Abstract:

By using a closed loop identification method, we build a mathematical model of the electron linacs plant of cooling water system in Shanghai Synchrotron Radiation Facility. Due to the impact of noise, which cannot be unmeasured, on the control input data through a two-stage closed feedback loop, identification is used to avoid bias of the obtained model. The algorithm transforms a closed loop identification problem to two open loop problems by constructing an intermediate signal without noise and pollution to replace the actual input, leading to de-correlation between input and noise. Using an existing open identification method, we obtain a consistent estimate of the system.

Key words: two-stage closed loop identification; direct identification; unmeasured noise