* If you want to update the article please login/register
The use of artificial intelligence technologies in the control domain has risen in importance in recent years, mainly due to the incorporation of human cognitive skills and improved capability to deal with the modeling and parameter uncertainty. The proposed PFTC architecture was based on the Feedforward Backpropagation Neural Network and the FFBPNN's preparation, and we used two key metrics normalized mean and residue data that can be generated from nonlinear level system model data.
Source link: https://doi.org/10.1007/978-3-031-16038-7_23
The growing demand for electricity has resulted in the adoption of electrical microgrids in power systems. Microgrids have many benefits for power plants, but their security is one of the most difficult problems in power systems. In MATLAB Simulink, a three-phase AC grid is modelled for standard and fault situations. As the speed of simulations in MATLAB differs from the central processing unit of PC, the same Simulink model is then implemented in OP4510 RT-LAB Simulator to compare the real time results with MATLAB results.
Source link: https://doi.org/10.1007/978-981-19-3590-9_25
It's unavoidable that intense earthquakes occur in cities located near active faults. Surface fault rupture is one of the many contributing characteristics of near-fault earthquakes that have caused significant loss of life and extensive structural damage. This characteristic is generally denied in most of the seismic design of buildings; in other specifications, fault avoidance zones are supposed to cope with surface rupture, but these zones are usually inadequate. Using numerical methods, the effect of foundation stiffness on the surface fault rupture path is investigated in this study. Through simulation of two different soil mediums, the potential effects of soil properties are considered. In addition, the mat foundations with various dimensional characteristics are estimated considering soil-foundation interactions. According to this FE report, the foundation stiffness has a significant effect on the rupture course; increasing the thickness or reducing the length of the foundation contributes to fault rupture path.
Source link: https://doi.org/10.1007/978-981-19-0507-0_18
The smart distribution network, as an essential component of the smart grid, is an important link between power companies and power consumers. This paper mainly discusses cloud computing as a key to building an intelligent power distribution network based on cloud computing, develops a distribution automation fault location network, and finally tests the device. The system test results show that the average result of the avalanche test findings is 0. 5, with the desired value being 0. 5.
Source link: https://doi.org/10.1007/978-981-19-3632-6_49
As the availability of network, mobility, and portability of cellular devices increase, network traffic will also be on the rise. The consistency of the network is determined by monitoring network parameters and finding a fault in a cellular network. The study is focused on real-time data of 3G cellular networks and includes various network characteristics like uplink threshold and identifies anomalous events in order to anticipate fault occurrences. The research, which includes bidirectional LSTM, vanilla LSTM, and stack LSTM, as well as time-distribution Conv1D, is based on various LSTM methodologies, including bidirectional LSTM, vanilla LSTM, and stacked LSTM.
Source link: https://doi.org/10.1007/978-981-16-9967-2_31
SDP images obtained can easily demonstrate the feature differences of different faults, allowing first, analysis of different faults can be used to determine the vibration signals of various faults, and then, the SDP images can be embedded into the CNN network for feature learning and state recognition; finally, validation was carried out using the Case Western Reserve University bearing dataset.
Source link: https://doi.org/10.1007/978-3-030-99075-6_34
The machine learning framework also includes feature selection in classification tasks as a result of the importance of finding a fault in turning machinery excessively. In any case, this method can be described as a hybrid device because it incorporates both the correlation algorithm as a filter method and an exhaustive search as a wrapper method. Pearson's correlation matric, which measures the correlation between each feature and the target of labeling normal and abnormal results, shows a breakdown in the rotating machinery under normal operating conditions.
Source link: https://doi.org/10.1007/978-981-19-0867-5_36
Aiming at the fact that the early fault characteristics of a rolling bearing are easily submerged by noise and are impossible to recognize, a fault diagnosis system based on weighted variational mode decomposition and cyclic spectrum slice energy is introduced. Secondly, the benefit of the CSSE, which can accurately mediate the fault information, is used to analyze the rebuilt signal, and then the reconstructed signal's fault characteristic frequency is extracted. Finally, the bearing simulation results and outer ring fault signal are used to show that the modified diagnosis procedure can properly isolate the rolling bearing's early fault characteristics.
Source link: https://doi.org/10.1007/978-3-030-99075-6_52
Accurate identification of fault types is a crucial component of sensor fault diagnosis. Particle Swarm Optimization algorithm is enhanced by adjusting the particle velocity with weight and introducing mutated particles, so as to optimize the algorithm's flexibility and optimize Support Vector Machine parameters, Particle Swarm Optimization algorithm is improved.
Source link: https://doi.org/10.1007/978-3-030-99075-6_60
Today, neural networks have become common in modeling. However, the model preparation needs a lot of data, long training time, and high hardware requirements. The images of 112 u00d7 112, 75, 56, and 56 groups can still be appropriate for modified VGG16 to be classified and achieve high accuracy while still reducing the training time, according to the training results. Therefore, it is recommended that you set a target accuracy first and start training from a small sample.
Source link: https://doi.org/10.1007/978-3-030-99075-6_37
* Please keep in mind that all text is summarized by machine, we do not bear any responsibility, and you should always check original source before taking any actions