Nonlinear control system using learning Petri network

According to recent understanding of brain science, it is suggested that there is a distribution of functions in the brain, which means that different neurons are activated depending on which sort of sensory information the brain receives. We have already developed a learning network with a function distribution which is called the Learning Petri Network (LPN) and have shown that this network could learn nonlinear and discontinuous mappings which the Neural Network (NN) cannot. In this paper, a more realistic application which has dynamic characteristics is studied. From simulation results of a nonlinear crane control system using LPN controller, it is clarified that the control performance of LPN controller is superior to that of NN controller. © 2000 Scripta Technica, Electr Eng Jpn, 131(3): 58‐69, 2000     

高精度伺服系统的CMAC学习控制研究

根据神经网络的特点，提出了基于CMAC网络的伺服系统在线学习控制方案，并给出了其再励学习算法。该方法无需离线训练，经在线学习得到系统的逆动态特性，并据此实时修改控制参数。引入鲁棒控制项，增强了学习算法的鲁棒性。在某单轴速率／位置转台上的应用结果表明，该方法可有效地改善伺服系统的跟踪精度和抗扰性能。     

基于深度强化学习的配电网负荷转供控制方法

随着城市规模的快速扩张以及电能替代的不断推进,配电网节点数大量增加,结构愈加复杂,发生故障后拓扑变化不确定性较大,传统负荷转供方法难以在短时间内给出高质量的解决方案。为此,提出基于深度强化学习的配电网负荷转供控制方法。将负荷转供过程视为一个马尔可夫决策过程,与配电网实时电气、拓扑数据进行交互,对联络开关与分段开关进行控制。为了提高算法的精度与泛化能力,针对算法动作策略加入了预模拟机制,调整了动作与学习的比例并采用自适应优化算法进行求解。算例分析表明,所提方法能够应对不同故障下配电网的拓扑变化,即时给出负荷恢复量、电网损耗、开关动作次数多方面最优的转供控制方案,这对于减小故障后的停电损失与提高用户满意度有着重要意义。     

基于卷积神经网络的电力系统暂态稳定预防控制方法

为了更好地实现电力系统暂态稳定预防控制,提出了基于卷积神经网络(Convolutional Neural Network, CNN)的电力系统暂态稳定预防控制方法。通过CNN模型输出变量灵敏度选择控制发电机并确定控制量,然后采用CNN和时域仿真相结合的暂态稳定评估方法进行控制方案校核,得到使系统在预想故障下稳定的控制方案。采用某省级电网算例进行预防控制效果验证。结果表明,采用所提出的预防控制方法,可以找到使系统恢复稳定的预防控制策略。     

基于迭代学习的矿井提升系统跟踪控制

矿井提升机在煤矿开采和生产过程中发挥着至关重要的作用。针对矿料提升过程具有较强重复性的特点,提出迭代学习的提升机速度和位置跟踪控制方法。设计了提升机的D型迭代学习控制器,同时考虑运行过程中出现的非重复性干扰设计了带有滤波器型迭代学习控制器。并采用λ范数证明了系统的收敛性,理论结果表明矿井提升机位置与速度跟踪误差可以收敛到0。同时仿真结果表明,经过30次运行后,跟踪误差几乎收敛到0,迭代学习控制算法可利用矿井作业的重复运行特性可以有效提高提升机的跟踪性能,带有滤波器的迭代学习控制算法可较好地抑制了非重复扰动的影响。     

Reinforcement learning based backstepping control of power system oscillations

This paper proposes a reinforcement learning based backstepping control technique for damping oscillations in electric power systems using the generators excitation systems. Decentralized controllers are first designed using the backstepping technique. Then, reinforcement learning is used to tune the gains of these controllers to adapt to various operating conditions. Simulation results for a two area power system show that the proposed control technique provides better damping than (i) conventional power system stabilizers and (ii) backstepping fixed gain controllers.     

