SPE-174935-MS

Offshore Membrane Deaeration as a Replacement for Vacuum Tower

Deaeration - A Comparative Study
Stephen Van Pelt, and Holly Churman, Water Standard

Copyright 2015, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in Houston, Texas, USA, 28 –30 September 2015.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
A membrane deaeration pilot study undertaken on seawater from an onshore seawater intake, successfully
demonstrated removal of dissolved oxygen to less than 10 parts per billion without the use of chemical
oxygen scavenger. To assess the commercial viability of adapting membrane deaeration technology to
offshore oil and gas applications, a comparative evaluation of size, weight, capital cost, and operating cost
for membrane deaeration versus conventional vacuum tower deaeration was performed.
The comparative evaluation considered membrane and vacuum tower deaeration technologies across a
range of injection flow rates, from 30,000 – 250,000 barrels per day. Multidisciplinary engineering and
design was conducted for each technology at specific injection flow rates within this range. Detailed 3D
models, material take-offs, and weight reports were produced for each design. The capital cost study was
prepared, with the goal to estimate direct field cost for installation of each technology. These costs were
derived using quotes from vendors and installed cost norms for similar offshore installations. Operating
costs were developed using quotes for consumables and estimations for power cost.
The comparative evaluation indicated that selection of membrane deaeration technology over conven-
tional vacuum towers resulted in the following offshore facilities impacts:
● 60 – 70% dry and operating weight savings across the entire range of injection flow rates studied.
● 30 – 60% total size savings, with increased savings at elevated injection flow rates.
● 10% footprint savings from 125,000 – 250,000 barrels per day. The footprint of membrane
deaeration is larger at small flow rates.
Selection of membrane deaeration technology over conventional vacuum towers results in the follow-
ing offshore economic impacts:
● 10 – 15% direct field cost savings from 30,000 – 125,000 barrels per day. Direct field cost savings
between 5 – 10% from 125,000 – 250,000 barrels per day, as economies of scale of vacuum towers
begin to take effect.
● 50 – 60% annual operating cost savings across the entire range of injection flow rates studied,
based partially on an assumed cost per kilowatt-hour for offshore power generation.
This study accounts for parameters which have not been addressed in previous literature to date:
2 SPE-174935-MS

● Costs related to the installation of each deaeration technology on offshore topsides

● Impacts on the entire topsides process due to selection of each deaeration technology (i.e.,
injection booster pumps, chemical injection facilities, etc.)

Introduction
For many years, offshore oil and gas operators have injected seawater into new and aging reservoirs to
improve oil recovery and offset production declines. In these waterflooding applications, removal of
dissolved oxygen from seawater has become an essential unit process to mitigate corrosion and biogrowth
potential, and subsequently preserve the integrity of water injection infrastructure and of the reservoir.
Vacuum towers have been widely used to achieve these objectives. However, vacuum towers are
suboptimal for several reasons: large size and weight, stability impacts from sea-state effects, inability to
treat to desired dissolved oxygen levels, propensity for biogrowth in the tower internals, the need for
injection booster pumps, and the need for large capacity vacuum pumps.
Membrane deaeration (MDA) technology offers an opportunity to overcome these challenges. Widely
used in boiler feed water, electronics, and other industries, MDA has only recently been tested for use in
seawater service, thereby extending its applicability to offshore oil and gas use. A pilot test conducted in
2014 validated the technical feasibility of MDA technology for the removal of dissolved oxygen from
seawater. Using these results, a comparative evaluation of size, weight, capital cost, and operating cost for
MDA versus vacuum tower deaeration was performed.

