Robot machining draws closer

6 min read

Steed Webzell attended the UK leg of a series of European seminars centred on machining with robots, discovering that this technology is now tantalisingly close

As the lead COMET project (see box, below) co-ordinator, Delcam (0121 683 1000) hosted the UK leg of a series of robot machining seminars that have been taking place across Europe. The aim of the event, staged at the Liverpool headquarters of robot system integrator CNC Robotics (0151 523 8009), was to help disseminate COMET results and findings as the project draws to a close. Since September 2010, the EU-funded COMET project has aimed to overcome the drawbacks of industrial robots by developing innovative robot machining systems that are flexible, reliable and predictable, and which offer significant cost savings, in comparison to machine tools. For example, according to Jan Willem Gunnick, manager R&D projects Europe at Delcam, a machining facility filling 3 m³ will cost in the region of €70,000, using a robot-based system, or between €220,000 and £335,000, using a machine tool. On the flip side, the drawbacks of machining with robots are well documented. Robots lack absolute positioning accuracy, are unable to reject disturbances, in terms of process forces, and are also minus reliable programming and simulation tools. While these critical limitations have, to date, prevented the wider adoption of robots in commonplace machining applications, things are changing. Machining with robots is no longer restricted to projects such as prototypes, sand moulds, resin sculptures and polishing/linishing. Instead, precision machining with robots is now a reality. Sure enough, the overlap with CNC machine tools is still relatively narrow, but this is certain to broaden with time. Able to illustrate some real-life precision machining examples was Francesco Leali of COMET partner SIR SpA, a robot integrator based in Italy. According to Mr Leali, among recent landmark cutting trials was the machining of an aluminium wing leading-edge bracket at TEKS in the UK. The robot machining cell at TEKS comprises an ABB IRB 2400-L robot, equipped with a 1.7 kW spindle mounted on a robot end-effector. In terms of accuracy, machining the wing bracket delivered typical deviations of less than 0.1 mm from nominal values. In Greece, system integrator Gizelis Robotics and end user Bazigos SA have been busy machining a bending machine component made from carbon tool steel. Using a Yaskawa Motoman ES165D robot, fitted with a spindle in the robot end-effector, the part was fixed on a work table. Linear dimensions of external shape and lateral slots again hit nominal values, with deviations seen of less than 0.1 mm. Mr Leali also commented that an aluminium automotive brake caliper had seen the achievement of "rigorous geometric tolerances" across 15 different machining operations, including drilling, face milling and chamfering. Typical face milling data recorded are said to have included 16,000 rpm spindle speed and 17 mm/s feed rate. Of course, the use of software to improve robot machining accuracy is a big theme of the COMET project. With this in mind, Delcam R&D project manager Johnny Van der Zwaag wanted to explain a little about the company's offline compensation software, which is currently in development. To compensate for anything, it is first necessary to identify the errors that need compensating. In a robot machining cell, these comprise two main areas: joint-based errors and kinematic deviations. Addressing joint-based errors, the Delcam offline compensation software has the ability to model slackness (backlash), as well as account for factors such as robot compliance and friction. To obtain the best results, process forces must be modelled, including force in the feed direction, force orthogonal to the feed direction and passive force in the Z direction. COMET research has resulted in a simple-to-use method to determine joint parameters. Unfortunately, the technology is patent-pending, so Mr Van der Zwaag was reluctant to explain further. With regard to measuring kinematics, the Delcam offline software is able to compensate for small differences in the length and orientation of robot links, for example. In terms of accuracy, Mr Van der Zwaag stated that the best results are achieved using a combination of good milling and robot strategies. Backlash is inherent in robots, so the simple answer is to avoid it. Sure, moving all robot axes simultaneously looks fantastic, but, if certain axes can be locked down, then all the better. From recent trials on resin, aluminium and steel workpieces, Mr Van der Zwaag was able to confirm machining accuracy results that were as good as ±0.1 mm. This is pretty impressive, particularly as the following speaker, Roger Holden, business development manager at Nikon Metrology (+32 16 740100), underlined that an uncompensated machining robot would typically deliver an error of up to 1 mm. Nikon offers a complementary compensation technology – an integrated metrology system named Adaptive Robot Control, which, although available in scalable solutions, is based on the Metris K600 optimal CMM. The K600 essentially deploys three linear CCD cameras that localise infrared LEDs incorporated into the carbon fibre housing of a touch probe or laser scanner. As a result, the system creates a 6D reference frame of the machining envelope. The set-up then measures the position of the tool in real time, compares it to the CAD nominal and guides it to the right place. Using high frequency streaming up to 1,000 Hz, instantaneous signals can be sent to the servos. Mr Holden said that, although this type of technology has been used in assembly robots for the past eight years, it is only now transferring to the