A beneficial, customer-oriented production is the goal of every company - and a great challenge at the same time: the requirements for quality, costs and production time must be met and at the same time reconciled with the issue of sustainability. In series production, the cooling lubricant supply has a significant influence on the process result and is often considered the cause of quality variations.
Record and analyse process factors
Nevertheless, the process factors on which these qualitative differences depend are rarely investigated. "Many companies don't even deal with this question," knows Tobias Kaufmann, research associate at the Machine Tool Laboratory of RWTH Aachen University and head of the Digital in NRW transfer project. "But the larger the supply chain becomes, the more machines are connected to each other, the more important it is to know how each individual system is supplied with operating materials such as compressed air, water, electricity or cooling lubricant, where the pressure is and what the temperature or germ load is. Only with the help of the quantitative recording of these status variables in a networked system can differences and errors in the process be detected and compensated for," says Dr.-Ing. Dirk Friedrich, Managing Director of grindaix GmbH. "However, many companies do not yet record these parameters despite their technical feasibility."

Development of a digital process monitoring system
In order to achieve clear predictions on the supply of operating materials in production, Digital in NRW and Grindaix are working on "Resilient process control through AI-based cooling lubricant supply during grinding", the official title of the transfer project. The idea behind this is the development of a digital process monitoring system. With the help of artificial intelligence, errors and deviations in the supply of the machines are to be detected, recorded and recommendations for action developed in the future. The transfer project with Digital in NRW sets the course for a sustainable cooling lubricant supply with high quality standards. "Predicting failures in the supply chain and taking direct action to avoid interrupting the production process is becoming increasingly important," says Kaufmann.
From the actual state analysis to testing
The experts from Digital in NRW first recorded the possibilities of a centralised or decentralised cooling lubricant supply of a machine tool as part of an ACTUAL analysis and identified communication interfaces. This was followed by the development of a target concept: in close cooperation with the company, a suitable sensor system for recording relevant coolant supply parameters was developed, as well as a technical solution concept for digitally networking the individual interfaces. "We are currently testing coolant monitoring with the latest fluid sensor technology in a research environment to generate and analyse data," explains Tobias Kaufmann.
Simulation on the demonstrator
In the machine tool laboratory in Aachen, work is being carried out on analogue test benches and a demonstrator with pipeline, pump and sensor technology is being set up to simulate the various process sequences and causes of faults: What happens when a pipe becomes clogged? What if the pressure is too low? "The challenge is to create a database that the algorithms can work with," explains Tobias Kaufmann. The connections between cause and effect must be researched, and recommendations for action must be developed. "To do this, we have to record scenarios from real operations and transfer them to the test bench in order to record their effects on the data," says the project manager. The core of the transfer project is to develop a methodology for plausibility checks that can later be transferred to various applications. High-quality plausibility checks provide "the basic prerequisites for subsequently derived recommendations for action, which should ultimately improve process control and not cause a machine crash." The more data available, the greater the later coefficient of determination of the algorithms.

More efficiency, less waste
Subsequently, three representative cases are to be transferred from the demonstrator to the real environment and the findings are to be used in real production. The plan is to use the developed concept in an assistance system on a series machine and prepare it for practical use. "Small and medium-sized manufacturing companies can certainly benefit from the development," Tobias Kaufmann emphasises. "The use of this system creates resource efficiency, more process reliability and the reduction of rejects, which in turn can save costs."