基于Netica的自学习贝叶斯网络的构建

针对贝叶斯网络构建时参数与结构难以自适应调整,提出基于Netica的自学习贝叶斯网络的构建方法。首先根据Netica要求处理样本数据集,然后运用Netica基础函数开发结构学习模块和参数学习模块,进而能够构建出自动学习样本数据集的贝叶斯网络。同时,开发了概率推理模块和证据敏感性分析模块以评估所建网络的有效性。以国家电网的短路故障样本数据为例建立其自学习贝叶斯网络,实验构建的自学习贝叶斯网络能够实现不确定性推理,表明所提方法是贝叶斯网络功能实现的一个新途径。     

A control strategy for leveling load power fluctuations with a successive learning fuzzy-neural network based on prediction of average load power

The effective usage of power facilities can be realized by leveling the fluctuating active power and compensating the reactive power. A fuzzy control strategy of superconducting magnetic energy storage (SMES) has been proposed for this purpose. The control results depend on the values of the scaling factors in fuzzy reasoning. Therefore, to obtain better control results, the scaling factor should be successively adjusted according to the load power fluctuations. In this paper, a control strategy based on autotuning of scaling factors and a fuzzy singleton reasoning method using backpropagation in a neural network is proposed for leveling load fluctuations. The prediction and revision of the teaching signal in terms of the energy of the SMES is proposed. The learning rate and the revision of the teaching signal are discussed. Better leveling of load power fluctuation is shown to be achievable by using fuzzy logic and neural networks. © 1997 Scripta Technica, Inc. Electr Eng Jpn, 120(2): 72–81, 1997     

基于长短期记忆循环神经网络的AGC实时控制策略

大量新能源的接入以及电网中冲击负荷数量的剧增,使得电网对自动发电控制(AGC)策略提出了新的要求.简化AGC的一般控制流程,对比不同AGC策略的控制特性,在每个考核周期内选择控制效果更优的控制策略,并充分发挥多种控制策略在各自优势工况下的性能,以得到优秀控制数据集;在此基础上,以长短期记忆(LSTM)循环神经网络为神经...     

基于2-D线性连续-离散系统理论的P型闭环迭代学习控制

将2-D线性连续一离散系统理论应用于迭代学习控制中，给出能很好反映迭代学习控制过程的数学模型(2-D线性连续一离散系统Roessor模型)。在2-D系统理论基础上证明了P型闭环迭代学习控制律的收敛性。根据该系统理论设计的闭环迭代学习控制器，受到的限制较小。     

Reinforcement learning based closed-loop reference model adaptive flight control system design

In this study, we present a reinforcement learning (RL)-based flight control system design method to improve the transient response performance of a closed-loop reference model (CRM) adaptive control system. The methodology, known as RL-CRM, relies on the generation of a dynamic adaption strategy by implementing RL on the variable factor in the feedback path gain matrix of the reference model. An actor-critic RL agent is designed using the performance-driven reward functions and tracking error observations from the environment. In the training phase, a deep deterministic policy gradient algorithm is utilized to learn the time-varying adaptation strategy of the design parameter in the reference model feedback gain matrix. The proposed control structure provides the possibility to learn numerous adaptation strategies across a wide range of flight and vehicle conditions instead of being driven by high-fidelity simulators or flight testing and real flight operations. The performance of the proposed system was evaluated on an identified and verified mathematical model of an agile quadrotor platform. Monte-Carlo simulations and worst case analysis were also performed over a benchmark helicopter example model. In comparison to the classical model reference adaptive control and CRM-adaptive control system designs, the proposed RL-CRM adaptive flight control system design improves the transient response performance on all associated metrics and provides the capability to operate over a wide range of parametric uncertainties.     

基于神经网络补偿的二自由度PID控制

提出一种基于神经网络补偿的二自由度PID控制方法，该方法将BP网络应用到传统二自由度PID控制中，克服了参数固定不变时，系统跟随性能和抗扰性能变差的缺点，有效地减弱了参数摄动对系统造成的影响，改善了系统的动态品质，仿真结果表明该方法 的有效性。     

基于神经网络的模糊控制在开关磁阻电机中的应用研究

把神经网络与模糊逻辑结合起来,利用神经网络的学习控制算法调节模糊逻辑隶属函数,通过对开 关磁阻电机运行特性的分析,提出了一种可应用于开关磁阻电机驱动系统的智能控制方法,理论和仿真结果均证明了这种基于神经网络模糊控制方法在开关磁阻电机驱动系统中应用的可行性和可靠性。     

