A Hybrid Neural Network Approach for the Development of Friction Component Dynamic Model

M. Cao; K. W. Wang; Y. Fujii; W. E. Tobler

doi:10.1115/1.1649980

Journal of Dynamic Systems, Measurement, and Control | Volume 126 | Issue 1 | TECHNICAL PAPERS

< Previous Article Next Article >

TECHNICAL PAPERS

A Hybrid Neural Network Approach for the Development of Friction Component Dynamic Model

M. Cao, K. W. Wang, Y. Fujii and W. E. Tobler

[+] Author and Article Information

M. Cao, K. W. Wang

Department of Mechanical and Nuclear Engineering,The Pennsylvania State University, University Park, PA 16802

Y. Fujii, W. E. Tobler

Research and Advanced Engineering, Ford Motor Company, Dearborn, MI 48121

J. Dyn. Sys., Meas., Control 126(1), 144-153 (Apr 12, 2004) (10 pages) doi:10.1115/1.1649980 History: Received May 14, 2002; Revised August 25, 2003; Online April 12, 2004

Article

References

Figures

Tables

Citing Articles

Purchase this Content

$25.00

Purchase

Learn about subscription and purchase options

Your Session has timed out. Please sign back in to continue.

References

Wu, H., 1978, “A Review of Porous Squeeze Films,” Wear, 47, 371–385.

Ting, L. L., 1975, “Engagement Behavior of Lubricated Porous Annular Disks,” Wear, 34, 159–182.

Natsumeda, S., and Miyoshi, T., 1994, “Numerical Simulation of Engagement of Paper Based Wet Clutch Facing,” J. Tribol., 116, pp. 232–237.

Jacobson, B., 1992, “Engagement of Oil Immersed Multi-Disc Clutches,” International Power Transmission and Gearing Conference, 2 , pp. 567–574.

Berger, E. J., Sadeghi, F., and Krousgrill, C. M., 1997, “Analytical and numerical modeling of engagement of rough, permeable, grooved wet clutches,” ASME J. Tribol., 119, 143–148.

Yang, Y., Lam, R. C., Chen, Y. F., and Yabe, H., 1995, “Modeling of Heat Transfer and Fluid Hydrodynamics for a Multi-disc Wet Clutch,” SAE Special Publication-Paper 950898.

Fujii, Y., Tobler, W. E., and Snyder, T. D., 2001a, “Prediction of Wet Band Brake Dynamic Engagement Behavior Part I: Mathematical Model Development,” Proc. Inst. Mech. Eng., Part D (J. Automob. Eng.), 215(D4), pp. 479–492.

Fujii, Y., Tobler, W. E., and Snyder, T. D., 2001b, “Prediction of Wet Band Brake Dynamic Engagement Behavior Part 2: Experimental Model Validation,” Proc. Inst. Mech. Eng., Part D (J. Automob. Eng.), 215(D5), pp. 603–611.

Fujii, Y., Tobler, W. E., Clausing, E. M., Megli, T. W., and Haghgooie, M., 2002, “Application of Dynamic Band Brake Model for Enhanced Drivetrain Simulation,” Proc. Inst. Mech. Eng., Part D (J. Automob. Eng.), 216(D11), pp. 873–881.

Fanella, R., 1994a, “Design of Bands,” Design Practices-Passenger Car Automatic Transmissions 3rd ed., 1994, (Soc. Automotive Eng., New York).

Fanella, R., 1994b, “Design of Friction Clutches,” Design Practices-Passenger Car Automatic Transmissions 3rd ed., 1994 (Soc. Automotive Eng., New York).

Cybenko, G., 1989, “Approximation by Superpositions of a Sigmoidal Function,” Math. Control, Signals, Syst., 2, pp. 303–314.

Funahashi, K., 1989, “On the Approximate Realization of Continuous Mapping by Neural Networks,” Neural Networks, 2(3), pp. 183–192.

Hornik, K., Stinchcombe, M., and White, H., 1989, “Multilayer Feedforward Networks are Universal Approximators,” Neural Networks, 2, pp. 359–366.

Narendra, K. S., and Parthasarathy, K., 1990, “Identification and Control of Dynamical Systems Using Neural Networks,” IEEE Trans. Neural Netw., 1(1), pp. 4–27.

Parlos, A. G., Atiya, A. F., Chong, K. T., and Tsai, W. K., 1992, “Nonlinear Identification of Process Dynamics Using Neural Networks,” Nucl. Technol., 97, pp. 79–96.

Funahashi, K., and Nakamura, Y., 1993, “Approximation of Dynamical Systems by Continuous Time Recurrent Neural Networks,” Neural Networks, 6, pp. 801–806.

