Today energy recovery using Micro-Hydropowers (MHPs) in Water Distribution Networks (WDN) is a well-known approach for recycling the wasted energy in infrastructures as a sample of circular economy. Likewise, in this study for the first time a framework for evaluation of WDN for energy harvesting have been designed with the application of statistical optimization, simulation, and artificial intelligence concepts. In this study, after modelling a WDN in Mashhad, Iran, with Environmental Protection Agency Network Evaluation Tool (EPANET) software, the potential of energy recovery using MHP technology was optimized with the application of Design of Experiment (DOE) methods, including Taguchi and Response Surface Methodology (RSM) and then the model prediction ability was improved by Artificial Neural Network (ANN) technique. Results of this investigation revealed that the combination of Taguchi and RSM methods could successfully optimize the energy recovery potential with consideration of improving the hydraulic parameters of WDN. With the application of RSM and Taguchi, high potential positions for MHP placement are detected and analyzed based on a high-performance operational decision-making methodology. According to Artificial Intelligence (AI) computations, energy harvesting and hydraulic responses can be estimated with more than a 99 % correlation coefficient. Also, it shows that the soft-operator can be executed to control the features of MHPs in WDNs. The outputs of this research demonstrated that MHP harvested energy is more than 400KW for the run time of this study with consideration of hydraulic parameters.

如今，在配水网络 (WDN) 中使用微型水电 (MHP) 进行能量回收是一种众所周知的方法，用于回收基础设施中的浪费能源，作为循环经济的一个样本。同样，在这项研究中，首次设计了用于评估 WDN 能量收集的框架，并应用了统计优化、模拟和人工智能概念。在这项研究中，在使用环境保护署网络评估工具 (EPANET) 软件对伊朗马什哈德的 WDN 建模后，通过应用实验设计 (DOE) 方法优化了使用 MHP 技术的能量回收潜力，包括田口和响应面方法（RSM），然后通过人工神经网络（ANN）技术提高模型预测能力。这项调查的结果表明，田口和 RSM 方法的结合可以成功地优化能量回收潜力，同时考虑改善 WDN 的水力参数。随着 RSM 和田口的应用，基于高性能运营决策方法的 MHP 安置的高潜力位置被检测和分析。根据人工智能 (AI) 计算，可以估计能量收集和水力响应，相关系数超过 99%。此外，它表明可以执行软算子来控制 WDN 中 MHP 的特性。本研究的结果表明，考虑到水力参数，在本研究的运行时间内，MHP 收集的能量超过 400KW。