GB2429797A

GB2429797A - A pipeline control system

Info

Publication number: GB2429797A
Application number: GB0517667A
Authority: GB
Inventors: Ian Robertson
Original assignee: Genesis Oil and Gas Consultants Ltd
Current assignee: Genesis Oil and Gas Consultants Ltd
Priority date: 2005-08-31
Filing date: 2005-08-31
Publication date: 2007-03-07
Anticipated expiration: 2025-08-31
Also published as: GB0517667D0; GB2429797B

Abstract

A pipeline control system 20 comprises a control valve 26, at least one measurement means 28, 30, 34 enable to measure a parameter of the pipeline flow and provide at least one parameter output and a flow controller 36 enabled to vary the opening of the control valve 26. The flow controller 36 varies the opening of the control valve 26 according to an intended flow rate and the intended flow rate is varied in response to the at least one parameter output. The flow controller 36 includes a fuzzy logic controller 38 utilising fuzzy logic based on parameters such as flow rate (from flow meter 34) and differential pressure (from pressure sensors 28, 30) to define an intended flow rate. The fuzzy logic controller 38 may take the form of a computer program product on a digital computer. The system is particularly useful for preventing slug flow induced in "riser" portions of a pipeline.

Description

1 PIPELINE CONTROL SYSTEM 3 The present invention relates to a pipeline

control 4 system and particularly, but not especially, to a pipeline control system for use in oil well pipeline 6 risers.

8 A pipeline carrying a multiphase flow, substantially 9 consisting of varying quantities of oil, water and gas, from an oil well below the sea, typically, has 11 a riser portion to carry the multiphase flow from 12 the seabed to a separator on an offshore 13 installation. The movement of different phases 14 through a pipe at the same time, combined with the compressibility of gas, gives rise to different flow 16 regimes within the pipe. The flow regime of interest 17 is the slug flow regime Slug flow can occur due to 18 hydrodynamic pressures or through the terrain that a 19 pipe crosses.

21 1-lydrodynamic slugging is a function of the liquid 22 and gas velocities in the line. Typically, as the 1 fluid travels down the pipe, the pressure will drop 2 and the gas density decreases. The gas velocity will 3 therefore increase. As the velocity increases, it 4 will pick up waves on the surface of the liquid. If these waves are sufficiently high to block the pipe, 6 a hydrodynamic slug can be formed.

8 It is frequently not possible to operate the 9 pipeline outside the hydrodynamic flow regime without adversely affecting oil production.

12 Terrain induced slugging is where there is a dip in 13 the line and liquid accumulates at the bottom of the 14 dip. The level will rise in the dip until either the line is blocked, or the available pipe bore is 16 reduced so that the gas velocity creates waves and 17 hence hydrodynamic slugs.

19 Hence a hydrodynamic slug can be the trigger that "kicks" a large terrain slug out of the dip in the 21 pipe into the downstream process.

23 During the flow from the seabed to the surface 24 within a riser portion, the multiphase flow may form slug flow. As shown in Fig. 1, slug flow in the 26 riser is caused when a liquid slug blocks the bottom 27 section. Gas then backs up behind the slug until the 28 slug is drawn further up the pipe.

Typically, severe riser based slugging occurs when a 31 pipeline slopes down gently to the base of a riser 32 and the rate of hydrostatic head build-up is greater 1 than the rate at which gas pressure accumulates 2 upstream of the blockage. Liquid starts to 3 accumulate, until, there is sufficient pressure 4 built-up upstream of the liquid plug to overcome the hydrostatic pressure and force the slug up the 6 riser. Once the liquid slug is produced, there is 7 then a gas surge into the topside plant equipment.

9 Riser based slugging can have an adverse effect on production as large quantities of liquid and gas are 11 alternately produced into the system within a short 12 time span. These quantities can be larger than the 13 available fluid processing capacity and can lead to 14 production problems including flaring or poor oil- and-water separation.

