An Early Cost Estimation Model for Hydroelectric Power Plant Projects Using Neural Networks and Multiple Regression Analysis

Abstract

Energy is increasingly becoming more important in today's world, whereas energy sources are drastically decreasing. One of the most valuable energy sources is hydro energy. Because of limited energy sources and excessive energy usage, cost of energy is rising. Among the electricity generation units, hydroelectric power plants are very important, since they are renewable energy sources and they have no fuel cost. To decide whether a hydroelectric power plant investment is feasible or not, project cost and amount of electricity generation of the investment should be precisely estimated. In this paper, fifty four hydroelectric power plant projects are analysed by using multiple regression and artificial neural network tools. As a result, two cost estimation models have been developed to estimate the hydroelectric power plant project cost in early stages of the project.

Early Bill-of-Quantities Estimation of Concrete Road Bridges: An Artificial Intelligence-Based Application

“…Similarly, Shehab et al (2010) concluded that the ANN developed for water projects cost prediction produced much more accurate results compared to the regression one and thus demonstrated superior capabilities in mapping relationships between inputs and outputs in limited data environments. Gunduz and Bayram Sahin (2015) also confirmed in their research that the ANN that was trained on cost data from 41 hydroelectric power plant projects had substantially higher prediction accuracy than the respective regression model developed. Similar performance superiority of the ANNs over linear regression has also been reported by Creese and Li (1995) for their timber bridge cost prediction model.…”

“…However, the validation error is not a good estimate of the generalization error, and thus, a third set of data, not used during the (Gunduz & Sahin, 2015) has also been highlighted. Other applications of ANNs include financial tools, electric load consumption prediction and intelligent manufacturing systems (Huang & Zhang 1994, Zhang, Patuwo & Hu (1998.…”

“…ANNs’ powerful capabilities for capturing and modeling complex interrelations in real-world data sets have been widely identified in the literature (e.g., Dimitriou & Hassan, 2013; Karlaftis & Vlahogianni, 2011; Marinelli, Lambropoulos, & Petroutsatou, 2014; Moselhi, Hegazy, & Fazio, 1992) and their suitability for cost estimation of different kinds of construction projects such as buildings (Gunaydin & Dogan, 2004; Kim, An, & Kang, 2004), tunnels (Petroutsatou, Georgopoulos, Lambropoulos, & Pantouvakis, 2012), highway projects (Al-Tabtabai, Alex, & Tantash, 1999; Hegazy & Ayed, 1998; Xin-Zheng, Xiao-Chen, & Jing-Yan, 2010), drainage projects (Alex, Al Hussein, Bouferguene, & Fernando, 2010), water projects (Taormina, Chau & Sethi, 2012; Shehab, Farooq, Sandhu, Nguyen, & Nasr, 2010),), and hydroelectric power plant projects (Gunduz & Bayram Sahin, 2015) has also been highlighted. Other applications of ANNs include financial tools, electric load consumption prediction, and intelligent manufacturing systems (Huang & Zhang, 1994; Zhang et al, 1998).…”

Cost estimation and prediction in construction projects: a systematic review on machine learning techniques

“…Recently, [15] have investigated forecasting hydroelectric power plant project's cost via ANN through which, three different architectures have been generated and examined, while seeking the best performance. The results have been compared with those of RA and concluded that the ANN shows better promising results.…”

“…On the other hand, RA can be classified as a data oriented technique that deals with just the data in hand and not the characteristics behind them and is divided to two linear and nonlinear models [15]. In addition, decision trees are widely used for solving classification problems.…”