direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Logo der TU Berlin

Inhalt des Dokuments

Automation Benchmark for Basic Control Layer Design of Chemical Plants: Process Transitions of a Vapour Recompression Distillation Unit

From Fachgebiet Regelungssysteme TU Berlin

Jump to: navigation, search



Transitions in chemical processes as startup, shutdown or load change operations are characterised by strongly coupled sub–processes and combine nonlinear continuous dynamics with a number of discrete–event switchings in plant equipment or control structure.

The implementation of stable plant operation is, on the one hand, an issue of the applied assets. But, equally important, it is also a matter of the design of the basic control layer, which involves hierarchically structured switches, sequences and continuous controllers. The design of this basic control layer poses a major task within the design of modern process automation. This layer is usually implemented in a digital control system (DCS).

Because of the complex dynamics and the large scale of the systems, detailed process modelling is usually expensive and, therefore, rarely applied. For this reason, control design has to be based on process data and intense process reviews. A clustering of control tasks is highly suitable, since control schemes shall be as simple as possible to reduce operation complexity and also commissioning and maintenance efforts. Plant downtimes due to, e.g., control commissioning or activating safety interlocks, are economically costly and pose safety risks and, therefore, have to be avoided. Therefore, when taken into operation, the control scheme should have as few errors as possible.

For these reasons, a sophisticated design approach is necessary. Under the given circumstances, this is not only a practical, but also a theoretical challenge. The definition of a benchmark for startup automation design of a complex chemical unit shall allow for investigation of theoretical design and practical engineering methods. This shall support the development of a comprehensive method for fast and efficient design, implementation, commissioning and maintenance as well as documentation of the basic control of complex plants in process industry.

The unit which was chosen for the benchmark is a distillation column with vapour recompression including the basic configuration of process variable measurements and safety interlocks. The automation task is the startup of the unit from an initial startup configuration to a production configuration under the presence of disturbances in the feed.

A distillation column with vapour recompression.

A simulation model of this unit is implemented in Matlab/Simulink. It has been tested for a wide range of operation conditions like startup and disturbance transients. The model was developed in a cooperation with BASF SE and is maintained by the Control Systems Group, TU Berlin, and is freely available via the download form below. It is meant to be part of benchmark examples in the following EU-projects:

  • Highly-complex and networked control systems (HYCON 2),
  • Embedded Optimization for Resource Constrained Platforms (EMBOCON).

A detailed derivation of the model equations as well as a description of the process and its operation is provided within the included report.


  • When requesting a download you obtain an email with a short description and a zip-file containing

- the simulation model files in Matlab and

- the plant modelling and operation report.

  • Here you can get the benchmark model files.
Privacy policy: By submitting this form you agree to the the following conditions:

Before providing the vapour recompression distillation model and a corresponding report, we collect the following user data:

  • your name
  • your e-mail address
  • your institution (company, academic institution etc.)

The data is solely used for scientific information exchange regarding the vapour recompression distillation application. This includes providing the model and the report via email, informing you about updates, corrections, etc. The data is also used for statistical purposes after being rendered anonymous.

The data is stored on a secure server and will not be passed on to third parties. The data is not used for any other purpose than indicated above. Once the data is no longer needed for the above purposes it will be deleted. The data will be stored for a maximum of 5 years.

I agree*
* required information
  • Feedback

Please send your feedback to: benchmark_support@control.tu-berlin.de


People involved

Recommend this page