Gas Gathering Network Hydraulics: Compressible Flow Formulations, Nodal Convergence & API RP 14E Debottlenecking Guide

Executive Summary & Hydraulic Architecture

Natural gas gathering systems represent the critical hydraulic bridge between upstream production wells and centralized gas processing or transmission facilities. Unlike cross-country transmission pipelines that operate under steady, high-pressure, single-phase conditions, gathering networks exhibit unique operational complexities: 1. Branched Tree Topologies: Multiple wellheads feed into lateral lines, converging at header junctions before discharging through a common trunkline into a central compressor station or slug catcher. 2. Variable Supply Pressures: Wellheads produce at varying reservoir depletion pressures, meaning low-pressure wells must overcome the common backpressure exerted by high-pressure wells sharing the same gathering manifold. 3. Compressible Flow Non-Linearities: Gas density varies continuously along the pipeline length as pressure drops, leading to non-linear pressure gradients and velocity escalation. 4. Hydraulic Bottlenecks: Undersized branch lines or constricted trunklines induce excessive backpressure, causing severe production curtailment on upstream wells.

This engineering guide provides a rigorous derivation of the fundamental compressible flow equations (Weymouth, Panhandle A/B), details the backward-marching nodal pressure convergence algorithm, examines API RP 14E erosional velocity constraints, and establishes capital-efficient debottlenecking strategies.

1. Fundamental Compressible Gas Flow Formulations

1.1 The General Steady-State Isothermal Energy Equation

For steady-state, one-dimensional, compressible gas flow in a circular conduit of diameter $D$ and length $L$, the mechanical energy balance (accounting for friction and gas expansion) is:

$$\frac{dP}{\rho} + \frac{v \, dv}{g_c} + \frac{g}{g_c} \, dh + \frac{f v^2}{2 g_c D} \, dL = 0$$

Assuming horizontal or gently undulating terrain ($dh \approx 0$), substituting the real gas law $\rho = \frac{P M_w}{Z R T}$, and integrating from upstream ($1$) to downstream ($2$):

$$Q = C_e \left( \frac{T_b}{P_b} \right) \left[ \frac{P_1^2 - P_2^2 - E_e}{S_g \cdot T_f \cdot L \cdot \bar{Z} \cdot \bar{f}} \right]^{0.5} D^{2.5}$$

where: * $Q$ = Gas volumetric flow rate at base conditions ($\text{SCFD}$ or $\text{m}^3/\text{d}$) * $T_b, P_b$ = Base temperature and pressure ($520\text{ }^\circ\text{R}, 14.73\text{ psia}$ or $288.15\text{ K}, 1.01325\text{ bar}$) * $T_f, P_1, P_2$ = Flowing gas temperature and pressures * $S_g$ = Gas specific gravity relative to air ($M_{\text{air}} = 28.964$) * $\bar{Z}$ = Average gas compressibility factor across the segment * $\bar{f}$ = Moody Darcy-Weisbach friction factor * $E_e$ = Elevation head loss term

1.2 The Weymouth Equation (Fully Rough Turbulent Flow)

Developed by Thomas R. Weymouth (1912) for short, high-pressure gathering networks ($D \le 16\text{ inches}$), this equation assumes fully rough turbulent flow where the friction factor depends solely on pipe diameter:

$$f_{\text{Weymouth}} = \frac{0.032}{D_{\text{in}}^{1/3}}$$

Substituting this into the general flow equation yields the standard field form:

$$Q = 433.5 \cdot E \cdot \left(\frac{T_b}{P_b}\right) \cdot \left[ \frac{P_1^2 - P_2^2}{S_g \cdot T_f \cdot L_m \cdot \bar{Z}} \right]^{0.5} \cdot D_{\text{in}}^{8/3}$$

Solving for upstream square pressure ($P_1^2$):

$$P_1^2 = P_2^2 + \left[ \frac{Q_{\text{scfd}}}{433.5 \cdot E \cdot (T_b / P_b) \cdot D_{\text{in}}^{8/3}} \right]^2 \cdot (S_g \cdot T_f \cdot L_m \cdot \bar{Z})$$

where $E$ is the pipeline hydraulic efficiency factor (typically $0.85\text{--}0.95$ for carbon steel gathering lines with fittings), $L_m$ is length in miles, and $D_{\text{in}}$ is internal diameter in inches.

