Intelligent dynamic simulation of a solar collector field

Linguistic Equation (LE) modelling approach has various applications in non-linear multivariable systems. Insight to the process dynamic operation is maintained, and automatic generation of systems, model-based techniques and adaptation techniques can be applied in developing and tuning systems for process modelling and control. The multimodel LE approach provides a compact modelling of more or less smooth input-output dependencies. The overlapping operating areas are obtained by fuzzy clustering. The Fuzzy-ROSA method (FRM) serves for a data-based rule generation to model a given input-output dependency and is efficient for modelling complicated local non-linear structures. These properties are combined in a hybrid data-based modelling concept applied to dynamic simulation of a solar collector field. The hybrid fuzzy LE simulator was tested in data-based modelling of dynamic behaviour of a solar collector field. The new adaptive controller tuned with this technique has reduced considerably temperature differences between collector loops. Efficient energy collection is achieved even in variable operating condition.