Sio K. C., and Lee, C. K., 1996, “Identification of a Nonlinear Motor System with Neural Networks,” International Workshop on Advanced Motion Control, Part 1, Mar 18–21, 1, pp. 287–292.

El-Gindy, M., and Palkovics, L., 1993, “Possible Application of Artificial Neural Networks to Vehicle Dynamics and Control: a Literature Review,” Int. J. Veh. Des., 14, pp. 592–614.

Lai J. S., 1995, “Intelligent Robust Control of Clutch-to-Clutch Shifts in Vehicle Transmission Systems,” Ph.D. Thesis In Mechanical Engineering, Penn State University.

Parvataneni, V. A., First Principle-based Hybrid Neural Network Clutch Dynamics Model. M.S. Thesis in Mechanical Engineering, the Pennsylvania State University, August, 1998.

Parvateneni, V., Cao, M., Wang, K. W., Fujii, Y., and Tobler, W. E., 1999, “Hybrid Neural Network for Modeling Automotive Clutches,” Proc. ASME Dynamic Systems and Control Division, November, DSC-67 , 255–263.

Crouch, R. F., and Cameron, A., 1961, “Viscosity-Temperature Equations for Lubricants,” J. Inst. Pet., 47, pp. 307–313.

Dowson, D., and Higginson, G. R., 1977, Elasto-Hydrodynamic Lubrication, Pergamon Press.

Press, W. H., Flannery, B. P., Teukolsky, S. A., and Vetterling, W. T., 1986, Numerical Recipes: The Art of Scientific Computing, Cambridge University Press, Cambridge, England.

Bose, N. K. and Liang, P., 1996, Neural Network Fundamentals with Graphs, Algorithms, and Applications, McGraw-Hill, Inc.

Bassani, R. and Piccigallo, B., 1992, Hydrostatic Lubrication, Elsevier Science Publishers B. V.

Figures

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m)

View Large | Save Figure | Download Slide (.ppt)

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m)

View Large | Save Figure | Download Slide (.ppt)

Testing Results of Hybrid Neural Network Model: –Measured Torque [[dashed_line]]Network Estimated Torque [[dotted_line]] Torque Estimated by the First Principle Model

View Large | Save Figure | Download Slide (.ppt)

Baseline Network Architecture z⁻¹=One step time delay

View Large | Save Figure | Download Slide (.ppt)

Measured Band Brake Torque with Different Oil Temperature, Apply Pressure and Slip Profiles

View Large | Save Figure | Download Slide (.ppt)

Hybrid Network Architecture

View Large | Save Figure | Download Slide (.ppt)

Friction Component Engagement Test Stand

View Large | Save Figure | Download Slide (.ppt)

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/sec (1500 rpm), Final Apply Pressure: 6.206e+4 N/m² (9 psi), Initial Oil Temperature: 394.1 K (250°F), Oil Flow Rate: 2.273e−5 m³/sec (0.3 gallon/min)

View Large | Save Figure | Download Slide (.ppt)

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/s (1500 rpm), Final Apply Pressure: 6.206e+4 N/m² (9 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m³/s (0.3 gallon/min)

View Large | Save Figure | Download Slide (.ppt)

Training Results of Baseline Neural Network Model Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m² (15 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m³/s (0.3 gallon/min)

View Large | Save Figure | Download Slide (.ppt)

Testing Results of Baseline Neural Network Model –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m) Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m² (15 psi), Initial Oil Temperature: 310.8 K (100°F), Oil Flow Rate: 4.546e−5 m³/s (0.6 gallon/min)

View Large | Save Figure | Download Slide (.ppt)

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated (N.m)

View Large | Save Figure | Download Slide (.ppt)

Tables

Errata

Web of Science® Times Cited: 8

Some tools below are only available to our subscribers or users with an online account.

Sections
PDF
Email

You must be logged in as an individual user to share content.

Return to: A Hybrid Neural Network Approach for the Development of Friction Component Dynamic Model

Copyright in the material you requested is held by the American Society of Mechanical Engineers (unless otherwise noted). This email ability is provided as a courtesy, and by using it you agree that you are requesting the material solely for personal, non-commercial use, and that it is subject to the American Society of Mechanical Engineers' Terms of Use. The information provided in order to email this topic will not be used to send unsolicited email, nor will it be furnished to third parties. Please refer to the American Society of Mechanical Engineers' Privacy Policy for further information.
Share
Get Citation

Cao MM, Wang KW, Fujii YY, Tobler WE. A Hybrid Neural Network Approach for the Development of Friction Component Dynamic Model. ASME. J. Dyn. Sys., Meas., Control. 2004;126(1):144-153. doi:10.1115/1.1649980.

