Papers by Keyword: RBF Neural Network

Abstract: In this paper, the relationship between microstructure, parameters of cyclic loading and high cycle fatigue property of Ti-6Al-4V alloy was established by artificial neural network (ANN) modeling. The back propagation (BP) neural network and radial basis function (RBF) neural network were established by MATLAB. The input parameters of these models were the primary α volume fraction, primary α size, cyclic loading frequency and stress ratio. The output parameter was high cycle fatigue strength. The neural networks were trained with dataset collected from the literature. The prediction results showed that both of the networks have good generalization ability. In addition, the BP neural network with Levenberg-Merquardt (LM) learning algorithm has better fault tolerance and versatility in dealing with high cycle fatigue property, which is able to predict the high cycle fatigue property with a high accuracy.

Abstract: This paper is addressed water pipe pressure control method, chose a water pipe with bypass valve as an object. The paper made the pressure and flow control of the object a mathematic model. This paper design a PID control algorithm which base on RBF neural network to control the pressure and flow of the object, and simulate the control process both RBF neural network PID algorithm and PID algorithm. The last part of the paper contrasts the two simulation result.

Abstract: With the interconnection of the large-scale wind power, wind power forecasting is particularly important to the dispatcher of power grid. Based on the historical data, this paper proposes a prediction method based on RBF (radial basis function) neural network. This method is based on the model taking the influence of the system input (wind speed, wind direction, historical power output data) on the predicting error into consideration to get the optimal input values. Examples with field data obtained from Northwest of China show the effectiveness and higher precisionof the proposed method.

Abstract: Dynamic optimization scheduling of the gas in iron and steel enterprises has great significance to reduce gas emission and the short-term forecast is the premise to realize the energy dynamic scheduling. Based on the characteristics that the influencing factors of blast furnace gas amount are complex and difficult to collect, a grey radial basis function (RBF) neural network forecast model is proposed to predict the gas amount for blast furnace in this paper. Combining grey theory, which is used to preprocess the historical data and obtain abundant information, with RBF neural network makes the effective trend forecast in the next 30 minutes come true. The model proposed in this paper is proved to be more accurate according to control experiments against the grey BP neural network.

Abstract: Using the C - C method to reconstruct the phase space of wind power time series, get the maximum wind power time series Lyapunov exponent, confirmed that the wind power time series have chaotic characteristics. Followed by the radial basis function (RBF) neural network model for wind power chaotic local multi-step prediction, results show that the prediction effect is better than that of the predicted effect of 48 hours for 24 hours.

Abstract: Research and analysis of RBF neural network structure and characteristics. Find out its shortcomings and propose an improved method for the deficiencies, then created a neural network model for using entropy-based clustering and competitive learning algorithm. Using MATLAB simulation tools for model simulation, confirmed the entropy clustering and competitive learning algorithm of FBF prediction neural network have high precision and generalization ability of stronger character.

Abstract: The study of existing deficiencies, on the basis of overhead transmission line fault detection based on radial basis function (RBF) neural network theory, the fundamental frequency power spectrum as characteristic vector of fault signal, this paper proposes a new method of transmission line fault type identification. The system with complex structure of 10 KV overhead transmission lines as the research object, on the basis of the transmission line model is established by using Simulink software, for different types of short circuit fault simulation sampling, extract fault features, combined with the zero sequence current, as the input vector, establish the RBF neural network for fault type. Results show that: the fundamental frequency in fault signal power spectrum as the feature vector is easy to extract, information is more concise, the RBF neural network in the feasibility in training high, identify accurately and quickly.

Abstract: Aiming at the issue of fault prediction technique of power electronic circuits, a method based on characteristic parameter data and Particle Swarm Optimization RBF(Radial Basis Function) Neural Network for the fault prediction of power electronic circuits was proposed. Taking the Buck converter circuit as an example,the fault prediction of power electronic circuits was achieved. Firstly,the output voltage was selected as monitoring signal, then the average voltage and ripple voltage were extracted as characteristic parameters. Lastly Particle Swarm Optimization RBF Neural Network was used to predict the fault. The experimental results show that the Particle Swarm Optimization RBF Neural Network is more accurate in predicting than the only RBF Neural Network.The new method can trace the characteristic parameters’ trend and can be effectively applied in fault prediction of power electronic circuits.

Abstract: From numerous approaches studying the prediction of stock price, this paper proposed a new approach which was the combination of RBF neural network and Markov chain to forecast the stock closing price of the Shanghai composite index. Markov chain was aimed at making the error between the actual price and predicted price obtained by RBF neural network correct. Besides, for higher prediction accuracy, genetic algorithm was used to optimize the state division of Markov chain. The experimental result confirmed its effectiveness and superiority in comparison with the other two methods in some time interval.

Abstract: Dam deformation is a multivariate complicated and nonlinear problem, it’s unable to establish accurate mathematical model.A dam deformation prediction model based on RBF neural network was constructed in this paper to enhance prediction accuracy. Three closely related factors in dam deformation are hydraulic components, temperature component and aging components ,they were selected as Input vector of RBF neural network, dam deformation measured value as a model target output. In Matlab 2011b simulation software,50 groups Fengman dam quantitative observation data from 2012 to 2013 as the sample data,45 groups were used in RBF neural network model training, other 5 groups were used in testing for the model. The simulation result shows that testing value is very close to the true value in this method, the average relative error close to 3%. Effectiveness of the dam deformation prediction based on RBF neural network is Verified by experiments.