PLACE FOR TITLE

Investigation of Steam Turbine blade
AUTHORS Mr. Yuki Nakamura
failure
Mr. Ashutosh Vengurlekar
ExxonMobil Research and Engineering, Singapore
Mr. Robert D Fisher
Mitsubishi Heavy Industries Compressor Corporation, Hiroshima Japan
Mr. Yuichi Sasaki
ExxonMobil Research and Engineering, Houston Mitsubishi Heavy Industries Compressor Corporation, Hiroshima Japan
Global Technology Sponsor (Machinery) Mr. Satoshi Hata
MHI Compressor International Corporation , Houston
Mr. Kyoichi Ikeno
Mitsubishi Heavy Industries Compressor Corporation, Hiroshima Japan

Presenter/Author bios (To be revised)
Mr. Ashutosh Vengurlekar
ExxonMobil Research and Engineering, Singapore
Discipline Technology Lead – Machinery Asia Pacific, ashutosh.vengurlekar@exxonmobil.com
Mr. Robert D Fisher
ExxonMobil Research and Engineering, Houston
Global Technology Sponsor (Machinery), robert.d.fisher@exxonmobil.com
Mr. Yuki Nakamura
Mitsubishi Heavy Industries Compressor Corporation, Hiroshima Japan
Engineering & Design Division, yuki4_nakamura@compressor.mhi.co.jp
Mr. Yuichi Sasaki
Mitsubishi Heavy Industries Compressor Corporation, Hiroshima Japan
Engineering & Design Division,yuichi2_sasaki@compressor.mhi.co.jp
Mr. Satoshi Hata
Mitsubishi Heavy Industries Compressor Corporation, Hiroshima Japan
Engineering & Design Division, satoshi_hata@compressor.mhi.co.jp
Mr. Kyoichi Ikeno
Mitsubishi Heavy Industries Compressor Corporation, Hiroshima Japan
Engineering & Design Division, kyoichi_ikeno@compressor.mhi.co.jp

Abstract Blade failure was observed on a backpressure steam turbine (driving a centrifugal compressor) after it was in service for more than one year. This paper presents details of observations. . inspections carried out and root cause analysis of the turbine blade failure.

Conclusion 10. Fretting stress analysis 8. Fracture surface observation by OEM 3. Heat transfer stress analysis 7. Possible cause 6.1.Contents 1.2. Review of operation data 5. Fracture surface observation 3. Turbine specification 2. Lessons Learned . Solution to avoid fretting 9. Fracture surface observation result 4. Observations during overhaul 3.

Turbine Specification • Turbine specification – Back pressure turbine driving ethylene compressor in a cracker plant – Power : 17MW – Speed : 4117 – 5085 rpm – Normal steam inlet condition • Inlet : 103 kg/cm2G . 503℃ • Outlet : 45 kg/cm2G • History 2013 March 2014 Plant Start Turbine trips on high vibrations Turbine Cross Section Restart attempted and unsuccessful due to high vibrations at low speeds.1. Decided to overhaul turbine. .

. Observations during overhaul (1/3) 1st stage <View from steam inlet side> <View from steam outlet side> First stage wheel : Broken blade at one location. No damage on other components.2.

Other components (2nd and 3rd stage) not damaged .2. and fracture was located at the 1st bearing portion of blade root. Observations during overhaul (2/3) steam outlet side Damaged Blade Blade root 1st bearing 2nd bearing 3rd bearing Good Blade One blade damaged out of 78 blades.

11 No.70 blade root Failed blade was No. MT inspection.2. Steam inlet side ○:Crack indication No. Steam inlet side Concave side. first/last of shroud.70 on blue colored circle part.22 Failed blade and crack indicated and No. No. Steam outlet side . No.22 Crack Crack Convex side.40 blade root in blades were located on the Concave side. Observations during overhaul (3/3) Failed blade No.40 Crack indication was Crack detected on No.11.

Fracture surface observations This portion was cut Concave side Ratchet mark Ratchet mark A Overview of fracture surface Magnification of A area Rough surface  Ratchet mark was observed on fracture surface. Initial crack was initiated by fretting fatigue.  Chemical components and hardness were satisfied the required specification.  Rough surface was observed around crack origin. and blade failed from high cycle low stress fatigue.  Crack was started from rough surface.3-1. .  No corrosion pit was observed. so the crack has multiple origin.