Membrane Deaeration Technology Background & Piloting

Membrane Deaeration Technology Overview

Membrane contactors are used to transfer dissolved oxygen from seawater efficiently from within a small
weight and footprint configuration. Membrane contactors are comprised of thousands of microporous
polypropylene hollow fibers knitted into an array and wound around a center tube. Water flows through
the inside of the hollow fiber membranes, while high purity nitrogen sweep gas flows outside of them in
a countercurrent directionality. As the fibers are hydrophobic, seawater does not penetrate the membrane
pores. Instead, the dissolved oxygen transfer between the seawater and the high purity nitrogen sweep gas
is governed by the difference in partial pressure of the oxygen gas between the two phases, according to
Henry’s Law (Wiesler, 1996)
Because the degassing surface area per unit volume is an order of magnitude higher than vacuum
towers, the greater surface area to volume ratio enables a reduction in system size and an improved level
of performance. Offshore, vacuum towers are typically able to achieve dissolved oxygen removal to below
50 parts per billion (ppb) as a standalone unit (Carlberg, 1976). They must employ oxygen scavenger
chemical, however, to achieve dissolved oxygen specifications of 10 ppb.
SPE-174935-MS 3

Figure 1—Membrane Contactor (Membrana, 2014)

Pilot Experimental Setup

A 5,000 bpd MDA pilot unit was designed, constructed, and installed in a marine facility in Florida, USA
in 2014 to assess the viability of the technology in seawater service. Seawater was drawn through a beach
well intake system on site, pretreated using cartridge filters (as needed), and transferred to the MDA
system for dissolved oxygen treatment. Water was treated through one of two membrane contactors. A
standby contactor installed in parallel provided redundancy in preparation for fouling or other potential
process upsets. High purity nitrogen sweep gas was injected counter-currently using compressed nitrogen
bottles. A vacuum pump was used to impart vacuum pressures on the gas side the membrane contactor.
A clean in place (CIP) unit comprised of a chemical feed tank and pump system was installed to remedy
biofouling events. Testing was divided into two phases: a performance validation, and a feasibility
assessment.

Figure 2—MDA Pilot Skid

4 SPE-174935-MS

Figure 3—MDA Technology Flow Diagram

Phase 1 – Performance Validation To extend the applicability of MDA to seawater service, dissolved
oxygen removal performance was tested at a variety of operating conditions. Key operating variables
included nitrogen flow rates, feed water flow rates, and vacuum pressure.
In general, as nitrogen flow increases, dissolved oxygen removal performance improves. Dissolved
oxygen removal performance was therefore tested between 0.85 and 1.7 normal cubic meters per hour
(Nm3/hr). If results indicated that dissolved oxygen removal was relatively insensitive to nitrogen flow
rates, then the technology could potentially be scaled up using smaller nitrogen generators, with lower
subsequent capital and operating costs.
As feed water flow rate increases, dissolved oxygen removal performance tends to decrease, as a
function of residence time within the membrane contactors; with less residence time, gas-water contact
time decreases. Feed water flow rate was tested between 13.6 and 25.0 cubic meters per hour (m3/hr) per
membrane contactor. If results indicated satisfactory oxygen removal performance at elevated flow rates,
then the technology could be designed with less membrane area for full scale facilities, with lower capital
costs.
As vacuum pressure increases, dissolved oxygen removal performance tends to increase, as a function
of Henry’s Law. Vacuum pressures between 50 – 75 torr were tested, according to manufacturer
recommendations. If results indicated that dissolved oxygen removal performance was adequate at lower
vacuum pressures, then the technology could be scaled up using smaller vacuum pumps for full scale
facilities. Space, weight, capital and operating cost requirements would be subsequently reduced.
Inlet oxygen concentration, seawater temperature, nitrogen sweep gas flow rate, seawater feed flow
rate, and vacuum pressure were tested as independent variables. Dependent variables included outlet
dissolved oxygen concentration and membrane differential pressure. All variables were measured with
both online and manual instrumentation. Fifteen experiments were conducted, each for a minimum of 10
hours. Dependent variables were compared to manufacturer projections for further validation.
Phase 2 – Feasibility Assessment To further extend the applicability of MDA to seawater service,
dissolved oxygen removal performance and membrane differential pressure were tested at varying,
controlled degrees of fouling potential. The key operating variable was the degree of prefiltration applied
upstream of the membrane contactors.
As the degree of filtration decreases, levels of suspended solids and organic foulants likely to reach the
membrane contactors increases. Levels of filtration of 1 micron and 20 micron were tested. If results
indicated that dissolved oxygen removal performance was sustained without dramatically increasing
differential pressure (an indicator of fouling), full-scale designs using lower degrees of filtration could be
used, thereby saving capital and operating cost.
Fouling variables, including total suspended solids, turbidity, and total organic carbon were closely
monitored. Various chemical cleaning options were available to address fouling, and the cleaning regimes
were evaluated for effectiveness. Inlet oxygen concentration, seawater temperature, and degrees of
SPE-174935-MS 5

filtration were tested as independent variables, while outlet dissolved oxygen concentration and membrane
differential pressure were monitored as dependent variables.
Pilot Results and Implications for Comparative Study
A sample of results from Phase 1 testing are indicated below.