基于迭代学习控制理论的励磁控制器设计

基于迭代学习控制理论提出了一种设计单机无穷大系统励磁控制器的新方法,克服了迭代学习控制在有限时间区间上实现完全跟踪的限制。将迭代学习控制对输出控制量u(t)的记忆与修正改成对期望控制ud(t)的记忆与修正,采用最小二乘法拟合控制器参数,使设定的控制与期望控制之差达到最小,求得控制律。用Matlab对发电机励磁控制系统进行仿真研究,结果表明所设计的励磁控制器具有较好的动态特性和较强的鲁棒性。     

基于云神经网络自适应逆系统的电力系统负荷频率控制

针对区域互联电力系统受到风电及负荷扰动后,系统频率会出现大幅度波动的问题,提出一种基于云神经网络自适应逆系统的多区域互联电力系统负荷频率控制方法。在分析单一区域电力系统有功输出特性的基础上,建立计及多区域有功输出的互联电力系统负荷频率控制模型。采用自适应逆控制有效解决系统响应和扰动抑制的矛盾。将云模型引入自适应逆系统构建云神经网络辨识器。利用云模型在处理模糊性和随机性等不确定性方面的优势,进一步提高神经网络的辨识能力。仿真结果表明,所设计的云神经网络自适应逆系统不仅可以得到好的动态响应,还可以使风电及负荷引起的扰动减小到最小。     

Adaptive multitask network based on maximum correntropy learning algorithm

Adaptive networks solve distributed optimization problems in which all agents of the network are interested to collaborate with their neighbors to learn a similar task. Collaboration is useful when all agents seek a similar task. However, in many applications, agents may belong to different clusters that seek dissimilar tasks. In this case, nonselective collaboration will lead to damaging results that are worse than noncooperative solution. In this paper, we contribute in problems that several clusters of interconnected agents are interested in learning multiple tasks. To address multitask learning problem, we consider an information theoretic criterion called correntropy in a distributed manner providing a novel adaptive combination policy that allows agents to learn which neighbors they should cooperate with and which other neighbors they should reject. In doing so, the proposed algorithm enables agents to recognize their clusters and to achieve improved learning performance compared with noncooperative strategy. Stability analysis in the mean sense and also a closed‐form relation determining the network error performance in steady‐state mean‐square‐deviation is derived. Simulation results illustrate the theoretical findings and match well with theory.     

网络控制系统基于LMI的时延优化方法研究

针对网络控制系统中存在的传输延迟,介绍了一种网络控制系统时延补偿方法,将线性矩阵不等式优化方法引入到控制器的设计中,使闭环系统具有较好的性能。仿真结果表明了所提出方法的有效性。     

基于IP网实现变电站监控系统的远程维护

针对变电站计算机监控系统软件维护量大、效率低等问题,提出了基于IP网实现监控系统的网络化远程维护方案,即在调度端增加远程维护终端,利用光纤通信IP网络和路由技术接入各套监控系统,并采用多种远程控制方法实现监控系统的远程维护.具体规划了该系统的总体框架,讨论了系统软件配置及远程维护方法,并提出了几个存在的问题.该方案已在漯河市电业局得到应用,解决了许多实际问题,降低了维护成本,提高了劳动效率,增强了网络的安全可靠性.     

Neural network based terminal iterative learning control for uncertain nonlinear non‐affine systems

In this paper, a novel neural network based terminal iterative learning control method is proposed for a class of uncertain nonlinear non‐affine systems to track run‐varying reference point with initial state variance. In this new control scheme, the non‐affine terminal dynamics are converted affine, and the unrealisable recurrent network is simplified into realisable static network. As a result, the effect of initial state and control signal on terminal output can be estimated by neural network. With this estimation, the proposed control scheme can drive nonlinear non‐affine systems to track run‐varying reference point in the presence of initial state variance. Stability and convergence of this approach are proven, and numerical simulation results are provided to verify its effectiveness.     

基于神经网络的过热汽温模型预测控制

通过神经网络系统辨识实现非线性系统的模型预测控制MPC(Model Predictive Contro1),建立了基于神经网络系统辨识的过热汽温模型预测控制系统。仿真结果表明,较之传统的过热汽温串级调节系统,采用该模型预测控制策略的过热汽温控制系统的控制性能有很大的提高。