Datasets
Standard Dataset
UHPC Beams and U shaped girder
- Citation Author(s):
- Submitted by:
- Tianbo Peng
- Last updated:
- Sat, 12/21/2024 - 05:56
- DOI:
- 10.21227/60v8-tc92
- License:
- Categories:
- Keywords:
Abstract
Dataset 1 include 100 rebar-reinforced rectangular UHPC beams data. The selection criteria for these datasets is stringent, including: (1) a minimum cubic compressive strength of UHPC of 110 MPa (16 ksi); (2) the use of three-point or four-point bending tests for flexural analysis; (3) the use of hot-rolled ribbed rebars for steel reinforcement; (4) a minimum steel fiber volume fraction of 0.5%; (5) balanced-reinforced failure of UHPC beams, ensuring neither over-reinforcement nor under-reinforcement (Balanced-reinforced failure indicates that the UHPC beam, with moderate steel reinforcements, simultaneously reaches both tensile and compressive limit strains upon failure). Whereas for Dataset 2, A computer algorithm has been developed using modified box complex method (63, 64). The algorithm works on the principle of replacing the worst design point with a new point called a reflecting point, which is found by reflecting the worst design point having maximum objective function through the centroid of the leftover design points in the complex [65], [66], [67]. During this process all the feasible design points are stored. There are nine independent features included in the dataset, specifically related to the geometrical configuration of U-shaped girders. The ultimate moment capacity of U-shaped girders is the dependent variable that serves as the output.
The selection criteria for these dataset 1 is stringent, including: (1) a minimum cubic compressive strength of UHPC of 110 MPa (16 ksi); (2) the use of three-point or four-point bending tests for flexural analysis; (3) the use of hot-rolled ribbed rebars for steel reinforcement; (4) a minimum steel fiber volume fraction of 0.5%; (5) balanced-reinforced failure of UHPC beams, ensuring neither over-reinforcement nor under-reinforcement (Balanced-reinforced failure indicates that the UHPC beam, with moderate steel reinforcements, simultaneously reaches both tensile and compressive limit strains upon failure).
Due to the strict limits used in generating dataset 2, the data is highly scattered and contains outliers. To address this, we employed the Interquartile Range (IQR) method to identify and remove outliers. The IQR is calculated by finding the difference between the first quartile (Q1) and the third quartile (Q3) of the data. Outliers are defined as data points that fall below Q1 minus 1.5 times the IQR or above Q3 plus 1.5 times the IQR. After filtering out these outliers, we retained 700 datasets that fall within the acceptable range.