Introduction:

Water distribution networks are vital urban infrastructure, yet a substantial share of the water they carry is lost before reaching consumers as non-revenue water, largely through real leakage from mains, distribution pipes and service connections. Leakage wastes scarce water, raises production and pumping costs, reduces reliability and shortens asset life. Hydraulic pressure is the dominant driver of leakage and pipe bursts, so pressure management is widely regarded as one of the most effective and economical ways to reduce real losses. Calibrated hydraulic models such as Water GEMS allow pressure-control strategies to be tested before field implementation. However, most published studies address large networks or specific hydraulic conditions, and small networks with pronounced elevation differences, particularly in Chaharmahal and Bakhtiari Province, remain little studied. In such networks a single fixed pressure setting can cause excess pressure in low-lying zones and deficits at high points. This study therefore develops and validates a Water GEMS model of the Pordanjan distribution network and compares three pressure-reducing-valve (PRV) strategies, namely fixed-outlet, time-modulated and flow-modulated control, in terms of pressure, leakage reduction and practical feasibility.

Materials and Methods:

Pordanjan, a town in Farsan County, Chaharmahal and Bakhtiari Province, is supplied by gravity from a single concrete ground reservoir at the highest point of the town. The network comprises about 37 km of pipes (25–160 mm; 78% polyethylene, 16% steel, 6% asbestos cement), 207 nodes, roughly 3,000 service connections and six demand zones, serving about 9,400 people (projected 13,700 by 2041). Digital maps, as-built drawings, elevation data, customer locations, demand patterns and field pressure records were obtained from the provincial water and wastewater company, drawn in AutoCAD, checked against GIS data and site surveys, and imported into Water GEMS, where geometric errors were corrected. Hydraulic analysis relied on mass and energy conservation, with the Hazen–Williams equation for head loss.

Real losses were estimated by minimum night flow (MNF) analysis. Night leakage was extended to 24 hours through the pressure–leakage relationship Q = C × P^N1, with N1 = 1.15 and C calibrated from hourly model pressures. Network performance was also expressed by the Infrastructure Leakage Index (ILI).

The model was calibrated by adjusting pipe roughness and demand patterns against pressures recorded every 15 minutes for 48 hours at eight points in zones 1,3, 4 and 6, and was then validated with independent pressure and flow data. Three scenarios were simulated over an extended 24-hour period: (1) constant PRV outlet pressure, selected by a sensitivity analysis of 30–55 m; (2) time-modulated outlet pressure following the daily demand pattern; and (3) flow-modulated outlet pressure adjusted to instantaneous flow and controlled through the critical node in zone 6. All scenarios were constrained by a minimum service pressure of 20 m. Indicators were mean, maximum, minimum and standard deviation of nodal pressure, minimum night flow, daily leakage, annual water saving, and the number of nodes above 50 m or below 20 m. The annual economic value of the saved water was examined by sensitivity analysis at 10, 20 and 30 thousand tomans per cubic meter.

Results and Discussions:

The calibrated model reproduced field behavior well: pressure RMSE was below 5%, and differences between simulated and measured pressure and flow were under 10%. The baseline simulation showed strongly non-uniform pressure, from 15.2 to 67.3 m (mean 41.2 m, standard deviation 11.9 m), with 116 nodes above 50 m and 38 below 20 m. Real losses were about 850 m³/day, roughly 22% of inflow and above the national average of 20–25%. MNF was 10.44 L/s, of which 66.5% (6.94 L/s) was real leakage, and the ILI was 3.8. Losses stem mainly from ageing pipes, corrosion-related cracks and weak joints aggravated by excess pressure, and daily pressure fluctuations impose continual mechanical fatigue on pipes and fittings.

All three scenarios improved on the baseline. A constant outlet pressure of 45 m proved optimal: lower values starved the upper zones, higher values limited leakage reduction. It cut mean pressure from 41.2 to 35.8 m, nodes above 50 m from 116 to 32, and daily leakage from 850 to 715 m³ (15.9%), saving 49,275 m³/year. Its weakness is that pressure stays at 45 m through low-demand night hours, when 30–35 m would suffice, and the setting must be revised as the population grows.

Time-modulated control lowered mean pressure to 33.6 m, night flow to 7.98 L/s and daily leakage to 655 m³ (22.9%), saving 71175 m³/year, while holding about 22 m in zone 6 at peak demand. Its limitations are the need for a programmable valve (20–50% costlier than a simple PRV), periodic re-tuning of the time pattern, and no response to sudden demand changes such as fires or pipe bursts. Flow-modulated control, with set points of about 35–38 m at low flow and 48–50 m at peak flow, gave the best technical performance: mean pressure of 32.1 m, pressure standard deviation of 7.8 m, only 8 nodes above 50 m, and daily leakage of 605 m³ (28.8%), saving 89425 m³/year. However, it requires intelligent controllers, flow sensors and the highest investment.

Pressure reduction did not compromise service, as no node fell below 20 m in any scenario. The sensitivity analysis gave annual benefits of about 0.49–1.48, 0.71–2.14 and 0.89–2.68 billion tomans for the three scenarios, and the ranking did not change with water value. Because equipment costs were unavailable, no payback period was calculated. Considering technical performance, cost and implementation constraints, the time-modulated scenario was judged the most suitable option under present conditions. The findings agree with earlier studies showing that pressure management can cut leakage substantially without extensive pipe replacement.

Conclusion:

A calibrated Water GEMS model proved an effective tool for diagnosing pressure problems and comparing control strategies in the Pordanjan network, where topography causes excess pressure in low zones and deficits at high points. All three PRV strategies reduced excess pressure and leakage (by 15.9%, 22.9% and 28.8%) while keeping service pressure above 20 m. Time-modulated control is recommended in the short term for its balance of leakage reduction and cost, with flow-modulated control as a longer-term upgrade. Continuous monitoring of pressure and minimum night flow is needed to sustain the gains. Future work should use long-term seasonal data, SCADA-based real-time control and life-cycle cost analysis.

Keywords: Leakage, Pressure Management, Water Distribution Network, Water Loss, Water GEMS

Author Contributions:

First author: Data collection and preliminary analysis, and manuscript writing.

Second author: Data analysis, guidance, editing and revision of the manuscript, and verification of results.

Third author: Data analysis, guidance, editing and revision of the manuscript, and verification of results.

Data Availability Statement:

The data of this study are part of the M.Sc. thesis of Seyyed Mohammad Kazem Banihashemi Emam Gheysi (Civil–Environmental Engineering, Faculty of Engineering, Qom University of Technology). Part of the data was obtained from the Chaharmahal and Bakhtiari Provincial Water and Wastewater Company.

Funding:

The authors thank the Chaharmahal and Bakhtiari Provincial Water and Wastewater Company for providing data and information and for financial support of this research.

Ethical Considerations:

The authors observed ethical principles in conducting and publishing this work, including avoidance of data fabrication, falsification, plagiarism and misconduct, and all authors confirm this