Big data: lubricant and raw material for Audi

BasicInfo Sub Page
11/17/2016
Back to overview

Data is the lubricant at any company – and an important raw material for the production of the future. The more systematically that data is prepared and evaluated, the more valuable it is. With big data, Audi can ease its employees’ work, make processes more efficient, predict mistakes or avoid them completely. The Audi Production Lab provides two case studies.

The people and machines working in production at Audi create an enormous amount of data – at an accelerating rate. This data contains valuable information and interconnections, which need to be discovered and evaluated, because it has great potential for the production processes – at the technical and financial levels, for the employees and for the customers.

In general, big data refers to

  • the strategic recognition of data as a resource with its own value,
  • the operational recognition of data as a separate means of production,
  • the active and consistent utilization of all product and process data,
  • the connection, linking and correlation of various data sources,
  • and a change in corporate culture, organization and departmental processes.

For the Audi Smart Factory, the overall meaning of big data is the changeover towards data‑driven and thus highly flexible and highly efficient production – towards the so‑called “data driven business.”


Big data case study 1: the screws‑and‑bolts analysis
Whether suspension, engine or interior – on average, every Audi has more than 1,000 screws and bolts tightened to a certain torque; at the Ingolstadt plant alone, about 500 million of them are tightened each year. With a small percentage of them, this procedure doesn’t immediately run smoothly, when there are metal filings in the thread for example. This isn’t really critical, because Audi’s pneumatic or electric screwdrivers permanently measure their operating angle and the torque being applied. If these parameters exceed the limits defined for each screw or bolt, the machine automatically switches off. That usually happens within two seconds of the start of the procedure.

But employees of the Audi Production Labs have now found out that so much time isn’t necessary. This finding was based on an analysis of anonymized screw‑and‑bolt data that had been recorded over several weeks. The result: A mistake can already be predicted with a high probability after 0.3 seconds. When the screwdriver then switches off immediately, the employer saves time and is less hurried for the next screw or bolt.

The new screw‑and‑bolt analysis is already in use in the Audi A3/Q2 assembly. This has resulted in an enormous efficiency advance for AUDI AG at the Ingolstadt plant, and the method will now be rolled out at the other plants. The next step is to expand the spectrum of evaluation and to find out whether the method can be transferred to other operations.


Big data case study 2: the project “CKD tool international logistics”
Columns of numbers are hard to read and understand. But from childhood onwards, we are all used to understanding and storing pictorial information quickly. This basic principle is followed by the project “CKD tool international logistics” (CKD = completely knocked down) in the Audi Production Lab. Its goal is to improve the flow of components in logistics by means of a new and easily understood pictorial presentation of data.

Audi’s CKD plants around the world carry out contract production by which cars are put together out of parts kits – preassembled groups of components. These are delivered to their destinations by the suppliers via so‑called consolidation points and packing locations. If the logistics experts want to analyze this flow of goods and examine its optimization potential, they have so far had to work through complex Excel spreadsheets.

The new tool, which is based on an existing tool, is based on other principles: the intelligent linking of previously separate databases and their graphic implementation in shapes and colors. With a mouse click, users can recognize on the computer screen interdependencies in the flow of goods that were not visible before. The simpler the depiction, the clearer and more comprehensible it is.

One of these new levels of visualization is a map on which suppliers are marked. Depending which packing location they are serving, they are marked in different colors; for example orange stands for Ingolstadt, blue for Duisburg and violet for Wunstorf near Hannover. A glance at the map already showed that there were several suppliers in southern Hesse within a radius of 20 kilometers that supplied different packing locations. Their delivery routes were then combined, saving Audi a lot of time and money, and also reducing CO2 emissions and noise.

Whether for planning truck routes or ensuring the constant utilization of the consolidation centers for many weeks – the “CKD tool international logistics” brings great added value. It has the potential to reduce transport costs by a large six‑digit amount. The prototype has already advanced to series maturity in the P‑Lab.


Assisted, augmented, virtual: Audi utilizes data glasses

In the smart factory of tomorrow, data glasses will provide targeted support for employees in assembly, as well as for planners and engineers. The brand with the Four Rings is at the forefront with the technology of data glasses. The Audi Production Lab is currently testing glasses technologies for assisted reality, augmented reality and virtual reality. Their areas of application and degrees of maturity differ greatly – but each technology has enormous potential.


