Pipe networks and flow in series and parallel

Series and parallel pipes, Dupuit equivalent pipe, branched and looped networks (Hardy Cross), pumps in networks and maximum power transmission.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Real piping is never one straight pipe. Engine cooling circuits split between the radiator and the bypass, fuel and lubrication systems branch to several consumers, and water-supply grids are loops. Engineers need to know how flow divides, what head the pump must supply and how to replace a complicated arrangement with one equivalent pipe. The rules are simple analogies of electric circuits, but the "resistance" is non-linear (h ∝ Q²).

Key ideas

  • Pipe resistance. For turbulent flow with a given friction factor, Darcy–Weisbach can be written h_f = k·Q², where k = 8fL/(π²gD⁵). This makes the circuit analogy clear: head loss plays the role of voltage drop and flow rate the role of current, but the drop rises with Q², not Q.
  • Pipes in series. The same Q flows through every pipe, and the head losses add (friction plus any minor losses at the joints, such as sudden expansions or contractions).
  • Equivalent pipe (Dupuit's equation). A series combination can be replaced by one pipe of uniform diameter D_e and the same total length that gives the same loss for the same Q. With equal f: L/D_e⁵ = Σ(L_i/D_i⁵).
  • Pipes in parallel. The head loss between the two junctions is the same in every branch, because each branch connects the same two points with the same total energy. The flows add up to the total. With h = k·Q² for each branch, Q_i ∝ 1/√k_i ∝ √(D_i⁵/(f_i·L_i)). A larger, shorter or smoother branch takes more of the flow.
  • Branched pipes (three-reservoir problem). At a junction, continuity requires Σ inflows = Σ outflows. The unknown is the head at the junction: guess it, compute each branch flow from the energy equation, and adjust until continuity balances. The direction of flow in the middle branch depends on whether the junction head is above or below that reservoir's level.
  • Looped networks (Hardy Cross method). Two laws apply, as in Kirchhoff's laws: continuity at every node, and the algebraic sum of head losses around every closed loop is zero. Assume flows that satisfy continuity, compute the loop correction ΔQ = −Σ(kQ|Q|)/Σ(2k|Q|), apply it to each pipe in the loop, and repeat until ΔQ is small. Computers now do this, but GATE may test the principle or one iteration.
  • Pumps in a network. The pump head appears as an energy gain in the energy equation. The operating point is where the pump's H–Q curve meets the system curve H = static lift + Σk·Q².
  • Power transmission through a pipe. Hydraulic power delivered at the outlet is ρgQ(H − h_f). It is maximum when h_f = H/3, giving a maximum transmission efficiency of 2/3 (66.7 %) at that point.
  • Empirical alternative. For water distribution, the Hazen–Williams formula (with a roughness coefficient C from tables) is used instead of Darcy–Weisbach. It applies only to water at ordinary temperatures.
  • Minor losses are often neglected in long networks (L/D > 1000) but must be included in compact systems such as engine cooling circuits.

Formulas

h_f = 8·f·L·Q² / (π²·g·D⁵) = k·Q² h_f head loss (m), f Darcy friction factor, L length (m), Q flow (m³/s), D diameter (m), g = 9.81 m/s²; k resistance (s²/m⁵).

Q = Q₁ = Q₂ = …, H = h₁ + h₂ + … Pipes in series.

L/D_e⁵ = L₁/D₁⁵ + L₂/D₂⁵ + … Dupuit equivalent pipe (same f; L = L₁ + L₂ + …).

Q = Q₁ + Q₂ + …, h₁ = h₂ = … Pipes in parallel.

Q₁/Q₂ = √[(D₁⁵·f₂·L₂) / (D₂⁵·f₁·L₁)] Flow split between two parallel pipes.

ΔQ = −Σ(k·Q·|Q|) / Σ(2·k·|Q|) Hardy Cross loop correction (n = 2 for Darcy–Weisbach).

P = ρ·g·Q·(H − h_f), η = (H − h_f)/H, P_max at h_f = H/3 Power transmitted by a pipe with supply head H (m).

h_f = 10.67·L·Q^1.852 / (C^1.852·D^4.87) Hazen–Williams (SI; water only; C from tables).

