The ongoing advancements in water distribution systems are characterized by a shift towards data-driven methodologies, particularly in the realms of leakage detection, valve operations, zoning, and sensor placement strategies. This study was undertaken to develop a hydraulic model within EPANET, with a focus on testing techniques that are dependent on both pressure and demand. The model is based on an existing water distribution network in Vadodara city’s town planning scheme, providing a realistic portrayal of intermittent water supply scenarios. It incorporates observations of both partial and full demand achievements, as well as pressure allocations during supply and non-supply hours. EPANET is configured with controls for two test scenarios: the Pressure Driven Approach (PDA) and the Demand Driven Approach (DDA). These scenarios are designed to ensure the delivery of demand at nodes and evaluate the system's capacity to simulate water distribution. The study underscores the significance of PDA over DDA, with the comparative outcomes serving as a foundational basis for decision support systems, clustering techniques, leakage detection capabilities, and pattern-recognizing strategies. These aspects align with the real-time dynamics of intermittent water supply systems. Simulation results are meticulously analyzed, employing the clustering technique of Machine Learning (ML), to classify performance indicators. This analysis offers valuable insights into demand distribution and pressure allocation within the system. Furthermore, the study provides decision-making notes for both approaches, particularly in the context of leak detection capabilities. It also identifies pressure-sensitive locations, contributing to a comprehensive understanding of system behavior and vulnerabilities.