The transportation of potable and wastewater through pipelines has become an essential aspect of modern society. Pipelines offer significant advantages such as ease of continuous operation, high efficiency, automation potential, and low energy requirements. However, the risk of pipeline bursts or leaks poses a substantial threat, leading to potential disasters. Consequently, the detection and localization of pipeline leaks have become a critical concern. This paper presents a methodology to detect and locate leaks within urban pipe networks. Towards this, investigations were carried out on a lab-scale water distribution network that provides a controlled environment for introducing leaks at various points within the pipe network and analyzing the effect of demand nodes on leaks by collecting real-time data on multiple flow parameters. It also has provision to study the effect of different pipe materials (ductile iron, galvanized iron, and polyvinyl chloride) with multiple leak points in 100 mm nominal pipe diameter, which better depicts reality in real-world urban pipe networks. The facility experimentally simulates leaks and collects real-time data on various fluid parameters. A digital twin was constructed using EPANET and WaterGEMS software to create a realistic simulation of the pipe network. The lab-scale water distribution network, in combination with the digital twin, offered a comprehensive and high-quality dataset, enabling the development of a robust leak detection methodology for urban pipe networks. By analyzing this dataset, an algorithm was trained to identify and localize leaks within the network under various flow rates (Reynolds number) to demonstrate the algorithm's effectiveness in leak detection and localization.