DLE Overview.pdf

g

GE Aircraft Engines

__________________________ Marine & Industrial Control Dynamics

LM2500 AND LM6000 DRY LOW EMISSIONS CONTROL OVERVIEW

TECHNICAL DATA GENERAL LICENSE TSU/OTS APPLICABLE The information ( including technical data) contained in this document is the property of GE. It is disclosed in confidence and the technical data therein is exported under a U.S. government license. Therefore, none of the information may be disclosed to other than the recipient, or used for purposes other than to render services to GE, without the express prior written authorization of GE. In addition, the technical data therein, and the direct product of the data, may not be diverted, transfered, re-exported or disclosed in any manner not provided for by the license without the prior written approval of the U.S. government.

Prepared by : Peter Harrison 27th March, 1997

‘Even-b800d\Data\Users’[S:]’\General\Common\Dle_fld\dleovr.doc

Dry Low Emissions Control Overview Table of Contents 1. OVERVIEW

6

1.1. Dry Low Emissions Combustor

6

1.2. Dry Low Emissions Control Components

10

1.3. Fuel System 1.3.1. Four Valve System 1.3.2. Three Valve System

12 12 13

1.4. Bleed System

14

1.5. Fuel Control

15

1.6. Flame Temperature Control 1.6.1. Bulk Flame Temperature 1.6.2. A and C Ring Flame Temperature

15 15 17

2. FLAME TEMPERATURE ALGORITHM

19

3. FUEL PROPERTIES

25

3.1. Effect of fuel properties

25

3.2. Ring fuel nozzle scalars

27

3.3. Calorimeter and Chromatograph Operation

27

4. STARTING

28

4.1. Start Sequence

28

4.2. Control Operation

29

5. OPERATION AT IDLE

36

6. OPERATION WITHIN A COMBUSTOR CONFIGURATION WINDOW

38

7. FUEL METERING

40

GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 2

8. COMBUSTOR STAGING

43

8.1. General

43

8.2. Starting

44

8.3. LM6000 Core Idle-Sync-Idle

44

8.4. Idle and Above Operation

44

8.5. LM6000 BC to AB zone avoidance

46

8.6. LM6000 BRNUL upper limit

46

8.7. ABC to AB stage down - LM2500

46

8.8. ABC to AB stage down - LM6000

47

8.9. Load drop/ overspeed

48

9. ACOUSTICS AND BLOWOUT AVOIDANCE (ABAL) LOGIC


48

TABLE OF FIGURES FIGURE 1.1 DLE COMBUSTOR

6

FIGURE 1.2 26-VALVE FUEL MANIFOLD

7

FIGURE 1.3 STAGING VALVE LAYOUT

7

FIGURE 1.4 COMBUSTOR CONFIGURATIONS

8

FIGURE 1.5 DLE ENGINE WITH COMBUSTOR STAGING VERSUS NON-DLE ENGINE

8

FIGURE 1.6 DLE ENGINE WITH COMBUSTOR STAGING AND BLEED MODULATION

9

FIGURE 1.7 DLE BLEED VALVE

11

FIGURE 1.8 FOUR VALVE FUEL SYSTEM

12

FIGURE 1.9 THREE VALVE FUEL SYSTEM

13

FIGURE 1.10 EFFECT OF COMBUSTOR CONFIGURATION AND COMPRESSOR BLEED

14

FIGURE 1.11 BULK FLAME TEMPERATURE WINDOW

16

FIGURE 1.12 SPECIFICATION LM6000 RING AND BULK FLAME TEMPERATURE SCHEDULES

17

FIGURE 2.1 LM2500 DLE FLAME TEMPERATURE ALGORITHM

19

FIGURE 2.2 FLAME TEMPERATURE ALGORITHM/AIRFLOW CONTROL INTERFACE

21

FIGURE 2.3 LM2500 FLAME TEMPERATURE SENSITIVITY

22

FIGURE 2.4 LM2500 FLAME TEMPERATURE SENSITIVITY DUE TO PS3 VARIATION

23

FIGURE 2.5 LM2500 FLAME TEMPERATURE SENSITIVITY DUE TO WFAGMV VARIATION

23

FIGURE 2.6 LM2500 FLAME TEMPERATURE SENSITIVITY DUE TO LHV

23

FIGURE 2.7 LM2500 FLAME TEMPERATURE SENSITIVITY DUE TO SG

24

FIGURE 4.1A TYPICAL LM2500 DLE START CHARACTERISTICS

32

FIGURE 4.2A TYPICAL LM6000 DLE START CHARACTERISTICS

34

FIGURE 6.1 LM2500 BLEED SEQUENCE

39

FIGURE 6.2 LM6000 BLEED SEQUENCE

40

FIGURE 6.3 LM6000 VBV BLEED MODULATION

40

FIGURE 7.1 RING FUEL FLOW DEMANDS

41

FIGURE 8.1 COMBUSTOR STAGING DURING LOAD ACCELS

45

FIGURE 8.2 COMBUSTOR STAGING DURING LOAD DECELS

45

FIGURE 9.1 ACOUSTICS/BLOWOUT AVOIDANCE LOGIC

49

FIGURE 9.2 BLOWOUT DETECTION ALGORITHM WF/PS3 ERROR CALCULATION

50

FIGURE 9.3 BLOWOUT DETECTION

51


List of Tables TABLE 1.1 ADDITIONAL LM2500 CONTROL COMPONENTS FOR DLE APPLICATION

10

TABLE 1.2 ADDITIONAL LM6000 CONTROL COMPONENTS FOR DLE APPLICATION

10

TABLE 1.3 FUEL CONTROL REGULATORS

15

TABLE 4.1 START-RUN SEQUENCER OUTPUTS

28

TABLE 4.2 LM2500 START-RUN ENGINE MODES AND CONTROL ACTIONS DURING START TO IDLE

29

TABLE 5.1 LM2500 TYPICAL CORE IDLE PARAMETERS

37

TABLE 5.2 LM6000 TYPICAL CORE IDLE PARAMETERS

38

TABLE 7.1 DISPLAYED FLAME TEMPERATURES

43

TABLE 8.1 STAGING CONTROL PARAMETERS

44

TABLE 8.2 LM6000 BRNUL T3 SWITCH POINTS

46

TABLE 9.1 ABAL CORRECTIVE ACTION

51

TABLE 9.2 ABAL SPECIAL FEATURES

52

TABLE 9.3 ABAL ACOUSTIC SPIKE DETECTION

53


1.

Overview 1.1.

Dry Low Emissions Combustor

The LM2500 and LM6000 Dry Low Emissions (DLE) gas turbines employ a triple annular combustor. Figure 1.1 shows the basic combustor configuration.

A B

C AFT LOOKING FORWARD

A B C

End view showing the 75 cups

Cross section of 3 cup assembly

Figure 1.1 DLE combustor Gas fuel is introduced into the combustor via 75 air/gas premixers packaged in 30 externally removable and replaceable modules. The premixers produce a very uniformly mixed lean fuel/air mixture. Half of these modules have two premixers and the other half have three. The 75 premixers, or cups, as they are often referred to for the DLE, are arranged in three rings or domes. The middle ring is referred to as the pilot or the B ring and has 30 cups. The pilot ring is always fueled. The inner ring is referred to as the C ring and has 15 cups, whereas the outer ring, which is referred to as the A ring, like the pilot has 30 cups. Unlike the pilot ring, fuel to the cups in the inner and outer rings has to be turned on and off by means of staging valves. This is because of the limited flame temperature (or fuel-air ratio) range over which the combustor can operate. The flame temperature range is limited by thermal stress limits on the high side and lean blowout on the low side. The minimum bulk or average flame temperature for an LM6000 ranges from approx. 3300 deg F at no load sync idle to approx. 2900 deg F at maximum power, whereas the maximum bulk or average flame temperature ranges from approx. 3450 deg F at no load sync idle to approx. 3000 deg F at maximum power. With such a limited flame temperature operating range, it is necessary to “stage” the combustor, i.e. it is necessary to turn sections of the combustor “on” and “off”. In the current design, 15 staging valves supply the inner ring, one cup per staging valve, and 10 staging valves supply the outer ring, three cups per staging valve. One additional staging valve, as described later, is used to control the fuel flow level to what was originally referred to as an enhanced lean blowout (ELBO) circuit, that is connected to 15 of the 30 pilot cups. This brings the total number of staging valves to 26. The staging valves are mounted on the fuel manifold assembly as illustrated in Figures 1.2 and 1.3.


Staging valves Section of LM6000 PB showing staging valves mounted on fuel manifold

Figure 1.2 26-valve fuel manifold

Forward looking AFT

1

2

26 25

24

3 4

23 Ignitor 22

5 Boroscope 6

21 7 20 19

8

Boroscope

18 Ignitor

9

outer 17 10

16 11

inner

15 12

13 14

ELBO

Figure 1.3 Staging Valve Layout


In the near future, as part of a cost reduction initiative, the LM2500 DLE will change to a new fuel manifold and staging valve configuration which uses fewer staging valves, 5 for the inner and 5 for the outer, which, when the ELBO staging valve is added, brings the total to 11. The two configurations are often referred to as the 26-valve and 11-valve (or 5/5/1) systems respectively. The staging valves allow different fueling configurations for the combustor, ranging from B-only for starting and idle operation, to fueling of all three rings (ABC) for operation at high power. As mentioned earlier, different combustor configurations are required to keep the combustor flame temperature within limits. The different combustor configurations are shown in Figure 1.4

(Starting only for LM2500) Figure 1.4 Combustor configurations Figure 1.5 shows the operating line of an LM6000 DLE engine employing combustor staging, compared with a conventional non-DLE engine.

LM6000 DLE ENGINE ZERO BLEED versus NON-DLE ENGINE 2600

ABC MAX TFLAME 2400

MIN TFLAME

AB

2200

NON DLE ENGINE 2000

BC

BC/2

1800

1600 .0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

45.0

GENERATOR MEGAWATTS

Figure 1.5 DLE engine with combustor staging versus non-DLE engine It is clear from Figure 1.5 that there is a limited operating power range for each combustor configuration. Operating at a higher power than intended in a given combustor configuration means exceeding the max allowable average flame temperature and can result in extensive damage to the combustor. Attempting to operate at a lower power than intended in a given combustor configuration means attempting to run below the min allowable flame temperature and can result in blowouts. From this illustration it can be seen that GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 8

there are “gaps” between each configuration, i.e. power regions in which the DLE engine could not run. This is overcome by using compressor bleed as illustrated in Figure 1.6.