1.3 The Panhandle Equations (Smooth and Medium Pipes)

For larger transmission lines and long gathering trunklines ($D > 12\text{ inches}$), the Panhandle equations incorporate Reynolds-number-dependent friction factors:

• Panhandle A: Designed for Reynolds numbers from $5 \times 10^6$ to $11 \times 10^6$: $$Q = 435.87 \cdot E \cdot \left(\frac{T_b}{P_b}\right)^{1.0788} \cdot \left[ \frac{P_1^2 - P_2^2}{S_g^{0.8539} \cdot T_f \cdot L_m \cdot \bar{Z}} \right]^{0.5394} \cdot D_{\text{in}}^{2.6182}$$
• Panhandle B (Modified): Designed for fully turbulent high-flow large-diameter lines ($Re > 1 \times 10^7$): $$Q = 737.0 \cdot E \cdot \left(\frac{T_b}{P_b}\right)^{1.02} \cdot \left[ \frac{P_1^2 - P_2^2}{S_g^{0.961} \cdot T_f \cdot L_m \cdot \bar{Z}} \right]^{0.510} \cdot D_{\text{in}}^{2.530}$$

Comparison of Hydraulic Formulations

2. Spatial Pressure Profiles & Average Line Pressure

Because gas expands as it flows down a pressure gradient, the velocity increases and the pressure drop per unit length ($dP/dx$) accelerates toward the outlet.

Spatial Pressure Distribution

Integrating from the inlet to distance $x$:

$$P(x) = \sqrt{P_1^2 - \left( P_1^2 - P_2^2 \right) \frac{x}{L}}$$

Notice that $P(x)$ is parabolic: the pressure drops gradually near the wellhead inlet and plunges steeply near the header junction.

Pressure (bar)
   ^
P1 |---------....
   |             ````---....
   |                        ```--..
P2 |                               ```\
   +-----------------------------------> Distance x (km)
   0                                   L

The Weymouth Spatial Average Pressure ($\bar{P}$)

Because pressure is non-linear with distance, an arithmetic mean $(P_1 + P_2)/2$ significantly overestimates average gas density and underestimates volumetric flow. Integrating the pressure profile over the line length yields the Weymouth two-thirds average formula:

$$\bar{P} = \frac{1}{L} \int_0^L P(x) \, dx = \frac{2}{3} \left[ P_1 + P_2 - \frac{P_1 P_2}{P_1 + P_2} \right] = \frac{2}{3} \left( P_1 + \frac{P_2^2}{P_1 + P_2} \right)$$

This average pressure $\bar{P}$ must be used to calculate the pseudo-reduced pressure $P_{pr} = \bar{P} / P_{pc}$ and the average compressibility factor $\bar{Z}$ for the segment.

3. Nodal Pressure Convergence in Branched Networks

A gathering network is modeled as a directed acyclic graph (DAG) where wellheads are source nodes, pipe segments are edges, and headers/compressor manifolds are junction nodes.

[Well Pad 1] ── Branch 1 ──┐
[Well Pad 2] ── Branch 2 ──┼──> [Header Junction] ── Main Trunkline ──> [Central Compressor (P_sink)]
[Well Pad 3] ── Branch 3 ──┘

Algorithm: Backward Hydraulic Marching

1. Fixed Boundary Sink: The network boundary condition is fixed at the delivery sink (Central Gathering Station suction pressure $P_{\text{cgs}}$).
2. Trunkline Resolution: The combined field flow rate is: $$Q_{\text{trunk}} = \sum_{i=1}^N Q_{\text{pad}, i}$$ Back-calculate the junction pressure $P_{\text{jct}}$ from $P_{\text{cgs}}$ using the Weymouth equation. Because $\bar{Z}$ depends on $\bar{P}$ (which depends on $P_{\text{jct}}$), iterate 2–3 times until $P_{\text{jct}}$ converges within $\pm 0.01\text{ bar}$.
3. Lateral Branch Marching: With $P_{\text{jct}}$ fixed at the downstream end of all converging lateral branches, back-calculate each individual wellhead flowing pressure $P_{\text{wh}, i}$: $$P_{\text{wh}, i}^2 = P_{\text{jct}}^2 + \left[ \frac{Q_i}{433.5 \cdot E \cdot (T_b / P_b) \cdot D_i^{8/3}} \right]^2 (S_g \cdot T_f \cdot L_i \cdot \bar{Z}_i)$$
4. Hydraulic Bottleneck Identification: Compare each calculated wellhead pressure $P_{\text{wh}, i}$ against the pad's Maximum Allowable Wellhead Pressure (MAWHP) or minimum reservoir drawdown deliverability. The segment exhibiting the highest pressure gradient ($\Delta P / L$) or causing a wellhead to breach MAWHP is the field bottleneck.

