@article{Kifli_Absa_Musyafa_2017, title={Prediction of Mechanical Properties of Light Weight Brick Composition Using Artificial Neural Network on Autoclaved Aerated Concrete}, volume={5}, url={https://www.ajouronline.com/index.php/AJAS/article/view/4754}, abstractNote={Indonesia with raw material limestone abundant, light brick of Autoclaved Aerated Concrete (AAC) is the most important component in the construction of buildings, so that needs light brick AAC qualified in mechanical and thermal properties of acoustic. In the research domain lightweight brick lifting who qualified mechanical properties. Advantages of light brick AAC low density of about 500 to 650 kg/m3, more economical, suitable for multi-storey buildings can reduce the weight of 30 to 40 % compared with conventional brick (clay brick). One of the problems found in the fabrication of lightweight brick is how to determine the composition of raw materials used. The composition of materials in lightweight brick can affect its mechanical properties which are an important parameter for building materials. In this research the prediction of the effect of elements composition and density on compressive strength (AAC) using neural network has been done. Furthermore, the simulation on the effect of each element composition and the density on compressive strength also have been done. The best network developed in this research using feed forward back propagation architecture and Levenberg-Marquardt algorithm is to use 8 hidden nodes, with MSE (mean square error) training of 0.001605667 and MSE validation of 0.01455. Simulation results show that composition of Ca, Si, O, and density are all directly proportional to compresssive strentgh, while composition of Al is inversely proportional. The compressive strength prediction results obtained for 4 AAC, and for sample AAC-a = are 4.80 MPa, AAC-b= samples 5.24 MPa, AAC-c=3.23 MPa, and AAC-d= 3.67 MPa. The result of prediction shows that the neural network developed can predict the effect of composition and density on compressive strength of AAC.}, number={3}, journal={Asian Journal of Applied Sciences}, author={Kifli, Zul and Absa, Munzir and Musyafa, Ali}, year={2017}, month={Jun.} }