Search

Gas-insulated switchgear (GIS) is one of the most important power devices in the power system. Once a fault occurs, it is difficult to determine and locate the fault due to its fully enclosed nature, which may lead to greater accidents and huge losses. Many GIS anomaly detection methods based on vibration and deep learning (DL) have been developed in recent years. However, these DL-based methods only utilize one single sampling point data on GIS, while not all faults are reflected at the points currently in use due to the signal attenuation. In this paper, an embedded position variational auto-encoder (EP-VAE) model for anomaly detection of GIS devices is presented, which can model multipoint vibration signals. We fuse the position information and the corresponding one dimension vibration signal to the two dimension grayscale images by Gram matrix. Then, the normal grayscale images are feed into the EP-VAE model, which is optimized by KL divergence and reconstruction error jointly. In the test phase, when one of these two losses is higher than the normal images, it is identified as an abnormal fault. We use the vibration data from the actual operation of GIS to carry out the anomaly detection experiments. The results show that our model EP-VAE has a detection accuracy of 93.12% for normal vibration grayscale images, and an abnormality detection accuracy of 87.94%.

At present, transformer protection usually operates after failures. With the raise of voltage level and capacity of transformer, a fault easily becomes an accident, resulting in huge economic losses. Transformer insulation failure is the main cause of transformer accidents. In this paper, a new idea of transformer protection based on insulation critical breakdown prediction is proposed. According to the requirements of transformer protection, a prediction method of oil-paper insulation critical breakdown under DC voltage based on partial discharge analysis is proposed by studying partial discharge characteristics of oil-paper insulation defects under DC voltage and the development process. The results show that when partial discharge critical breakdown occurs in oil-paper insulation under DC voltage, the distribution of partial discharge lagging recovery time is concentrated, and dense discharge clusters are easy to form, which can be used to predict the critical breakdown of oil-paper insulation.

Magnetic valve controllable reactor (MCR) is a kind of dynamic reactive power compensation device. The exciting current of MCR is one of the important indexes of operating condition of MCR. This paper firstly analyzed the principle of MCR, and deduced the relationship between the exciting current and the output current, verified by MATLAB/SIMULINK. And then an effective method to detect the exciting current was proposed by this paper. The method has the advantages of optical fiber isolation, high detection precision, strong anti-interference, etc. Finally, a 220v/1200var MCR was experimented in detail. The results demonstrate the accuracy of the theoretical analysis.

Magnetic controlled reactor (MCR) is a novel reactive power compensation device. MCR can supply continuous reactive power. But harmonic current is usually generated by MCR. The principle of harmonic current generation of MCR is analyzed in this paper. And then a method based on extension triangle structure MCR is proposed to suppress the particular order harmonic current. An additional winding is connected in series with the three phase windings. In the proposed method, no additional filters are needed and the control method is simple. At last, simulation is done to verify the correctness of this method.

Power transformer is one of the key electrical apparatus in power system, whose reliability has a directly correlation on the safety of power system. The concentration of dissolved gas in oil can be got by DGA online monitoring technology, and then the latent fault of transformer could be found in time. However, with the widespread use of on-line monitoring technology, the scale of monitoring data is also expanding, which is up to TB magnitude, forming on-line monitoring large data. At the same time, complex operating environment results in the emergence of a large number of invalid monitoring devices, which reduces the reliability of the monitoring system. Therefore, how to extract valid data from massive data and evaluate the monitoring device is an urgent problem to be solved. A validity assess method of DGA sensors in power transformer oil based on chaos theory is proposed in this paper. Firstly, the chaotic characteristics of oil chromatographic monitoring time series are qualitatively demonstrated. And the Lyapunov exponent is used as quantitative judgment. Then, according to the contraction characteristics of the chaotic phase space, the selection method of the optimal analysis data is given to solve the problem of excessive data. Finally, an assess method of DGA monitoring devices based on Lyapunov exponent and the optimal length selection is proposed. Verification test over a real case in the field shows that the method proposed in this paper is able to accurately evaluate the validity of DGA monitoring devices, whose detection rate of the failure monitoring device is up to 100%. Relative to traditional methods, the optimal length selection method proposed in this paper can effectively reduce the computational data, improve calculation efficiency and the efficiency of DGA monitoring devices validity assess method.

The reliability of the power system depends directly on the operating state of the transformer. As equipment manufacturing processes and maintenance levels are continuously improved, the number of machine downtimes caused by faults is significantly reduced. Transformer faults are dominated by overheating faults and internal discharges, and energy is released at the same time as the fault occurs. The insulation material of the transformer is decomposed and cracked in the above process. Alkanes are produced and dissolved in transformer oil. A large amount of data indicates that there is a correlation between the condition of the equipment and the dissolved gas content in the oil. The time sequences data of oil chromatography shows small-scale fluctuations in the local area. In order to reduce the impact of data fluctuation on the prediction model, this paper proposes a combination prediction method of power transformers based on the feature gas arrangement diagram and grey model. Compared with the traditional grey correlation analysis, this model can solve the model divergence caused by data fluctuations. In addition, a method based on the feature gas arrangement diagram is proposed, which corrects the deviation of the prediction result based on the actual operation state of the equipment. This method improves the efficiency of the genetic algorithm and allows the model to obtain more accurate predictions. In the actual case analysis, the model proposed in this paper obtains a good application effect, especially for small sample prediction nroblems with high accuracy.

The open-pit mine vehicle scheduling problems are discussed to assign the mine resource reasonably, adjusting the balance of supply and demand effectively and improving the transportation efficiently, and provide necessary technical support for complete open-pit operational scheduling measures. Meanwhile, we design natural number coding genetic algorithms with the technologies of adaptive adjustment of maximum reserved exchange, exchange rate and mutation rate, which achieves good performance in the experiments of solving transportation problems. The experimental results show the approach proposed in this article is feasible and effective, providing a novel technical measure for the optimization of open-pit mine vehicle scheduling.

This paper mainly discusses the mechanism for FRA method on bushing insulation detection. It builds several RC models of the bushing instruction in MATLAB, as well as does some simulation work to observe the change of the FRA curve with different parameters. It also takes the influence of the valve wire into consider and get some useful conclusions. Simulation results show that the obvious indentation in the FRA curve will appear when the bushing has some insulation defects. The sum and change state of the indentation in FRA curve is related to the degree and relaxation polarization of the insulation defects of the bushing. It is found that the simulation and on-spot detection curves have tallies with each other, which verifies the feasibility of FRA method in this field.