TMF Lite Process Development

Transcription

1 TMF Lite Process Development Qingzhong Li Dong Wei Bill Moser Scott Thompson Don Sit Caterpillar Inc. SCC May, 2012

2 Table of Contents 1. Project Background / Objective 2. TMF Lite Model Development /Verification Lite Model Development Abaqus User Subroutine Verifications 3. Project Summary

3 Temperature(K) 1 Background/ Objective TMF --- Thermal Mechanical Fatigue analysis (mainly used in exhaust manifold) Cylinder Head Turbo-housing Gas temperature Rated gas temp (full load) Idle gas temp t1 t2 t3 t4 Time (s) Typical Test Cycle Center section Exhaust Manifold C High-Mount Temperature Temperature cycles profiles at on different the thermocouples locations t0 t4 Time(Sec) t8 tc11 tc10 tc9 tc8 tc7 tc6 tc5 tc4 tc3 tc2 tc1

4 Background/ Objective Overview of Standard TMF Process CFD Analysis -Create CFD model -Run Thermal Cycle Analysis -CFD/FEA Post Process FEA thermal stress Analysis -Create FEA model -ABAQUS Assembly/ Verify Contacts -Thermal stress analysis Fatigue Analysis -Convert Stress format (skin mesh result) -Run COBRA for TMF - Post Processing Back to Table of Content

5 Project Background / Objective Challenges of New HPL EGR Strategy for Tier4i Multiple Tier4a manifold designs need to assess the structural integrity quickly during the design stage.

6 Validation Challenges of New Designs for Tier4i More than 6 configurations had to use totally new manifold designs More configurations to analyze. Limited validation time. Project Objective This study is to develop a simplified heat transfer and TMF life analysis procedure which can simplify the analysis process and achieve reasonable results at the same time. This simplified procedure sacrifices overall accuracy for analysis speed and is intended to complement the full version of TMF. Back to Table of Content

7 2. Lite Model Development Full CFD (Fluent) Full Structural (ABAQUS) Lite transient heat transfer analysis model (ABAQUS) Lite structural model (ABAQUS) Same mesh

8 Lite Model Initial Iterations Outer surface: Htc=0.05 * head-manifold: htc=1.0 Htc unit: mw/mm^2-c Same as internal surface/ or no BC HTC Full load htc Internal surfaces: constant htc 0.70 Abaqus usage: *sfilm,amplitude=temp_amp_2, film amplitude=htc_int_long2 internal_long_surf_b, F, 127,0.70 Idle load htc t1 t2 t3 t4 Step Time (s)

9 Temperature(K) Temperature Comparison 1000 C High-Mount Temperature profiles on the thermocouples ABAQUS simplified model CFD tc11 abaqs-tc11 abaqus-v61 v13 v14 Idle temperature contour from CFD The goal is to map CFD steady results with a set of heat transfer BC v15 v17 v18 v t t t Time(Sec)

10 User Subroutine film.f Development Advantage: define different thermal boundary conditions without extra definitions of surfaces (one surface definition but different htc and sink temperature values at different locations) All cylinders have 1 exhaust pulse every 2 revs of the engine. The head ports see 1 exhaust pulse every 2 engine revolutions. Physics says this area HTC should be ~1/3 of the high value. This area sees 2 exhaust pulses every 2 engine revolutions. Physics says this area HTC should be ~2/3 of the high value. t0 HTC Full load htc Idle load htc t1 t2 t3 t4 Step Time (s) This area sees 3 exhaust pulses every 2 engine revolutions. Assume this area gets the high HTC value.

11 Inputs to Film User Subroutine - Coordinates of nodes at port interfaces A1 B1 A2 B2 C1 C2 Head Flange Surface 3 A3 C3 B3 Abaqus usage: *sfilm internal_long_surf, FNU, 127, 0.61 This should be the high htc of the port

12 HTC Verification using ABAQUS User Subroutine UVARM ** It was verified that the ports were correctly divided into different regions and assigned with correct htc.

13 Temperature Diff (C) Temperature Comparison between Full Model and Lite Models Temperature Difference between CFD and Lite Models old- h=0.70 New - h=0.61 user subroutine 20 0 Thermal Couple and Other Critical Locations Used for Comparison -20 TC6, 7, 8 TC 1 TC 3 TC8 TC11 top edge radius TC9 TC4, 5 TC11 TC11 TC1 TC1 TC3 TC3 TC2 TC10 TC3 Top edge Radius

14 TMF Life Comparison between Different Models and Full CFD Results Left View TMF Comparison Locations R1 Right View L3 R3 L2 L1 R2 Top View Bottom View T4 B1 T3 B3 B2 B4 T1 T2 Head Flange View B5 B6 H1 H2

15 Cycles - Lite Lite/Full TMF Life Comparison 1.E+07 10x 2x 1.E x 1.E x 1.E+04 1.E+03 Lite Analysis -User subroutine 1.E+02 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 Cycles - Full CFD Most data fall in the range of one order of magnitude More than 70% of data are below Full TMF results; the data which exceed full TMF results are less than 2x of the CFD results, therefore the model is generally conservative

16 4. Project Summary A TMF Lite Procedure has been developed and verified; Its advantages / limitations defined; Lite Process significantly reduces the TMF analysis time to 3 ~ 5 working days (Transient FEA + Stress FEA +TMF) from about 2~3 months for the full TMF; This process has been incorporated as an important part of the overall manifold validation strategy; TMF Lite procedure has been used in many Tier4a exhaust manifold new designs; More than 20 new designs were evaluated; Back to Table of Content

17 Thanks! Questions?