LM6000 DLE Combustor Operating Modes 2600

ABC EE D

2400

BL

BLEED

M

AX

AB

NO

BL E

ED

2200

Typical Mode Maps at T2=47F

2000

BC

Based on OrangeCo

1800

BC/2

1600 .0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

45.0

GENERATOR MEGAWATTS

Figure 1.6 DLE engine with combustor staging and bleed modulation Another important requirement for the triple annular DLE combustor is the ability to independently vary the combustor flame temperature of each ring. This is achieved by individual control of the total fuel flow to each ring.


1.2.

Dry Low Emissions Control Components

The DLE application for both the LM2500 and LM6000 requires additional control components over and above those required for their non DLE counterparts. These additional control components are listed in Tables 1.1 and 1.2 . Description

Type

Used for

Engine inlet temp (T2)

One dual element RTD

Flame temperature control

Compressor exit temp. (T3)

One dual element TC


Flame detector

Two UV detectors

Combustor lightoff detection

Acoustic sensor (PX36)

Two piezoelectric charge sensors

Staging valves

Twenty six solenoid operated valves with switch position f/b

Flame temperature trim Combustor staging

Eighth stage (ST8) bleed valve One hydraulically operated valve with LVDT dual f/b


Compressor discharge (CDP) bleed valve


One hydraulically operated valve with LVDT dual f/b

Table 1.1 Additional LM2500 control components for DLE application

Description

Type

Used for

Compressor exit temp. (T3) (new location for DLE)

Two dual element TC’s

Flame temperature control Power management

Compressor exit press (PS3) (new location for DLE)

Two transducers

Flame temperature control IGV scheduling Power management Stall detection

Acoustic sensor (PX36)

Two piezoelectric charge sensors

Staging valves

Twenty six solenoid operated valves with switch position f/b

Flame temperature trim Combustor staging

Eighth stage (ST8) bleed valve One hydraulically operated valve with LVDT dual f/b


Compressor discharge (CDP) bleed valve


One hydraulically operated valve with LVDT dual f/b

Table 1.2 Additional LM6000 control components for DLE application The pressure and temperature sensors are conventional and are described in more detail in the Control System Specifications (M50TF3740 and M50TF3731 for the LM2500 and LM6000 respectively) and the Installation Design Manuals (MID-IDM-2500-10 and MID-IDM-6000-3 for the LM2500 and LM6000 respectively). Note that the LM6000 T3 sensor and PS3 pressure tap locations for the DLE engine are different, being located downstream of the combustor diffuser at station 32. The acoustic sensors are unique to the DLE application and are mounted on the compressor rear frame. These transducers are GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 10

piezoelectric charge devices similar to vibration monitoring accelerometers but are used to sense dynamic pressures in the combustor. A purchaser-supplied monitoring system is used in conjunction with the pressure transducers to provide a signal to the control system. The staging valves mounted on the gas manifold are electrically activated and are de-energized open. There are two suppliers for the staging valves, and although valves from the two suppliers look different, they can be intermixed. The eighth stage and compressor discharge bleed valves are located off-engine and each comprise a torque motor servo valve, actuator, LVDT, and air valve. The LM2500 uses 4.0 inch diameter air valves for both eighth stage bleed and compressor discharge bleed. The LM6000 on the other hand uses a 2.5 inch diameter air valve for compressor discharge bleed and a 6.0 inch diameter air valve for eighth stage bleed. A bleed valve assembly is shown in Figure 1.7. LVDT

Airflow

Torquemotor connector

Hyd supply and return

Figure 1.7 DLE bleed valve


1.3.

Fuel System

For the DLE application a three-ring high accuracy fuel system, with associated sensors, is required. Two different fuel system configurations are currently in service. The first DLE gas turbines used a four valve system. More recent units have a three valve system. 1.3.1. Four Valve System The four valve fuel system comprises a single main, or total, metering valve and three trim, or delta P regulator, valves, as illustrated in Figure 1.8. The system is described in more detail in Section 7.

Outer Manifold Trim Valve

A Manifold

Ten sets per engine

Cup

(Outer Ring)

Pilot Manifold Trim Valve

B Manifold

FUEL SHUT-OFF (Provided by Packager)

PGAS Measurement

Cup

Outer Manifold Pressure (GP3O)

GP2 Measurement GP1, TFUEL Measurement

Cup Staging Valve

Fuel Metering Valve

Pilot Manifold Pressure (GP3P)

Thirty per engine Cup

(Pilot Ring)

(Lean Blow-out Enhancement)

Staging Valve

Fifteen per engine Inner Manifold Pressure (GP3I)

C Manifold

LM2500 Fuel System Schematic Four Valve Trim Fuel System

Every other Pilot Cup

Staging Valve

Cup

(Inner Ring)

Fifteen sets per engine Outer Manifold Trim Valve

Figure 1.8 Four valve fuel system The main metering valve is positioned in response to a total fuel flow demand, whereas the trim valves are positioned in response to delta P demands. As described later, the delta P demands are calculated in the control based upon the relative fuel flows required in each ring.


1.3.2. Three Valve System The three valve fuel system is more straightforward in that it employs three metering valves that are independently positioned in direct response to the fuel flows required in each ring. The three valve fuel system requires two orifices to be connected, one between the pilot manifold and the outer manifold, the other between the pilot manifold and the inner manifold. These orifices limit the manifold pressure build up in a non flowing ring. This reduces the initial fuel flow pulse, and therefore flame temperature, when a ring is first fueled (i.e. the first staging valve is opened). The three valve fuel system configuration is shown in Figure 1.9 and described in more detail in Section 7.

GP1O Outer Inlet Pressure

Outer Metering Valve

TFUEL Measurement

Optional I/F

GP2P - Pilot Manifold Pressure

Cup Staging Valve

Cup

Ten sets per engine

Cup

Inner Metering Valve

Thirty per engine Cup

(Pilot Ring)

)(

Optional I/F

PGAS Measurement Pressure Relief Orifice

(Outer Ring)

)( B Manifold

FUEL SHUT-OFF (Provided by Packager)

GP1P- Pilot Inlet Pressure

GE

A Manifold

FUEL SYSTEM PACKAGER GP2O - Outer Manifold Pressure

)(

(Lean Blow-out Enhancement)

Staging Valve

Fifteen per engine

GP2I - Pilot Manifold Pressure

Pilot Metering Valve

Every other Pilot Cup Optional I/F C Manifold

GP1I- Inner Inlet Pressure

Staging Valve

Cup

(Inner Ring)

Fifteen sets per engine

LM2500 Fuel System Schematic Three Metering Valve System

Figure 1.9 Three valve fuel system


1.4.

Bleed System

As mentioned earlier, in order to limit the variation in combustor flame temperature, the combustor configuration is changed from B-only for starting and idle operation to ABC for high power operation. However, this alone is not sufficient to keep the combustor flame temperature between the blowout and thermal stress limits. Changing combustor configuration changes the local fuel-air ratio in each cup by changing the fuel flow to each cup. Another way to change a cup fuel-air ratio is by varying compressor bleed in order to change the combustor airflow. By changing combustor configuration and modulating compressor bleed, local fuel/air ratio, and hence flame temperature, can be kept within limits across the entire power range, as illustrated in Figure 1.10. ABC

INCRE

ASING

BLEED

Max Tflame AB Max Bleed BC

BC/2 mode

Min Bleed Min Tflame

B

Staging Transition Points POWER Figure 1.10 Effect of combustor configuration and compressor bleed

For the DLE gas turbine, two bleed valves are added (eighth stage compressor bleed and compressor discharge bleed). For the LM6000 an existing variable bleed valve (VBV) is also used to provide additional bleed air modulation.


1.5.

Fuel Control

Fuel control, in the context of the DLE control system, refers to that part of the digital control system that determines the total combustor fuel flow demand (WF36DMD). The total fuel flow demand is subsequently split into three ring fuel flow demands based upon individual ring (A and C) combustor flame temperature demands. The DLE fuel control is very similar to previous non DLE (single annular combustor) LM2500 and LM6000 fuel control systems. The only real change is the addition of maximum and minimum fuel flow limits that correspond to maximum and minimum bulk flame temperature limits. These limits are encountered primarily during starting, decelerations in B-only configuration, operation at maximum power in ABC configuration and briefly during rapid transients. The fuel control comprises a set of regulators and fuel flow limiters, that through a series of min/max selects, often referred to as the priority selection logic, output a single fuel flow demand (WF36DMD). Regulators adjust fuel flow to regulate an engine variable (power turbine speed, gas generator speed etc.), whereas fuel flow limiters directly apply upper or lower fuel flow limits to the fuel flow demand (min fuel flow, max fuel flow). Only one regulator or fuel flow limiter can be in control at any time. The regulators and fuel flow limiters for the LM2500 and LM6000 are listed in Table 1.3. The “REGULATOR” parameter, as shown in this table, is available in the control and can be monitored to determine which regulator (1 thru 8 for the LM2500 or 1 thru 10 for the LM6000) is active at any time. LM2500

LM6000

REGULATOR Power turbine speed 1 Gas generator speed 2 Gas generator decel speed rate 3 Gas generator accel speed rate 4 Min fuel flow 5 Max turbine temperature 6 Max gas generator speed 7 Max fuel flow 8 May be used by control vendors 9 for application-specific purposes 10

Power turbine speed Core speed Core decel speed rate Core accel speed rate Min fuel flow Max turbine temperature Max compressor discharge press. Max compressor discharge temp. Max core speed Max fuel flow

Table 1.3 Fuel Control Regulators

1.6.

Flame Temperature Control

The original DLE control strategy, proposed for the LM6000, provided control of the average or bulk flame temperature. During development engine testing it became clear that potentially damaging high dynamic pressures in the 300 Hz to 700 Hz frequency range could occur with the DLE combustor. To avoid these high dynamic pressures, often referred to as combustor acoustics, and also ensure each fueled ring remained lit, it became necessary to control the flame temperature independently in each of the rings. The net result was a strategy that independently controlled the A ring and C ring flame temperatures, as well as the bulk flame temperature. 1.6.1. Bulk Flame Temperature For the bulk flame temperature for each combustor configuration a flame temperature window is defined as illustrated in Figure 1.11.

