Hammerstein–Wiener identification of industrial plants: A pressure control valve case study

Abstract

Abstract Hammerstein–Wiener (H–W) identification problem is investigated under practical considerations for industrial plants. The main consideration in these cases is inability to apply arbitrary input signals. An algorithm is proposed to identify H–W model for industrial systems, in the presence of noise and disturbance. The identification problem is changed into a constrained optimisation problem and is solved by minimising the non‐linear functions modelling error. To improve the efficiency of the proposed method, the known information about non‐linear functions are used. To make the proposed method applicable for general systems with hard non‐linearities, non‐linear functions are described by some ordered pairs. A pressure control valve that is used for anti‐surge protection of an air compressor in the second refinery of South Pars Gas Complex is considered as the case study to demonstrate the effectiveness of the proposed algorithm. Seventy‐five percent of the gathered data during the time is used for identification and the remaining 25% is used for verification of the resulted model. Verification of the obtained H–W model showed that accuracy of the identified model is more than 95%.

Keywords

• Optimisation techniques
• Pressure and vacuum control
• Nonlinear control systems