atr-ftir spectroscopy based on pls-r multivariate calibration technique for assessing transformer oils samples to predict of dielectric dissipation factor (ddf) in power industry

Abstract: transformer is one of the most applied equipment’s in electrical circuit. transformers change the voltage from one level to another one by using magnetic field effect. it is cooled down by using oil in power industry for reducing current for transforming electrical energy. increasing the temperature of oil transformer leads to damage and aging of this oil. the most important parameters of transformer oil are about the water content, breakdown voltage, furfural concentration, color, viscosity, surface tension, sulfur content and solid content. nowadays, the quality of a power transformer is determined by measuring the dielectric dissipation factor (tan-delta) as a primary parameter for evaluating the deterioration of an oil transformer. the dielectric dissipation factor (ddf) of oil samples was determined under the iec 60247:2004 standard testing method [1].in this research, attenuated total reflection fourier-transform infrared (atr-ftir) spectroscopy is employed as an analytical technique based on multivariate calibration techniques to determine the ddf for oil samples used in the power industry. the partial least squares regression (pls-r) as a linear multivariate calibration technique was used to forecast the ddf in oil samples using successive projection algorithm (spa) variable selection as suitable models. as a result, in the spa-pls-r chemometric method, the correlation coefficient (r2pred) and root mean square prediction error (rmsep) were 0.9773 and 0.0045, respectively. thus, using mid-ir spectroscopy, pls-r combined with spa can be a practical method for reliable, rapid, and simple quantitative analysis of ddf in the power products industry.