MAX BULK FLAME TEMP/FUEL FLOW LIMIT

“STAGING UP” TRANSITION POINT

MAX BLEED AIRFLOW CONTROL REGULATION “STAGING DOWN” TRANSITION POINT

INCREASING TFLAME GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 MIN BLEED 15 INCREASING BLEED

T4

MIN BULK FLAME TEMP/FUEL FLOW LIMIT POWER,T3

Figure 1.11 Bulk flame temperature window The upper boundary (TFLMAX) in general indirectly determines the maximum pilot flame temperature and hence also the maximum NOX level, whereas the lower boundary (TFLMIN) in general indirectly sets the pilot lean operating line. The bulk maximum and minimum flame temperatures (TFLMAX and TFLMIN) are scheduled in the control as a function of combustor configuration and T3. The left-hand or low power boundary is defined by the maximum compressor bleed capability, and the right-hand or high power boundary corresponds to zero bleed. As power varies the control adjusts bleed so that the bulk flame temperature is maintained at a demanded level between the min and max limits, until either the maximum or zero bleed limit is reached. This control concept is used for both the LM2500 and LM6000. For the LM2500 the bulk flame temperature is maintained at the “50% level” (i.e. mid way between the min and max limits), whereas for the LM6000 the bulk flame temperature is maintained at the “50% level” for operation in the B, BC/2, and BC configurations, but is reduced to the “25% level” (i.e. closer to the min limit), whenever possible, in the AB and ABC configurations. During power increases, bleed is progressively decreased, until zero bleed is reached, whereupon bulk flame temperature increases toward the maximum limit. Just before the maximum limit is reached, unless already in the ABC configuration, staging to the next combustor configuration “up” is initiated. During power decreases, the bulk flame temperature is maintained at the demanded level until maximum bleed is reached, whereupon the bulk flame temperature decreases toward the minimum limit. Just before the minimum limit is reached, unless already in the B configuration, staging to the next configuration “down” is initiated.


1.6.2. A and C Ring Flame Temperature As mentioned earlier, to avoid high combustor dynamic pressures etc., it became necessary to control the flame temperature independently in the A and C rings. Unlike the bulk flame temperature, which is controlled between maximum and minimum limits, the A and C ring flame temperatures always track reference schedules. The ring reference schedules are, like the bulk flame temperature min and max schedules, programmed in the control as a function of combustor configuration and T3. Figure 1.12 shows a typical set of ring and bulk flame temperature schedules for a LM6000.

3500

B Mode

TFLAME MAX

3400

Tflame (deg F)

3300 3200 3100 3000 TFLAME MIN

2900 2800 2700 2600 0

100

200

300 T3 (deg F)

400

500

600

TFLAME MAX

3500

BC/2 Mode

3400 3300 Tflame (deg F)

TFLAME MIN

3200 3100 3000 2900 2800

TFLAME INNER

2700 2600 200

300

400

500 T3 (deg F)

600

700

800

Figure 1.12a Specification LM6000 ring and bulk flame temperature schedules - B and BC/2 modes


3700 3600

B C M o de TFLA M E M A X

Tflame (deg F)

3500 3400 3300

TFLA M E M IN

3200 3100 3000 T FLA M E INNE R

2900 2800 400

500

600

700 T3 (deg F )

800

900

1000

3500

A B M o de

Tflame (deg F)

3400 3300

TF LA M E M AX

3200

TF LA M E M IN

3100 3000 TFLA M E O U TE R

2900 2800 2700 2600 600

700

800

900 T3 (deg F)

1000

1100

1200

3500 3400

A B C M o de

Tflame (deg F)

3300 3200 TFLA M E IN NE R

3100 TFLA M E M A X

3000 TFLA M E M IN

2900 TFLA M E O U TE R

2800 2700 2600 600

700

800

900 T 3 (d eg F )

1000

1100

1200

Figure 1.12b Specification LM6000 ring and bulk flame temperature schedules - BC, AB, and ABC modes


2.

Flame Temperature Algorithm

Because the combustor flame temperature cannot be measured directly in a reliable and accurate manner, it is estimated based on fuel flow demands and a “physics” based calculation of combustor airflow. The algorithm comprises four main sections, as illustrated in Figure 2.1. Engine inlet pressure

P2SEL

Engine inlet temperature

T2SEL

Comp. discharge pressure

P3SEL

Comp.discharge temp

T3SEL

CDP bleed flow

WB3Q

Stage 8 bleed flow

WB26Q

Gas gen speed

NGGSEL

Gas gen speed rate

AIRFLOW CALCULATION

NGGDOT

Fuel lower heating value Total fuel flow demand

LHVSEL WF36DMD

F_PFL

F_H3

F_WA36

T3 TFLMAX BRNDMD TFLMIN

BULK TFLAME TO FUEL FLOW

TFLOREF

TFLIREF

OUTER & INNER RING TFLAME TO FUEL FLOW

WFMX WFMN

to staging control and airflow control (lm6000

}

WFOREFABC

WFIREFABC

}

SWCOMB SWCOMBO

Control Display TFLAMEPCT = F_TFLCYC - TFLMIN *100 TFLMAX - TFLMIN

SWCOMBI

WF36DMD

FUEL FLOW TO BULK TFLAME

F_TFLCYC

Figure 2.1 LM2500 DLE flame temperature algorithm


to fuel flow split demands

}

air flow splits

to airflow control (lm2500 only)

Engine sensor and control information is used to calculate the combustor airflow (F_WA36) based on an assumed HP turbine flow function. Once combustor airflow is known, then the combustor fuel flows (WFMX and WFMN for the bulk limits, and WFOREFABC and WFIREFABC for the ring demands) can be calculated for the given scheduled flame temperatures (bulk TFLMAX and TFLMIN, and ring TFLOREF and TFLIREF respectively), and also bulk flame temperature (F_TFLCYC) can be calculated based on the current bulk or total fuel flow demand (WF36DMD). Note that the LM2500 uses the bulk flame temperature F_TFLCYC as feedback to the airflow control, whereas in the LM6000 the bulk flame temperature is for monitor purposes only and feedback to the airflow control effectively is derived from WFMX and WFMN. The differences between the LM2500 and LM6000 flame temperature algorithm/airflow control interface are illustrated in Figure 2.2.


[from fuel control] WF36DMD

LM6000 IMPLEME PS3DMD

N +

P+I regulator

-

D

LM2500 IMPLEMENTATION:

PS3ERR

TFLDMD

WF/PS3 DEMAND [from Tflame alg.]

TFLERR +

-

P+I Regulator

DWB36 [total bleed demand]

PS3EST [from sensor compensation]

F_TFLCYC [from Tflame algorithms]

LM6000 EQUIVALENT TO PS3ERR [from fuel control] WF36DMD WFMX

[from Tflame algorithm]

{

N

+

D

“WFMID”

WFMN

P+I regulator

-

1.0

PS3EST

Position in Tflame window [0.5 - 0.25]

Figure 2.2 Flame temperature algorithm/airflow control interface


DWB36 [total bleed

Primary flame temperature influences include combustor fuel flow, fuel lower heating value (LHV), compressor discharge temperature (T3) and compressor discharge pressure (PS3). Note that errors in flame temperature due to fuel flow are a result of errors in the fuel metering system (i.e. differences between actual and demanded fuel flows). The accuracy, or more importantly, the consistency of the calculated flame temperature is of course influenced by all of the algorithm inputs. Sensitivity studies performed during the design and development of the LM2500 and LM6000 control systems illustrate the relative importance of all algorithm inputs. Errors in these inputs can cause unpredictable or erratic behavior of the overall system. Figure 2.3 provides a chart summarizing the influence of all control variables on flame temperature for the LM2500. This chart shows the average variation in bulk and ring flame temperatures for specific perturbations in each of the control variables. The perturbation magnitudes chosen are based on control specification accuracies. This chart clearly shows that compressor discharge pressure (PS3), fuel flow (WFAGMV) and lower heating value (LHV) have the biggest effect on flame temperatures. Figures 2.4, 2.5, and 2.6 show the sensitivity to these parameters for each combustor configuration. In the field, errors in SG as well as LHV are often encountered. Figure 2.7 shows the sensitivity to SG error. It should be noted that although variable stator vanes (VSV) appear to have a large effect on bulk flame temperature, this only occurs when the airflow is commanding the bleeds fully closed, and under these circumstances bulk flame temperature is not being regulated by the control.

Tflame Sensitivity cdpsel (0.707%) Fpv(wfagmv) (1.0%) gp3isel (0.75 psi) gp3osel (0.75 psi) gp3psel (0.75 psi) K(wfagmv) (1.0%) lhvsel (1.0%) nggsel (10 rpm) p2sel (1.003 psi) ps3sel (3.162 psi) Parameter

sgsel (1.0 %) st8sel (0.707%)

Bulk

swcmbi (1.0%)

Inner Outer

swcmbo (1.0%) swcmbp (1.0%) t2sel (1.414 degF) t3sel (6.5 degF) tfuelsel (2.83 degF) wb26q wb3q (5.0%) wf36innz (1.0%) wf36otnz (1.0%) wf36plnz (1.0%) wfagmv (3.0%) VSV (3 deg) 0

5

10

15

20

25

30

35

Tflame Variation (deg F)

Figure 2.3 LM2500 Flame temperature sensitivity High “indicated” PS3 = low “indicated” Tflame

PS3 Effect on Tflame 3.162 psia Change in PS3 160 140 120 100 80 deg F 60 40 20 0

Bulk Inner Outer

IDLE B

NO B

HI BC

LO BC

NO BC

HI AB

LO AB

NO AB

HI LO NO ABC ABC ABC


Figure 2.4 LM2500 Flame temperature sensitivity due to PS3 variation

High “indicated” fuel flow = high “indicated” Tflame Metering Valve Fuel Flow Effect on Tflame 2.0% Change in WFAGMV for WF>2700 pph 3.0% Change in WFAGMV for WF<2700 pph 70 60 deg F

50

Bulk

40

Inner

30

Outer

20 10 0 IDLE B

NO B

HI BC

LO BC

NO BC

HI AB

LO AB

NO AB

HI ABC

LO ABC

NO ABC

Figure 2.5 LM2500 Flame temperature sensitivity due to WFAGMV variation

High “indicated” LHV= high “indicated” Tflame

LHV Effect on Tflame 1.0% Change in LHV 0 -5

IDLE B

NO B

HI BC

LO BC

NO BC

HI AB

LO AB

NO AB

HI ABC

LO ABC

NO ABC

deg F

-10 Bulk

-15

Inner -20

Outer

-25 -30 -35

Figure 2.6 LM2500 Flame temperature sensitivity due to LHV GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 23

High “indicated” SG = high “indicated” Tflame

Specific Gravity Effect on Tflame 1.0% Change in SG

10 5

deg F

0

Bulk Inner

-5

Oute r

-10 -15 -20 IDLE B

NO B

HI BC

LO BC

NO BC

HI AB

LO AB

NO AB

HI ABC

LO ABC

NO ABC

Figure 2.7 LM2500 Flame Temperature sensitivity due to SG


3.

