ENGINEERED SOLUTIONS

Don’t accept poor

starting reliability
By Brent A Gregory, Creative Power Solutions

S
ome power producers want
to believe that gas-turbine-
based generating facilities
don’t require a rigorous
engineering effort at the design/
specification stage of a project.
Given that the track record of gas
turbines for land-based generation
service industry-wide is excellent,
and that GTs are “standard” facto-
ry-assembled machines, it’s no won-
der that executives who majored in
business and finance think buying
a gas turbine is not much different
than buying a new corporate limo:
Sign the papers, pay, and you’re
ready to go.
But the “standard” engine often
doesn’t quite match actual site and
operating conditions and, conse-
quently, the new plant may not meet
the owner’s expectations regarding 1. Libyan combined-cycle plant has three 2 × 1 power blocks. First four Als-
starting reliability, availability, effi- tom 13E2 gas turbines were installed in 2001, the third block five years later
ciency, etc—critical factors in the
competitive generation business. It’s
rare that a modest investment in 100
engineering and hardware improve-
ments can’t correct the situation. The Rotor speed (100% = 3333 rpm)
90
experience of the General Electricity
Co of Libya (Gecol) that follows is a 80
case in point. Control valve
The utility’s GT fleet includes stroke, torch 2
70
23 Alstom machines—including
these models: 13D, 13E1, 13E1M,
Percent of value

60
13E2MXL, and 8C. One of its
combined-cycle plants, a facility Control valve
50 stroke, torch 1
with three 2 × 1 power blocks and
equipped with distillate-fueled 13E2 TAT (turbine
engines, faced recurring problems 40 exit temperature)
during wintertime starts. It turned Temp of torch 1 tip
to Creative Power Solutions (CPS) 30
for a root-cause analysis and per-
manent solution (Fig 1). Note that 20 Temp of torch 2 tip
Libya, like most Middle Eastern Flame (1 = off; 10 = on)
nations, does not release the names 10
and locations of important indus- TSOV (10 = closed; 0 = open)
trial facilities. 0
0 50 100 150 200 250 300 350 400
For readers unfamiliar with the Time, sec
13E2, it is a robust frame engine with
a 21-stage compressor, five-stage tur- 2. Successful startup, mapped here, often was elusive because of ignition issues
92 COMBINED CYCLE JOURNAL, Second Quarter 2010
 ENGINEERED SOLUTIONS
bine, and a 72-burner annular com- neously: distillate and water for NOx

Velocity of fuel at burner nozzle

bustion system. Last has two ring- control. A drawback of this arrange-
type fuel manifolds, each with 36 ment is that during ignition, NO x
burners. Arrangement of burners is Normal operation water is not used and the velocity of
such that there’s one located every 5 with water injection injected fuel is low, which is condu-
deg around the combustion chamber. cive to poor atomization (Fig 3).
The OEM claims this design assures The large droplets of fuel resulting
even temperature distribution and from poor atomization vaporize slow-
avoids what it considers problem Start cycle, no ly and mixing of the oil vapors with
zones in competitive can-annular water injection air likewise is poor. Such conditions
systems—such as cross-firing tubes produce an erratic flame which can
and transition pieces. self-extinguish. A strong pilot flame
Relative rotor speed, Relative load
The Libyan 13E2s have two igni- can help assure reliable ignition in
tion torches located 120 deg apart. 3. Fuel velocity was low on startup such situations.
They are ignited electrically and making proper atomization virtually Engineers thought another reason
operate on bottled gas; flame is sup- impossible for the failed starts might be high
ported with air from the instrument flash-point temperature, so they ana-
air system. lyzed distillate samples from many
Flash point