Assisted reality: pilot application in engine assembly
The production of the future will be increasingly complex and the responsibility of each assembly employee will grow. Against this backdrop, data glasses that support employees with their work are gaining importance. At the engine plant in Győr, Hungary, Audi is currently testing data glasses from Google: Google Glass. They are being used in the area of assembly of original parts. There, all engines are built by hand in accordance with customers’ specifications. The complexity is immense: The production process takes several hours; some shelves in the area of engine assembly have up to 200 sections for small parts. Many of these parts look very similar.

Data glasses help here. With a plant pass and a QR code, the assembly worker checks into his or her workplace, Google Glass then receives the individual assembly order from an order server. In the form of a picture‑and‑text presentation, the individual steps of assembly for the respective engine version are shown over the right eye. By voice or via the touch‑sensitive frame of the data glasses, the employee can run through the work stages and confirm them or, if required, call up detailed information such as a training video. When the employee is standing in front of the shelf, he or she sees how many parts with which serial number have to be taken from which section. However, the employee only sees this information when he or she looks up – otherwise, the field of visions remains free.

Google Glass is a suitable wearable for everyday use – it is robust and visually inconspicuous. Its technical complexity is similar to that of a smartphone: It includes a processor, memory, microphone, loudspeaker, camera, radio module and battery in a light frame. It can be used for several hours at a time, and for an eight‑hour shift if an external battery is connected. A scientist is accompanying the pilot project in Győr and is recording anonymized data. She compares that with data from the current, monitor‑based assistance: Do the glasses save walking distance? Do they reduce assembly times? How secure is their operation? Do the employees feel at ease with them?


Augmented reality: fusion of reality and simulation
With augmented reality (AR), one’s real surroundings are merged with data from the computer. The Audi Production Lab is working on this with specially designed, completely new data glasses: the HoloLens from Microsoft. Through them, the user continues to see the real surroundings, but can augment them with virtual images as holograms – for example with a welding robot from the body shop.

The Hololens integrates the computer power of a tablet, including the battery. Visually, it is similar to a helmet visor with a strap that surrounds the head. The strap controls the Hololens via voice input and with finger gestures. Loudspeakers aimed at the ears play sound effects.

The AR glasses from Microsoft use several cameras and 3D sensors to permanently measure the room in which the wearer is located. In the map of the surroundings thus created, it projects the respective hologram onto the desired place – and it stays there no matter how the wearer moves. The hologram appears on special displays in the visor, slightly offset for each eye, creating the effect of stereoscopic vision with depth.

This shows what is possible with the Hololens: When the glasses are fed with the right 3D models, it is even possible to plan entire body shops. The user can superimpose as many virtual robots as desired, and can place them so that their gripper arms and other moving parts do not collide with each other. Several experts can follow, assess and alter the procedure at the same time.

The HoloLens offers new potential also with regard to cooperation between different locations. For example, if there is a malfunction at a plant, employees can use the glasses to send images and data to an expert providing assistance from another plant. In this way, they can together identify the cause step by step and can then rectify the malfunction.


Virtual reality: full integration in the simulation
The third data‑glasses project of the Audi Production Lab is a virtual‑reality (VR) application that fully integrates the user into the scenery presented. Like in a gamer world, the technical core consists of common VR glasses of the type HTC Vive. Two tracking stations placed in the room communicate with sensors on the glasses and the hand controllers. In this way, the system recognizes the user’s movements and adjusts the presentation with practically no delay. For that purpose, the VR glasses are connected to a high‑performance laptop that the user carries in a backpack, avoiding inconvenient cables. Several users can experience the virtual world together and simultaneously.

In the data world, Audi employees will be able to go anywhere –into a virtual workshop for example where the cockpit of an Audi Q2 is installed on a scale of 1:1. They will meet there to identify possible problems with installation of the infotainment control unit in the storage compartment already in the design phase.

The hand controllers can be used for various actions: They move the infotainment control unit or make marks where needed; they transfer working stages into three‑dimensional simulation. For example, when they cut through the front of the cockpit, the control unit is revealed; further back, parts of the air‑conditioning and the passenger airbag can be seen, all presented in technical colors.

Product verification, factory and process planning, employee training – VR technology has interesting solutions for all of these applications. It also allows meetings in the virtual space. Within a short time, factory planners and machinery manufacturers can meet across continents in a factory that might not yet even exist. In the 3D simulation, complex interrelations become clearer and easily comprehensible, facilitating a reliable assessment.

Audi is now starting the first pilot application of the data glasses, to be gradually followed by further applications. The first VR meetings involving several locations will soon take place. In the Audi Production Lab, the next step is being considered: The participants’ faces will be seen in the simulation in the future. Because also in the digital age, a look often says more than a thousand words.