The water content in oil is closely related to the deterioration performance of an insulation system, and accurate prediction of water content in oil is important for the stability and security level of power systems. A novel method of measuring water content in transformer oil using multi frequency ultrasonic with a back propagation neural network that was optimized by principal component analysis and genetic algorithm (PCA-GA-BPNN), is reported in this paper. 160 oil samples of different water content were investigated using the multi frequency ultrasonic detection technology. Then the multi frequency ultrasonic data were preprocessed using principal component analysis (PCA), which was implemented to obtain main principal components containing 95% of original information. After that, a genetic algorithm (GA) was incorporated to optimize the parameters for a back propagation neural network (BPNN), including the weight and threshold. Finally, the BPNN model with the optimized parameters was trained with a random 150 sets of pretreatment data, and the generalization ability of the model was tested with the remaining 10 sets. The mean squared error of the test sets was 8.65 × 10⁻⁵, with a correlation coefficient of 0.98. Results show that the developed PCA-GA-BPNN model is robust and enables accurate prediction of a water content in transformer oil using multi frequency ultrasonic technology.

Transformer oil is an important insulating material in power transformers. The detection of transformer oil is an important means to ensure the normal operation of the power system, which is recognized by the public. Interfacial tension (IFT) is an important parameter to characterize the flow properties of liquids. Too low interfacial tension of transformer oil can cause major accidents in transformers. Therefore, it is of great practical significance to achieve effective detection of interfacial tension. This work proposes the measurement of the IFT of transformer oil using multi frequency ultrasonic and a support vector machine that was optimized by particle swarm optimization algorithm (PSO-SVM). 210 samples, which were detected by multi frequency ultrasonic and the Ring Method, were divided into training sets (200 samples) and test sets (10 samples). Then multi-frequency ultrasonic data is subjected to MDS dimensionality reduction to obtain 11 dimensional low-dimensional data which as the input of SVM, and the IFT, which was detected by the Ring Method, as the output of SVM. After that, a particle swarm optimization algorithm (PSO) was incorporated to optimize the parameters (C, g) for a support vector machine (SVM).Then, the SVM model with the optimized parameters (g=1.356 and C=2.831) was trained with the training sets, and the model was verified with the test sets. Results show that the model, which proposed in this article, that describes the nonlinear relationship between multi frequency ultrasonic data and the interfacial tension of transformer oil shows higher accuracy.

Timely identification of the aging stage of the transformer can effectively ensure the safe operation of the transformer and prevent the aging fault of the transformer. This work proposes the aging diagnosis of transformer using the Raman spectroscopic recurrence plot combined with learning vector quantization (LVQ) neural network. Oil-paper insulation samples at different aging stages are obtained by accelerated thermal aging tests. Acquisition of Raman spectra of insulating oil is from 800 to 3000 cm ⁻¹ and converts to the Raman spectroscopic recurrence plot. The gray-level cooccurrence matrix (GLCM) is used to extract the characteristic information of the Raman spectroscopic recurrence plot. The optimal number of neurons in competitive layer of LVQ neural network is determined by the experimental method. A transformer aging diagnosis model based on the Raman spectroscopic recurrence plot is established. The results show that the diagnostic accuracy of the proposed diagnostic model reached 95%. The Raman spectroscopic recurrence plot has more abundant aging “fingerprint” characteristic information and the proposed diagnostic method shows potential for monitoring oil-paper insulation.

In order to achieve accurate quantitative evaluation of the ageing state of oil paper insulation, an evaluation method of polymerisation degree (DP) of oil–paper insulation based on Raman spectroscopic plot is presented. Samples covering the full ageing stage of oil–paper insulation were obtained by accelerated thermal ageing experiments. The samples were tested by Raman spectra, and the Raman spectra were converted into Raman spectroscopic plot using Gramian angular field. Feature extraction was performed on the images using local binary pattern (LBP), and 28 feature variables were selected by competitive adaptive reweighted sampling and uninformative variable elimination. Finally, the DP prediction model was constructed by long short‐term memory neural network (LSTM). The training accuracy exceeded 90%, and the mean absolute deviation of DP prediction for 45 test samples was only 40.70. A new visual detection method for transformer ageing diagnosis is presented.

Acetone is a reliable index to evaluate the aging state of oil paper insulation. It is of great significance to study a fast, highly sensitive and accurate method of detecting acetone in oil to ensure the safe and stable operation of oil-immersed transformers. In this paper, a method based on surface enhanced Raman spectroscopy for the detection of acetone in oil is investigated. Based on the mechanism of surface-enhanced Raman spectroscopy (SERS), a simulation model of surface-enhanced substrate was constructed using COMSOL software, and it was determined that the enhancement effect was better when silver nanowires/zinc oxide nanorods (Ag/ZnO) composite structure was selected. Ag/ZnO substrates were prepared by solvothermal method and experimentally tuned and characterized to obtain SERS substrates with high enhancement properties. The substrate coupled partial least square (PLS) model was used to establish a method for the detection of acetone in insulating oils, and its limit of quantification for acetone was 0.003 mg/g, which meets the requirements for engineering testing.