Figure 4 —MDA Phase 1 – Key Results

Membrane contactor performance was validated against manufacturer projections in seawater service
at water flow rates between 13.6 and 18.2 m3/hr. The MDA pilot removed dissolved oxygen below 10 ppb
without chemical oxygen scavenger throughout this range, and aligned with upper and lower projection
bounds. However, between 22.7 and 25.0 m3/hr of water flow, actual outlet dissolved oxygen concen-
trations were consistently higher than manufacturer projections by approximately 1 to 6 ppb. Therefore,
the projections underestimated oxygen removal performance losses at water flow rates above 22.7 m3/hr.
At 1 micron pre-filtration, the data showed no relative loss in expected percentage of oxygen removal
performance throughout most experimental runs. The contactors did not experience an increase in
differential pressure throughout Phase 1, indicating that particulate or biological fouling did not occur.
Therefore, 1 micron cartridge filter pre-filtration was deemed sufficient to sustain performance for
membrane contactors in seawater applications.
Scaling could take place in the membrane contactor due to removal of carbon dioxide gas from the
seawater, leading to an increase in pH. Data indicated that pH did not increase by more than 0.1
throughout testing. Coupling this finding with consistent dissolved oxygen removal performance without
increases in differential pressure, membrane contactors were determined to be applicable in seawater
without significant scaling concerns.
Nitrogen sweep gas flow tests indicate that there is a point of diminishing returns when rates were
greater than 1.3 Nm3/hr. Above this point, only slight improvements in oxygen removal performance were
observed. If nitrogen supply is scarce, then it is possible to reduce nitrogen consumption to 0.43 Nm3/hr
per contactor and make only small sacrifices in oxygen removal performance.
A sample of results from Phase 2 testing are indicated below.
6 SPE-174935-MS

Figure 5—Key Phase 2 Results

Phase 2 testing implemented 20 micron cartridge filter pre-filtration. During operation, dissolved
oxygen removal performance gradually worsened, from 7 ppb to 30 ppb, over 250 hours of run time. A
caustic/acid CIP fully restored performance to 7 ppb.
Data indicates that, using 20 micron cartridge filter pre-filtration, membrane contactors are subject to
performance losses as a function of seawater quality. Identification of red-orange deposits in the cartridge
filters during an inspection illustrated that iron was present in the feed water, and likely a contributor to
fouling. Successful cleaning showed that acid/caustic CIPs are effective methods for maintaining con-
tactor performance in seawater service.
A build-up of condensation on the fibers of the membrane contactors occurred during Phase 2, and
remained for 24 hours of testing. This resulted in a short-term rise in outlet dissolved oxygen concen-
tration by 10 ppb. A short purge with nitrogen sweep gas removed condensation on the fibers of the
contactors. Purging with nitrogen sweep gas was found to be effective for removing condensation on the
fibers.
Based on the various levels of pre-treatment provided throughout Phases 1 and 2 of testing, it was
concluded that 1 micron pre-filtration should be provided prior to membrane contactors in MDA units.
Further long-term testing should be done to validate whether 5 micron filtration is acceptable.