17 Typically, riser based slugging has been reduced by 18 control of a topside control valve. Plow controllers 19 are known which vary the topside control valve in response to varying flow measurements. For example, 21 an intended flow rate is chosen, which is known as 22 the set point (SP) . If the measured flow rate is 23 greater then the SP the opening of the topside 24 control valve is reduced. If the measured flow rate is less then the SP the opening of the topside 26 control valve is increased.

28 In practice, constantly varying individual flow 29 rates of the different phases of the multiphase flow require that the set point is varied, so that 31 production can be maximised.

i US 6,595,294 discloses a system for stabilisation of 2 gas lifted oil wells based on a model of the system 3 and directed towards control of the topside control 4 valve and a gas injection valve.

6 According to a first aspect of the present invention 7 there is provided a pipeline control system 8 comprising: 9 a control valve; TO at least one measurement means enabled to 11 measure a parameter of the pipeline flow and provide 12 at least one parameter output; and 13 a flow controller enabled to vary the opening 14 of the control valve, wherein the flow controller varies the opening 16 of the control valve according to an intended flow 17 rate and the intended flow rate is varied dependent 18 on the at least one parameter output.

Preferably, the flow controller comprises a fuzzy 21 logic means enabled to output the intended flow rate 22 dependent on the at least one parameter output.

24 Preferably, the at least one measurement means comprises a flow meter or other device for measuring 26 the flow rate, the flow rate being used as a 27 measured parameter.

29 Preferably, the at least one measurement means comprises a differential pressure means, wherein the 31 differential pressure across the control valve is 1 measured, the differential pressure being used as a 2 measured parameter.

4 Preferably, the fuzzy logic means varies the intended flow rate dependent on the measure 6 differential pressure value and the measured 7 flowrate value.

9 Preferably, the pipeline control system controls a riser portion of a pipeline.

12 Preferably, the at least one measurement means and 13 the control valve are located downstream of the 14 riser portion. Typically, a riser is used from a seabed to an oil platform and downstream of the 16 riser portion is located "topside" on the oil 17 platform.

19 According to a second aspect of the present invention there is provided a method of controlling 21 a pipeline comprising the steps of: 22 (i) receiving from at least one measurement 23 means a parameter of the pipeline flow and provide 24 at least one parameter output; and (ii) controlling in response to the at least 26 one parameter output a control valve, 27 wherein sinp (ii) varies the opening of the 28 control valve according to an intended fLow rate and 29 the intended flow rate is varied dependent on the at least one parameter output.

1 According to a third aspect of the present invention 2 there is provided a computer program product 3 directly loadabiLe into the internal memory of a 4 digital computer comprising software code portions for performing the method of the second aspect of 6 the present invention.

8 Embodiments of the present invention will now be 9 described with reference to the accompanying drawings, in which; 12 Pig. 1 illustrates build of a liquid slug in a 13 pipeline riser; Pig. 2 illustrates a riser control system according 16 to the present invention; 18 Fig. 3 illustrates a membership function of fuzzy 19 possibilities for differential pressure; 21 Fig. 4 illustrates a membership function of fuzzy 22 possibilities for flow; and 24 Fig. 5 illustrates a membership function of fuzzy possibilities for intended flow rate.

27 Referring to Pig. 1, as explained above, riser based 28 slugging occurs when a liquid blockage occurs at the 29 riser base 1. The liquid blockage grows 2 until there is enough hydrostatic force to more the slug 31 up the riser. Liquid production 3 initially occurs 32 allowing the pressure behind the slug to move the 1 remainder of the slug at a faster rate, giving fast 2 liquid production 4. The gas, which has built up 3 behind the slug, is then released producing a gas 4 blow down 5.