Fuel Properties 3.1.

Effect of fuel properties

As already mentioned fuel lower heating value (LHV) directly affects the relationship between fuel mass flow and flame temperature. A 1% variation in LHV has the same effect on flame temperature as a 1% variation in fuel mass flow, i.e. it is Btu/hr that is important as far as flame temperature is concerned. Other fuel properties affect fuel mass flow metering. The DLE application requires precise control of metered fuel flow, which in turn means precise characterization of individual metering valves. A gas metering valve is generally characterized by the following equations: P1 P2 T1 k Sg Ae Wf

Pr = max

2

Inlet absolute pressure [lbf/in ] 2 Discharge absolute pressure [lbf/in ] Inlet fuel temperature [deg R] Ratio of specific heats Cp/Cv [-] Specific gravity [-] 2 Metering valve effective area [in ] Fuel mass flow [lbm/hr]

P2 P1

2 1+k

,

unchoked

k/(k-1)

choked

2/k Wf = 3953.73 * Ae * P1*

k Sg T1 * k-1 *

Pr

(k+1)/k -

1/2 unchoked or choked flow

Pr

At idle and low power conditions for choked flow conditions, i.e. when simplifies to:

P2 < P1

2 1+k

k/(k-1)

1/2 GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 25

the above equation

Wf = 2797.52 * Ae * P1 *

Sg * k T1

*

2 1+k

(k+1)/(k-1) choked flow

In order to achieve the required fuel metering accuracy the fuel system supplier has to characterize each main metering valve in a four valve fuel system and the three metering valves in a three valve fuel system. The control has to be programmed with specific characterization tables for each installation. Units are often a source of confusion, particularly those for LHV. In the GE control specification LHV has units of Btu/lbm. Often when working at field installations LHV is calculated in different units and requires conversion for use with GE control algorithms. The following conversions can be used under these circumstances: 1 joule = 0.7376 lb-ft = 1 N-m 1 Kj = 737.60 lb-ft 1 Btu = 778.169 lbf-ft scf = standard cubic feet at 60 deg F and 14.696 psia scm = standard cubic meter at 60 deg F and 14.696 psia 1 scm = 35.3198 scf ncm = normal cubic metre (standard for European compressor industry) at 0 deg C and 14.696 psia Sg = specific gravity of gas Wa = weight density of air at 60 deg F and 14.696 psia = 0.0764 (lbm/scf) Wa = weight density of air at 0 deg C and 14.696 psia = 0.0807 (lbm/ncf) Wg = weight density of gas = Sg * Wa


3.2.

Ring fuel nozzle scalars

Gross errors in fuel properties affecting fuel metering accuracy, if not so large as to result in unstable control operation, may be found by monitoring the total ring fuel flow scalar WFNOZTOTFF for a 4-valve fuel system or the individual ring fuel flow scalars WFNOZOTRFF, WFNOZPILFF and WFNOZINRFF for a 3valve fuel system. These variables are calculated in the control as the ratio of the demanded fuel flow to the estimated fuel nozzle flow. The estimated fuel nozzle flows are calculated based on assumed fuel nozzle flow functions. With a perfect metering system and exact fuel nozzle flow functions these scalars would be = 1.0. Values in the range 0.9 to 1.1 can be considered typical. Once errors in fuel metering accuracy have been ruled out, by monitoring the fuel nozzle flow scalars as described above, gross errors in LHV if not so large as to result in unstable control operation may be found by comparing bulk flame temperature F_TFLCYC with the measured gas turbine temperature T54SEL or T48SEL

3.3.

Calorimeter and Chromatograph Operation

The DLE engine has a requirement for lower heating value and specific gravity inputs to schedule fuel flow accurately. The fuel system supplier might also have requirements to have other gas property inputs. Two types of commercially available instruments which provide some of these inputs are calorimeters and chromatographs. A calorimeter takes gas from the pipeline and air and burns them under closely controlled conditions. The resulting exhaust gas temperature is measured by a precise sensing system. A controller regulates changes in the quantity of combustion and cooling air to the burner. This serves to maintain the exhaust gas at a constant temperature. So, if the heating value of the gas is changing, the required air flow will be modified by the controller to maintain constant exhaust gas temperature. By measuring the change in airflow, fuel gas properties can be determined. Typical outputs from a calorimeter are specific gravity, calorific value, and the Wobbe index. From the calorific value and specific gravity, the lower heating value can be calculated. Calorimeters do not provide a gas constituent breakdown - i.e. mole percent methane, mole percent O2, etc. Since calorimeters are continuously burning gas from the pipeline, they respond quickly to gas property changes. The time for calorimeters to respond to changes in gas properties is usually less than twenty seconds. Calorimeters would be used when changes in gas properties occur frequently. A chromatograph provides a breakdown of the gas constituents. From the breakdown, the lower heating value and specific gravity can be calculated. A chromatograph typically consists of the following components: 1.

A carrier gas system for transporting the component through the column at a constant flow rate.

2.

A chromatographic column for separating the sample into individual components.

3.

A detector for detecting the components in the gas.

Gas from the pipeline is injected into the chromatograph. Inside the chromatograph, the pipeline gas is mixed with a carrier gas such as helium. The carrier gas takes the pipeline gas and flows it through the chromatographic column and the detector. The carrier gas needs to be inert to avoid interaction with the pipeline gas. The chromatographic column consists of around 30 feet of quarter inch diameter tubing. The column acts as a barrier to the gas flow. As the carrier and pipeline gas mix flow through the column, the gas constituents are broken down. The lighter gas constituents are able to move through the column more easily than the heavier ones. Thus, they are the first to exit the column. So as the gas sample exits the column, it is grouped by gas constituents according to their molecular weight, the lighter constituents exiting first. The sample then goes to the detector. Here, the type and percentage of each constituent is determined. There are different type of detectors used among chromatograph manufacturers. One method GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 27

(used in GE test cells) is called thermal conductivity. Here the gas sample from the column is passed through a form of Wheatstone Bridge where the four arms of the bridge are heated to a precisely controlled, high temperature. When the constituents pass by one pair of the bridge elements, the wires lose heat which translates to a bridge resistance change. The amount of resistance change determines the type and percentage of each constituent. From the constituent type and percentage, the lower heating value, specific gravity, Wobbe index, ratio of specific heats, and compressibility can be calculated using a few formulae and the molecular weights of the constituents. The typical processing time for a chromatograph is around 5 minutes. Chromatographs should be used only where the gas properties are not changing rapidly or where knowing the other gas property information (ratio of specific heats, compressibility) is required.

4.

Starting 4.1.

Start Sequence

A description of the overall start sequence for the LM2500 and LM6000 can be found in Section 20.0 of the respective Installation Design Manual. Outputs from the start-run sequencer are shown in Table 4.1. ENGMODE

- Engine mode condition

STGVLVOPEN

- Enable normal control of staging valves

LM2500

Z_OP_STRTR Z_OPEN_STRTR

- Operate starter

LM2500 LM6000

Z_IGNDMD Z_IGNITORDMD

- Turn ignitor(s) on

LM2500 LM6000

Z_VENTDMD Z_VENT1DMD Z_VENT2DMD

- Open gas vent valve(s)

LM2500 LM6000 LM6000

FUELON

- Command fuel on

LM2500

Z_FSOV1DMD Z_FSOV2DMD

- Open shut off valve 1 - Open shut off valve 2

Table 4.1 Start-Run Sequencer Outputs Start-run engine modes, and control actions during start to idle, are similar for the LM2500 and LM6000 and are described in detail in Section 20.0 of the respective Installation Design Manual and are summarized for the LM2500 in Table 4.2.


ENGMODE 0

CONTROL ACTION

LOGIC ACTIONS

PRESTART Engine and control prestart checks

2

CRANK Check gas supply pressure above minimum start pressure

3

START Enable normal control of staging valves Turn on starter Turn off starter and STOP if NGGSEL does not reach 2000 rpm within 20 sec Open vent valve Purge for 2 minutes (purge time and speed set by site requirements) Close vent valve

4

Close shutoff valves, turn ignitor off and perform purge if light off not detected within 10 sec Turn ignitor off 10 sec after light off

Z_VENTDMD = False

Z_IGNDMD = True Z_FSOV1DMD = True Z_FSOV2DMD = True FUELON = True LITEOFF = True? Z_IGNDMD = False

NGGSEL>=4500? Z_OP_STRTR = False

RUN2 ESHUTDOWN if idle not achieved within 2 minutes

8

Z_OPSTRTR = False Z_VENTDMD = True

RUN1 ESHUTDOWN if starter cutout speed not reached within 90 sec Turn starter off when starter cutout speed reached

6

STGVLVOPEN = False Z_OP_STRTR = True

IGNITE Turn ignitor on After 2 sec open shutoff valves

5

PGAS>200

NGGSEL>=NGGIDL?

RUN3 Optional 5 minute idle warm -up

9

RUNNING ESHUTDOWN if “check power turbine rotation at idle” required and NPTSEL has not reached 350 rpm If “check power turbine rotation at idle” not required then raise NGG speed ref. as required by application DECEL TO IDLE if NPTSEL has not reached 350 rpm within 60 sec and ESHUTDOWN if NPTSEL has not reached 350 rpm within a further 5 minutes

10

SHUTDOWN

11

PURGE

12

ESHUTDOWN

13

STOP

Table 4.2 LM2500 Start-run engine modes and control actions during start to idle

4.2.

