Time delay estimation is one of the most important issues in system control and signal processing.Through establishing an appropriate fitness function
the time delay estimation of soft-sensor model could be formulated as a multi-dimension nonlinear functional optimization problem
which could be solved by hybrid differential evolution (HDE) algorithm.Numerical simulation results on two benchmarks demonstrated the effectiveness and robustness of HDE.Taking a crude atmospheric distillation unit in oil refinery as an example
a numerical application of the soft-sensor model for the kerosene flash point was made.The application results demonstrated that the accuracy of soft-sensor model was greatly improved with the introduction of time delay estimation.Besides
the effectiveness of the proposed HDE was confirmed.