Top Banner
ELEMENTAL SULPHUR FORMATION IN NATURAL GAS TRANSMISSION PIPELINES APIA Research Program Project Number 2000-09 CRC-WS Project Number 01-301 Phase 2 Investigations June 2006. David J. Pack CRC-WS
60
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Sulphur

ELEMENTAL SULPHUR FORMATION IN NATURAL

GAS TRANSMISSION PIPELINES

APIA Research Program Project Number 2000-09

CRC-WS Project Number 01-301

Phase 2 Investigations

June 2006.

David J. Pack

CRC-WS

Page 2: Sulphur

Elemental Sulphur Formation in Natural Gas Transmission Pipelines

Index

______________________________________________________________ 1.0 Introduction. Page 2 2.0 Scope. Page 3 3.0 Terminology. Page 4 4.0 The “Elemental Sulphur” Formation / Deposition Issue. Page 6 5.0 Pipeline Particle Deposition Processes. Page 11 6.0 Operation of the Sulphur Vapour Map. Page 17 7.0 Application of Physical Data. Page 21 8.0 Gas Quality Determination and Application. Page 28 9.0 Two-Phase Flow at Pipeline “T” Junctions. Page 33

9.1 The Dynamics Associated with Fluid Mixing at Pipeline “T” Junctions. Page 33

9.2 CFD Modelling Studies into a Transmission Pipeline “T” Junction. Page 39

9.3 Liquid Formation and Hold-up within a Transmission Pipeline. Page 42

9.4 Particle Deposition Studies in Natural Gas

Transmission Pipelines Page 48 10.0 Steps to Minimise the Threat. Page 50 11.0 Acknowledgements. Page 51 12.0 Recommendations for Further Research Work. Page 52 13.0 References. Page 53

Page 3: Sulphur

1.0 Introduction Within this report are the findings of the phase 2 “elemental sulphur” studies. These studies have drawn on the results of the phase 1 study, which were successfully concluded late 2004. From the phase 1 studies a number of processes have been identified that will minimize the threat and impact of “elemental sulphur” formation and deposition in natural gas transmission pipeline systems. A number of recommendations have been made from these studies for further research. Two of the main recommendations have formed the basis for the phase 2 studies.

The phase 2 studies have been split into two sections, namely:

Part (i) – The generation of a Technical Note. This technical note, which has already been submitted, has focused on the conclusions and recommendations made in through the phase 1 report.

Part (ii) – Actual phase 2 studies and report. The three main topics studied within the phase 2 studies have focused on:

1 The potential contribution of flow dynamics at pipeline “T” junctions on the observed preferential site selection of sulphur contaminated sites,

1 The impact of pressure regulation equipment on the “elemental sulphur”

formation and deposition processes, and

2 The determination and application of gas flow and quality information. Both theoretical and practical studies have been conducted. Computer fluid dynamic (CFD) studies have also been performed. These studies have been in relation to the investigations on flow behaviour at pipeline “T” junctions. Unfortunately, the full scope of the proposed phase 2, part (ii) studies could not be undertaken due to non-availability of specialized analysis equipment and administration problems associated with the advanced CFD studies. The pipeline “T” junction investigations have been based on a number of actual transmission pipeline situations. For each of the selected mainline “T” junction sites, there was a downstream pressure reduction / metering site on the lateral that was known to have continuous, or near continuous, contamination of “elemental sulphur”. The phase 2 studies conducted have demonstrated that the flow dynamics at pipeline “T” junctions have the potential to contribute to “elemental sulphur” formation and deposition at selected downstream sites. At pressure reduction facilities, the type of valve cage plays a pivotal role in the “elemental sulphur” formation / deposition process. Labyrinth type (whisper trim) valve cages are not recommended for sites likely to be conducive to the formation of “elemental sulphur”.

Page 4: Sulphur

1.0 Scope. The information contained within this report applies to high-pressure natural gas transmission pipelines, associated equipment such as compressors, and industrial gas consumer equipment such as gas turbines and gas dispensing facilities. It will also have application to pipeline upstream operations such as gas processing and gas gathering facilities and in particular downstream processing operations sourced with natural gas from medium pressure gas distribution networks. This report provides guidance to natural gas pipeline operators and others in relation to minimizing the “elemental sulphur” formation and deposition processes. It also presents recommendations with respect to the design and operation of susceptible pipeline facilities, such as pressure reduction stations, custody transfer metering facilities, and fuel gas facilities.

Page 5: Sulphur

3.0 Terminology Agglomeration A mass of differing sized particles ANSI American National Standards Institute APIA Australian Pipeline Industry Association Brownian Motion Brownian motions result from impacts by molecules of the fluid, and the suspended particles require the same mean kinetic energy as the molecules of the fluid. The suspended particles are extremely large compared to the molecules of the fluid and are being continually bombarded on all side by them. If the particles are sufficiently large and the number of molecules is sufficiently great, equal numbers of molecules strike the particles on all sides at each instant. For smaller particles and fewer molecules the number of molecules striking various sides of the particle at any instant, being merely a matter of chance, may not be equal; that is, fluctuations occur. Hence the particle at each instant suffers an unbalanced force causing it to move randomly. Coagulation The precipitation of particles from solution. Condensation The union of molecules. CRC-WS Co-operative Research Centre for Welded Structures Desublimation. By definition, desublimation is the passage of substance from the gaseous state directly to the solid state. Electrophile. A reagent that takes an electron pair is called an electrophile and the reaction is electrophilic. Elemental Sulphur Elemental sulphur occurs naturally as an ore. It is generally recognised by a characteristic yellow colour or by its generation of sulphur dioxide when burned in air.

Page 6: Sulphur

Elemental sulphur can also occur in colours other than yellow. The term ‘elemental sulphur’ is used in the text to describe the samples found within the pipeline systems. Entrainment Fraction Entrainment fraction is defined as the ratio of the liquid flow rate in the form of droplets to the total liquid flow rate. Entrainment Rate Entrainment rate is defined as the mass of liquid in the form of droplets detached from a liquid film per unit area per unit time. EOS Equation of State ESEM Environmental Scanning Electron Microscope GC-MS Gas Chromatograph – Mass Spectrometer HYSYS A process engineering software package developed by Hyprotech. It provides an integrated steady state and dynamic simulation capability. ICP-MS Inductively Coupled Plasma – Mass Spectrometer Liquid Loading Liquid loading is defined as the liquid volumetric flow rate divided by the gas volumetric flow rate at in-situ conditions. Nucleation The process of the formation, within an unstable supersaturated solution, of the first particles of precipitate of spontaneous growth into large crystals of a more stable solid phase. Nuclei Minute particles formed spontaneously. The nuclei are the centre or ‘seed’ from which crystals begin to grow during solidification. Nucleophile A reagent that brings an electron pair is called a nucleophile and the reaction is nucleophilic. RESS Rapid Expansion of Supercritical Solutions. Sulphur Vapour Map Concept developed during Phase 1 research that relates the sulphur vapour concentration (in ppm) in suspension within the gas composition to flowing gas pressure and temperature conditions.

Page 7: Sulphur

Superficial Gas Velocity. Superficial gas velocity is defined as the volume flow of gas divided by the cross sectional area of the pipe. Vena Contracta The vena contracta is the portion of a flow stream where fluid velocity is at its maximum, and fluid static pressure and the cross sectional area at their minimum. In a control valve, the vena contracta normally occurs just downstream of the actual restriction. 4.0 The “Elemental Sulphur” Formation / Deposition Issue The information contained within this section of the report is essentially similar to the information presented in the Phase 1 Technical report. It is reproduced, in an updated format, to demonstrate the complexity of the “elemental sulphur” formation and deposition processes and provide background to the research work performed up to date. The formation of ‘elemental sulphur’ within natural gas transmission pipeline systems has demonstrated to be a process that has the potential to severely impact the continuity, and hence reliability, of high-pressure natural gas supplies. This report builds on the information presented within the Phase 1 Report and Phase 2 Technical Note. Collectively, the reports review the causes for, and provides an understanding of this growing phenomenon, which is now impacting, in varying degrees, the operation of a significant number of natural gas transmission pipeline systems within Australia as well as overseas. Since the publication of the Phase 2 Technical Report, considerable work has been undertaken with a number of overseas pipeline operators. The concept of the “sulphur vapour map”, as discussed in previous reports, has been proven. The details of the operation of the “sulphur vapour map” is represented and discussed in a following section of this report. Although the report references “elemental sulphur” as the cause of the pipeline deposition processes, this element is but a small part of the overall make-up of the extensively analysed contamination deposits. The impurities found within a natural gas transmission pipeline system can be many and varied. These impurities can be in gaseous, liquid or solid state. Under properly maintained pipeline operating conditions such impurities will generally be present in trace quantities. The more common impurities found within transmission pipeline natural gas compositions are hydrogen sulphide, water vapour and carbonyl sulphide. Aromatics, such as benzene, and mercaptans may also be present in trace quantities. If the pipeline is internally uncoated, then there is a reasonable probability that iron oxide deposits will be found. Although reasonably rigid quality specifications apply to the natural gas composition transported in transmission pipeline systems, adverse processes that lead to the formation of unwanted

Page 8: Sulphur

contaminants can, and do at times occur within such pipelines. One such process is, as already referenced, the formation of elemental sulphur. This elemental sulphur formation can be through chemical reactions or by a desublimation process, that is, the sulphur present in the gaseous state is being converted to solid state by a particular mechanism, or series of mechanisms within the pipeline system. The desublimation process is shown through this research work to be the dominant elemental sulphur formation process within natural gas pipelines. The recorded observations of elemental sulphur within natural gas transmission networks are increasing. This is probably due to two key factors, which are:

(a). An increase in the operating pressure of transmission pipelines, and (b). Greater awareness of the elemental sulphur problem by pipeline operators. Prior to the commencement of the APIA / CRC-WS sponsored research work into transmission pipeline elemental sulphur formation, limited theoretical studies had been made based on basic information and a small number of on-line tests and analyses. While a number of theories have been developed from these prior studies, further data and more in-depth analyses needed to be performed to ensure:

- That sulphur deposition occurrences can be quantified, and reliably related to system operating parameters (flow, temperature, pressure, composition) and

- That appropriate theories and models can be developed to account for the

observations, and hence provide guidance for developing minimisation and mitigation strategies.

Interestingly, at the commencement of these studies, the “elemental sulphur” formation problem was thought to only exist in a couple of Australian pipelines, together with a small number of overseas facilities. Through the ability to present technical papers on this research work during the life of the project, awareness has been generated that this problem is internationally widespread within natural gas transmission networks. ‘Elemental sulphur’ formation and deposition processes have also been reported to have adversely impacted the proper operation of related facilities such as gas processing plants, gas turbines used for power generation, and natural gas flow- testing facilities. This knowledge now suggests that the better understanding of the “elemental sulphur” formation process is of interest and value, not only to pipeline owners and operators, but also to gas producers, gas processing plant operators and large industrial gas consumers. The formation and presence of the so-called ‘elemental sulphur’ (orthorhombic sulphur) deposits in natural gas streams can have serious consequences for gas production, processing, transportation and end-user operations. Within recent years the formation of the ‘elemental sulphur’ deposits within high pressure natural gas transmission pipelines has become quite wide spread and is creating significant operating and maintenance problems for pipeline operators. Indeed, some pipelines may have this problem without realising it, or it may be disguised as the commonly referred to ‘black powder’ problem. This research has shown that the deposition processes of ‘elemental sulphur’ and more

Page 9: Sulphur

common ‘black-powder’ have many similarities. However, there are still a number of very unique and complex features associated with the ‘sulphur deposition’. These features are discussed in the following sections of this report and elsewhere [1]. Sulphur is a very complex element and can have many different forms depending upon pressure and temperature conditions. Sulphur vapour is also soluble, to varying degrees, in a number of the common natural gas components. The clogging of well-bore tubing and underground natural gas reservoirs by elemental sulphur, especially with sour-gas compositions, is well documented.

A large pressure reduction, and hence consequent temperature quenching, within a flowing natural gas stream containing sulphur vapour in solution, provides the mechanism for the sulphur vapour to become supersaturated, and is hence conducive for the sulphur desublimation process. This situation occurs commonly within high-pressure natural gas transmission pipeline systems. The transition of the sulphur vapour to solid state (commonly referred to as S8) occurs because at normal pipeline operating conditions the partial pressure of the sulphur vapour is well below the triple points (sulphur has four triple points unlike most elements which have one). The sulphur particles are formed by nucleation; therefore the presence of other particles and liquid droplets in the gas stream will assist with this desublimation process. Figure 4.1 demonstrates ‘elemental sulphur’ deposition on a control valve cage. As the deposited material is distinctively yellow this indicates absence of significant co-deposited hydrocarbons. Also the location of the deposits is at a point of high gas velocity, due to what appears to be a fine consistent ‘powder’ deposit.

