天然气露点测量

′

MOSITURE MEASUREMENT IN NATURAL GAS

Rolf Kolass, Michell Instruments GmbH, Friedrichsdorf, Germany

Chris Parker, Michell Instruments Ltd, Cambridge, UK

ABSTRACT

The measurement of moisture content in natural gas is extremely important, from a technical perspective and in order to ensure conformance to contractual specifications.

Typically, this measurement is one of the most difficult to perform successfully - natural gas sources are generally dirty, corrosive, heavily moisture laden and at high pressure. Moisture removal is a key stage of natural gas processing prior to sale and its efficiency is important in order to satisfy fiscal contractual obligations, to ensure that the gas is safe to transport and has the right properties for subsequent use.

This paper discusses the technology of natural gas processing and the demands it puts on the humidity measurement industry, both technically and commercially. This paper also describes the methods that may be employed to effect a measurement in natural gas and to ensure integrity and longevity.

Actual case histories are referred to within the paper as examples of good and bad practice.

INTRODUCTION

Natural gas extracted from underground sources is saturated with liquid water and heavier molecular weight hydrocarbon components. In order to meet the requirements for a clean, dry, wholly gaseous fuel suitable for transmission through pipelines and distribution for burning by end users, the gas must go through several stages of processing, including the removal of entrained liquids from the gas, followed by drying to reduce water vapour content. The dehydration of natural gas is critical to the successful operation of the production facility and the whole distribution train through to the end user. The presence of water vapour in concentrations above a few 10s of parts per million has potentially disastrous consequences. The lifetime of a pipeline is governed by the rate at which corrosion occurs which is directly linked to the available moisture in the gas which promotes oxidation. In addition, the formation of hydrates can reduce pipeline flow capacities, even leading to blockages, and potential damage to process filters, valves and compressors. Such hydrates are the combination of excessive water vapour with liquid hydrocarbons, which may condense out of the gas in the course of transmission, to form emulsions that, under process pressure conditions, are solid masses. Furthermore, in the processing of gas prior to transmission, a cold temperature separator is most often used to extract the heavier molecular components to avoid the formation of such hydrocarbon liquid condensates at prevailing pipeline operating temperatures that change with climate. The drying of natural gas to a dew point lower than the operating temperature of the chiller plant is of obvious importance to prevent freeze up problems, causing flow restriction, with resulting consequences in terms of plant operating efficiency.

′

For these reasons it is standard practice at natural gas production facilities, both on- and offshore, to measure the moisture content in natural gas on a continuous, on-line basis at critical points to ensure successful processing and efficient, reliable plant operation. The successful design, installation and operation of industrial hygrometers for such applications requires special consideration to be given to the particular nature and composition of the gas being measured and the processing techniques being utilised.

Dehydration Process

The most common processing technique for drying natural gas is that of simple mechanical separator, to divide the gas from the liquids of the two phase flow coming from the gas field, followed by glycol dehydration. Here a riser tower has an array of spray nozzles around it’s circumference through which glycol is injected, as a liquid desiccant, into the gas stream flowing up through the tower. The adsorption process results in moisture-laden glycol that coalesces into globules that are naturally forced, through flow dynamics, outward towards the wall of the tower. The liquid glycol is collected in trays, piped out of the tower and is regenerated by heating to evaporate the absorbed moisture prior to re-injection in a continuous operating, re-circulating loop. Such glycol contactors, as they are termed, are specified to achieve a moisture content of less than 3 Lb./MMSCF (pounds of moisture per million standard cubic feet of gas) under normal operating conditions.

The high flow velocity of gas through the contactor leads to possible carry through of glycol mist. Consideration must be given to this characteristic if the application of a hygrometer is to be successful in monitoring the performance of the dehydration process. Contamination of the moisture sensor or sample handling system results in a serious deadening in response for the analyser due to the moisture adsorption/desorption capacity of the glycol. A conventional coalescing filter with fibre element positioned at the front end of the sampling system can effectively protect the moisture sensor from contamination but will not solve response problems unless any collected liquids are flushed out of the filter housing by a continuous flow from the drain port (Fig. 1). A membrane type filter offers the best protection in such glycol applications but is restricted to a maximum operating pressure of 10 MPa. These filters also work on a bypass flow arrangement but use a micro-porous membrane of fluorocarbon material to offer superior protection.