6 Referring to Fig. 2, a pipeline control system 20 7 has an input from a pipeline 22 having a riser 24.

8 The control system 20 has a topside control valve 26 9 positioned after the riser 24. A first pressure sensor 28 is situated on the upstream side of the 11 control valve 26 and a second pressure sensor 30 on 12 the downstream side of the control valve 26. A 13 differential pressure means 32 receives inputs from 14 the first and second pressure sensors 28, 30. A flow meter 34 is situated downstream of the control valve 16 26. A flow controller 36, which includes a fuzzy 17 logic controller 38, receives inputs from the 18 differential pressure means 32 and the flow meter 19 34. The controller 36 has a control output 40 which operates the control valve 26.

22 It should be appreciated that although, in this 23 example, the measurements taken are differential 24 pressure and flow rate, the control system may be adapted to receive other measurements which describe 26 the current state of the pipeline flow.

28 In the example in Fig. 2, the controller 36 has to 29 achieve the following tasks: 1. If the flow measured at the flow meter 34 is 31 too low (i.e. not enough production is coming 1 out of the pipeline) , the intended flow rate, 2 or flow setpoint (FSP), needs to be increased.

3 2. If The pressure drop across the valve is too 4 high (i.e. the pipeline is operating at too high a pressure), the setpoint needs to be 6 increased to depressurise The pipeline.

7 3. If the pressure drop across the valve is 8 dropping too low (i.e. the pipeline has now 9 been depressurised to its desired operating point), the setpoint needs to be reduced.

12 In this case, there are two variables which need to 13 be controlled, the differential pressure (P) across 14 the control valve 26 and the volumetric flow (F) These values must be controlled only by the control 16 valve 26 position.

18 It is possible for a cascade arrangement with the AP 19 loop as the outer master controller and the flow loop as the inner slave controller to allow both 21 flow and AP to be adjusted by a single valve.

22 However, since both the master and slave controllers 23 will be immediately affected by the valve position, 24 a conventional cascade arrangement is not suitable -- in practice such a control arrangement would not 26 reach a steady sLate value and continually "hunt".

28 The controller 36 must decide on the required 29 control valve FSP based on a knowledge of the current valve AP and the flow rate.

1 The Fuzzy logic means 38 is used to perform this 2 decision making process. The concept of fuzzy logic 3 is based on fuzzy set theory. Fuzzy sets attempt to 4 deal with the imprecision, uncertainty and vagueness In a manner analogous to how humans deal with the 6 same. Unlike conventional set theory which requires 7 an element to either belong or not belong to a set, 8 fuzzy sets also allow for an element to have partial 9 membership or a degree of belonging to more than one set.

12 A fuzzy set, A, on a universe of discourse, S, is 13 defined as a membership function mA(x) which is a 14 mapping from the universe S onto the unit interval: tA(x) :S 40,1] 17 There are a number of differently shaped membership 18 functions that may be used to define a fuzzy set in 19 the continuous domain. Figures 3, 4 and 5 define the membership functions used in this example.

22 In this case, the following nine rules were set-up 23 for the Fuzzy Logic.

1. If control valve P is low and the flow rate 26 is tow, then FSP needs to be High.

27 2. It valve P is low and the flow rate is 28 normal, then FSP needs to be low.

29 3. If valve P is low and the flow rate is high, then FSP needs to be Low.

31 4. If valve AP is normal and the flow rate is 32 low, then FSP needs to be Normal.

1 5. If valve P is normal and the flow rate is 2 normal, then FSP needs to be Normal.

3 6. If valve P is norma and the flow rate is 4 high, then FSP needs to be Normal.

7. If valve t2P is High and the flow rate is low, 6 then FSP needs to be normal.

7 8. If valve P is High and the flow rate is 8 normal, then FSP needs to be high.

9 9. If valve AP is High and the flow rate is High, then FSP needs to be low.

12 For the fuzzy logic to work, it needs to consider if 13 the flow rates and APs are low, normal or high.

The fuzzy logic means requires information relating 16 to what flow rates or APs would be considered low, 17 normal or high. This information is dependent on the 18 particular pipeline of interest and can be set by 19 simulation or during a pre-commissioning trial prior to full operation.

22 For example, as shown in the differential pressure 23 membership function Fig.3, "low" APs are on a 24 sliding scale up to 2.5 bar, normal APs are on a sliding scale from 1. 0 to 4.0 bar and high AP5 are 26 on a sliding scale from 2.5 to 4.0 bar and above.

28 In Fig. 3, the vertical axis ji(AP) is a measure of 29 how "low" or how "normal" the fuzzy logic considers the AP to be.

1 It should be appreciated that in Figures 3, 4 and 5, 2 = "possibility", (in the range 0 to 1) . This is a 3 relationship used for the fuzzy logic only, and is 4 not related to probabilities or likelihoods.

6 The values for the horizontal axis are defined from 7 the pre-comrnissioning trial or from simulation of 8 the pipeline.

Referring to Fig. 3, there are three "fuzzy set U centres" for the differential pressure measurement.

12 In this case, a fuzzy set centre defines the values 13 of P which are nominally low, normal and high.

High is the value obtained from the pre- 16 commissioning trial or simulation, with the control 17 valve 26 gagged in, or substantially closed, to 18 eliminate slugging. For example, in one pre- 19 commissioning trial the high value was 4 bar with the valve closed in to 40% open.

22 Low is the value at which control may be lost and is 23 the minimum P across the control valve 26 for the 24 controller 36 to operate fully. This is typically 1 to 2 bar.

27 The normal P may be an arbitrary value, typically 28 mid-way between the high and low Ps defined above.

Referring to Fig. 4, the vertical axis 1j(F) is a 31 measure of the fuzzy logic probability for the 1 "fuzzy set centres" for the flow rate measurement 2 and the horizontal axis is flow rate.

4 In this case, low was set below the stabilised flow rate measured or simulated during the pre- 6 commissioning phase. For example, in a previous pre- 7 commissioning trial, with the control valve 26 at 8 40% open, the system stabilised with a flow rate of 9 360 m3/h. The control valve 26 pressure drop was 4 bar. Therefore low fuzzy set centre was taken as 350 11 m3/h, slightly below the stabilised flow value of 12 360 m3/h. The normal flow setpoint was set at 10% 13 above this value at 400 m3/h. The high setpoint was 14 set at another 10% higher up, rounded to 450 m3/h

for this example.

17 Referring to Fig. 5, the vertical axis p(FSP) is a 18 measure of the fuzzy logic probability for the 19 "fuzzy set centres" for the required overall production from the system and the horizontal axis 21 is flow rate.

23 In this case, the range of flow rates will have a 24 smaller spread than in Fig. 4 for the flow rate fuzzy set centres. This is because the controller 36 26 has to receive measurements and react to peaks in 27 flow rate but we do not want to set the flow 28 controller 36 to introduce high peaks into the 29 system.

31 In this example, the low FSP fuzzy set centre has 32 been set as with the 1.ow flow rate fuzzy set centre 1 (350 m3/h) . The normal FSP fuzzy set centre was set 2 at the steady state flow rate expected from the 3 pipeline once the system is stabilised. For example, 4 the valve AP from the pre-commissioning trial is 4 bar and the final steady state valve AP is 1 bar, 6 hence the reduction in pipeline back-pressure is 3 7 bar. Using an approximation, 1% more production for 8 each 1 bar drop in back-pressure giving a rounded 9 estimate for the normal fuzzy set centre of 375 m3/h. The high fuzzy set centre may be defined as 11 the maximum rate at which the pipeline could be 12 depressurised. 10% above the pre-commissioning trial 13 flow rate is may be a reasonable estimate (400 m3/h 14 in this case) 16 In all the above cases, it must be emphasised that 17 these are fuzzy set centres and not exact flow 18 setpoints. As long as the fuzzy set centres are 19 carefully chosen, the fuzzy logic will compensate and correct any errors accordingly.

