آب و فاضلاب (Feb 2021)

Risk Assessment of Using Urban Treated Wastewater for Agricultural and Industrial Application with Bayesian Networks Model

  • Javad Shafiee Neyestanak,
  • Abbas Roozbahani,
  • Seyyed Ebrahim Hashemi Garmdarreh

DOI
https://doi.org/10.22093/wwj.2020.237516.3044
Journal volume & issue
Vol. 31, no. 7
pp. 162 – 179

Abstract

Read online

Urban treated wastewater is a useful resource for allocation to various non-potable reuses because its quantity and quality can be controlled and managed by humans, but it has its limitations. Improper use of these resources without treatment can lead to environmental and health risks and, as a result, can cause social dissatisfaction and sometimes economic problems. This paper examines the risks of limitations associated with the use of treated wastewater. Qualitative information of 26 wastewater treatment plants in the country was collected to create a proper database for risk assessment of using treated wastewater for agricultural and industrial application by taking into account the major factors such as environmental, socio-cultural, economic, and technical. For risks calculation, multiple risks associated with the use of treated wastewater were identified by reviewing various available publications and categorizing them according to the above four risk factors. Hierarchical networks were prepared to calculate the main and the overall risk. Bayesian networks were used to model risk-based structures. Bayesian relationships are causal and helpful in expressing the links of the nodes in a probabilistic way. Research results show that the risk of using treated urban wastewater in agricultural sector is generally higher than its application in industrial use, and hence it will make it more suitable and more acceptable for industrial use under current treatment. Bayesian risk-based agricultural and industrial structures were suitably modeled according to R2, RMSE, and MAPE indicators. The average values of the above indicators for the calibration of Bayesian model using treated wastewater in agricultural and industrial risked-based structure were 0.993, 0.202 and 0.637; 0.988, 0.980 and 2.731, respectively. From this study it can be concluded that Bayesian networks have a modeling capability in assessing the risk associated with the use of treated wastewater in agricultural and industrial sectors. The method presented in this paper can be used by the wastewater treatment plant managers and end users when assessing the potential risk of using treated effluent for agricultural and industrial applications.

Keywords