Spatial resolution of space charge distribution measurement using PEA (Pulsed electro-acoustic) method at high temperature was improved by applying a thin piezo-electric sensor film with a glass substrate. Partial discharge generation is one of problems to be solved in motor winding when the driving voltage of motor is elevated to downsize it. In the high-voltage drive of motors, the space charge accumulation in the thin insulating layer of the winding is said to affect the partial discharge generation. Therefore, it is necessary to evaluate the space charge accumulation in thin insulating materials of the winding. Especially, since high temperature characteristics of the partial discharge is said to be important, the measurement of the space charge distribution for the thin insulating layer is required at high temperature. However, the spatial resolution of the ordinary PEA measurement system was not enough to measure the space charge distribution in such thin insulating materials for windings, especially at high temperature. Therefore, the PEA measurement system was improved using a thin sensor film, which is available at a relatively high temperature, as a sensor of the PEA measurement system with a high heat-resistant glass substrate. In this report, a typical measurement result of the space charge distribution observed in 25 µm-thick polyimide sample at 80°C is introduced.
List of Conference Committee Chairs, Program Committee Chairs and International Technical Committees are available in this pdf.
Renewable energy-based power systems will experience a larger number of harmonics in voltages and currents due to the increased use of converters, hybrid lines, and non-linear loads in the future. These harmonics could lead to additional power and heat losses, increase operational expenses and affect the life expectancy of components. This study aims at analyzing the increased losses and temperature rise in a 25kVA oil-filled distribution transformer under well-defined harmonics. The data obtained are then used to calculate the power losses for odd and even harmonic contents and a comparison is made between the two cases. The results contribute to a better understanding of the power losses for different harmonic contents. A direct correlation between theoretical and experimental results is established, showing the negative impact of harmonics on power losses and the corresponding temperature rise in a distribution transformer.
The dissolved gas analysis (DGA) is one of the main diagnostic techniques for oil-filled electrical equipment. Though self-contained fluid filled (SCFF) cables have been maintained by DGA, some faults occurred on SCFF cable joints in Japan. Therefore, it is important to improve DGA for SCFF cables. In this viewpoint, the characteristics of decomposition gases by partial discharges in the insulation system for SCFF cables were experimentally investigated. The insulation layer of test samples was composed of sheets of oil-impregnated insulating paper and an oil gap that simulated a butt gap in SCFF cables. Partial discharges were generated in the oil gap in short or long periods, and the oil gap and light emission due to partial discharges were observed. In addition, after partial discharge generation, the insulating oil in the oil gap was taken and hydrogen and hydrocarbon gases in the insulating oil were detected by DGA. As the result, in the short period case, partial discharges occurred through insulating oil, and main decomposition gases were H-2 and C2H2. On the other hand, in the long period case, partial discharges occurred in bubbles, and main decomposition gases were H-2, CH4 and C2H6. It was estimated from this result that the difference of the main decomposition gases between two cases depended on whether partial discharges mainly occurred through insulating oil or in bubble.
Dielectrically graded insulation (DGI) is a potential approach in next-generation electric insulation systems due to its effective electric field (E-field) regulation, high insulation performance and reduced geometrical size. However, it is challenging to find a construction method of DGI with facile procedures, high accuracy, high flexibility, and good scalability. Fortunately, the rapid developing of additive manufacturing (also known as 3D printing) provides a novel approach – lattice material – for DGI construction. The lattice material is made of ordered, periodical cellular structure (unit cell), in which the feature size or topological structure of every unit cell can be tailored to build non-uniform, functionally graded objects. Since the dielectric properties of almost all the insulation materials is strongly related to its internal structure, it is reasonable to build FGM insulation by lattice materials. In this paper, the concept of lattice-material DGI (LM-DGI) is proposed, which is based on the description and analysis of research progress in DGI. Finite element simulation is then conducted on a 2D example to verify its effectiveness in E-field regulation. The results indicate that despite certain amount of fluctuation, the LM-DGI still has adequate performance in electric field regulation. Moreover, the distribution of electric field can be flexibly modified by changing the feature size (e.g. hole radius) of unit cells. We believe that LM-DGI is promising for facile, accurate and large-scale DGI insulations, making it ideal for industrial usage.
