Objectives
1-To study various AI Models available in literature for prediction of concrete behavior.
2-To Collect dataset from literature.
3-To apply suitable AI techniques for prediction of compressive strength of recycled aggregate concrete.
Socio-Economic Benefit
1-Environmental Sustainability
2-Waste Reduction and Recycling
3-Cost Savings
4-Time Saving
Methodologies
1-Data Collection
We begin with the crucial step of data collection, where we meticulously gather a dataset that includes a 11 independent variables cement ,water, water to cement ratio, sand , natural coarse aggregate, recycled coarse aggregate , density of natural coarse aggregate, density of recycled aggregate , water absorption of natural coarse aggregate, water absorption of recycled coarse aggregate, and age and it contain dependent variable compressive strength for the collection of data set we have reviewed more than 100 research papers. Our model is designed to predict outcomes based on this rich dataset.
2-Architecture Selection
Next, we delve into the architecture of our neural network. We explore two architectures: the feedforward and the cascadeforward. Each has its unique strengths,
3-Training the Model
With our architecture in place, we proceed to train the model. For training of our model we have firstly applied grid search on feedforward architecture contains 5 different training function trainbr, trainlm, traincgp, traincgb, traincgf, and we have applied different no of neurons and then we repeat all of this procedure with cascade forward net After grid search we reached to best no of neurons which is [30 10] and best architecture which is feedforward net and best training function is trancgp.
4-Evaluation and Iteration
For EVALUATUION we used RMSE , MSE, R and R2 but not all first attempts are perfect. When our initial training doesn't meet our standards, we employ techniques like early stopping and data filtering to improve accuracy.
5-Retraining
We retrain our model. This time, it exceeds our expectations, achieving the level of accuracy we strive for.
6-Saving and Loading the Model
Our successful model is then saved for future use. We ensure that it can be easily loaded and utilized whenever needed, making it a reliable tool for ongoing analysis.
7-User Interface Development
Finally, we create a user-friendly graphical interface. This allows users to interact with our model effortlessly, making complex predictions accessible to everyone.
Outcome
1-The development of accurate predictive models for the behavior of recycled aggregate concrete using artificial intelligence techniques. The model will predict the 2-compressive strength of Recycled Aggregate Concrete.
3-Reduction of environmental impact through decreased demand for natural aggregates and diversion of construction and demolition waste from landfills.
4-Realization of cost savings for construction projects by utilizing recycled aggregate concrete. With optimized mix designs and reduced material costs, contractors and developers can achieve economic benefits while meeting sustainability goals.
The outcomes of our project have the potential to drive positive change in the construction industry, promote environmental stewardship, and contribute to the achievement of sustainable development objectives.
