Modelling Using STAR-CCM+ Gary Yu, Martin Timmins and Mario - - PowerPoint PPT Presentation

DENSO MARSTON LTD.

DNMN Product Development

40-Ton Articulated Truck Cooling System Modelling Using STAR-CCM+

Gary Yu, Martin Timmins and Mario Ciaffarafa

DENSO Marston Ltd, Bradford, BD17 7JR, UK

DENSO MARSTON LTD.

DNMN Product Development

• Founded in 1904
• Acquired by DENSO in 1989
• Located in Shipley, West Yorkshire
• Designs and Manufactures engine

cooling modules for Heavy Duty Cooling applications

• Product Range includes radiators, oil

coolers, inter-coolers and condensers

DENSO Marston

DENSO MARSTON LTD.

DNMN Product Development

• Research project partly funded by the UK government - Validation of

Complex Systems (VOCS);

• DNMN was part of a consortium of UK based companies participating in the

project which was led by a major off highway original equipment manufacturer (OEM);

• The project covered many aspects of vehicle simulation, with DNMN

focussing on cooling system simulation and validation. Acknowledgement:

This work was co-funded by the Technology Strategy Board (TSB) and One North East, UK under the Validation of Complex Systems (VOCS) grant programme. The Technology Strategy Board is an executive body established by the United Kingdom Government to drive innovation. It promotes and invests in research, development and the exploitation of science, technology and new ideas for the benefit of business

• increasing sustainable economic growth in the UK and improving quality of life.

Project Background

DENSO MARSTON LTD.

DNMN Product Development

Outline

• 1. 1D KULI simulation of cooling system – coupled with STAR-CCM+ single

stream heat exchanger model ;

• 2. 3D CFD simulation of cooling system – STAR-CCM+ dual stream heat

exchanger model;

• 3. Comparison of 1D KULI and 3D CFD results.

Charge air cooler (CAC) and cooling fan Radiator module including radiator, condenser, oil cooler and cooling fan Tier 3 Off Highway Truck

DENSO MARSTON LTD.

DNMN Product Development

Computational Domain and Boundary Conditions

Vehicle speed (m/s) Wind tunnel size 22 X 13 X12 L X W X H (m^3) Top and Sides Pressure outlet Ground Wall Ambient temperature (oC) 20 Radiator fan rpm 1500 Charge Air Cooler fan rpm 2100

• Vehicle overall size: 10.889 m (L) x 3.43 m (W) x 3.745 m (H);
• Vehicle modelled at stationary and idle condition.

DENSO MARSTON LTD.

DNMN Product Development

Truck Model Meshing

Import CAD Surface Wrap Surface Remesh Volume Mesh

volume mesh refinement

Mesh size : 8mm (in dense area) Number of cells : ~21.6 million

DENSO MARSTON LTD.

DNMN Product Development

Numerical Model

1. 3D, steady state flow, κ-ε model for turbulence; 2. Cooling air, charged air modelled as ideal gas, whilst radiator coolant, oil cooler oil as constant density; 3. Segregated flow temperature; 4. Heat exchangers modelled as porous media for fluid flow, dual stream model for heat rejection calculation; 5. Moving reference frame (MRF) fan model, momentum source model used but with convergence problem.

DENSO MARSTON LTD.

DNMN Product Development

Coupling of KULI with CFD

• 1. Air flow distribution is uniform in

traditional 1D KULI model across heat exchanger face;

• 2. 3D distribution based on velocity

profile from CFD simulation.

Velocity distribution

KULI is a 1D tool used for vehicle cooling system simulation

Images courtesy of Magna Powertrain

DENSO MARSTON LTD.

DNMN Product Development

Coupling of KULI with CFD

CFD velocity field (isothermal) KULI resistance matrix KULI Model CFD single stream heat exchanger model CFD velocity field (thermal) STOP

Q

• It is very expensive for CFD volume cells > 20M;
• Dual stream heat exchanger model in

STAR-CCM+ can avoid this.

• 1. Initial CFD isothermal simulation

results;

• 2. KULI generation of resistance

matrix;

• 3. KULI modelling for heat rejection

rate, Q;

• 4. Update of CFD model with Q values

for all heat exchangers;

• 5. Calculation of thermal velocity field

by CFD;

• 6. Feedback to KULI for next iteration

until converged.

Typical procedure:

DENSO MARSTON LTD.

DNMN Product Development

Dual Stream Heat Exchanger CFD Model

Test Data Required: Cold stream side: Q Table for heat rejection rate and mass flow rate; Hot stream side: Mass flow rate in the test should be the same as the condition modelled in CFD.

