Istraživanje uticaja promene temperature mineralnog ulja na budžet merne nesigurnosti sveakustičnog neiterativnog algoritma za lociranje parcijalnog pražnjenja
Creators
- 1. Elektrotehnički institut Nikola Tesla
- 2. Elektrotehnički fakultet Begrad
Description
U radu je predstavljen postupak za unapređenje proračuna merne nesigurnosti neiterativnog algoritma za lociranje parcijalnog pražnjenja u mineralnom ulju sveakustičkom metodom. Specifičnost datog algoritma je aproksimacija (pretpostavka) da je temperatura mineralnog ulja (brzina akustičkog signala) konstantna po volumenu materijala. Sa obzirom da je u relanim uslovima eksploatacije, temperatura mineralnog ulja nestacionarna, nehomogena i zavisna od uticajnih spoljnih faktora, ova pretpostavka može biti veliki izvor merne nesigurnosti pri određivanju lokacije parcijalnog pražnjenja datim algoritmom.
Neiterativni algoritam ima ukupno 19 parametara, odnosno relativno veliki devetnaesto-dimenzionalni prostor stanja koji je potrebno istražiti u cilju kvantifikacije uticaja promene temperature ulja na budžet kombinovane merne nesigurnosti. Za dati problem, generisanje i istraživanje prostora stanja, se može efikasno postići primenom Monte Karlo metode iz oblasti veštačke inteligencije.
U ovom radu, za određeni položaj senzora, istražuje se uticaj oblika i dimenzija oblasti ispunjene mineralnim uljem (u kojoj je moguća pojava parcijalnog pražnjenja) na doprinos temperature ulja standardnoj kombinovanoj mernoj nesigurnosti neiterativnog algoritma.
Takođe, diskutovano je o mogućnostima primene stečenih uvida (uzimajući u obzir savremena ekonimična senzorska rešenja) u konceptima kontinualnog nadgledanja pojave parcijalnih pražnjenja u minernom ulju u okviru Industrijskog interneta predmeta.
Abstract (English)
The paper presents a procedure for improving the calculation of the measurement uncertainty of a non-iterative algorithm for locating partial discharge in mineral oil by the all-acoustic method. The specificity of the given algorithm is the approximation (assumption) that the temperature of the mineral oil (the speed of the acoustic signal) is constant over the volume of the material. Since in actual exploitation conditions, the temperature of mineral oil is non-stationary, inhomogeneous, and dependent on influential external factors, this assumption can be a great source of measurement uncertainty in determining the location of partial discharge by a given algorithm.
The non-iterative algorithm has 19 parameters, i.e., a relatively large nineteen-dimensional state space that needs to be investigated to quantify the impact of changes in oil temperature on the combined measurement uncertainty budget. For a given problem, the generation and exploration of state space can be effectively achieved by applying the Monte Carlo method from the field of artificial intelligence.
In this paper, for a specific sensors' position, the influence of the shape and dimensions of the area filled with mineral oil (in which partial discharge can occur) on the contribution of oil temperature to the standard combined measurement uncertainty of the non-iterative algorithm is investigated.
Also, the possibilities of applying the acquired insights (taking into account modern economic sensory solutions) in the concepts of continuous monitoring of the occurrence of partial discharges in mineral oil within the Industrial Internet of Things were discussed.
Files
D1 04.pdf
Files
(1.4 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:d4b374ed94feb5285e7342f5df635890
|
1.4 MB | Preview Download |
Additional details
Additional titles
- Translated title (English)
- Investigation of the influence of change in mineral oil temperature on the measurement uncertainty budget of all-acoustic non-iterative algorithm for location of partial discharge