xBerry Case studies Automatic Grinder

Automatic Grinder

Revolutionizing Grinding in Industrial Transformation

Watch the video

In the realm of digital transformation, precision and efficiency are paramount. Few robots exist today that can automate grinding, primarily due to the inherent variability in each weld. Achieving automation that meets the requirements related to precision poses a significant challenge. Our newest project, aptly named Automatic Grinder, aims to tackle the challenges of automating the grinding process for our client, a leading supporter of industrial innovation in Poland.

Scroll pageg icon

Challenge

Our client sought to automate the grinding process by using a robotic arm equipped with a grinder. We knew that the task would not be easy – it is not without reason that there are not many robots in the world that automate the grinding process. However, our aim remained clear: achieve satisfying accuracy and processing speed by leveraging widely available robotic arm solutions on the market.

 

The mission called for an innovative application that could enable precise control of the robot, gather essential information, and simulate operations before executing them in real-time.

Goals

xBerry set out to create a cutting-edge application that would meet the client's expectations.

The Automatic Grinder application was precisely designed, with a focus on extending publicly available robotic arms through the integration of additional sensors and a laser scanner. This innovative approach enabled the creation of an application capable of planning and controlling the grinding trajectory.

Our primary goal was to achieve an accuracy of 0.2 mm. The main focus remained on delivering precision and efficiency, revolutionizing the industrial grinding landscape.

Solution

  • We based our solution on the Robot Operating System (ROS) and MoveIt motion planning framework, which provided numerous tools, libraries, and capabilities to develop robotic applications and integrate them in production environments.

  • We added 3d laser scanner and numerous sensor to enable environment perception and by extent - automatic trajectory generation with collision anticipation.

  • We used Open3D, OpenCV and PyVista for rapid development of 3D laser sensor data processing pipeline.

  • To automatically set and tune machining parameters, we use machine learning algorithms.

Results

Imagine a flat plate awaiting its transformation. Through an intuitive graphical interface, operators input the appropriate parameters, such as the abrasive wheel type or the type of the material welded. The detail is then scanned and compared to the 3d reference CAD model. The Automatic Grinder generates suggested trajectory, ensuring utmost precision. To ensure absolute safety, the proposed trajectory is then tested in a simulation environment.

Automatic Grinder’s exceptional capabilities enable the automation of intricate grinding tasks, maintaining an astonishing accuracy level of up to 0.2 mm. This precision not only saves precious time but also yields substantial financial benefits for our esteemed client.

Our solution goes beyond mere automation, offering a comprehensive solution for data collection and analysis. Operators can effortlessly gather vital grinding data, including depth, speed, and stock removal, to meticulously assess the effects of the process as well as monitor status of all robots components. By examining the detail before and after grinding, our client gains invaluable insights, assessing whether the results fall within tolerance limits and identifying potential areas for enhancement.What’s more, the collected data will be used to implement machine learning, which can then improve the grinding parameters based on this data and ultimately provide better results with each subsequent detail

The versatility of the xBerry solution knows no bounds.The Automatic Grinder application has the potential to be seamlessly integrated with a wide range of industrial robots, allowing for adaptability in various manufacturing environments. However, it’s important to note that adapting this solution would require a separate project.

Tech Stack

Python
ROS
MoveIt
Open3D
OpenCV
PyVista
Docker
Sklearn

Planning a digital project?

Contact us Arrow icon