TEPCO Power Grid owns about 15,000 km of overhead transmission lines. When conducting inspections, helicopters are used to film certain portions of overhead transmission lines in mountainous areas which cannot be otherwise easily confirmed by skilled workers. The workers subsequently check and review such portions in their office via VTRs. This involves checking the soundness of the overhead transmission lines via slow-motion reproduction and is unfortunately exceptionally time-consuming work. Accordingly, we developed an overhead transmission line VTR diagnostic system by combining helicopter-based transmission line inspection technology and AI (deep learning) technology accumulated to date. Based on the judgment rate achieved by failure diagnosis AI, we see potential to further streamlining conventional inspection methods using this system. Previously, around 1,330 hours a year were needed for inspection work, which involved playing back the filmed VTR in slow motion and checking for failure in conductors. However, using this system, the work is expected to take about 660 hours a year, boosting efficiency by about 50%.
This paper gives examples of the application of condition monitoring to large power transformers. In many reported cases of condition monitoring the analysis of data, conclusions drawn, and decisions made are fairly straightforward. We show such a ‘straightforward’ case relating to traditional dissolved gas analysis (DGA), but we also show cases where more analysis of the context of the main tank partial discharge (PD) measurements and bushing power factor measurements are needed to draw appropriate and actionable conclusions. Condition monitoring may provide significant benefits if data is evaluated in context, and if it is considered as part of, and embedded in, the operational and asset management processes of the organization in order to ensure positive outcomes.
We have developed a continuous remote monitoring system that can monitor partial discharge (PD) trends for a long period of time, which greatly saves the labor in operating. An Edge terminal uploads a signal obtained from the high frequency CT attached to the power cable to a cloud system. In this process, a PD-like signal is effectively discriminated from a large amount of acquired data by using a noise filtering function. The signal data uploaded on the cloud system is PD-judged by AI, and a learning cycle function improves the AI judgment accuracy.
For high-voltage rotating machines used in power plants, partial discharge (PD) monitoring is performed during maintenance and inspection periods. However, nowadays, an online PD diagnostic system is generally employed as the tool to provide early warning of an incipient fault in high-voltage stator windings. In this study, an algorithm is described that can classify the defect types of stator windings. The classification process is based on analysis of the phase-resolved PD spectrum obtained by a capacitive slot coupler (CSC) sensor inserted between the stator core and bottom of the stator winding. The first technique to utilize the CSC sensor is the application of sweep frequency response technique (SFRA) to monitor the degradation level of interturn insulation. The second is the adoption of machine learning technique to classify the type of insulation defect in stator windings. After reviewing several artificial intelligence techniques, a multiclass SVM was selected as the main classification algorithm, and six defect types of stator winding were trained as multiple classes and then tested. Based on the Python Scikit-learn library, both polynomial-type and radial basis function (RBF) type kernels were utilized for the SVM analysis by changing the dimensional level to find the optimal hyperplane. In the RBF kernel method, the optimal classification accuracy was derived by varying the gamma and cost parameters. Finally, the classification accuracy was 90% for the polynomial kernel and 94% for the RBF kernel. This classification algorithm will be adopted in the novel on-line diagnosis unit as an initial step toward the digital power plant.
One of the most popular insulating liquids is mineral oil, which is used in many high-voltage apparatuses, especially in transformers. For a few decades, the concept of the use of alternative insulating liquids has been presented. Natural ester is a vegetable-based insulating liquid. Pressboard is a type of electrical insulation paper made of cellulose. Frequency Domain Spectroscopy (FDS) has been used to assess the insulation characteristics such as oil-impregnated paper. In this paper, the FDS of pressboard impregnated with natural ester under thermal stress and pressboard impregnated with natural ester were studied. In the experiment, the pressboard test specimens were heated at 100 degrees Celsius in a vacuum oven for 48 hours to reduce the moisture. Then, the test specimens were impregnated with natural ester for 0, 8, 16, and 24 hours. After that, they were simulated with thermal stress by heating at 150 degrees Celsius in a vacuum oven for 0, 30, and 90 days. To analyze their dielectric integrity, FDS was performed by DIRANA. The test cell for solid insulation testing, according to JIS C2111 was prepared and used for the FDS experiment. The result shows that the characteristics of the natural ester-impregnated pressboard that had been subjected to thermal stress were slightly different when compared with the same impregnation time of the pressboard.
OF cables of Korea have been installed since the 1970s, and the maximum operating years among OF cables in the field is estimated to be 45 years. In this paper, it was proposed to determine the replacement priority using the health index, by which we can quantitatively check the condition of the cable. In order to consider the importance of the line as well as the condition of the cable, a Monte Carlo Simulation was conducted when the operating years were 40 or 50 years using a risk matrix.
