In this paper we have studied the failures in medical electrodes such as electroencephalogram electrodes (EEG) being used for collecting brain signals. As those electrodes have to guarantee high level of reliability it is important to explore and predict the possible occurrence of failures in there structure. The electrode tip (needle) made of stainless steel is covered with thin oxide film acting as a dielectric and determing the total electrode resistance. In fact, studying the fluctuations of that resistance gives the insight into defects of the whole structure. The electrical properties of the oxide layer are characterized by charge hopping mechanism and the total resistance could be modeled by implementing random resistance network (RRN) methodology.The applied computational algorithm is based on Monte Carlo simulation procedure with direct and iteration methods. The obtained simulation results show non-gaussian Bramwell-Holdsworth-Pinton (BHP) distribution of the total resistance fluctuations, and they verified by the experiments.
Key words: Cold solder, oxide layer, failure detection, biased percolation, RRN simulationModeliranje i detekcija kvara u medicinskim elektrodama. U ovom je radu razvijen model za simulaciju predviđanja defekata u posebnim elektrodama koje se koriste pri mjerenju i prikupljanju signala mozga. Ključni dio elektrode je njenčelični vrh prekriven tankim oksidnim slojem koji ujedno djeluje kao dielektrik. Upravo fluktuacije vrijednosti otpora dielektrika odnosno njihovo rasipanje od normalne do BHP raspodjele, korištene su kao osnova predloženog modela. Pritom je primjenjena metodologija mreže nasumičnih otpora (RRN), algoritam temeljen na Monte-Carlo simulacijama kao i teorija usmjerene perkolacije. Definiran je i poseban parametar kao indeks intervala valjanosti za svaku elektrodu. Simulacije su potvrđene mjerenjem u laboratorijskim uvjetima na komercijalnim EEG elektrodama.