Early Stage Cost Modeling of Drip Irrigation Systems

Allocation of budget to pressurized irrigation projects is supported by the government to manage, saving water resources and increase agricultural production and is considered as the ideal project of the country. Cost-effectiveness and awareness of the cost of an irrigation system is important and necessary for the government. Therefore, estimating the initial and final cost of the project, especially irrigation systems, is one of the project management tools that allow project managers to make more accurate decisions at different stages. Finding a model to identify the important factors affecting the final cost of an irrigation system, as well as formulating it for use throughout the country and regions with different characteristics, is what the present study seeks.

The aim of this study was to estimate the cost of drip irrigation projects in the early stages of design using genetic programming technique, using data from 100 drip irrigation projects, in four sections including; cost of pumping station and central control system (TCP), cost of on-farm equipment (TCF), cost of installation and performing on-farm and pumping station (TCI) and total cost (TCT). First, a database containing 39 important and influential variables on the cost of the mentioned sections was prepared and the prices of the projects were updated for the base year of 2019. In the next step, the most important features that had the highest impact on the costs were selected using Eureqa Formulize software and using genetic programming. In the last stage, different models were presented in each section to estimate the cost and the best model in each section was introduced based on the statistics of accuracy and complexity.

The results of correlation analysis between independent variables and dependent variable (cost of each section) show that in TCP section PP variable (pump power required), in TCF section L16mm variable (total lateral length), in TCI section SR variable (plant row spacing) and in the TCT section, the HP variable (pumping height) with the correlation coefficient (R) equal to 0.766, 0.638, 0.355 and 0.429, respectively, had the highest value among the other variables and all were significant at the level of one percent confidence. Also, the results of cost modeling for drip irrigation system showed that in the TCP section, a model with an evaluation criterion of R equal to 0.449, MAE (average absolute error value) equal to 27236333 Rials and complexity of 22 was the best model. These evaluation criteria for the best model in the TCF section were equal to 0.848, 21198257 Rials and 13, in the TCI section equal to 0.770, 45483996 Rials and 11, and in the TCT section these criteria were 0.743, 77220845 Rials and 15, respectively.

In this study, the cost estimation of different parts of the drip irrigation system was modeled using genetic programming algorithm, and the obtained results showed that the presented models had excellent accuracies in each part. The results of this study can be a very useful tool for researchers, managers, students, consultants, contractors and those who are concerned in the water industry. By conducting similar research, it is possible to make an economic estimate with a high accuracy before the implementation stage. AcknowledgementThis article is extracted from the Ph.D. dissertation, the first author of the article. For this purpose, the authors of the article would like to thank the Department of Irrigation and Reclamation Engineering, Faculty of Agriculture Engineering and Technology, College of Agriculture and Natural Resources, University of Tehran, Karaj, Iran for their excellent cooperation, providing the necessary facilities for this research and the preparation of relevant articles. Keywords: Pressurized Irrigation, Cost Estimation, Genetic Algorithm, Feature Selection, Eureqa Formulize.