Worked examples

Example 1 (standard) — pipes in series and equivalent pipe. Given: Q = 0.08 m³/s of water flows through pipe 1 (L = 300 m, D = 0.3 m) and then pipe 2 (L = 200 m, D = 0.2 m), both with f = 0.02. Neglect minor losses. Find the total head loss and the equivalent diameter for a single 500 m pipe.

  1. k₁ = 8 × 0.02 × 300/(π² × 9.81 × 0.3⁵) = 204.0 s²/m⁵ → h₁ = 204.0 × 0.08² = 1.306 m.
  2. k₂ = 8 × 0.02 × 200/(π² × 9.81 × 0.2⁵) = 1032.8 s²/m⁵ → h₂ = 1032.8 × 0.08² = 6.610 m.
  3. Total H = 1.306 + 6.610 = 7.916 m.
  4. Dupuit: 500/D_e⁵ = 300/0.3⁵ + 200/0.2⁵ = 123 457 + 625 000 = 748 457 → D_e = (500/748 457)^(1/5) = 0.2317 m. Answer: H ≈ 7.92 m; equivalent diameter ≈ 0.232 m. The smaller pipe causes most of the loss, because h ∝ 1/D⁵.

Example 2 (GATE level) — two pipes in parallel. Given: a total flow of 0.12 m³/s divides between pipe A (L = 500 m, D = 0.2 m, f = 0.02) and pipe B (L = 400 m, D = 0.15 m, f = 0.025). Find the flow in each and the head loss between the junctions.

  1. Equal head loss → Q ∝ √(D⁵/(fL)).
  2. a = √(0.2⁵/(0.02 × 500)) = √(3.2 × 10⁻⁵) = 5.657 × 10⁻³; b = √(0.15⁵/(0.025 × 400)) = √(7.594 × 10⁻⁶) = 2.756 × 10⁻³.
  3. Q_A = 0.12 × a/(a + b) = 0.12 × 5.657/8.413 = 0.0807 m³/s; Q_B = 0.12 − 0.0807 = 0.0393 m³/s.
  4. Check: h_A = k_A·Q_A² = [8 × 0.02 × 500/(π² × 9.81 × 0.2⁵)] × 0.0807² = 2582 × 0.006511 = 16.81 m; h_B = 10 881 × 0.0393² = 16.81 m. ✓ Answer: Q_A ≈ 0.081 m³/s, Q_B ≈ 0.039 m³/s, head loss ≈ 16.8 m.

Common mistakes

  • Adding head losses for parallel pipes or adding flows for series pipes — it is the other way round.
  • Splitting parallel flow in proportion to area or D², instead of √(D⁵/(fL)).
  • Forgetting that h ∝ Q², so doubling the flow quadruples the loss (series) and a branch's share does not scale linearly.
  • In the three-reservoir problem, fixing the flow direction in the middle pipe before finding the junction head.
  • Applying Hardy Cross corrections with the wrong sign for anticlockwise flows.
  • Using Dupuit's equation when the friction factors of the pipes differ.

For GATE ME

Expect: flow split between two parallel pipes, equivalent length or diameter of a series combination, head loss when a pipe is replaced by two parallel pipes (for example, with the total length or diameter changed), condition for maximum power transmission (h_f = H/3, η = 66.7 %), and the principles of Hardy Cross. Practise expressing everything in k·Q² form — it makes ratio questions one-line.

Quick check

  1. In parallel pipes, which quantity is common to all branches?
  2. Two identical parallel pipes replace one pipe of the same size. If f is the same, by what factor does the head loss fall for the same total Q?
  3. Two parallel pipes of equal length and f have D = 0.2 m and 0.1 m. What is Q_large/Q_small?
  4. What is the maximum efficiency of power transmission through a pipe?
  5. Which two laws does the Hardy Cross method enforce? Answers: 1. Head loss. 2. Each carries Q/2, so h falls to 1/4. 3. 2^2.5 = 5.66. 4. 66.7 % (h_f = H/3). 5. Continuity at nodes and zero net head loss around each loop.

