INTRODUCTION

Applying multibody dynamics simulation to the design of robotic arms can offer several advantages. MBD can provide a comprehensive understanding of a robotic arm’s kinematics and dynamics. Adjusting link lengths, joint types, or actuator parameters is quick and easy which allows engineers to test and optimize robot arm performance to meet design requirements. During the validation phase, FEA can be combined with MBD to understand component stresses throughout a pick-and-place robot’s cycle. Lastly, control engineers can test and refine control algorithms to reduce cycle time and improve precision.

In this example, we will build a model of a pick-and-place robotic arm with a suction cup end-effector. The robot transfers metal sheets from one pallet to another. This type of analysis could be modified to accommodate vehicle body panels, glass panes, or other objects moved via suction cup. The model is used to review the stresses on the metal sheet and compute the motor torques necessary throughout the pick-and-place operation.

Understanding the Model Definition in MotionSolve

The MotionSolve model is built from simple geometry but complex geometry can be imported from CAD data. Joints are built between each segment of the robotic arm to mathematically capture the relationships between each link. In the real-world, suction cup end-effectors often use a vacuum to actuate the suction cups. In the MBD model, the suction cups are modeled with force entities. Once the suction cup is in contact with the panel and a set amount of time has passed, a force is applied to grip the panel. After the robot reaches the drop location, the force decreases to zero allowing the panel to drop.

The contacts between the panel and pallets are modeled using forces dependent on the interference between parts and contact stiffness. Aerodynamic forces are modeled across the surface of the panel dependent on the velocity and surface area of the panel. Together contact forces and aerodynamic forces accurately model the operating conditions the robot arm must overcome.

Understanding the Model Definition in Twin Activate

The MotionSolve model uses functional mock-up units (FMU) to model the motors and motor controllers responsible for actuating the robot arm. FMUs are one method used to simulate complex, multi-physics models. An FMU can be built in a program which accurately models one domain. Then the FMU can be exported to another program which accurately models a different domain. FMUs are available in two versions: Model exchange and co-simulation. Model exchange (ME) provides the equations and algorithms necessary to run a multi-physics simulation. Co-simulation (CS) runs two programs in parallel that pass data between each other. Co-simulation FMUs allow users more control over the internal algorithms but not all programs support co-simulation FMUs. MotionSolve supports both model exchange and co-simulation. Two models are included in the attached zip file to demonstrate both options.

In this case, the FMU is built in Twin Activate, an intuitive 1D modeling software for multi-physics and controls. The FMU receives the reference position, actual position, and velocity of each motor from MotionSolve. A PID controller and motor model calculate the torques to apply to each motor. MotionSolve receives the torques from FMU and uses them to calculate the new positions and velocities of each body in the model. This process repeats for each timestep until the robot arm has completed one cycle.

SOFTWARE REQUIREMENTS

MotionView (2024 or newer)

MotionSolve (2024 or newer)

Twin Activate (2024 or newer)

MODEL FILES

Pick_and_Place_Robot.zip (See Attachments)

MODEL SETUP & SIMULATION STEPS

1. Open Robot_FMU_CS.mdl or Robot_FMU_ME.mdl in MotionView.
2. Select run analysis and choose an appropriate output directory.
3. Open the resulting h3d file in HyperView to review the animation.

RESULTS

The results can be animated in HyperView. Generating a stress contour plot shows the sheet reaches 119 MPa during the pick and place cycle. Material properties can be used to determine a desirable limit on the stress parts experience. If the stresses are too high, the speed of the robot can be decreased to prevent parts from failing.

HyperGraph allows users to plot and analyze performance indicators. Torques can be plotted to determine the specifications a motor must meet to reach to achieve the desired system level performance.

CONCLUSION

Multibody dynamics simulation can provide significant value to engineers designing robot arms throughout the design cycle. From rapidly testing different arm designs to predicting the torques necessary to achieve efficient operation, MBD can save time and cut costs. Combining MBD with finite element analysis, the robot arm or objects picked up by the robot arm can be analyzed for stresses. This data can be used to minimize cycle time and to minimize parts damaged during pick-and-place operations.

AUTHORS

John Dagg, Systems Engineering Intern

Christopher Fadanelli, Solutions Engineer – Systems Integration

Ananth Kamath Kota, Global Technical Manager – Systems Integration