KIWA

AI-Precision – AI-controlled Compact Precision Water Jet Recycling System

Schematic diagram of the machining head with water and cutting material treatment.
© Fraunhofer AZOM
Schematic diagram of the machining head with water and cutting material treatment.
Schematic diagram of the system for sludge and water treatment after waterjet processing.
© Fraunhofer AZOM
Schematic diagram of the system for sludge and water treatment after waterjet processing.

Motivation

Water jet cutting enables the precise separation of nearly all materials without heat input. However, in emerging precision machining (< 10 µm), compact systems for the treatment of water and abrasive sand are lacking. Currently, the abrasive sludge is mostly disposed of, which is costly and ecologically problematic. Recycling systems exist only for larger facilities and cannot reliably meet the high quality requirements of precision applications, as the quality of water and corundum changes complexly and dynamically due to the process.

Objectives and Approach

The aim of the collaborative project KIWA is to develop a compact, AI-controlled precision water jet recycling system that processes the abrasive sludge generated during the cutting process, allowing water and corundum to be reused over at least 100 cycles without compromising machining quality.

To this end, a laboratory and later a pilot plant will be established, consisting of a flocculation system, inclined clarifier, vortex and membrane filters, UV treatment, and comprehensive inline sensors (optical matrix, spectroscopy, turbidity, conductivity, pH, temperature, differential pressure). Based on the sensor data, Fraunhofer IWS will develop AI algorithms for real-time characterization of corundum and process water, as well as for automatic plant control. Maschinenbau Müller GmbH will realize the plant construction and control technology, WHZ will develop the optical sensors and spectroscopy, and IfU will specialize in chemical-physical sensors. Systematic measurement series will be conducted to collect datasets for different sludges, define limits, and verify the algorithms up to the pilot plant.

Innovations and Perspectives

KIWA combines a compact sludge treatment plant with an AI-based, physically motivated process control for precision water jet processes. Noteworthy innovations include the inline characterization of corundum structure (particle size, shape), spectral analysis of dissolved salts and metals, the fusion of classical control technology with neural networks, and automatic decision logic for the reuse, mixing, or disposal of materials. Material savings of approximately 60% for abrasive materials and significant savings of process water are expected, along with compliance with stringent environmental regulations. A new product segment with international market potential is anticipated. Fraunhofer IWS and WHZ will strengthen their roles as technology providers for AI-supported optical measurement technology in industrial processes.