This feature article discusses the primary enhancement of RAM Concept CONNECT Edition V6 Update 5 (release 6.5.0): Post-Tensioning (PT) Optimization. This feature is available as a technology preview in this release. The article below outlines the basic steps required to perform an optimization and highlights tools that have been added in v6.5.0 to facilitate optimizations.
Please watch our learning videos here for more help associated with this feature.
Licensing and Usage
The post-tensioning optimization feature in RAM Concept is a Bentley Cloud Service. All possible design solutions that are searched during the optimization process are analyzed in the cloud.
A cloud-based solution is necessary for the following reasons:
Cloud usage is tracked by ACUs (Analytical Compute Unit), the standard unit of consumption for Bentley analytical cloud compute resources. ACU credits are required to start an optimization.
Click here for more information on ACU entitlements and how to obtain additional ACU credits.
Steps to Complete an Optimization
Step 1 - Sign-in to Bentley CONNECT and associate a CONNECT project with the file.
Because the optimization feature is a Bentley Cloud Service, CONNECT sign-in and CONNECTED project associations are required. A project must be registered as a CONNECTED Project before it can be assigned to a model. Only site administrators or users assigned to the “CONNECT Services Admin” role by their organization’s site administrator can register new CONNECTED projects. Once a project is registered, all users in the organization can assign it to a model regardless of their assigned user roles. Click on Bentley Cloud Services - Project Portal within the RAM Concept window for links to register a new project and a link to a learning video showing how to register a project.
Site administrators can click here for step-by-step instructions on modifying user roles for individual user in their organization
Step 2 – Model tendon layout with tendon parameters, design strips, and punching shear checks.
This step would also be completed when designing PT floors without using the optimization feature. The exception is that you will typically work only with objects on the tendon parameters layer and not the manual tendons layer when optimizing tendons, because manual tendons are not optimized by the program. If you are not familiar with tendon parameters, a quick reference is available at the following link:
Automated Tendon Generation in RAM Concept
When models with manual tendons are optimized, the user will be prompted by the program to keep them in the model or delete them entirely. Manual tendons that are kept in the model are considered in the optimization calculations and can affect the results of other optimizable objects.
In previous versions, jacks and PT loss calculations were available for manual tendons only. In version 6.5.0, jack regions were added to the tendon parameter layer and allow for PT loss calculation of generated tendons.
Step 3 - Define optimizable objects and set the ranges of tendon quantities and profile elevations
Optimizable objects are the objects that are automatically modified by the optimizer. In RAM Concept, banded tendon polylines, distributed quadrilaterals, and profile polylines are optimizable objects.
To designate tendons as optimizable objects, view the properties of the banded tendon polyline or distributed quadrilateral, click on the Optimization tab, check the Optimize box, and set the range and increment values (see below). Non-integer values can be used for the specified strand number. The minimum value should be calculated based on code required minimum precompression. The maximum value should be based on practical maximum precompression limits.
To designate a profile polyline as an optimizable object, view the properties of the polyline, click on the Optimization tab, check the Optimize box, and set the range and increment values (see below). Typically, support profile polylines (high points) will be set to the highest possible elevation respecting cover requirements and defined with the Optimize box unchecked. An exception is the high point at a cantilever, which may need to be reduced from the maximum value so the cantilever is not overbalanced. The span profile polylines (low points) will typically be defined as optimizable.
Note that the elevation values set in the optimization tab refer to the “Elevation Reference” set in the general tab. For example, if “Elevation Reference” is set to “Above Soffit” for a support polyline, then the values will be measured from the bottom of the slab. We recommend using a wide range for the profile elevations, as counterintuitive values may be associated with the optimal design.
Eliminating profile polylines that you do not need is a good idea as it can reduce the number of optimizable objects and decrease optimization time. The Adjust Profile Polylines tool () on the Tendon Parameters Layers was added in v6.5.0 to automatically manipulate polylines and prepare them for optimization. This tool can automatically extend profile polylines to the slab edge, trim profile polylines to the slab edge, delete short polylines, and connect nearby endpoints for polylines with like properties (which merges them into a single polyline).
Step 4 - Define optimization regions
A new Optimization Layer was added to the program for defining optimization regions. These objects serve the following purposes:
For best performance, we recommend limiting the number of optimizable properties in any one optimization region to 50. The program enforces a hard limit of 75. If an optimization region has more than 75 optimizable properties, you will need to either adjust the optimization regions or make a change to the optimizable objects. If the whole model has 75 optimizable properties or less, then no optimization regions need to be defined. The number of optimizable objects that are associated with an optimization region can be displayed using the Optimization tab in the Visible Objects. The number of objects is also displayed in the Start Optimization dialog that appears on screen when an optimization is started.
