Papers by Keyword: RBF Neural Network

Paper TitlePage

Authors: Xiao Dong Yuan, Tie Ying Jiang, Bing Li
Abstract: As the problem that loitering unit is vulnerable to stall because of air turbulence in the initial period of posture adjustment process, this paper is based on the classic flight control and the active-disturbance-rejection control principle, and uses the method of ADRC which based on RBF neural network. It designs the ADRC for the attitude adjustment of loitering unit in the initial period, and compares it with the classic flight control algorithm and the active-disturbance-rejection algorithm. Through the contrast of the control simulation for the three control algorithms we can obtain that, RBF ADRC is superior to ADRC control method and the classic flight control method.
Authors: Bo Zhu, Min You Chen, Rui Lin Xu, Xin Xu, Qian Wang
Abstract: As a renewable energy, wind power is considered to be an important non-exhaustible energy in the times of energy crisis. Therefore, in electric power system, wind turbine generators have become important generation alternatives. As wind power penetration increases, power forecasting is crucially important for integrating wind power into a conventional power grid. A short-term wind farm power output dynamic prediction model is presented using RBF neural network with error discriminant function. Based on the wind data from a wind farm in Inner Mongolia of China, a power forecasting map is illustrated, and the errors of the model are analyzed to present the differences between dynamic model and conventional prediction model.
Authors: Wen Ke Jiang, Yu Juan Chen, Jing Zhang
Abstract: A new fault diagnosis method based on artificial immune network is proposed. The network combined aiNet with radial basis function (RBF) NN. The structure of the network proposed is the same as RBF NN. The training samples are clustered first by the improved aiNet algorithm. The centers of the clustering are saved as the centers of the hidden layer, therefore, the amount and positions of nodes in the hidden layer can be determined automatically. The weight matrix is determined by least squares (LS) algorithm. The network is applied to fault diagnosis of CJK6136 spindle gear case. The results of the experiments confirm the performance of the proposed network through comparing with RBF NN under the same conditions. The diagnosis success rate for the network proposed was 99%, while that for RBF NN is 89.5%.
Authors: Yu Min Pan, Cheng Yu Huang, Quan Zhu Zhang
Abstract: In order to improve the precision of gas emission forecasting,this paper proposes a new forecasting model based on Particle Swarm Optimization (PSO).PSO is a novel random optimization method which has extensive capability of global optimization.In the model, PSO is used to optimize the weight,width and center of RBF neural network and the optimal model is applied to forecast gas emission.The diversified factors analysised with grey correlation,MATLAB is employed to implement the model for gas emission forecasting.The simulation results show that the gas emission model optimized by PSO is more accurate than the traditional RBF model.
Authors: Huang Lin Zeng, Xue Fei Tang
Abstract: An intelligent computation is proposed for a matching selection of cutter tools in this paper. A matching selection of cutter tools based on an intelligent computation and knowledge reasoning is realized by three steps. At the first, a cutter tool is roughly selected by inference machine of an expert database. Then a matching scheme of cutter tools will be realized by a RBF neural network. Finally, a matching selection scheme of cutter tools is determined by fuzzy inference machine. It is shown that method proposed here is of a higher efficiency for a matching selection of cutter tools in processing and manufacturing of CNC machine.
Authors: Ming Xiang Pang, Xue Zhen Cheng, Xiao Chao Qian, Mao Yong Cao
Abstract: In coal mines fire consists of one of the main disasters, which usually take place for the reason that the water content of coal is over low. Over low water content of the coal transported with belt more likely brings about flying coal dust, which, when accumulated to some degree, will triggers explosion. Given that in China now coal is mainly transported with belt in coal mines, the author in this paper proposes a way to measure water content of coal transported with belt by use of microwave attenuation method and improve the measure accuracy through RBF neural network algorithm. This method is proved to be scientifically reasonable through laboratory simulation and experimentation. The theoretical basis and technical support are provided to increase the accuracy measuring water content of coal transported with belt by this method.
Authors: Dong Dong Liu
Abstract: Rolling mills process is too complicated to be described by formulas. RBF neural networks can establish finishing thickness and rolling force models. Traditional models are still useful to the neural network output. Compared with those finishing models which have or do not have traditional models as input, the importance of traditional models in application of neural networks is obvious. For improving the predictive precision, BP and RBF neural networks are established, and the result indicates that the model of load distribution based on RBF neural network is more accurate.
Authors: Heng Jie Li, Xiao Hong Hao, Xi Ping Pei
Abstract: Improved clonal selection algorithms and RBF neural network are used for solving nonlinear optimization problems and modeling respectively in iterative learning control, and a nonlinear optimal iterative learning control algorithm (NOILCA) is proposed. In this method, an improved clonal selection algorithm is used for solving the optimum input for the next iteration; another one is used to update the RBF neural network model of real plant. Compared with GA-ILC, NOILCA has faster convergence speed, and is able to deal with the problem of inaccurate plant model, can obtain satisfactory tracking through the few several iterations.
Authors: Fa Hong Yu, Mei Jia Chen, Wei Zhi Liao
Abstract: There are many learning evaluation methods, but most of them are subjective, which contains a lot of man-made factors. This paper presents a new learning evaluation method based on radial basis function (RBF) neutral network. By analysis the orthogonal least squares for RBF and determines the center of the basis functions, the model of RBF neural network was constructed. Experimental studies show that the Method Based on RBF Neural Network is effective for learning Evaluation.
Authors: Xiao Qiang Qin
Abstract: In this paper a prediction model of mine gas emission base on continuous ant colony algorithm was presented aimed at the difficulty in prediction of mine gas emission. By comparing the measured data with the data calculated by the prediction model, indicated the model was accurate and credible.
Showing 1 to 10 of 159 Paper Titles