Scribd
Upload
51 views1 page

Artificial Intelligence Techniques To Predict The Compressive Strength of Sustainable Eco-Friendly Strain-Hardening Concrete Composites

Prediction - ANN

Uploaded by

Reva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
51 views1 page

Artificial Intelligence Techniques To Predict The Compressive Strength of Sustainable Eco-Friendly Strain-Hardening Concrete Composites

Prediction - ANN

Uploaded by

Reva
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

Engineering for Self Reliance and Sustainable Goals

Artificial Intelligence Techniques to Predict the


Compressive Strength of Sustainable Eco-friendly
Strain-hardening Concrete Composites
K K Yaswanth, J Revathy, P Gajalakshmi
Department of Civil Engineering, B S Abdur Rahman Crescent Institute of Science & Technology, Chennai,Tamil Nadu
[email protected]

ABSTRACT
A novel building material that breaks the frailty criteria of the brittle nature of concrete, comprises of eco-friendly sustainable
materials as ingredients is called as “Engineered Geopolymer Composites (EGC).” In the scenario of the material development
stage of EGC, researchers need to focus on the application of various sustainable materials that are locally available, in
the production of EGC, without compromising the mechanical performance. A prediction tool, that helps to forecast the
performance of the designed material with the mix-design criteria, would help the researchers for more advanced research;
and practicing engineers to decide upon structural applications in the field. ANN and ANFIS have been used in this study
to develop the predictive models to estimate the compressive strength performance of EGC. Totally 110 data sets have been
prepared from EGC material researches with 6 mix-design influences as inputs, to facilitate the training and validation process
of the predictive models. The validated results are analyzed. The research study reveals that the ANN model has predicted the
compressive strength of EGC with the accuracy of 78% with RMSE 3.3 whereas ANFIS showed a prediction of about 86%
accuracy with RMSE 3.03.
Keywords : ANFIS; ANN; Engineered geopolymer Composites; Flexible concrete; Prediction; Strain hardening composites.
INTRODUCTION
Interdisciplinary research has become popular in almost all kinds of research fields. Especially the usage of electronics and
Information technology in accounting, material design, machine designs, labour managements, navigation etc., The most
notable thing in the field of the construction industry is the application of artificial intelligence in concrete material design.
The development of predictive models with various soft-computing techniques (SCT) or machine learning models have
helped various researchers and practicing engineers to prejudge the outcome of design or research progress without actually
experimenting it. Significant SCTs include Neural Networks, Fuzzy logic, Genetic algorithm etc., that facilitated to predict the
mix-design, strength and workability parameters in various kinds of special concrete composites.
Artificial Neural Networks (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) are the most popular tools among
AI, particularly in the application of concrete technology from the early 2000s. These are based on real-life concepts, generated
artificially, to solve problems on optimization, prediction, forecast etc. ANN works on the principle of interconnected neurons
in the human brain. ANN network consists of an input layer, hidden layer and an output layer, interconnected with neurons.
Speaking mathematically, starting with an input variable across the hidden layer bridged with ‘weights’ towards output variable
forms an empirical equation. During the training process, the weights upon the variables keep on changing continuously until
a suitable equation with maximum accuracy will be obtained. With that established equation, prediction can be done. ANFIS
is a combination of ANN and Fuzzy logic (FL). FL works on the principle of ‘IF-THEN’ rule sets. FL has various types and
numbers of membership functions (MF), based on which it forms different ‘rule sets’, e.g., “IF input-1 is very low, input-2 is
medium and input-3 is high; THEN the output is very low”
Thus, based on the number of inputs; the number of MF and the type of MF, the ANFIS system generates large ‘rule sets’ from
which the output values will be predicted.
Researchers focus on two major drawbacks in cement concrete in terms of environmental and material strength aspects. Usage
of cement binder in concrete affects the environmental conditions near the manufacturing unit of cement, as it releases huge
greenhouse gases like carbon-di-oxide into the atmosphere, when raw materials are burnt at very high temperature in the
clinker. In ‘strength of materials’ point of view, concrete is weak upon tensile force due to its brittle nature, because of which
the deformation in concrete occurs instantly thereby giving least warning to the occupancies.

Technical Volume of 35th Indian Engineering Congress, December 18-20, 2020 371

You might also like