CPS engineers began the proj- Small droplets, shipments for viscosity and flash
ect by examining data provided by rapid vaporization point. Fig 4 shows that high-flash-
Gecol for both successful and failed Large droplets, point fuels also have high viscosities
slow vaporization
starts. Based on this work, hypoth- and, therefore, are more difficult to
eses were put forward and an action atomize than low-flash-point/low-vis-
plan developed to test the hypoth- Kinematic viscosity cosity oils. Variability in flash-point
eses and identify possible solutions. 4. Small droplets of fuel produced by temperatures across the samples was
Next, the preferred solution would good atomization vaporize quickly 36 deg F, confirming in the minds of
be implemented on one engine and, investigators that high viscosity was
after validation, would be installed the underlying reason for at least
on all six 13E2s. detail by analyzing the following: some failed starts.
How the engine is supposed n Fuel properties. To sum up, the CPS investigators
to start (follow curves in Fig 2). n Operating conditions. concluded that the primary reason
First steps in a successful start: The n Ignition torch tests. for low starting reliability was poor
ignition sequencer lights the igni- Here’s what engineers learned: ignition caused by weak gas pilot
tion torch and later presets the oil 1. Low-quality atomization of liq- flames and poor atomization. The
control valves to the system filling uid fuel during engine startup. ideal oil for startup, they agreed,
stroke. When the flame is stable, as 2. Wide variability—36 deg F—in would be one with a low viscosity
indicated by an increase in torch tip the flash point temperature of the to achieve fine droplet sizes during
temperature, the trip shut-off valve light fuel oil burned. atomization and a low flash point to
(TSOV) opens and oil fills the system 3. Ignition torches produced a facilitate ignition. Also that the pilot
up to the sector (pilot) valves. weak flame at the established operat- flame must be strongly “attached”
After a timed delay to fill the ing parameters. to its lance and be sufficiently hot to
system, control valves go to ignition 4. Rapid depletion of bottled pilot vaporize and ignite all possible oils
stroke, the burners ignite, and moni- gas caused a weak torch flame dur- delivered to the plant.
tors detect the main flame. Then the ing startup. (Perhaps the gas bottles Another benefit of prompt igni-
torch is shut off and its tip tempera- were too small or were not being tion and complete combustion dur-
ture drops. refilled fully.) ing startup is the elimination of oil
That’s the plan. But analysis of residue in the combustion chamber
startup data revealed the following The details from failed starts. When that residue
scenarios which caused or contrib- ignites during a successful start the
uted to the failed starts: Fuel nozzles for the Gecol GTs are additional heat may trip the unit on
n Ignition pilot flame either did not designed to inject two fluids simulta- “high TAT.”
ignite or failed after ignition.
n Combustor ignited and the tur-
bine exhaust temperature (TAT)
rose to the alarm limit, accelera- Atomizing
tion dropped below its limit, and air system
the engine tripped. Orifice
n Combustor ignited and operation or valve Ignition torch
was normal—that is, until the
TSOV closed (data offered no clues Pressure gage
as to why this happened) and the
engine tripped on “no fuel.”
n Combustor ignited and opera-
tion was normal—that is, until Flowmeter
the flame monitors no longer saw
the flame (weak signal) and the Fuel gas
bottle
engine tripped on “flame-out.”
Several hypotheses proposed
for the various startup failures
observed were investigated in 5. Test apparatus installed by CPS engineers was simple, effective
COMBINED CYCLE JOURNAL, Second Quarter 2010 93
 ENGINEERED SOLUTIONS

Blow-off titative analysis to determine the throughput on flame strength. This

optimum operational envelope for work was done by calculating flow
Overheat the igniters. The parameters used to rates for different orifice sizes based
indicate flame condition included the on air and fuel line pressures. Plots
Air flow

Ideal ignition Detached

conditions following: for theoretical air and fuel flow rates
Optimal n Torch metal temperature. A ther- as a function of line pressure and
window mocouple was used for this pur- orifice size is shown in Figs 7 and 8,
pose. respectively.
Fuel flow n Combustion-zone temperature. In Fig 7, the ellipse defined by
6. Flame conditions are shown for Measured by a thermometer to the dashed line presents the optimal
various air and fuel flow rates using a assess the impact of gas-flame window for air flow. It was deter-
single ignition torch; no orifices were radiation on fuel-oil vaporization. mined using the minimum air pres-
installed in the supply lines. Best con- n Flame strength. A small-diame- sure expected from the instrument
ditions are within the dashed ellipse ter, high-velocity jet of compressed air system—about 70 psig—and the
air was blown across the flame to blow-off point for the igniter flame,
The solution qualitatively determine its stabil- as determined during the workshop
ity/strength. Photos were taken tests and plotted in Fig 6.
Obvious from the engineering evalu- during the tests at various oper- The takeaway from this effort was
ation was that igniter flame strength ating conditions to record visual that an orifice in the air line of 4 mm
had to be increased to improve start- observations. or larger is undesirable. Further,
ing reliability. A more robust flame Mapping of test results identi- that use of a small orifice (1 mm), as
would help even large oil droplets fied the air and fuel flows that pro- suggested by the OEM’s design crite-
vaporize and ignite quickly. A fully duced both high flame strength and ria, also is not in your best interest.
instrumented test rig that would temperature (dashed oval in Fig 6). Optimal window for fuel flow
allow CPS engineers to vary air and This information was verified as revealed that a moderate orifice size
gas flow to one engine’s ignition suitable for conducting a full-scale was the best option—too small and
torches was assembled in the plant proof-of-concept test on one of the flame temperature would be too low,
workshop (Fig 5). station’s GTs with the expectation too large and the flame would blow
First step was to identify the opti- of success. off.
mum flow conditions, with no orifices With an optimum operating win- The ignition flame produced when
installed, using the flowmeters on dow identified, engineers conducted using OEM-recommended orifice
both the air and gas lines. The data a quantitative analysis to demon- sizes in the gas and air lines is too
gathered provided the basis for quan- strate the importance of air and gas lazy (Fig 9). Modest enlargement of