Control Operation GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 29

Control operation for starting to core-idle is similar for the LM2500 and LM6000. There are two phases. The first phase, which applies to the majority of the start, until the core approaches idle speed, is performed with the airflow control disabled and no eighth stage or compressor discharge bleed. The bulk Tflame schedules, during this initial phase have no effect. Instead a start fuel control calculates upper and lower bulk (or total) fuel flow limits (WFMAX = WFMAXSI and WFLBO = WFLBOSI respectively) based on independent max. and min. equivalence ratio schedules. These schedules were originally intended to correspond to combustor thermal stress and lean blowout limits but were ultimately adjusted during the LM6000 development engine testing to provide reliable starts (blowout-free) with acceptable combustor acoustic levels. Note that for the LM6000, the VBV’s throughout a start, are scheduled just as for their non DLE counterparts, i.e fully open (100%) once LP rotor speed reaches 1250 rpm. As far as the overall fuel control is concerned, during the first this initial start phase, in addition to the upper and lower fuel flow limits, two other fuel flow regulators/limiters can come into play, i.e. a core speed acceleration rate regulator and a max. WF/PS3 accel schedule limit. The WF/PS3 accel schedule limit exists in both the LM2500 and LM6000 controls, and is based on their non DLE counterparts, but in general is encountered only on the LM2500. The schedules when originally developed for the non DLE engines were intended to provide compressor stall protection. The WF/PS3 accel schedule limit is “merged” with the start fuel control upper limit WFMAXSI through a Min select to form the final WFMAX upper limit. By virtue of the fuel control priority selection logic, the WFMAX upper limit will always override the WFLBO lower limit, which means that it is possible for the WF/PS3 accel schedule limit to override both the upper and lower start fuel control fuel flow limits. A leaking or badly calibrated PS3 pressure transducer, resulting in a low sensed pressure, could in turn result in the WF/PS3 accel schedule inadvertently lowering the final fuel flow and producing a hung or aborted (flame out) start. Note that although the start fuel control upper and lower fuel flow limits are, like the “idle-and-above” Tflame algorithm limits, a function of T3 and PS3, because of accuracy concerns in the start region the T3 and PS3 are from internal model estimates, rather than the sensors. Therefore, although errors in sensed PS3 will affect the WF/PS3 max. fuel flow limit, errors in neither PS3 nor T3 sensed values will affect the start fuel control upper and lower fuel flow limits. LHV is an input to the start fuel control and so errors in LHV will affect the start fuel control upper and lower limits! When the second phase of the start to core-idle is entered the complete airflow control/bulk flame temperature control strategy is enabled, and the upper and lower fuel flow limits come from the Tflame algorithm as opposed to the start fuel control. Transition from the first phase to the second phase is strictly a function of core speed. At a specific core speed (N25SEL = N25SIATV = N25SI + N25SIJA = 6300 rpm for the LM6000 and NGGSEL=NGGSI =4900rpm for the LM2500) the airflow control is enabled and as the core approaches that same specific speed the fuel flow upper and lower limits transition from the start fuel control limits to the Tflame algorithm limits. This occurs over the core speed range of 6200 to 6300 rpm for the LM6000 and 4800 to 4900 rpm for the LM2500. Other DLE-specific control actions occur during the first phase of the start. When the IGNITE mode of the start is entered, in addition to the opening of the shutoff valves and the energizing of the ignitor, in order to ignite the fuel, the outer staging valve(s) that supplies fuel to the three combustor cups alongside the energized ignitor(s) is opened. At this point all of the inner staging valves are closed. The outer staging valve(s) is open for the complete ten seconds of the IGNITE mode. Note that both the LM2500 and LM6000 have provision for two ignitor locations. The staging control logic will open staging valve #22 and/or #9 depending upon whether ignition demands IGN1DMD and/or IGN2DMD are set during the IGNITE mode. When either or both of these staging valves are open during the IGNITE mode, the outer ring fuel flow is determined just as it is for operation above idle in AB or ABC mode, i.e. as described earlier, the outer fuel flow WFOREFABC is calculated in the Tflame algorithm based on a scheduled ring flame temperature TFLOREF. This fuel flow represents the outer fuel flow per staging valve and, depending upon the number of “ignition” outer staging valves open (one or two), is translated into a total outer ring fuel flow demand (WFOREF). Being that the outer fuel flow is derived from the Tflame algorithm, it will be influenced by errors in any Tflame inputs, in particular PS3 and T3. So, during the IGNITE mode, a bulk or total fuel flow is demanded (WF36DMD) , and from this is subtracted the outer ring fuel demand (WFOREF) to give a resultant pilot ring fuel flow demand (WFPREF). The inner ring is not fueled during the IGNITE mode, only the pilot ring and three or six of the outer ring combustor cups (one or two of the outer staging valves). As the start progresses, part of the inner combustor ring can also be fueled. The logic that determines this is fairly straightforward and functions as follows: If the demanded fuel flow (WF36DMD), under the influence of the core speed accel rate regulator, in attempting to track the core speed accel rate schedule, is forced onto the start fuel control upper limit (WFMAXSI) for more than three seconds then the combustor configuration transitions from B (pilot-only) to BC/2 (pilot plus 8 of the 15 inner staging valves or combusor cups). The start fuel control operates to control the avaerage fuel-air ratio in each combustor cup. Increase GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 30

the number of cups from thirty plot cups to thirty pilot cups plus eight inner cups, and the upper and lower fuel flow limits will increase accordingly. This allows WF36DMD to increase and thereby “speed-up” the start. The inner ring fuel flow is determined just like the outer ring, i.e. in the same manner as it is for above-idle operation in BC/2 (LM6000 only), BC or ABC mode. In summary: All three rings can be fueled during a start with total fuel flow being determined by the core rate regulator and limited by Tflame max. and min. fuel flow limits. Errors in fuel properties (SG, fuel temp. and Cp/Cv) affect mass flow metering accuracy and errors in LHV affect max. and min. fuel flow limits - problems with either of these can hang-up a start or prevent a lite-off completely - it’s important to recognize that a 5 % error in fuel flow can mean approximately a 150 deg F error in Tflame! Remember that max fuel flow has an overriding WF/PS3 limit and that PS3 sensor calibration or leaks resulting in low PS3 can cause the WF/PS3 accel limit to hang-up a start. Also remember that PS3 and T3 sensors do not affect bulk Tflame limits in the initial phase of the start before the airflow/Tflame control is enabled, but they do affect the outer ignition fuel flow/Tflame and the inner fuel flow/Tflame if staging to BC/2 occurs. The outer ignition fuel flow is essential for lite-off to occur - the appropriate staging valve must be opened, i.e. the one that fuels the cups alongside the energized ignitor. At a core speed of 4900 rpm for the LM2500 and 6300 rpm for the LM6000 the airflow/Tflame control is enabled and the fuel control fuel flow/Tflame limits come from the Tflame algorithm. At this point the bulk Tflame min. and max. schedules become effective. Fuel metering or fuel property errors can at this transition point result in a blowout! Typical start characteristics for both the LM2500 and LM6000 are shown in Fig 4.1 and 4.2 respectively.

WF36DMD

WF36DMD - pph

1600 1200 800 400 0 0

20

40

60

80

100

120

140

80

100

120

140

Time - sec

NGGSEL

NGGSEL - rpm

8000 6000 4000 2000 0 0

20

40

60 Time - sec


NPTSEL

NPTSEL - rpm

3000 2500 2000 1500 1000 500 0 0

20

40

60

80

100

120

140

Time - sec

Figure 4.1a Typical LM2500 DLE start characteristics

T54

T54SEL - deg F.

1000 800 600 400 200 0 0

20

40

60

80

100

120

140

100

120

140

Time - sec

BRNDMD 10 BRNDMD

8 6 4 2 0 0

20

40

60

80 Time - sec


DWB36PCT

DWB36PCT - %

4 3 2 1 0 0

20

40

60

80

100

120

140

Time - sec

Figure 4.1b Typical LM2500 DLE start characteristics

PX36SEL - psi peak to peak

PX36SEL 5 4 3 2 1 0 0

20

40

60

80

100

120

Time - sec

Figure 4.1c Typical LM2500 DLE start characteristics


140

N25SEL

N25SEL - rpm

8000 6000 4000 2000 0 0

20

40

60

80

100

120

80

100

120

Time - sec

WF36DMD

WF36DMD - pph

2000 1500 1000 500 0 0

20

40

60 Time - sec

Figure 4.2a Typical LM6000 DLE start characteristics


T48SEL 1200

T48SEL - deg F.

1000 800 600 400 200 0 0

20

40

60

80

100

120

80

100

120

80

100

120

Time - sec

BRNDMD 10 BRNDMD

8 6 4 2 0 0

20

40

60 Time - sec

N2ROTOR

N2ROTOR - rpm

2500 2000 1500 1000 500 0 0

20

40

60 Time - Sec

Figure 4.2b Typical LM6000 DLE start characteristics


DWB36PCT

DWB36PCT - %

100 80 60 40 20 0 0

20

40

60

80

100

80

100

120

Time - sec

PX36SEL - psi peak to peak

PX36SEL

5 4 3 2 1 0 0

20

40

60

120

Time - sec

Figure 4.2c Typical LM6000 DLE start characteristics

5.

Operation at Idle

5.1 Core Idle Control operation at core-idle is very similar for both the LM2500 and LM6000. The combustor operates in the pilot-only (B) mode (BRNDMD = 0) and the total fuel flow ( = pilot fuel flow) is adjusted by the fuel control core speed regulator (REGULATOR = 2) to set the core speed at the core-idle reference setting. For the LM2500, core idle speed is set to a nominal physical speed setting of NGGFLOOR = 6800 rpm. For the LM6000, core idle speed varies as a function of T2, decreasing as T2 increases ( 7819.3 rpm at 0 deg F, 7678.0 rpm at 48 deg F, 7409 rpm at 80 deg F). The airflow control is active and adjusts bleed in order to regulate bulk flame temperature. Both the eighth stage and compressor bleed may be used depending on the level of bleed required. For the LM6000 the VBV, depending on T2, may be scheduled fully open irrespective of the level of bleed required. The bleed sequencing is described in more detail in section 6.0. Typical characteristics to expect at core idle, for key parameters are given in Tables 5.1 and 5.2 for the LM2500 and LM6000 respectively, ENGINE FACTORY TEST 2/28/96

T2

(deg F)

34


LHVSEL

(BTU/lbm)

20696

BRNDMD

(-)

0

NGGSEL

(rpm)

6800

WF36DMD

(lbm/hr)

1462

TFLMIN

(deg F)

2723

TFLMAX

(deg F)

3586

TFLCYC

(deg F)

3158

DWB36PCT

(%)

59

T3SEL

(deg F)

350

T54SEL

(deg F)

794

PX36SEL

(psi peak-peak)

0.5

Table 5.1 LM2500 Typical core idle parameters


GE SIMULATION

ENGINE FACTORY TEST 9/20/94

ENGINE 109-208 SILKEBORG SITE 4/23/96

T2 (deg F)

59

76.7

57

LHVSEL

20400

-

-

BRNDMD

0

0

0

N25SEL (rpm)

7617

7442

7717

WF36DMD (pph)

2479

1972

1977

T3SEL (deg F)

508

443

511

TFLMIN (deg F)

2900

3000

2700

TFLMAX (deg F)

3400

3500

3200

TFLCYCS (deg F)

3205

3273

2966

DWB36PCT (%)

0

68

53

T48SEL (deg F)

918

848

848

PX36SEL (psi p-p)

-

0.34

0.58

Table 5.2 LM6000 Typical core idle parameters 5.2 LM6000 Core Idle to Sync Idle Transition The original staging logic design assumed that there may not be “overlap” between core idle and sync idle, i.e. as accelerating from core idle to sync idle it may not be possible to transition directly from zero bleed B mode to high bleed BC/2, if required, and stay within the bulk Tflame limits. Therefore the staging logic was developed to provide partial staging when transitioning from core idle to sync idle. The logic functions as follows - as the gas turbine is accelerated from core idle and zero bleed/max. bulk Tflame is reached one inner staging valve is opened (BRNDMD is incremented) which results in bulk Tflame reducing. This process is repeated every time zero bleed/max. bulk Tflame is reached until BC/2 configuration is reached ( 8 inner staging valves open - BRNDMD = 8). At that point staging logic strategy changes to the scheme that is used from synch idle to max power as described in Section 8.

