This article briefly discusses the purpose of the PT Design Optimization feature in RAM Concept and its intended use in engineering practice.
Other important information about the PT Design Optimization feature can be reviewed using the links below:
Feature Overview
The PT Design Optimization feature helps post-tensioned (PT) concrete designers develop economical floor designs in less engineering time than traditional, manual design approaches.
A common manual design method for PT concrete floors involves the following:
- Determine a PT layout that satisfies minimum precompression limits (if applicable) and maximum spacing requirements.
- Drape tendons using a load balancing approach, where the tendon profiles are set to balance a predetermined percentage of the gravity loads.
- Incrementally add PT strands and check flexural tension stress until code stress limits are satisfied.
- Document final strand quantities and tendon profiles.
- Calculate and detail required mild reinforcement and punching shear reinforcement.
Potential shortcomings with this design approach include:
- It can be tedious and time consuming due to its iterative nature.
- It forces engineers to judge the economy of their design based on experience or intuition, because schedule and budget constraints allow for only a few design iterations to be completed and compared.
The PT Design Optimization feature in RAM Concept is designed to save hours of engineering time and find an economical design that minimizes material and labor cost. It uses intelligent search algorithms to automatically compare thousands of design alternatives with varying strand quantities, tendon drapes, rebar quantities, and punching shear reinforcement requirements. It automatically weeds out invalid alternatives that do not satisfy design code requirements and systematically converges to an economical solution that minimizes material and labor cost.
Using the PT Design Optimization feature, engineers can easily compare many different solutions side-by-side and select the best design for the situation. Rather than spending hours or days producing a valid PT concrete floor design, engineers can spend a few minutes setting optimization parameters/ranges and let the PT Design Optimization feature complete the design work. Because the optimized design can be compared to thousands of other intelligently selected alternatives, the engineer can be confident that an optimal design is found.
The PT Design Optimization process also allows engineers to easily consider design alternatives that are not intuitive or that are not feasible to review with a typical, manual design approach. For example, the optimization search does a good job of adjusting profile elevations to balance the moments at column joints, which then reduces or eliminates punching shear reinforcement and decreases the demand on the columns. This is impractical for engineers to do manually because of the engineering time required.
Cloud Solution
Cloud computing and a specialized genetic algorithm are two unique aspects of the PT Design Optimization feature. The optimization calculations are based in genetic algorithms, which takes a pool of the best trials found thus far, mutates (modifies) them, and then crosses them with each other to create a new generation of trials. The best trials from this generation are identified, and this new “elite” group is used to calculate the next generation by mutating them and crossing them with each other. This process is repeated until the improvements over a number of these cycles becomes smaller than a specified convergence tolerance. Since this process requires many trials to be evaluated, cloud computing is employed to calculate many of these trials in parallel. This decreases the optimization computation time and frees the user's desktop for other processes.
The Convergence Chart in RAM Concept's Optimization Manager window () is helpful for visualizing how the optimization eliminates possible solutions with design code failures and systematically converges to an optimal design. The chart plots total design cost, which represents the total material and labor cost of post-tensioning, mild reinforcement, and stud rail reinforcement.
The image below shows the Convergence Chart that was generated when optimizing the PT Design Optimization Tutorial model (access using Help - Tutorials and Sample Models - PT Design Optimization). Note the following when reviewing this chart, which are typical observations that will apply to most successful optimizations.
- During the first few iterations, only a red line (design cost with failure penalty) and purple line (design cost without failure penalty) are plotted. At this point, the optimization search is focused on eliminating design code failures and reducing the failure penalty, as indicated by the decreasing distance between the red and purple lines during this period.
- During the 3rd iteration, a green line, which indicates that a valid design with no failures has been found, appears in the chart.
- After a valid design is found, the optimization search continues to reduce design cost.
- During the 10th iteration, red and purple lines reappear in the chart. This indicates that the lowest cost designs have one or more design code failures and shows that search investigates possible solutions that are at or just over the design code limits even after a valid solution is found.
- The optimization search steadily reduces design cost with valid solutions again until it converges to an optimal design after 70 iterations.
Feature Video
Click below to watch a short feature video: