oil recovery. Change in ... internal and external gas drive processes can be different: Egermann, P. ... permeability data based on external drive experiments!
Modelling of Tail Production by Optimizing Depressurization
Arne Skauge*, Dag Standnes, and Øystein Pettersen Centre for Integrated Petroleum Research, Univ. of Bergen Bergen, Norway Centre for Integrated Petroleum Research University of Bergen, Norway
Main effects of depressurization influencing oil recovery Change in fluid properties Change in fluid saturation Change in phase mobilities Change in rock properties
++
Centre for Integrated Petroleum Research University of Bergen, Norway
Main effects of depressurization Change in fluid properties Pressure decline will give:
- larger phase viscosities and densities - G/O capillary pressure will increase ( because gas-oil IFT is larger at lower pressure )
Change in fluid saturation Pressure decline will give:
- shrinkage of oil due to mass exchanges - gas expansion - immobile gas (Sgc)
Pressure below bubble point
Depletion rate (dP/Pb)/dt (1/hr)
0.1
High Pressure
0.01
Lower Pressure
0.001
0.0001 1
Centre for Integrated Petroleum Research 2
3
4
5
6
7
Critical Gas Saturation (%)
University of Bergen, Norway
Production below bubblepoint
General trend from literature data
Phenomenological events
higher drawdown rate
Pressure decline below saturation pressure
higher supersaturation
supersaturation nucleation treshold
more bubbles nucleated
nucleation
higher Sgc and Sgt
bubble formation bubble coalesence
0.1
continuous gas phase
Depletion rate (dP/Pb)/dt (1/hr)
boyancy movement of gas ganglia 0.01
0.001
0.0001 1
2
3
4
5
6
7
Critical Gas Saturation (%)
Centre for Integrated Petroleum Research University of Bergen, Norway
Example of an oil field with poor pressure communication and therefore high Sgcr around production Wells Sgc = [log (dP*/dt) – log C)] / D Example: C = 5.4x10-6 and D = 25.06 The pressure decline rate used to calculate the critical gas saturation is defined as pressure decline from time t1 (hours) to time t2 (hours) relative to Pb per hour : 250
dP*/dt = [ P(t1) - P(t2) ] / [ Pb x (t2-t1) ]
Using a variable Sgcr gives better match of production GOR Examples from matching GOR in 3 production wells
GOR (Sm3/Sm3)
200
150
measured varied Sgc Sgc=0.02 no Sgc
100
50
0 A
B
C
Centre for Integrated Petroleum Research University of Bergen, Norway
Source: Egermann et al, SCA 2000
Change in phase mobilities
Gas injection
Discontinous gas from pressure decline
Gas originating from depressurization has mobility that may be 2 orders of magnitude lower than injection gas Centre for Integrated Petroleum Research University of Bergen, Norway
Background Experimental data show that relative permeability from internal and external gas drive processes can be different:
Egermann, P., Vizika, O. (2000). A new Method to Determine Critical Gas saturation and Relative Permeability During Depressurisation in the Near-Wellbore Region. Soc. Core Analyst. Proceedings, Oct. 18-22, 2000.
Centre for Integrated Petroleum Research University of Bergen, Norway
Visualisation of network model results Many small pores, many bubbles, low coordination number:
Many large pores, few bubbles, high coordination number:
oil gas
gas oil
Gas phase is poorly connected ⇒ krg low & Sgc high
Gas phase is well connected ⇒ krg high & Sgc low Centre for Integrated Petroleum Research University of Bergen, Norway
Summary pore structure and rel perms 1)kro is controlled by coordination number rather than by the process, the pore size distribution or the bubble density 2)krg is sensitive to both coordination number, process type, pore size distribution and bubble density ¾ ¾
krg(external) ≈ krg(internal) in well connected network with many large pores and at low bubble density krg(external) >> krg(internal) in poorly connected network with many small pores and a high bubble density
Warning - Be careful about generating internal drive gas relative permeability data based on external drive experiments! ”Comparison of Relative Permeability Resulting From Internal and External gas Drive Processes – A Network Modelling Study” Susanne Poulsen1, Steven McDougall2, Ken Sorbie2 and Arne Skauge3 1. DTU, Denmark 2. Heriot-Watt U. UK 3. CIPR, UoB (Norsk Hydro) Norway SCA 2001
Centre for Integrated Petroleum Research University of Bergen, Norway
Can pressure depletion altered flow pattern caused by change in rock properties? If so, what are the condition needed to change the flow pattern?
Centre for Integrated Petroleum Research University of Bergen, Norway
Theory vs. experimental data • Young’s modulus vs. effective stress for typical sandstones 30
Young's modulus (GPa)
25
20
15
10
5
Rock A
Rock D
Rock B
Rock C
0 0
100
200
300
400
500
600
700
800
Effective stress (bars)
Centre for Integrated Petroleum Research University of Bergen, Norway
Permeability vs. load – measured data
Proposition: Permeability reduction is “proportional to” initial value, i.e. Medium is homogenized by loading WS: weak sandstone US: unconsolidated sand
6000
5000
4000
Permeability (mD)
Weak sands appear to have greater loss of conductivity than strong sands
GF A-23 test
GF A5H US
Texaco M, US
3000
Mobil U, US
Mobil E, WS
GF A5H WS
2000
1000
0 0
50
100
150
200
250
Eff. stress (bars)
Centre for Integrated Petroleum Research University of Bergen, Norway
Permeability vs. load – assumption log(K) linear w. σ, K(σ*) = 1 mD
10000
GF A-23 test
σ* = 1000 bars
WS: weak sandstone US: unconsolidated sand
Texaco M, US Mobil U, US
Permeability (mD)
fits GF WS
GF A5H US
1000
Mobil E, WS GF A5H WS Teor. 1
100
Teor. 2 Teor. 3 Teor. 4 Teor. 5
10
Teor. 6 Teor. 7
1 0
50
100
150
200
250
Eff. stress (bars)
Centre for Integrated Petroleum Research University of Bergen, Norway
Theory vs. experimental data • Permeability vs. effective stress 10000
ln k 0 ln k = − ⋅ σ'+ ln k 0 1000
1000
Permeability (m D)
Well tes t A
US B
US C 100
US D
WS E
WS F Theory 10
1 0
50
100
150
200
250
Eff. stress (bars)
Centre for Integrated Petroleum Research University of Bergen, Norway
Oil production
Water production
Centre for Integrated Petroleum Research University of Bergen, Norway
Remaining oil in place for different rock formations
Centre for Integrated Petroleum Research University of Bergen, Norway
Moderate channel perm, low background perm
Centre for Integrated Petroleum Research University of Bergen, Norway
Moderate channel perm, low background perm, Permeability reduction model σ* = 1000
Centre for Integrated Petroleum Research University of Bergen, Norway
Pressure distribution “some time” into depressurization, consequence for wells’ PI
Centre for Integrated Petroleum Research University of Bergen, Norway
Inferences – effects of compaction during depressurization + Reduces pressure reduction rate – energy supply + Reduces permeability contrasts – stops water cycling through high permeability channels - Reduces well productivity - Reduces general conductivity - May induce pore collapse / sand production ? Production scheme may be critical ? Late stage injection may be advantageous ? Stimulation, fracturing, gas lift may be necessary
Centre for Integrated Petroleum Research University of Bergen, Norway
Summary of change in rock properties • Empirical data show that permeability reduction due to increase in stress is stronger for high initial permeability • Selective permeability reduction may occur in reservoirs comprised of weak high-permeability soils in a background of stronger low-permeability soils This permeability homogenization may increase the sweep efficiency by reducing water cycling through high permeability layers
Centre for Integrated Petroleum Research University of Bergen, Norway
Guidelines for simulation of depressurization Capillary forces will vary with pressure decline, and may affect the buoyancy movement of gas Correct pVT data for the pressure decline range is needed The simulator has to handle gas relative permeabilities coming from different sources (influx versus internal generated gas in the grid block) Rock material with; - low coordination number - low permeability (more exact description - many small pores) - experiencing high depletion rate will have the largest demand for the complex gas rel perm model Sgcr should be coupled to grid block historical pressure decline rate The necessary geomechanical modelling (improved modelling of rock compaction) needs to be coupled to the reservoir simulator
Centre for Integrated Petroleum Research University of Bergen, Norway