6.

Operation within a combustor configuration window

6.1 Flame temperature control As previously described in Section 1.6, a combustor configuration window is defined in terms of bulk flame temperature upper and lower boundaries and bleed upper and lower boundaries. The airflow control adjusts bleed in order to regulate the bulk flame temperature, until either max or min. bleed is reached. Bleed decreases to min. as power is increased. Further increases in power at that point will cause the bulk flame temperature to increase toward the max. upper limit. Conversely, bleed increases to max. as power is decreased, and once max bleed is reached further decreases in power will cause the bulk flame temperature to decrease toward the min. lower limit. This was illustrated in Figure 1.8. Note that in other than pilot-only (B) mode, the respective ring flame temperatures (outer (A) and/or inner (C) ) are also being controlled. Unlike the bulk flame temperature, which is regulated by varying bleed, ring flame temperatures are controlled by varying the fuel flow split between the fueled rings. Unlike the bulk flame temperature, ring flame temperatures continue to be controlled when min. or max. bleed limits are reached. This means that as power is increased when the bleed is at the min. limit, the ring flame temperature(s) continue to follow the scheduled demand, but the bulk flame temperature increases toward the max. limit. This results in the pilot ring flame temperature increasing. The max. bulk flame temperature GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 38

schedule and the ring flame temperature schedule(s) at this min. bleed condition therefore determine how hot the pilot ring will get. Raising the bulk max flame temperature schedule or lowering the ring flame temperature schedule(s) at min. bleed will raise the pilot ring flame temperature. Similar but opposite effects occur at max. bleed. In summary, rasing or lowering bulk flame temperature at constant power results in increasing or decreasing bleed, and increasing or decreasing pilot flame temperature with very little change in the inner and/or outer flame temperatures; whereas raising or lowering either the inner or outer flame temperature at constant power results in very little change in bleed and bulk flame temperature, but results in decreasing or increasing pilot flame temperature. 6.2 Bleed Sequencing Varying bleed between min. and max. levels involves use of eighth stage compressor bleed (ST8) and compressor discharge bleed (CDP). In addition, for the LM6000 the VBV’s are modulated between min. and max. schedule limits. The bleeds are operated sequentially. The airflow control/bulk flame temperature regulator outputs a total bleed flow demand (DWB36) which can vary between zero and a maximum allowable limit (DWB36MAX). The total bleed demand is generally monitored as a percentage of max (DWB36PCT) and varies between 0 and 100%. The total bleed demand is translated into bleed valve position demands. The LM2500 bleed sequence is simpler than the LM6000. The LM2500 uses 0 to 12% CDP(percentage of core airflow W2) followed by 0 to 3% ST8. The LM6000 uses 0 to 2% ST8 (percentage of core airflow W25), followed by min. to max. VBV, followed by 2 to 10% ST8 and finally 0 to 3.5% CDP. The sequencing is illustrated in Figure 6.1 and 6.2 for the LM2500 and LM6000 respectively. In practice, the ST8 bleed on the LM2500 has proved to be very ineffective. Operation in the LM2500 ST8 modulating region tends to be very unstable with ST8 bleed valve scheduling either min (“off”) or max (“on”) bleed. VBVs are used on the LM6000 to provide additional bleed modulation as illustrated in Fig 6.3. DWB36MAX upper limit

TFLERR

P+I Regulator

DWB36

CDP 0-12%

ST8 0-3%

Stage 8 Bleed

Zero lower limit

CDP Bleed Control Display DWB36PCT=

DWB36 *100 DWB36MAX

Figure 6.1 LM2500 bleed sequence


Total Bleed Demand DWB36 DWB36MAX upper limit

P+I regulator PS3ERR

ST8 0 ->2%

VBV min->max

ST8 2 ->10%

CDP 0 ->3.5%

CDP BLD VLV DMD

ST8 BLD VLV DMD

Zero lower limit

VBV DMD Control Display DWB36PCT=

DWB36 *100 DWB36MAX

Figure 6.2 LM6000 bleed sequence

`

MAX

VBV [%]

MIN INCREASING BLEED

SUMP PRESSURIZATION LIMIT

Limits vary as a function of T2 also

BOOSTER STALL MARGIN LIMIT

N25R2 [ RPM]

Figure 6.3 LM6000 VBV bleed modulation

7.

Fuel Metering

7.1 Fuel System Demands As described in Section 1.3 there are two different fuel system configurations currently in service on the LM2500 and LM6000. The first DLE gas turbines used a four valve system, whereas more recent units have a three valve system. Schematics of the two systems are provided in Section 1.3. Both systems independently meter fuel to the three combustor rings in response to outer, pilot and inner combustor fuel flow demands WFOREF, WFPREF and WFIREF. As illustrated in Figure 7.1, WFOREF and WFIREF are GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 40

calculated directly from the flame temperature algorithm outputs WFOREFABC and WFIREFABC. WFOREFABC and WFIREFABC are the outer and inner demanded fuel flows per staging valve, and are multiplied by the respective number of staging-valves-open variables OTREST and INREST to determine the total outer and inner ring fuel flow demands WFOREF and WFIREF. These final inner and outer ring fuel flow demands are subtracted from the total fuel flow demand WF36DMD to provide the pilot ring fuel flow demand WFPREF. These three demands represent the required fuel flows at the combustor.

TOTAL FUEL FLOW DEMAND WF36DMD

TFLAME A REF

TFLAME TO FUEL FLOW CONVERSION

EXTENSION OF TFLAME BULK ALGORITHM

TFLAME C REF

TFLAME TO FUEL FLOW CONVERSION

+

+

-

-

WFOREFABC

FUEL FLOW PER STAGING VALVE

WFPREF

WFOREF

OTREST Number of A StagingValves Open

WFIREFABC

TO FUEL METERING SYSTEM

FUEL FLOW PER RING

WFIREF

INREST Number of A StagingValves Open

Figure 7.1 Ring fuel flow demands The subsequent logic that is used to translate the three combustor fuel flow demands into final fuel system demands varies significantly between the three valve and four valve systems, but in both cases employs a model-based gas volume dynamics compensation scheme.. The three valve system is more straightforward, but places heavier demands on the control processor because independent gas volume dynamics compensation is provided for each of the three rings (the four valve system has a single overall gas volume dynamics compensation algorithm). With the three valve scheme each of the three combustor fuel flow demands are input to a gas volume dynamic compensator and the outputs represent the metering valve fuel flow demands WFOTRDMD, WFPLTDMD and WFINRDMD that are passed to the fuel system supplier’s fuel metering system. For the four valve system the logic is not so straightforward. Referring back to the schematic of the four valve system provided in Section 1.3, one can see a main metering valve downstream of which are three trim or delta P regulator valves. The main metering valve controls the total combustor fuel flow, and the three trim valves vary the fuel flow split between each combustor ring. The total combustor fuel flow demand WF36DMD is input to a single gas volume dynamics compensator, the output of which represents the main metering valve fuel flow demand WFMVDMD that is passed to the fuel supplier’s fuel metering system. The individual ring fuel flows are controlled by varying the pressure drop across each of the trim valves. The pilot trim valve delta P follows a predefined schedule that is a function of total fuel flow demand. Using assumed flow functions for each of the three combustor staging valve/premixer fuel circuits the inner and outer trim valve demanded delta P’s relative to the pilot are calculated based on the the three combustor fuel flow demands WFIREF, WFPREF and WFIREF as illustrated in Figure 7.2. The resultant outputs from this logic are the three trim valve delta P demands DP2P3ODMD, DP2P3PDMD and DP2P3IDMD that are passed to the fuel supplier’s metering system. The actual pressure drops across the trim valves are measured using pressure taps located on the gas manifold (GP3OSEL, GP3PSEL and GP3ISEL) and located upstream of the trim valves (GP2SEL).


WF36DMD

WFMVDMD

VOL DYN COMP

WFMVDMD / WF36DMD

+

-

WFOREFABC TFLOREF

+

WFPLTDMD

WFPREF

-

TFLAME TO WF CALC

WFOTRDMD

OTREST WFIREFABC TFLIREF

TFLAME TO WF CALC

WFINRDMD

INREST WFPREF PILOT FUEL NOZZ FF (total)

GP3PREF

OUTER FUEL NOZZ FF (per stg vlv)

GP3OREF

(GP2 - GP3O) DMD = (GP2 - GP3P) - (GP3OREF-GP3REF)

WFOREFABC

-

-

DP2P3ODMD

+ +

WFIREFABC INNER FUEL NOZZ FF (per stg vlv)

GP3IREF

DP2P3IDMD

+ +

WF36DMD PILOT DELTA P SCHED

DP2P3PDMD (GP2-GP3P)

Figure 7.2 Trim valve delta P demands

7.2 Monitor Fuel Flows and flame temperature Included in the control, for monitor purpose only, are estimated ring fuel flow calculations. Fuel flow at the combustor is calculated for each ring using the measured pressure ratio across the staging valve and premixer fuel circuit. (GP2X/PS3 for a 3-valve fuel system and GP3X/PS3 for a 4-valve fuel system - where X=O, P or I) together with other relevant parameters that include the temperature of the gas fuel and the compressor discharge temperature (T3). The outputs from this calculation are estimated “raw” fuel flows for each ring - WFOTR, WFPIL, and WFINR. These “raw” fuel flows are estimates based on assumed flow functions for each staging valve + premixer fuel circuit. These “raw” fuel flows are corrected using the ring fuel flow scalars WFNOZTOTFF (for a 4-valve fuel system) and WFNOZOTRFF, WFNOZPILFF and WFNOZINRFF (for a 3-valve fuel system), that were described in section 3, to provide a best estimate of each ring fuel flow - WFOTRM, WFPILM, and WFINRM. From these ring fuel flows estimated ring flame temperatures are calculated - F_TFLODF, F_TFLPDF, and F_TFLIDF. These appear on the control display with the leading “F_” prefix omitted and with an “S” suffix added to indicate that these are smoothed (with respect to time) variables. The various flame temperatures that generally appear on the control display are summarized in Table 7.1. TFLMAX

Max bulk Tflame demand

(deg F)


TFLMIN TFLDMD TFLCYCS TFLAMEPCT

TFLOREF TFLIREF

Min bulk Tflame demand (deg F) Bulk Tflame demand (deg F) Estimated Actual Bulk Tflame (smoothed) (deg F) TFLCYCS relative to TFLMIN and TFLMAX 0% when TFLCYCS = TFLMIN, 100% when TFLCYCS = TFLMAX (%) Outer Tflame demand (deg F) Inner Tflame demand (deg F)

TFLODFS TFLPDFS TFLIDFS

Estimated actual outer Tflame (smoothed) (deg F) Estimated actual pilot Tflame (smoothed) (deg F) Estimated actual inner Tflame (smoothed) (deg F)

Table 7.1 Displayed flame temperatures

8.

