Artificial neural network based method for temperature correction in FDS measurement of transformer insulation

Mousavi, Seyedali; Sedighizadeh, Mostafa; Hekmati, Arsalan; Bigdeli, Mehdi

doi:10.1088/1361-6463/ab62c2

Cited by 8 publications

(10 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To test the moisture content of power transformer insulation, diagnostic methods based on relaxation changes in the time domain using the recovery voltage method (RVM) [3][4][5][6] and polarization depolarization current (PDC) method [7][8][9][10] are used. In the frequency domain, however, the values of the loss factor tgδ and system capacity at low frequencies are determined using the frequency domain spectroscopy (FDS) method [11][12][13][14][15].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Precise Measurements of the Temperature-Frequency Dependence of the Conductivity of Cellulose—Insulating Oil—Water Nanoparticles Composite

et al. 2020

View full text Add to dashboard Cite

This article presents direct σDC and alternating σ(f) current conductivity measurements obtained by the frequency domain spectroscopy (FDS) method on cellulose-transformer oil–water nanoparticle composite with a moisture content of (5.0 ± 0.2)% by weight in a temperature range from 293.15 to 333.15 K with step of 8 K. The uncertainty of temperature maintenance during measurements was below ±0.01 K. The sample was prepared for testing in a manner as close as possible to the cellulose insulation moisturizing process in power transformers. For the analysis of the results obtained, a model of alternating and direct current hopping conductivity was used, based on the quantum phenomenon of electron tunneling between the potential wells and nanodrops of water. It was observed that on the d(logσ)/d(logf)-derived waveforms there was a clear low-frequency maximum, and a tendency to reach the next maximum in the high-frequency area was visible. On this basis it was established that the increase in conductivity takes place in two stages. It was found that the position of σ(f) waveforms in the double logarithmic coordinates is influenced by the temperature dependence both of the conductivity and of the relaxation time of the conductivity. These relationships are described with the appropriate activation energies of the conductivity and relaxation time of conductivity. Based on the analysis of experimental data using Arrhenius diagrams, average values of the activation energy of conductivity ΔWσ ≈ (0.894 ± 0.0134) eV and the relaxation time of conductivity ΔWτσ ≈ (0.869 ± 0.0107) eV were determined. The values were equal within the limits of uncertainty and their mean value was ΔW ≈ (0.881 ± 0.0140) eV. Using the mean value of the activation energy, the frequency dependence of conductivity, obtained at different temperatures, was shifted to 293.15 K. For this purpose, first the waveforms were shifted along the horizontal and then the vertical axis. It was found that after the shift the σ(f) waveforms for the different temperatures overlap perfectly. This means that the shape of the frequency dependence of the conductivity is determined by the moisture content of the pressboard. The position of the waveforms in relation to the coordinates is determined by the temperature relationships of the conductivity and the relaxation time of the conductivity.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Formulas (14) and (16) show that the DC or low-frequency conductivity for electron tunneling amounts to:…”

mentioning

confidence: 99%

Precise Measurements of the Temperature-Frequency Dependence of the Conductivity of Cellulose—Insulating Oil—Water Nanoparticles Composite

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Machine learning (ML) has become increasingly popular in recent years across a variety of fields [8], including the analysis of images [9,10], videos [11], and spectra [12][13][14][15][16][17][18]. The features in the data can be successfully extracted using ML algorithms [19], and the extracted features can be used for regression [6,[20][21][22][23][24], classification [25,26], and error detection.…”

Section: Introductionmentioning

confidence: 99%

Using transfer-learning-based algorithms as data reduction strategies for volatile organic compounds classification using plasma spectroscopy