4) The failed blade steel is made of good quality. 1) The blade failed from high cycle fatigue. This and the presence of ratchet marks support a high cycle low stress fatigue mode.70 blade . 5) No evidence of an external factor related to steam quality was found. 2) Fatigue cracks start in 4 areas on both sides of the blade root.3-2. Fretting fatigue cracks start from these locations. 3) In the fatigue crack initiation locations multiple fretting marks are present.70 blade.1 Fracture surface of No. Fracture surface observation result The followings are observed from fracture surface of No. Fig.

This20event had occurred approximately 1 month before the failure of the turbine 100 blade. FVJ501. SHP Steam Pressure [BarG] / Torquemetter [KN SHP Steam Temperature [Deg-C] / Flow [T/H 80 400 4. FVF551. reduction Blade failed 600 FVP501. Torque meter 100 500 Inlet Temp. 200 Inlet press. Inlet 20 temp.:503℃ 100 SHP Steam Pressure [BarG] / Torquemetter [KNM] 90℃ SHP Steam Temperature [Deg-C] / Flow [T/H] 80 1 month 400 0 0 2014/2/25 5:45 2014/2/25 6:43 Feb. 2014 2014/2/25 7:40 2014/2/25 8:38 2014/2/25 9:36 March.Torquemeter FVT517. Review of operation data • Review 60 of Operation data 300 Inlet Flow SHP Steam Condition 120 Inlet steam temp.SHP Inlet Press. .SHP Inlet Temp.SHP Inlet Flow 40 20min. 2014 2014/2/25 10:33 Date 60 300 Inlet Flow The temperature changed from approximately from 500 degree C to 410 degree C in 40 200 roughly 20 minutes and then recovered to rated temperatures in approximately Torque meter 30 minutes.:103 kg/cm2G Inlet Press.

by improper machining × blade outline Groove dimension is within design. surface Blade 1st stage blade stress due to NPF. and/or assembly Corrosive Insufficient quality of steam Observation of environment Failed blade material is made of good Defects of material fracture surface quality. Excessive static stress Observation of No indication of contact with other parts Unexpected damage nozzle and blade × was observed by visual. 2×NPF failure on Re-check of 1st and shock load are satisfied with our 1st stage stage blade Excessive exciting force design criteria. Possible cause (1/2) PHENOMENON ANALYSIS OF FAILURE PHENOMENON HOW TO VERIFICATION Result & Assessment Review of Inlet steam temperature goes down from Out of allowable operation operation data 500℃ to 410℃ in 20 min. material property material △:Possibility ×:Low possibility . 5.(ERC turbine is drive 3D FEM analysis vibration 2 shock road condition) steam stress Large stress concentration turbine Observation of Blade root dimension is within design. However. valve sequence is of Comp. × No evidence of an external factor related Inferior Insufficient mechanical Inspection of blade to steam quality was found. strength by using Excessive different from PRC turbine.

5. Possible cause (2/2) Operation Stress Fracture Number of cycles Steam inlet temp. Vibratory stress exceed fretting Step 1 90℃ reduction (Shock) thresh-hold for crack initiation (-4℃/min) ⇒ Slip on blade root ⇒ initial cracks formed Vibratory stress ⇒ Step 2 High cycle low stress cause crack Rated Operation fatigue propagation even with no slip .

Operating Time Operating Surface of Blade & Shroud Reduction Reheating Reduction Reheating Total Speed Steam Temperature Nor.4 ℃/min 22 min 38 min 60 min (4700 rpm) Disc Front & Back of Disc Heat Insulated Circumferential Section Heat Insulated Rotor Surface & Inside of Rotor 465℃ (Constant) (Initial / Steady state) Steam Temperature Condition Analysis Model & Thermal Boundary Condition .6. Heat transfer stress analysis (1/2) <Analysis Model> <Analysis condition table> Blade & Shroud Steam Temp. -4 ℃/min +2.

180deg .6. Heat transfer stress analysis (2/2) 静圧分布(Pa) 流速分布(m/s) 0deg ~ 90deg 90deg ~ 180deg 0deg ~ 90deg 90deg ~ 180deg Rated temperature High 180deg ~ 270deg 270deg ~ 360deg 180deg ~ 270deg 270deg ~ 360deg High contact load Low 90 ℃ temp reduction of blade Low contact load 全圧分布(Pa) 0deg ~ 90deg 90deg ~ 180deg 入口(弁4) 入口(弁1) 0deg The contact area shifts from 1st bearing surface to 3rd bearing surface on end blades 270deg 90deg 180deg ~ 270deg 270deg ~ 360deg as temperature is reduced. Makes the end blades susceptible to vibrations.

