Model-based approach for elevator performance estimation. (February 2016)

• Record Type: Journal Article
• Authors: Esteban, E.; Salgado, O.; Iturrospe, A.; Isasa, I.
• Is Part Of: Mechanical systems and signal processing. Volume 68/69(2016)
• Page Start: 125
• Page End: 137
• Publication Date: 2016-02
• Subjects: Elevator -- Dynamic modeling -- State space models -- Kalman filters -- State estimation; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2015.07.005
• Languages: English
• ISSNs: 0888-3270

Abstract

Abstract: In this paper, a dynamic model for an elevator installation is presented in the state space domain. The model comprises both the mechanical and the electrical subsystems, including the electrical machine and a closed-loop field oriented control. The proposed model is employed for monitoring the condition of the elevator installation. The adopted model-based approach for monitoring employs the Kalman filter as an observer. A Kalman observer estimates the elevator car acceleration, which determines the elevator ride quality, based solely on the machine control signature and the encoder signal. Finally, five elevator key performance indicators are calculated based on the estimated car acceleration. The proposed procedure is experimentally evaluated, by comparing the key performance indicators calculated based on the estimated car acceleration and the values obtained from actual acceleration measurements in a test bench. Finally, the proposed procedure is compared with the sliding mode observer. Abstract : Highlights: The elevator performance is monitored by model-based system identification approach. The dynamics of an elevator installation is modeled in the state space domain. We monitor an elevator system based solely on signals extracted from the regulator. Calculated five key performance indicators are validated experimentally. The Kalman filter and the sliding mode observer estimations are compared.