4. API RP 14E Erosional Velocity Screening

High gas velocities in gathering pipelines cause acoustic vibration, pressure drop spikes, and erosional-corrosive wall thinning, particularly when entrained sand grains or liquid droplets are present.

Erosional Velocity Limit Formula

API Recommended Practice 14E provides the standard empirical criterion for fluid velocity:

$$v_e = \frac{C}{\sqrt{\rho_g}}$$

where: * $v_e$ = Maximum allowable erosional velocity ($\text{ft/s}$ or $\text{m/s}$) * $\rho_g$ = In-situ gas density at flowing conditions ($\text{lb/ft}^3$ or $\text{kg/m}^3$) * $C$ = Empirical design constant: * $C = 100$: Continuous solids-free gas service (carbon steel) * $C = 125$: Intermittent service without solids * $C = 150\text{--}200$: Corrosion-resistant alloys (CRAs) or inhibited systems

Actual In-Situ Gas Velocity

The actual superficial gas velocity at flowing conditions ($T_f, P$) is:

$$v_{\text{actual}} = \frac{Q_{\text{actual}}}{A_{\text{pipe}}} = \frac{Q_{\text{scfd}} \cdot (P_b / P) \cdot (T_f / T_b) \cdot Z}{\frac{\pi}{4} D_{\text{pipe}}^2 \cdot 86400}$$

In field units ($Q$ in MMSCFD, $P$ in psia, $T$ in $^\circ\text{R}$, $D$ in inches):

$$v_{\text{actual}} = 60.0 \cdot \frac{Q_{\text{mmscfd}} \cdot Z \cdot T_f}{P_{\text{psia}} \cdot D_{\text{in}}^2} \quad (\text{ft/s})$$

Screening Decision Rules

• $v_{\text{actual}} / v_e < 0.70$: Optimal hydraulic operating range (conservative).
• $0.70 \le v_{\text{actual}} / v_e \le 0.85$: Warning zone: elevated acoustic noise and accelerated carbon dioxide/hydrogen sulfide sweet-gas corrosion.
• $v_{\text{actual}} / v_e > 0.85$: Severe hydraulic risk: pipe wall thinning, acoustic fatigue on small-bore connections, and risk of catastrophic piping breach. Requires line looping or diameter expansion.

5. Network Debottlenecking: Looping vs. Booster Compression

When a gathering lateral or trunkline reaches its hydraulic capacity, two capital mitigation strategies exist:

Original Single Line:
[Inlet P1] =======================================> [Outlet P2]   (Flow = Q1)

Looped Pipeline Segment:
           ==================== (Loop Line) =======
          /                                        \
[Inlet P1] ========================================> [Outlet P2]   (Flow = Q2 > Q1)

Mathematical Theory of Pipeline Looping

When a parallel pipe ("loop") of length $L_{\text{loop}}$ and diameter $D_{\text{loop}}$ is installed along an existing pipeline of length $L$ and diameter $D_1$: 1. The effective hydraulic resistance of the looped section drops dramatically because flow splits between parallel paths according to: $$\frac{Q_1}{Q_{\text{loop}}} = \left( \frac{D_1}{D_{\text{loop}}} \right)^{8/3}$$ 2. For a 100% full-length loop of identical diameter ($D_1 = D_{\text{loop}}$), each line carries $50\%$ of the total flow. Because pressure drop is proportional to $Q^2$: $$\Delta P_{\text{looped}} = \left(\frac{1}{2}\right)^2 \Delta P_{\text{original}} = 0.25 \cdot \Delta P_{\text{original}}$$ The pressure drop decreases by $75\%$ (or alternatively, field throughput can be increased by $200\%$ for the exact same pressure drop).

Capital Decision Matrix: Looping vs. Compression

6. Summary Design Checklist for Field Gathering Engineers

1. Verify Gas Composition: Confirm specific gravity $S_g$ and evaluate $\text{CO}_2 / \text{H}_2\text{S}$ acid gas fractions to set appropriate $Z$-factor correlations and API 14E $C$-values.
2. Model Downstream Pressure Boundary: Accurately measure compressor suction pressure fluctuations under full and partial load.
3. Screen Erosional Limits at Outlets: Always check $v_{\text{actual}} / v_e$ at the lowest pressure point of each segment (the outlet), where gas velocity reaches its peak.
4. Monitor Wellhead Pressure Margins: Maintain at least a $10\text{--}15\text{ bar}$ buffer between wellhead flowing pressure and reservoir shut-in pressure to ensure sustained well production without liquid loading.
5. Debottleneck Strategically: Utilize ChemProCal's Gas Gathering Network Balancing Analyzer to simulate branch diameter upsizing and partial looping scenarios prior to committing field capital.