Thermal output (kW) CAC cold stream flow rate (kg/min)

hot stream mass flow rate @ actual working condition

Heat exchanger 3D heat transfer map from KULI Use KULI to obtain actual working condition

DENSO MARSTON LTD.

DNMN Product Development

Determination of Porous Media Pressure Drop Coefficients

• Pressure drop over heat exchanger core is calculated by DP/DL=αv2+βv , α

and β are determined based on test data, v is the inlet flow velocity;

• CFD under predicts the pressure drop at a fixed mass flow rate because it

does not use the correct throughway area, so needs to be corrected using these equations.

Real Geometry

CFD

αcfd= (ρcfd /ρin)2(Acfd /Atest)2αtest αcfd= (Acfd /Atest)2αtest βcfd= (Acfd /Atest)βtest βcfd= (ρcfd /ρin) (Acfd /Atest)βtest

Charge Air Cooler Radiator, Oil cooler

DENSO MARSTON LTD.

DNMN Product Development

Example Results from Dual Stream Model: Charge Air Cooler Charged Air Temperature Distribution

• Better cooled in the core center corresponding to the

cooling air velocity profile on front face;

• Charged air temperature drop: ΔT≈115o C.

Air flow distribution matches front grille

DENSO MARSTON LTD.

DNMN Product Development

Example Results from Dual Stream Model: Radiator Coolant Temperature Distribution

• Better cooled on one side due to the cooling air

velocity profile on front face;

• Coolant temperature drop: ΔT≈6o C.

Air flow distribution matches layout (more complex than CAC)

DENSO MARSTON LTD.

DNMN Product Development

Comparison of 1D KULI and 3D CFD results

• Cooling air velocity and temperature fields from CFD dual

stream heat exchanger model used by KULI to compare the heat rejection and pressure drop predictions;

• Working conditions modeled in CFD and KULI:
• Mass flow rate and inlet temperature of internal fluid in

each heat exchanger e.g. coolant, charge air and oil;

• Ambient temperature.

DENSO MARSTON LTD.

DNMN Product Development

Comparison of CFD and Initial KULI Model Results

• Big discrepancy in Oil Cooler predictions

DPcold stream DPhot stream Heat transfer rate (%) (%) (%) KULI CFD KULI CFD KULI CFD Radiator 103 100 99 100 95 100 Oil Cooler 140 100 99 100 136 100 Charge Air Cooler 96 100 90 100 95 100

DENSO MARSTON LTD.

DNMN Product Development

KULI Model Set-up Based on CFD Results

Uniform cooling air temperature

Basic 1D KULI model requires Ambient temperature plus uniform warm up STAR-CCM+ Simulation 2D temperature map T1 T3 T2 T4 Improved 1D KULI model 4 blocks with targeted temperatures from CFD

DENSO MARSTON LTD.

DNMN Product Development

KULI Model Set-up Based on CFD Results

• Dummy plane in front of heat exchangers to show cooling air inlet

temperature distribution;

• From CFD results, 4 separate inlet temperature targets / zones are required in

the KULI model.

DENSO MARSTON LTD.

DNMN Product Development

KULI Model Set-up Based on CFD Results

mcfd = m1 = 0.24 kg/s mKULI = m1 + m2 = m3 = 0.404 kg/s

m1 m2 m3 Oil Cooler RAD

• In basic KULI model, cooling air mass flow through oil cooler is m3 which is higher

than real condition & CFD;

• From CFD results, a separate mass flow target is required for the oil cooler.

DENSO MARSTON LTD.

DNMN Product Development

KULI Model Set-up Based on CFD Results

1. Two mass flow targets, one of which is for Oil Cooler; 2. Four blocks; 3. Four cooling air inlet temperature targets; 4. Three resistance matrixes; 5. No resistance matrix for oil cooler.

DENSO MARSTON LTD.

DNMN Product Development

Comparison of CFD and Improved KULI Model Results

• Good correlation between KULI and CFD models

DPcold stream DPhot stream Heat transfer rate (%) (%) (%) KULI CFD KULI CFD KULI CFD Radiator 105 100 99 100 103 100 Oil Cooler 94 100 99 100 95 100 Charge Air Cooler 96 100 90 100 95 100

DENSO MARSTON LTD.

DNMN Product Development

Conclusions and Summary

• Standard procedure for coupling KULI and CFD requires

multiple iterations and is not practical in this case;

• The dual stream heat exchanger model in STAR-CCM+ is

efficient and was used successfully to simulate the cooling system of the 40-ton truck;

• STAR-CCM+ can be used to help generate an improved

KULI model;

• STAR-CCM+ and KULI predictions agree well.

DENSO MARSTON LTD.

DNMN Product Development

THANK YOU