Bosch plans to roll out AIoT software for industrial automation in 50 powertrain plants worldwide this year, connecting it to more than 800 production lines.
More than one billion data messages will be stored on the analysis platform every day. Bosch plans to subsequently deploy the AI solution across the company at its roughly 240 plants.
The AI software was developed at the Bosch Centre for Artificial Intelligence (BCAI) to detect anomalies and malfunctions in the manufacturing process at an early stage. Bosch pilot plants where the AIoT technology is already in use are saving between one and two million euros per year. The plant in Hildesheim, for example, was able to identify and eliminate disruptions in process flows with the help of AI. As a result, the cycle times of the lines dropped by 15 percent.
“The use of artificial intelligence will make factories more efficient, more productive, more eco-friendly – and will make products even better. Our new AI solution will save plants millions in costs,” said Dr. Michael Bolle, CTO of Bosch.
Industry 4.0 in particular would benefit from the widespread use of AI in Germany, says Bolle. The greatest cost-saving potential is to be found in AI-assisted production where it can cut costs by more than 50 percent, or €182bn, according to a report by the Association of the Internet Industry and Arthur D. Little in 2019. However more than half of all German companies (58 percent) see disruptive potential in artificial intelligence, but only one in seven (14 percent) is currently using AI for Industry 4.0 according to Bitkom last year.
The pilot user of the new AI analysis platform is Bosch’s Mobility Solutions business sector.
Over the next few years, Bosch will invest €500m in digitalization and connectivity for its plants and expects to save twice that: €1bn by 2025. An integral part of the project is the use of artificial intelligence.
Next: Universal AIoT
Collaboration between BCAI and the division’s plants has resulted in universal AIoT software for manufacturing that uses Bosch Connected Industry’s Nexeed Manufacturing Execution System (MES) to automatically collect, process, and analyze data from a variety of sources in near real time.
Sensor data from machines serves as the basis for the system, determining fluctuations in a wide range of manufacturing processes. The Nexeed Industry 4.0 software captures and visualizes the data, and the AI system makes a recommendation for action, and the associate decides how to proceed. The main tools in this process are dashboards, individually configured and tailored to local use cases and the corresponding AI analysis.
This setup makes it easier to find potential causes of errors. Self-adapting processes for machines and assembly lines can be integrated as well. If, for example, a drill hole deviates from the defined placement, the AIoT system independently initiates the necessary steps. At times, the AI system receives support from cameras that are positioned along the production lines and record the manufacturing process. On the basis of patterns it has learned, the system identifies deviations, and action can be taken immediately. In addition, field and customer data is linked to the platform in individual cases. This helps the system understand even better how products behave in the field, enabling it to detect defects in good time and predict impending failures.
While the manufacturing industry in Germany has grasped the situation in theory, AIoT is still lagging behind in practice, says Bolle.
“Artificial intelligence is an epoch-making technology, comparable to the invention of letterpress. It will revolutionize manufacturing. With the help of artificial intelligence, machines and products learn how to be smart and anticipate,” he said. In addition to projects in its own plants, Bosch is launching AI-based solutions on the market. with automated visual inspection of workpieces, software for intelligent production management, and sophisticated energy management.
By 2025, the aim is for all Bosch products to either contain AI or have been developed or manufactured with its help. To do this, the company aims to train 20,000 associates in AI by the end of 2022 usin BCAI.
“We research and offer AI that is safe, robust, and explainable,” said Bolle. “Our plants manufacture a wide variety of products – from refrigerators and power tools to powertrains and assistance systems for the automotive industry and automation technology for use in factories. We are now adding AI algorithms to this expertise,” he said.
Related AIoT articles
- BATTLE FOR THE AI TOOLFLOW
- PREDICTIVE MAINTENANCE WITH CLOUD-CONNECTED SENSOR KIT
- BOSCH PREPS ITS BILLION EURO VEHICLE COMPUTER BUSINESS
- BOSCH STARTS 5G AUTOMATION TESTS AT DRESDEN WAFER FAB
Other articles on eeNews Europe
- TDK to close Scottish sensor factory
- Europe needs to wake from its 30-year semiconductor sleep
- FD-SOI breakthrough boosts performance
- NXP shows first details of edge AI i.MX9 processor
- Fateful Choices For ICs and Taiwan
- Semiconductor showdown approaches