Industrial Digitalization

A producer of pulp and paper aims to further enhance equipment effectiveness to strengthen its competitiveness. To achieve this, modern mathematical methods and AI algorithms are to be used to predict failure conditions and plan appropriate maintenance actions.

Together with the customer’s management team, d-fine develops a strategic vision for predictive maintenance, identifies suitable pilot use cases, and assesses feasibility and added value at a specific production facility. After an analysis of machine data, d-fine supports technical implementation and operational deployment of an algorithmic fault detection system for high-stress machine components. Subsequent functional extensions allow to differentiate fault types and estimate remaining useful lifetimes.

The system immediately leads to a substantial reduction in false alarms and serves as a data-driven assistant tool for the maintenance team. After successful pilot implementation and training of the internal maintenance teams, the application is rolled out to additional production lines and facilities.

To gain insights into complex relationships and implement digital value-added services, industrial companies need to consolidate, analyze, and quickly and clearly visualize shop floor data from a variety of machines and systems. The implementation of commercial IoT solutions often involves multi-year projects with uncertain added value and high costs. Companies at the beginning of their "digital journey" and research institutions, in particular, need to achieve rapid results while adapting their IoT solutions flexibly and cost-effectively to new developments.

The IoT platform d-scover combines proven open-source tools for collecting, storing, visualizing, and analysing industrial data into a flexible integrated system. The container-based deployment supports both cloud and on-premises use, allowing secure data exchange across organisations. d-scover can be flexibly integrated into company-specific systems and combined with commercial identity solutions and platform services. Stored data can be searched in natural language via API access to Large Language Models.

d-scover is used by research partners and industrial clients to connect machinery, analyse and visualize project data, and test innovative digital applications.

An internationally operating company in the metal industry aims to accelerate its maintenance processes. In particular, employees should be relieved of the time-consuming task of searching for maintenance information. For this purpose, relevant documents and other data sources in various languages need to be searched, with results reliably substantiated.

d-fine develops an intelligent maintenance assistant for the customer, based on large language models. This assistant searches through text, handwritten notes, and other types of data and can be queried in all languages spoken on the shop floor. Results are consolidated and presented in the language of the user making the query. To increase trust in the AI, source data is provided as evidence, and users receive dedicated training to ensure high application efficiency and a positive user experience.

The system is initially tested on selected equipment, receives excellent user feedback and leads to a significant improvement in the efficiency of maintenance processes.

A medical technology company wants to offer its customers Vision-AI solutions for end-of-line quality assurance on highly specialised inspection systems. Customers should be able to prepare their specific datasets, train individual neural networks, and subsequently install these on their respective systems.

d-fine implements a highly efficient AI pipeline for data augmentation, training, and evaluation of various deep learning methods, and fully integrates the solution into the client’s software stack. Over the course of a long-term collaboration, the system is enhanced with additional features, such as heatmaps for plausibility checks of the results, and is further developed into an AI-as-a-Service solution.

The company’s customers are provided with an individual training setup on a container-based cloud infrastructure, including user-friendly configuration mechanisms and a web front-end.

The client’s robots must perform complex maneuvers without endangering humans or causing damage to obstacles. A dedicated in-house team is responsible for developing the robot's control systems.

To improve the software lifecycle, the project focuses on quality control, reliability, maintainability, systematic processes for requirements engineering and development, as well as automated testing. The in-house software team and d-fine enter into a long-term collaboration to establish best practices for requirements management, architecture design, implementation, and testing processes.

A stringent CI/CD pipeline, seamless collaboration between internal and external teams, and best-practice software development processes increase efficiency, reduce workload, and deliver immediate value.

A manufacturer of infrastructure components with a high degree of variety for road construction wants to provide workers with an effective assistant to support the verification of component selection and the assembly of products, guiding them through complex work steps.

d-fine implements an augmented reality assistant that can be installed on mobile devices and AR glasses. Through the system, employees have access to digitalized work instructions and parts lists, which are made available for every work step. Navigation through the documentation can be done via menus or voice control.

Users can navigate complex construction instructions with the worker assistant. Additional products can be flexibly integrated without special requirements for digital images or document materials.

An aircraft manufacturer aims to digitally map production processes across its global locations and certify numerous processes to meet aviation safety standards. The focus is on tool, machine, transport, and production data, as well as the coordination of workflows. Commercially available applications lack the necessary functionality and desired flexibility.

d-fine supports the customer with a team of technically experienced consultants over several years in the design and implementation of data models and IT architectures, the development of suitable microservices for the system, and the analysis of machine, production, financial, and logistics data.

The solution ensures regulatory-compliant versioning and subsequent archiving of the production process, tool management, and monitoring of production workflows. It also enables data exchange between machines and measurement instruments.