Membrane Deaeration vs. Vacuum Tower Comparative Study

Overview and Technology Comparison
The results of the MDA pilot were used to perform a comparative evaluation of size, weight, capital cost,
and operating cost for MDA versus conventional vacuum tower deaeration. Through piloting, the MDA
design basis was validated. Key MDA design parameters for the comparative study include seawater
flowrate per membrane contactor, nitrogen sweep gas purity, nitrogen sweep gas flow rate, and vacuum
pump operating pressure. The design basis for vacuum tower deaeration is already well established due
to the maturity of the technology; thus, a standard offshore design basis was assumed using several past
offshore projects as references. A general comparison of each of the technologies is provided in below.
SPE-174935-MS 7

Table 1—Vacuum Tower vs. Membrane Deaeration Technology Comparison

Parameter Vacuum Tower Membrane Deaeration

Vacuum package Approx. 15 torr vacuum pressure. Requires 50 torr vacuum pressure. Smaller pumps with
large pumps and ejectors. 50% less power required and no ejectors
required.
Oxygen removal performance 50 – 100 ppb mechanically ⬍ 10 ppb mechanically
Chemical oxygen scavenger Required downstream to meet oxygen Not normally required.
specification.
Anti-foam chemical Required upstream of tower inlet. Not required.
Biocide chemical Dedicated biocides required to control growth Biocides not normally required.
in tower internals.
Injection booster pumps Required due to low pressure operation of Can be eliminated due to pressurized operation.
vacuum tower.
Life cycle consumables No regular consumable costs. Requires membrane replacement every 3 – 7
years, depending on pre-treatment.
Technology readiness level Field proven offshore for many years – Field proven in onshore, industrial applications,
(API RP 17N, 2009) technology readiness level (TRL) 7. but not offshore - TRL 5.
Operability and maintainability Long operational history offshore with Requires periodic chemical cleaning to maintain
established maintenance procedures. oxygen removal performance. Limited
operational history offshore.

Size and Weight Study Methodology

The size and weight evaluation considered membrane and vacuum tower deaeration technologies at three
injection flow rates: 30,000, 125,000, and 250,000 bpd. In order to perform a complete comparison, the
impacts on the entire topsides process due to selection of each deaeration technology were considered (i.e.
injection booster pumps, chemical injection facilities, utility requirements). Multidisciplinary engineering
and design was conducted for each technology at these three injection flow rates. 3D models, material
take-offs, and weight reports were produced for each design. The engineering scope included process
design, mechanical and electrical equipment selection, piping design and layout, and structural analysis.
The scope of MDA technology that was considered in the analysis included a membrane skid with
pressure vessels and membrane elements; a vacuum skid with liquid ring vacuum pumps and separator;
oxygen scavenger chemical metering skid with tote tank and metering pumps; CIP skid with tank,
cartridge filter, and pumps; nitrogen generation skid with pressure swing adsorption nitrogen generator
and receiver. All instrumentation, controls, piping, valves, and structural steel were included to make the
system a fully functional offshore module. A MDA module scope of supply summary is below.
The scope of vacuum tower deaeration technology considered in the analysis included a vacuum tower
with internals and access ladders/platforms; a vacuum skid with liquid ring vacuum pumps, separator, and
ejectors; chemical metering skid with tote tanks & metering pumps for oxygen scavenger, biocide, and
anti-foam; high pressure injection booster pump skid. All instrumentation, controls, piping, valves, and
structural steel were included to make the system a fully functional offshore module. A vacuum tower
module scope of supply summary is below.
In developing the above engineering and design, certain assumptions were made. Major assumptions
include:
● Filtered seawater feed with outlet dissolved oxygen specification of less than 10 ppb.
● Corrosion resistant material selection for seawater service (fiberglass reinforced plastic piping,
glass-flake lined carbon steel vacuum tower, fiberglass reinforced plastic membrane pressure
vessels, nickel aluminum bronze vacuum pumps, super duplex stainless steel injection booster
pumps).
● Non-hazardous area classification for electrical and instrument selection.
● Non-API industry codes and standards due the nature of the process fluid (seawater, not hydro-
carbons).
8 SPE-174935-MS

