Successful Pilot of Thermosyphon Process Heater Reduces GHG Emissions
Successful Pilot of Thermosyphon Process Heater Reduces GHG Emissions and
Operating Costs
W.A. Arnold, SPE, PanCanadian Resources, and J.I. Neulander, Hudson Products Corporation
Abstract
PanCanadian originally pursued thermosyphon technology to
lower operating costs. However, it is now apparent that a
more significant benefit is reduction of greenhouse gas (GHG)
emissions. Other benefits include loss management, increased
safety, improved operability and maintenance cost reduction.
A prototype ThermfloTM thermosyphon process heater was
developed by Hudson Products Corporation (HPC) as a joint
effort with PanCanadian Petroleum Ltd. The unit was
installed in a 320 m3 atmospheric heavy oil tank in
northeastern Alberta in November ’97. Performance testing
was carried out over the next several months and evaluated
using API Standards & Recommended Practices.
The testing was carried out in two phases. In phase one, we
compared the performance of the thermosyphon to a naturally
drafted venturi style nozzle burner in a firetube located in an
adjacent tank. Both units were fired using sales utility gas.
Increased combustion efficiency and improved heat transfer
reduced fuel gas consumption by 33% and cut GHG emissions
by 40%. In phase two of the testing, the prototype
successfully burned solution gas normally vented to
atmosphere. This translates to an annual operating cost saving
of about $40,000 and GHG emission reduction of about
10,000 tonnes per year.
Introduction
The thermosyphon process heater is a means of transferring
heat from an external combustion chamber to liquid inside a
process vessel or tank. It does this without having a flame
within a firetube inside the vessel or tank and is therefore,
significantly more efficient and safer than a firetube.
Employment of a prototype heater in a heavy oil field
production tank pilot project confirmed the concept and the
improved efficiency lead to substantial reduction in GHG
emission, lower operating costs and improved safety.
Background
Production Methodology. The preferred strategy of cold
heavy oil reservoir exploitation in the Western Canadian
Sedimentary Basin is to produce emulsion into field
production tanks from wells drilled on a high density pattern,
often at ten to twenty acre spacing. The emulsion is then
trucked to a centralized cleaning facility, where it is processed
to pipeline specification of 0.5% BS&W and 350 centistoke
viscosity in preparation for transportation to refineries.
Pipelining the emulsion directly to the cleaning plants is
generally not feasible due to the sand cut, high viscosity and
entrained gas. The threat of bottom or side water
breakthrough also introduces risk into the return on investment
for pipelining.
Producing emulsion into field production tanks for trucking to
cleaning plants has its own set of challenges. A property of
low temperature heavy oil emulsion is high viscosity, which in
turn lifts large amounts of sand and entrains gas with the oil.
The entrained gas is in the form of expanding microbubbles
that enhance fluid flow from the reservoir1,2. The
microbubbles continue to expand in the tank, as the pressure
drops to atmospheric. Here the gas slowly breaks out causing
foam to sit on the oil in the tank. This foam layer can be
several feet thick, increasing the risk of the tank overflowing
and causing a spill. It also raises trucking costs by reducing
the pay load. Cold viscous emulsion increases load/offload
time, which also raises trucking costs.
Emulsion Processing. Heating the emulsion breaks down the
foam, drops the sand to the bottom of the tank and separates
the water out of the oil. Introduction of chemical deemulsifier
into the tank or gathering line can assist in this
process. The type and amount of chemical is dependent on
water, clay and sand cut, oil characteristics, cleaning plant
ownership and availability, water disposal options, etc. The
apparent treating cost savings from reduced water cut may be
outweighed by the increase in “slop” (tight interfacial
emulsion) disposal costs that don’t surface for several weeks
or months. The total costs must be included in the analysis.
However, chemicals only enhance the process while heat is
essential.
The heat source usually takes the form of a naturally drafted
venturi style nozzle burner in a firetube. The burners
nominally range in size from 75 to 300 kW (250,000 BTU/hr
to 1,000,000 BTU/hr). The firetubes can be straight through
(shotgun) or U-tube, and range from 100 to 300 mm in
diameter. The nominal BTU rating refers to combustion
capability, not the actual heat transfer. Burner manufacturers
most often report combustion efficiency. The actual heat
transfer is affected by liquid residence time, heat transfer
surface area and fouling of this surface. The actual firetube
heat transfer efficiencies are typically in the 35-65% range3.