Combustor Staging 8.1.

General

The combustor staging logic controls the opening and closing of the 10 outer (A) and 15 inner (C) staging valves, as well as the single enhanced lean blowout (ELBO) staging valve. The inner and outer staging valves are opened and closed in accordance with the required combustor configuration. As described previously, there are five combustor configurations, viz B, BC/2 (starting only for the LM2500), BC, AB and ABC. Transitioning from one combustor configuration to another involves increasing or decreasing bleed in conjunction with opening and closing of staging valves. Because of the finite response of the airflow (bleed) control and because of the small combustor flame temperature windows it is not possible to switch immediately from one combustor configuration to another. Therefore, a series of intermediate, or partial, staging configurations are required when going from one steady-state, or permanent combustor configuration to another. In the control there are two key variables BRNREQ and BRNDMD that specify the steady-state combustor configuration target (BRNREQ) and the current combustor configuration demand (BRNDMD). BRNDMD can assume any integer value between 0 and 40, whereas BRNREQ can only assume the values 0, 8, 15, 25, 40 that correspond to the permanent combustor configurations B thru ABC respectively. BRNDMD is translated through look-up tables in the control into inner and outer staging valve commands (INRCMDID and OTRCMDID). This information is summmarized in Table 8.1. INRCMDID and OTRCMDID specify the inner and outer staging valve patterns. For each value of INRCMDID and OTRCMDID, specific inner and outer staging valves are opened. The staging patterns are different for the LM2500 and LM6000 and are defined in the Output Signal Processing section of the Control System Specifications M50TF3740 and M50TF3731 respectively. BURNER CONFIG. B

B+C/2

B+C

A+B

BRNREQ

0

8

15

25

BRNDMD

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

INRDMD INRCMDID INRSTSOP 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 13 12 10 9 7 6 4 3 1 0 1 2 3 4 5 6

OTRDMD OTRSTSOP

OTRCMDID BRNREQ

- Burner Config. Steady-State Target

BRNDMD - Burner Config. Demand FOR BRNDMD < 16 SEE IGNITION CONTROL

INRDMD

- No. Inner Staging Vlv. Open Demand

INRCMDID - Inner Staging Vlv. Pattern ID INRSTSOP - No. Inner Staging Vlv. Open Feedback OTRDMD - No. Outer Staging Vlv. Open Demand

1 2 3 4 5 6 7 8 9 10 10 10 10 10 10 10

1 2 3 4 5 6 7 8 9 10 10 10 10 10 10 10

OTRSTSOP - No. Outer Staging Vlv. Open Feedback OTRCMDID - Outer Staging Vlv. Pattern ID OUTER STAGING VALVE IGNITION CONTROL IF (BRNDMD < 16) THEN TABLE II Z_IGN1DMD Z_IGN2DMD OTRDMD OTRCMDID OTRSTSOP F F 0 0 T F 1 11 F T 1 12


A+B+C

40

32 33 34 35 36 37 38 39 40

7 8 9 10 11 12 13 14 15

10 10 10 10 10 10 10 10 10

10 10 10 10 10 10 10 10 10

Table 8.1 Staging control parameters Staging control is similar for the LM2500 amd LM6000, and is described in the following paragraphs for the various phases of engine operation.

8.2.

Starting

In B mode at fuel-on. Outer staging valves #9 and/or #22 alongside the ignitor(s) are opened during the IGNITE mode. If the max sub-idle Tflame fuel flow limit is encountered during a start for > 3 seconds then the combustor is staged from B to BC/2 (BRNDMD increments from 0 to 8) and remains in BC/2 mode until (as described in section 4.2) the airflow control is enabled as the core speed reaches a sub-idle switch setting (4900 rpm for the LM2500 and 6300 rpm for the LM6000). When the airflow control is enabled the LM2500 staging control switches into the idle and above mode and the LM6000 switches into the core idlesync-idle transition mode.

8.3.

LM6000 Core Idle-Sync-Idle

Because for the LM600, as described in section 5.2, when transitioning between core-idle and sync idle there is not necessarily overlap between the B and BC/2 modes, partial staging is allowed. At core-idle the LM6000 operates in B mode. As the core is slowly accelerated until no-load synch-idle is reached, the combustor progressively stages from B to BC/2 (BRNDMD slowly increments from 0 to 8 and usually overshoots until it finally settles at 8). It is possible depending upon T2 and the bulk flame temperature schedules for synch-idle to be acheived in B mode. Transitioning from synch-idle back to core-idle results in the combustor progressively staging (if not already in B mode) from BC/2 to B mode.

8.4.

Idle and Above Operation

Staging between configurations occurs at extreme corners of a combustor operating window. When accelerating, as illustrated in Figure 8.1, staging is initiated at max. bulk Tflame, min. bleed. When this occurs BRNREQ switches immediately to the new configuration value (8 (LM6000 only) , 15, 25 or 40) and BRNDMD increments to the BRNREQ value. INRCMDID and OTRCMDID follow BRNDMD and inner and/or outer staging valves are progressively opened or closed. ABC

BC

B


Staging Transition Points

AB

BC/2 mode (lm6000 only)

Figure 8.1 Combustor staging during load accels Similar actions occur when decelerating except, as illustrated in Figure 8.2, staging is initiated at min. bulk Tflame, max. bleed.

ABC

AB

BC

ABC to AB stage down logic can cause “early” stage down

BC/2 mode (lm6000 only)

B

Staging Transition Points POWER

Figure 8.2 Combustor staging during load decels GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 45

8.5.

LM6000 BC to AB zone avoidance

The LM6000 includes zone avoidance logic which has two functions, both associated with the BC to AB region. The first function is intended to overcome lack of overlap between BC and AB modes, whereas the second function is intended to avoid high pilot flame temperatures (and therefore Nox) typically seen in the BC mode. Although not strictly part of the staging control, the zone avoidance logic does indirectly force staging from BC to AB. The logic, which produces a megawatt demand bias, is only active in grid mode and has no effect in isochronous (island) mode. The first function which is activated when a BC to AB transition is initiated (BRNREQ>15) adds in a nominal 1 megawatt bias (ZAMWBIASJ) that ramps up at a nominal rate of 0.07 megawatt/sec (ZAMWRATEJ). The second function is activated in BC mode (BRNREQ=15) if T3 reaches a nominal threshold of 760 deg F (T3MAXBCJ) and starts ramping in a megawatt bias at a nominal rate of 0.1 megawatt/sec (ZAT3RTUPJ) until the BC to AB transition is initiated (BRNREQ>15). Once the BC to AB transition is initiated, the megawatt bias is ramped out at a nominal rate of -0.1 megawatt/sec (ZAT3RTDNJ). Note that this second function has to be enabled by setting the control adjustment ZAT3ENAJ = TRUE. The nominal setting for ZAT3ENAJ is specified = FALSE.

8.6.

LM6000 BRNUL upper limit

The LM6000 staging logic includes an upper limit BRNUL that is applied to BRNREQ and BRNDMD. BRNUL is switched 8 to 15 to 25 to 40 as a function of T3SEL. This logic was added to limit the severe effects of partial blowouts that were experienced during early LM6000 engine tests. When a partial blowout occurs the fuel control power turbine speed will increase the demanded fuel flow WF36DMD to compensate for the unburned fuel. When WF36DMD increases, the bulk Tflame/airflow regulator decreases bleed to maintain bulk Tflame. Without the BRNUL limit logic, this could result in the control reaching max bulk Tflame and min bleed and then staging to the next configuration. This situation could repeat until the ABC mode is reached, even if, for the current delivered power, the combustor should have been in B or BC/2 mode! However there are some pitfalls with the BRNUL limit logic. The T3 switch point settings may not be consistent with the bulk Tflame schedules which could result in BRNUL not allowing BRNREQ and BRNDMD to “switch up” on an accel, or could result in BRNREQ and BRNDMD “switching down” prematurely! Therefore, if erratic staging occurs, along the lines just described, it may necessary to adjust the T3 switch points. The nominal T3 switch point settings are defined in Table 8.2. BRNUL switching

T3 switch parameter T3 nominal switch setting (deg F)

8 to 15

T3BCJA

665.0

15 to 25

T3ABJA

783.0

25 to 40

T3ABCJA

874.0

Table 8.2 LM6000 BRNUL T3 switch points

8.7.

ABC to AB stage down - LM2500

The maximum power of the LM2500 DLE is set by either T54 or gas generator speed. For T2’s greater than 10 deg F, the T54 limit of 1535 deg F will limit the maximum power. Unfortunately, for a DLE engine, maximum power is usually not the place where the highest T54 is reached. For any gas turbine engine, the exhaust gas temperature will increase as compressor bleed is increased. This also applies for the DLE engines. Section 1.6.1 described how the bulk flame temperature is modulated between minimum and maximum limits by varying bleed. In ABC mode, it is possible for the left side of the window, i.e. the high GE Proprietary Information - Technical Export License TSU/OTS March 27th, 1997 46

bleed side, to have a higher T54 than the right side of the window, i.e. the low bleed side of the window. If the T54 limit is reached with the total bleed level (DWB36PCT) greater than 0%, the engine will start to decel due to the T54 fuel control regulator reducing fuel flow demand. Keeping in mind the bulk flame temperature window of section 1.6.1, deceling means moving from right to left across the window. As the engine moves from right to left, the total bleed levels increases. As the total bleed level increases, T54 increases. This causes the engine to decel even further. This turns into an unrecoverable cycle. One way to avoid this situation is to keep the bulk flame temperature as low as possible in ABC mode. The field mapping procedure contains the instructions on how to do this. Additionally, the control logic has been modified with two special features to help this situation. First, the control logic will automatically stage down to AB mode if T54 is within 5 degrees of the limit and the total bleed level is greater than 70%. This has been named the ABC to AB stage down logic. Once in AB mode, the control logic locks out staging back to ABC until the calculated flame temperature percentage (TFLAMEPCT) is less than 80%. This interlock prevents the engine from cycling back and forth between AB and ABC modes. Second, the T54 limit schedule has been modified to reflect the trend of high T54s at high bleed. The T54 limit schedule consists of three parts. One part is the maximum power, zero bleed limit of 1535. Another part raises the limit with increasing bleed to 1550 deg F. The final part of the T54 limit schedule raises the T54 limit by 50 deg F only when transitioning from AB to ABC modes. After 20 seconds of operation in ABC mode, this 50 degree bias drops out. In summary, field engine operation to date have shown a potential for hitting the T54 limit in ABC mode prior to reaching zero bleed. This problem can be eliminated by reducing the bulk schedules in ABC mode. To give added margin, the control logic contains the following features to prevent this undesired situation: 1. ABC to AB stage down logic 2. Modified T54 limit schedule - raise T54 limit by 15 degrees at higher bleed levels and bump T54 limit by 50 degrees when staging into ABC mode

8.8.