Figure 4.1. Whisper trim cage with “elemental sulphur” deposits.

Page 10: Sulphur

The trend with new natural gas pipelines is to have them operating at higher pressures (ANSI class 900 now not uncommon), therefore it is anticipated that the occurrence and magnitude of this ‘sulphur deposition’ process will increase. Although elemental sulphur is referenced as the deposition element, there are clearly many other elements and compounds involved. Extensive studies have been undertaken into the potential chemical reactions within natural gas transmission pipeline systems with respect to the formation of sulphur and its related compounds. The results of this research work, as discussed in detail in the following sections, demonstrate that the ‘elemental sulphur’ formation / deposition process is very complex. It will be shown that the majority of the deposited material results from a desublimation process of the supersaturated sulphur vapour due to the rapid cooling of the flowing gas stream through a pressure regulator, or similar pressure reduction / control device. It must be emphasised that other contaminants already in the gas stream, together with potential chemical reactions upstream of the pressure reduction stage can also contribute to the observed particle formation / deposition processes. The following points provide a simplified overview of the ‘elemental sulphur’ formation and deposition process for a transmission pipeline pressure reduction facility that would have the necessary gas composition and operating conditions.

1. Sulphur vapour already in the gas stream in sub ppm levels. 2. The sulphur vapour becomes supersaturated due to the swift cooling of the

gas mixture rapidly expanding through the pressure control valve cage mechanism, nozzle or like pressure restriction/control device.

3. The supersaturated sulphur vapour molecules form nuclei, which are minute

particles. This very rapid conversion of the supersaturated vapour to minute solid particles is the nucleation process.

4. Concurrent possibility of retrograde condensation occurring for some of the

heavier hydrocarbon components in the gas stream. This is also due to the rapid cooling of the gas stream.

5. Other molecules (retrograde condensation components) are attracted to the

sulphur particle surface through the mechanism of condensation.

6. The resulting larger particles, which will have a very high velocity, will collide with other particles in the gas stream forming larger particles. This is the coagulation process.

7. There may be other deposits on the internal pipe-walls or fittings, or traveling

within the gas stream. Due to the high gas velocities and turbulence, there will be a high probability of collision between these particles. The growth in particle size due to the collision processes is the commencement of what is termed the agglomeration phase.

Page 11: Sulphur

Therefore, the mechanism is particle formation through nucleation and condensation and particle growth through further condensation and coagulation. The nucleation process will ascertain the particle numbers, with condensation determining the mass of the particle. Coagulation will, on the other hand, decrease the number of particles through combination. Agglomeration will be the overall final mass formation process. Other components in the gas stream can also be in minute quantities, yet significantly impact on the proper operation of a pipeline. As for the more common “black-dust” problem, the “elemental sulphur” deposition mechanism also has demonstrated that there can be preferential site conditions. That is, at certain locations along a given transmission pipeline system, a metering / pressure regulation site ‘A’ may be known to be a location for “elemental sulphur” deposition, yet a similar facility ‘B’ on the same pipeline receiving the same natural gas at near identical flowing pressure and temperature conditions will not be impacted, or be minimally impacted, by the elemental sulphur deposition phenomena. Studies undertaken during the course of this research project into these observed situations have identified two possible explanations. This selective site situation is known to impact several Australian natural gas transmission pipelines as well as a number of overseas. Studies undertaken so far has been theoretical, however modeling performed has been based on actual operating data obtained for a couple of the impacted pipeline systems. The results obtained to date are encouraging, however further investigations are required. The two identified explanations, which are discussed in following sections of this report, both of which are considered plausible for the field case situations are:

3 The dynamics associated with two-phase flow at pipeline “T” junctions, and 4 The generation of pipeline section liquid hold-up.

The following example illustrates what appears to be an insignificant quantity of an unwanted contamination component within a transmission pipeline system can, over time, grow into an appreciable quantity which can directly impact the proper operation of a pipeline. A pipeline has hydrogen sulphide at 1 part per million (approx 1.43 mg/m3). Natural gas flow is 210 TJ/day with the gas calorific value being 40 MJ/m3. If all the hydrogen sulphide (H2S) is converted to iron sulphide (FeS) then approximately 6,600 kg per annum of FeS will be produced in this pipeline.

Page 12: Sulphur

5.0 Pipeline Particle Deposition Processes. As defined through the Phase 1 research, elemental sulphur particle formation within natural gas transmission pipeline system is a result of a desublimation process. The Phase 2 studies have taken these findings further and investigated the potential particle deposition mechanisms with the aim to see if there are any practical control mechanisms that can be applied. Due to the rapid pressure drop across the control valve cage, or nozzle, as applicable, the velocity of the natural gas increases rapidly. This means that a large amount of heat is transformed into the kinetic flow energy, resulting in the temperature of the gas being lowered rapidly. As the temperature is lowered, a critical supersaturation is reached at some point at which nucleation begins. This nucleation process will occur suddenly. It is believed that the deposition (desublimation) process for the observed particle matter is through the action of simultaneous nucleation, coagulation and/or condensation – with similarities to the gas to particle formation process referenced by Wu et al [2]. These researcher studied condensational growths of monodisperse aerosols in a system with limited vapour, concluding that nucleation and condensation are the mechanisms for gas to particle formation, and condensation and coagulation the particle growth mechanisms. Nucleation can be classified as being either homogeneous or heterogeneous. The investigations undertaken indicate that both classifications can occur within gas transmission pipeline systems. Homogeneous nucleation is the nucleation of vapour in the absence of foreign substances whereas heterogeneous nucleation is the nucleation

Page 13: Sulphur

on a foreign substance, surface, or a solid particle. In addition, nucleation processes can be homomolecular (involving a single species) or heteromolecular (involving two or more species); again both processes are believed to occur within the pipeline systems. A definition [3] of nucleation is given as the formation within an unstable, supersaturated solution of the first particles of precipitate capable of spontaneous growth into large crystals of a more stable solid phase. These first viable particles called nuclei, may either be formed from solid particles already present in the system (heterogeneous nucleation), or be generated spontaneously by the super saturated solution itself (homogeneous nucleation). As demonstrated in the Phase 1 results, hydrocarbon liquid droplets entrained within the gas stream appear to play a major role in the development of the “elemental sulphur’ deposits. A number of researchers have investigated the occurrence and magnitude of liquid condensate in natural gas pipelines. Adewumi et al [4], one such researcher, states that liquid condensate commonly occurs because of the multi-component nature of the transmitted natural gas and its associated phase behaviour, as well as the inevitable temperature and pressure variations that occur along the pipeline. Although it is not suggested that hydrocarbon liquids are always present in a given transmission pipeline system, the analysis of the “elemental sulphur” samples does clearly show that small quantities of liquid (probably as small droplets entrained in the gas stream) are present at pressure reduction stations that are delivering gas in accordance with contractual specifications. The gas quality calculation results given in Section 8.0 of this report demonstrate how hydrocarbon liquids can be present within a pipeline operating at normal conditions. As already discussed, when a vapour / gas mixture is cooled, there is the potential for it to become supersaturated and condensation can occur on any surfaces within the pipeline. This can also include any foreign particles. Clearly it may not be possible to reduce condensation occurring at a pipeline facility. However, as the coagulation rate is proportional to the square of the number of particles present, there is clearly a threshold, below which the number of particles nucleated and/or already present within the gas stream is not sufficient for significant coagulation to occur. Obviously controlling the number of particles present will control the formation of the “elemental sulphur” deposits – unfortunately a considerable amount of further research needs to be performed to determine the particle number threshold levels. Considerable work has been performed on the determination of particle size and numbers formed principally by the rapid expansion of supercritical solutions (RESS) process. This process, under strictly controlled conditions, uses a critical flow nozzle and saturated solution to generate fine particles. Freidlander et al [5], for example, has shown that particle size can be controlled by varying the velocity, temperature and aerosol precursor concentration at the nozzle, and the temperature of the ambient gas. It was also determined that when particles are formed at very high concentration near the nozzle exit, the particles may collide and coalesce; the total number of particles continues to decrease with distance from the nozzle. It was found that the average particle size increases according to the following equation:

Page 14: Sulphur

52

0

03

1

656

⎥⎦

⎤⎢⎣

⎡+⎟

⎠⎞

⎜⎝⎛= a

cx

UV

Ad p π …… (a)

21

61

6434.0 ⎟

⎟⎠

⎞⎜⎜⎝

⎛⎟⎠⎞

⎜⎝⎛=

p

kTAρπ

where pd = volume average particle diameter

0V = volumetric concentration of aerosol material at the nozzle exit

0U = gas velocity at the nozzle exit k = Boltzmann’s constant T = absolute temperature

pρ = particle density x = axial distance from the nozzle exit c = ratio of spread of matter to momentum ≈ 1.4

a = constant which can be neglected at long distances from the nozzle exit

Equation (a) holds for the fully developed region of the turbulent jet, that is, dx

>> 10 where d is the diameter of the nozzle. The total number of particles produced in the turbulent jet per unit time in the coagulation limit at a distance x from the nozzle exit is given by:

56

0

0

0

165

⎥⎦

⎤⎢⎣

⎡+

=

axcU

VA

VNT

The value of TN also represents the maximum number of particles that can be generated. When the concentration of condensable vapour is low, and/or very high saturation ratios are not attained in the jet, particle formation by homogeneous nucleation leads to relatively low number concentrations. In this mode, particle formation is nucleation limited. Particles do not collide, which may be advantageous in avoiding the formation of “hard” agglomerates of solid particles. According to classical nucleation theory however, the rate of particle formation depends

primarily on the group ,log23 ST not on S alone, where S is the saturation ratio in the jet and T the corresponding temperature.

Page 15: Sulphur

Although there are strong parallels with the RESS process and pressure reduction stages on pipelines, it is not possible to apply the particle concentration / size equations to the pipeline situation as:

5 the concentration of the nucleated material (sulphur) cannot be controlled, and indeed is unknown,

6 the gas flow, pressure and temperature conditions cannot be precisely controlled,

7 the presence of other particle matter within the gas stream is unknown,

8 the rate and/or concentration of liquid hydrocarbons is unknown, and

9 the geometry of the pressure reduction stage is unlikely to be symmetrical.

Therefore, particle growth and deposition processes have been turned to as a means to understand and hence help control the formation of the “elemental sulphur” deposits. According to Helfgen et al [6] potential pipeline particle growth / deposition mechanisms can be classified into the following:

1 Thermophoresis 2 Turbulent eddy impaction, 3 Brownian diffusion, 4 Turbulent eddy diffusion, 5 Electrostatic precipitation. 6 Gravitational settling

Details of the above referenced particle deposition mechanisms are now briefly discussed.

(i) Thermophoresis. This is the condition in which a temperature gradient in a gas cause any suspended particles to migrate in the direction of decreasing temperature. This will be most effective where there is a very pronounced temperature gradient, however the surface of any object within the pipe approaches the temperature of the flowing gas, so the thermophoretic force will decline. The thermophoretic force is a function of the Knudsen number. Due to the high Reynolds number at the point of pressure reduction, it is considered unlikely that thermophoresis will be very effective. However, for the situation where there is a labyrinth type control valve cage is in use and already has some particle deposition on the solid surface of the cage, thermophoresis may be applicable.

(ii) Turbulent eddy impaction. Turbulent eddy impaction is believed to have

the potential to considerably contribute to the growth of particles through impaction, especially for particles that have already obtained a reasonable size and for particles that are within the turbulent environment of the pressure reduction stage.

(iii) Brownian diffusion. This applies to very small particles, generally less

than 0.1μm in diameter. Believed to be a contributing factor to the

Page 16: Sulphur

observed deposition on and around pressure control valves.

(iv) Turbulent eddy diffusion. This is the turbulent diffusion of heat, mass or momentum caused by the random eddy motions along the axis or radius of the pipe. Believed not to have any major influence on the observed particle growth / deposition processes.

Figure 5.1. As for Figure 4.1 – Close-up view.