Download citation file:

RIS (Zotero)

EndNote

BibTex

Medlars

ProCite

RefWorks

Reference Manager

© Copyright
Get Alerts

Processing your request... Please Wait.

A Hybrid Neural Network Approach for the Development of Friction Component Dynamic Model

References

Figures

CPU Time Comparison: Hybrid ANN versus First Principle Model

Cylindrical Co-ordinates for Reynolds Equation: Plate-type Friction Component

X-Y-Z Co-ordinates for Reynolds Equation: Band-type Friction Component

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m)

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m)

Testing Results of Hybrid Neural Network Model: –Measured Torque [[dashed_line]]Network Estimated Torque [[dotted_line]] Torque Estimated by the First Principle Model

Baseline Network Architecture z⁻¹=One step time delay

Measured Band Brake Torque with Different Oil Temperature, Apply Pressure and Slip Profiles

Hybrid Network Architecture

Friction Component Engagement Test Stand

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/sec (1500 rpm), Final Apply Pressure: 6.206e+4 N/m² (9 psi), Initial Oil Temperature: 394.1 K (250°F), Oil Flow Rate: 2.273e−5 m³/sec (0.3 gallon/min)

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/s (1500 rpm), Final Apply Pressure: 6.206e+4 N/m² (9 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m³/s (0.3 gallon/min)

Training Results of Baseline Neural Network Model Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m² (15 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m³/s (0.3 gallon/min)

Testing Results of Baseline Neural Network Model –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m) Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m² (15 psi), Initial Oil Temperature: 310.8 K (100°F), Oil Flow Rate: 4.546e−5 m³/s (0.6 gallon/min)

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated (N.m)

Tables

Errata

Related Content

Journals

Conference Proceedings

eBooks

A Hybrid Neural Network Approach for the Development of Friction Component Dynamic Model

References

Figures

CPU Time Comparison: Hybrid ANN versus First Principle Model

Cylindrical Co-ordinates for Reynolds Equation: Plate-type Friction Component

X-Y-Z Co-ordinates for Reynolds Equation: Band-type Friction Component

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m)

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m)

Testing Results of Hybrid Neural Network Model: –Measured Torque [[dashed_line]]Network Estimated Torque [[dotted_line]] Torque Estimated by the First Principle Model

Baseline Network Architecture z−1=One step time delay

Measured Band Brake Torque with Different Oil Temperature, Apply Pressure and Slip Profiles

Hybrid Network Architecture

Friction Component Engagement Test Stand

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/sec (1500 rpm), Final Apply Pressure: 6.206e+4 N/m2 (9 psi), Initial Oil Temperature: 394.1 K (250°F), Oil Flow Rate: 2.273e−5 m3/sec (0.3 gallon/min)

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/s (1500 rpm), Final Apply Pressure: 6.206e+4 N/m2 (9 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m3/s (0.3 gallon/min)

Training Results of Baseline Neural Network Model Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m2 (15 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m3/s (0.3 gallon/min)

Testing Results of Baseline Neural Network Model –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m) Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m2 (15 psi), Initial Oil Temperature: 310.8 K (100°F), Oil Flow Rate: 4.546e−5 m3/s (0.6 gallon/min)

Training Results of Hybrid Neural Network Model: –Measured Torque (N.m) [[dashed_line]]Network Estimated (N.m)

Tables

Errata

Related Content

Journals

Conference Proceedings

eBooks

Baseline Network Architecture z⁻¹=One step time delay

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/sec (1500 rpm), Final Apply Pressure: 6.206e+4 N/m² (9 psi), Initial Oil Temperature: 394.1 K (250°F), Oil Flow Rate: 2.273e−5 m³/sec (0.3 gallon/min)

Training Results of Baseline Neural Network Model Initial Slip: 157.1 rad/s (1500 rpm), Final Apply Pressure: 6.206e+4 N/m² (9 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m³/s (0.3 gallon/min)

Training Results of Baseline Neural Network Model Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m² (15 psi), Initial Oil Temperature: 271.9 K (30°F), Oil Flow Rate: 2.273e−5 m³/s (0.3 gallon/min)

Testing Results of Baseline Neural Network Model –Measured Torque (N.m) [[dashed_line]]Network Estimated Torque (N.m) Initial Slip: 314.2 rad/s (3000 rpm), Final Apply Pressure: 1.034e+5 N/m² (15 psi), Initial Oil Temperature: 310.8 K (100°F), Oil Flow Rate: 4.546e−5 m³/s (0.6 gallon/min)