The costs for managing the automation system represent a significant proportion of the overall engineering costs of an automated production facility. Many of the working steps required generate little added value and should therefore be automated. Under the motto “Generating rather than programming”, methods for the automated project planning of automation systems are being developed at the Nuremberg Campus of Technology. The aim is to generate large parts of the basic automation, as well as HMIs and simple models for virtual commissioning (VIBN) from existing planning data, within a machine process.

Shortening the time from receipt of order to delivery of a plant is one of the central demands on the mechanical and plant engineering industry. In the field of factory automation, plants/machines consist of a number of stations which are commissioned individually. A real commissioning usually starts with simple components such as pneumatic actuators, electric drives and various sensors. It is only when all the individual components are working properly that an overarching process (e.g. joining) can be programmed. This commissioning phase is characterized by many optimization loops with hardware/software adaptations. Significant improvements have already been achieved in mechanics and electrics by the parallel production of assemblies which are then joined together in the final assembly. However, this is not yet the case in the field of software commissioning.

A virtual plant model, also called a digital twin, is used to achieve the following results, among others:

• In project planning, the virtual representation of the plant provides all participants with the same understanding and enables detailed coordination. Specification deficiencies are detected earlier and coordination errors are avoided.
• In software development, the simulation serves as a test environment. It enables software errors to be detected earlier, thus avoiding expensive reworking.
• During operation, shorter downtimes are possible during planned plant conversions. Equally, set-up times can be minimized, and planned products and formulae can be tested on the twin first and quality statements made.
• The functional quality of the software can be demonstrated without risk in the office. Travel to the plant location is reduced. The real commissioning time is shortened, thus reducing personnel costs.
• Large plants such as those in the mining industry or ship lifts have to work as soon as they are first switched on. The software can be tested cost-effectively using the digital twin.
• Resources and consumption can be optimized by rearranging the individual stations, and the continuous optimization of systems becomes possible by means of tests on the twin. 
• OTS (operator training) systems – virtual training and testing of malfunctions without danger Training with the digital twin can be offered as a service.

How is a digital twin created?

Option 1 (discrete manufacturing):
3D CONSTRUCTION DATA is SUBSEQUENTLY enriched with information. This creates a model that contains geometry, kinematics and behaviour.

Option 2 (discrete manufacturing and process industry):
The starting point is a newly created BEHAVIOURAL MODEL. This model can be coupled with a 2D or 3D representation if it is an example of discrete manufacturing. This requires two model components: “Behaviour” and “Geometry”.

Option 3 (discrete manufacturing and process industry):
Already EXISTING PLANNING DATA are processed and important information is extracted. This data, from the process control system, from HMI graphics, from P&I flow diagrams or circuit diagrams, is used to generate the digital twin in a partially automated way.

Common to all the twins mentioned above is their application in such a way that the industrial controller is switched to either the real or the virtual plant. This results in the numerous possible uses.