Rıza Polat

Effect of heat treatment temperature on ground pumice activation in geopolymer composites

Abstract In emerging countries, the driving elements for sustainable development are greenhouse and global warming concerns and the need for the development of low-CO2 cements as replacement for Portland cement. Pumice is an aluminosilicate-type material that can be condensed with NaOH and Na2SiO3 solution and can be used for green building with reduction in CO2 footprint. The present paper highlights the effect of curing temperature on Hasankale pumice activation. Four curing temperatures have been investigated in this paper, 25°C, 45°C, 65°C, and 85°C, and 65°C has been confirmed as the best temperature for ground pumice activation. Furthermore, the aging effect has been studied at different curing temperatures. The aging of the samples before 28 days has a remarkable effect on compressive strength gain, but after 28 days this effect is inconsiderable for all heat treatment temperatures.

Safety factor determining for space trusses by non-linear analysis and artificial neural network method

Abstract Determining a feasible safety factor for space trusses is an important phase in structural analysis that could have economic benefits. We know there are many kinds of imperfections in structural elements, which include both material and geometric flaws. Predicting factual behavior of structures is very difficult and occasionally impossible. Elements with initial geometric imperfections in space trusses are a common phenomenon, in addition, equivalent initial geometric imperfections can be applied for modeling of residual stresses or eccentric loading effect. The number of members in the space structures is usually high as is the diversity in the kind of initial imperfection. Therefore, there is a high likelihood that models must be analyzed. The structure must be analyzed with non-linear methods, making these approaches time consuming, and potentially uneconomical. In this study, we selected 30 cases for random analysis based on Monte Carlo methods to find the bearing capacity of the space truss. We attained results from the LUSAS program LUSAS Modeller, Version 13, UK program and these were then exported as input data to the Artificial Neural Network (ANN) program. A reasonable neural network has been found of predicting another 30 cases for load bearing capacity without any analysis and only based on the neural network program. Finally, a new approach for determining the load capacity of the space trusses was extracted and we predicted the occurrence possibility of the convenience load bearing capacity in 60 cases.