Self-programming Paint Booth for Batch Size One

Project SelfPaint

Automated painting of individual pieces

Within the Fraunhofer research project, a painting cell was developed over a period of three years (2016-2019) that independently detects, measures and paints objects.

The fully automated painting cell makes it possible to process any objects. "SelfPaint" covers the entire painting process and is therefore suitable for numerous industries and fields of application. The idea and the project's successes to date show how future paint applications can be made more flexible and at the same time more resource-saving. The modular structure allows the use of the developed software and technologies beyond the self-programming painting cell and can also be integrated into existing painting systems.

The project has been completed. At the end of the project, a demonstrator was set up in the Fraunhofer IPA painting pilot plant. We will present the project at the PaintExpo in April 2020 in Karlsruhe.  

Modules of the Painting Cell and Steps for Optimal Painting

The architecture of the self-programming painting cell/process is divided into the following modules:

  1. 3D Object Acquisition and Creation of Virtual Twins: 
    In 3D object acquisition, the Fraunhofer ITWM uses the Kinect fusion algorithm to digitally capture a point cloud of the object to be painted. Using adapted algorithms, we then perform a position comparison between the CAD drawing of the component and the detected point cloud. The recognition of the known CAD model in the detected point cloud enables the position of the component in the painting cell to be determined using cost-effective hardware. 

  2. Spray Simulation:
    Based on experimentally determined paint parameters - particle size, density, etc. - the spraying behavior of the used spray head (»near-bell simulation«) can be simulated with the software developed by Fraunhofer IPA. Resist quantity, spray head rotation speed and air flow play a decisive role and are covered by the developed models to be helpful as a starting point for the subsequent »laydown« simulation.

  3. »Laydown« Simulation: 
    Based on the results of the near-bell simulation, the Fraunhofer-Chalmers simulation tools are used to calculate the spraying behavior at greater distances. Here, electrostatic effects as well as air flow and the Magnus effect are taken into account, so that this simulation must be carried out online.

    In order to create paint lines with high precision in the shortest possible time, Fraunhofer-Chalmers FCC has combined simulation methods. The physics-based simulation is divided into a simulation near the high rotation atomizer (near-bell) and a dynamic layer thickness simulation (laydown) (see picture). The path optimization is done by machine learning.

  4. Computer-Aided Paint Optimization: 
    From the simulation data, the system determines the best possible robot path for the painting process. In order to achieve a homogeneous paint layer, precise knowledge of the spray mist behaviour and optimised painting paths are necessary. The projection method uses a static, simulated spray pattern based on the principle of a flashlight to generate a layer thickness distribution. With the information from the path optimization, the component is automatically painted by a robot.

  5. Final Inspection With Non-Contact Coating Thickness Measurement:
    Is the thickness of the paint layer as desired? Terahertz coating thickness measurement technology is used for quality control to reliably determine the thickness of both wet and dried paint. The system developed by the Fraunhofer ITWM can measure the object by means of a robot with up to 50 measurements per second in order to detect even curved components.
Physics based simulation divided into near-bell and laydown simulation.
© Fraunhofer ITWM
The physics-based simulation is divided into a simulation near the high rotation atomizer (near-bell) and a dynamic layer thickness simulation (laydown).
3D object detection of a component.
© Fraunhofer ITWM
3D object detection of a component

Three-Dimensional Object and Position Detection

In order for the painting cell to know the position of the object for simulation and painting, it must be recorded in three dimensions. The self-programming painting cell uses 3D sensors that were originally developed to control video games - a global mass market. The accuracy of the position detection is more exact than the deviations of common components from the CAD data and is therefore within the manufacturing and positioning tolerances. This high resolution is made possible by tailor-made algorithms for data processing.

Schematic representation of the 3D scanning process
Schematic representation: 3D scanning process, in this example for a chair.
Schematic representation
Schematic representation: Painting the component, in this example, a chair.
Schematic representation
Schematic representation: Quality checked via terahertz technology.

Terahertz Technology for Quality Control

In the final process step of automated painting, the quality is checked: Is the thickness of the coating as desired? For this quality-control we use terahertz waves. With this technology developed by us wet and coloured lacquers can be measured without contact. The quality of the paint layers can already be checked during or after the painting process. Substrates - the basis of the coating layers - do not have to be metallic, but can also consist of other materials.

Partners and Responsibilities

Layer Thickness Measurement with Terahertz Waves
© Fraunhofer ITWM
Layer Thickness Measurement with Terahertz Waves

Video: Layer Thickness Measurement with Terahertz Waves

Privacy warning

With the click on the play button an external video from is loaded and started. Your data is possible transferred and stored to third party. Do not start the video if you disagree. Find more about the youtube privacy statement under the following link:

Video: Detailed Summary of the Selfpaint Project