New Methods in Predicting Concrete Chloride Resistance Tested in Recent Study
The utilisation of machine learning techniques to effectively predict the chloride migration coefficient of concrete has shown potential in a recent study undertaken by the Babol Noshirvani University of Technology.
Machine learning involves artificial intelligence and technology to learn from previous experiences.
This way of testing is financially viable and is a less complex substitute for labour-intensive experimental evaluations.
Chloride migration is described as the movement of ions caused by an external electrical field.
Analysing this allows researchers to understand and enhance the durability of concrete structures and reduce the risk of corrosion.
Existing models have two definitive issues, which involve constraints imposed by limited data, and the absence of certain input variables.
These problems contribute to a decrease in effectiveness of the current models.
The proposed machine learning model incorporates various input variables, such as water content, cement content, slag content, fly ash content, silica fume content, fine aggregate content, coarse aggregate content, superplasticizer content, fresh density, compressive strength, age of compressive strength test, and age of migration test.
This new model was effectively validated, with findings indicating superior performance compared to other models.
Algorithms ‘XGBoost’ and ‘SVM’ had high R2 scores of 0.94 and 0.91.
Models ‘Random Forest’, ‘LightGBM’ and ‘XGBoost’ displayed the highest levels of accuracy, with 0.93, 0.96, and 0.97, respectively.
With these findings, improvements in concrete chloride resistance testing can be fast-tracked in order to attain more reliable and economically practical calculations.
To read the full study, visit the link below.