- Two-phase gas-liquid flow is not simply a mixture of single-phase flows — the interaction between phases creates flow regimes, pressure drop behaviour, and valve loading conditions that are fundamentally different from either phase alone, and sizing a valve using single-phase methods in a two-phase system produces results that range from inaccurate to dangerously wrong.
- Slug flow — the intermittent, high-momentum slugs of liquid separated by gas pockets that occur in horizontal and near-horizontal pipelines — generates impulsive mechanical loads on valves that can be orders of magnitude higher than the forces from steady two-phase flow, and failure to account for slug loads is a leading cause of valve body cracking, actuator damage, and pipeline flange leaks.
- Control valves in two-phase service require special trim selection, velocity limits, and noise assessment because the sudden vaporisation or condensation of the liquid phase as pressure changes across the trim produces cavitation-like damage mechanisms, extreme noise, and vibration that destroy conventional trims within weeks.
Understanding Two-Phase Flow Regimes
Before sizing any valve for two-phase service, you need to know what the flow actually looks like inside the pipe — not just what the average flow rates are, but what regime the flow is in. The flow regime is the spatial arrangement of the gas and liquid phases relative to each other, and it determines virtually everything important about pressure drop, heat transfer, and the loads that the flowing mixture imposes on piping components including valves.
In horizontal pipelines, the primary flow regimes from low to high gas velocity are: stratified flow, where gravity separates gas above and liquid below in distinct layers with a smooth or wavy interface; slug flow, where liquid bridges periodically fill the pipe cross-section and are propelled by the gas pressure, creating alternating regions of liquid slug and gas bubble; annular flow, where liquid forms a film on the pipe wall and gas flows rapidly through the core; and mist flow, where liquid is dispersed as droplets in the continuous gas phase. In vertical upward flow, the sequence is different: bubble flow at low gas fractions, slug flow or churn flow at intermediate gas fractions, and annular flow at high gas velocities.
The flow regime for a given set of conditions can be estimated using flow regime maps — Baker charts for horizontal flow, Hewitt and Roberts maps for vertical flow — which plot a pair of coordinates derived from the gas and liquid mass fluxes and physical properties to identify the regime boundary. More sophisticated two-phase flow simulators (OLGA, LedaFlow, PIPESIM) can calculate regime transitions with greater accuracy, accounting for pipe inclination, diameter, and fluid property variations. For any new two-phase flow valve application, knowing the flow regime is the starting point for all other calculations.
Two-Phase Pressure Drop Through Valves
The pressure drop across a valve in two-phase flow cannot be calculated accurately using single-phase methods even if you use the average fluid properties of the mixture. The reason is that the two phases behave very differently as they accelerate through the valve trim: the gas phase, being compressible, accelerates more than the liquid phase, and the interaction between the phases as they negotiate the tortuous flow path through the trim creates additional pressure losses not captured by any single-phase model.
Several methods have been developed for two-phase pressure drop calculation through valves. The homogeneous model treats the two-phase mixture as a single pseudo-fluid with mixture properties (density, viscosity) calculated from the component phase properties weighted by the volume fractions. It is the simplest approach and works reasonably well in the mist flow regime where the phases are truly well-mixed. It significantly underestimates pressure drop in stratified or slug flow regimes where the phases are not homogenised.
The separated flow models — most importantly the Lockhart-Martinelli correlation — treat each phase as flowing independently and calculate a pressure drop ratio (the two-phase multiplier) that accounts for the interaction between phases. The Lockhart-Martinelli multiplier is a function of the Lockhart-Martinelli parameter X, which is the square root of the ratio of the liquid-only to gas-only pressure drops. This approach works better than the homogeneous model for stratified and slug flow regimes but still involves significant uncertainty — empirically, two-phase pressure drop predictions are rarely better than plus or minus 30 percent, even with sophisticated methods.
For control valve sizing in two-phase service, the IEC 60534 standard provides a two-phase extension of the single-phase liquid and gas sizing equations. The approach uses the flowing quality (mass fraction of gas) to weight between liquid and gas limiting cases. It is not a rigorous two-phase model but provides a systematic framework that is more defensible than an ad-hoc estimate. Many control valve manufacturers have their own two-phase sizing methods, often based on proprietary empirical data from flow tests, and it is worth using these alongside the IEC method and comparing results.
