Where accuracy is concerned, robots have traditionally relied on repeatability. In the past, robotic accuracy has not been developed to a level of maturity acceptable to standard production processes. Critical aerospace manufacturing techniques such as fastening and drilling were historically not held to tight tolerances. Typical tolerances for airframe assembly fastening were in the +0.030" range. The standard is set by the positional requirement for drilling of fastener holes, which is a key target application for robotics in manufacturing.

FANUC Robotics' Learning Vibration Control (LVC) allows a robot to learn its vibration characteristics for higher accelerations and speeds. After a path is taught, an accelerometer is added to the tool, and the path is run several times in order for LVC to collect learning data. Using the retrieved data, the robot optimizes the motion to have shorter accelerations while keeping vibration to a minimum. The accelerometer is only used during the learning process.
Because there are many factors that influence robot accuracy, it is important to define the accuracy requirements for the system. Different levels of accuracy require different solutions; the higher the accuracy required, the more factors that must be considered, adding to the cost and complexity of the solution. The level of accuracy should be defined according to the process requirements. Some processes will only require positional accuracy while others require path accuracy, and some applications will require both.

Recently, for example, manufacturers have begun demanding that high-wear parts that require frequent maintenance and replacement be replaced seamlessly with identically manufactured parts. Inconsistent and inaccurately machined or assembled replacement parts might traditionally have meant time lost due to trimming, deburring, or other adjustments. Reducing fastener tolerances not only improves the reproducibility of an assembled component, but also allows for a reduction in overall structure weight due to reduced fastener size and weight. Eliminating these adjustments by machining or assembling precisely formed parts allows for predictable and timely part replacement, reducing costs and downtime, and allowing for parts to be interchanged repeatedly without any interruption in production. The introduction of robotic accuracy into the manufacturing process guarantees that this replacement is smooth, does not interrupt the manufacturing process, and is cost-effective and highly accurate.

High accuracy is also critical in data-driven applications — those developed using offline programming methods. For example, an advanced deburring process can begin offline with PC-based simulation software that allows users to import 3D representations of robots, parts, and system peripherals to create realistic “virtual” workcells. Part features can be selected from 3D CAD data of the part. Robot programs can then be generated automatically from these selected features. Even vision programming can be accomplished off-line. Until recently, it was not easy to offline-program a robot without having to perform extensive, manual touch-up of the programs on the production floor. Because of the large number of points required for airframe drilling, it is mandatory to offlineprogram without extensive program touch-up.

Robot Accuracy and Repeatability

Robot accuracy is a measure of how close a robot can attain a known position. It is required for systems where the paths are taught offline or if the process requires changing the robot position dynamically using vision or another means. Robot repeatability is a measure of the robot’s ability to return to a known position. High robot accuracy during manufacturing ensures that parts are precisely manufactured with predictable results, even after changes are made to the process. High-accuracy robots are becoming valuable tools for many processes in aerospace manufacturing such as drilling and fastening, deburring and trimming, and a variety of others such as non-destructive inspection, coatings, and composite layup. Manufacturers can enjoy significant cost savings as a result.

Systems that combine processes like drilling, routering, and material removal require both positional accuracy and path accuracy. Positional accuracy is a measure of how accurately the robot can achieve a commanded position. Positional accuracy is required for processes like drilling, where the robot moves to a position, stops, and holds that position while the process is completed. Path accuracy is a measure of how accurately the robot follows a line between two points. Path accuracy is required for processes like laser cutting where the process is taking place while the robot is moving between points.

Robot accuracy is improved when the work zone is defined as localized as possible. It is important to define where in the robot’s work envelope the process will take place. This is called the process work zone. A higher level of accuracy is achievable if the process work zone is defined and the calibration is restricted to this zone. When defining a process work zone, there are three considerations to follow. First, the process work zone needs to include all processes that require accuracy. Second, make the zone only as large as the process requires. Third, limit robot configuration (orientation) changes in the process work zone as much as possible.

A FANUC R-2000iB robot simulates drilling an aerospace panel. Absolute accuracy in a given area allows the robot to be accurately positioned. Secondary encoders allow the robot to further enhance the accuracy by being able to control the robot position within the backlash band.
Different levels of accuracy require different solutions. The required level of robot accuracy determines the number of options and calibration tools required to achieve that accuracy. The more calibration tools required, the more complex and expensive the solution will be.

Repeatable robot paths and tool execution means critical material cost savings in removal applications. An added benefit of using accurate robots for aerospace manufacturing is the inherent repeatability of robotic processes, allowing for better predictability and control of process parameters. This makes it easier to identify and refine process parameters that affect component quality. In addition, robots can execute complex or repetitive processes at very high speeds.

