Cost Prediction of Aeration & Clarifier Tanks on Ground Level using Artificial Neural Networks

Design and cost estimation of high capacity Aeration & Clarifier reinforced concrete tanks are time consuming task, which requires a great deal of expertise. Many times it is required to know the cost of a tank of known capacity and geometry before its detailed design. Quick prediction of cost is a challenge before less experienced design engineers. In this study, artificial neural networks have been employed in cost prediction of rectangular, square and circular tanks. A back propagation neural network has been considered in the present solution of the analysis problem using A NEURO SOLUTION FOR EXCEL Toolbox. Authors have developed software D-CEGWT (Design and Cost Estimation Ground Water Tank) for design, estimation and costing of square and circular on ground water tank. The main benefit is using a neural network approach is that the network is built directly from experimental or theoretical data using the self-organizing capabilities of the neural network. The input parameters for the software are wall height, length, width, unit cost of labor and material for tank (rectangular, square) tank and wall height, diameter, unit cost of labor and material for circular tank, etc. Using the software, rectangular, square and circular tanks, of large size, have been designed and estimated. Using these input data, artificial neural networks have been trained. The trained networks have been further used to predict the costs of new tanks of different input variables. It is observed from the study that network predictions are quite close to the true ones. Such methodology will be very useful for quick prediction of the cost of the tanks without carrying out whole design process i.e. preliminary design stage. It has been observed that the result given by the network has good approximation when compared with the conventional solution.