An integrated scheme of neural network and optimal predictive control

An approach of adaptive predictive control with a new structure and a fast algorithm of neural network (NN) is proposed. NN modeling and optimal predictive control are combined to achieve both accuracy and good control performance. The output of nonlinear network model is adopted as a measured disturbance that is therefore weakened in predictive feed-forward control. Simulation and practical application show the effectiveness of control by the proposed approach.     

Industrial applications of advanced control techniques

This paper discusses two industrial control applications using advanced control techniques. They are the optimal-tuning nonlinear PID control of hydraulic systems and the neural predictive control of combustor acoustic of gas turbines. For hydraulic control systems, an optimal PID controller with inverse of dead zone is introduced to overcome the dead zone and is designed to satisfy desired time-domain performance requirements. Using the adaptive model, an optimal-tuning PID control scheme is proposed to provide optimal PID parameters even in the case where the system dynamics is time variant. For combustor acoustic control of gas turbines, a neural predictive control strategy is presented, which consists of three parts: an output model, output predictor and feedback controller. The output model of the combustor acoustic is established using neural networks to predict the output and overcome the time delay of the system, which is often very large, compared with the sampling period. The output-feedback c     

Decoupling Control Method Based on Neural Network for Missiles

In order to make the static state feedback nonlinear decoupling control law for a kind of missile to be easy for implementation in practice, an improvement is discussed. The improvement method is to introduce a BP neural network to approximate the decoupling control laws which are designed for different aerodynamic characteristic points, so a new decoupling control law based on BP neural network is produced after the network training. The simulation results on an example illustrate the approach obtained feasible and effective.     

Neuro-fuzzy predictive control for nonlinear application

Aiming at the unsatisfactory dynamic performances of conventional model predictive control （MPC） in a highly nonlinear process, a scheme employed the fuzzy neural network to realize the nonlinear process is proposed. The neuro-fuzzy predictor has the capability of achieving high predictive accuracy due to its nonlinear mapping and interpolation features, and adaptively updating network parameters by a learning procedure to reduce the model errors caused by changes of the process under control. To cope with the difficult problem of nonlinear optimization, Pepanaqi method was applied to search the optimal or suboptimal solution. Comparisons were made among the objective function values of alternatives in initial space. The search was then confined to shrink the smaller region according to results of comparisons. The convergent point was finally approached to be considered as the optimal or suboptimal solution. Experimental results of the neuro-fuzzy predictive control for drier application reveal that the proposed control scheme has less tracking errors and can smooth control actions, which is applicable to changes of drying condition.     

Study on Adaptive Control with Neural Network Compensation

A scheme of adaptive control based on a recurrent neural network with a neural network compensation is presented for a class of nonlinear systems with a nonlinear prefix. The recurrent neural network is used to identify the unknown nonlinear part and compensate the difference between the real output and the identified model output. The identified model of the controlled object consists of a linear model and the neural network. The generalized minimum variance control method is used to identify parameters, which can deal with the problem of adaptive control of systems with unknown nonlinear part, which can not be controlled by traditional methods.Simulation results show that this algorithm has higher precision, faster convergent speed.     

Neural network modeling and control of proton exchange membrane fuel cell

A neural network model and fuzzy neural network controller was designed to control the inner impedance of a proton exchange membrane fuel cell (PEMFC) stack. A radial basis function (RBF) neural network model was trained by the input-output data of impedance. A fuzzy neural network controller was designed to control the impedance response. The RBF neural network model was used to test the fuzzy neural network controller. The results show that the RBF model output can imitate actual output well, the maximal error is not beyond 20 m-, the training time is about 1 s by using 20 neurons, and the mean squared errors is 141.9 m-2. The impedance of the PEMFC stack is controlled within the optimum range when the load changes, and the adjustive time is about 3 min.     

Wind speed prediction by chaotic operator network based on Kalman Filter

A novel prediction network composed of some chaotic operators is proposed to predict the wind speed series.Training samples are constructed by the theory of phase space reconstruction.Genetic algorithm is adopted to optimize the control parameters of chaotic operators to change the dynamic characteristic of the network to approach to the predicted system.In this way,the dynamic prediction of wind speed series can be completed.The wind acceleration series can also be predicted by the same network.And the prediction results of both series can be fused by Kalman Filter to get the optimal estimation prediction result of the wind speed series,which is superior to the result obtained by each single method.Simulation results show that the prediction network has less computation cost than BP neural network,and it has better prediction performance than BP neural network and autoregressive integrated moving average model.Kalman Filter can improve the prediction performance further.     

Hybrid dynamic model of polymer electrolyte membrane fuel cell stack using variable neural network

The polymer electrolyte membrane(PEM) fuel cell has been regarded as a potential alternative power source,and a model is necessary for its design,control and power management.A hybrid dynamic model of PEM fuel cell,which combines the advantages of mechanism model and black-box model,is proposed in this paper.To improve the performance,the static neural network and variable neural network are used to build the black-box model.The static neural network can significantly improve the static performance of the hybrid model,and the variable neural network makes the hybrid dynamic model predict the real PEM fuel cell behavior with required accuracy.Finally,the hybrid dynamic model is validated with a 500 W PEM fuel cell.The static and transient experiment results show that the hybrid dynamic model can predict the behavior of the fuel cell stack accurately and therefore can be effectively utilized in practical application.     

Sensor Registration Based on Neural Network in Data Fusion

The contents of sensor registration in the multi-sensor data fusion system are introduced, and some existing methods are analyzed. Then, one approach to sensor registration based on BP neural network is proposed. Here the measurements from radar are transformed from the polar coordinate system to the Cartesian coordinate through a BP neural network. With this approach, the systematic errors are removed as well as the coordinate is transformed. The efficiency of this method is demonstrated by simulation, and the result show that this approach could remove the systematic errors effectively and the DAR are closer to real position than DBR.     

