Detection and quantification of valve stiction based on normality test and Hammerstein system identification

Abstract

Thermodynamic analysis of the effect of reservoir temperature change on the performance of organic rankine cycle system on geothermal power plant AIP Conference Proceedings 1755, 110005 (2016) Abstract. The presence of static friction or stiction in a valve can lead to poor performance in industrial automation and control system. This paper presents a method to detect and quantify stiction based on normality test and Hammerstein system identification. Detection and quantification are showed by the and parameters, where the value of and greater than 0 indicate the valve suffer from stiction. The proposed method is implemented to a confirmed sets of industrial data as a validation. The result show that the proposed method can accurately detect the presence of stiction and distinguish it to other problem in control loops.

Valve Stiction Detection and Quantification Using a K-Means Clustering Based Moving Window Approach

“…A bootstrap Hammerstein system identification procedure was developed by Yan et al 31 to derive control limits for stiction parameters. Hutabarat et al 32 claimed that the cause of control loop oscillations can be found to be either external oscillatory disturbances or poor controller tuning if PV is normally distributed. The authors (Hutabarat et al 32 ) used the Hammerstein system based approach to detect and quantify stiction if PV is not normally distributed.…”

“…Hutabarat et al 32 claimed that the cause of control loop oscillations can be found to be either external oscillatory disturbances or poor controller tuning if PV is normally distributed. The authors (Hutabarat et al 32 ) used the Hammerstein system based approach to detect and quantify stiction if PV is not normally distributed. In all the Hammerstein model based methods discussed so far, the sticky valve was modeled by a data-driven model which was derived based on the input−output behavior shown in Figure 3.…”

Comparison of current state of control valve stiction detection and quantification techniques

“…The Hammerstein methodology is used as a model framework and a two-stage optimization is used to identify model parameters. Hutabarat et al (2016) developed a stiction quantification method using a normality test followed by Hammerstein model methodology. First, a PV normality test is performed to find out whether the process is normally distributed on the basis of skewness and kurtosis values of PV data.…”

Implementasi Metode Deteksi Friksi Statik pada Katup Kontrol menggunakan Pencocokan Elips pada Standar IEC 61499

“…Hutabarat dkk. [8] melakukan pengembangan metode deteksi dan kuantifikasi friksi statik pada katup kontrol berbasis uji normalitas dan identifikasi sistem Hammerstein. Penelitian ini melakukan uji normalitas pada data PV lalu melakukan deteksi dan kuantifikasi friksi statik menggunakan pemodelan Hammerstein.…”