The transformer bushing failure occupies an over 30% or more of the transformer failure and it is categorized into a severe fault that can occur in a transformer because it causes a fire accident with a high probability of failure. With respect to a bushing failure prevention, a leakage current measuring using bushing tap adapter is adopted to diagnose the insulation condition of bushing. However, there are some problem that the conventional method can be determined only when a certain or more deterioration proceeds, and detection is difficult at the initial stage of the bushing defect. This paper deals with the development of bushing diagnosis system based on the high frequency (HF) partial discharge measurement technology. Based on the analysis result on the types of early-stage of bushing defect, the fault type that can occur in the transformer bushing was classified. In addition, the diagnostic system and new type of bushing sensor based on high-frequency partial discharge technology was developed. The performance verification of developed system was performed based on IEC-60270 partial discharge test.
Power transformer diagnostics and condition assessment are essential measures to increase the transformer's lifetime. In this regard, the dissolved gas analysis (DGA), the most common method worldwide in predicting the incipient faults in the transformer, is applied. Recently, the optical spectroscopy methods have attracted most of the attention towards effective monitoring of the state of power transformer insulation. Therefore, in this paper, Fourier Transform Infrared (FTIR) spectroscopy was employed to discriminate between the electrical and thermal faults that frequently happen in oil insulation. Regarding the electrical fault, it was simulated by applying repeated high voltage impulses to a fresh oil sample while the thermal fault was imitated through a localized overheating in another fresh oil sample with the help of a heating coil. Firstly, the two samples were examined via FTIR spectroscopy to obtain their FTIR spectra. Then, the effect of ageing on the spectrum was studied comprehensively for determining the degree of deterioration as well as any change that happen in the concentration or structure of the oil molecules. After that, both of them were analyzed by DGA to measure the concentration of the dissolved gases in the oil in order to confirm the FTIR spectroscopy results. In the final analysis, it was obvious that the implementation of the optical method is considered a promising tool to monitor the faulted oil and distinguish between the electrical fault and the thermal one making the FTIR spectroscopy a superior alternative for DGA.
Accurate and efficient maintenance inspections are essential for the rational replacement of stator windings of aged hydropower generators. In practice, the accuracy of maintenance inspection depends on engineers' experience. The shortage of senior engineers due to the aging population is a big problem, and technology transfer has been an essential concern in Japan in recent years. Various industries are trying to introduce systems for studying large amounts of maintenance and inspection data using machine learning technology to detect problems of apparatuses automatically. Machine learning technology can also be applied to the maintenance and inspection of stator windings to help engineers determine the maintenance policy. In this study, a database is developed; the database contains partial-discharge data measured at stator coils, which simulate winding problems in real stators.
Electric power demand (EPD) varies depending on weather conditions, social activity, industrialization, and so on. For reliable operation of the electric power system, including renewable energy (RE) sources, it is necessary to monitor the EPD for generation and weather conditions to forecast the EPD as well as the generation of the RE sources. Since there were few studies on EPD changes over time or with weather conditions in Mongolia, it is necessary to clarify the change in EPD and the relationship between the EPD and temperature that is representative of the weather condition for future monitoring of the EPD and the weather condition. The authors are also interested in identifying differences in EPD change between Mongolia and Japanese characteristics. As a result, the linear relation between EPD and temperature was seen in Mongolia, while “U-shape” characteristics in Japan. Also, differences in electricity usage between summer and winter seasons were clearly seen as the term for premature dependence.
This paper investigates the influences of cellulose bridge formation on the Lightning Impulse Breakdown Voltage (LIBDV), electrical filed strength and thermal generated of Palm Fatty Acid Ester (PFAE). A standard lightning impulse voltage (SLIV) waveform of 1.2/50 µs with DC superimposed was set up in accordance to IEC 60230. The commercial cellulose powder were dispersed into PFAE as contaminated by-product to replicate a bridge skeleton with concentrations of 0.004%, 0.008% and 0.012% by weight. The rising-voltage method was used to measure the LIBDV in accordance with IEC 60897 test method. Weibull cumulative breakdown probability to present breakdown results statistically and the results was compared with clean oil (CO). Electrical field strength was simulated using Finite Element Analysis (FEA) and thermal profile was obtained via numerical calculation. The breakdown results showed that the cellulosic particles at 0.004 wt% reduced the LIBDV by 16%. The influence of cellulose concentration on LIBDV become more prominent during bridge formation, with a further reduction for 29% and 31% at 0.008 wt% and 0.012 wt% respectively. The electrical field strength along the bridge lines is dominant factor for the breakdown occurrence, subsequently the cellulose contamination; contributing up to 32% to the instantaneous breakdown of the PFAE. Moreover, the $L_{ITG}$ indicated that, the prominent effect of thermal evacuation process influence by concentration level, where the cellulose accumulation acts as a heat scavenger in electrically stressed by inhibited the thermal to conduct easily by 21%, thus reducing the thermal performance of the oil. Therefore, the finding potentially contributes to the formulation of guidelines for condition assessment and contamination monitoring to minimize common issues in power transformer failures that are attributable to the insulation system, and consequently, achieve optimum transformer insulation integrity by extending HVDC converter transformer life span.