(v) Electrostatic precipitation. This process is the removal of suspended particles in a gas stream through the application of an electric charge and the subsequent attraction to a charged surface. Considered not applicable to pipeline situations.

(vi) Gravitational settling. This process will most likely be more dominant

downstream of the pressure reduction stage where the velocity of the gas is reduced and particle coagulation has occurred. It would only have a minor influence at the pressure reduction stage.

Of the six referenced pipeline particle growth / deposition mechanisms, only thermophoresis, turbulent eddy impaction and Brownian diffusion are regarded as having the ability to contribute to the overall “elemental sulphur” particle deposition process. Each mechanism would most likely contribute to the deposition process under differing conditions. As the influence of Brownian diffusion is limited to very small particles, it would be expected that this mechanism would operate during the initial particle nucleation phase. As the particles go through the resulting condensation and coagulation phases and therefore grow in size, the influence of Brownian diffusion would diminish. The only mechanism to control Brownian diffusion is to minimize the number of minute particles forming through nucleation. Obviously for a situation that presents a significant solid surface area to the gas stream, such as a labyrinth type control valve cage, particle

Page 17: Sulphur

‘capture’ on the solid surface (cage walls and orifices) will be high. Therefore, in a pressure regulation situation where there is a high probability of nucleation occurring it is advisable not to use a labyrinth type control valve cage. A quick acting or equal percentage type cage would be preferable with respect to the minimization of particle deposition. Turbulent eddy diffusion has the potential to be the dominant mechanism for the formation of the “elemental sulphur” deposits. Due to the complexity and number of differing elements that may contribute to the particle formation process, each formed particle (post nucleation) will most likely have a unique mass. Therefore, even within a hypothetical two-dimensional stream flow there would still be coagulation occurring due to the differing momentum of each particle. As for Brownian diffusion, the only control mechanism is to minimize the number of particles within the gas stream. Thermophoresis may be a contributing factor where there is a pronounced temperature gradient between the gas stream and a solid surface. Again a labyrinth type cage would provide favourable conditions for particle deposition through the mechanism of thermophoresis where the temperature of the solid surface is considerably cooler than the actual temperature of the flowing gas. Control of the thermophoresis mechanism is through heating of the solid surface. The mechanisms for the deposition of particles onto solid surfaces from a turbulent gas stream are complex and must still be regarded as inconclusive. However considerable experimentation has been conducted into this phenomenon. The experimental results of Romay et al [7] have indicated that there is a complex coupling of turbulent and thermophoretic deposition mechanisms for the deposition of particles onto a solid surface from a gas stream. The results also demonstrated that turbulent deposition dominated over thermophoresis for larger particle sizes. Lin et al [8] found that for all particle sizes, the particle deposition efficiency decreased with increasing pipe wall temperature and gas flow rate. It was also noted that particle deposition was suppressed completely when the pipe wall was heated slightly above the flowing gas temperature. For natural gas transmission pipeline operating conditions with high Reynolds numbers, Owen et al [9] found evidence that turbulent diffusion will assist particles in the gas stream overcome the thermophoretic force. Although there could be some influence from the thermophoretic force, it was found to be less effective at high Reynolds numbers.

Page 18: Sulphur

Figure 5.2 View of Labyrinth type Valve Cage with “Sulphur” Deposition.

6.0 Operation of the Sulphur Vapour Map.

Page 19: Sulphur

The purpose of this Section is to demonstrate the application of the developed ‘Sulphur Vapour Map’. An important assumption in the application of this map is that the natural gas supply is saturated with sulphur vapour. With reference to Figure 6.1, let us assume that the sulphur vapour concentration in the natural gas is 0.1 ppmv. The pipeline operating conditions immediately upstream of the pressure control valve is 8,000 kPa @ 320.2 K (point A). The contractual supply pressure is 2,000 kPa. From HYSYS, the gas temperature immediately downstream of the control valve will be about 290.7 K (point B).

0

2000

4000

6000

8000

10000

12000

14000

16000

273 283 293 303 313 323 333 343Temperature (K)

Pres

sure

(kPa

.)

1.0E-03 ppmv

1.0E-02 ppmv

1.0E-01 ppmv

3.0E-02 ppmv

5.0E-02 ppmv

OOA

OB

OC D

OE

Point A: 8,000 kPa @ 320.2 KPoint B: 2,000 kPa @ 290.7 KPoint C: 6,600 kPa @ 314.4 KPoint D: 6,600 kPa @ 334.7 KPoint E: 2,000 kPa @ 323.3 K

Figure 6.1: Elemental Sulphur Equilibrium Map: (273 - 343 K)

Concentration of S8 for Given Natural Gas Composition.

That is, the operating conditions go from point A directly to point B. Now according to the ‘sulphur vapour map’, sulphur deposition can be expected, with the degree of deposition being 0.1 – 0.001 ppmv = 0.099 ppmv. To help reinforce the minute levels of elemental sulphur formed, if we take this elemental sulphur deposition value of 0.099 ppmv, expressing the level of 0.099 as 1.0 m3, the actual ratio would be 1.0 m3 of elemental sulphur to 10,101,000 m3, at standard conditions, of all remaining components in the gas stream. If a two-stage pressure reduction is introduced then two possible operating examples can be given, as provided below. The ‘sulphur vapour equilibrium map’ values are in good agreement with the results derived by Wilkes et al [10]. Example 1: Having the ability to heat the gas between pressure reduction stages.

Assume pipeline operating conditions immediately upstream of the pressure reduction facility is at 8,000 kPa abs and 320.2 K (80 bar abs and 47 OC). The

Page 20: Sulphur

referenced inlet conditions are point A on Figure 6.1. The conditions of 8,000 kPa abs and 320.2 K correspond to a sulphur vapour saturation level of 1.0E-01 ppmv (1 part in 10-7)

Now the required outlet (supply) conditions at this facility are 2,000 kPa abs (20 bar abs.). This condition corresponds to a sulphur vapour saturation level of 1.0E-03 ppmv (1 part in 10-9). Now if a single pressure cut were to be made at these conditions then the amount of sulphur vapour desublimating out of the gas stream would be:

= 1.0E-01 – 1.0E-03 ppmv = 0.99E-01 ppmv (the majority of the sulphur vapour)

This situation corresponds to point B on Figure 6.1.

If the gas is reduced in pressure, through the first pressure reduction stage, to 6,600 kPa abs @ 314.4 K (66 bar abs @ 41.3 OC), which corresponds to a sulphur vapour saturation level of 5.0E-02 ppmv (point C on Figure 6.1), then the amount of sulphur vapour desublimating out of the gas stream at the first pressure cut would be:

= 1.0E-01 – 5.0E-02 ppmv = 0.5E-01 ppmv (half of the sulphur vapour)

Now the gas supply is re-heated to 334.7 K (61.6 OC), which corresponds to point D on Figure 6.1. The gas is then directed to the second pressure reduction stage (point E on Figure 6.1), which supplies the gas at 2,000 kPa abs @ 323.3 K (20 bar abs @ 50.2 OC)

The amount of sulphur vapour desublimating out of the gas stream at the second pressure cut would be:

= 5.0E-02 – 3.0E-02 ppmv = 2.0E-02 ppmv.

This means for this given pressure reduction situation the amount of sulphur vapour desublimating out of the gas stream would be:

= 0.5E-01 ppmv (first stage) + 2.0E-02 ppmv (second stage) = 0.52E-01 ppmv (total)

This compares with 0.99E-01 ppmv had there been a single pressure cut.

Therefore, the amount of “elemental sulphur” forming at the upstream pressure reduction valve has been approximately halved for the case study.

Example 2: Not having the ability to heat the gas between pressure reduction stages.

For this situation the first pressure reduction stage would provide the same results as for example 1, that is the pressure would be reduced to 6,600 kPa abs @ 314.4 K (66 bar abs @ 41.3 OC), which corresponds to a sulphur vapour

Page 21: Sulphur

saturation level of 5.0E-02 ppmv (point C on Figure 6.1). However, the

second stage would have inlet conditions of 6,600 kPa abs @ 314.4 K and outlet conditions of 2,000 kPa abs @ 290.7 K (20 bar abs @ 17.6 OC), which corresponds to a sulphur vapour saturation level of 1.0E-03 ppmv (point B on Figure 6.1).

This means the amount of sulphur vapour desublimating out of the gas stream at the second pressure cut would be:

= 5.0E-02 – 1.0E-03 ppmv = 4.9E-02 ppmv. = 0.49E-01 ppmv (nearly equal to the first stage desublimation rate)

As can be seen from this example, by applying a dual pressure reduction stage, the amount of sulphur vapour desublimation out of the gas stream can be proportioned between the stages.

Note: Any heat recovery / heat loss between the pressure reduction stages has not been accounted for within the above worked examples. Also, the greater the distance between the pressure reduction stages the more effective will be the dual pressure reduction configuration in reducing the “elemental sulphur” formation.

The given examples apply for a particular natural gas composition, having the following properties:

1 Molecular mass of approximately 19.45 kg/kmol 2 Density at MSC of 0.825 kg/Sm3 3 Real relative density of 0.673 4 Gas higher heating value of 40.805 MJ/Sm3.

Page 22: Sulphur

0

1000

2000

3000

4000

5000

6000

7000

273 278 283 288 293 298 303 308 313

Temperature (K)

Pres

sure

(kPa

.)

O

O

1.0E-02 ppmv

Point A: 5,600 kPa @ 302.6 KPoint B: 3,000 kPa @ 288.5 K

A

B

1.0E-03 ppmv

Figure 6.2: Simplified Elemental Sulphur Equilibrium Map –

Concentration of S8 for Given Natural Gas Composition.

Figure 6.2 represents the identical gas composition to that applied to the Figure 6.1 case study. However, for the Figure 6.2 case study the selected pipeline operating condition is 5,600 kPa at 302.6 K. These flowing gas conditions happen to be on the 0.01 ppmv sulphur vapour equilibrium concentration line (conditions represented by point A). Now if the required flowing gas pressure cut is to be 3,000 kPa, the flowing gas temperature will drop to 288.5 K according to the HYSYS calculation. Conveniently, this new operating condition lies on the 0.001 ppmv sulphur vapour equilibrium concentration line (conditions represented by point B). Therefore in accordance with the sulphur vapour equilibrium map, the amount of elemental sulphur (assuming saturation conditions) that will desublimate out of the natural gas vapour will be 0.009 ppmv. This amount of elemental sulphur appears to be insignificant, however if it is assumed that this gas flow rate is the supply to a gas turbine and is being metered by a turbine meter, then over a 12 month period of a continuous natural gas flow rate a reasonable S8 deposition level could be expected. Now if a gas mass flow rate of 10,000 kg/h (240 Tonnes/day) is applied, which would approximate to 11.9 TJ/day based on the gas composition given in Table 6.1, the amount of elemental sulphur that would form over one year would be approximately 1.3 kg or 0.000628 m3. This volume would represent a cube with sides of 8.56 cm. The concept of the “sulphur vapour map” has been successfully applied to an overseas transmission pipeline pressure reduction station that continually experienced “elemental sulphur” deposition problems. The particular site required the cleaning of the pressure regulator cages on a fortnightly basis. However, on application of the pressure settings,

Page 23: Sulphur

per the “sulphur vapour map” concept to the two-stage regulation set-up there was no noted deposition on the cages after two months of continuous operation. It is probable that some “elemental sulphur” could be occurring, however it would seem from the result that the amount forming would be below the threshold level.

Component Mole percent Methane 83.155 Ethane 7.05 Propane 3.514 n-Butane 0.554 i-Butane 0.412 n-Pentane 0.048 i-Pentane 0.074 n-Hexane 0.012 n-Heptane 0.022 Nitrogen 3.106 Carbon Dioxide 2.075

Total 100.000

Table 6.1 Gas Composition Applied to Provided Examples. 7.0 Application of Physical Data. This section critically reviews the application of physical data, and in particular the Joule-Thomson effect, to pipeline pressure reduction stages. As immediately downstream of pipeline pressure reduction stages is the most common location for the “elemental sulphur” deposits, the understanding of the kinetics associated with the pressure reduction has been a very important feature of this research. It is common practice within the natural gas industry to apply a ‘rule-of-thumb’ temperature / pressure correlation at a pressure reduction stage of 5 OC drop in temperature for every 1,000 kPa drop in pressure. Although this value may be valid for some gas compositions and a small range of operating pressure and temperature conditions, for the majority of applications it really is an approximation. It is also generally assumed that the pressure / temperature reduction is linear, with the minimum pressure and temperature conditions as recorded by installed instrumentation or derived by calculation. The true minimum pressure will be at the vena-contracta, with the observed and applied minimum pressure being the actual or approximate recovered pressure value. The following calculations and examples demonstrate how the true minimum pressure and temperature conditions may vary considerably from the indicated values, and why hydrocarbon liquid formation may be more prevalent at pressure reduction stages than is normally anticipated.

