Managing Quality Control with Autonomous Robots
Posted on Wednesday, May 12, 2021
By Robert Ravensbergen
Robotic coating has been a well-understood component of automotive paint and other mass manufacturing processes, but specific to powder coating and as applied to a variety of circumstances, robots have seen rare adoption in the space.
Why haven't we seen mass adoption of robots in the powder coating industry? A few reasons come to mind: robot integration is a time-consuming and highly determined process. It requires specific programming, jigging, and involves coordination constraints that make it difficult to apply in scenarios where powder is most often used.
At the same time, powder is a “forgiving” medium when compared to other coating materials. Powder coaters have in many circumstances “gotten away with” less precision and refinement in application due to the electrostatic nature of the process, but many more have still suffered from overcoating, rework, angular limitations, and a variety of productivity issues that could conceivably be rectified through robotics.
Many of these challenges can finally be addressed with autonomous robots. Autonomous robots can use 3D vision and other applicable sensing mechanisms to automate the programming process in real time, making a variety of robotic production scenarios far more cost-effective. They can generate highly coherent strategies for a variety of part shapes, save on significant skilled labor risks, and leverage articulated arm capability in order to further work “around” many of the limitations still present in existing powder automation systems.
While it may sound too good to be true, there are of course still some limitations to autonomy: certain 3D reconstruction scenarios, cavity coating, and full-lifecycle payback for smaller shops all need to be taken into consideration.
From Traditional Programming to Autonomy
An autonomous robot can be defined as any automation system that is capable of performing tasks in the world by themselves, without explicit human control. For example, while self-driving cars can be considered a form of autonomous robotics, the technology is already present and more widespread than one may realize. Robot autonomy is more accessible the more constrained or specialized its environment, tasks, or objectives become, and given the value added to industrial robots, the ability to allow a robot to effectively “program itself ” is critical to bringing new forms of automation to the floor.
Robot programming has been a significant barrier in the growth of robotic installations. This isn’t simply the case because it is time consuming, but also because of the jigging requirements and the need to generate a precise program for each part. This has been particularly challenging for high- mix manufacturers. High-mix manufacturing makes up approximately 80% of industrial firms, and can be defined as processing more than 100 SKUs per year or working with batches under 1000 pieces.
From this high-level view, robotics is fundamentally unusable if programming and jigging is required. Too much changeover and too many possible sources of imprecision (particularly in materials handling) will only explode the waste and rework experienced here. Autonomous robots change this by introducing principles from automatic machine control to a broader set of robotic applications. In powder coating, these principles include 3D vision, the ability to generate robot motion and tool path strategies from primitive know-how of the task at hand, and the qualification of whatever program an autonomous robot generates with the particular limitations or process constraints known to powder coaters (like Faraday caging).
It’s important to note that, up to this point, robotics has not simply been a non-entity when it comes to the application of electrostatic powder. Nordson worked with OBUK to create an ABB-based robotic coating cell for large architectural doors. This program achieved high efficiency and consistency in coating and came with an HMI (Human- Machine Interface) that made it possible to define the coordinates and constraints of any door being processed. This program was highly effective, and the ability to add greater autonomy to this kind of existing expertise will allow an even greater variety of more complex and unexpected part shapes to be processed in increasingly cost-effective ways.
Unique Benefits to Powder
For powder coaters, depending on the complexity of the part shapes you’re working on, unexpected costs can add up. These costs include: 1. employee training and turnover; 2. rework due to lack of skill or inconsistent operation; 3. waste due to contamination, excess coating, or inconsistent application; and 4. significant up-front costs to large automation systems.
In these cases, autonomous robots offer a happy medium that allows powder coaters to overcome each particular problem.
For instance, with process know-how and the ability to generate a coherent strategy for each part, an autonomous robot eliminates the need for human workers to conduct direct application themselves. This not only reduces health and safety risks as well as total labor costs, but also reduces turnover and makes working in powder shops less “dirty” and more attractive to the younger, less skilled cohorts in the workforce.
Those same coherent strategies also allow for machine-level consistency at every application point, with consistent gun and pressure control that can be optimized to maximize the transfer efficiency of coating for every type of part. This also allows for extensive optimization, where the minimum possible coating can be applied with that same maximal consistency. While an industrial robot itself can cost between $80,000 and $120,000, total installation costs for autonomous technology can be comparable to reciprocating arm and similar systems, but again permit greater flexibility when it comes to unique or never-before-seen part shapes.
New technology is never perfect and autonomous robots are no exception to this principle. While the value-add they provide is tremendous, there are still particular process constraints to take into consideration.
With respect to powder coating, Faraday caging is always a challenge. The ability of an articulated robot arm to move with six degrees of freedom can allow for more effective Faraday cage avoidance when it comes to complex part shapes and crevasses—or any scenario where an electromagnetic box might be otherwise unavoidable. This, however, is not the exclusion of existing means of Faraday cage avoidance when it comes to mass, voltage, pressure, and more forms of modulation, and the use of a robot arm without considering other factors may not necessarily produce a prefect result. Fortunately, automatic forms of Faraday cage detection may one day be possible.
Secondly, Faraday cage avoidance is essentially dependent on the existing 3D reconstruction capability used in autonomous robotics. This feature allows for the generation of a 3D image of each part being processed. That image is used in a digital twin to generate the robot motion that an autonomous robot carries out.
In certain cases, 3D reconstruction may not provide a good enough shape for path planning. For instance, deep cavities or inner hidden volumes are hard to perceive without restrictive technologies like x-ray scanning. Other limitations may come from an application’s characteristics, but in evaluating the practicality of an application, the degree of difficulty related to flow speed, shape complexity, and tooling constraints must all be considered. Many such scenarios that exist today can be addressed in a cost-effective way with currently available AI and processing power. For some, more fundamental advancement in the component technologies used may be necessary to adequately address the process.
At the same time, this technology is also new and relatively costly, which means smaller shops may not benefit from bringing in new robotic and vision hardware until more processes can be addressed, effectively maximizing the uptime and value derived from new equipment. This could change in the future, as the technology can support any process with the right know-how added, but such capabilities won’t be available within the next two years.
Opening the Door to Autonomous Robotic Coating
Robotic coating has been a new frontier in powder coating for some time. Both robots and powder coating add tremendous value to their own unique stakeholder, but general- purpose technology has not yet emerged to enable the benefits of the two practices to be combined and hence maximized.
With autonomous robotics technology, this can not only change, but be done in a way that is purpose-built to meet the needs of industrial manufacturers and custom coaters alike who are looking to maximize the quality and consistency of their output while minimizing waste, improving working conditions, and eliminating unexpected costs
Robert Ravensbergen is marketing director at Omnirobotic.