- This tech paper is filed under:
Reducing Cycle Time and Increasing Throughput in Robotic Material Handling Applications
Advantages of using robots in material handling applications include speed, payload capacity and consistent productivity. Increases in throughput are typically necessary to keep pace with production demands without installing additional robot cells. The techniques presented here will enable you to attain higher productivity from a given amount of floorspace and capital investment. Shorter cycle times can be achieved through the use of the following practical cell layout, hardware and software considerations both during the design phase and after installation.
The layout of your system will have a major effect on the overall cycle time. Some of these items must be addressed before the cell is designed and others can be implemented on existing systems.
The location of the robot and peripheral equipment such as conveyors, machine tools, and pallet racks should be done with the sequence of operation in mind. Minimize the distance that the robot must travel from one program step to the next, especially on moves that require large movements of the slowest robot axes. For instance, the first axis, or base axis, of a six-axis vertically articulated robot is typically the slowest axis. If the robot is oriented such that a 270 degree rotation of the base axis is required to pick up a part and place it on a pallet, turn the robot around 180 degrees. Now the robot will only need to rotate its first axis 90 degrees during this motion. In robot applications that require many straight line motions, try to avoid operating near the extremes of the robot's reach or near singularities caused when multiple wrist axes line up with each other.
Explore alternate mounting orientations for the robot such as a wall or ceiling mount. A shelf-mounted robot with the conveyors running underneath it has become popular due to the wider variety of this robot type available today. If you have access to a robot simulation package, try a variety of positions for the robot and its peripheral equipment to find the best option with the lowest cycle time.
Look closely at the flow of parts into and out of the workcell. In a multiple robot system, if an assembly pallet enters and exits the robot workcell on a single conveyor spur, time will be lost shuttling pallets in and out. A better solution is to use a conveyor loop through the workcell so that pallets only flow in one direction. When a pallet is finished, it is released and the next pallet will enter the workcell right behind it. This provides better traffic control for the pallets and provides a buffer that minimizes the robot's waiting time.
As a first step, review the overall program flow. If the robot system has been operating for a while, the program may have been modified numerous times. Remove any motions or program steps that are no longer necessary. On six-axis articulated robot applications, minimize large rotations of the wrist axes, especially axis 6. Many times the wrist joints, axes 4 through 6, will be traveling through large rotations due to intermediate points having been added to the program. Also, end effector cable and hose routing considerations may necessitate large rotations to allow the cables to unwind from one move to the next. Reprogramming these positions to minimize rotation can make a major improvement in cycle time.
Correct use of your robot's built-in software functions is a cost-effective way to minimize cycle time. Various brands of robot controllers treat these parameters differently. Some of these programming techniques can be written into the application program code and set to optimize a given motion. This is the preferred method for most parameters. Others are global parameters that affect all motions for that robot and can't be modified during program execution.
Path accuracy settings control how tightly the robot will follow a given path between points, typically a straight line. Settling time or position level settings determine how closely the robot must get to a programmed point before executing the next step in the program. Set the path accuracy and settling time parameters as loose as possible (i.e. COARSE/CONTINUOUS instead of FINE). Use tight settings for critical motions, such as picking up a part, setting a part down, triggering an output signal or waiting for an input signal. This guarantees that the robot is in the correct position before the end effector closes or opens. On other motions, such as those between pick-up and set-down, set the parameters to a looser setting or use a continuous path setting so the robot does not stop or slow down. Modify these parameters carefully from tighter to looser to make sure the robot doesn't interfere with any fixturing, conveyors or guarding. Test the effect of these settings in both TEACH mode and PLAY/AUTO mode because these parameters can behave differently as the playback speed increases.
If the robot control software will allow, adjust the acceleration and deceleration settings for each motion. Ideally these parameters should be set at their maximum levels but this is not always feasible, especially when handling fragile, large, or long parts. In these cases, increasing the acceleration and deceleration time may allow for a higher velocity to be reached on long moves. All "air-cut" motions, non-process related moves with empty end effectors, should be run at the maximum possible velocity and acceleration. Avoid large changes in acceleration and deceleration from one motion to the next since these can cause the robot to slow down drastically or stop between motions.
In straight-line or linear paths, try to keep the robot away from singularity positions. As a robot travels near or through a singularity, it will typically fault out due to an axis over-speed condition. This can be avoided by reducing the robot velocity in these areas but this is detrimental to throughput. Better solutions include programming around the singularity position by adding more positions or using joint-interpolated moves since these are not affected by singularities. Also, joint motions allow the robot to achieve higher velocities than linear motions and should be used wherever possible. A straight line is the shortest path between two points but this does not always result in the shortest cycle time. Keep in mind that all robot axes do not travel at the same velocity during a linear motion and a particular axis may be limiting the maximum achievable velocity.
