P. Teufel and A. Böhmer, ABB Turbo Systems, SIMULIA Customer Conference Thrust Collar Bearing Optimization using Isight

Transcription

1 P. Teufel and A. Böhmer, ABB Turbo Systems, SIMULIA Customer Conference 2012 Thrust Collar Bearing Optimization using Isight May 23, 2012

2 Thrust Collar Bearing Optimization Using Isight Contents Turbocharging: Principle and Customer Value ABB Turbo Systems: Products and Range of Application Dealing with Power: Minimizing Risk for each Component Dealing with Thrust: Axial Bearing Arrangement Prediction of Mechanical and Thermal Loading Optimization Using Isight: Targets and Restrictions Parameterization and Challenges Loop and Optimization Techniques Results Summary and Outlook May 23, 2012 Slide 2 filename

3 Turbocharging: Principle and Customer Value Main purpose: increase the specific power output of the engine Further advantage: lower specific fuel consumption Key performance figures: achievable compressor pressure ratio πc turbocharger efficiency ηtc specific airflow Customer value: 400% more power 10% lower specific fuel consumption reduced emissions May 23, 2012 Slide 3 filename

4 ABB Turbo Systems: Products and Range of Application Turbocharged power per turbocharger: ,000 kw Traction: Off Road Vehicles, Locomotives Stationary: Power Plants Marine propulsion and onboard power generation May 23, 2012 Slide 4 filename

5 Dealing with Power: Minimizing Risk for each Component Tip velocity: > 500 m/s 1,800 km/h Centrifugal force: 97 tons / blade 3,200 tons / all blades Revolutions: 9,900 rpm Kinetic energy equivalent to a large tractor trailer passing at about 100 km/h More than ten tons of thrust May 23, 2012 Slide 5 filename

6 Dealing with Thrust: Axial Bearing Arrangement The axial bearing arrangement has to: cope with ever increasing demands in terms of performance, efficiency and cost fit inside the given design space withstand varying thrusts up to several tons and broad ranges of temperature under all service conditions ensure controlled, constant oil gaps over the whole operating range Floating disk Thrust collar Axial bearing Thrust May 23, 2012 Slide 6 filename Rotor shaft

7 Prediction of Mechanical and Thermal Loading Mechanical loading Thermal loading Centrifugal load Pressure distribution from the oil film interference fit assembly Heat flux from the oil film The assembly related loading and the centrifugal forces are easy to predict. The pressure distribution and the heat flux from the axial bearing is more complicated to predict and coupled to the deformation of the bearing surface. May 23, 2012 Slide 7 filename

8 Prediction of Mechanical and Thermal Loading The oil film properties can be calculated using CFD Turbine side Compressor side Temperature [K] Pressure [Pa] Oil velocity [m/s] Courtesy of Dr. B. Rembold, May 23, 2012 Slide 8 filename

9 Prediction of Mechanical and Thermal Loading The stationary solution for the mechanical and thermal interaction between the oil and the thrust collar can be calculated using a staggered FSI simulation Heat flux in y-direction Temperature Master Thesis F. Schmid, May 23, 2012 Slide 9 filename

10 Prediction of Mechanical and Thermal Loading The stationary solution predicted by the staggered FSI approach fits acceptably well to measurements executed for several working points (WP) May 23, 2012 Slide 10 filename Master Thesis F. Schmid

11 Optimization Using Isight: Targets and Restrictions Targets: Minimization of inclination and deflection of the bearing surface Generation of a slight inclination towards the floating disc in emergency stop situations Minimization of equivalent von Mises stresses under all service conditions Restrictions: Range of inclination and deflection of the bearing surface Maximum acceptable equivalent stresses and contact stresses Required minimal contact stresses May 23, 2012 Slide 11 filename

12 Optimization Using Isight: Parameterization Parameterization within Abaqus/CAE Direct access from Isight Seven complex variable parameters have been defined A DoE sensitivity analysis revealed four parameters with the most significant influence May 23, 2012 Slide 12 filename

13 Optimization Using Isight: Challenges Some parameter combinations might be invalid due to: physical inconsistence: negative volume contact violations being too complex in the manufacturing process inacceptable mesh quality These parameter combinations should not cause expenditure in terms of time, computational power and licensing! May 23, 2012 Slide 13 filename

14 Optimization Using Isight: Loop and Optimization Techniques Simplified optimization loop in Isight: Java script for setting the penalty function or a failed run flag Feasibility design check Geometry update, analysis and post-processing Additional postprocessing Evaluation of more complex targets and restrictions A genetic optimization algorithm could be used due to: the restricted number of parameters acceptably low simulation time achieved (axisymmetric model and excellent convergence rate) May 23, 2012 Slide 14 filename

15 Optimization Using Isight: Typical Simulation and Results Typical simulation containing several service conditions to be evaluated with regard to the optimization targets Equivalent stress distribution, deformations enlarged May 23, 2012 Slide 15 filename

16 Optimization Using Isight: Results Arbitrary initial design: Interesting improved design: Inclination range exceeded May 23, 2012 Slide 16 filename

17 Optimization Using Isight: Results Normalized deflection of the bearing surface for different service conditions May 23, 2012 Slide 17 filename

18 Conclusions and Outlook The adaptation of Isight to our computational environment appeared to be far from a plug and work solution. In optimization the investment in the pre-processing pays itself back with every iteration made (or skipped) in the loop. Optimization does not change the challenge of finding an acceptable compromise in competing targets. Unusual solutions might be found that might open the engineers mind for new features. The successful optimization loop is currently applied on the collar bearing design of the new two-stage turbocharging system. For these applications the staggered FSI analysis will be included in the optimization loop. May 23, 2012 Slide 18 filename

19 May 23, 2012 Slide 19 filename