●Oxygen scavenger chemical is not required for the MDA system, but it was included as a
safeguard.
● The high pressure injection booster pumps can be eliminated in the MDA system.
● Dry air supply for MDA nitrogen generation is available from the common utility system.
● Only one vacuum tower is required for the 250,000 bpd case. Depending on the application,
sometimes two are required.
Capital Cost Study Methodology
The goal of the capital cost comparison was to estimate direct field cost for installation of each deaeration
technology on offshore topsides. Direct field cost (DFC) includes purchased equipment, material, and
labor for installation of equipment and material in the field. Comparing equipment cost only would lead
to inaccuracies in the true capital cost due to the module related savings realized with selection of MDA
technology.
The capital cost evaluation considered membrane and vacuum tower deaeration technologies at the
three injection flow rates: 30,000, 125,000, and 250,000 bpd. The module designs developed in the size
and weight study formed the basis for the capital cost estimate. Vendor and fabrication quotations were
received for all major mechanical/electrical equipment and skids within the scope listed in Tables 2 and
3 above. Module costs were derived using material take-offs for interconnecting piping, valves, wiring,
structural steel, and access ladders/platforms. Offshore norms and cost factors from a historical project
database were applied to the material take-offs to estimate the installed module costs.

Table 2—MDA Module Scope of Supply

Item Quantity

Membrane Skid 1x100% (Number of membranes dependent on flow)

Liquid Ring Vacuum Pump Skid 2x100% Pumps, 1x100% Separator
Pressure Swing Adsorption Nitrogen Generation Skid 2x100% Adsorbent Vessels, 1x100% Receiver Vessel
CIP Skid 1x100% Tank, Pump, & Filter
Oxygen Scavenger Chemical Skid 2x100% Pumps, 1x100% Tank
Instrumentation and Local Control Panel N/A
Module Interconnecting Piping, Valves, & Wiring N/A
Skid and Module Structural Steel with Access Ladders and Platforms N/A

Table 3—Vacuum Tower Deaeration Module Scope of Supply

Item Quantity

3 Stage Vacuum Deaeration Tower with Internals 1x100%

Liquid Ring Vacuum Pump Skid 2x100% Pumps, 2x50% Air Ejectors, 1x100% Separator
Injection Booster Pump Skid 3x50% Pumps
Oxygen Scavenger Chemical Skid 2x100% Pumps, 1x100% Tank
Anti-Foam Chemical Skid 2x100% Pumps, 1x100% Tank
Biocide Chemical Skid 2x100% Pumps, 1x100% Tank
Instrumentation and Local Control Panel N/A
Module Interconnecting Piping, Valves, & Wiring N/A
Skid and Module Structural Steel with Access Ladders and Platforms N/A

The capital cost evaluation excluded the following components:

● Indirect field costs (IFC): Temporary construction facilities; construction services, supplies and
consumables; field staff and subsistence expense; payroll, benefits, insurance; construction equip-
ment and tools
SPE-174935-MS 9

● Home-office costs (HOC): Project management, controls, and estimating criteria, procurement,
construction management, engineering and design, and home-office expenses
● Other project costs (OTC): Including commissioning and contingency

Total installed project cost (TIPC) is equal to the sum of DFC, IFC, HOC and OTC.
Other capital cost items that are not directly accounted for in this study but should be considered in
evaluation of total installed cost for deaeration technology include:
● Shipping and delivery of equipment and skids from the fabrication facility to near-shore module
integration yard.
● Craneage required for transfer of equipment to the near-shore module integration yard, placement
into module, and the topsides lift.
● Any special engineering or lift studies required for transfer of a large vacuum tower.
● Any safety improvements due to elimination of large, high-risk lifts associated with vacuum
towers modules.

Operating Cost Study Methodology

The operating cost evaluation considered membrane and vacuum tower deaeration technologies at the
three injection flow rates: 30,000, 125,000, and 250,000 bpd. The operating costs included in the
evaluation were major consumables, chemicals, and power. These costs were determined using a
membrane replacement frequency of five years, vendor quotations for chemicals, and a 10 cent per
kilowatt-hour estimate for power. Specific operating cost components for each technology are broken
down below.

Table 4 —Operating Cost Evaluation Breakdown

Parameter Vacuum Tower Membrane Deaeration

Major Consumables None Replacement membrane elements

Chemicals Oxygen scavenger, anti-foam, and biocides Acid and caustic for clean-in-place
Power Vacuum pumps, chemical metering pumps, Vacuum pumps, control panel, and
injection booster pumps, control panel, instruments
and instruments

The operating cost evaluation excluded the following components:

● Cost of shipping and handling chemical totes to and from the offshore facility.
● Operating and maintenance labor cost.
● Minor replacement parts and consumables such as gaskets, fittings, and lubrication oil.

