Solving The Wet Gas Flow Measurement Challenge

July 2010 Vol. 237 No. 7

Bryan Trostel


One of the biggest problems encountered in measuring natural gas flows occurs when liquids flow with the gas. Traditional gas meters are not designed to cope with such wet gas flows. The main issues when dealing with wet gas flows include (1) knowing how wet the gas is, (2) knowing how wet gas affects flow measurement and (3) knowing what systems exist to correctly measure the wet gas flow. Wet gas flow metering is of increasing importance to the natural gas production industry. Therefore, wet gas flow testing facilities are also of increasing importance to the industry. State-of-the-art wet gas flow testing facilities allow the research and development and verification of wet gas flow metering technologies by meter manufacturers, operators and regulatory authorities. This article describes the operation of the wet gas testing facility at Colorado Engineering Experiment Station, Inc. (CEESI).

What is wet gas flow? It is a flowing mixture of gas and liquid where the liquid makes up a relatively small part of the mixture by volume. The liquid can be made up of hydrocarbons and/or free water. The flow conditions dictate how the liquid phase is dispersed throughout the pipe. The description of the physical distribution of the liquid phase with the gas phase is termed the flow pattern (or the flow regime). The flow pattern has a considerable influence on the reaction of most meters to the wet gas flow.

At relatively high pressures and flow rates for horizontal or vertical flow, the flow pattern could be mist, where all the liquid flows in small droplets entrained in the gas. At relatively low pressures and flow rates for horizontal flow, the flow pattern could be stratified, where the liquid flows at the base of the pipe with the gas flowing above. However, in many cases moderate pressures and flow rates produce complicated and transient flow patterns that are difficult to predict theoretically due to the fact that they are influenced by many factors: meter orientation, fluid velocity, liquid properties, pipe size, liquid/gas ratios and others. This inability to predict the flow pattern theoretically drives the need for wet gas flow facilities to replicate actual field conditions in order to test wet gas meter systems.

Several terms are commonly used to describe the relative amounts of liquid and gas in a flowing stream. A qualitative term for describing the amount of liquid with the gas is the “liquid loading.” There are several quantitative terms: gas volume fraction (GVF) is the volume of the gas flow divided by the total volume of fluid flowing. Liquid load is the ratio of liquid-to-gas-mass flowrates. Another term commonly used is the Lockhart-Martinelli parameter. This is a non-dimensional method of describing the relative wetness of a gas. The higher the Lockhart Martinelli parameter the wetter the gas flow.

Industry has found the development of accurate wet gas flow meters a difficult task. Attempting to meter wet gas flow with a gas flow meter can cause many problems.

How does wet gas flow affect measurement? Many gas flow meter designs have been tested with wet gas flows. The most commonly tested and used gas meter with wet gas flow is the differential pressure (DP) meter (e.g. orifice, cone, Venturi meters). Liquid presence with a gas flow induces a DP meter to produce a higher differential pressure than would exist if the gas flowed alone. This results in an over measurement of the gas flow. According to many technical papers, the gas flow rate errors induced by a liquid’s presence with a gas flow can be 10 and greater percent.

Other meter types such as Coriolis, turbine and ultrasonic meters have been tested with wet gas flows. Wet gas flow causes all these meter types to have significant measurement issues. The generic Coriolis, turbine and ultrasonic meters will–in general–incur significant gas flow rate measurement errors for trace liquid entrainment with the gas flow. At moderate to high liquid loading wet gas flows they can fail completely and give no flow rate predictions whatsoever. Alternatively they can give highly inaccurate random flow rate predictions that are not reproducible.

For gas meters that do exhibit repeatable and reproducible wet gas flow responses, a considerable amount of flow testing is required to gather the required data to form a mathematical prediction method to account for the liquid’s effect. This wet gas flow correction method is required to take account of the liquid effects on the gas flow meter and therefore accounts for the flow pattern effects. This fact makes it impractical to attempt to estimate liquid induced gas flow rate prediction errors through theoretical means only, and facilitates the requirement for wet gas flow facility testing.

Even when gas meters have low-uncertainty wet gas correction factors available, these corrections still require that the liquid flow rate be known externally to the wet gas meter system. For this reason meter manufacturers are developing and marketing wet gas metering systems. Again, these systems are heavily dependent on the services of wet gas flow facilities for research and development and verification.

What can be done to correctly measure the wet gas? Through the extensive use of wet gas flow test facilities manufacturers are having significant success in developing wet gas flow meters. The key to accurately measuring wet gas flow is to understand how a metering system reacts to liquids flowing with the gas stream. In order to understand this, it is necessary to conduct testing over the full range of conditions that the meter will see in the field.