Page 24: Sulphur

It is well known that when a natural gas stream expands through a restriction such as a pressure control valve or nozzle, the gas stream goes from high pressure to low pressure and there is also a corresponding change in the flowing gas temperature. This process occurs under conditions of constant enthalpy and is known as the Joule-Thomson, or JT, expansion. This process can also be considered as an adiabatic effect (heat not added or subtracted) because the pressure change occurs too quickly for significant heat transfer to take place. The temperature change is related to pressure change and is proportional to the JT coefficient. Of interest are the determination of the temperature change, and particularly the rate of change of temperature (temperature quench conditions). To obtain the temperature

change ( )TΔ , the Joule-Thomson (JT) coefficient gas to be determined. The JT

coefficient is represented by the term JTμ The following calculations demonstrate how the Joule-Thomson coefficient is derived – the important issue is that there are a number of assumptions that are made in this derivation which may not be truly valid for a given pipeline operational conditions. Let us assume a pressure reduction facility associated with a high-pressure natural gas pipeline. Let the upstream pressure of a parcel of gas be assigned the pressure

value 1p , temperature value 1T , and volume value 1V . After passing through the

restriction the pressure will have a new pressure value of 2p and volume of 2V . Now if the kinetic and potential energy of the gas going from state 1 to state 2 is negligible, the first law of thermodynamics, which states that the net amount of energy added to any system must equal the net change in energy within that system, can be approximated by equation (i):

WQUU +=− 12 ……. (i)

where 2U = internal energy of the system at state 2

1U = internal energy of the system at state 1

Q = heat added to the system, and W = work done by the system. If we assume that the system is operating in a steady state flow condition and that heat

added is negligible, that is ≈Q 0, then the work done by the system can be expressed as:

2211 VpVpW −= ……. (ii) Substituting equation (ii) into (i):

Page 25: Sulphur

221112 VpVpUU −=−

222111 VpUVpU +=+ ……. (iii) Now enthalpy is defined by the equation:

pVUH += ……. (iv) Let equation (iv) represents 1 mole of a homogeneous gas of constant composition Therefore equation (iii) can be expressed as:

21 HH = ……. (v)

where 1H = enthalpy at state 1

2H = enthalpy at state 2 For a condition where equation (v) is valid, means that the particular gas flow parcel can be considered as a constant enthalpy or isenthalpic process. A further term is now considered, that of entropy. Entropy, which is given the symbol “ S ”, is fundamental to the second law of thermodynamics and can be considered as a measure of the usefulness of the energy content of the system. The change in entropy of a system is defined as:

TdQdS =

……. (vi) therefore

dQTdS = ……. (vii) differentiate equation (iv):

VdppdVdUdH ++= ……. (viii)

now dWdUdQ += therefore:

dWdQdU −= where pdVdW = ……. (ix) [additional terms may apply to the right hand side of the equation; however they will be small and are neglected] substitute equations (vii) and (ix) into (viii)

VdppdVdWdQdH ++−=

Page 26: Sulphur

VdppdVpdVTdSdH ++−=

VdpTdSdH += ……. (x) Equation (x) provides a relationship between enthalpy, entropy and internal energy with pressure, temperature and volume.

By dividing equation (x) by dp at constant T , the following equation is derived:

V

pST

pH

TT

+⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

……. (xi) Applying the Maxwell equations (reciprocity relationship for exact differentials)

pT TV

pS

⎟⎠⎞

⎜⎝⎛∂∂

−=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

equation (xi) can be transformed into:

V

TVT

pH

pT

+⎟⎠⎞

⎜⎝⎛∂∂

−=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

……. (xii) By taking the enthalpy as a function of temperature and pressure, obtain:

dp

phdT

Thdh

Tp⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

+⎟⎠⎞

⎜⎝⎛∂∂

= ……. (xiii)

now pTH⎟⎠⎞

⎜⎝⎛∂∂

is the heat capacity at constant pressure pc

Substitute for pc and equation (xii) into equation (xiii):

dp

TVTVdTcdH

pp

⎥⎥⎦

⎢⎢⎣

⎡⎟⎠⎞

⎜⎝⎛∂∂

−+= ……. (xiv)

By dividing equation (xiv) by dp with h constant, equation (xv) is obtained:

Page 27: Sulphur

⎥⎥⎦

⎢⎢⎣

⎡−⎟

⎠⎞

⎜⎝⎛∂∂

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂ V

TVT

cpT

pph

1

pph TV

TcT

pT

⎟⎠⎞

⎜⎝⎛

∂∂

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂ 2

ppH TZ

pcRT

pT

⎟⎠⎞

⎜⎝⎛∂∂

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂ 2

……. (xv)

Now HJT p

T⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

=μpp T

Zpc

RT⎟⎠⎞

⎜⎝⎛∂∂

=2

……. (xvi)

where JTμ = is the JT coefficient R = Molar gas constant T = Temperature (K)

pc = Specific heat capacity (J/kg K) p = Pressure (MPa) Z = Compression factor For the determination of the JT coefficient equation (xvi) is applied, however it will be noted that in the derivation of this equation there are a number of assumptions made. Although these assumptions may be reasonably valid, they may not necessarily be precise, accurate and unbiased. If we now extend this equation to actual natural gas compositions and pipeline p and T conditions, the following results are obtained. Inlet pressure: 6,500 kPa @ 30.0 OC Outlet pressure: 2,000 kPa @ 4.75 OC (determined from Hysys 3.2) Figure 7.1 provides a range of JT coefficients, which have been derived through the application of equation (xv), for a series of flowing conditions. Two differing gas compositions have been applied. The standard gas composition is that given in Table 7.1 with the high CO2 gas composition being the composition given in Table 7.1 modified with an additional 10 mol % of CO2 at the expense of the methane content. It will be noted that there is a reasonable shift in the JT coefficient, for the same operating conditions, between the standard and high CO2 content natural gas. If it is assumed that the designed maximum gas velocity at the valve is Mach 0.2 (IEC established guideline is to limit the gas velocity at the valve to Mach 0.3 or less), then taking the gas composition per Table 7.1, the following speed of sound values, as given by Table 7.2, are determined for the given pressure and temperature conditions.

Page 28: Sulphur

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

270 275 280 285 290 295 300 305

Temperature (K).

JT c

oeffi

cien

t (K

/MPa

).

Standard gas @ 2 MPa.Standard gas @ 10 MPa.High CO2 gas @ 2 MPa.High CO2 gas @ 10 MPa.

Figure 7.1. Joule-Thomson coefficient (K/MPa) for a standard and a high carbon dioxide natural gas composition for given temperature condition.

Component Mole percent Methane 83.155 Ethane 7.05 Propane 3.514 n-Butane 0.554 i-Butane 0.412 n-Pentane 0.048 i-Pentane 0.074 n-Hexane 0.012 Nitrogen 3.106 Carbon Dioxide 2.075 Molecular Weight 19.4532 kg/kmol Density @ MSC 0.825 kg/Sm3 Relative Density 0.6732 HHV 40.8047 MJ/Sm3

Table 7.1. Natural Gas Composition Applied to Calculations

Page 29: Sulphur

Normal composition High CO2 composition

Speed of Sound (m/s). Speed of Sound (m/s). 10.0 MPa

abs. 2.0 MPa abs. 10.0 MPa abs. 2.0 MPa abs.

273 K 362.23 374.25 333.05 348.31

283 K 371.13 382.20 341.28 355.86

293 K 380.41 389.80 350.21 363.07

303 K 389.58 397.07 359.16 369.97

Table 7.2. Speed of Sound values (m/s) for a standard and a high carbon dioxide natural gas composition for given temperature condition.

Pressure control valves used at natural gas pressure reduction facilities may use specially designed valve cages to reduce the noise associated by the turbulence created by the high pressure reduction. The valve cage design may utilize multiple orifices (labyrinth type) that act to break up the turbulent gas stream. The number of orifices in the cage may mount up to many hundreds in number. Although significant noise reduction can be achieved through the application of such trims, they are very susceptible to sulphur deposition. For pressure regulation facilities with a significant pressure reduction, there is generally a requirement to minimize the noise from the pressure regulation valve. This is particularly the case for pipeline facilities that are in, or near built-up areas. The noise, or more precisely the aerodynamic noise, generated at the valve results from the turbulent flow of the flowing stream. The main source of the turbulence is the rapid expansion of the gas stream, however obstructions in the path of the gas stream, together with directional changes of the stream also contribute to the noise generated. Case Study. Take a common transmission pipeline pressure reduction facility experiencing “elemental sulphur” deposition and is fitted with a pressure control valve having a labyrinth type cage. Assuming that the pressure reduction is significant and the gas composition and flow conditions at the valve are such that some retrograde condensation may occur. Initially the gas flow rate at the outlet of the control valve will vary according to the position of the valve plug. That is, the control valve is operating as expected. If the valve cage trim consists with multiple orifices then there will be a slow build up of “elemental sulphur” around and in the orifices. Initially, this build up will not impact the operation of the valve, however at some point in time the position of the valve plug will cease to control the flow rate due to the flow restrictions (contamination build-up) collectively across the cage trim orifices. This contamination build up will start to restrict

Page 30: Sulphur

the required flow and although the valve plug may go to the wide open position in response to the control system there will be no resultant change in the flow rate. With the flowrate now essentially controlled by the trim cage orifices, a new set of conditions will apply to the “elemental sulphur” desublimation process. Due to the increased restrictions for the gas flow path through the trim cage, there will be an increase in the pressure drop and hence increase in the gas velocity. This will result in an increase in the sulphur desublimation, coagulation and deposition rates and hence lead to blockage of the cage orifices. Figure 7.2 demonstrates the pressure drop across a valve cage wall orifice. It will be noted that there will most likely be a minimum pressure point that is actually lower than the required output pressure. This point is the ‘vena contracta’. Valve manufacturers have generally met the requirement for reducing the noise from pressure control valves through the design of the valve cage. The design criteria generally applied is to have the cage wall consist of many small diameter holes. The principle being that through the application of these small holes, the peak frequency of the resulting noise will be shifted to outside the audible range of humans. The smaller the diameter of the holes, the higher will be the resulting frequency. A reduction in the diameter of the cage holes from a nominal bore size of 6 mm to 3 mm could be expected to reduce the valve noise level around 5 to 7 dBA. This reduction of noise could mean the compliance with operating / environmental requirements for a transmission pipeline pressure regulation station. Although the valve cage design will considerably assist in the reduction of noise levels, it unfortunately contributes to favourable conditions for the formation and deposition of “elemental sulphur”. Orifice Gas flow Pin Cage wall Pout Vena contracta

Figure 7.2 Pressure Drop across a Valve Cage

Page 31: Sulphur

Figure 7.2 is not drawn to any scale – it is presented for demonstration purposes only. With any build up of contamination on the inlet to the orifice, or within the orifice bore, the gas stream flow will become distorted with the location of the vena contracta most likely shifting, probably in both magnitude and position. This is an area that requires further investigation. 8.0 Gas Quality Determination and Application.

The purpose of this section is to demonstrate, through the application of realistic natural gas compositions and operating conditions, how two-phase flow situations may develop within natural gas transmission pipelines. Although only very small quantities of hydrocarbon liquid may be generated by the given examples (less than 0.0005 mol fraction), such volumes are more than sufficient to meet the “elemental sulphur” deposit criteria. As found through the Phase 1 studies, and with reference to Figures 6.1 and 6.2, sub ppm levels of sulphur vapour are sufficient for the formation of the observed “elemental sulphur’ deposits.

