Using Artificial Neural Network to Predict the Restraint in Concrete Culvert at Early Age

• Estimation of restraint is very important for accurate prediction of the risk of concrete cracking at early age. The present study predicts the restraint in 324 walls and 972 roofs for a concrete culvert. A parametric study included the thickness and width of the roofs, thickness and height of the walls, thickness and width of the slab, and length of the structures. Each parameter increased or decreased the restraint in the walls and the roofs. The calculation of the restraint was done elastically by the finite-element method (FE). The results were used by an artificial neural network (ANN) tool, where firstly an influential percentage was investigated as input parameters on the restraint prediction. Equations have been derived by the ANN model to calculate the restraint in the walls and the roofs. It was then used in an Excel sheet to calculate the restraint and compare the result with the result from the finite-element calculations giving high accuracy between the ANN model and the FE calculations

Subject headings

TEKNIK OCH TEKNOLOGIER  -- Samhällsbyggnadsteknik -- Infrastrukturteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Civil Engineering -- Infrastructure Engineering (hsv//eng)

Keyword

Structural Engineering

Konstruktionsteknik

Publication and Content Type

ref (subject category)

art (subject category)