Thermal process model identification based on radial basisfunction neural networks

In order to accurately reflect the dynamic behavior and realize the whole optimal control of the thermal process, a novel modeling method of the RBF NN (Radial Basis Function Neural Networks) model is proposed to build nonlinear model. This method is based on entropy clustering and competitive learning algorithm, combined with nonlinear autoregressive moving average (NARMA) model to identify the RBF NN stucture, and the power vector is gotten by the least square algorithm. Two simulation experiments show that the proposed method of the identification based on NARMA model and RBF NN can accurately describe the non linearity of the process and has less hidden nodes.     

An artificial immunity based predictive method for neural networks and its applications

A radial basis function (RBF) neural network learning algorithm based on immune recognition was proposed to improve the low forecast precision and the slow convergence speed of such networks. In the algorithm, artificial immunity was used to determine the center and width parameters of the Gauss basis function. The recognized data were regarded as antigens and the compression mapping of antigens were taken as antibodies, i.e., the centers of the hidden layer. The recursion least square algorithm (RLS) was employed to determine the output layer weights. The algorithm improved the convergence speed and precision of the RBF neural networks. The model was applied to the blast furnace of a large iron and steel company. The results show that the model has forecast precision far superior to existing models and requires less training time than they do.     

Nominal electric field calculation at the ground level of UHV DC transmission lines based on radial basis function method

Nominal electric field at ground level is one of the important factors for the design of transmission lines. FEM meets two big challenges in electromagnetic field computation. One comes from the mesh generation, which is a big difficulty especially in complex geometry. The other is the long time calculation. So the radial basis function method is introduced to calculate the nominal electric field at the ground level of UHVDC transmission lines. Not only the definition and ideas, but also the detailed implementation procedure of radial basis function method are presented. Computer programs are developed to simulate the model and the influences of conductor configuration on nominal electric field intensity are analyzed. Examples of calculation show that the results are accorded with those available literatures, and agreed satisfactorily with analytic solutions of the coaxial cable model. The height of the polar lines and conductor inter-spacing has a remarkable influence on ground nominal electric field intensity while the sub-conductor radius and number of sub-conductors have little influence.     

A radial basis function neural network based self adapting predictive decoupling control system for gas collector pressure in coke ovens

Pressure is measured at different levels in the loop layer because self adapting predictive decoupling control systems are strongly coupled, disturbed, and non linear and there is a long time delay for gas collector pressure systems in coke ovens. By combing the traditional neural network control and proportional integral differential(PID) controllers based on radial basis function(RBF) neural network identification, the gas collector pressure is ensured to reach the desired technology range. The prediction model of an RBF neural network is used for advanced prediction of the actual output pressure to overcome delays in general gas collection. The simulation results and application indicate that the method can obtain ideal control results.     

Feature Extraction and the Design of Classifier in Small Set Handwritten Chinese Character Recognition

This paper presents basic concept of feature extraction of handwritten Chinese character, and proposes a new feature extraction named superposition mesh weighting factors strokes extracting algorithm to obtain feature of small set handwritten Chinese character. Basing on the analysis model of RBFNN, an integration RBF classifier is used for small set handwritten Chinese character recognition. Then, the hybrid optimize strategy, which combines the genetic algorithm and the simulated annealing, is adopted to train RBFNN.     

Fast algorithm for apparent resistivity calculation based on back propagation neural networks

To avoid the complex numerical calculation for the electromagnetic field and determine underground abnormality, a neural network based method is proposed. In consideration of turn off transmitter current, the effect of a linear ramp turn off current on transmitter is corrected. The characteristics of transient expression and the traditional calculation algorithm for apparent resistivity are analyzed, and a predigest structure of network is obtained based on the kernel expression. The three layer back propagation(BP) neural network is trained by using sample data in homogeneous half space, and its number in hidden layer was determined. The method proposed is compared with two traditional calculation methods with simulation experiments. The result demonstrates that BP neural network has a high speed of processing data and is useful in explanation of the transient electromagnetic method.     