Figure 1. Schematic Diagram and Photograph of a Typical Moisture Analyser

for Natural Gas Dehydration Plant

In addition to protection against liquid contamination, it is suggested by some suppliers of instrument sample filters for natural gas applications that an absorbent material, in the form of an in-line column of activated charcoal through which the sample flows, is used to remove

′

glycol vapour. In our experience the presence of glycol vapour, which has a low vapour pressure (130Pa at 53°C) and corresponding low maximum trace concentration, has little detrimental effect on the performance of sensing technologies like the Ceramic Moisture Sensor used for such measurements.

The measurement principle of such sensors is adsorption/desorption of water molecules into a hygroscopic layer between two conductive electrical plates - A substrate layer beneath and a porous top plate exposed to the flowing sample and through which moisture molecules freely permeate to maintain a natural equilibrium of moisture content. The variation of moisture adsorbed into the hygroscopic layer results in a corresponding change in the dielectric between the conductive plates and thus the ability to use this principle for continuous on-line measurement. However, the Glycol molecule has similarities to that of water in that both have polar covalent bonds with an unequal sharing of electrons between a bond of oxygen and hydrogen atoms which, as a result, become negatively and positively charged respectively.

As such both water and glycol molecules possess the potential to cause a response in the sensor if adsorbed into the hygroscopic layer as the oxygen atoms are attracted to the positively charged regions of the hygroscopic layer. However, the maximum possible concentration of glycol vapour is extremely low relative to that of moisture which means that, given effective filtration to remove glycol mist as prescribed above, any effects to measurement accuracy are negligible. However, the use of in-line adsorption cartridges, in addition to a filter, to remove such glycol vapour can have serious detrimental side effects. Activated charcoal is an effective desiccant with a capacity to adsorb moisture in addition to glycol vapour and as such will cause significant damping of the changes in moisture content of the analysis sample in response to process variations. If the process gas becomes drier, the activated charcoal will tend to act as a water source at the moisture analyser inlet and if the process gas gets wetter, the charcoal will desiccate it, giving a falsely dry reading.

Hydrocarbon dew point

For natural gas there are two dew-point temperatures of relevance, the water dew point, as we discuss here, and the hydrocarbon dewpoint. The latter is quite simply the temperature at which liquid hydrocarbons condense out of the gas upon cooling. Such liquid hydrocarbons comprise the heavier molecular weight components of the gas composition, typically butane and higher. This parameter, as with water dew point, requires dedicated processing plant (in the form of condensing chillers) and purpose designed measurement instrumentation. However, the significance to the measurement of water dew point arises if a moisture analyser using a condensing dew-point measurement technique is utilised such as a Bureau of Mines Apparatus1. This form of manual visual cooled mirror dewpointmeter, and any other type of automated, condensing dew-point analyser, may give confusing results when used for water dew-point measurement in natural gas. This is because of the difficulty in observing the water dew point separately from that of hydrocarbons and glycol that are highly likely to condense on the mirror surface at a higher temperature than the water dew point (Fig. 2). The use of a sensor based on a non-condensing measurement principle, such as the Ceramic Moisture Sensor, avoids this difficulty as it does not employ a condensation measurement technique and therefore will not suffer from such cross-measurement effects.