An industry benchmarking study4 reports utilities as the third
highest cost driver. Tank heating can comprise one half of
utility costs. Obviously low heat transfer efficiency raises
operating costs, but it also greatly increases GHG emissions.
Firetube Fouling. A serious problem encountered with
firetubes is fouling or coking which leads to the development
of a “Hot Spot”. This hot spot normally occurs due to flame
impingement baking the formation sands, clays, salts,
asphaltenes, etc. on to the outside of the firetube. This coke
product acts as an insulator preventing heat dissipation and
eventually causing the metal to become so hot the tube
collapses. This has the potential for personal injury or death,
equipment damage and production loss. Fires have resulted
due to the emulsion entering the firetube through cracks, and
in some instances, the roof of the tank has blown off. In any
case, significant costs are incurred cleaning up the spill
resulting from the emulsion leak and replacing the damaged
firetube. The tube we replaced for the pilot had significant
buildup that we were unaware of. This tube likely would have
failed within two months.
Integrated Solution. PanCanadian recognized these problems
and implemented a firetube preventive management program
in 1994. We also chose to use an integrated risk management
approach to solve this problem. The integrated solution of the
problem would:
! Lower Operating Costs
! Improve Safety
! Improve “Operability”
! Improve Loss Control Management
! Lower Green House Gas Emissions
Table 1) outlines the methodology of the Hazop Process we
followed. Table 2) contains the results of the exercise that
compelled us to pursue heat pipe technology. (Appendix C
provides a sample calculation to show how we arrived at cost
of keeping emulsion at 85°C.)
PanCanadian started experimenting with heat pipe technology
in 1989. Two heat pipe systems were installed in pressurized
process vessels during 1994 with limited success. The
application of this technology developed very slowly due to
supplier economic stability, technology ownership issues and
failures of heat pipes in unrelated industries. In 1996
PanCanadian approached HPC to develop an alternative
heating system for our atmospheric heavy oil storage tanks.
At the time HPC was a supplier of heat pipes for use in heat
exchangers but not a manufacturer of an externally heated
system. Working as a team, we set out to develop a
thermosyphon process heater.
Prototype Description
Gravity-assisted heat pipes or thermosyphons5 are selfcontained
devices for thermal energy transport. They are
characterized by their ability to transport heat at high rates
over considerable distances, yet require only small
temperature differences to drive the heat flow. The basic
principles of thermosyphon operation are relatively simple.
Referring to Figure 1, the working fluid inside a
thermosyphon exists as liquid and vapor at nearly equilibrium
thermodynamic conditions. During operation, two transport
mechanisms exist – heat and mass transport. When a
thermosyphon is exposed to heat source and sink, the working
fluid equilibrates to an average saturation temperature and
pressure. The temperature of the working fluid is nearly
uniform through out the tube between the temperatures of the
heat source and sink. Heat addition causes the liquid within
the tube to evaporate and enter the vapor phase. Heat removal
at the condenser section condenses the vapor, returning it to
the liquid state. This simultaneous influx and efflux of vapor
creates the driving pressure differential for the nearly isotherm
flow of vapor from the evaporator to the condenser. To
complete this thermodynamic and transport cycle, the liquid or
condensate must be transported from the condenser to the
evaporator. Condensate return is accomplished via gravity
force.
In 1996, HPC began development of the thermosyphon
process heater for use in heavy oil storage tanks. This was a
very challenging application of the technology given the
stagnant fluid and temperature limits of 100oC within the
storage tanks. Drawing on past and recent work on
thermosyphon technology, in mid 1997 a prototype heater was
designed and constructed.
The inherent design advantages of the thermosyphon process
heater over naturally drafted venturi style nozzle burners
include:
! External low pressure burner can utilize lower quality gas
as a fuel source
! External burner and lower surface temperature of
thermosyphons, is safer
! Higher effective heat transfer allows for quicker heat
times, more consistent emulsion temperatures throughout
the tank, lower operating costs and less GHG emissions
Test Design
The field trial was designed using API Recommended
Practices & API Standards to maximize the repeatability of the
results. We chose a site at our Marwayne Field located in
northeastern Alberta that had sixteen wells drilled on ten acre
spacing, with eight wells producing into each of two 320 m3
tanks. We proposed replacing one firetube with the
thermosyphon prototype, allowing direct comparisons of
performance by mitigating the effects of varying ambient
temperature, inconsistent oil properties and differing
production rates on the resultant data. Equipment
specifications are summarized in Appendix A.
