Research projects

Project specification

The aim of the joint research project is to develop an innovative, software-controlled, miniaturised and high-precision spraying system with precise and traceable quantity dosing - based on the extrusion system.
The planned development is intended to achieve significant advantages over current technology:

• Reduction of deviations in quantity output from approx. 15% to max. 5%
• Traceability of the quantity output
• Virtual measurement of the viscosity of the operating medium
• Adaptive control behaviour in relation to the viscosity of the operating medium
• Graded spraying of components with linearly variable volume output
• Achievement of high dynamics of the spraying system with a control response in less than 50 ms

Focus of the tasks of the project partners

• Development of an innovative design for the dosing element
• Implementation of the dosing element design in a demonstrator
• Development of the overall design of the spraying system

• Development of an intelligent nozzle control with recalibration
• Development of a viscosity adaptation with virtual sensor

• Development of a drive unit consisting of drive actuators and control electronics
• Validation and calibration of the nozzle functions and model development

Information on our partners

HIESSL Schmiertechnik GmbH (Lubrication Technology)

Fraunhofer IWU - Institut für Werkzeugmaschinen und Umformtechnik (Institute for Machine Tools and Forming Technology)

Project specification

The aim of the project is to develop a new, AI-based modular assembly solution with autonomous workstations in order to offer a greater variety of products without any loss of efficiency in production. This is to be achieved through a simulative mapping of the production chain, the collection of cross-machine production data in real time and an AI-based analysis of the data with an interface to an overarching order-picking system. The control algorithm, which is supported by machine learning, should also enable SMEs with primarily manual production to react flexibly to various changes in the production order. Furthermore, employees will be provided with all relevant information about their production order in a new digital working environment.

Initial situation

Current assembly lines with defined cycle times can no longer meet growing customer demands due to a lack of flexibility and are proving to be less and less economical.

Objectives and solutions

The planned development is a novel, AI-based modular assembly solution with autonomous workstations to achieve a higher variety of products.

Expected results

• Reduction of the processing time of the products by 20%
• Reduction of buffer slots by 50%
• System reliability of 99%

Focus of the tasks of the project partners

• Testing on an assembly line that was specially provided for the project
• Determination of relevant production data for the simulation and development of the control algorithm
• Development of an interface to the ERP system for access to relevant work schedule and order material data (in cooperation with iMes)
• Definition of requirements and target parameters to be met
• Provision of premises for the staff of the Kempten University of Applied Sciences for on-site appointments and tests
• Investments for the realisation of the project

• Development of the neural network that supports the control algorithm for the production line
• Creation of the necessary interfaces to connect the control system with the existing ERP system at Otto Martin Maschinenbau GmbH
• Integration of the control system into the production line
• Development of an intuitive user interface for the staff (so that no special training is required)
• Implementation of automated data exchange for certain processes (for example, control data for quality assurance)

• Creating a simulation of the production line of Otto Martin Maschinenbau GmbH that is as close to reality as possible and an algorithm to optimize it
• Checking the data for plausibility and quality
• Determination of the probability of occurrence for various events that have an influence on the production process and, based on this, creation of a result distribution
• Development of an algorithm that analyses and controls production predictively as well as prescriptively on the basis of synthetically generated target data, including any special runtimes
• Testing the resulting algorithm using further simulations to ensure the robustness of the results

Information on our partners

Otto Martin Maschinenbau GmbH & Co. KG

Hochschule für Angewandte Wissenschaften Kempten

Project specification

The aim of the research project is to develop an "intelligent crystallisation sensor" to provide premium quality assurance in injection moulding and to test it with end users (Exipnos GmbH, Mercedes Benz AG). The "intelligent crystallisation sensor" should be able to minimise distortion-related rejects in the production process and to take countermeasures in the event of quality-relevant process parameter fluctuations. The innovative basis is the combination of an ultra-fast infrared (IR) temperature sensor with an inline hardware-integrated crystallisation model into the "crystallisation sensor".

In a second step, an AI module will be added to the crystallisation sensor in order to create an "intelligent crystallisation sensor". This module uses the data from the crystallisation sensor and the continuously recorded injection moulding process and environmental parameters to generate a quality prediction for each individual component inline and also generates suggestions for adjusting injection moulding parameters.

The basis for this is training the AI system during the usual sampling phase when the injection mould is started. 3D measurement will be required only during this learning period of the AI system; in the production process itself, this is to be replaced by the AI system for quality assurance, which offers significant time and cost advantages.

Focus of the tasks of the project partners

• Development of a crystallisation model suitable for inline use, including the
• Implementation in an algorithm that outputs crystallisation parameters (crystallisation temperature and heat) and meets the requirements of a routinely usable inline crystallisation sensor in terms of runtime and robustness

• Hardware integration of the developed crystallisation model by
• combining it with the own IR temperature sensor to form a "crystallisation sensor", which determines and outputs crystallisation parameters inline and makes them available for further evaluation

• Development of an AI system that is trained in a sampling phase with 3D component measurement and which subsequently provides predictions on component quality and suggestions on process parameter adjustment

• Systematic testing of the "intelligent crystallisation sensor" under real-life conditions, including the
• feedback with the project partners necessary for its optimization