Orifice size, mm
Blow-off
Fuel flow

5 Orifice
Air flow

size, mm
Flow too high
4 3
Optimal window Optimal window
3 Inoperable
sector valve 2
2 Flow too low
1 1
Air pressure Fuel pressure
7, 8. Dashed ellipse shows the optimal window for air flow (left), fuel flow (right)

Unit 4 modification CPS limit

3 by plant operators
Normalized flame strength

2
Preferred region
of operation (CPS
recommended)

1 Unit 1 modification
by plant operators
OEM design

0
0 0.2 0.4 0.6 0.8 1
Normalized flame temperature
9. Ignition flame produced using the OEM’s gas and air
orifice settings is much too weak to achieve a high start- 10. Operating in the preferred region (upper right)
ing reliability on oil assures high starting reliability
94 COMBINED CYCLE JOURNAL, Second Quarter 2010
 plant because the fuel did not meet
the manufacturer’s spec and there
was insufficient gas to create a strong
flame.
Use of orifice sizes specified by
CPS created the flame shown in Fig
11. Use of these orifices on all six
engines restored starting reliability
to fleet-wide numbers. All engines
started on the first attempt. ccj

11. Ignition flame using orifice sizes Brent A Gregory (brent.gregory@cpsu-

sainc.net) is president of Creative Power
recommended by CPS provides the Solutions’ US operations, headquar-
thrust needed for successful starts tered in Fountain Hills, Ariz; the com-
pany also has offices in Switzerland and
the fuel and air nozzles by the plant’s the UAE. CPS has
operations team improved flame a broad clientele: it
strength somewhat, but not enough works with utilities
to solve complex
for the required task.
problems affect-
Orifice sizes for the gas and air ing turbomachinery
lines suggested by calculations per- and steam genera-
formed by CPS engineers were then tors, with OEMs to
tested using the engine ignition help develop equip-
torches. Stable flame conditions were ment upgrades and The authoritative
obtained in the region marked in the next-generation
hardware, and with
information resource for
upper right of Fig 10.
The stability of a flame depends government to pro- owner/operators of gas-
on its thrust: The higher the thrust, vide strategic thinking on future tech-
nologies—such as gasification.
turbine-based
the more heat available to vaporize CPS’s executive team has over a cen- peaking, cogen, and
the liquid fuel. For a given torch, tury of engineering/management experi-
thrust is increased by boosting the ence with engine manufacturers Rolls- combined-cycle plants.
total flow of gas and air through Royce, GE, and Alstom, and Honeywell Subscribe via the Internet, access:
the ignition system. The OEM igni- and Combustion Engineering. www.psimedia.info/subscriptions.htm
tion system performed poorly at this

Did you read an article in the COMBINED

CYCLE Journal that you’d like your
D FR
OM
customers and prospects to have?
INTE
REPR
Use reprints to help raise the visibility of
ate your products and services among buying
e stiming
d e r
’t un of st men
o r t influences at the plant level and
Don ands t equip
dem rplan term
e - headquarters locations
pow long

Call Susie Carahalios

for an estimate
303-697-5009
.
2004
Inc,
edia
SI M
ht P
opyrig
2004. C
Fall
nal,
Jour
YCLE
ED C
BIN
COM

susie@carahaliosmedia.com
the
om
ission fr
perm
with
rinted
Rep

COMBINED CYCLE JOURNAL, Second Quarter 2010 95