Component Normal composition

(1)

Normal composition plus C16 (2)

Vapour phase (1)

Vapour phase (2)

Liquid phase (2)

Methane 0.8309 0.8308 0.8309 0.8309 0.2431 Ethane 0.0705 0.0705 0.0705 0.0705 0.0736 Propane 0.0351 0.0351 0.0351 0.0351 0.0948 i-Butane 0.0041 0.0041 0.0041 0.0041 0.0222 n-Butane 0.0055 0.0055 0.0055 0.0055 0.0388 i-Pentane 0.0007 0.0007 0.0007 0.0007 0.0103 n-Pentane 0.0005 0.0005 0.0005 0.0005 0.0084 n-Hexane 0.0003 0.0003 0.0003 0.0003 0.0126 n-Heptane 0.0002 0.0002 0.0002 0.0002 0.0214 n-Octane 0.0002 0.0002 0.0002 0.0002 0.0398 n-Nonane 0.0001 0.0001 0.0001 0.0001 0.0560 n-C16 - 0.0001 - - 0.3590 Nitrogen 0.0311 0.0311 0.0311 0.0311 0.0040 Carbon dioxide 0.0208 0.0208 0.0208 0.0208 0.0158

Total 1.0000 1.0000 1.0000 1.0000 1.0000

Table 8.1. Gas Composition with and without simulated Oil Fraction

Table 8.1 provides the composition, expressed as mole fraction, of a typical natural gas mixture up to and including nonane (C9), as given in column 2. Column 3 is the column 2 gas composition with 0.0001 mole fraction of n-hexadecane (C16). n-hexadecane has been selected as it has a molar mass o 226.4 kg/kmol and density (in liquid form) of 773.2 kg/m3 – these properties being representative of light compressor oil.

Page 32: Sulphur

The purpose of making reference to these two nearly identical natural gas compositions is to demonstrate the results of performing a flash calculation, using the Peng Robinson equation of state (PR-EOS), at a typical transmission pipeline operating condition. The conditions selected are:

6 Pipeline conditions of 7,000 kPa with flowing gas temperature of 25.0 OC. 7 Pressure reduction from 7,000 kPa to 2,000 kPa.

8 Pressure reduction facility has an on-line gas heater delivering gas at

45.0 OC. An allowance of 10 OC has been made for temperature loss between the heater and single pressure regulation stage.

Columns 3 to 6 inclusive of Table 8.1 provide the flash calculation results for the two gas compositions at the specified pipeline operating conditions of 7,000 kPa and 25 OC. It will be noted that the gas composition without the n-hexadecane (simulated oil deposit) has the entire fluid composition in the vapour phase (column 4). As would be expected, the gas composition with the n-hexadecane is a two-phase fluid; that is there is a very predominant gaseous phase and a very small liquid phase. Expressed as a vapour/phase fraction, the vapour phase fraction is 0.9997 with the liquid phase being 0.0003. The two phase fractions for the n-hexadecane gas are represented by columns 5 and 6 in Table 8.1. There point of interest here is that there is nominally only 0.0001 mol fraction of n-C16, however the overall liquid fraction is 0.0003, a factor of three times greater. If a natural gas stream with the composition given in Table 8.1 has its n-hexadecane content increased to 0.0003 mole fraction (value increased to provide better calculation resolution), is heated to 45 OC by the pressure regulation station heater, with the gas stream then directed to the pressure regulation facility, the following results are obtained from the relevant flash calculations:

Component Compositionat heater

inlet

Vapour phase at

7,000 kPa

Liquid phase at

7,000 kPa

Vapour phase at

2,000 kPa

Liquid phase at

2,000 kPa

Methane 0.8306 0.8309 0.2184 0.8311 0.0890 Ethane 0.0705 0.0705 0.0599 0.0705 0.0424 Propane 0.0351 0.0351 0.0722 0.0351 0.0780 i-Butane 0.0041 0.0041 0.0162 0.0041 0.0240 n-Butane 0.0055 0.0055 0.0276 0.0055 0.0454 i-Pentane 0.0007 0.0007 0.0070 0.0007 0.0157 n-Pentane 0.0005 0.0005 0.0056 0.0005 0.0137 n-Hexane 0.0003 0.0003 0.0079 0.0003 0.0277 n-Heptane 0.0002 0.0002 0.0127 0.0002 0.0574 n-Octane 0.0002 0.0002 0.0221 0.0001 0.1054 n-Nonane 0.0001 0.0001 0.0294 - 0.1112

Page 33: Sulphur

n-C16 0.0003 - 0.5038 - 0.3817 Nitrogen 0.0311 0.0311 0.0039 0.0311 0.0011 Carbon dioxide 0.0208 0.0208 0.0132 0.0208 0.0075

Total 1.0000 1.0000 1.0000 1.0000 1.0000

Table 8.2. Gas Composition with 0.0003 mol fraction n-hexadecane The following vapour / phase fractions are determined for the given gas composition:

Vapour phase fraction at 7,000 kPa and 45 OC: 0.9994

Liquid phase fraction at 7,000 kPa and 45 OC: 0.0006 Vapour phase fraction at 2,000 kPa and 8.2 OC: 0.9992 Liquid phase fraction at 2,000 kPa and 8.2 OC: 0.0008

The above gas composition examples do provide a plausible explanation as to why the analysed “elemental sulphur” samples have such a high make-up proportion of hydrocarbons. Although the flash calculations do indicate that small hydrocarbon liquid fractions will occur at the given pipeline conditions, the flash calculations are as good as the provided gas composition and pressure temperature conditions. As a process on-line gas chromatograph system will sample the natural gas stream at a pressure that is significantly reduced from the pipeline operational conditions, the gas composition given by column 5 in Table 8.2 would be more indicative of the sampled gas than would be the composition given in column 2. Although only marginally different, the variation in the cricondentherm value between the two gas compositions is: Gas composition column 3, Table 8.2: Cricondentherm = 10.6 OC Gas composition column 5, Table 8.2: Cricondentherm = - 4.4 OC - a difference of 15.0 OC. The results have been derived from the application of Hysys version 3.2 using the Peng Robinson equation of state. The above results demonstrate that the care may need to be exercised when directly applying the gas composition, as determined from an on-line gas chromatograph system, to the calculation of gas quality parameters. The information given in Table 8.3 provides a valuable insight into the detailed analysis of transmission quality natural gas. This information has been sourced from a paper by Luijten et al [11], into multi-component nucleation and droplet growth in natural gas. These authors investigated the condensation behaviour of natural gas consisting of over 30 components. The experimental results provide information on homogeneous nucleation and droplet growth in a multi-component gas-vapour mixture. Of particular interest to the findings and postulations of the “elemental sulphur” formation and deposition research work is the determined natural gas compositions applied by

Page 34: Sulphur

Luijten et al [12]. to their studies. The natural gas compositions applied to the condensation behaviour studies, were derived directly from the Dutch natural gas distribution system. The experiments were performed using three different types of gas cylinders, denoted as A, B and C in Table 8.3. Each cylinder was filled from the same gas supply. It is to be noted that careful isobaric flushing of the test cylinders was observed. It was assumed that through this process any possible wall adsorption of heavy hydrocarbons would be equilibrated, thus ensuring that the test gases had the same molar fractions as the source gas. Although it is not suggested that the gas compositions given in Table 8.3 are atypical of all natural gas transmission systems, it does demonstrate that higher paraffins, some aromatics and naphtenes, in trace quantities, can be present. Also the finding by Luijten et al [13]. demonstrate the importance of appropriate sampling techniques.

Component Type A Type B Type C Methane 8.13E-1 8.13E-1 8.13E-1 Ethane 2.85E-2 2.84E-2 2.84E-2 Propane 3.95E-3 3.92E-3 3.92E-3 n-Butane 7.31E-4 7.32E-4 7.25E-4 2-Methylpropane 6.37E-4 6.43E-4 6.33E-4 n-Pentane 1.62E-4 1.63E-4 1.60E-4 2-Methylbutane 1.68E-4 1.69E-4 1.66E-4 2,2-Dimethylpropane 7.20E-5 7.34E-5 7.21E-5 n-Hexane 6.03E-5 6.06E-5 5.97E-5 3-Methylpentane 2.14E-5 2.15E-5 2.12E-5 2,2-Dimethylbutane 5.62E-5 5.73E-5 5.64E-5 2,3-Dimethylbutane 3.56E-5 3.58E-5 3.51E-5 n-Heptane 2.86E-5 2.88E-5 2.83E-5 i-Heptane 4.38E-5 4.76E-5 4.70E-5 (n+i)-Octane 3.30E-5 2.69E-5 3.23E-5 (n+i)-Nonane 1.60E-5 1.40E-5 1.78E-5 (n+i)-Decane 1.12E-5 1.11E-5 1.10E-5 (n+i)-Undecane 2.60E-6 2.90E-6 2.90E-6 (n+i)-Dodecane 2.00E-7 2.00E-7 2.00E-7 Benzene 1.77E-4 1.71E-4 1.67E-4 Toluene 3.96E-5 3.62E-5 3.95E-5 Xylenes 1.34E-5 1.33E-5 1.28E-5 Cyclopentane 1.30E-5 1.31E-5 1.31E-5 Cyclohexane 1.94E-5 2.02E-5 1.99E-5 Methylcyclohexane 2.46E-5 2.44E-5 2.41E-5 Helium 5.00E-4 5.00E-4 5.00E-4 Nitrogen 1.41E-1 1.42E-1 1.42E-1 Carbon dioxide 9.90E-3 9.90E-3 9.90E-3

Total 0.999 1.000 1.000

Page 35: Sulphur

Table 8.3. Natural gas composition (molar fractions) per sample cylinder.

Figure 8.1. Further view of the Contaminated Labyrinth type Valve Cage.

Page 36: Sulphur

Figure 8.2. Pressure Control Valve - Equal Percentage Cage with “Elemental Sulphur” deposition.

[Although deposition still occurs on the valve cage in Figure 8.2, the valve will still provide reasonable performance.] 9.0 Two-phase Flow at Pipeline “T” Junctions. From the extensive study of the preferential site selective process of “elemental sulphur” deposits within two transmission pipeline systems, it was noted that the most severely impacted sites are located on a lateral downstream of a mainline “T” junction. This initiated an extensive study into fluid flow behaviour at “T” junctions, with particular emphasis on two-phase flow. As the analysed “elemental sulphur” deposits demonstrated a high concentration of hydrocarbon components (C11 +), the phase 1 studies concentrated on the potential source of liquid hydrocarbons. The potential to have preferential flow of liquids into a lateral from a mainline “T” junction clearly provided a plausible source / mechanism for the abnormal liquid hydrocarbon concentrations at the affected pressure regulation / metering sites. Unfortunately the majority of research conducted into two-phase flow at pipe “T” junctions has used water and air as the fluid under study. Also pipe diameters have been limited to a few centimetres with operating pressures generally less than 1,000 kPa. The main and offtake, or branch pipe diameters in most cases had the same diameter. Although these physical and operating conditions are not that of a transmission pipeline system, there is however believed to be some relevance of the “T” junction two-phase flow behaviour results to natural gas pipeline situations. However it has to be acknowledged that the “T” junction two-phase flow characteristics that have been determined for small diameter pipes may not necessarily apply to large transmission sized pipes.

Page 37: Sulphur

Extensive CFD modelling has been conducted for a 26 inch diameter mainline / 12 inch diameter lateral offtake. The information applied to the modelling used the pipeline operational data spanning a continuous period of nearly 12 months. Although an average gas superficial velocity at the referenced “T” junction was able to be determined, an average liquid superficial velocity had to be assumed. However, results were checked for validity against published or otherwise verified information on gas and liquid superficial velocities in natural gas transmission pipelines. 9.1 The Dynamics Associated with Fluid Mixing at Pipeline ‘T’ Junctions’. Results derived from both the phase 1 and 2 research work have demonstrated that hydrocarbon liquids and derivatives are present in natural gas transmission pipeline systems. This observation has backing from a number of extensive research projects and publications into this matter. The diversity of the hydrocarbon liquids present plus the potential quantity under certain conditions has exceeded general expectations. Retrograde condensation would, no doubt, be a major contributing factor. It is quite feasible in certain locations on pipeline systems for there to be a two-phase hydrocarbon flow. However it must be appreciated that the ratio of liquid to gas flow volume will be an extremely low fraction. This two-phase flow behaviour is of particular interest at mainline-lateral or ‘T’ junctions on natural gas transmission pipelines. Research [14] into this phenomenon has demonstrated that if a certain fraction of the gas phase is being extracted into a lateral, it is uncommon for the liquid component fraction to remain the same as in the upstream mainline section. What occurs is an uneven ‘phase splitting’ between the two resulting downstream flows of the pipeline junction node. Further, it has been found that any variations in the pipeline operating conditions can significantly alter the preferred route for the liquid flow. The gas/liquid flow in the mainline and lateral in the vicinity of the pipeline junction node can be modelled in accordance with Bernoulli’s equation. Obviously the upstream gas velocity in the mainline will be greater than the downstream gas velocity, assuming the pipe diameter remains constant. This Δv will result in a static pressure increase in the mainline and a static pressure decrease in the lateral entry due to friction. The pressure difference between the mainline and the lateral will be the deciding factor for the path of the liquids. The resulting pressure difference will be dependent on flow velocities, the liquid/gas ratio and the physical characteristics of the pipeline junction node. This is clearly an important criterion for the understanding of why there are selective sites for the ‘elemental sulphur’ formation and deposition processes.

