On the monitoring task of solar thermal fluid transfer systems using NN based models and rule based techniques

Supervising thermal fluids transfer systems applied in solar parabolic through fields deals with a typical non linear process which suffers from lack of detectability when analytic model based techniques are applied on fault detection, isolation and reconfiguration tasks. This work describes the implementation of a supervision strategy to be applied in thermal fluids transfer systems used on solar parabolic through based power plants for which massive neural networks based functional approximation techniques associated to recursive rule based techniques on the basis of parity equations has been applied. Experimental results carried out on a test rig show that diagnosis applied to the thermal fluid transfer problems can be carried out under acceptable determinism and reliability. •Supervising complex heat transfer fluids and transfer systems.•Accurate functional approximation by means of feedforward neural networks.•Rule based diagnosis.