′

8070H2O DewpointPressure, Bara6050403020100-40-30-20-10oHC Dewpoint0+10Temperature, C

Figure 2. Hydrocarbon and Water Dew Point Variation with Pressure

for a Typical Natural Gas Composition

Conversion from measured dew point to moisture content

The pressure of natural gas is typically 4 to 8 MPa in processing plant and on-shore transmission whilst gas entering offshore pipelines is often compressed to 16 MPa or higher. In any dew-point analysis the influence of gas pressure must be considered. The Michell Ceramic Moisture Sensor as with its predecessors, the older aluminium oxide technologies, adsorb moisture in equilibrium with the gas sample flow to which it is exposed and thus exhibits a response to variations in water vapour pressure. Water vapour pressure is directly related to dew point, which enables such sensors to be calibrated accurately and easily in the parameter of dew point. The relationship between partial pressure of water vapour and dew point remains consistent irrespective of total gas pressure and the composition of the dry gas components. Thus such a sensor calibrated by the instrument manufacturers on known dew point calibration gases, usually performed at atmospheric pressure, can be applied to accurately determine the dew point of any process gas at any chosen analysis pressure.

The parameter of water dew point is the most widely used parameter to stipulate this element of gas quality in contractual supply specifications between gas producers and pipeline operators through to end customers. However, in some specifications for process plant such as glycol dehydration contactors as well as pipeline operations it is more common for a maximum permissible moisture content to be stipulated. The conversion from measured dew point to moisture content needs specific consideration to be given to the non-ideal behaviour of high-pressure natural gas that requires the use of enhancement factors when performing the conversion from measured dew point at known analysis pressure to moisture content.

A number of sources of such conversion data, originating mainly from work carried out at IGT2, Chicago, in the 1950’s are in common usage today and are reproduced in a current ASTM standard3. Data is only provided down to -40°C dew point, limiting its applicability in colder climate regions, where specifications for moisture content are stricter in order to avoid

′

the potential for condensation to occur with the associated problems discussed earlier. Furthermore, where desiccant columns are used as second stage dehydration plant, which is a necessity, fore instance, on a natural gas liquefaction plant to prevent freeze up inside the cryogenic process, then moisture levels need to be less than 1 ppmV, typically lower than 0.1 ppmV, which equates to less than -70°C dew point at process conditions. Extrapolations of the IGT data have been used by Michell Instruments and other moisture analyser manufacturers but this introduces increased uncertainty of measurement and can lead to disputes at custody transfer between gas producers, pipeline operators and end customers if different conversion data is used. Major European natural gas companies are now tending to use a new harmonised standard4 covering the full range of measurement. The conclusion of this work, leading to a consistent approach amongst gas companies and instrumentation suppliers, has an added importance given the increased prevalence of inter-country pipelines.

For applications requiring units of moisture content to be indicated then the conversion is usually performed in the measurement electronics of the hygrometer where the measured dew point, at a known pressure, is converted to the measurement unit desired by the natural gas company. Units in regular use include Lb./MMSCF (gas industry in USA and also world wide users of American designed processing plant) and mg/std.m³ (European specifications). Knowing the pressure of analysis for the primary dew point measurement is critical to the accuracy of the unit conversion as a percentage error in defining the analysis pressure will directly transpose into the same percentage error in calculating the moisture content. The analyser pressure should therefore either be fixed by a ‘peak shaving’ pressure regulator set to the minimum line pressure level or by on-line measurement of the analysis condition using a pressure transducer to provide a real time input of the variations with line pressure into the calculation (Fig. 1).

Sour natural gas

Further difficulty may also be experienced in interpreting water dewpoint measurements made in natural gas if a conversion to units of moisture content is required and if the gas composition contains a significant amount of carbon dioxide and/or hydrogen sulphide. Such ‘sour’ natural gas, as it is termed, is found in many gas fields in current production. The amount of moisture required to reach saturation water vapour pressure in H2S and CO2 is considerably higher than for moisture in methane or a ‘sweet’ natural gas composition at the same temperature. As such the water dew point measured in a sour gas, irrespective of the measurement principle applied, will be significantly lower than for a sweet gas containing the same moisture content. This needs to be compensated for, using published data5, and thus requires knowledge of the concentration of sour gas components.