Page 38: Sulphur

Figure 9.1 Baker’s Flow Map

By applying the averaged flow conditions for the studied transmission pipeline “T” junction (average gas velocity of 6.0 m/s) and assuming a superficial liquid velocity of 0.0030 m/s, an annular flow pattern within the pipeline at the referenced “T” junction is determined in accordance with the Baker’s flow map criteria. The averaged flow conditions at the referenced “T” junction on the transmission pipeline is represented by the red circle on Figure 9.1

Parameter Gas Liquid

Density (kg/m3) 63.48 727.8

Viscosity (cP). 0.01321 1.134

Table 9.1. Fluid Properties for Determination of Flow Pattern. For the determination of the fluid pattern, per the Baker’s flow map as given by Figure 9.1, the parameters given in Table 9.1 were applied, together with the following operating conditions:

Pressure: 6,500 kPa. Gas temperature 20 OC Gas Composition: per Table 7.1 with small amount of liquid hydrocarbons

representative of compressor oil (MW = 226.4 kg/kmol).

The following equations have been applied to determine the flow pattern of the liquid within the pipeline, (per Baker’s flow map):

Page 39: Sulphur

333.023.6273

⎥⎥⎦

⎢⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛⎟⎠⎞

⎜⎝⎛=Ψ

LL ρ

μσ

⎟⎟⎠

⎞⎜⎜⎝

⎛⎟⎠⎞

⎜⎝⎛⎟⎠⎞

⎜⎝⎛=

3.62075.0LG ρρ

λ

where GG = gas mass velocity

LG = liquid mass velocity σ = liquid surface tension

Lμ = liquid viscosity

Gρ = gas density

Lρ = liquid density It is generally considered that a two-phase flow with a gas superficial velocity greater than 6 m/s, but less than 60 m/s, will have the liquid flow in annular mode. An annular flow is one in which the liquid flows as a film around the pipe internal wall and the gas flows as a core. A portion of the liquid may be entrained within the gas stream as a spray. However, as the pipeline superficial velocity is at the lower end of the annular flow velocity range, any gas stream entrained mist flow would be minimal. Although it is not suggested that there is a continuous thin film of hydrocarbon liquid coating the internal pipe walls, this exercise does provide an estimate of the liquid flow patterns should there be any liquid present. In most natural gas pipeline situations, any annular flow conditions would be intermittent. Indeed, with reference to Figure 9.1 it will be noted that the pipeline liquid flow pattern is bordering on the wavy flow pattern. The determined liquid annular flow pattern supports the hypothesis that there exists, under certain pipeline operating conditions, a mechanism that supports preferential flow of any liquids within a mainline to a lateral that is fed from a horizontal (or near horizontal) pipeline “T” junction. It is important that the liquid flow pattern be annular and not spray or dispersed. For spray or dispersed flow nearly all of the liquid is entrained as fine droplets in the gas stream. Due to the momentum of the droplets in the gas stream, it would be unlikely that such droplets would change direction at the “T” junction and enter the lateral, unless the gas flow into the lateral was much higher than the mainline flow. As determined from the Phase 1 studies, only ppm levels of hydrocarbon liquid are required at the point of the sulphur vapour desublimation for the formation of the observed and analysed “elemental sulphur” deposits. This means that the amount of liquid in the mainline just needs to be a very small fraction of the gas flow. To illustrate the quantity of hydrocarbon fluid required to meet the analysed composition of common “elemental sulphur” samples, the following example is applied as per Pack [15]. In this example it was determined that the amount of elemental sulphur required,

Page 40: Sulphur

expressed as a volumetric fraction, was 0.009 ppmv. Case study:

Metering and pressure regulation station on a lateral. Daily gas flow in lateral approximately 1.7 % of mainline flow, with the majority of lateral gas being delivered through the referenced pressure regulation facility. Assume very small quantities of hydrocarbon liquid can be intermittently present at the mainline “T” junction. The flow pattern of the hydrocarbon liquid is annular; therefore it is traveling very slowly as an intermittent film around the pipeline walls. Taking the lateral gas fraction intake as 0.017 and assuming the lateral liquid fraction intake to be 0.1. If the volume of hydrocarbon liquid is to be taken as 1 ppmv (a very generous 100 fold increase on the sulphur requirement), then the volumetric fraction of liquid in the mainline will be 0.15 ppmv. This simple calculation does not take into account that, in all probability, there will be some retrograde condensation occurring at the pressure regulation facility. This means that the given requirement of 1 ppmv of liquid in the lateral is very conservative.

As already referenced, research into the behaviour of two-phase fluid flow at pipe “T” junctions has been reasonably widely studied, however not at natural gas transmission pipeline conditions or with natural gas / condensate as the fluids applied to the studies. The research conducted has shown that when liquid and gas flows in a pipeline encounter a T-junction, the two phases very rarely split in the same ratio. Some results have shown that all the liquid may be diverted into the lateral, with other sets of conditions showing that all the liquid may maintain its flow in the mainline. The fraction of liquid diverted into the branch can be very different from that of the gas. Oranje [16] is credited with being the first person to investigate the phenomenon of gas / liquid splitting at a “T” junction. Through laboratory experiments, this researcher suggested that the mechanisms controlling the liquid route selectivity are a relatively lower pressure in the branch arm, liquid inertial forces, upstream flow pattern and the branch arm geometry. According to Penmatcha et al [17] this unpredictable splitting of the liquid phase between the branch (lateral) and the run arms (mainline) is complicated due to the large number of variables that influence it. Geometry of the tee arms, flow pattern upstream of the T-junction, the inclination of the branch arm, the gas and liquid flow rates, and the gas fraction diverted into the branch are given as the important variables that determine the liquid splitting between the two arms of the tee.

Page 41: Sulphur

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.2 0.4 0.6 0.8 1.0

Branch Liquid Fraction Intake

Bra

nch

Gas

Fra

ctio

n In

take

0.0030 m/s horizontal

0.0051 m/s horizontal

0.0051 m/s 1.0 degupward0.0051 m/s 5.0 degdownward

Figure 9.2. Experimental results for “T” junction splitting ratios [Information derived from Penmatcha et al. (1996)]

The experimental results derived by Penmatcha et al.,[18] were for a superficial gas velocity of 6.1 m/s and pipe mainline and branch diameters equal at 51 mm. The results shown in Figure 9.2 are for branch pipe orientation of horizontal, 1.0 degree upwards and 5.0 degrees downward. Operating pressure was 295 kPa (abs). The superficial liquid velocities being 0.0030 or 0.0051 m/s. It is interesting to note that pipeline “T” junctions encountered in the field are seldom truly horizontal; also they can very often have an inclined branch arm. The true geometry of the studies pipeline “T’s” is unknown; however both are situated within an undulating countryside. As indicated by the results from Penmatcha et al [19], the inclination of the mainline and/or lateral may have an influence on the liquid fraction split at the “T” junction. Other researchers, such as Chen et al [20], have formulated mathematical models in an effort to define the nature of the problem involving liquid yields in natural gas pipeline systems and in particular downstream of T-junctions. These works were generally motivated by the observation that thermodynamic considerations of the retrograde condensation alone could not fully account for liquid yields, and that the dynamics of gas mixing pertains to such rate processes as nucleation / droplet growth, gas absorption by liquid droplets carried over from upstream processing units, and heat and mass transfer between the bulk gas phase and liquid droplets. Although large amounts of data have been acquired and a number of models developed for the two-phase splitting phenomena at pipeline “T” junctions, unfortunately no one model appears to be able to predict the liquid split accurately or provide a complete

Page 42: Sulphur

understanding to the fluid splitting criteria. The preferential appearance of condensate, or other liquids, at some pipe nodes or actual pipeline delivery stations are commonly referenced within the works studies.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.2 0.4 0.6 0.8 1.0

Branch Liquid Fraction Intake

Bra

nch

Gas

Fra

ctio

n In

take

.

0.063 kg/s horizontal

0.0126 kg/s horizontal

Figure 9.3. Experimental results for “T” junction splitting ratios [Information derived from Azzopardi. (1992)]

The data used to generate the plots given in Figure 9.3 has been derived from the experiments conducted by Azzopardi [21]. The applied data has a gas flow rate of 0.101 kg/s, with two differing liquid flow rates of 0.063 kg/s and 0.0126 kg/s. The orientation of the pipework is horizontal with the mainline having a diameter of 38 mm and the branch pipe a diameter of 25 mm. Operating pressure was 300 kPa abs. Of particular interest is the branch gas / liquid fraction ratio at low gas fraction intake values. The experimental work conducted by Shoham et al [22] has demonstrated that for annular flow conditions the liquid phase tended to flow preferentially into the branch (lateral) line. It was observed that for low branch gas fraction intake, the liquid splitting ratio was higher than the gas splitting ratio. It was concluded that there is an apparent branch liquid fraction threshold, since the liquid phase tended to flow into the branch even for negligible gas fraction intake. These observations support the modelling and theoretical studies conducted through the Phase 2 research into why there is preferential site selection for the formation / deposition of “elemental sulphur”. Interestingly, if the liquid flow pattern is stratified, then it appears that the reverse of annular flow driven conditions occurs at the “T” junction; that is the liquid will tend to continue to flow in the mainline. Although the geometry of the pipework and fluids used for the majority of the “T” junction

Page 43: Sulphur

flow studies do not reflect natural gas transmission pipeline conditions, the noted distinct differences between flow patterns (e.g. stratified and annular) does indicate the complexity of this situation and provides credibility to the theory that selective flow of any liquid deposits within a natural gas transmission pipeline can, and does, occur. Obviously the preferential liquid flow behaviour at pipeline “T” junctions is controlled by competing inertial and centripetal forces that act on the liquid phase. Under annular flow conditions, the centripetal forces would be dominant resulting in the liquid phase preferentially flows into the lateral. 9.2 CFD Modelling Studies into a Transmission Pipeline “T” Junction. As already referenced, one of the main aims of the Phase 2 studies was to conduct extensive modelling studies into the two-phase fluid behaviour at an actual transmission pipeline “T” junction. The particular pipeline “T” junction studied has a known “elemental sulphur” contaminated site situated at the termination point of a lateral. The following Figures show the results from these studies.

Figure 9.4. Contours of Velocity for a Gas-Split of 0.02

To provide a visible representation of the liquid flow within the pipeline, the given Figures have an exaggerated liquid flow rate. As will be appreciated, the actual determined volumes of liquid are so small they would just not appear on any graphical presentation of the pipeline situation. The CFD package used was Fluent. The results are in the form of contours that represent various physical properties of the fluid under investigation. Figure 9.4, which applies an exaggerated gas-liquid split of 0.02, demonstrates the velocity contour for the

Page 44: Sulphur

two-phase fluid for the averaged pipeline conditions of the “T” junction under study. Figure 9.5 is a cross-section of the mainline immediately downstream of the “T” junction. The lateral offtake is on the right hand side of the pipe. Although a little difficult to visualize, it will be noted that the CFD program has predicted that for the pipeline liquid annular flow pattern, an arc of approximately 90O of the liquid film has been directed into the lateral. The mainline to lateral gas flow ratio being set at 60:1

Figure 9.5. Contours of Liquid Film Thickness – Mainline Downstream of “T”

Figure 9.6 represents the gas-liquid flow in the lateral pipe immediately downstream of the “T”. Note the orientation of the ‘exaggerated’ liquid flow on the left hand side of the Figure. The CFD program has clearly predicted a much higher liquid fraction going into the lateral. The left hand side of the lateral is on the upstream section of the mainline.

Page 45: Sulphur

Figure 9.6. Contours of Liquid Film Thickness – Lateral immediately downstream of the “T”.

Figure 9.7 represents a cross section of the mainline upstream of the “T”. A near uniform annular flow pattern of the liquid is predicted

Page 46: Sulphur

Figure 9.7. Contours of Liquid Volume Fraction in Mainline (upstream).