The marketed wet gas flow meters can generally be grouped into three generic types. All three are developed with extensive wet gas flow testing. The first generic design uses a DP meter with the additional measurement of the permanent pressure loss (or “head loss”). The relationship between the two measurements, i.e. the traditional DP and the head loss gives enough information to predict the two unknowns, i.e. the gas and liquid flow rates. The second generic design uses two dissimilar gas meters with different responses to wet gas flows in series. The different meter performance allows a comparison between the two meter outputs and a measurement by difference technique to be employed. The third generic design uses a DP meter with phase fraction devices (e.g. gamma ray, microwave, capacitance devices etc), embedded in the meter body. Cross referencing the output of the multiple instruments allows derivation of the gas flow, liquid flow and water cut. For all three generic designs the manufacturers are heavily dependent on wet gas flow facilities for research and development. Operators and regulatory authorities are heavily dependent on wet gas flow facilities to verify all wet gas meter designs performance claims.

Wet Gas Testing Facility
Colorado Engineering Experiment Station Inc (CEESI) operates a state-of-the-art wet gas flow test facility. The facility can flow liquid hydrocarbon and/or water with natural gas. As seen in the figure, the facility is a loop in which the gas and liquids are circulated. Distribution quality natural gas is compressed into storage. This stored gas is used to pressurize the loop. The gas is driven around the loop using circulation compressors. The natural gas flow rate is metered using a gas turbine meter located in a dry gas section of the system. The liquid measurements take place prior to the liquid injection into the gas flow. Liquid hydrocarbon and water flows are metered separately in their own supply pipe work. All liquid flows are metered with a bank of Coriolis meters to accurately meter the wide liquid flow rate turndowns as well as liquid density information. Flowrates of each component fluid can be individually adjusted to capture the ranges that will be seen in the field installation. Following the liquid injection into the gas stream, the liquids are driven through the test section by the gas flow.

Nominal test section pipe size is 4-inch, although other pipe sizes can be tested depending on desired volume flow rate ranges. The CEESI wet gas flow loop has a substantial test section length to allow for replicating a variety of field installations or to allow for multiple meter installations. Replicating the field piping arrangement as closely as possible reduces any installation differences between the test facility and the field application. Installation differences can cause wet gas flow meters to alter performance as the installation alters the flow pattern entering the meter. Testing multiple meters in series can considerably reduce the cost of testing per meter.

After passing through the test sections, the fluid stream enters a gas/liquid separator. The gas is returned from the gas/liquid separator to the dry gas portion of the loop where it is driven at the desired flow rate by the gas compressors and metered by the gas turbine meter. The liquids exit the gas/liquid separator and enter a liquid/liquid separator. The water and liquid hydrocarbon fluids are then separated. They are then individually driven at the desired flow rate by dedicated liquid pumps and metered by the Coriolis meters.


Diagram of the wet gas meter testing loop.

During a test point, care is taken to ensure steady state conditions. Initially, the entire system is pressurized to the desired pressure. The circulation compressors then drive the natural gas flow through the loop. The heat of compression is used to raise the temperature of the system to the desired temperature. Heat exchangers then moderate this temperature as desired. Due to the natural pressure loss through the loop, the pressure naturally varies across the system. Therefore a constant mass flow rate of fluids has a varying volume flow rate through the loop. Hence, the reference flow rates are considered in mass flow. In steady state re-circulating flow the mass flow references of the gas, water and hydrocarbon liquid flows are known to be true values at every point in the system. Volume flow rate references are obtained by local pressure and temperature measurements and PVT calculations converting this known mass flow to volume flow values.

The CEESI wet gas loop has a possible pressure range of 200 to 1,100 psia and can operate at temperatures between 70 and 125 degrees F. The natural gas flow range is 2,020 to 21,600 actual cubic feet per hour. The water and liquid hydrocarbon can move at rates of 0.04 to 60 gallons per minute and 0.2 to 60 gallons per minute respectively.

Conclusions
Wet natural gas metering is very important to the oil and gas industry. The effect of wet gas flow on single phase flow meters make it difficult to get accurate flow measurements using traditional single phase flow measurement techniques. The effect of wet gas flow on single phase flow meters is influenced by several parameters including meter type, amount of liquids, and flow pattern. Some gas meters do have repeatable wet gas flow performance and in such cases this performance can be characterized. Alternatively manufacturers have developed, and are still developing, various wet gas flow meters systems. For both these wet gas flow metering methods, manufacturers, operators and regulatory bodies are dependent on wet gas flow facilities to both research and develop and verify the performance of wet gas flow meter technologies.

The Author
Bryan Trostel is a staff engineer with the Colorado Engineering Experiment Station Inc., in Nunn, CO. Ph: 970-897-2711, btrostel@CEESI.com, www.ceesi.com.

Find articles with similar topics