A typical example for a sour natural gas project can be taken from a major producer in Northern Germany. Here a specification for maximum permissible moisture content is set at 50 mgH2O/std.m³ (equivalent to around -12°C dew point at line pressure of 7 MPa for sweet gas but lower than –20oC dew point for gas which is very sour). This limit is set by safety standards to control the severe corrosion problems that are associated with sour gas and is enforced by The Bergamt, a German Federal authority. For natural gas from such sour fields the concentration of H2S can be as high as 33% (mol) in extreme cases and 9 to 15% (mol) is average. Sour natural gas production facilities include de-sulphurisation plant at an early stage in the processing with sulphur production being a major part of the operation. Handling such gases prior to de-sulphurisation means careful consideration to the design of the processing plant and moisture analyser alike. The obvious corrosion difficulties are compounded by the characteristic of H2S to promote sulphide stress cracking6 in metallic materials and the potentially fatal consequences of any leakage of sour gas as H2S, which

′

cannot be detected by the human sense of smell above 200 ppm concentration, is highly toxic attacking the nervous system.

The application of a moisture analyser for sour gas measurement requires a sensing technology that can perform satisfactorily in these most aggressive of gas media. The Ceramic Moisture Sensor has been successfully used in these applications for some years. The materials selected for the active device of this sensor are exclusively ceramics and base metals that are inert by nature and offer good resilience to chemical attack. This combined with the robust design and careful selection of materials for the construction of the other gas-wetted parts of the sensor affords reliable service in this application. Experience has been gained over the last five years with two types of installation method.

The first installation type is the conventional method used for a natural gas installation, that of remote sampling with a sample conditioning system (Fig. 1) but in such sour gas applications all sample wetted components must be selected strictly in accordance with NACE6 requirements, with significant cost implications. In this case the sample flow exhausting from the system is taken to a flare where the toxic gas is rendered safe by burning. The alternative installation type is direct insertion ‘in-line’ with the Ceramic Moisture Sensor being mounted at the tip of a specially adapted, stainless steel probe assembly that inserts directly into the process pipeline. The advantages of this installation arrangement is that the sour gas remains in the pipeline, so removing the risks and costs associated with deploying a conventional sampling in such a sour gas application. The speed of response for such an installation is extremely fast but a major disadvantage is the difficulty involved in removing the sensor probe assembly from the pipeline that is required for periodic maintenance of the sensor calibration. This, as for all sour gas applications of the Ceramic Moisture Sensor, is recommended on a six monthly schedule, which, as these sensors are fully interchangeable, is achieved by exchange of the sensor in use for a freshly calibrated sensor carrying a detailed certification of calibration. A further disadvantage is the lack of protection to glycol contamination that is afforded by such direct insertion.

Some Examples of Successful Applications

Glycol Dehydration – Underground Gas Storage Facility Fiscal Metering at Gas Transfer Points – European Pipelines

Michell’s Ceramic Moisture Sensor has been implemented successfully to measure water dew point in high-pressure natural gas stored in large, natural underground storage caverns in mainland Europe. Storage pressures of up to 30 MPa mean that the gas must be very dry in order to prevent condensation occurring. Both water and hydrocarbon dew-point temperatures are measured during filling (pressurisation) and usage (de-pressurisation) cycles. The storage facility is used to meet demand in peak periods and may be non-active in the summer months. Therefore reliability is a vital factor, as the most likely time for a moisture analyser to fail is when there is no gas flow and a corrosive, moist or contaminated, stale sample is presented to the sensor for an extended period of time.

Michell’s Cermet and Cermet II IS Hygrometer systems are used for the continuous measurement of the moisture content in natural gas that is transferred from one transmission company to another, normally across country borders. In one particular case, natural gas from a Russian source is measured after glycol dehydration at its transfer point into Germany, where the dew point must be lower than 0 oC at any line pressure, up to 10MPa. Performance of the on-line Cermet II IS measurement system is validated by regular reference against a regularly-calibrated CERMAX IS portable hygrometer. At any such

′

transfer point it is vital that the gas dew point is measured, particularly if there is to be mixing of gases from different sources. Whilst it is possible to calculate the resultant dew point from a volumetric mixture of two or more sources of gas, practical verification is required for contractual conformity.