Figure 9.8 provides a snapshot of an enhanced section of the gas-liquid and liquid-pipe wall interfaces.

Figure 9.8. Enlargement of Gas-Liquid Interface per Section in Figure 9.7 Although the CFD studies were a little more limited than planned, the information gained through the demanding modeling exercises have provided further valuable information on the projected fluid behaviour at a pipeline “T” junction. This in formation is in agreement with observations made and other theoretical studies and “elemental sulphur” analysis results. 9.3 Liquid Formation and Hold-up within a Transmission Pipeline. From the basic modelling of a typical transmission pipeline section, with inlet to the section fed from a gas processing plant or an on-line pipeline compressor, an estimate has been made of the liquid drop-out profile along the pipeline section. The model assumes warm gas enters the inlet section, at between 308 and 313 K. The gas enters at pressure P1 and exits at pressure P2. P1 is always greater than P2.

The differential pressure value between P1 and P2 is due to the overall pipeline friction factor along the pipeline section, and is a function of flow rate. It is assumed that the temperature of the warm gas entering the pipeline section will decay at an inverse natural logarithmic rate function to very near the mean ground temperature along the

Page 47: Sulphur

pipeline section. The gas composition, as given in Table 9.2 has been used for the modelling exercise. Figure 9.9 gives an indication of the position of the liquid hold-up profile for a given flow and gas composition. The liquid hold-up is expressed in Sm3/h.

Component Mole percent Methane 83.1280 Ethane 7.0500 Propane 3.5140 n-Butane 0.5540 i-Butane 0.4120 n-Pentane 0.0480 i-Pentane 0.0740 n-Hexane 0.0120 n-Heptane 0.0245 n-Octane 0.0015 n-Nonane 0.0010 Nitrogen 3.1060 Carbon Dioxide 2.0750

Total 100.000

Table 9.2 Gas Composition Applied to Liquid Hold-up Example

The quoted mass flow rate of 540,000 kg/h equates to a daily gas energy flow rate of 641 TJ/day, using the Table 9.2 gas composition. This would approximately represent the daily flow rate within the two largest throughput natural gas transmission pipelines in Australia. The liquid hold-up profile will be dependent on the gas flow rate, gas composition and operating pressure and temperature conditions. The shape of the curve is a function of the components that are subject to retrograde condensation. The finite number of points generated for the graph is probably also a factor in the curve shape.

Page 48: Sulphur

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

0 20 40 60 80 100 120 140

Pipeline Segment Distance (km)

Liqu

id V

olum

e Fo

rmed

(Sm

3 /h)

Inlet Outlet

Conditions:Inlet gas at 8,000 kPa abs @ 313 KOutlet gas at 5,645 kPa abs @ 293 KMass f low rate: 540,000 kg/hGas composition per Table 5 w ithaddition of 0.005 total mol fraction ofC7, C8 and C9.

Figure 9.9. Pipeline Segment Liquid Hold-up Profile

Taking the gas flow conditions, as given in Figure 9.9, then during the transition period of the flowing gas stream going from the warm 313 K to near the ground temperature conditions (example taken as 293 K), there is the potential for sulphur vapour in solution to become supersaturated and hence desublimate out as elemental sulphur. Therefore for the given scenario of the liquid hold-up profile, not only will there be selective retrograde condensation within the pipeline segment, but also the possibility of concurrent elemental sulphur formation. Should both the hydrocarbon liquid and elemental sulphur formation occur concurrently, then depending upon the flow ratios and gas velocities at the junctions between any laterals and/or off-takes and the mainline, the liquid/elemental sulphur agglomerate will selectively travel along one of the pipeline systems. It is to be appreciated that the discussion presented with respect to liquid hold-up is for a section of pipeline only. According to Meng et al [23] for low liquid loading in gas pipelines, the most important parameters governing the flow behaviour are pipe geometry (inclination angle and diameter), operating conditions (flow rate, pressure and temperature), and physical properties of the gas and liquid (density, viscosity and surface tension). Pipeline inclination angles of just up to ± 2 % (reference) have been shown to have an influence on the flow behaviour. Even when single-phase gas enters a pipeline, condensate traces can be formed by retrograde condensation. The presence of condensate traces can lead to a significant increase in pressure loss along a pipeline. For instance Hart [24], states

Page 49: Sulphur

that condensate traces as small as 0.5% by volume can result in a 30% greater pressure loss than in a single-phase gas flow. As already referenced, the most commonly encountered flow patterns in low liquid loading pipelines are stratified and annular flow. With annular flow, the flow pattern of interest in this paper, a number of distinctly differing flow patterns have been observed and documented by various researchers. No doubt the conditions of the fluid and the pipe characteristics will have a significant influence on the resulting flow patterns. Annular flow has been observed with a very slow moving liquid film along the upper half of the pipe wall with a relatively fast moving film, resembling stratified flow, occurred along the lower half of the pipe wall. Another type of annular flow reported is characterized by roll waves existing all around the inside pipe wall.

Pipe inclination (deg) 2.0O 1.0O 0.0O -1.0O -2.0O Gas superficial velocity (m/s)

5.0 0.090 0.054 0.045

7.5 0.208 0.088

10.0 0.080 0.060 0.068 0.049 0.043

15.0 0.044 0.036 0.043 0.032 0.029

20.0 0.034 0.028 0.024 0.025 0.022

Table 9.3 Liquid Hold-up as a Function of Pipe Inclination and Gas Superficial Velocity.

[Adapted from Meng, et al [25]] Again the majority of liquid hold-up information has been derived for water-air systems. Unfortunately high-pressure natural gas transmission pipeline situations with a gas – hydrocarbon liquid flow are poorly documented. Table 9.3 provides information derived by Meng [26], for liquid holdup versus gas (compressed air) velocity for an in-situ liquid volume loading of 1,800 m3 for pipe inclination angles of 2O, 1O, 0O, -1O and -2O. As compressed air was used for the gas phase, the results are indicative only for a natural gas pipeline situation as any impact of retrograde condensation or hydrocarbon gas / liquid interaction is not accounted for. The main purpose of Table 9.3 is to demonstrate the change in liquid hold-up to the defined variations. A high liquid loading has been selected to help emphasis the impact of the change in the pipe inclination to the liquid hold-up value. Hart et al [27] developed the below equation for the calculation of the liquid holdup in a pipeline system. It is claimed that the results determined from the application of this equation are in good agreement with experimentally determined values.

Page 50: Sulphur

⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

⎥⎥⎦

⎢⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛+=

−−

5.0363.0Re4.101

1 G

LSL

G

L

L

L

ρρ

υυ

εε

where =Lε liquid holdup

=Lυ superficial liquid velocity

=Gυ superficial gas velocity

=SLRe superficial liquid Reynolds number

=Lρ liquid density

=Gρ gas density The amount of liquid within a natural gas transmission pipeline system will vary between each pipeline system, from constant liquid presence through to intermittent flows to no liquid at all. In many cases the presence of liquid will be event driven with ‘one-off’ situations, such as gas processing plant upset conditions, being the cause of the liquid. Liquids, such as lubricating oils, from on-line or external upstream compression facilities and gas treatment plants (gas dehydration fluids such as glycol), and / or products from retrograde condensation may be present in the pipeline, and / or be entrained within the gas stream. Liquids may be ‘dormant’ at a pipeline depression for long periods. Such liquids may only accompany the gas flow during high flow rates. With respect to liquid entrainment, a definition of what constitutes a ‘dry’ and ‘wet’ natural gas pipeline has tried to be determined. Although there does not appear to be a common break point between the two categories, the liquid entrainment value applied for a ‘dry’ gas pipeline appears to be 10 bbls/MMSCF. This figure translates to 1.0 m3 liquid for 17,800 Sm3 gas. For the following example 2, the ratio will be around 1.0 m3 of liquid for 226 Am3 of gas at the referenced operating conditions. This means that the given examples, with a gas volume phase fraction at operating conditions of 0.9994, are well within the ‘dry’ gas pipeline criteria. To demonstrate the potential impact of liquid holdup, two examples are given based on actual pipeline situations.

1 Example 1. Liquid holdup due to retrograde condensation at a pressure reduction station, and

2 Example 2. Liquid holdup due to existing hydrocarbon based liquid within a

transmission system. Example 1.

If the gas composition given in Table 7.1 is slightly modified to include additional hexane and small amounts of the heavier hydrocarbons of heptane, octane and nonane (addition of 0.00070 mole fraction of heavier paraffins at the expense of the

Page 51: Sulphur

same mole fraction of methane), then the gas composition given in Table 9.4 is derived. For the following pipeline operational conditions a phase fraction (volume basis) of 0.9994 is obtained. This has been derived from a flash calculation using Hysys 3.2 Conditions at pressure reduction:

Inlet pressure: 7,000 kPa abs. Inlet temperature: 30 OC

Outlet pressure: 2,000 kPa abs. Outlet temperature: 2 OC Gas mass flow rate: 10,000 kg/h Density of gas: 18.42 kg/m3. Density of liquid: 684.5 kg/m3. Diameter of pipe: 0.1524 m (6 inches) Viscosity of liquid: 0.004516 Pa.s Gas velocity: 8.25 m/s Liquid velocity: 0.01 m/s (assumed)

[Outlet temperature and phase fraction determined by Hysys version 3.2 for given gas (Table 9.4) composition].

The liquid holdup is calculated as 0.012

If the liquid superficial velocity is reduced to 0.003 m/s, the liquid holdup value is reduced to 0.0035.

This gas composition applied is regarded as a realistic natural gas pipeline composition. The upstream gas composition is for the gas prior to the pressure reduction. Downstream gas composition is after the 5,000 kPa pressure reduction.

The cricondentherm value for the upstream gas composition is 11.71 OC at 3,229 kPa, whereas the downstream gas cricondentherm value is 3.88 OC at 3,272 kPa

Component Mole percent (up stream)

Mole percent (down stream)

Methane 83.085 83.104 Ethane 7.05 7.05 Propane 3.514 3.512 n-Butane 0.554 0.552 i-Butane 0.412 0.411 n-Pentane 0.048 0.048 i-Pentane 0.074 0.074 n-Hexane 0.030 0.029 n-Heptane 0.022 0.020 n-Octane 0.018 0.013 n-Nonane 0.012 0.005

Page 52: Sulphur

Nitrogen 3.106 3.107 Carbon Dioxide 2.075 2.075

Total 100.000 100.000

Table 9.4 Natural Gas Composition Applied to Example 1 Calculation.

Example 2.

The gas composition per Table 7.1 is applied; however 0.0001 mol fraction of hexadecane (n-C16) is applied. This component is selected as being the average molecular weight hydrocarbon for typical compressor oil that may be found within a transmission system. Conditions: In this case there is a normal flow situation with a set fixed pressure of 7,000 kPa abs and flowing temperature of 30 OC. From Hysys version 3.2, the gas phase fraction (volume basis) is determined as 0.9994 for the composition given in Table 9.5. For this case, the Hysys flash calculation determines the liquid phase to consist of just over 52 % of n-C16.

Component Mole percent Methane 83.145 Ethane 7.05 Propane 3.514 n-Butane 0.554 i-Butane 0.412 n-Pentane 0.048 i-Pentane 0.074 n-Hexane 0.012 n-Hexadecane 0.010 Nitrogen 3.106 Carbon Dioxide 2.075

Total 100.000

Table 9.5 Overall Gas / Liquid Composition Applied to Example 2 Calculations.

Component Mole percent gas

Mole percent liquid

Methane 83.156 23.1469 Ethane 7.05 6.792 Propane 3.513 8.5883 n-Butane 0.5535 1.9981

Page 53: Sulphur

i-Butane 0.4117 3.4607 n-Pentane 0.0479 0.9119 i-Pentane 0.0738 0.7357 n-Hexane 0.0119 0.4398 n-Hexadecane 0.0005 52.0085 Nitrogen 3.1065 0.3958 Carbon Dioxide 2.0751 1.5223

Total 100.0000 100.0000

Table 9.6 Gas & Liquid Compositions Derived from Example 2 Flash Calculation.