ABC to AB stage down - LM6000

The LM6000 has similar characteristics to the LM2500 in the ABC mode, although in the case of the LM6000, T48 and corrected (to station 25) core speed (N25R) increasing with increasing bleed (decreasing power) is of concern. For the LM6000, ABC to AB stage down is initiated if the following conditions persist for > 2 seconds: In ABC or AB-ABC transition region (BRNDMD>25) and max core speed regulator (REGULATOR=9) or max T48 regulator (REGULATOR=6) or “throttle” core speed regulator (REGULATOR=2) is encountered and bleed (DWB36PCT) is > 30 %. Once an ABC to AB stage down has occurred then, like the LM2500, to prevent cycling back and forth between AB and ABC modes, staging back to ABC is inhibited, in the case of the LM6000, until bleed level (DWB36PCT) is increased above 30% in AB mode or until staging down to BC mode (BRNDMD<25) has occurred. The nominal reference settings for max corrected core speed, max physical core speed and max T48 are 9525 rpm, 10711 rpm and 1590 deg F respectively. The max T48 setting, as for the LM2500, is raised, in the case of the LM6000 to 1650 deg F when staging from AB to ABC, and remains raised until the bulk Tflame drops from the 50% to 25% level in ABC mode. Also for the LM6000, high temperatures at the gas turbine exhaust in the ABC high bleed region have proven to be of concern in customers’ installations (can result in exceeding boiler temperature limitations etc.). Therefore the LM6000 stage down logic includes an “external” customer forced stage down input (Z_ABC2AB to force stage down, Z_ABUNLCK to unlatch stage down).


8.9.

Load drop/ overspeed

Both the LM2500 and the LM6000 include additional staging logic to optimize speed holding and load drop capability for power generation isochronous operation. This logic anticipates the need to change combustor configuration before the normal combustor window switching conditions are satisfied (min or max bulk Tflame and max or min bleed), and as a result can minimize the variations in power turbine speed when large changes in load occur. For example load drops on an LM2500 can cause the staging control to switch from ABC to B mode almost instantaneously.

9.

Acoustics and Blowout Avoidance (ABAL) Logic

LM2500 and LM6000 combustors have to be mapped in the factory and field to measure acoustics and blowout boundaries and establish bulk and ring flame temperature schedules. The margin or window between a blowout and acoustic condition is often small. If acoustics or blowout boundaries drift with time we do not want to have to constantly remap engines. This is where ABAL can help. ABAL automatically lowers or raises A or C ring temperatures (or bulk Tflame when in pilot-only mode) when high combustor acoustics or partial blowouts are detected. ABAL can also reduce the high combustor acoustics or partial blowouts that are common when transitioning to a new combustor mode. Combustor acoustics (dynamic pressures) are measured, as mentioned in section 1.2, using two sensors mounted on the compressor rear frame. After appropriate signal conditioning and bandpass filtering, the control sees the acoustics as a dynamic pressure level (peak-peak psi). Partial blowouts are not so easy to detect and certainly can’t be directly measured! In this case a model-based blowout detection (BOD) algorithm is used. The ABAL interfaces are shown in simplified form in Figure 9.1.


ACOUSTIC / BLOWOUT AVOIDANCE LOGIC

ACOUSTIC SENSOR PX36SEL

TFLAME BULK MAX ADJUST

TFLAME BULK MIN ADJUST TFLAME A REF ADJUST

BLOWOUT DETECT WF / PS3 ERROR

TFLAME C REF ADJUST

Figure 9.1 Acoustics/blowout avoidance logic The blowout detection input is in the form of a WF/PS3 error. The blowout detection algorithm includes a model of the engine core (HP compressor - combustor - HP turbine) for which boundary conditions are defined from available engine sensors together with combustor fuel flow from WF36DMD. The algorithm assumes all the fuel to being burning and estimates an HP rotor speed and a compressor discharge pressure PS3. The WF/PS3 error is calculated using sensed PS3 (or PS32 for the LM6000) and estimated PS3 as shown in Figure 9.2.


WF36DMD

B_NGG BOD ALGORITHM B_PS3

START WF36DMD PS3SEL B_PS3

B_WFQPS3 :=F_WF36DMD / B_PS3; WFQPS3SEL :=F_WF36DMD / PS3SEL; B_WFQPERR :=((WFQPS3SEL - B_WFQPS3)/WFQPS3SEL) * 100.0;

B_WFQPERR

END Figure 9.2 Blowout detection algorithm WF/PS3 error calculation THE BOD logic has provision for correcting or calibrating the WF/PS3 error B_WFQPERR for each engine as part of the mapping process. B_WFQPERR has a bias added to it (subtracted for the LM6000) to provide the corrected WF/PS3 error WFQPERRCOR. The bias is calculated as a function of combustor mode (BRNDMD) using an adjustable table (T259A_1). This table then can be adjusted for each engine. The ideal value for WFQPERRCOR is zero, but in practice a target of 0.0 to 3.0, when the combustor is fully lit, is aimed for for all combustor modes. ABAL compares WFQPERRCOR with a threshold BWFQPERRSP that is also calculated as a function of BRNDMD (blowout detection setpoint table T259A_2). This threshold does not change from engine to engine. If the corrected WF/PS3 error WFQPERRCOR becomes greater (more positive) than the threshold BWFQPERRSP, then ABAL assumes a partial blowout has occurred and takes action (see Figure 9.3). ABAL works in the same way for high acoustics. In this case the acoustics are measured (PX36SEL), but as for the blowout detection, their level is compared with an acoustics threshold P_ACSTICJ, which is calculated as a function of BRNDMD (acoustic detection setpoint table T259A_3). This threshold is not intended to change from engine to engine.


B_WFQPERR

LM2500 DLE BLOWOUT DETECTION

WFQPERRCOR ADJUSTABLE T259A_1 Blowout Detection Ref. Table

+

+

X GE Y

BRNDMD

ADJUSTABLE T259A_2 Blowout Detection Set Pt., Table BWFQPERRSP

FUNCTION T259A_1 BLOWOUT DETECTION REFERENCE TABLE - ADJUSTABLE X RANGE:NOT ADJUSTABLE, Z RANGE:MIN=-50.0, MAX=50.0 FUNCTION T259A_2 BLOWOUT DETECTION SET POINT TABLE - ADJUSTABLE X RANGE:NOT ADJUSTABLE, Z RANGE:MIN=0.0, MAX=50.0 BRNDMDR (REAL) 0. 8. 15. 25. 40.

LM2500

BWFQPERRSP (%) 15.000 15.000 15.000 15.000 15.000

BRNDMDR (REAL) 0. 8. 15. 25. 40.

LM2500

BWFQPERRSP (%) -11.000 0.000 -1.500 -4.500 -3.000

CALIBRATE EACH ENGINE

Figure 9.3 Blowout detection

The corrective action ABAL takes for either high acoustics or partial blowouts is described in Table 9.1 B mode - Tflame bulk min and max decreased for high acoustics - Tflame bulk min and max increased for partial blowouts BC/2 and BC mode - Tflame C reference decreased for high acoustics - Tflame C reference increased for partial blowouts AB mode - Tflame A reference decreased for high acoustics - Tflame A reference increased for partial blowouts ABC mode - Tflame A then C reference decreased for high acoustics - Tflame A then C reference increased for partail blowouts Table 9.1 ABAL corrective action


Special features of ABAL are described in Table 9.2 ABAL is limited to: +/- 100 on LM2500 (adjustment TACSTCLMTJ, TBLOWLMTJ) +/- 150 on LM6000 (adjustment TACSTCLMTJ, TBLOWLMTJ) Acoustics detected when: PX36SEL greater than acoustic threshold P_ACSTICJ which is scheduled as a function of BRNDMD (Table T259A_3). Nominal value of threshold is 3.3 psi peak-peak except for BRNDMD=0 which has a nominal threshold of 8.0 psi peak-peak. Acoustics or blowout must persist for 0.48 sec or longer Adjustment PERS_ATMRJ, PERS_BTMRJ When acoustics are detected: The appropriate Tflame reference is reduced by 10 deg F (TABALAJ) every 0.48 seconds until the condition clears or the ABAL limit is reached. Note that LM6000 has C ring reversal feature that allows, through adjustment (Table T310_1), C ring to be raised when high acoustics are encountered in BC or ABC mode. This is intended to accomodate a scenario where acoustics are induced by high pilot flame temperature. When partial blowout is detected: The appropriate Tflame reference is increased by 20 deg F (TABALBJ) every 0.48 seconds until the condition clears or the ABAL limit is reached ABAL logic is disabled during burner transitions ABAL adjustments are interpolated back to 0.0 during burner transitions ABAL always starts at 0.0 when entering a new burner configuration ABAL adjustments are “backed out” if no high acoustics or partial blowouts occur for 60 seconds If high acoustics or partial blowouts recur during backout, then backout logic is disabled until combustor mode changes Table 9.2 ABAL special features


Under certain circumstances short duration high level acoustic spikes could persist, which over time could be harmful, but would go undetected by the normal acoustic detection logic. Therefore ABAL also includes an acoustic spike detection feature as described in Table 9.3 Spikes detected when 2 spikes within 20 seconds of > 10 psi p-p in B mode, otherwise > 4 psi p-p Spike detection enabled when: In permanent combustor mode (BRNDMD=BRNREQ) for more > 60 seconds > 10 seconds since last spikes detected ABAL corrective action for acoustic spikes same as for normal persistent high acoustics Table 9.3 ABAL acoustic spike detection


DLE Overview.pdf

Recommend Documents