Perform several operations in parallel instead of in sequence. An example is opening the end effector during an approach motion instead of stopping and waiting at a point near the part to be grasped. In a palletizing system, an example would include turning on the vacuum supply to the suction cups at the program step just before the cups contact the corrugated case instead of waiting until they touch the case. If many operations or program steps need to be executed simultaneously, a separate PLC or internal soft PLC should be used.
Synchronization of the robot's movements with a conveyor may allow for a faster cycle time because this reduces the amount of starting and stopping of both the robot and the products being handled. If this software function is available with your robot controller, it can be implemented with the addition of an encoder to measure the conveyor speed and appropriate sensors to trigger the start of the measurement process.
Communications and networking delays should be taken into consideration, especially if the robot must transfer information during every program cycle. Handshaking delays after every piece of data is transferred can add up if a large quantity of data is being sent. If possible, transmit larger strings made up of multiple data bits instead of sending individual bits of data, one bit at a time.
All hardware used in the workcell should be analyzed for potential to increase the throughput of the system. This includes not only end effectors, but also conveyors, feeders, escapements, doors, pneumatic slides and vacuum systems.
End effector cylinders should use sensors for detecting position instead of relying on time delays. When time delays are used, they must be set longer than the actual time it takes for the gripper to open or close to guarantee that the motion has completed.
In applications that require rotation of the end effector, keep the end effector as compact as possible to reduce the rotational inertia. The lower the rotational inertia of the end effector, the faster the robot wrist will be able to accelerate and decelerate. Mount pneumatic valves, vacuum generators, electrical junction boxes and other large masses as close as possible to the center of rotation to reduce inertia or remove them from the end effector and mount them on the robot arm instead.
Gripping multiple parts will greatly improve throughput. In packaging applications, such as case packing and carton loading, it is common to handle 20-30 products per cycle. Machine tool load/unload applications use dual end effectors to unload a finished part from the spindle and then immediately load the next unprocessed part to maximize the utilization of the machine tool.
Pneumatic system flow rates will pace the actuation of cylinders and valves. Correct sizing of the air supply components, from the compressor to the workcell's main filter-regulator unit, must be done to make sure there is enough flow to allow the system to run at its maximum design speed. Using solenoid valves with insufficient flow ratings or undersized air lines and fittings will slow down gripper actuation. The use of too many fittings and elbows will reduce the flow rate in both pneumatic and vacuum systems. Quick-exhaust valves can be used on large volume cylinders for faster actuation.
Vacuum system design is another area to analyze to remove restrictions that cause increased cycle time. Minimize the volume that needs to be evacuated by locating vacuum generators as close to the vacuum cups as possible. Two areas of flow must be analyzed-the air supply to the vacuum generators and the vacuum supply to the suction cups. On venturi-style vacuum generators, use air supply lines that are large enough to supply the high flow rates that are typically required. Some robots' internal air lines are not large enough and will restrict air flow. In this case, use multiple internal lines to feed the vacuum generators or use a larger diameter, externally mounted supply line. The use of sensors to detect vacuum level at the suction cups is again preferred over software time delays. In applications where corrugated cardboard or other porous materials are being handled with vacuum cups, a high vacuum flow rate to the cups is necessary. This is in contrast to the high vacuum level/low vacuum flow requirements encountered when handling non-porous materials such as steel sheets. This is addressed by using the correct vacuum generator since different versions are designed specifically for each application.
Using an alternate device other than the robot to position a box or product will reduce the requirements of the robot. For example, a case tipper is typically used in a bottle packaging system to position the case at an angle. In high-speed packaging applications, a collation device such as a timing screw or racetrack is used to position multiple products while the robot is placing a group of products.
Maintenance of system components is critical to keeping performance at its maximum possible level. Worn seals and loose fittings on pneumatic cylinders will cause air flow and pressure losses that can increase gripper actuation times. Dirty cylinder rods and linear bearings can cause components to stick. Loose bolts can cause sliding components to bind and wear out. All these situations can cause the overall cycle time to increase over time and may result in someone speeding up the robot motion speeds to make up the difference. This will put unnecessary wear and tear on the robot drivetrain components.
Like any continuous improvement method, each of these techniques will not make a major improvement in system throughput by itself. Each technique may only save a fraction of a second, therefore implementation of multiple techniques is required. These techniques may require some trial and error to achieve optimum settings depending on the brand and style of robot and peripherals. Regardless, this effort will pay off through increased throughput in your robotic material handling systems. Even if you do not need to reduce your system cycle time, these techniques can be used to run your systems smoother and more efficiently by reducing wear and tear on robot, tooling and peripheral components.