Slug Flow: The Most Dangerous Regime for Valves
Slug flow deserves extended discussion because it is both common in industrial pipelines and seriously underestimated in its mechanical impact on valves. A liquid slug travelling through a pipeline at gas velocity — which may be several metres per second — has significant momentum. When that slug arrives at a valve, fitting, or change of direction, its momentum must be absorbed by the piping structure. The force involved is the product of the liquid mass flow rate and the velocity change, and for large slugs in gas-driven systems it can be enormous — tens of kilonewtons on the valve body and adjacent fittings.
This slug impact force is impulsive in nature — it rises quickly to a peak as the slug front arrives and then falls away as the slug passes. The dynamic response of the piping structure to these repeated impulsive loads can create resonance if the slug frequency matches a natural frequency of the piping span. A pipeline that appears to be stably supported under steady-state conditions can experience large-amplitude vibration when slug flow is present, with accelerations and displacements that fatigue support structures, loosen bolted connections, and eventually crack pipe welds and valve bodies at stress concentrations.
Designing for slug flow requires several approaches applied together. Slug characterisation — estimating slug frequency, slug length, and slug liquid holdup — should be performed at the design stage using a two-phase flow simulator. The resulting slug force time history is then used as input to a piping structural analysis to identify whether resonance is possible and what the maximum structural response will be. Valve bodies and adjacent pipework must be designed with adequate strength and toughness to withstand the calculated slug forces with appropriate safety margin. Pipe supports must be designed not just for static loads but for the dynamic loads from slug impacts — clamps and guides that are adequate for steady-state operation may fail rapidly under slug conditions.
Slug catchers — large vessel or finger-type separators — are installed at production manifolds and gas processing facilities to capture arriving slugs before they enter the processing plant. The slug catcher absorbs the incoming liquid holdup from the production pipeline and releases it to the downstream process at a controlled rate. A properly designed slug catcher protects downstream valves and equipment from the worst slug impacts. Sizing the slug catcher for the maximum expected slug volume is critical — an undersized slug catcher that overflows during a large slug event passes the slug directly to downstream equipment, potentially with even higher velocity than before.
Control Valve Trim for Two-Phase Service
The control valve trim in two-phase service must handle the coexistence of liquid and gas and the potential for phase change as pressure drops across the trim. When liquid at near-saturation conditions passes through the low-pressure zone immediately downstream of the control valve trim, the local pressure can drop below the saturation pressure, causing a portion of the liquid to flash to vapour. This flashing process is energetically violent — the specific volume of steam or vapour is hundreds of times greater than the specific volume of the liquid it was formed from — and the resulting high-velocity two-phase jet erodes valve trim components at a rate that destroys standard trim in very short service periods.
The velocity of the two-phase jet leaving the trim is the primary driver of erosion rate. Trim selection for two-phase service must limit the maximum velocity to levels where erosion rate is acceptable — typically below 30 to 50 metres per second for liquid-continuous two-phase flows with moderate solid content, and lower still for slurry or solid-laden two-phase streams. Multi-stage pressure reduction trim — cage or tortoise-shell style trims that reduce pressure in multiple steps rather than in a single orifice — reduces the velocity at each stage and dramatically reduces both erosion rate and noise generation.
For severe two-phase erosive service — gas well choke valves handling sand-laden gas-liquid mixtures, for example — the trim material is as important as the trim geometry. Tungsten carbide inserts in the choke orifice provide excellent erosion resistance for moderate conditions. For the most severe erosive conditions, silicon carbide, polycrystalline diamond (PCD) coated components, or sintered ceramic orifice inserts may be specified. These ultra-hard materials can provide service lives ten to fifty times longer than standard alloy steel trim in high-velocity erosive service, justifying their substantial cost premium.
Noise and Vibration in Two-Phase Valve Applications
Two-phase flow through a control valve generates noise through multiple mechanisms: turbulence in the two-phase jet downstream of the trim, bubble collapse if the flow passes through a cavitation cycle (pressure below vapour pressure followed by pressure recovery above vapour pressure), and flashing vapour expansion. In gas-continuous flows at high velocity, aerodynamic noise from the high-velocity gas jets contributes as well. The combination of these mechanisms can produce sound pressure levels of 110 to 130 dB at 1 metre from the pipe surface — high enough to cause immediate hearing damage, to cause pipe wall vibration fatigue, and to alarm or disturb sensitive instrumentation in adjacent systems.