Results
The outcome of the engineering and design for each technology is summarized below for the 125,000 bpd
injection flow rate case. 3D models for the 30,000 bpd and 250,000 bpd injection flow rate cases were also
produced, but are excluded from this summary for conciseness.
10 SPE-174935-MS

Figure 6 —125,000 bpd Vacuum Tower Deaeration Module – Elevation View

Figure 7—125,000 bpd Vacuum Tower Deaeration Module – Isometric View

SPE-174935-MS 11

Figure 8 —125,000 bpd Membrane Deaeration Module – Elevation View

Figure 9 —125,000 bpd Membrane Deaeration Module – Isometric View

The outcome of the size, weight, capital cost, and operating cost comparative evaluation are summa-
rized in Figures 10 through 14 below for the 30,000, 125,000, and 250,000 bpd injection water flowrates.
12 SPE-174935-MS

Figure 10 —Membrane deaeration vs. vacuum tower weight comparison

Figure 11—Membrane deaeration vs. vacuum tower size comparison

SPE-174935-MS 13

Figure 12—Membrane deaeration vs. vacuum tower footprint comparison

Figure 13—Membrane deaeration vs. vacuum tower capital cost comparison

14 SPE-174935-MS

Figure 14 —Membrane deaeration vs. vacuum tower yearly operating cost comparison

The results of the above size and weight comparative evaluation indicate that selection of MDA
technology over conventional vacuum towers results in the following offshore facilities impacts:
● 60 – 70% dry and operating weight savings across the entire range of injection flow rates studied.
● 30 – 60% total size savings, with increased savings at elevated injection flow rates.
● 10% footprint savings from 125,000 – 250,000 barrels per day. The footprint of MDA is larger at
small flow rates.
The results of the above capital and operating cost comparative evaluation indicate that selection of
MDA technology over conventional vacuum towers results in the following offshore economic impacts:
● 10 – 15% direct field cost savings from 30,000 – 125,000 bpd. Direct field cost savings between
5 – 10% from 125,000 – 250,000 barrels per day, as economies of scale of vacuum towers begin
to take effect.
● With power cost included, 50 – 60% annual operating cost savings across the entire range of
injection flow rates studied.
● With power cost excluded, 5% annual operating cost savings from 125,000 – 250,000 barrels per
day. Operating cost for MDA is higher at small flow rates.
Conclusions
Membrane deaeration technology as an alternative to conventional vacuum tower deaeration demonstrates
significant promise for use in offshore seawater injection applications for dissolved oxygen removal.
MDA was successfully de-risked through piloting in a simulated offshore environment in 2014. Through
piloting, superior oxygen removal performance compared to vacuum towers was demonstrated, optimal
MDA operating parameters were selected, and minimum levels of pre-treatment were specified.
SPE-174935-MS 15

MDA outperformed vacuum towers in nearly all aspects of the size, weight, capital cost, and operating
cost comparative evaluation. Size and weight savings of up to 60% were demonstrated through detailed
engineering and design of offshore deaeration modules at a wide range of injection flowrates. These size
and weight savings, when combined with performance efficient improvements, translated to direct field
cost savings of up to 15%. Operating costs of MDA systems were up to 60% lower than comparable
vacuum tower systems when power cost was included in the analysis. In practice, power costs offshore
are highly variable and often aren’t considered in offshore operating cost analysis. When power is
excluded from operating cost, MDA operating costs are near equal that of comparable vacuum tower
systems.

References
ElSherif, Kerim, Membrana “Marketing Contact and Contactor Pricing.” Message to Stephen Van
Pelt, 11 September 2014. E-mail.
Wiesler, F. 1996. Membrane Contactors – An Introduction to the Technology. Ultrapure Water.
UP130427.
API RP 17N, Recommended Practice for Subsea Production System Reliability and Technical Risk
Management. 2009. Washington, DC: API.
Carlberg, B. L. 1976. Vacuum Deaeration – A New Unit Operation for Waterflood Treating Plants.
Presented at SPE Annual Fall Technical Conference and Exhibition, New Orleans, Louisiana, 3-6
October. SPE-6096-MS. http://dx.doi.org/10.2118/6096-MS