Prediction of Compressive Strength of Concrete using Artificial Neural Network, Fuzzy System Model and Thermodynamic Methods (A Comparative Study)
| dc.contributor.advisor | Gebreyouhannes, Esayas (PhD) | |
| dc.contributor.author | Siraj, Nebiyu | |
| dc.date.accessioned | 2018-07-26T11:07:27Z | |
| dc.date.accessioned | 2023-11-11T12:56:19Z | |
| dc.date.available | 2018-07-26T11:07:27Z | |
| dc.date.available | 2023-11-11T12:56:19Z | |
| dc.date.issued | 2015-10 | |
| dc.description.abstract | Utilization of concrete, as a material for building structures has been one of the major triumphs of mankind. Driven by increased standard of living and ever-rising population, the need for ‘better’ concrete is now the current trend. This in turn necessitates a representative measure of concrete quality that can be used as a benchmark. Amongst the possible alternatives, compressive strength of concrete is the main parameter used in design and assessment of quality in massive structures built with the likes of reinforced concrete, composite materials etc. Such endeavor is the challenge of the structural engineer while coming up with appropriate material and structural designs and controlling the quality of the concrete throughout the construction stage. This thesis aims at adressing these challenges. The need to bypass the 28-day or accelerated compressive strength testing at the laboratory, especially in speedy constructions is duly recognized. Hence, compressive strength predicting models using the principles of artificial intelligence are proposed. These models apply fuzzy logic and artificial neural networks as a tool to predict the compressive strength of concrete at a given day. The Matlab software was also used to create a program that employs a user-friendly Graphical User Interface tool. To help in the comparative study, regression analysis was done to represent a statistical approach. There have been many reported incidents where a structure encounters material failure before it reaches its capacity. Knowing the behavior of the concrete is therefore very important. Hence, one needs to strike the balance between structural and material design. This is apparent in massive structures like dams or concrete structures built in extreme weather conditions. In an attempt to study of the behavior of the concrete, thermodynamic method is adopted. This was done with the help of a software called DuCOM-COM 3D. Here, hydration of the concrete is closely studied and a compressive strength predicting model has been proposed. The models were found to give good predictions of compressive strength. Sugeno type fuzzy inference system performed better under the fuzzy logic approach and the artificial neural network approach mapped the compressive strength outputs with a much better accuracy. The thermodynamic approach also gave very good results for specific cement types and provided a much elaborate analysis for a mechanistic study of concrete properties such as degree of hydration, temperature rise, free water content, porosity and most importantly, stepwise development of compressive strength. Keywords: compressive strength, fuzzy logic, artificial neural network, thermodynamics | en_US |
| dc.identifier.uri | http://etd.aau.edu.et/handle/12345678/10082 | |
| dc.language.iso | en | en_US |
| dc.publisher | Addis Ababa University | en_US |
| dc.subject | compressive strength; fuzzy logic; artificial neural network; thermodynamics | en_US |
| dc.title | Prediction of Compressive Strength of Concrete using Artificial Neural Network, Fuzzy System Model and Thermodynamic Methods (A Comparative Study) | en_US |
| dc.type | Thesis | en_US |