At a site in Germany, a special application has been successfully fulfilled, for measurement in sour natural gas. At this site the NH3 and H2S concentrations are as high as 33%, but the Michell Ceramic Moisture Sensor has been proven in practice to operate successfully over a number of years. Special materials of construction have been used to ensure that electrical contacts to the Moisture Sensor are not corroded by the high NH3 and H2S content, though the sensor active surface is of the standard Michell design. In the past, other analysers had a maximum operating life of only a few weeks and in some cases just hours or days. With the Michell Ceramic Moisture Sensor the calibration interval has been extended to six months.

Sour Gas Moisture Measurement LNG Production Gas exporting countries have a need to verify the quality of natural gas prior to liquefaction, for bulk transportation by sea. Similarly, the importing gas company will need to verify the quality of the LNG as-delivered. Michell has successfully implemented its Ceramic Moisture Sensor technology on many LNG plants, particularly in the Middle East region, for measurement of sub-ppm moisture levels at relatively high line pressures. This application is reasonably simple, as the LNG has been processed to remove heavy hydrocarbons, most of the moisture and is usually very low in H2S and other corrosive components.

Often forgotten, the end-point for much natural gas is in the consumer chain at low pressure. Here, there is normally a specification that requires the dew point to be lower than –26 oC. Two factors are important. The first is to ensure that the calorific value of the end-user gas meets specification. Excess moisture will effectively reduce the CV and therefore must be kept to an acceptably low level. Second, the dew point must be low enough to ensure that no condensation (and therefore subsequent pipeline freezing and possible fracture) can occur in winter conditions. In some countries where winter temperatures dictate, the specification may be even more stringent. Michell’s CERMAX IS portable hygrometer is used to rapidly determine the moisture content in low-pressure consumer gas lines to determine contractual conformance and to help trace leaks. Michell has also supplied many gas authorities with humidity calibration equipment to enable local calibration of field instruments, maintaining a traceability chain back to UK, US and other national humidity standards.

CONCLUSION

The application of moisture analysers for the measurement of natural gas is not simple and straightforward. There are many aspects to be considered that are unique to natural gas and that can greatly affect the reliability of both the instrument in service and of the measurement data that it provides. Detailed consultation between instrument manufacturer and their customers in the natural gas industry is required to device the best solution to each individual application.

Low Pressure Measurements at Consumer Points

′

REFERENCES 1. Deaton, W.M., Frost E.M., Jr., “Bureau of Mines Apparatus for Determining the Dew Point of Gases Under Pressure”. Bureau of Mines Report of Investigation 3399, May 1938. 2. Bukacek, R.F., “Equilibrium Moisture Content of Natural Gases”, Research Bulletin 8. Institute of Gas Technology, Chicago, USA, 1955. 3. ASTM designation D 1142-95, “Standard Test Method for Water Vapor Content of Gaseous Fuels by Measurement of Dew-Point Temperature”. American Society for Testing and Materials, Philadelphia, USA, 1995. 4. Oellrich, L.R., Althaus, K., “Relationship Between Water Content and Water Dew Point Keeping in Consideration the Gas Composition in the Field of Natural Gas”. GERG Technical Monograph TM 14. Fortschritt-Berichte VDI, Nr 679, 2002. 5. Robinson, J.N., Wiekert E., Moore R.G., Heidemann R.A., “Charts help estimate H2O content of sour gases”. The Oil and Gas Journal (USA), February 6, 1978, pages 77-78. 6. NACE Standard MR0175-97, “Standard Material Requirements: Sulfide Stress Cracking Resistant Metallic Materials for Oilfield Equipment”. National Association for Corrosion Engineers, Houston, USA, 1997.