NN adaptive pole placement method for sintering finish point control

A neural network model with a special structure, which is divided into linear and nonlinear parts, was proposed for identification of a nonlinear system. In this model, the nonlinear part of the object is treated as a measured disturbance, and is compensated by a feed forward method; an adaptive pole placement algorithm is used to control the linear part of the object. The simulation results show that the identification efficiency and accuracy are improved when the new controller is applied to sintering finish point control.     

基于自校正神经网络的预测控制系统设计

为改善数控线切割加工过程伺服控制性能,提出了一种基于自校正神经网络预测控制系统总体结构,建立了用于离线训练控制器的对象模型以及可以进行多步前向预测的预测模型,设计了基于预测模型提供的梯度信息进行学习训练的神经网络控制器.以数控线切割机床加工过程控制为研究对象,对该系统进行了计算机仿真并进行了结果分析,结果表明本文所提出的基于自校正神经网络预测控制系统具有较好的控制效果.     

神经网络预测-PID串级控制在同步发电机励磁控制中的应用

为提高发电机励磁控制系统的稳定性,分析了同步发电机的自并励励磁系统的结构和数学模型,介绍了神经网络预测控制的结构和算法,分别基于PID控制、神经网络预测控制和神经网络预测-PID串级控制算法对自并励励磁系统进行了仿真分析.通过仿真结果的对比分析,说明神经网络预测-PID串级控制在励磁控制中的应用提高了励磁系统的动态性、稳定性和抗干扰能力.     

基于RBF神经网络的非线性模型预测控制

提出了一种基于径向基函数(RBF)神经网络的非线性模型预测控制系统,利用RBF神经网络的非线性拟合性,构建一个神经网络预测器(NNP)来预测模型未来时刻的输出值.然后利用神经网络控制器(NNC)实现基于模型的预测控制.仿真结果表明此方法具有较好的控制效果,并且在有扰动和模型失配的情况下,表现了良好的鲁棒性.     

神经网络预测器在铝电解过程控制中的应用

目的采用神经网络预测控制方法来解决铝电解过程中存在的时变和大时滞问题，提高其控制性能．方法提出了一种基于铝电解过程的神经网络预测控制算法，建立了神经网络预测模型，将神经网络和预测控制算法相结合，结果实现了铝电解过程的最优控制．神经网络预测模型的输出能够很好地跟踪铝电解生产过程，预测效果好．结论笔者提出的控制方案能够使铝电解过程很快进入稳态，超调量较小，提高了铝电解过程的动态和稳态性能．     

基于模糊神经网络的非线性系统预测函数控制研究

针对预测函数控制难以很好地实现非线性系统控制的问题,将模糊神经网络与预测函数控制相结合,设计一种基于模糊神经网络的非线性系统的预测函数控制器。用模糊神经网络辨识非线性系统的模型,辨识结果送到预测函数控制中,从而得到预测模型,最终得到最优的控制量。通过Matlab计算机仿真,可以看出此控制器对于非线性系统具有良好的控制效果和鲁棒性。     

基于神经网络的风力发电机组变桨距复合控制

在分析风力发电机组系统特性和变桨距控制要求的基础上,提出了一种基于神经网络的分段复合控制方法进行变桨距控制,以解决风力发电系统的多干扰、时滞性、非线性等控制问题。该方法利用神经网络进行模型辨识,再依据运行工况分别进行模型预测控制和前馈控制,不但解决了风力发电机组系统模型难以精确建立的困难,而且去除了可测量的主要外扰——风速随机变化对系统动态控制品质的影响,从而提高了变距系统的响应快速性和抗干扰能力。最后,利用考虑尾涡效应的机组动态模型作为应用实例对复合控制方法进行仿真,结果表明提出方法的有效性与实用性。     

基于模块化神经网络的铁水硅含量预测系统

目的 提出基于模块化神经网络的铁水硅含量预测方法,改善系统控制性能指标.方法 采用模块化神经网络预测控制策略,建立模块化神经网络预测模型,按输入物理量的性质构成4个神经网络模块,再由预测神经网络输出铁水硅含量的预测值,从而控制炉温.结果 提高了学习效率和泛化能力,有效地改善了模型的预测精度.结论 模块化神经网络铁水硅含量预测模型,将同类量进行了模块划分,通过对铁水硅含量预测,可提高炉温的控制精度和动态跟踪能力,具有结构简单、实时性好、预测精度高等特点.     

多模型预测控制及其在火电厂过热汽温控制中的应用

针对火电厂存在的过热汽温问题,设计了多模型预测控制系统.根据若干建模工况点,离线训练局部人工神经网络模型,利用贝叶斯估计的方法在线计算每个局部神经网络模型概率,加权计算出模型预测输出值.根据预测控制的原理,利用Newton-Raphson迭代法得到控制信号,从而得到了仅含一个控制器的多模型预测控制系统.仿真结果表明,在负荷大范围变化的工况下仍能保持良好的控制性能,具有较强的鲁棒性.     

基于Tchebycheff正交神经网络的非线性多步预测控制

针对多变量非线性系统 ,提出了一种基于 Tchebycheff正交神经网络的多步预测控制方案 ,采用 Tchebycheff正交神经网络离线建立预测模型 ,并以偏差补偿和模型修正相结合的方式对预测模型进行误差补偿 ,经在线校正用于预测控制。同时对性能指标中的偏差项和控制项加权 ,进一步改善预测控制性能。仿真结果表明了控制算法的有效性