The most significant game-changing process in aerospace manufacturing is carbon fiber layup. In this process, carbon fibers are combined with a resin or epoxy material to create a lightweight, but strong composite. This material is highly suited for the aerospace industry because it can reduce the weight of the airplane in order to achieve better fuel economy without sacrificing strength or durability. Robotic accuracy is important in this process because the placement of the carbon fiber strands relative to each other is critical to the structural integrity of the component.

One important feature used to achieve high accuracy in robots is the use of secondary encoders. This reduces omnidirectional repeatability to nearly zero, and has been validated via laser tracker while exploiting the combined effects from moving all axes. Secondary encoders connected directly to the controller are installed on the output side of each axis drive train to measure and control the true position of each axis. This allows the robot to control position, eliminating errors due to backlash and essentially improve the ability of the robot to achieve a commanded position. This is ideal for applications that require high precision or need to compensate for external forces.

Deflection compensation and advanced motion planning tools are also critical in the manufacture of large aerospace parts where large tools are mounted to the robot, the robot tooling is required to contact the part, and where offline programming is critical. Some additional applications that have benefited from high-accuracy robots are aerospace engine components manufacturing and airframe painting/depainting.

Several factors contribute to robot accuracy and must be considered: robot foundation, mounting, and environmental considerations; end of arm tooling; software; maintenance and repair; and validation.

Foundation, Mounting, and Environmental Considerations

One of the first considerations when designing a highly accurate system is how the robot is mounted or anchored to the floor; this is a critical consideration and cannot be underestimated. If the robot is not securely anchored to the floor, the robot can pitch or yaw in different directions or shift from the inertia of the robot’s movement and payload. The amount of effort spent on making the robot accurate will not matter if the very foundation to which it is mounted is not secure. This includes the floor thickness, how it is anchored, if it is isolated, the number of anchors, floor flatness, and the thickness of the base plate. The riser construction needs to be as ridged as possible to ensure only minor deflections. Some options to consider are to create the riser as a hollow tube and fill the tube with concrete. Be sure to use ample gussets and the proper thickness for the base plates and robot mounting plate.

The effects of thermal expansion should be considered for the applications that require the highest level of accuracy. There are two factors that directly influence thermal expansion. The first is thermal expansion due to ambient temperature — the temperature in the atmosphere surrounding the robot. The second is thermal expansion due to self-heating; reducers and motors will heat up during robot operation causing the robot castings to expand, which could alter robot accuracy. The ambient temperature cannot control thermal expansion due to self-heating. There is one factor that will directly affect the magnitude of the thermal expansion in both cases. The coefficient of thermal expansion of the material used to construct the robot is a major factor when allowing for thermal expansion. The coefficient of linear thermal expansion of aluminum is 23, where the coefficient of linear thermal expansion of steel is 11 to 13, depending on composition. This means that the effects of thermal expansion for a robot constructed of aluminum will be approximately double that of a robot constructed out of steel.

End of Arm Tooling (EOAT)

The next consideration is the design of the End of Arm Tooling (EOAT), payload, and dress out. The mass properties of the payload must be accurately defined. This is also true for any valve packages that are mounted on the robot’s arm. The better the payload is defined, the better the robot software will be able to correct the robot’s position due to gravitational effects. For calibration proposes, it is best to do the calibration with the actual EOAT and dress out.

When designing the EOAT, considerable attention needs to be paid to minimize EOAT deflection. The design needs to eliminate the possibility of the EOAT shifting or moving either from the robot’s own inertia, or from any process forces that will transfer back to the robot. This is done by designing an EOAT that uses dowel pins to eliminate the possibility of the EOAT moving or shifting. Also, the effects of the dress out cannot be ignored. Large and improperly designed dress outs affect robot accuracy by pulling on the EOAT.

Software Settings

Software defines the relationship between the part or fixture and the robot base frame. It is important that it is established accurately to account for any variance in the location of either the part or the fixture. This variance will be important when trying to maintain an accurate system. Depending on the application, several software packages can be used to set up this relationship accurately.

Maintenance and Repair

A maintenance and recovery plan is an essential part of an accurate robotic system. Just like there are three components that contribute to the overall accuracy of a robotic system, the maintenance and recovery plan must address all three of these components.

Validation

Not all options and not all applications will require accuracy validation, but accuracy validation is achieved using metrology such as a laser tracker to measure the robot’s actual position, and compare it to the commanded position. A laser tracker, tooling ball reflectors, and a knowledgeable user will be required for most validations.

This article was contributed by FANUC Robotics America Corp., Rochester Hills, MI. For more information, Click Here .