Dürr delivers one of biggest paint shops in China to SAIC Volkswagen; AI and big data
15 April 2021
Despite all the restrictions caused by the coronavirus pandemic, Dürr has handed over one of the biggest paint shops in China to SAIC Volkswagen (SVW) on schedule. The new plant in Anting boasts a production capacity twice that of a standard paint shop at 120 bodies per hour, and behind the intelligent production is software and artificial intelligence on a hitherto unknown scale.
A smart painting line records around 3,500 digital data points for every individual body. On top of this, sensors supply many gigabytes of data in process values for temperature, pressure, and humidity, for one example. This information is the raw material for increasing the overall equipment effectiveness (OEE) of a paint shop.
Dürr supplied all of the plant engineering and the most comprehensive portfolio yet of its proprietary DXQ software products for the highly digitized factory.
From the beginning, mechanical engineering and IT expertise went hand in hand in the SVW project. The digital solutions were developed in parallel with planning and implementing the comprehensive plant engineering—from the RoDip dip coating system through paint booths with EcoDryScrubber dry separation to the oven and the air pollution control.
SVW’s digital requirements catalog essentially lined up with the development plan of Dürr’s Digital Factory with the result that many new DXQ products found their way into the plant. Over the 18-month project duration, six international product development teams from Dürr worked with up to ten developers in each case on the new technologies. They coordinated closely with SVW in short sprints, as is customary in agile development.
To ensure as smooth an IT installation as possible in China, all the software functions were first tested in the Dürr headquarters in Bietigheim-Bissingen in an environment that faithfully replicated the conditions in Anting.
DXQcontrol for controlling the mega paint shop. SVW uses nearly all modules from the feature-rich DXQcontrol software solution for higher-level plant control. It enables the life cycle of each body to be tracked from beginning to end.
This starts with receipt of a production order, whereby the order information is translated into concrete production steps. In addition to order data management, the DXQbusiness.intelligence module from the DXQcontrol product family is also used in a new form. For the first time, consumption data such as energy, water, or air usage can be evaluated historically over long periods.
This lays the foundation for sustainable plant operation. A further innovation is the use of mobile apps provided by various DXQ products. For example, employees can manage and filter alarm functions directly at the plant’s point of origin using tablets if the line comes to a stop.
One piece of maintenance software for all components in a paint shop. For SVW, mobile versions also have the advantage that the employees can access maintenance data using tablets by scanning QR codes on plant components. In this way, information can be called up directly at the point of origin. This enables the employees to work more flexibly than is possible with stationary PCs.
At SVW in Anting, the DXQequipment.maintenance software is being used for the first time for a complete paint shop, and as a mobile version. The software has interfaces to the plant equipment’s almost 130 controllers, and uses them to determine each piece of equipment’s need for maintenance based on up-to-date information like operating hours or counter readings. SVW can also incorporate plant components from other suppliers here.
Optimizing the painting process with big data. Several terabytes of digital data per year are collected, saved, and evaluated by the intelligent algorithms of the DXQplant.analytics application in the new paint shop. If, for example, employees detects an irregularity on the paint surface during the quality inspection, they enter it into the tablet. DXQplant.analytics correlates recorded quality results with order data (production number, derivative, color, etc.) and process data based on artificial intelligence to identify patterns.
A broad base of data makes it possible to track the fault causes in a targeted way and define measures. In this way, the system can be permanently optimized. For the first time, process data from other suppliers’ components and systems are also evaluated at SVW in Anting using DXQplant.analytics, for example in Application Technology.
Comments