Line A (cracked point).  After steam reheated. Contact force Contact force Cracked point 3rd brg. Contact force Average of contact pressure (at Blade⑥ / Concave side) .Contact force reduction (Focused on first or last blade of shroud group)  At the 1st brg. 1st brg. the contact pressure significantly reduce after steam cooled. the contact pressure become nearly equal to initial condition again. 2nd brg.

Slip evaluation result (Focused on first or last blade of shroud group / 1st bearing)  Slip between rotor groove and blade was possible as the vibratory forces exceeded contact force on blade root during temperature excursion event Fnominal : Contact force Groove side μ×Fnominal Fvib: Blade side Vibratory force Movement force Steady state (0 sec) 90℃ Reduction (1320sec) Fvib > μ×Fnominal ⇒ Slipped Fvib < μ×Fnominal ⇒ Not Slipped Evaluation portion Slip level of contact area .

No fretting zone Relative slip S Schematic of fretting crack Fvib > μ×Fnominal ⇒ Slipped Contact pressure Fvib < μ×Fnominal ⇒ Not Slipped Fnominal Image of contact pressure distribution Groove side Direction y μ×Fnominal Crack propagation Fvib Blade side Stress amplitude Slip Adhesive Slip Movement force region region region (Not slip) . Fretting zone Relative slip S is. Fretting stress analysis (1/2) Fretting map Fretting evaluation Relative slip & contact stress is plotted on fretting criteria Contact stress (based on experimental data in OEM).7.

Fretting stress analysis (2/2) Result Rated speed: 4700 rpm Steam inlet 3600sec Fretting zone Temp. Enough time to initiate fretting crack .6×105 cycles = 157Hz×1000sec Commonly.7. cyclic number for fretting fatigue crack initiation is 104 ~ 105 cycle. Contact stress 500℃ Fretting start Fretting stop Re-heating 410℃ Cooling 1320sec Steady state time (0sec) No fretting zone 1000sec Relative slip Fretting estimated time : 1000sec Dominant cycle : 157Hz = 4700rpm/60 × 2shock/round ⇒ Cyclic number : 1.

1 No. 50% FFT (Fast Fourier Transform) is applied to above lower stepped force σv = A×σb σv: Shock stress σb: Bending stress . Original External No.3 No.4 force Rotating direction 2 time external force occurs on blade during one rotation After modification External No.1 No. Solution to avoid fretting (1/2) Vibratory stress can be reduced about 50% due to GV opening location change for No.4 force Rotating direction 1 time external force occurs on blade during one rotation Approx.2 and 4.2 No.8.3 No.2 No.

.8. Solution to avoid fretting (2/2) Comparison of fretting evaluation Fretting analysis result Fretting analysis result Original case Improvement case Fretting zone Fretting zone Contact stress Contact stress Relative slip Relative slip Fretting occurs from 750sec Fretting does not occur.

Reduction of contact pressure at the first/last blade of shroud group due to temperature change.9. This was possible by modification of governing valve sequence in case of this turbine. Effect of change has been studied and model results show that fretting can be avoided even in case of temperature excursions by reducing vibratory stresses. 5. and crack location are matching with the analysis result. Slip between rotor disk and blade was possible as the vibratory forces exceeded contact pressure on blade root during temperature excursion event. Conclusion 1. . 4. Contact pressure at the origin of crack changed significantly during inlet steam temperature excursion. Vibratory stresses have to be reduced to be lower than contact pressure as a solution to avoid fretting. 3. 2. Temperature event combined with number of cycles during low temperature excursion was adequate to cause fretting and to initiate fretting cracks.

10. Lessons Learned  Operation  Plant operation can have significant impact on performance of steam turbines.  Establish guideline for fretting on Goodman diagram to avoid fretting Guideline for fretting on Goodman diagram Where: σY .Yield strength σEA . Stable temperature must be maintained for long term reliability.  Design  Robust design should consider potential operation out of normal operating ranges.Fatigue limit in pure steam σT – Actual breaking stress .

Thank you for your attention .