Zhang

Hsu

2023

J. Phys. D: Appl. Phys.

View full text Add to dashboard Cite

This work presents the use of transfer-learning-based algorithms as data reduction strategies for classification of volatile organic compounds (VOCs) using optical emission spectroscopy (OES) of plasmas. The plasma used is generated with a home-made microplasma generation device (MGD) ignited in the mixtures of Ar and VOCs. The spectra are acquired from 10 MGDs. The VOC tested are methanol, ethanol, and isopropanol. VOCs are classified using a convolutional neural network (CNN). In addition, gradient-weighted class activation mapping (Grad-CAM) is used as the explainable artificial intelligent technique. It ensures the model classification is based upon rational plasma physics by considering appropriate wavelengths. The VOC concentrations are then quantified using linear regression and an artificial neural network (ANN). Transfer learning-based algorithms tested are parameter transfer, REPTILE, and self-training. Spectral data from 10 MGDs are grouped into source and target datasets. Ten MGDs are tested individually using a model that was trained on the rest nine MGDs. The three MGDs with the lowest accuracy are chosen as the target dataset, while the other seven MGDs make up the source dataset. The original target dataset has 22500 spectra and is further reduced to 12600, 9000, 1800, 225, and 22 spectra to test the behavior of each algorithm. With 225 spectra used for training, the model trained with the random initial (RI) model shows an accuracy of 0.82. The models trained with parameter transfer and REPTILE have accuracies of 0.98 and 0.95, respectively. Finally, an ANN model is used to quantify the VOC concentration with an R2 value of 0.9996. The results demonstrate the potential using transfer-learning-based algorithms as the data reduction strategies for classification of spectroscopic data.

show abstract

“…With the development of artificial intelligence technology, some scholars have begun to use machine learning to evaluate the moisture content of oil-paper insulation equipment. Mousavi et al [13] used genetic algorithm to find the equivalent circuit parameters of transformer and eliminated the influence of temperature on FDS curve through neural network. Ma et al [14] used support vector machine (SVM) to evaluate the moisture content of transformers, where the complex permittivity in the FDS curves was the sample attribute, and the moisture content of oil-paper insulation was the mark.…”

Section: Introductionmentioning

confidence: 99%

“…Mousavi et al. [13] used genetic algorithm to find the equivalent circuit parameters of transformer and eliminated the influence of temperature on FDS curve through neural network. Ma et al.…”

Section: Introductionmentioning

confidence: 99%

Evaluation method for moisture content of oil‐paper insulation based on segmented frequency domain spectroscopy: From curve fitting to machine learning

Yao

Ding

et al. 2021

IET Science Measure & Tech

View full text Add to dashboard Cite

In recent years, frequency domain spectroscopy (FDS) is often used to evaluate oil paper insulation state in power transformer bushing. But it is still very difficult to evaluate the moisture content accurately and quickly. In order to solve this problem, this paper proposes an intelligent algorithm based on random forest regression (RFR) to construct an efficient evaluation method through segmented FDS curves. Furthermore, the characteristics of FDS curves were studied and the intelligent method was compared with support vector regression (SVR) and deep neural networks (DNN). The results show that the dielectric loss, the real part and imaginary part of complex capacitance all move upward with the moisture increasing, so they can be used as the input feature of the evaluation model; The moisture content evaluation accuracy of the RFR model in the whole frequency band is higher than that of SVR and DNN models; With the increase of lower cut off frequency (FDS test stop frequency), the FDS test time is greatly shortened, and the accuracy of the RFR model can still meet the evaluation requirements. Therefore, the data in a compromise frequency band can be used to evaluate the moisture content of oil paper insulation accurately and quickly. How to cite this article: Yao H, Mu H, Ding N, et al.: Evaluation method for moisture content of oil-paper insulation based on segmented frequency domain spectroscopy: From curve fitting to machine learning.

show abstract

Artificial neural network based method for temperature correction in FDS measurement of transformer insulation

Cited by 8 publications

References 30 publications

Precise Measurements of the Temperature-Frequency Dependence of the Conductivity of Cellulose—Insulating Oil—Water Nanoparticles Composite

Precise Measurements of the Temperature-Frequency Dependence of the Conductivity of Cellulose—Insulating Oil—Water Nanoparticles Composite

Using transfer-learning-based algorithms as data reduction strategies for volatile organic compounds classification using plasma spectroscopy

Evaluation method for moisture content of oil‐paper insulation based on segmented frequency domain spectroscopy: From curve fitting to machine learning

Contact Info

Product

Resources

About