22 The fuzzy logic means 38 will now have the 23 information required to process measurement. When 24 the system is operational, the first step in the fuzzy logic means is to identify which of the nine 26 rules identified above apply.

27 For example, if the measured values at the start 28 are: flow rate = 380 m3/h, valve AP = 1.8 bar, then, 29 according to Fig. 3 and Fig. 4, the flow could be viewed as low and/or as normal and the AP could be 31 viewed as low and/or as normal.

1 So, applying the fuzzy possibilities rules: 2 From Fig. 3: 3 Possibility of AP being low = 0.4667 4 Possibility of P being normal = 0.5333 Possibility of P being high = 0 7 From Fig. 4: 8 Possibility of flow rate being low = 0.4 9 Possibility of flow rate being normal = 0.6 Possibility of flow rate being high = 0 12 The following fuzzy rules therefore apply: 13 1. If valve AP is low and the flow rate is low, then 14 the FSP needs to be High.

2. If valve AP is low and the flow rate is normal, 16 then FSP needs to be Low.

17 4. If valve AP is normal and the flow rate is low, 18 then FSP needs to be Normal.

19 5. If valve AP is normal and the flow rate is normal, then FSP needs to be Normal.

22 The possibilities of these rules applying, is 23 calculated from the minima of the possibilities of 24 the APs and flow rates.

26 In the above example, 27 Possibility rule 1 applying = minimum of 0.4667 and 28 0.4 = 0.4.

29 Posshility rule 2 applying = minimum of 0.4667 and 0.6 = 0.4667.

31 Possibility rule 4 applying = minimum of 0.5333 and 32 0.4 = 0.4.

1 Possibility rule 5 applying minimum of 0.5333 and 2 0.6 = 0.5333.

4 The new FSP is calculated from the possibilities of the rules applying multiplied by the fuzzy set 6 centre of the FSP for that rule. In the above 7 example, the possibility of rule 1 appLying was 0.4.

8 Rule 1 states that if the valve AP is low and the 9 flow rate is low, then the FSP needs to be High. The FSP High set centre was 400 m3/h, so this equates 11 to: 0.4 X 400 = 160 m3/h.

13 This is applied to all the rules and then normalised 14 (divided by the sum of all the possibilities of the rules that apply) . In the above example, this works 16 out as: 0.4x 400+ 0.4667x 350+ 0.4x 375 + 0.5333 x 375 17 F. = ________________- =374.07m3/h 0.4 + 0.4667 + 0.4 + 0.5333 19 As explained above, the fuzzy logic changes the FSP of the controller 36, the controller 36 then 21 maintains that flow rate through operation of the 22 control valve 26.

24 The PSP should preferably not be changed too frequently. In a long pipeline, there will be a 26 time-lag between the control valve 26 at the riser 27 changing and the impact on the flow conditions in 28 the upstream part of the pipeline. If the FSP is 29 changed before the upstream part of the pipeline has adjusted to the new conditions, the pipeline could 31 end up in an unstable condition.

2 The minimum update period for the fuzzy logic to 3 change the fLow setpoint can be calculated from: Length of pipeline Update period = Speed - of - sound - in fluid 7 The invention as described above may use other 8 measurements other than flow rate and differential 9 pressure. Although these measurements are currently the most significant for detecting rier based slug 11 flOW, it may be possible to derive indications of 12 situations indicated by these measurements through 13 other means.

Furthermore, the application of this invention is 16 not restricted to riser based slugging but to any 17 application where a balance must be achieved between 18 flow rate through a pipeline and the pipeline 19 conditions.

21 Improvements and modifications may be incorporated 22 without deparLing from the scope of the present 23 invention. 1*7

Claims

1 CLAIMS 3 1. A pipeline control system comprising: 4 a control valve; at

least one measurement means enabled to 6 measure a parameter of the pipeline flow and 7 provide at least one parameter output; and 8 a flow controller enabled to vary the opening of 9 the control valve, wherein the flow controller varies the opening 11 of the control valve according to an intended flow 12 rate and the intended flow rate is varied 13 dependent on the at least one parameter output.

