DigiFab4KMU – Digital Factory for SMEs: Utilization of Point Clouds

There are usage scenarios for digital factory models in which manual post-modeling in CAD tools is not necessary. Instead, the point clouds of the factory objects can be attributed directly, eliminating the most time-consuming step 3.

In this context, a usage scenario describes the way in which a user performs a specific task with the help of a system. A usage scenario is implemented in reality through its application in an actual use case.

References:

[1] Bermpohl, F.; Schäfer, S. F.; Hohmann, A. et al.: Short-cycled factory planning – motivation and existing challenges Procedia CIRP (130), p. 1708-1713
[2] Geissbauer, R.; Bruns, M.; Wunderlin, J.: PwC Digital Factory Transformation Survey 2022, 2022
[3] Feldmann, K.; Reinhart, G.: Simulationsbasierte Planungssysteme für Organisation und Produktion. Modellaufbau, Simulationsexperimente, Einsatzbeispiele. Berlin, Heidelberg, s.l.: Springer Berlin Heidelberg 2000
[4] NavVis GmbH: NavVis Digital Factory Survey 2021. A time of change for manufacturing industries, Munich, 2021
[5] Klinc, R.; Jotanović, U.; Kregar, K.: POINT CLOUDS FOR USE IN BUILDING INFORMATION MODELS (BIM). Geodetski vestnik 65 (2021) 4, p. 594–613

In order to make the potential of digital factory models accessible to SMEs, »DigiFab4KMU« set out to research a method for the direct utilization of point clouds. To this end, usage scenarios for digital factory models were identified that can be implemented without manually remodeled factory objects. The usage scenarios resulted in functional requirements for the implementation of the process steps of point cloud segmentation and attribution, as well as visualization of the overall model in the inventory data viewer.

The implementation of a process step was subsequently referred to as a solution component. Various usage scenarios are being tested with the application partners. Existing solution approaches were adapted for this purpose and any necessary interfaces were developed. The researched method was intended to serve as a decision-making aid for SMEs by helping them select existing solution approaches. It was methodically proven that point clouds can be directly utilized. The prerequisites for the transferability of the method were then examined. Future research and development needs were systematically identified. These sub-aspects formed the basis for the structure of the »DigiFab4KMU« project. As shown in Figure 2, this was divided into a total of three work packages. The partners involved are indicated by logos in the work program.

Based on data collection and analysis, a new method for utilizing point clouds was developed.

The process begins with the selection of a usage scenario. Once all these requirements have been defined, the point cloud can be recorded. At the same time, digital tools for implementing the sub-areas of »segmentation«, »attribution«, and »visualization« must be selected. This method is supplemented by a decision support tool that suggests suitable digital tools or a combination of several of these tools based on the requirements for the three sub-areas. The decision support tool is Excel-based and is shown in Figure 3 as an example.

If the digital tools identified by using the decision support tool are not available in the company, a procurement process may need to be carried out. Subsequently, both the point cloud of the factory area and suitable digital tools selected for the following steps based on the requirements are available.
The point cloud is then segmented according to the specified geometric abstraction level, and alphanumeric data is recorded in the required LoI from the information sources.
The linking and visualization of this data is the final step in the method. All steps and their interrelationships are visualized again in Figure 4.

More planning flexibility for SMEs thanks to point cloud approach

The presented method for the direct utilization of point clouds demonstrates how a robust, digital planning basis can be created without having to invest significant effort in modeling individual factory objects. By combining point cloud data with alphanumeric information, users can implement usage scenarios for digital factory models without having to go through the time-consuming step of post-modeling in CAD tools.

Applying the method to the three SMEs showed that this approach can be used in different production environments. The method worked well in planning-focused scenarios, like redesigning layouts, and in operations-focused scenarios, like displaying and linking operational data at the plant level.

The respective implementations illustrate that different usage scenarios can be implemented by combining existing software solutions. No specially developed system environments are required. This method therefore offers SMEs in particular a practical introduction to the use of point clouds.

Some manual effort remains

However, the applications also revealed the current limitations of the method and the available technology. These include, in particular, the manual effort required for the segmentation and attribution of point clouds. Even though this effort is significantly less than that required for modeling, there is great potential for optimization in the automation of these steps. In addition, it has become apparent that the compatibility of individual software solutions can be an obstacle to the implementation of certain usage scenarios. Nevertheless, in all three cases, added value was generated in terms of data availability, transparency, and planning reliability.

Future outlook: Further automation and simplified integration

For future research, this highlights the need to further automate sub-processes of the method. The use of machine learning methods offers great potential here, especially for the classification of complex factory objects. In addition, the development of standardized interfaces must be pursued in order to seamlessly integrate attributed point clouds into existing planning, operating, and simulation environments. In this way, point clouds from factories could be of greater use than merely serving as a basis for modeling. Extending the method to include immersive visualization technologies, such as virtual or augmented reality applications, could also further improve intuitive use and collaboration between different stakeholders.

Overall, the method developed shows that the expanded use of point clouds is a practical option for digitizing factories. It helps to make the potential of digital factory models accessible to SMEs as well, thereby contributing to the digitization of German small and medium-sized enterprises.