Which design method do you prefer?
Hello, everyone. Sorry I’ve been away for so long but I’ve gotten busy. I had to take over teaching a Judo class for a little bit and school has also been very busy for me this last semester. Along with that, I want to make my blog post a bit more technical while releasing fewer of them. This will fall a little bit shorter than I expected in that department; however, it is something that I think is a critical base of knowledge to have as an engineer. In this post, I want to talk about the advantages and disadvantages of different types of design methods. I will compare four main types: allowable stress design, strength based design, performance based design and prescriptive design.
Allowable Stress Design has been the most common design method in recent engineering practice. The design process involves using structural analysis methods to find stress imparted on the structure by the design loads and comparing that to a stress limit for the structure. A safety factor is applied to the stress limit to ensure that the design is safe. This method is the most common design method – especially for older buildings. The benefits are that it is a shorter and simpler design calculation and it is easier to select a design based on shape alone. This approach also accounts for more variability in material. The drawbacks are that it requires elastic behavior to be assumed and it can be overly conservative when compared to strength based design.
Strength Based Design, also known as Load and Resistance Factor Design (LRFD) in civil engineering, is a strength based design where the required strength is calculated from the load on the structure using structural analysis methods. This strength requirement is compared to the design strength that the structure can carry. A factor is applied to the load and structure strength to ensure that the design is safe. These factors account for different levels of uncertainty for different loading and failure scenarios and allow for more efficient design better than ASD. This method also has the added benefit accounting for inelastic behavior which is better for extreme events like earthquakes and extreme weather events. The drawbacks are that it is a longer design process and that the methods may not work well with older designs that use ASD. Along with that, it requires consistent material properties – if adequate information for that can’t be found ASD has to be used.
Performance Based Design is displacement based design process where structural analysis is used to calculate how the structural element will react to the service loads. This is very useful for certain aspects of design. A common example is the deflection of a bridge deck or floor. A strength approximation may be possible, but a displacement based design process checks for this issue in a much more intuitive way. However, there is no strength check executed as part of this design process. Along with that, there are some situations where the use of displacement based design is impractical.
Prescriptive Design is an empirical method based off of previous designs and calculations. This method relies on a good quality design guide and the design is calculated by using the necessary structure information to create an adequate design specification. This design method is very good for common structures and can be executed a lot faster than the previous methods mentioned. The drawback is that there may not be adequate information for strength, stress and displacement checks. Along with that a lot of research and testing in field engineering practice to make this method feasible; by this nature, this may exclude a lot of extreme events like earthquake or extreme weather.
My personal preference between the four methods is LRFD combined with deflection check for serviceability design. However, in the field of civil engineering field, soil analysis is unpredictable enough that we are still not able to use LRFD; so in that case ASD is still used. As for prescriptive design, I think it is great for finding a starting point. Then from there, a check for the critical issues can be performed on what should be a good quality estimate. What is your preferred method for design and why is that? If it differs, what do you usually use in your engineering practice and why do end up having to use a different method?