Beam pumping process modeling and parameters optimization based on generalized regression neural networks for energy conservation

This paper presents a data-mining-based beam pumping unit process modeling and parameters optimization method to solve the problem of inefficiency and energy-intensive of beam pumping unit. The ideality of process parameters is one of the main factors influencing system efficiency and energy consumption, while the effectiveness of mode plays a key role in process parameters choosing. Beam pumping unit system is a complicated nonlinear system, and is hardly to be precisely described by precise mathematical models. Generalized regression neural network (GRNN), which is powerful in nonlinear mapping and generalization, is suitable for nonlinear systems. Therefore, GRNN is proposed to model the beam pumping unit in this paper, and the experimental results show that the fitness is good. Then the trained model is applied to optimize the decision parameters by vector evaluated particle swarm optimization based on Pareto (VEPSO-BP), and at last the resulting parameters are applied to the production. Experimental results show that after using the optimal parameters, the efficiencies and energy consumptions increase more than 6.6% and decrease more than 4.1% respectively, which illustrates the feasibility and effectiveness of the proposed method.     

基于布谷鸟搜索神经网络的微波加热温度预测模型

微波加热是一种与被加热物直接相互作用的选择性加热方式,具有清洁、节能、减排等特点。针对工业物料作为微波加热负载时,其温度非线性变化的特点,以微波工业加热过程中的多维、海量参数为研究对象,基于泛函接神经网络模型提取样本数据的深度特征,提出了一种基于布谷鸟搜索算法,优化BP神经网络的网络参数,建立了以"数据驱动"为手段微波加热工业物料温度模型。仿真实验结果证明了所提出模型的准确性、实时性。     

Cost estimate of power line projects based on grey relational analysis and neural networks

To accurately estimate the cost of a power line project,a method based on grey relational analysis (GRA) and neural networks (NN) is presented and studied. Grey relational analysis technologies are used to analyze the features of the transmission line project and ten main features which affect the project cost most are selected. Then, the main features are used as input neural cell of neural networks, and a model of GRA-ANN is built. To verify the method, the cost data of a 110 kV power construction project are used to train and test the model. Results show the model’s maximum relative error of static investment is 3.72% and the minimum is 1.85%, and its accuracy is high. The LM-BP algorithm and the traditional BP algorithm are used respectively to train the GRA-ANN network, and results show the error declining rate of LM-BP algorithm is faster and the overall error is lower.     

基于无迹卡尔曼滤波神经网络的光伏发电预测

针对光伏发电系统在不同天气状况下发电功率预测精度不高的问题,在分析传统方法的基础上,提出一种无迹卡尔曼滤波神经网络光伏发电预测方法。该方法利用无迹卡尔曼滤波实时更新神经网络模型的权重,以直流电压和电流作为系统的输入,以有功功率和无功功率作为系统的输出,分别建立两个独立的双输入单输出功率预测模型。实验结果表明：所提出的方法对有功功率和无功功率的预测精度分别为97.3%和94.2%,并且对天气具有良好的鲁棒性。     

Fault diagnosis of rotating machinery based on self-learning fuzzy spiking neural network

For fuzziness classific boundry of fault diagnosis of rotating machinery and traditional neural network algorithms difficulted to solve contradiction between application problems example scale and netwok scale,a methord of self-learning fuzzy spiking neural network is put forward. The methord overcomes unavailability of cluster analysis on classific boundry of fault diagnosis of rotating machinery by species encoding of pulse sequence and unsupervised learning. The method shows that it effectively solves boundary fuzziness problem on fault diagnosis of rotating machinery,and greatly improves efficiency of fault diagnosis.     

Mechanical property prediction model of AZ31 magnesium alloys based on an artificial neural network with parameter optimization

The tensile strength, yield strength and elongation of AZ31 magnesium alloys on different annealed conditions are tested by mechanical properties experiments. A model of corresponding mechanical properties is built by applying artificial neural network, and it is optimized by a new method,namely all permutations and combinations training of parameters. The results show that the network model has an excellent performance, which is based on optimal parameters obtained from all permutations and combinations training. Compared with traditional model, whose parameters are obtained from conventional heuristic, the improved model has higher average correlation coefficient and lower average error. Therefore, it can predict the mechanical properties of AZ31 magnesium alloy on different annealed conditions more accurately.