Field Description. The Marwayne Field is drilled into the
Mannville group of zones at a depth of about 700 meters. The
wells are drilled with slant rigs, keeping the boreholes as
straight as possible to minimize tubing and rod wear. The oil
is lifted using PCP 15-1400EL pumps (15 m3/d @ 1400 m
head) with 22kW electric motors c/w VFD’s. The two zones
are commingled and yield an emulsion at surface of about 11°
API and 35,000 cP viscosity. Emulsion temperature at surface
is 27° C. The sand cut ranges from up to 30% initially, to 1-
3% after flow stabilization in one to three months.
Test Objectives
Prior to test initiation, HPC and PanCanadian met to confirm
that each company’s objectives were mutually attainable. The
field test was designed to ensure operability and to provide
information about the equipment specifics.
Here is the compiled list of test objectives that we intended to
establish:
! overall thermal efficiency (fuel based)
! heat input and heat transfer rates (tank emulsion heating)
! firing, safety, environmental and operating cost
improvements by implementing the prototype for indirect
heating with burner equipment firing outside the storage
tank causing less coking/build up on firetubes
! operational advantages in implementing thermosyphons
for transferring heat into the tank with high rate of sand
deposits
♦ better temperature control of emulsion
♦ quicker emulsion heating times
♦ improved burner reliability
! energy conservation, fuel saving and decreased CO2 and
NOX emissions on equipment burners
! ability to burn solution gas with no significant decrease in
preceding objectives
Test Procedure
The thermosyphon prototype was assembled at the
manufacturer’s Beasley, Texas facility. The unit was designed
for retrofit installation into the 320 m3 atmospheric storage
tank through the existing 300 mm firetube flange.
Thermosyphon performance was verified by testing the
prototype for equipment and control system operability in a
tank simulator at the Beasley site in Oct ’97. The simulation
was successful, and the prototype was shipped.
The prototype was installed at the field site on November 13,
1997. The assembly and installation of the unit took about six
hours using a maintenance crew, picker truck and a welder.
We were very pleased with the time considering this was the
first time the operations personnel had seen, let alone handled
this equipment. As a prototype, the thermosyphon employed a
number of features to ensure test reliability, which increased
the size of the unit. Plans are already underway to reduce the
size of the external heat processor. In future, standardized
flanges and procedures could further shorten the installation
time.
The testing was scheduled to run approximately three months.
Preparation for and testing was carried out according to API
RP #532, Sect 2, Par. 2.1 & 2.2. The following measurements
were taken during the test using API RP #532, Par. 15:
! Fuel flow rate
! Process flow rate
! Process fluid inlet temperature
! Process fluid outlet temperature
! Process fluid inlet and outlet pressure
! Fuel pressure at the burner
! Flue gas draft profile
To evaluate firing, combustion, heat generation and heat
transfer performance of the thermosyphon, a testing method
was developed in four stages. In stage 1, we would monitor
the flue gas side and measure: fuel gross efficiency; net
efficiency in the fire box; and, combustion efficiency.
Expected duration was one week. In stage 2, we would
monitor the process side and measure: heat transfer; and, heat
absorbed by the process. Expected duration was one week. In
stage 3, we would monitor overall heat balance and
measure: combustion efficiency, heat transfer efficiency, and
relative waste. Expected duration was five weeks. In stage 4,
we would burn solution gas normally vented to atmosphere up
the casing. Expected duration was six weeks.
Stage 3 & 4 ran extended times to gain experience operating
the prototype under field conditions and allow for
optimization of:
! Available heat/ radiant-convection section
! Heat picked up by process / liquid in pipes
! Distribution of heat in lowest row and in between pipes
! Flue gas circulation and distribution in between pipes at
different bundle height inside the fire box
! Stack draft and damper effect
! Size and equipment dimensions to optimize heat transfer
Evaluation
The performance testing compared tank operation,
temperature and energy consumption between the
thermosyphon process heater installed in tank “A” and the
firetube and natural draft burner in tank “B”. The objective
was to compare and investigate Tank A and Tank B operation
by evaluating the following steps:
1. Fuel Consumption - Measuring the differences in
operating Tank A vs. Tank B by maintaining the product
in tanks at the same temperature level.
