not only bridges

SAP2000 vs StaadPro

StaadPro is the most widely used bread-and-butter structural software in Asia and the UK while SAP2000 seems to be the most widely used one in Latin America, Portugal, Italy and Spain.

The grids provided in SAP makes it easier to create the geometric input than with the StaadPro. Dynamic analysis is stronger in SAP2000 (earthquake force applied in any direction, automatic lumping of masses for earthquake, live load reduction, bridges transient loads, eigenmodes and ritz modes, etc...). Sap2000 has facilities for creep and shrinkage of concrete. SAP2000 program has the ability to solve heterogeneous soil-structure interaction which is not possible in StaadPro. Also, SAP2000 can import the geometry and the material properties of StaadPro models, I am afraid that the opposite is not possible.

However, StaadPro has some great advantages. First and foremost, the license is cheaper. The connection with visualization and drawing software is much easier, not in vane StaadPro is a product of Bentley (Microstation, PDS, etc...). StaadPro provides additional modules for designing steel joints and foundations. The text editor facility for the model input file makes StaadPro is more versatile for model modification. It is possible to work with the ASCII file of SAP but it is cumbersome and it is necessary to leave the model and reload.

Now, the choice is yours.

Data intensive engineering

Data Intensive Engineering at Slideshare

Dam-break, numerical modelling, examples and consequences

Dam-break link on Scribd

Fast square roots

I have always been puzzled about the way square roots calculations are taught in the elementary school. I remember a very cumbersome method in which each decimal was extracted with hard work.

In college, especially in my studies in computer science, I have not always been permitted to use a calculator during the examination time, so I learnt to calculate square roots with pencil and paper in a sufficiently approximate but fast way. In my case, I use to apply the Newton-Raphson algorithm to the equation f(x) = x-root(K) = 0

Suppose we need to calculate the root of a real number K. To calculate the root of K, first assume an educated guess x[0].

For small numbers my usual estimate is x[0] = K/2
Up to 1000, to start with x[0] = 8+(K/40) usually works well. In other cases, bearing in mind some perfect squares helps.

Once the first estimate is made, it is refined with the formula
x[i+1] = (1/2) * (x[i] + (K / x[i]))

For example, if you are on-site and you do not have a calculator near you and a worker asks you about the side of a slab of 500 sq meters you know that the side L = root(500) = 10*root(5) = 10*x
then
x[0] = 5/2 = 2.5 (error 12%)
x[1] = 9/4 = 2.25 (error 1%)
x[2] = 161/72 = 2.2361 ... (error 0.002%)

therefore the side is L = 22,361 m

The difference between the classic algorithm and the Newton-Raphson one is clear in the appended images below.