Case Study

Multi-Physics with MotionSolve Innovation Intelligence®

Rajiv Rampalli & Keshav Sundaresh, Altair

Copyright © 2012 Altair Engineering, Inc. Proprietary and Confidential. All rights reserved.

Presentation Outline • Need for Multi-Physics simulations • Case Study: Control of a Wind Turbine • How other customers are doing Multi-Physics simulations

2


Real Models Are Complex

Hydraulics H d li & Control systems

Virtual tires & digitized roads

Flexible Fl ibl & Non-linear N li components

Virtual driver models

3


Industrial Processes Are Complex TIRE & ROAD

MBD

FATIGUE G

POST

DRIVER

FE DRIVER

FE 4


The Conundrum

Systems

Software

Multiple Domains

Single Domain

Diverse Phenomena

Single Formulation

Many Technologies

One Integrator

5


Introducing MotionSolve… • A natural environment for multi-disciplinary studies • Open architecture for coupling domains • Many different choices available for coupling

6


Model Connection Methodologies Master System Equations

S-1

S-2

S-2 Co-Simulation (Sharing signals)

Sharing Equations Simulator-1

Simulator-2 Shared Data

Sharing Linear Models 7


Sharing Linear Models: With Simulink

M ti S l MotionSolve

Matlab Simulink

MotionSolve exchanges state matrices with Matlab/Simulink 8


Sharing Linear Models: With FE

HyperMesh Radioss

MotionView MotionSolve

MotionSolve shares data with a variety of FE codes 9


Sharing Equations: via User Subroutines

Object Code

User MotionSolve Python Script Subroutine DLL

MotionSolve

User subroutines allow users to to U b ti ll t link li k their th i own equations ti t MS 10


Sharing Equations: via Code Export

Object Code

Motion Solve

Motion Solve DLL

“C” Code

RTW

Simulink Controller Co simulation Co-simulation

Native RTW code import from Simulink 11


Sharing Signals: MotionSolve + Simulink

System Model MotionSolve model

Controller

S-Function (MS ) Simulink Window

MotionSolve is an S-Function in Simulink 12


Sharing Signals: MotionSolve + DSHplus

System Model

Hydraulic System

User Block (MS ) DSHplus Window

MotionSolve is a user block in DSHplus 13


Sharing Signals: MotionSolve + Driver

System y Model Controls & Switching Signals

Speed Controller

Gearbox Controller Feedback Signals

Controller Outputs Steering Controller

Vehicle Model ((MS))

Driver Model

MotionSolve can include complex driver models via co-simulation 14



15


Co-simulation – Wind Turbine Emergency Stop

“Grid Grid loss” loss – Sudden drop in generator torque Reaction: Immediate blade pitching and activation of the brake disk. E-stop event •

0-60s 120.0s 120.4s 135.0s

Ramp up to full effect, 1MW. “Grid loss” 100-0% in 0.1s Maximum brake pressure & reversing pitch angle Pitching complete

Mechanical system, Wind loads – lift and drag forces v/s pitch angle, Tower shadow - %ge drop of wind forces in front of the tower

16


Initial Results • Grid loss and braking generate massive rotor t torque t variations i ti • Large blade vibrations during overshoot • Blades vibrate during initial braking, causing high frequency variations of rotor torque.

17


Control Requirements • Implement a non-linear braking pressure to: •

Reduce blade deflection

•

Eliminate overshoot

Brake controller: Blade deflection v/s Brake magnifier Brake release: Rotational speed v/s release factor Low pass filter to the braking signal – to avoid rapid vibrations

18


Control Model in Simulink

Rotor Velocity

C t lP Control Parameter t Plant Outputs

Flex Sensors

Flex Sensors Plant Outputs Rotor Velocity

Controller MS Model

Flex Sensors

Rotor Velocity

1

1

Sensors

z

z Blade Deflection

19


Co-Simulation Results

100% Rotor Speed Simulink Rotor Speed

100% R Rotor t S Speed d

Si Simulink li k Rotor R t Speed S d 100% Rotor Speed Simulink Rotor Speed

Brake controller: Blade deflection v/s Brake magnifier Brake release: Rotational speed v/s release factor

20



21


Faurecia – Glass guidance simulation • Challenge: •

Find the cable positions to meet requirements on normall fforces ffor different diff t combinations bi ti off friction.

• Solution: •

Mechanism analysis with MotionSolve to find the normal forces

•

Design optimization with HyperStudy to find the optimal cable positions.

• Benefits: •

An optimal design that meets all conflicting requirements.

•

A fast & automated design process leading to fewer errors & better designs.

“Co-simulation of window regulator with MotionSolve and Matlab/Simulink gives a full mechatronics model of the system, allowing simultaneous design of both mechanism and actuator” – Dany Desrus, Leader Product Line EE & Mechatronics, Faurecia Interior Systems22


Summary • Real systems are complex • A full understanding of these requires multi-disciplinary simulation • MotionSolve has an architecture that facilitates coupling of different solvers to enable multi-disciplinary simulation y different options for coupling g are available • Many • The simulation requirements and coupling difficulty dictate which choice should be used.

Thank You!

23

Rajiv Rampalli Innovation

Intelligence®

Keshav Sundaresh([email protected])