2. A system as claimed in claim 1, wherein the flow 16 controller comprises a fuzzy logic means enabled 17 to output the intended flow rate dependent on the 18 at least one parameter output.

3. A system as claimed in claim I or claim 2, 21 wherein the at least one measurement means 22 comprises a flow meter or other device for 23 measuring the flow rate, the flow rate being used 24 as a measured parameter.

26

4. A system as claimed in any preceding claim, 27 wherein the at least one measurement means 28 comprises a differential pressure means, wherein 29 the differential pressure across the control valve is measured, the differential pressure being used 31 as a measured parameter.

1

5. A system as claimed in claim 4, wherein the 2 fuzzy logic means varies the i.ntended flow rate 3 dependent on the measure differential pressure 4 value and the measured flowrate value.

6 6. A system as claimed in any preceding claim, 7 wherein the pipeline control system controls a 8 riser portion of a pipeline.

7. A system as claimed in claim 6, wherein the at 11 least one measurement means and the control valve 12 are located downstream of the riser portion.

13 Typically, a riser is used from a seabed to an oil 14 platform and downstream of the riser portion is located "topside" on the oil platform.

17 8. A method of controlling a pipeline comprising 18 the steps of: 19 (i) receiving from at least one measurement means a parameter of the pipeline flow arid 21 provide at least one parameter output; and 22 (ii) controlling in response to the at least one 23 parameter output a control valve, 24 wherein step (ii) varies the opening of the control valve according to an intended flow rate 26 and the intended flow rate is varied dependent on 27 th at least one parameter output.

29 9. A computer program product directly loadable into the internal memory of a digital computer 31 comprising software code portions for performing 32 the method of claim 9.

Amendments to the claims have been filed as follows 1 CLAIMS 3 1. A pipeline control system comprising: 4 a control valve; at least one measurement means enabled to 6 measure a parameter of the pipeline flow and 7 provide at least one parameter output; and 8 a flow controller enabled to vary the opening of 9 the control valve, wherein the flow controller varies the opening 11 of the control valve according to an intended flow 12 rate and the intended flow rate is varied 13 dependent on the at least one parameter output.

3. A system as claimed in claim 1 or claim 2, 21 wherein the at least one measurement means 22 comprises a flow meter for measuring the flow 23 rate, the flow rate being used as a measured 24 parameter.

26 4. A system as claimed in any preceding claim, 27 wherein the at least one measurement means 28 comprises a differential pressure means, wherein 29 the differential pressure across the control valve is measured, the differential pressure being used 31 as a measured parameter.

1 5. A system as claimed in claim 4, wherein the 2 fuzzy logic means varies the intended flow rate 3 dependent on the output value of differential 4 pressure means and the output value of the flow meter.

7 6. A system as claimed in any preceding claim,

8 wherein the pipeline control system controls a 9 riser portion of a pipeline.

11 7. A system as claimed in claim 6, wherein the at 12 least one measurement means and the control valve 13 are located downstream of the riser portion.

14 Typically, a riser is used from a seabed to an oil platform and downstream of the riser portion is 16 located "topside" on the oil platform.

18 8. A method of controlling a pipeline comprising 19 the steps of: (1) receiving from at least one measurement 21 means a parameter of the pipeline flow and 22 provide at least one parameter output; and 23 (ii) controlling in response to the at least one 24 parameter output a control valve, wherein step (ii) varies the opening of the 26 control valve according to an intended flow rate 27 and the intended flow rate is varied dependent on 28 the at least one parameter output.

9. A computer program product directly loadable 31 into the internal memory of a digital computer

LI

1 comprising software code portions for performing 2 the method of claim 9.