RESUMEN
Affected by the new coronavirus (COVID-19) pandemic, global energy production and consumption have changed a lot. It is unknown whether conventional short-term load forecasting methods based on single-task, single-region, and conventional indicators can accurately capture the load pattern during the COVID-19 and should be carefully studied. In this paper, we make the following contributions: 1) A mobility-optimized load forecasting method based on multi-task learning and long short-term memory network is innovatively proposed to alleviate the impact of the COVID-19 on short-term load forecasting. The incorporation of mobility data and data sharing layers potentially reduces the difficulty of capturing the load patterns and improves the generalization of the load forecasting models. 2) The real public data collected from multiple agencies and companies in the United States and European countries are used to conduct horizontal and vertical tests. These tests prove the failure of the conventional models and methods in the COVID-19 and demonstrate the high accuracy (error mostly less than 1%) and robustness of the proposed model. 3) The Shapley additive explanations technology based on game theory is innovatively introduced to improve the objectivity of the models. It visualizes that mobility indicators are of great help to the accurate load forecasting. Besides, the non-synchronous relationships between the indicators' correlations and contributions to the load have been proved.
RESUMEN
The distribution behavior of furfural in insulation systems is influenced by the oil-pressboard mass ratio. In addition, the equilibrium distribution of furfural between oil and pressboard will be disturbed after oil replacement. Therefore, it is of great significance to study the distribution ratio of furfural in oil with various oil-pressboard mass ratios after oil replacement. In this research, an accelerated thermal aging experiment and oil replacement experiment were conducted in the lab. Furthermore, the equilibrium characteristics of furfural dissolved in oil with various oil-pressboard mass ratios were studied. Multiple regression analysis was used to analyze the relationship between the oil-pressboard mass ratios and the distribution ratio of furfural in oil. The equilibrium distribution model of furfural was thus obtained. Afterwards, the modified furfural distribution model under oil replacement conditions was established. A novel scheme is provided for analyzing the equilibrium characteristics of furfural under various oil-pressboard mass ratios after oil replacement. The work of this paper is expected to improve the accuracy of furfural analysis.
RESUMEN
Frequency-domain spectroscopy (FDS) is demonstrated to be affected by electrode polarization and conductance behavior in the low-frequency ranges, which causes the unreliable prediction results of transformer cellulose insulation. In order to solve this issue, a fingerprint database based on the dielectric modulus is reported to predict the degree of polymerization (DP) and moisture content of cellulose insulation. In the current work, the relevant fingerprints that characterize the insulation conditions are obtained by studying the dielectric modulus curves of cellulose insulation with various insulation conditions, as well as the DC conductivity of transformer oil. Then, the dielectric modulus fingerprint database is established in the lab, and the accuracy of the reported fingerprint database is later verified. As a potential tool, the dielectric modulus fingerprint database is tested by several samples, and the results demonstrate that the accuracy of this method is more than 80%. In that respect, an interesting discovery of this paper is that the dielectric modulus fingerprint database may be a helpful tool for conditions prediction of the transformer cellulose insulation system.
RESUMEN
The support vector machine (SVM) combined with the genetic algorithm (GA) has been utilized for the fault diagnosis of transformers since its high accuracy. In addition to the fault diagnosis, the condition assessment of transformer oil-immersed insulation conveys the crucial engineering significance as well. However, the approaches for getting GA-SVM used to the moisture prediction of oil-immersed insulation have been rarely reported. In view of this issue, this paper pioneers the application of GA-SVM and frequency domain spectroscopy (FDS) to realize the moisture prediction of transformer oil-immersed insulation. In the present work, a method of constructing a GA-SVM multi-classifier for moisture diagnosis based on the fitting analysis model is firstly reported. Then, the feasibility and reliability of the reported method are proved by employing the laboratory and field test experiments. The experimental results indicate that the reported prediction model might be serviced as a potential tool for the moisture prediction of transformer oil-immersed polymer insulation.
RESUMEN
The aging kinetics model is of great interest to scholars since it is capable of describing the variation law between the degree of polymerization (DP) and the aging duration of transformer polymer (cellulose) insulation. However, it is difficult to determine the moisture content inside the transformer polymer insulation without destroying it, so that the model parameters cannot be confirmed. Such limitation greatly restricts its application. It is interesting to note that as long as the moisture content of the transformer polymer insulation could be characterized (replaced) by a certain feature parameter, the above issue will be solved naturally. The existing researches indicate that the Frequency Domain Spectroscopy (FDS) is sensitive to moisture. Consequently, the feature parameter that could characterize the moisture inside transformer polymer insulation (extracted from the FDS curve) can be used to report a modified aging kinetics model, which could perform the aging condition prediction of transformer polymer insulation under various test conditions, including aging duration, aging temperature, and initial moisture. In that respect, the average relative error of prediction results of prepared samples is equal to 7.41%, which reveals that the reported model might be serviced as a potential tool for the aging condition prediction of transformer polymer insulation.
RESUMEN
It is a fact that the frequency domain spectroscopy (FDS) curve at different temperatures can be corrected by the shift factor (αT) extracted from the master curve. However, the αT and master curve reported by previous works are distinctive due to the difference in the construction algorithm. Therefore, it is of great significance to report a universal approach for extracting αT. In this work, the unaged oil-immersed pressboards with different moisture content (mc%) are firstly prepared and selected as the research specimen. Then, the αT of FDS curves on the above pressboard is extracted based upon the master curve technique. The influence mechanism under the various test temperature (T) and mc% is therefore analyzed so as to establish a universal model for predicting the αT. The present findings reveal that the αT value extracted from FDS curves is both temperature-dependent and moisture-dependent. In addition, the predicted αT is not only suitable for temperature correction on FDS curve of same type pressboard with different insulation conditions (moisture contents and aging degrees), but also maintains considerable accuracy when applied to different types of pressboard. Therefore, the obtained conclusions will provide a universal method for temperature correction on FDS curve of transformer polymer insulation.