machining arena. He indicated a recent application at an aerospace manufacturer, involving a 21 kg payload Fanuc robot used to perform single-shot drilling and countersinking operations through 6 mm composite. Another component of the COMET accuracy compensation solution comes in the form of the High Dynamic Compensation Mechanism (HDCM) from Fraunhofer IPA, Germany. This high frequency mechanical actuator works by adjusting the cutting tool spindle (mounted in the HDCM) in 3D to dampen process oscillations. A total of 0.5 mm compensation/travel is possible, with the aim of removing the final pieces of error at a high frequency level to deliver an ultimate machining accuracy, using robots of 0.05 mm or better. However, application of the HDCM is intended only for situations where the robot holds the workpiece, manipulating this against a spindle. Turning to robot programming and this subject was delivered by Franck Messmer, senior analyst at Delcam France. There are many ways to program a robot for machining, such as point-by-point using a teach pendant, the robot's own simulation software or offline simulation of CAM data. However, all of these established approaches have their weaknesses, whether speed, accuracy or a lack of toolpath strategies. What conventional CAM systems lack is knowledge about the kinematics and dynamics of the robot system. SINGLE SOFTWARE SOLUTION According to Mr Messmer, by far the best option is to deploy a single software solution, such as Delcam's PowerMill with Robot Interface (see box, below), that can accommodate both CAM and simulation tasks. Making the transition from CAM to robot demands a list of operations that begins with toolpath calculation and loading of the robot. This is followed by defining the part position, as well as the robot starting configuration. It is then possible to simulate the toolpath and analyse the simulation. Is it optimised and safe? No, then go back and recalculate/edit the toolpath. When the simulation is correct, the tool and component origins can be defined. Ultimately, the program is generated and run on the robot. Among the robots currently supported by PowerMill are ABB (01908 350348), Fanuc (02476 639669), Kuka (0121 505 9970), Yaskawa Motoman (01295 272755) and Stäubli (01952 671917). So what of the wider COMET system? Well, Panagiotis Lagios, project manager at COMET consortium partner Gizelis Robotics, pointed out some of the critical factors that determine robot machining cell integration. For instance, material hardness will typically establish spindle power requirements, while external dimensions will need checking in line with the robot's working range and component weight will help assess payload. Configuration is another issue. Small parts can be gripped by the robot and addressed to a separate, stand-alone spindle. For larger parts, it is more usual for the robot to accommodate the spindle and approach a stand-alone workholding table. The need for an automatic tool change (ATC) unit also needs pondering, as does additional equipment, such as rotary table, lubrication system, robot part grippers, extra spindles (perhaps for roughing and finishing), tool check system, additional robot memory for large programs, PC connection for downloading programs, linear track to expand the robot's working envelope and a gantry to help accommodate exceptionally large parts. In essence, the message from Mr Lagios was clear: "Make your life easy – appoint a robot system integrator. Such companies can advise on everything from the robot and CADCAM system, through to the tooling and perimeter guarding." Box item 1 COMET project in brief The €8 million COMET project aims to overcome the challenges facing European manufacturing industries by developing innovative machining systems that are flexible, reliable and predictable, with an average of 30% cost efficiency savings in comparison to machine tools. The COMET platform consists of a combination of new technologies that increase robot accuracy and enhance the span of the industrial application. of robots: newly developed path programming and simulation software; modelling of the kinematic and dynamic behaviour of the robot; an optical tracking system to feed back positioning corrections to the controller; and an active compensation mechanism that further improves the accuracy beyond the structural capability of the robot system. COMET is led by Delcam as co-ordinator, but has 14 partners drawn from eight countries. These partners include technology suppliers, systems integrators, universities and end users, with varying areas of expertise. Box item 2 Robot interface Delcam's commercial product name for the programming and simulation environment (PSIR) being used in the COMET project is the PowerMill Robot Interface, which is now fully embedded inside the company's PowerMill CAM system as a plug-in. The product's core functionality consists of three main steps: programming, simulation (including analysis) and the creation of robot programs. Following programming, it is now possible to simulate the complete machining operation, controlling the robot's movements through different variables, such as axis limits, axis priorities and workplane constraints. Various aspects within the configuration of the robot cell, such as axis limits, tool constraints and home position, can be defined and the simulation of the robot completed within those constraints. Once the results of the simulation have been reviewed – and modified if necessary – the program can be output in the appropriate robot native language, eliminating the need for third-party translation software. Acceleration, smoothing values and other robot-specific parameters can be defined as part of the output. First published in Machinery, June 2013