7.6 million people work in the manufacturing industry in Germany. Despite the importance of this sector, the digitalisation and networking of companies is only progressing slowly. Reasons for the low level of digitalisation in companies include a lack of skilled workers as well as a lack of technologies and strategies.
This is where the Factory-X project comes in. One of the main objectives of the project is to develop intelligent machines and systems and integrate them into a standardised data ecosystem in order to ensure cross-company networking in the future. As part of this project, the Institute of Production Engineering and Machine Tools (IFW) at Leibniz University Hannover is developing software applications for intelligent machines and systems in order to drive forward automation in production planning. The focus is on automated order processing and dynamic capacity planning.
This work will increase the efficiency of planning processes and achieve resilience to disruptions in the process chain. This leads directly to an increase in the competitiveness of German companies. In addition to competitiveness, proof of the CO2 footprint will also play a major role for German companies in the future. The IFW’s work can reveal energy consumption and potential savings. The IFW is also developing energy-optimised production planning.
Automated order processing …
Automated order processing begins as soon as a production order is received and concludes with the creation of a machine-readable code, an NC code. As soon as a production order is received, work steps and the required tools and machines are derived from the order. As part of the Factory-X project, the IFW is developing a method for automating this process. This is referred to below as CAM automation.
Scientists at the IFW are developing a programming interface that makes it possible to connect an application to the design programme. This application uses artificial intelligence (AI) such as decision trees to assign suitable work processes, machines and tools to the component. To do this, features are extracted from the 3D model of the order. Features describe product characteristics such as the workpiece geometry and production characteristics, for example tolerances. This information can be used to calculate the costs of the order. This makes it possible to analyse the profitability of the order even before production begins. In the future, the application will also automatically generate an NC code.
For some operations, such as the production of gears or polygons, the programming effort required to create the NC code is very time-consuming. This effort can be reduced by using technology cycles. Technology cycles already contain partial work steps and requirements for certain operations such as gear skiving. As part of the project, the IFW is developing an AI assistant (see Figure 2), which selects a suitable technology cycle based on the features already described. The AI assistant uses methods such as a decision tree. In addition to selecting the technology cycle, a profitability analysis is also carried out. The AI assistant evaluates whether production with the technology cycle leads to faster and more efficient processing than without the use of a technology cycle. A possible cost reduction is then displayed.
… and dynamic capacity planning
The knowledge gained from CAM automation can also be used to optimise capacity planning. The scheduling of rush orders is particularly important here, as this type of order often appears to be particularly lucrative. However, unexpected costs often arise after acceptance of these orders due to increased set-up costs.
The IFW is developing an algorithm to evaluate rush orders before they are accepted and to determine whether sufficient capacity is available for the rush order. If, for example, the costs for retooling exceed the profit generated by the rush order, a rejection of the rush order is suggested. This information can then be fed back to a matching platform, which assigns the orders to a suitable manufacturing service provider with free capacity (see Figure 3).
The IFW expects the AI assistants for automated order processing and dynamic capacity planning to increase the efficiency and performance of production planning. In addition, the IFW is developing applications to promote sustainability in production.
Optimising energy consumption in production …
Energy efficiency is a key factor in promoting sustainable production. Optimising energy consumption is not only a cost factor, but also a key aspect of sustainability.
A large proportion of energy consumption in production is caused by machine tools. They account for around 4 % of the global demand for electrical energy. The IFW is therefore developing and researching an energy digital twin (eDT-X), which will be used to transparently display and model the energy requirements of machine tools. With its help, potential energy savings can be identified and the expected energy requirements in production can be forecast.
The eDT-X is technically implemented using an asset administration shell (AAS) or an administration shell server, which is hosted on an edge device on the machine (see Figure 4). The AAS is a digital image of a physical asset in accordance with the Industry 4.0 concept and includes all relevant information and functions of the respective asset. It acts as a standardised interface for communication and interaction with other systems, enabling efficient and interoperable networking of machines and systems.
… with the help of a digital twin
In future, the basic structure of the digital twin will be generated and provided by the machine manufacturer. This is done as automatically as possible from CAE data (computer-aided engineering) of the installed components. The framework comprises individual administration shells of the installed machine components and an administration shell of the machine tool. The asset administration shells contain various sub-models in which different data or information is provided.
A distinction is primarily made here between static and dynamic information. Static information includes, for example, metadata that describes the machine or the components. This also includes existing models that can represent the energy states. This data is provided by the machine and component manufacturers. The sub-models with dynamic data, which include time series on the current power consumption, for example, are also provided as digital components by the machine manufacturer in the asset administration shell. However, these are only filled with data during the utilisation phase of the machine at the factory operator.
During this utilisation phase, the eDT-X will initially fulfil two main functions. These include component-resolved energy monitoring and an energy demand forecast of the workpieces to be produced on the machine. In order to be able to save energy effectively, the current energy consumption and its distribution across the individual components of the machine tool must first be determined. Current solutions from the industry mainly focus on the total energy consumption, so that energy saving potentials of individual components can sometimes not be identified. A holistic energy monitoring system is therefore being developed for machine tools at component level.
Coordinating production with expected electricity prices
In addition, forecasting models are being created that determine the expected energy consumption for individual workpieces. This means that statements can already be made about energy consumption when planning production orders and planning can be optimised in terms of energy consumption based on the forecasts.
In future, production sequences can be created in line with the expected electricity prices or load peaks can be avoided. The development of these demand forecasts will be iterative and will become more detailed as the development time increases. Initially, the forecast is based on an extrapolation of average performance values for various operating states. The further developed method analyses the existing NC code and determines the temporal proportions of other machine states. In addition to the main processes, such as standby or machining, secondary processes, such as tool changes, are also analysed. Here too, an extrapolation is made on the basis of reference power measurements for the respective machine states. In the final stage of development, the energy requirement is forecast using data-based models with the aid of a process simulation in order to determine energy-relevant intervention variables.
The next steps in the development of the energetic digital twin are focussed on an initial prototype implementation of an eDT-X on a DMG MORI turning/milling machine in the IFW test field. For this purpose, the administration shells are first created manually and provided on an AASX server hosted on an edge device via the BaSyx middleware. In addition, OPC UA clients run on the edge device, which access process and sensor data from the machine tool and store this data in the asset administration shell.
The goal: more sustainable production
To summarise, the energy digital twin has great potential to promote more sustainable production. In addition to increasing the transparency of energy consumption, it can also strengthen the focus on energy efficiency as a target variable in production planning.
