Auburn Hills, Mich.
Few images warm the hearts of companies as much as that of a robot at work. Tireless, accurate, and uncomplaining, robots have been a fixture in manufacturing for almost 30 years. A downside is high programming costs. That's because most robots work as individuals instead of synchronized in teams. This is a major roadblock to the growth of robotic manufacturing.
Part of the problem is an antiquated software-programming method. In typical manufacturing facilities, each robot runs proprietary software and production engineers must program them one at a time. For facilities with, say, 300 robots, that's a lot of downtime. Engineers thus try to program at a frantic pace, with little chance to fine-tune code. And because engineers usually don't have an overall view of production lines and entire facilities, it's difficult to detect collisions or other common robot and work-cell incidences.
Digital-manufacturing software is one component to solving the problem. For starters, the software lets engineers program offline before using code on the shop floor. The software also provides simulation tools that let users coordinate operations between robots, work cells, and production lines. Ideally, the software is part of a PLM system that encompasses design, analysis, and manufacturing processes. Here, robotic programs deliver fast, high-quality results on a large scale.
In this scenario, production engineers first make 3D models of manufacturing facilities and then use the models to lay out individual machines, work cells, and production lines. The models are not simply animated illustrations — they are exact representations of the facility, in real-life scale and proportion, created from equipment geometries and the facility's physical dimensions. Virtual 3D parts and assemblies that move along digital production lines are created from original design data.
Everything in 3D simulation happens exactly as it would in the physical world. This gives engineers the chance to see how production lines and whole facilities will work before they commit to physical construction. Users can manipulate multiple robots and work cells at one time to avoid collisions and also anticipate safety issues such as a robot moving too far into human work areas.
A good example to illustrate differences between current programming methods and those based on digital-manufacturing software comes from a common, heavily automated, automotiveindustry process — painting bumpers. Engineers program robots to paint a precise number and length of strokes to cover the entire surface, yet must avoid overpainting and waste. In an ideal scenario, manufacturers would smoothly shift capacity between painting booths, according to demand, merely by modifying the robotic programming.
Instead, engineers must work for hours before programming begins. They first collect information such as offset numbers from controllers and current productionline layouts. Next, they use this information along with CAD partshape data to build virtual models of robotic-production procedures. In this scenario, engineers program in complex, proprietary controller-software languages.
After loading data into the robot controllers, it's up to operations engineers to handle problems that arise. These might include, for instance, a welding arm not fitting a certain hole. The trouble is, production engineers are seldom aware of modifications at the operations level. Thus, when reprogramming painting robots, production engineers often startfrom inaccurate base-lines.
In contrast, digital manufacturing software such as Delmia (Digital Enterprise Lean Manufacturing Interactive Application) from Dassault Systemes, used with PLM, eliminates or automates painting tasks. Parts information is already in the PLM system, which unifies design and manufacturing-planning processes. Digital manufacturing software lets engineers at offline workstations program a robot just as if they were sitting at the robot's control console. Simulation models show if robots collide, or damage parts by handing them off too quickly. Lastly the 3D facility model acts as a single, authoritative information repository for production and operations engineers. For instance, the model communicates modifications back to production engineers so there are accurate baselines for future programming projects.