Search

The leakage currents are appropriate for determining the contamination level of insulators in the power distribution system, which are efficiently cleaned or replaced during the maintenance schedule. In this research, the hybrid convolution neural network and gated recurrent unit model (CNN-GRU) are developed to categorize the leakage current pulse of the 15 kV HDPE insulator in the transmission towers in Taiwan. Many weather parameters are accumulated in the online monitoring system, which is installed in different transmission towers in coastal areas that suffer from heavy pollution. The Pearson correlation matrix is computed for selecting the high correlative features with the leakage current. Hyperparameter optimization is employed to decide the enhancing framework of the CNN-GRU methodology. The performance of the CNN-GRU is completely analyzed with other deep learning algorithms, which comprise the GRU, bidirectional GRU, LSTM, and bidirectional LSTM. The developed CNN-GRU acquired the most remarkable improvements of 79.48% CRE, 83.54% validating CRE, 14.14% CP, 20.89% validating CP, 66.24% MAE, 63.59% validating MAE, 73.24% MSE, and 71.59% validating MSE benchmarks compared with other methodologies. Therefore, the hybrid CNN-GRU methodology provides comprehensive information about the contamination degrees of insulator surfaces derived from the property of leakage currents. [ABSTRACT FROM AUTHOR]

This study proposes a new superior hybrid algorithm, which is the particle swarm optimization (PSO) and gene algorithm (GA)-based neural network to predict the leakage current of insulators. The developed algorithm was utilized for the online monitoring systems, which were completely installed on the 69 kV and 161 kV transmission towers in Taiwan. This hybrid algorithm utilizes the local meteorological data as input parameters combined with the extracted enhanced data: the percentage of spark discharge areas and the brightness change in the image of the discharge phenomenon. These data with a high correlation with the leakage current are utilized as input vectors to improve the accuracy and effectiveness of the developed hybrid model. The performance of the developed algorithm is compared with a traditional PSO-based neural network and backpropagation neural network (BPNN) to evaluate and analyze. The comparative simulation results prove the effectiveness of the combination of hybrid PSO-GA-based neural network and surface discharge data, which achieved a maximum improvement of 38.54% MSE, 10.62% MAPE, and 3.41% R square for 161 kV data and 39.28% MSE, 12.62% MAPE, and 1.61% R square for 69 kV data. Moreover, the data with enhanced inputs outperform the traditional data in most benchmark factors, improving the accuracy and effectiveness in defining the deteriorative insulators. The developed methodology with a noticeable improvement was utilized in the online monitoring system to reduce the operational and maintenance cost of transmission lines in Taiwan Power Company. [ABSTRACT FROM AUTHOR]

An inter-metal dielectric (IMD) is deposited between metal layers to provide isolation capability to a device and separate the different metal layers that are unnecessary in the conduction of electricity. Owing to the complicated input/response relationships of the IMD process, the void problem results in electric leakage and causes wafer scraping. In the current study, we combined neural networks, genetic algorithms (GAs) and the desirability function in order to optimise the parameter settings of the IMD process. Initially, a backpropagation (BP) neural network was developed to map the complex non-linear relationship between the process parameters and the corresponding responses. Moreover, the desirability function and GAs were employed to obtain the optimum operating parameters in respect to each response. The implementation of the proposed approach was carried out in a semiconductor manufacturing company in Taiwan, and the results illustrate the practicability of the proposed approach. [ABSTRACT FROM AUTHOR]

• To develop hybrid methodology CNN-LSTM architecture to predict the 15kV insulator leakage current, which has not been serious investigated. • To employ the hyperparameter optimization to construct the most appropriate architecture of CNN-LSTM. • To analyze and evaluate the performances of hybrid CNN-LSTM with other deep learning LSTM, Bi-LSTM methods with K-fold cross-validation. • An online monitoring system was developed with TensorFlow AI library on the web-based service. The paper introduced a deep learning algorithm, convolutional neural network Long Short-Term Memory (CNN-LSTM), which has yet to be deeply researched for classifying the leakage/discharge current on the web-based service. The four data models with the leakage current surge have been collected hourly for more than 14 months in Yunlin County, Taiwan, which was seriously affected by the salt-for pollution. The weather data, which are temperature, humidity, dewpoint, wind direction, wind speed, air pressure, rainfall, and solar illuminance, were utilized as input parameters. Consequently, the Convolution Neural networks' long short-term memory has been developed to predict the leakage current classification for one-hour intervals. Moreover, the hyperparameter optimization will enhance the proposed model's performance and accuracy by varying configurations. The optimized structure of the Convolution neural network Long short-term memory is also compared with other persistent models, the traditional Long Short-term memory and Bidirectional long-short term memory, to evaluate the performance. The simulation results proved that the optimized convolutional neural network long short-term memory is an appropriate neural network to predict the leakage current classification with a maximum improvement of 54.940% and 82.536% category error; 16.184% and 23.287% accuracy; 27.787% and 27.608% precision in training and validating data, respectively. [ABSTRACT FROM AUTHOR]

