Traditional industrial automation uses workcells designed to meet requirements that are defined by physical constraints. Tasks are laid out so that robots can execute identical tasks on identical parts. If the robot finds those parts in the same location and orientation each time, it performs the task reliably and efficiently.
These workcells rely on relatively unintelligent robots doing the same task millions of times. When designed and fielded properly, this style of workcell offers efficient and capable execution with quick cycle times at high volumes. But these workcells don’t tolerate variation, making them rigid and fragile. If part locations, orientations, or even the parts themselves fall out of expected tolerances, the workcell begins to fail regularly.
As product updates lead to changes in process requirements, adapting these workcells to execute the same task using the latest parts can often cost more than simply ripping them out and replacing them with yet another bespoke automation design.
SymbioDCS is a software platform that provides advanced robotic control through visual and force-torque feedback creating flexible automation that can adapt to variation in the manufacturing environment in real time. It harnesses the power of modern computing practices—networked devices, servers, machine learning, and widely-used programming languages—to enable industrial robots to gather information about their environment and then quickly act on that data to solve tasks.
At a large domestic automotive assembly plant – a major OEM designed and implemented a workcell on their factory floor to fasten two sheet metal parts into subassemblies used as electronics casings for various car models. The robot tightened bolts that were manually inserted by human operators. By automatically recognizing the model of the part placed on the jig, the robot went to pre-programmed locations, where it expected to find bolts. Upon introduction, the automation failed to meet reliability and quality requirements as the sheet metal parts tended to shift and bend in the fixtures designed to maintain the parts in exactly the same place. Additionally, with time the part orientation changed so significantly that the robot required re-programming and ultimately forced a return to manual operation.
Originally, the automotive OEM chose this task for automation because it was a monotonous task with poor ergonomics. The task led to high employee turnover, and the manual operation never fully met cycle time requirements. Additionally, the workcell’s prime location on the factory floor meant that any reduction in footprint would release valuable floor space for use by other applications.
In trying to automate the task, the OEM encountered “more bad days than good.” Part variability led to frequent failures, and each new vehicle model required time-consuming and error-prone programming.
Symbio adapted its technology to work with the existing equipment comprising Universal Robots model UR10/CB3 with Acradyne nutrunner, installed on the end effector and Allen-Bradley GuardLogix Controller (PLC). Symbio introduced an ATI Axia 80 force/torque sensor (FTS), installed between end effector and nutrunner together with a Dell Embedded Box PC 5000.
SymbioDCS was fully integrated with the PLC allowing the OEM to continue to use their existing hardware and sensors without the need for extensive reprogramming. The cell did not need to be recertified for safety. Additionally, SymbioDCS was fully integrated with the existing human machine interface (HMI) system. From the perspective of factory-floor operators, the SymbioDCS-powered solution behaves exactly the same as the failed legacy system this no retraining was required, and the workcell’s integration into the factory server was unchanged.
Because of the frequent need to accommodate new models, SymbioDCS uses the bolt positions for a vehicle model as parameters. Symbio provides an offline GUI enabling manufacturing engineers to create a parameter set for new models in under 30 minutes and put that profile into operation immediately. With each cycle, SymbioDCS receives a vehicle model identification code from the PLC and automatically chooses the correct profile to use for that set of parts.
To overcome variations in individual parts, SymbioDCS uses real-time feedback from the FTS to find the surface of the bolt. It then performs a spiral search to center the nutrunner on the bolt and tighten it. SymbioDCS collects and stores process data. A browser-based dashboard enables factory personnel to track key metrics and to tune the process when necessary.
The Symbio solution has been shown to improve operations across a number of parameters, including quality, cycle time, downtime, cost and flexibility. The upgraded system was shown to reduce cycle time by over 30%, increase reliability to 99% from less than 60%, reduce labor by 1 head per shift in comparison to manual operation and decrease footprint by removing the need for a re-work station. Since retrofitting, the OEM has experienced the benefit of the more flexible automation solution through increases in reliability and reductions in cycle times and manual labor.
For more information: www.symb.io