sections one by one, from the lightest. It is slower, but goes through the whole list.

See The Designer’s Process below for how each method works.

The Designer’s Process

The Designer runs in the following way:

1. Runs an initial analysis/design to ensure model setup is correct and to determine current results
2. Starts testing the first section:
   1. Ranks all the sections in the specified library in order of cross-sectional area. If the current section already passes, heavier sections are left out
   2. Changes a section in the model and re-runs the analysis (depending on the Analysis Interval setting)
   3. With the new analysis results (bending, deflection etc..) we test the next section, picked by the Search Method (see below)
   4. Store the results and run the next analysis
   5. With the Sequential search method, the sections are tested one by one, from the lightest. For speed and performance, any unnecessary tests will be skipped. These are if (a) the last 5 tests have produced worse results or (b) we have already found the best possible utility ratio (c) we’re failing by some large number, we will skip 5 sections
3. Start on the next section, with the updated sections from previous iterations carried through

Bisection search

With the Bisection search method, the sections are ranked by cross-sectional area, with a small weighting given to their elastic section moduli, so that sections of similar area are ordered by their bending capacity. The Designer then:

1. Starts from the current section. If it passes, it searches the lighter sections; if it fails, it searches the heavier ones
2. Tests the section halfway between the lightest section known to pass and the heaviest section known to fail, and keeps the half of the range that holds the answer
3. When a tested section fails but its utility ratio is close to the limit, it moves in smaller steps, so that more of the lighter sections around the limit get tested
4. Stops when the range can’t be narrowed any further, and keeps the lightest passing section it found

The jumps through the list are large at the start and get smaller as the search closes in. Since a large change of section changes the forces in the model the most, the Designer re-runs the analysis more often after large jumps: before the next test after a jump of more than 20 sections in the list, and at least every 2 or 3 tests for smaller jumps (or more often if your Analysis Interval is shorter). With Fast Run, the analysis is never re-run, whatever the size of the jumps.

Order of the sections

The order in which the sections are designed matters, since we will update the model as we go along if you end up with lighter elements from previous section optimizations, they will affect the results in subsequent section optimizations. If it helps, users can review the log of exactly what the Designer did under View Log when the results come back.

Consider the case where the columns are optimized first. They may pass, but then as you start designing the beams above (perhaps making them thicker/heavier) then when you go back to recheck the columns, they may fail. However, optimizing them in the opposite order, shouldn’t have this issue.

Results

The results will display in a popup. For each section, it shows the original and optimized section, the number of tests, a note, the utility ratio, the savings (the reduction in cross sectional area) and a checkbox to commit it. When the Designer made the model lighter, it also shows the total material savings (the weight of the model before the Designer run compared to the optimized model). In the below example, all 3 sections of the model were optimized, for a 26% material saving:

The Notes column tells you what happened to each section:

• No Change: no change was made to your section. Either it is the best option, or there is no passable section in the tested range.
• Optimized: this is the best option found in the tested range.
• Now Passing: the section was failing before, but now passes.
• Not Passable: no section in the range is passable.
• Best Option: this is the best option found in the tested range, but it may still be failing.
• Skipped: the section could not be optimized and was left unchanged. The reason is shown above the table.

We can take a closer look at all the runs and their results by clicking the icon under Details. This will display the following detailed list of runs. For instance, selecting section 3 (SHS100x100x5) will show the following:

From here we can sort by Max UR Ratio or cross area by clicking the column headers, and download the list with Download CSV. The original section and the selected section are labeled. This allows users to review the optimization runs to make sure the right section has been selected. In the above example, we kept the default maximum Utility Ratio of 0.95 and the Designer minimized the cross section area: it selected SHS100x100x4, the lightest section tested that still meets this design criteria.

Committing your Results

Once you’re happy with the results, you can hit Back and click Commit Changes to accept the sections that are ticked under Commit. The software will commit these shapes in both the Design module, and your Structural Model. The software will also re-solve the model and re-run your design check to give you the most up to date results. So in our above example, after committing we can see the sections are all still passing, with a governing utility ratio of 0.779: 

Changing your section from the design mode, will mean that your analysis results are out of alignment with your current model. If you change sections in any other way, we highly recommend going back and resolving your model and re running the design check to ensure your analysis and design results are in sync.

Designing Failing Members

The Designer doesn’t only make passing members lighter: when a section fails, it searches the heavier sections for the lightest one that passes. To show this, we took the optimized model from above and increased the first load combination from 1.2D + 1.6L to 1.6D + 2.4L. After solving and re-running the check, the middle columns and the long beams fail, with a governing utility ratio of 1.137:

Running the Designer again with the same settings, the failing columns and beams are moved up to heavier sections (HE 140 AA to HE 160 AA, and IPE O 330 to IPE O 360) and marked Now Passing. Since the sections got heavier, their Savings are negative. The bracing still passes, so it is left as No Change:

After committing, all the members pass again, with a governing utility ratio of 0.860:

Troubleshooting

If you cannot get the Designer to run, you should check the following:

• Are your sections loaded from the SkyCiv section database? The sections need to be loaded from the Section Builder database, this way the Designer knows which catalog of sections to test. If a section was loaded from a library table that is no longer in the current section library (a legacy table), the Designer skips it and says so: re-select the section from the Section Builder library and run the Designer again
• Can you run a Member Design “Check”? This will confirm that the sections are all designable and your model is setup correctly. Since the Designer iterates through this, it will need to run for your base model in order to function
• Do you have any special elements (tapers, rigid members)? The UI should prevent this, but in certain cases this can still slip through. Please remove those special elements from the Designer Settings and try run the Designer again
• Check your materials match the design code being tested. Again, this should be blocked by the UI but in some cases this can occur. Go back to your model and check those sections align with the design code you are trying to run. For instance, NDS should have proper NDS materials or AISC should have a steel material assigned to the section.
• Is your model an opposing unit system to the design standard unit system? We recommend running US design codes in Imperial and AS,EN,BS,CSA etc.. standards all in Metric