Therefore, for the hypothetical case of some compressor lubrication oil finding its way into a natural gas transmission pipeline, the flash calculation for the pipeline operational conditions, as given for example 2, shows that the amount of liquid is nearly doubled in volume. The amount of lubrication oil may only be a few hundred litres, however if this liquid does slowly and collectively travel downstream in an annular flow pattern and encounter a “T” junction, then there can be a multiplying effect within the lateral with respect to the liquid level. If the liquid split at the “T” junction is 6:1, with the higher liquid to gas fraction entering the lateral, then the resulting gas phase fraction (volume basis) is determined as 0.9961. The flash calculation has been determined using the modified overall pipeline fluid composition at the same transmission operating pressure and temperature conditions. If the gas stream is now subjected to a pressure reduction of 5,000 kPa, then the resulting flash calculation indicates that there will be little difference in the overall gas composition across the pressure reduction stage. However, if it is assumed that the majority of the hydrocarbon liquid in the lateral passes through this particular pressure reduction stage, then there is a reasonable amount of liquid present. For a gas flow of 10,000 kg/h, the amount of hydrocarbon liquid present at the pressure reduction stage will be at least 20 litres per day. This is more than sufficient to meet the “elemental sulphur” deposition composition criteria. This now raises the issue – does the presence of hydrocarbon liquid that already exists in the pipeline have a greater influence on the formation of “elemental sulphur” deposits than does the retrograde condensation occurring due to the pressure reduction at the pressure regulator? This is clearly an issue that needs further investigation. The above two examples, through the application of realistic data, demonstrate how hydrocarbon liquid deposits can exist and form within a transmission pipeline system. They also demonstrate the complexity of the problem. 9.4 Particle Deposition Studies in Natural Gas Transmission Pipelines The amount of research work performed on particle deposition in natural gas transmission pipelines is extremely limited. However, due to concerns about polychlorinated biphenyl (PCB) compounds contaminating a number of USA natural gas

Page 54: Sulphur

pipeline systems, some studies have been conducted to determine the levels of contamination. Field studies in the USA have demonstrated that the PCB contamination levels can vary considerably within the same pipeline system, this being particularly noted between the inlet and outlet levels at branching “T” junctions. According to Martinez et al [28], PCB’s tend to concentrate in the liquid phase, hence it was expected that for higher liquid-split conditions the branch would have higher concentrations of PCB. Figure 9.10 demonstrates a set of results from the hydrodynamic model developed by these authors. Martinez et al [29] have drawn on the limited set of experimental data, as generated by Oranje [30], for natural gas / condensate flow splitting at transmission system “T” junctions. A number of anomalies were found with respect to the PCB concentrations and the ratio of the mainline to branch pipe diameters. It was thought that the smaller the branch pipe diameter is to the mainline pipe diameter, the higher the potential gradient, and hence greater difficulty would exist for any liquid to enter the branch line. However, it was found that when the branch diameter is one-half the mainline diameter, the liquid split into the branch is higher than for branch pipe diameters that are less than the mainline diameter but greater than one-half the mainline diameter. The purpose of referencing the works of Martinez et al [31] is not to analyse the results, but demonstrate that uneven phase splitting at pipeline “T” junctions has been observed and modeled. Also the presence of liquid within natural gas transmission pipeline systems is acknowledged. The results from one particular survey of a PCB contaminated transmission pipeline are shown in Figure 9.10. The findings of the PCB concentrations, both from field observations and through the results of the application of a hydrodynamic model, further demonstrate the complexity and potential magnitude of the selective flow of small quantities of hydrocarbon fluids within natural gas transmission pipeline systems. 0.007 ppm PCB Q = 1.416 x 106 Sm3/d 0.17 ppm PCB P = 10.342 MPa abs. T = 15.6 OC G/L = 0.6/0.1 0.92 ppm PCB G/L = 0.5/0.96 G/L = 0.8/0.99 0.01 ppm PCB G/L = 0.72/0.72 2.35 ppm PCB G/L = 0.33/0.1 G/L = 0.4/0 0.54 ppm PCB 0.01 ppm PCB

Page 55: Sulphur

G/L = 0.2/0 2.67 ppm PCB 0.009 ppm PCB 0.009 ppm PCB

Figure 9.10 Preference Route of Liquid Phase in PCB Contaminated Natural Gas Transmission Pipeline

The coloured line in Figure 9.10 represents the preferred liquid route. Note how the PCB contamination levels also tend to reinforce this preferential flow selection criterion a pipeline “T” junctions. 10.0 Steps to Minimize the Deposition Threat Clearly it is just not practical, or economically feasible to eliminate all sources of sulphur and sulphur compounds, together with the identified contributing / enhancement factors, from a natural gas transmission pipeline system. Listed below are some means to minimize the ‘elemental sulphur’ formation / deposition process. The items are not presented in any ordered form.

When a large pressure reduction is required use a dual stage pressure cut. Due to the potential for temperature recovery between stages, a much lower temperature gradient at each stage, and lower retrograde condensation rate (if applicable) will decrease the overall nucleation, condensation and coagulation rates. This recommendation has now been implemented and proven.

Reduce the potential for retrograde condensation occurring at a pressure

reduction facility.

Minimize the entry and transmission of liquids and solid particles in the gas stream and in particular for the incoming gas supply to an affected facility.

Minimize the sources of water (moisture) and oxygen that can enter the gas

stream.

At commissioning of new pipelines ensure dewatering processes are complete and thorough. Do not permit ‘puddling’ of odorants or other like additives. Make sure pipe work is free of particle matter.

Ensure carry-over from glycol processing plants is kept to a minimum.

Review type of material used in molecular sieve beds does not favour the

conversion of H2S to COS.

Page 56: Sulphur

Minimize H2S levels.

Minimize site conditions suitable for the colonization and maintenance of SRB.

Try to minimize large temperature excursions along pipeline route.

Maintain flowing gas temperature as high as practically possible. This will not

only help maintain the sulphur in the gaseous form, and hence in solution, but will also assist in the minimization of retrograde condensation.

Care needs to be exercised in the interpretation of on-line and wellhead analysis

results. This may be due to sampling techniques, absorption of components by the sampling apparatus (tubing, valving and container), or due to the variations between sampling and analysis pressures (and temperatures).

1 Do not use labyrinth type valve cages at sites susceptible to the formation of

“elemental sulphur”. 11.0 Acknowledgements. The support of the Australian Pipeline Industry Association (APIA), many of its member companies and the Co-operative Research Centre (CRC) for Welded Structures is gratefully acknowledged. The support for, and belief in this research project (both the phase 1 and 2 studies) by the APIA Research and Standards Committee, and in particular Mr. M. Kimber and Mr. I. Haddow, is sincerely appreciated. Ongoing encouragement and support provided by a number of industry colleagues has been a major source of inspiration for this work. I am particularly indebted to Mr A. Chesnoy, of Paris, France, for his long time valued support and encouragement, without which this project would not have got of the ground. The hard work and dedication by Mr John Hamersley, (engineering honours student) in developing the CFD program and results for fluid mixing dynamics at pipeline “T” junctions has been of great value to this project. Finally, I would like to gratefully acknowledge a small group of ‘professionals’ within the industry, both within Australia and overseas, who willingly with patience and courtesy contacted and/or discussed issues of interest to the project with me.

Page 57: Sulphur

12.0 Recommendations for Further Research Work. Clearly, the topics studies for the phase 2 research program need further research. The areas recommended for further research and of value to pipeline operators, gas producers and large industrial gas consumers are: - Further understanding of the fluid dynamics at pressure reduction facilities and the impact of particular pressure regulation equipment design, - Determination of threshold levels for the nucleation processes, - Further modelling of the fluid dynamics at pipeline “T” junctions

Page 58: Sulphur

13.0 References. [1] Pack, D.J., Chesnoy, A.B., Bromly, J., White, R., 2000. Formation of Elemental

sulphur in Natural Gas Transmission Pipelines. The Australian Pipeliner. Number 100. January. pp 51-53

[2] Wu, C., Biswas, P., 1998. Particle Growth by Condensation in a System with

Limited Vapor. Aerosol Science and Technology. Vol 28. pp 1-20 [3] McGraw-Hill Encyclopedia of Science & Technology. 1997. 8th Edition, Vol 12. [4] Adewumi, M., Mucharam, L., 1990. Compositional Multiphase Hydrodynamic

Modelling of Gas / Gas-Condensate Dispersed Flow in Gas Pipelines. SPE Production Engineering, Paper No. 17076-PA, pp 85-90.

[5] Freidlander, S., Windeler, R., Weber, A., 1994. Ultrafine Particle Formation by

Aerosol Processes in Turbulent Jets: Mechanisms and Scale-up. NanoStructured Materials, Vol. 4, No. 5, pp 521-528

[6] Helfgen, B., Hils, P., Holzknecht, Ch., Turk, M., and Schaber, K., 2001.

Simulation of particle formation during the rapid expansion of supercritical solutions. Journal pf Aerosol Science. Vol 32. pp 295-319.

[7] Romay, F., Takagaki, S., Pui, D., Liu, B., 1998. Thermophoretic Deposition of

Aerosol Particles in Turbulent Pipe Flow. Journal of Aerosol Science. Vol. 29, No. 8, pp 943-959

[8] Lin, J., Tsia, C., Chang, C., 2004. Suppression of particle deposition in tube flow

by thermophoresis. Journal of Aerosol Science. Vol. 35, pp 1235-1250 [9] Owen, I., El-Kady, A.A., Cleaver, J.W., 1989. Deposition of Sub-Micrometre

Particles onto a Heated Surface with Widely Spaced Roughness Elements. Journal of Aerosol Science. Vol. 20, No. 6, pp 671-681

[10] Wilkes, C., Pareek, V., 1999. Sulfur Deposition in a Gas Turbine Natural Gas

Fuel Control System. ASME Joint Power Conference. San Francisco. [11] Luijten, C., van Hooy, R., Janssen, J., van Dongen, M., 1998. Multicomponent

nucleation and droplet growth in natural gas. Journal of Chemical Physics. Vol. 109, No. 9, pp 3553-3558.

[12] Luijten, C., 1998. op cit. [13] Luijten, C., 1998. op cit. [14] Fortuin, J., Hamersma, T., Hart, J., Smit, H., Baan, W., 1991. Calculations predict

condensate movement at ’T’ junctions. Oil & Gas Journal. Jan 21. pp 37-40

Page 59: Sulphur

[15] Pack, D.J., 2005. Elemental Sulphur Formation in Natural Gas Pipelines. PhD thesis. University of Western Australia.

[16] Oranje, L., 1973. Condensate Behavior in Gas Pipeline is Predictable. Oil & Gas

Journal. Vol. 73, No. 27, pp 39-44 [17] Penmatcha, V., Ashton, P., Shoham, O., 1996. Two-Phase Stratified Flow

Splitting at a T-Junction with an Inclined Branch Arm. International Journal of Multiphase Flow. Vol. 22, No. 6, pp 1105-1122.

[18] Ibid. [19] Ibid. [20] Chen, S., Ou, J., Dekat, A., Murthy, J., 1990. Dynamics of Fluid Mixing at a T-

Junction with Implications on Natural Gas Processing. Industrial & Engineering Chemistry Research. Vol. 29, No. 8, pp 1690-1695

[21] Azzopardi, B., Smith, P., 1992. Two-Phase Flow Split at T Junctions: Effect of

Side Arm Orientation and Downstream Geometry. Vol. 18, No. 6, pp 861-875 [22] Shoham, O., Brill, J., Taitel, Y., 1987. Two-Phase Flow Splitting in a Tee Junction

– Experiment and Modelling. Chemical Engineering Science. Vol. 42, No. 11, pp 2667-2676

[23] Meng, W., Chen, X., Kouba, G., Sarica, C., Brill, J., 2001. Experimental Study of

Low-Liquid-Loading Gas-Liquid Flow in Near-Horizontal Pipes. SPE Production & Facilities. Paper SPE 74687. pp 240-249

[24] Hart, J., Hamersma, P., Fortuin, J., 1989. Correlations Predicting Frictional

Pressure Drop and Liquid Holdup during Horizontal Gas-Liquid Pipe Flow with a Small Liquid Holdup. International Journal of Multiphase Flow. Vol. 15, No. 6. pp 947-964.

[25] Meng, W., et al. 2001. op cit. [26] Ibid. [27] Hart, J., et al. 1989. op cit. [28] Martinez, F., Adewumi, M., 1997. Two-phase Gas-Condensate Flow in Pipeline

Open-Network Systems. SPE Production & Facilities. Paper SPE 30995. [29] Ibid. [30] Oranje, L., 1973. op cit. [31] Martinez, F., et al 1997. op cit.

Page 60: Sulphur