Ventilation and Leak Dispersion in CCGT Enclosures

1
Ventilation and Leak Dispersion in CCGT
Enclosures

• Patrick Phelps ( Flowsolve )
• and
• Douglas Wylie (GEC Energy Services)

IPUC 7 - Luxembourg - May 2000
2
Ventilation and Leak Dispersion in CCGT Enclosures

• Industrial Context
• Health and Safety Issues
• Application to an Existing Power Station
• Application to New Enclosure Designs
• Conclusions
• Experimental Verification

3
Industrial Context - 1

• Combined Cycle Gas Turbine (CCGT) Plants
• Gas turbines drive an electricity generator
• Engine exhaust waste heat recovered by a boiler
  to produce steam.
• Steam turbine generates further output.

4
Industrial Context - 2

• Turbines are fuelled by gas at very high
  pressures
• Liquid fuel system as back-up
• Turbines are contained within acoustic
  enclosures.
• Enclosures tend to be installation-specific
  designs

5
Frame 6 Turbine Generator
6
Industrial Context - 3

• Each enclosure is divided into a number of
  compartments containing
• Auxiliary equipment
• Gas turbine and exhaust plenum
• Reduction gearing and the generation equipment.
• Auxiliary/GT/exhaust compartment is usually self-
• contained, with a dedicated ventilation system

7
Ventilation and Leak Dispersion in CCGT Enclosures
8
Health and Safety Issues - 1

• Enclosure ventilation system removes some heat
  from the turbine casing
• Enables operatives to carry out readings and
  routine maintenance under operating conditions.
• However, enclosures remain a thermally hostile
  and noisome environment .

9
Health and Safety Issues - 2

• Gas from leakages can accumulate to flammable
  proportions in poorly ventilated regions of the
  enclosures - dead zones
• OUTCOME - Big Bang
• REMEDY - Use the ventilating air to safely dilute
  and disperse any gas leakage.
• MOTIVATION - Legislation

10
UK Safety Assessment of CCGT Enclosures - 1

• Identify nature and potential sources of
  hazardous material releases
• Determine leak frequency and inventory of
  releases
• Investigate airflow characteristics
• identify "dead zones"
• Predict dispersion consequences safety-critical
  release scenarios
• Implement remedial measures

11
UK Safety Assessment of CCGT Enclosures - 2

• Experimental investigation of airflow
  characteristics is difficult within the confines
  of a turbine generator enclosure, especially
  under operational conditions.
• HSE promote computer simulation as the most
  appropriate technology in this case.

12
UK Safety Assessment of CCGT Enclosures - 3

• CFD models can
• simulate the dispersion consequences of releases
  under a variety of operating conditions
• compare the efficiency of alternative ventilation
  strategies, to achieve the desired dilution /
  dispersion result

13
Safe Dispersion Criteria - 1

• Santon Criterion
• The ventilation arrangements within the turbine
  enclosure must be such as to ensure the safe
  dilution/dispersion of gas releases prior to
  activation of mitigation/shutdown systems by the
  gas detection system. The criteria to be applied
  are that the envelope of the 50 LEL
  concentration contour should not occupy more than
  0.1 of the free volume of the enclosure, for a
  gas leak of sufficient magnitude to trigger the
  gas detection system.

14
Safe Dispersion Criteria - 2

• G V (0.01 E) (0.01 S).
• Gas concentration level (S) detected by sensors
  for activating emergency response systems
  (typically 10)
• Lower Explosive Limit (E) for the turbine fuel
  gas is around 5.
• Maximum undetected leak is thus of magnitude 10
  of LEL
• For compliance, ensuing flammable envelope (of
  the 50 LEL surface) must not exceed 0.1 of the
  compartment free volume.

15
Safe Dispersion Criteria - 3Compartment
free volume
16
Reference Leak Scenario

• Zero momentum leak source
• corresponds to jet release impinging immediately
  on an obstruction (casing, flange body
• No net directionality imparted to release.
• A directional release would require additional
  assumptions ..

17
Application to an Existing Power Station - 1

• A CFD-based simulation study
• commissioned by
• IVO Generation Systems
• and
• Regional Power Generators Ltd

18
Application to an Existing Power Station - 2
19
Application to an Existing Power Station - 3

• Over 100 simulations performed
• Studies to determine
• air flow distribution
• worst case operating condition (hot,cold)
• worst case leak location
• efficiency of alternative retrofit ventilation
  strategies, to achieve HSE compliance

20
Air flow Distribution at inlet toTurbine
Compartment
21
Application to an Existing Power Station
22
Application to an Existing Power Station -
Parameter Studies

• Over 100 simulations performed
• Studies to determine
• air flow distribution
• worst case operating condition (hot,cold)
• worst case leak location
• efficiency of alternative retrofit ventilation
  strategies, to achieve HSE compliance

23
Hot Operating ConditionsEnvelope volume -
0.69
24
Cold Operating ConditionsEnvelope volume -
2.28
25
Application to an Existing Power Station -
Parameter Studies

• Over 100 simulations performed
• Studies to determine
• air flow distribution
• worst case operating condition (hot,cold)
• worst case leak location
• efficiency of alternative retrofit ventilation
  strategies, to achieve HSE compliance

26
Worst Case Leak Location

• Under both hot and cold conditions, the worst
  case leak location was found to be in the pit
  region, in front of the lowest combustor flanges