For obtaining knowledge about the changes in electrical insulation behavior of a cable during the normal operation of a nuclear power plant (NPP), two cables insulated with the same flame-retardant ethylene-propylene-diene rubber were removed from two NPPs. A new cable insulated with the same rubber was also subjected to the experiment as it is and after a severe aging treatment. It has become clear that the new cable and the two removed ones have the highest and lowest electrical conductivity, respectively, among these four cables. The suppression of the carrier transport is likely caused by the hardening of the rubber induced by the heating and irradiation with gamma rays. This result, in turn, indicates that no serious concerns are needed about the degradation of insulation as long as similar cables are used in similar environments. The severity of two artificial aging treatments to simulate two types of serious accidents in NPP is also discussed.
Medium Voltage (MV) electrical assets are no longer just distribution networks, since also electrified transportation is moving towards kV, if not tens on kV, of target operating voltage and renewable power generation, being largely distributed, is also often delivered through MV links. A characteristic aspect of the evolution of MV assets is the type of power supply, which is moving towards a hybrid concept: voltage can be DC or modulated AC, that includes voltage transients, ripple, and harmonics. As regards insulation systems, however, this trend can be fatal for life and reliability, since the type of stress to be withstood, mostly electrical and thermal (but also mechanical and environmental, for example in aerospace or electric vehicles), could be far beyond the reference used for decades for the design of electrical insulation, which has been based on AC sinusoidal electrical stress. The implication is that even good and conservative design could not work anymore to avoid premature breakdown, due both intrinsic and, especially, extrinsic accelerated aging. Hence, besides relying upon the best available design practice, monitoring the health conditions of MV asset components is becoming crucial to get grid/asset working effectively for the whole planned operation life. However, measuring/monitoring diagnostic quantities often relies upon expert interpretation of diagnostic monitoring results, which for matters of cost, data amount and timing, is not feasible for MV assets on a large scale. This paper has the purpose to discuss the feasibility of an innovative, automatic approach to PD monitoring which has the potentiality to allow MV asset components to self-diagnose their health conditions and interact smartly with asset and maintenance managers. The fundamentals of the algorithms developed for such a smart component and examples of its application on electrical and electronics asset components (as PCB and spacers) are presented in the paper.
The partial discharge measurement in the cable is usually carried out by electrically connecting the detector at the terminal or insulation joint. The system has a complicated transfer function from the discharge point to the measurement point. The observed waveform is a convolution of the transfer function and the partial discharge waveform at the discharge point, resulting in a distorted waveform. On the other hand, the waveform characteristics of the partial discharge reflect the electron avalanche process in the discharge space. It may be possible to estimate the condition of defects based on the waveform characteristics. In this study, we constructed the mock model of the insulation joint and reconstructed the partial discharge waveform at the discharge point by deconvolution processing. To obtain the transfer function, it is necessary to input a known pulse and analyze its response. In a cable at a site, it is difficult to input the pulses because the sheathed terminals of the insulation joint are insulated with an insulation coating to ensure system reliability and safety. Therefore, the transfer function was obtained by measuring the impedance from the detector side and analyzing the equivalent circuit. The validity of the transfer function was shown in comparison with the transfer function when directly inputting the pulse to the discharge point. The characteristics of the reconstructed waveform were almost the same regardless of the acquisition method of the transfer function. Furthermore, the difference in the discharge points simulated deterioration can be found based on the characteristics of the reconstructed waveform.
In this study, the discharge characteristics of an insulator creepage surface with a back electrode under various humidity conditions were experimentally investigated. Partial discharge (without sparks) was detected at a lower voltage than that of the arc discharge (short-circuit failure with sparks). The partial discharge inception voltage (PDIV) under the DC voltage was increased at high humidity. Meanwhile, the PDIV under the AC voltage was decreased at high humidity and was lower than that under the DC voltage. This finding indicates that the amount of charge on the creepage surface induced by the applied voltage increased under the high-humidity condition: The surface charge relaxes the applied electric field under the DC voltage, whereas it enhances the applied electric field under the AC voltage due to the delay in the surface charge distribution against the voltage polarity change.