Noise prediction for two-phase control valves is less well developed than for pure gas or pure liquid service. The IEC 60534-8 standard covers noise prediction for liquid and gas service but does not directly address two-phase flow. Empirical experience and proprietary manufacturer models are typically used. For high-noise applications, the mitigation options are the same as for single-phase noise: multi-stage trim to reduce pressure drop per stage, downstream diffuser devices that expand the valve outlet velocity over a larger cross-section, acoustic pipe insulation to limit external radiation, and silencers on gas-venting systems.
Valve body vibration from two-phase flow must also be addressed. A control valve body that vibrates excessively fatigues the actuator mounting brackets, loosens the positioner feedback linkage, works the packing gland loose, and eventually cracks the body-to-pipeline connections. Adequate pipe support adjacent to the valve — particularly supports that resist lateral movement in the plane of the slug or two-phase flow direction — is essential. If the installed valve vibrates visibly in service, do not simply increase support stiffness without understanding the source of the vibration: increasing stiffness in a resonance situation can worsen the response if it shifts the natural frequency closer to the excitation frequency rather than further away.
Choke Valves for Wellhead Two-Phase Service
The choke valve — a specialised control valve for oil and gas wellhead and wellhead manifold service — operates in two-phase conditions virtually by definition. Reservoir fluids arriving at the wellhead are at high pressure and contain dissolved gas that partially flashes to free gas as pressure is reduced across the choke. The choke must handle this two-phase mixture at extremely high differential pressures (hundreds of bar in some high-pressure wells), high gas velocities, potentially significant sand production, and highly variable flow conditions as well production rates change over the producing life of the well.
Choke valves for wellhead service are typically needle-and-seat or cage-and-orifice designs using extremely hard materials (tungsten carbide or polycrystalline diamond for the orifice and seat) because the combination of high velocity, abrasive sand, and phase change is among the most erosive services in the entire process industry. Choke trim is consumable in the most severe sand-producing wells — replacement intervals of weeks rather than years are not uncommon, and the economics of the choke trim maintenance program must be built into the well’s operating cost model. Smart choke valves with integrated erosion monitoring — using electrical resistance probes or acoustic emission sensors — allow maintenance intervals to be based on actual measured erosion rather than fixed time intervals, potentially extending trim life while reducing unplanned failures.
Pigging and Two-Phase Flow Valve Interaction
Pipelines carrying two-phase gas-liquid flows are pigged regularly to remove liquid accumulations from low points in the line, to remove wax or hydrate deposits, and to inspect the pipe interior. A pig travelling through a two-phase pipeline accumulates liquid ahead of it and pushes it as a large liquid slug at the pig’s velocity — which may be 2 to 5 metres per second or more. When this pig-induced mega-slug arrives at the downstream end of the pipeline and passes through the receiving facility valves, the slug force is much larger than anything the steady-state two-phase flow would produce.
Valve design for pig receiving duty must account for the pig passage force and for the surge of liquid flow that follows the pig. The receiving facility is normally designed with a pig receiver vessel and manual divert valves that capture the pig and isolate it from the downstream process. Control valves downstream of the pig receiver may see the accumulated liquid surge even if the pig itself is captured. Slug catchers at pig receiving facilities serve double duty — capturing pig-driven slugs as well as production slugs — and must be sized for the maximum liquid volume accumulated by the pig over the full pipeline pigging interval.
Measurement and Metering in Two-Phase Lines
Measuring flow in two-phase lines is genuinely difficult, and the presence of a two-phase flow regime at any metering element — orifice plate, venturi, vortex meter, Coriolis meter — produces measurement errors that can be substantial. Most custody transfer metering systems require single-phase flow at the measurement point, which means the two-phase mixture must be separated before metering, metered individually for each phase, and the results combined to give a total hydrocarbon flow rate. This is the basis of the test separator systems used in oil and gas production allocation metering.
Where true two-phase metering is required — in subsea multiphase flow meters, for example, where installing separation equipment on the seabed is impractical — dedicated multiphase metering technology using venturi-based pressure measurements combined with gamma-ray or microwave hold-up measurements is used. These technologies are improving rapidly but still carry larger measurement uncertainties than single-phase meters, which has direct commercial implications for production allocation between co-mingled well streams.
Any valve upstream of a metering point in a two-phase system must be designed to minimise flow disturbance. A partially open control valve immediately upstream of an orifice plate creates a disturbed velocity profile that can double or triple the measurement error of the orifice meter compared to fully developed pipe flow. The minimum straight pipe run requirements specified in ISO 5167 for orifice plates and venturi tubes assume single-phase, fully developed flow — in two-phase service these requirements should be treated as absolute minimums and additional straight run specified where at all possible.