2. Temperature Gain - Compare operating temperatures in
Tank A vs. Tank B at the same rate of fuel input
comparing temperature increase and response time.
3. Operational Difference - Compare Tank A vs. Tank B by
firing each tank at its maximum firing capacity and
measuring the CO2 and NOX emissions, coking/buildup,
equipment reliability.
Performance and operability of the thermosyphon was
evaluated according to API Standards and Recommended
Practice issued for fired heaters covering the subjects:
" Measurement of the Thermal Efficiency on Fired Process
Heaters - API RP #532 (Aug ’82)
" Fired Heaters for General Refinery Services - API
Standard #560 (Jan ’86)
" Instrumentation and Control System for Fired Heaters
And Steam Generators - API RP #556 (May ’97)
" Process Analyzers - API #555, 550 part II
The flue gas was analyzed according to API RP #532, Par.
1.4.2 using a portable analyzer for oxygen and combustible
gases.
The measured data was interpreted for GHG emissions using
VCR Reporting Guideline Rev 98/April/8, published by the
Canadian Association of Petroleum Producers (CAPP). The
portion of the report containing relavent information is
reproduced in Appendix B. This report converts total
greenhouse gases to a carbon dioxide equivalent (CO2E).
All parameters, fuel consumption, temperature and firing
capacity were recorded on survey forms and the tank heating
efficiency was calculated using computer models and heat
balance calculation forms developed by HPC. All activities
including collection of data and analysis were conducted
according to a Standard Practice Manual certified under ISO
9001. The objective of this quality management system is to
ensure that all project work is performed on time and within
budget and the project will achieve the intended R&D
objectives and is understood, and, if necessary could be
reproduced successfully by others.
Test Results:
Table 3) summarizes the results. Appendix C provides some
sample calculations. The thermosyphon process heater had a
combustion efficiency of 83%, which is extremely high for a
naturally drafted burner. This was 25% higher than the
firetube burner. The effective heat transfer was double that of
the firetube system. The prototype burned 33% less fuel while
providing more heat to the emulsion. GHG emissions were
reduced by 5,000 tonnes per year.
Please note that in this test we adjusted the firetube burner on
a weekly basis to maintain maximum combustion efficiency.
Most field tank burners are lucky to receive annual tune-ups.
For each 1% drop in combustion efficiency, the CO2E
increases by 2%. Or, given the above fuel rates, a fire tube
burner at 50% combustion efficiency would have a 30%
higher CO2E than we reported. This assumes there is no
increase in fuel consumption due to decreased heat transfer
effectiveness. There are several operators who use two fire
tubes per tank to achieve the required heat input.
We were forced to make one change after initial results were
analyzed. In designing the test, we overlooked the fact that
the water cut in the emulsion going in to each tank was
significantly different. Therefore we added a step to the test
by switching which tank and burner system the emulsions
flowed into. This required some mechanical rework of the
gathering lines and extended the test one month, but it did
verify the initial test results.
In the last phase of the pilot, we burned solution gas in the
thermosyphon process heater that is normally vented to
atmosphere out the casing. The heating value was fairly
similar to the fuel gas we were purchasing, but being solution
gas, it was much richer. The richness and moisture content of
solution gas normally causes problems in the nozzles of
firetube venturi burners. We experienced no operational
problems and the thermosyphon process heater performance
was equivalent to using commercial gas. The GHG emission
reduction in this case is 10,000 tonnes per year.
Summary:
All objectives of the pilot were met. PanCanadian Petroleum
Ltd. is very pleased with the THERMFLO Tank Heater
supplied by Hudson Products Corp. Perhaps the most
significant measure of the success of this pilot is operator
endorsement. From the day the unit was fired, it has
performed trouble free and beyond all expectations. The unit
is operator friendly, safe and efficient. The prototype is still in
use. It has the potential of saving about $40,000 per year in
fuel gas costs on this pad, and is the equivalent of parking 9
cars per BOPD.
Acknowledgments
We thank Bob Giammaruti, George Millas and F.G. Russel
with HPC for valuable discussions on the interpretation of the
pilot results and strong R&D support; Jack Whittaker, Steven
Lee, Jody Kissick and Dale Cussack with PanCanadian for
their assistance in executing the pilot, and their helpful
feedback