Since optimizable objects are automatically split at the borders of optimization regions, it is best to use as few optimization regions as possible when optimizing an entire floor. When regions are required, a good rule of thumb is to define the optimization regions to match expected slab pours. This has the added benefit that tendon quantities can change at region boundaries (due to object splitting), which is also normally possible at pour breaks.
Step 5 - Review material and labor costs for PT, rebar, and SSR defined in the Estimate window
The optimal design is the design that minimizes the total material and labor cost of PT, rebar, and stud rails and has no design code failures. To calculate the material and labor costs, the program multiplies the PT, rebar, and SSR quantities by the cost factors that are defined in the Estimate window (Reports – Estimate). Because changing the cost of a single material (due to availability, labor issues, etc.) may result in a different most economical solution, it is important to set the costs in the estimate as accurately as possible for the job and location being designed.
Step 6 - Start the Optimization
A new optimization can be started by selecting the Optimize Tendons tool ( ) or clicking on Optimization > Optimize Tendons. This kicks off a preprocess that does a series of checks, and splits the tendon and profile polyline objects (if necessary). If this preprocess changes the file, you will be prompted with a “save as” dialog and can save the file with a new file name so that you can keep the original version and the modified version. After the preprocess, the Start Optimization dialog will appear (see below).
In general, the default values for the Stopping Criteria in the Start Optimization dialog will produce good results in all models.
Other important options in this dialog are the Trial Diversity slider and the “Seed with model’s current parameters” option.
The Trial Diversity slider tells the optimizer how hard to look for the best solution, with the left end representing a normal search with the lowest usage cost and the right end representing the most intensive search with the highest usage cost. For most situations, the slider can be placed at the left end end with a good chance of finding the optimal solution at the lowest cost. Although the cloud ACU consumption rate (ACU/hr) may remain the same as you move the slider to the right, the total analysis time and, in turn, the total cloud consumption will increase.
The “Seed with model’s current parameters” tells the optimizer to use the current tendon and profile parameters as a starting point in the optimization. This may reduce the total number of iterations needed to complete the optimization. You could check this option if you have manually iterated to what you think is a good design and you want the optimizer to see if it can improve it.
Step 7 - Monitor the optimization
The optimization calculations are completed off your desktop and in the cloud. The feature is designed so that it runs in the background without the need for watching or monitoring the optimization. This frees time to focus on other engineering tasks. While an optimization is running, you can also open another RAM Concept window on your machine and start work on another model.
During the optimization process, you may want to monitor convergence or cloud consumption. This can be accomplished using the Optimization Manager, which is launched by selecting the Optimization Manager tool ( )or clicking Optimization > Optimization Manager.
The Optimization Manager window shows the status of the optimization (running, stopped, run failure), the number of iterations completed, % improvement, cloud ACU consumption, and details of the total cost of each trial completed during the process. The total number of ACU’s consumed for a single optimization is calculated based on the total optimization time and the Resource Consumption Rate noted in the Trial Diversity box of the Start Optimization dialog.
The Stop button can be used to stop the optimization process at any time. You may want to stop an optimization if you see most of the trials have failures or if the optimization process seems too slow to converge to a solution. The Resume button can be used to continue trials that were stopped.
The Convergence Chart tabulates the Total Cost (y-axis) versus Total ACU’s consumed (x-axis) for the selected scenario. Three lines are plotted: Best Overall Design (w/ penalty for failures), Best Overall Design (w/o failure penalty), Best Valid Design (no failure). You can zoom into areas on the convergence chart by drawing a window over the lines in the chart.
The Trials Table at the bottom of the Optimization Manager displays the cost information and failures for each trial analyzed during the optimization. The values in the tables can be sorted by clicking on the column headers. For example, click once on one of the Cost column headers to sort the trials from ascending or descending values. If you click on the Failures column first and then Total Cost, the trials will be sorted by number of failures from least to greatest and also sorted cost by total cost. So, all trials with 0 Failures will be listed at the top of the dialog and all trials with 0 failures will also be sorted from least to greatest cost.
Scenarios and trials are saved with the .cpt file as long as the Scenario files created by the program have not been deleted. After a file is closed and reopened, you can launch the Optimization Manager to review information from previously created scenarios and load trials.
Step 8 - Load Results
During the optimization or, more often, after the optimization completes, you can use the Load Best button to load the best trial into the model. After loading the trial, generate tendons, run the calculations, and then review the tendon profiles and designs. In some cases, you may see results that are unexpected. You can make small adjustments to the tendon profiles and polylines as needed and then re-run the calculations for a final design.