Try answering each one aloud before you open it.

  1. 1.What is a pipe network in fluid mechanics?Concept

    A pipe network in fluid mechanics is a system of interconnected pipes used to transport fluids from one location to another. These networks can include various components such as pumps, valves, and reservoirs, and are designed to handle specific flow rates and pressures. Pipe networks can be arranged in series, parallel, or a combination of both, depending on the application and desired flow characteristics.

  2. 2.Explain the difference between series and parallel flow in pipe networks.Concept

    In a series flow, the fluid passes through each pipe or component sequentially, meaning the flow rate is the same through each section, but the pressure drops add up. In parallel flow, the fluid is divided among multiple paths, so the pressure drop across each path is the same, but the flow rate is divided among the paths. Series flow is typically used when a single path is needed, while parallel flow is used to distribute flow evenly or to reduce pressure drop.

  3. 3.Why are parallel pipe networks often used in cooling systems?Application

    Parallel paths let one pump feed several heat loads, such as engine block passages, a heater core and an oil cooler, and they lower the overall flow resistance. In parallel the head loss is common and the flows add, so the combined resistance is smaller than any single branch. Each branch's share depends on its resistance (Q ∝ √(D⁵/(fL))), so designers size branches or add orifices and thermostatic valves to get the intended split. Parallel paths also give redundancy if one passage is blocked.

  4. 4.What happens if a valve in a series pipe network is partially closed?Application

    If a valve in a series pipe network is partially closed, it increases the resistance to flow, leading to a higher pressure drop across the valve. This results in a reduced flow rate through the entire network, as the flow rate in a series network is the same throughout. The increased resistance can also lead to higher energy consumption by pumps trying to maintain the desired flow rate.

  5. 5.How does the addition of a pump in a series pipe network affect the system?Application

    Adding a pump in a series pipe network increases the pressure head, which can help overcome pressure losses due to friction and elevation changes. This allows for a higher flow rate or the ability to transport the fluid over longer distances. The pump effectively boosts the energy of the fluid, enabling it to move through the network more efficiently.

  6. 6.Calculate the total head loss in a series pipe network with three pipes, each having a head loss of 5 m.Numerical

    In a series pipe network, the total head loss is the sum of the head losses in each pipe. Therefore, the total head loss is 5 m + 5 m + 5 m = 15 m.

  7. 7.If two pipes in parallel have flow rates of 3 m³/s and 2 m³/s, what is the total flow rate in the network?Numerical

    In a parallel pipe network, the total flow rate is the sum of the flow rates in each path. Therefore, the total flow rate is 3 m³/s + 2 m³/s = 5 m³/s.

  8. 8.Explain how the Hazen-Williams equation is used in pipe network analysis.Concept

    Hazen–Williams is an empirical friction formula for water in pressure pipes. In SI units it is h_f = 10.67·L·Q^1.852 / (C^1.852·D^4.87), with h_f and L in m, Q in m³/s, D in m, and C a roughness coefficient from tables (about 100–150). It is popular in water-distribution design and Hardy Cross network calculations because C does not depend on Reynolds number, so no iteration on f is needed. It is valid only for water at ordinary temperatures in turbulent flow; for other fluids, use Darcy–Weisbach.

  9. 9.What are the advantages of using a looped pipe network?Application

    Looped pipe networks provide multiple paths for fluid flow, which enhances reliability and flexibility. If one path is blocked or requires maintenance, the fluid can still flow through alternate routes, minimizing disruptions. Additionally, looped networks can help balance pressure and flow rates across the system, improving overall efficiency and performance.

  10. 10.Describe the role of a surge tank in a pipe network.Concept

    A surge tank is used in a pipe network to absorb sudden changes in pressure, such as those caused by rapid valve closure or pump failure. It acts as a buffer, reducing the risk of water hammer, which can cause damage to pipes and equipment. By providing a space for excess fluid to temporarily accumulate, surge tanks help maintain stable pressure levels and protect the integrity of the network.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?