In the initial training phase of our AI models, process data is recorded at high frequency and workpieces are measured conventionally. This contextualized data is used to train the models, creating a "virtual AI measurement system." In series production, we use exclusively the recorded process data and the AI model to calculate manufacturing metrics such as diameters, lengths, or roughness with high precision in real time after the workpiece has been produced. This allows customers to quickly identify problems and take action to ensure quality.

During the production of a workpiece, we collect process data from the machine tool's control system, drives, and sensors at high frequency. After machining is complete, an individual digital signature is created for each workpiece. The workpieces are then measured conventionally, and it is crucial that each digital signature is assigned to the appropriate measurement protocol (track & trace).

The process data and measurement protocols are stored in pairs and then processed by the neural network to create a high-precision machine learning model. The number of required data pairs depends on the required manufacturing tolerances and the accuracy desired by the customer. Based on the data volume, the model learns factors such as tool wear, vibrations, temperature fluctuations, and material variations, enabling the transfer of production data to quality data.

Model training is complete once all manufacturing tolerances can be calculated with the required accuracy. The more precise the digital model, the higher the so-called model quality, which describes the statistical error.

After the training phase is complete, the neural network can use continuously acquired process data to precisely calculate manufacturing tolerances using the trained machine learning model. This AI-based virtual quality control enables all manufacturing tolerances to be determined with high accuracy based solely on the process data and the machine learning model.

Together we will record your current status, such as workpiece, machine and control type, etc. Once all necessary components are in place, we will install the data recording system on your machine.

During the production of a workpiece, we collect process data from the machine tool's control system, drives, and sensors at high frequency. After machining is complete, an individual digital signature is created for each workpiece. The workpieces are then measured conventionally, and it is crucial that each digital signature is assigned to the appropriate measurement protocol (track & trace).

The process data and measurement protocols are stored in pairs and then processed by the neural network to create a highly precise machine learning model.

Big Data in Manufacturing GmbH is an innovative company specializing in the integration of advanced data analytics and artificial intelligence (AI) into manufacturing processes. Our goal is to sustainably increase efficiency, quality, and productivity in industry by leveraging machine data and modern technologies. We rely on AI models such as neural networks to revolutionize quality assurance. The focus of our development is on solving problems of conventional quality assurance and creating technologies that enable cost-effective, 100% AI quality control and zero-defect production.