Pipe Network Solver

Hardy-CrossEPANET
Network Configuration
Build your pipe network or select a preset configuration

Start with a predefined network configuration

Choose friction head loss equation

C

Water temperature for viscosity calculation

Maximum Hardy-Cross iterations

L/s

Flow correction threshold for convergence

Frequently Asked Questions

Common questions about this calculator

Pipe network analysis determines flow rates and pressure distribution in interconnected piping systems. Unlike simple series/parallel calculations, networks with loops require iterative methods like Hardy-Cross to find equilibrium where flow continuity (into each node = out) and energy balance (pressure drops around loops = 0) are satisfied.

Hardy-Cross is an iterative technique for analyzing pipe networks. For each loop, calculate net head loss, then apply flow corrections until head losses balance. Formula: ΔQ = -Σ(hf)/Σ(2hf/Q) for each loop. Continue iterations until corrections are negligible (<0.1%). Converges in 3-6 iterations typically.

Define: pipe diameters, lengths, roughness coefficients; node elevations and demands; reservoir/tank levels. Assign initial flow distribution satisfying continuity. Apply Hardy-Cross or gradient method to find balanced flows. Verify velocities (0.5-3 m/s) and pressures (20-60 mWC at outlets) meet standards.

Pressure imbalance results from: undersized pipes creating excessive friction loss, unbalanced demands between zones, elevation differences without pressure regulation, closed or partially closed valves, and pump head mismatches. Use pressure reducing valves (PRVs) to equalize zones and prevent excessive pressure at low elevations.

Common tools include: EPANET (free, EPA-developed), WaterGEMS/WaterCAD (Bentley), InfoWorks WS (Autodesk), and KYPIPE. These handle complex networks with pumps, tanks, valves, and time-varying demands. For simple networks, spreadsheet implementations of Hardy-Cross work well.

Initial sizing: use velocity method (2-3 m/s for mains, 1-2 m/s for branches). Then analyze network for pressure distribution. Resize if velocities exceed 3 m/s (noise, erosion) or pressures fall below minimum (typically 15-20 mWC at fixtures). Balance pipe costs vs pumping costs over system lifetime.

Learn More

The Hardy-Cross method solves looped pipe networks where water can flow through multiple paths simultaneously. Unlike branched systems with unique solutions, looped networks require iterative balancing to satisfy two fundamental hydraulic laws: continuity (mass conservation at nodes) and energy conservation (head loss balance around loops). This algorithm, developed by Hardy Cross in 1936, remains the foundation for modern water distribution analysis software.

Network Topology: A pipe network consists of nodes (junctions, reservoirs, tanks) connected by pipes. Junctions have demands (water consumption), reservoirs provide unlimited supply at fixed head, and tanks have finite volume with variable head. Pipes have properties: length, diameter, material (affecting roughness), and minor loss coefficients for fittings. The network forms a graph where loops create multiple flow paths.

Continuity Equation: At every junction, inflow must equal outflow: ΣQin=ΣQout+D\Sigma Q_{in} = \Sigma Q_{out} + D, where D is node demand. This ensures mass conservation - water entering a node equals water leaving plus consumption. Reservoir nodes automatically supply whatever flow the network demands.

Energy Conservation: Around any closed loop, the sum of head losses must equal zero: Σhf=0\Sigma h_f = 0. Head loss in the clockwise direction must equal head loss in the counter-clockwise direction. This principle drives the Hardy-Cross iteration - if loop head loss isn't zero, flows need adjustment.

Hardy-Cross Iteration: The algorithm starts with initial flow guesses satisfying continuity. For each loop, it calculates: ΔQ=Σhfn×ΣhfQ\Delta Q = -\frac{\Sigma h_f}{n \times \Sigma \left|\frac{h_f}{Q}\right|}, where n=1.85 for Hazen-Williams or n=2 for Darcy-Weisbach. This correction adjusts all pipe flows in the loop simultaneously. Iterations continue until corrections fall below tolerance.

Head Loss Equations: Hazen-Williams: hf=10.67×L×Q1.852C1.852×D4.871h_f = \frac{10.67 \times L \times Q^{1.852}}{C^{1.852} \times D^{4.871}} meters, where C is roughness coefficient (150 for smooth pipes, 100 for rough). Darcy-Weisbach: hf=f×LD×V22gh_f = f \times \frac{L}{D} \times \frac{V^2}{2g}, where f is friction factor from Moody diagram based on Reynolds number and relative roughness.

Pressure Calculation: Once flows converge, pressure head at each node is calculated from the energy equation: Hj=HsourceΣhf(path)H_j = H_{source} - \Sigma h_f(path), where the path traces from source to junction. Pressure in meters of water head converts to kPa by multiplying by 9.81.