Modelling of integrated industrial processes has been and is still challenging because of the huge number of details needed to be considered whence modelling the target plant. The resulting large set of heterogeneous equations to be solved during the simulation imposes rigorous requirements on both the solution methods and the computing power.
 
The experiences presented in the lecture are based on numerous full scale application projects in co-operation with global industrial players in the fields of power plants, pulp and paper processes, and control systems engineering.  The target plants are situated all over the world. In many cases the construction of large scale pilot plants has been avoided when making consistent use of modelling and simulation. Much time and money has been saved. One referred application case comprises of an extensive engineering effort whereas a complete DCS system is tested against a computerised simulation model of the target plant before it is taken into use at the real plant. Typically, less configuration errors and shorter commissioning time is experienced.

The reason for a still halting take-up in Europe of the modelling and simulation as a working method in the complete plant life-cycle engineering processes is that simulation has usually been separated from other engineering disciplines.  Accordingly, for instance the manual extraction of required model specification data from various sources to a simulation tool requires availability of dedicated modelling experts and it is very time consuming, expensive and error prone.  

The advent of DCS standards for real time communication like OPC, already adopted by most automation system vendors, makes it easy to connect the control system to a process model. Another benefit is that the control system not is needed to be simulated if a training simulator is requested; a copy of the real DCS system can be used. Further, the introduction of semantic specification standards for process design data makes it possible to develop intelligent tools enabling automated simulation model specification for design evaluation purposes as well as tools enabling proactive maintenance optimisation. Recent OPC developments provide for semantic specifications, which opens up interesting possibilities for introduction of new control system functionalities.

The required computer power for extensive simulation is in many cases already available on hand held computers. The solution methods, however, need to be developed to suite automated simulation model extraction from semantic plant data bases. Also the content of the plant data bases need to be extended to include such parameters that are needed for the model specification. Above issues will be dealt with in required detail in the lecture. The typical software engineering issues of concern can not be neglected in this context.

Examples of real commercial implementations will be presented.