• to develop an optimized Bi-GRU architecture to categorize the 15 kV HDPE insulator leakage current, which has not been seriously researched before; • to utilize the hyperparameter optimization to construct the most appropriate architecture of Bi-GRU and evaluate the effectiveness of changing a setting parameter on the accurate performance; • the performances of Bi-GRU are compared and evaluated with other deep learning LSTM, GRU, and RNN methods with K-fold cross-validation; • A correlation matrix of all collected weather features and target output is calculated and illustrated as a summarizing diagram for advanced analyses of 15 kV HDPE insulators. Cleaning the polluted insulators is the most efficient way to enhance distribution lines' reliability and reduce contamination flashovers. The measurement of leakage current quantity is a critical issue that can be implemented. Many previous research and laboratory experiments prove that the leakage current levels could fully reflect the contamination developments. To predict the contamination levels of insulators and prevent flashover accidents, the multilevel categorizing leakage current system is installed to simulate the impact of different weather factors. In this project, the leakage currents of 15 kV HDPE insulators accumulate with varying weather elements in coastal areas of Taiwan. The Bidirectional Gated Recurrent Unit (Bi-GRU) is presented to multilevel categorize the leakage currents, which are utilized to evaluate the contamination flashovers. The hyperparameter optimization is employed to develop the most optimum deep learning Bi-GRU model architecture. Moreover, the effectiveness and stabilization of Bi-GRU are investigated and analyzed with other deep learning methodologies, which include the recurrent neural network (RNN), the long short-term memory (LSTM), and the gated recurrent unit (GRU). The comparison performance demonstrates that the addition of bidirectional layers in the GRU method has outperformed the LSTM algorithm, which improves 5.27 %, 14.03 %, 0.35 %, and 0.06 % in category cross-entropy (CCE) accurate benchmarks in the training and testing operations. The prediction of multilevel leakage current could estimate the safety degree of HDPE insulators before the development of pollution flashover accidents, and the cleaning or replacing maintenance services are operated on time. [ABSTRACT FROM AUTHOR]

• Predicting the leakage current level for 15 kV and 25 kV distribution insulators which applied the convolution neural network bidirectional long short-term memory (CNN-Bi-LSTM) is utilized. • The hyperparameter optimization was employed to optimize the structure of CNN-Bi-LSTM. • The performance of CNN-Bi-LSTM was compared with the GRU, LSTM, CNN-LSTM and Bidirectional-LSTM. In this project, the online monitoring systems to classify the leakage current of 15 kV and 25 kV distribution insulators were installed in mountain areas in Taiwan. The convolution neural network bidirectional Long Short-Term Memory (CNN-Bi-LSTM) was employed to categorize the leakage current levels based on the field experiences of the TaiPower company. The sequential weather factors and the different distribution insulators leakage current were collected every hour and utilized in the proposed model's learning process. In addition, a simple grid search parameter optimization was implemented to identify the optimum architecture of CNN-Bi-LSTM for each data collection. Moreover, the optimized CNN-Bi-LSTM structure performances were evaluated and compared with the gated recurrent unit (GRU), LSTM, CNN-LSTM, and Bi-LSTM neural networks. The experiments results proved that the proposed CNN-Bi-LSTM improved the category cross-entropy error, the accuracy, and the precision metric with a maximum enhancement of 71.331%, 14.382%, and 14.042% in the training data and 91.250%, 17.529%, and 17.343% in the validating data, respectively. Moreover, the experiment outcomes in different data models prove that the 24-sequential weather parameters have a significant impact on the achievement of CNN-Bi-LSTM in categorizing the leakage current. The optimized CNN-Bi-LSTM structures were applied in the online monitoring systems to provide a better schedule for maintenance operations at the Taipower company. [ABSTRACT FROM AUTHOR]

This article presents the effects of different grounding strategies (ungrounded, direct grounded, and diode grounded), bonding, and insulation on rail potential and stray currents in the Taipei rapid transit systems (TRTSs). The TRTS is a direct current transit railway and its running rails are used as the return conductor for traction currents. The advantage is that no dedicated return conductor is required, whereas the disadvantages are rail potential and stray-current problems. Thus, the analysis of grounding strategies is important and necessary. Moreover, the effects of cross-bonding of the running rails and stray-current collector cables (traction earth conductors) are investigated since bonding is usually used to balance traction negative return currents, to reduce resistances of the negative return circuit, aswell as to decrease rail potential. Meanwhile, the effects of insulation on stray currents are also described. Simulation results for the Blue line are presented in TRTS. [ABSTRACT FROM AUTHOR]

Taipei Metro adopted the diode-grounded scheme for stray current collection in construction of its cross rail network. During operation of the network, a high rail-to-earth potential (Vrail) has been observed at the east end of the Blue Line (i.e., stations BL13-BL16). To find effective countermeasures, a series of field tests in a step-by-step development nature was conducted from 1999-2000, which led to the decision of disconnecting the impedance bond at Gil of the tie line so that the negative return current of the Blue Line cannot flow to the rails of the Red-Green Line, and vice versa (detailed in Sections 111-A and V-C). This decision was implemented through contract-out work in 2003. Since then, the Vrail has been lowered by almost half before disconnection. To gain the insight characteristic before the contract-out, numerical simulations were also conducted by simulating the multi-train and multisection features of the cross transportation network. The simulation results (for Vrail and stray current, or Istray) were consistent with the field-test results. This paper presents the design of these field tests and their test results in comparison with the simulation results, based on which the countermeasures for reducing Vrail and the present status after Vrail reduction at Taipei Metro are presented. [ABSTRACT FROM AUTHOR]