27
Application to an Existing Power Station -
Parameter Studies

• Over 100 simulations performed
• Studies to determine
• air flow distribution
• worst case operating condition (hot,cold)
• worst case leak location
• efficiency of alternative retrofit ventilation
  strategies, to achieve HSE compliance

28
Alternative Ventilation Strategies 1 -
Abject failures

• Increasing ventilation rate
• overhead pendant baffles
• twin outlets
• blowing air into the pit region
• sucking air from the pit region
• EGT wavewall idea

29
Alternative Ventilation Strategies 2 -
Heroic failures

• Reversed flow system
• air supply through existing outlet
• air extract to TG inlet plenum
• Lateral side-gust system
• air supply through side door
• air extract through existing outlet
• other inlets blocked off

30
Alternative Ventilation Strategies 3 -
Final Success !

• The Corkscrew Strategy
• Close all existing inlets
• plate over grated walkway tops
• Single non-symmetric outlet
• Two inlet slots , one high, one low, cut in
  connecting doors
• 30-degree deflector plates create corkscrew
  effect

31
Corkscrew Ventilation Scheme
32
Corkscrew Ventilation Scheme
33
Corkscrew Ventilation Scheme
34
Corkscrew Ventilation Scheme
35
And so . . . . . . . . . .

• This led on to . . . . . . . .

36
Application to New Enclosure Designs

• A CFD-based simulation study
• commissioned by
• the Thermal Power Division
• of Kvaerner Energy Ltd

37
Application to New Enclosure Designs
38
Enclosure Geometry - Elevation
39
Enclosure Geometry - End View
40
Application to New Enclosure Designs
41
Turbine combustor flanges and associated pipework
42
Geometry Representation - 1
43
Geometry Representation - 2
44
Application to New Enclosure Designs -
Workscope

• Over 25 different simulations performed
• Studies to determine sensitivity to
• nodalisation level distribution
• leak location
• ventilating flowrate
• presence of internal geometric features
• inlet flow manipulation.

45
Findings - 1

• The worst case leakage scenario, under cold
  start-up conditions, was a zero-momentum leakage
  from the flanges in front of the lowest can
  combustor
• The flammable gas cloud with the reference
  ventilation arrangement was twenty five times
  larger than the target value (11 times larger if
  the accessory compartment volume was included)

46
Reference ConfigurationFlammable volume 2.8
TC
47
Reference ConfigurationFlammable volume 2.8
TC
48
Reference ConfigurationFlammable volume 2.8
TC
49
Reference ConfigurationFlammable volume 2.8
TC
50
Reference ConfigurationFlammable volume 2.8
TC
51
Application to New Enclosure Designs -
Workscope

• Over 25 different simulations performed
• Studies to determine sensitivity to
• nodalisation level distribution
• leak location
• ventilating flowrate
• presence of internal geometric features
• inlet flow manipulation.

52
Findings - 2

• Increasing the ventilating air flow by 25 had
  little effect on ventilation efficiency ( gas
  leak size increased proportionately ).

53
Application to New Enclosure Designs -
Workscope

• Over 25 different simulations performed
• Studies to determine sensitivity to
• nodalisation level distribution
• leak location
• ventilating flowrate
• presence of internal geometric features
• inlet flow manipulation.

54
Atomising Air System Flammable volume 1.9 TC
55
Atomising Air System Flammable volume 1.9 TC
56
Atomising Air System Flammable volume 1.9 TC
57
Atomising Air System Flammable volume 1.9 TC
58
Atomising Air System Flammable volume 1.9 TC
59
Application to New Enclosure Designs -
Workscope

• Over 25 different simulations performed
• Studies to determine sensitivity to
• nodalisation level distribution
• leak location
• ventilating flowrate
• presence of internal geometric features
• inlet flow manipulation.

60
Findings - 3 Inlet Flow Manipulation

• Improved results (decrease in flammable cloud
  size) obtained by
• adding a scoop diverter at inlet
• throttling - increasing the inlet velocity
• concentrating the incoming flow towards the axial
  centreline
• blocking off the outermost inlet hole(s) on each
  side and
• biasing the massflow distribution.

61
Final ConfigurationFlammable Volume 0.2 TC
62
Final ConfigurationFlammable Volume 0.2 TC
63
Final ConfigurationFlammable Volume 0.2 TC
64
Final ConfigurationFlammable Volume 0.2 TC
65
Final ConfigurationFlammable Volume 0.2 TC
66
Conclusions

• Flammable cloud volumes of less than 0.1 of the
  free volume of the combined compartments are
  attainable (at least conceptually) .
• This would satisfy the current UK dispersion
  criterion.
• Compliance achieved without major structural
  changes to enclosure or to air delivery system.

67
Future Efforts

• Confirm high velocities do not provoke thermal
  shock problems in hot operation
• Confirm required manipulation of inlet air supply
  is technically feasible (!)
• Use model to predict likely over-pressures
  arising from deflagration of the confined gas
  plume.

68
Experimental Verification

• No quantitative data available.
• Qualitative comparison - video footage of smoke
  tests on installations with similar internal
  geometries.
• Smoke trails confirm magnitude and direction of
  airflow in the region in front of the can flanges
  and around the front upper part of the turbine
  barrel

69
Ventilation and Leak Dispersion in CCGT Enclosures

• Thank you for your attention .
• When I count to three,
• you will awake
• and
• remember nothing ..

70
Ventilation and Leak Dispersion in CCGT
Enclosures

• Patrick Phelps (Flowsolve Ltd)
• and
• Douglas Wylie (GEC Energy Services)

IPUC 7 - Luxembourg - May 2000