Intended results

• Development of a robust inline algorithm based on a crystallisation model that can provide real-time information on the local crystallisation state of injection moulded plastic components
• Hardware integration of the developed algorithm into a "crystallisation sensor" for the output of crystallisation parameters such as crystallisation temperature and heat in real time
• Development of an AI system which, after a learning phase with 3D measurement, allows statements to be made on component quality - based only on the data from the crystallisation sensor and the accompanying conventional process data
• Combination of "crystallisation sensor" and AI system to form an "intelligent crystallisation sensor" that operates robustly under production conditions and reliably meets the above requirements
• Exploring the possibilities of generating useful suggestions for adjusting injection moulding process parameters based on the data from the crystallisation sensor and the accompanying process data collected by means of an AI system

Information on our partners

FOS Messtechnik GmbH

Exipnos GmbH

Fraunhofer-Institut für Mikrostruktur von Werkstoffen und Systemen IMWS

Project specification

The aim of the research project is the development of an intelligent, image-based control of micro-dispensing valves for the application of fluids in the nanolitre range.
Within the scope of this project, a novel image analysis system is to be developed, which will make target-actual comparisons of the fluid droplets during application and surface wetting on the basis of graphic pattern recognition and display them in images. This innovative analysis tool should make it possible to analyze the optimum discharge quantity and the optimum degree of wetting of the fluid to be applied. In this way, the automated application of industrially used solder pastes for highly miniaturized electronic circuits on three-dimensional surfaces will be made possible for the first time.

Focus of the tasks of the project partners

• Analysis of the rheological properties of technical fluids (solder pastes, oils, lubricants, lacquers, fluxes, etc.) to determine the correct setup parameters for the application of different fluids with varying viscosities
• On the basis of the determined parameters, a wide range of fluids should be able to be processed via the application system to be developed by means of slight interventions in the system, and ideally only in the software
• Design of the valve including valve and plunger geometry as well as any radii, surfaces and materiality
• Development of the application actuator (definition of the plunger speed when processing individual fluids in order to generate the corresponding application quantity)
• Research into the impulse stroke of the plunger to generate an optimum droplet
• Analysis of the valve, depending on the viscosity of the fluid, for the application of specific quantities

• Development of a suitable novel and variable actuator for controlling the dosing valve, which can be coupled in real time to the feedback of the sensor system (electromechanical, pneumatic or hydraulic approach)
• Implementation of the sensor technology developed by iMes Solutions GmbH for inline detection of the fluid volume and the dosing quality in the dosing system
• Assembly of all mechanical and electronic components in a compact dosing system
• Development of a prototype

• Implementation of an image-based measuring system that can evaluate the fluid volume and the application quality of the dosing system on the basis of graphical evaluations - consisting of a mono camera and an external computing unit
• Conversion of an image-based hardware based on a 2-D image into a valid measurement result with regard to fluid volume and application quality
• Training the program with numerous mathematical and graph-based algorithms

Intended results

The image-based approach to application checking of fluids is a complete novelty in the field of dosing technology. This innovative analysis tool will make it possible to analyse the optimum dispensing quantity and the optimum degree of wetting of the fluid to be applied. Thus, for the first time, the automated application of industrially used solder pastes for highly miniaturised electronic circuits on three-dimensional surfaces will be possible. Furthermore, precisely controlled quantities of grease or oil can be applied to miniaturised components such as micro gears.

Information on our partners

Bahner Feinwerktechnik GmbH (precision engineering)

TH Köln (Cologne University of Technology)

Project specification

The aim of the project is the technological and innovative development of an intelligent progressive tool that enables permanent condition analysis and active, continuous readjustment during production.

The condition of the tool components affected by wear and tear should be recorded by sensors providing information about service life and necessary readjustments. The digital processing of this information through data recording, analysis, visualisation and evaluation in the form of data management is to be taken over by an algorithm to be developed innovatively as well as with new software components that enable the intelligent use of this data and automation as well as self-learning processes.

Focus of the tasks of the project partners

• Development and production of a highly sensitive sensor system in the clean room for use in progressive metal forming tools
• Determination of different sensor concepts for innovative data evaluation of defects, wear and tear in the forming tool
• Development of a micro-sensor technology system for the analysis of dimensional deviations on the sheet metal part to be processed

• Draft and design of an innovative force-distance diagram for the intelligent tool to record pressure and pressing as well as movement effects

• Design and development of the sheet metal forming tool as an intelligent progressive tool and implementation of the sensor technology for data analysis
• Development of the appropriate sensor placement in the innovative tool without blocking the operation processes
• Research into the processes of each individual station in the progressive tool suitable for a self-learning mechatronic process and development of the corresponding process parameters
• Development of the highly complex progressive tool design and connection of the information and communication technology to perform the role of the production optimiser

• Calculation and development of new algorithms from the data of the intelligent tool processes
• Development of an information and communication system for processing the sensor data and initiating quality measures
• Processing of sensor data and visualisation in a user-friendly and innovative software system

Results

The integration and implementation of the new tool using detailed parameters, measurement and limit values in the manufacturing processes by all partners involved was successful.

In summary, the advantages of the technological developments presented here can be described as follows:

• Ensuring product quality
• Increasing the output quantity
• Reducing defective production
• Increasing the lifetime of tools

Information on our partners

Brabant & Lehnert Werkzeug und Vorrichtungsbau GmbH

IMSAS - Institute for Microsensors, Actuators and Systems of the University of Bremen