Self-Learning and Its Application to Laminar Cooling Model of Hot Rolled Strip

The mathematical model for online controlling hot rolled steel cooling on run-out table （ROT for abbreviation） was analyzed, and water cooling is found to be the main cooling mode for hot rolled steel. The calculation of the drop in strip temperature by both water cooling and air cooling is summed up to obtain the change of heat transfer coefficient. It is found that the learning coefficient of heat transfer coefficient is the kernel coefficient of coiler temperature control （CTC） model tuning. To decrease the deviation between the calculated steel temperature and the measured one at coiler entrance, a laminar cooling control self-learning strategy is used. Using the data acquired in the field, the results of the self-learning model used in the field were analyzed. The analyzed results show that the self-learning function is effective.

Self-Learning of Multivariate Time Series Using Perceptually Important Points

In machine learning,positive-unlabelled(PU)learning is a special case within semi-supervised learning.In positiveunlabelled learning,the training set contains some positive examples and a set of unlabelled examples from both the positive and negative classes.Positive-unlabelled learning has gained attention in many domains,especially in time-series data,in which the obtainment of labelled data is challenging.Examples which originate from the negative class are especially difficult to acquire.Self-learning is a semi-supervised method capable of PU learning in time-series data.In the self-learning approach,observations are individually added from the unlabelled data into the positive class until a stopping criterion is reached.The model is retrained after each addition with the existent labels.The main problem in self-learning is to know when to stop the learning.There are multiple,different stopping criteria in the literature,but they tend to be inaccurate or challenging to apply.This publication proposes a novel stopping criterion,which is called Peak evaluation using perceptually important points,to address this problem for time-series data.Peak evaluation using perceptually important points is exceptional,as it does not have tunable hyperparameters,which makes it easily applicable to an unsupervised setting.Simultaneously,it is flexible as it does not make any assumptions on the balance of the dataset between the positive and the negative class.

Where Have Network-based Self-learning Classes Gone?——Reflections & Expectations on the Employment of Network-based Self-learning Classes

To respond to the further development of college English reforms,many universities employed network-based selflearning classes to aid the traditional classroom teaching,especially in teaching listening,but as time went by,some universities gradually gave them up.The paper intends to reflect on the employment of network-based self-learning listening classes,analyz ing the learning with and without its aid,and meanwhile introduce the need to re-employ it,and discuss how we can improve the network-based self-learning classes to help with students' listening.

SELF-LEARNING FUZZY CONTROL RULES USING GENETIC ALGORITHMS

This papcr presents a new genetic algorithms(GAs)-based method for self-learniag fuzzy control rules. An improved GA is used to learn to optimally select the fuzzy membership functions of the linguistic labels in the condition portion of each rule, and to automatically generate fuzzy control actions under each condition. The dynamics of the controlled system is unknown to the GA. The only information for evaluating performance is a failure signal indicating that the controlled system is out of control. We compare its performance with that of other learning methods for the same problem. We also examine the ability of the algorithm to adapt to changing conditions. Simulation results show that such an approach for self-learning fuzzy control rules is both effective and robust.

Mathematical model for cooling process and its self-learning applied in hot rolling mill

Control precision of coiling temperature is one of the key factors affecting the profile shape and surface quality during the cooling process of hot rolled steel strip.For this reason,the core of temperature control precision is to establish an effective cooling mathematical model with self-learning function.Starting from this point,a cooling mathematical model with nonlinear structural characteristics is established in this paper for the cooling process of hot rolled steel strip.By the analysis of self-learning ability,key parameters of the mathematical model could be constantly corrected so as to improve temperature control precision and adaptive capability of the model.The site actual application results proved the stable performance and high control precision of the proposed mathematical model,which would lay a solid foundation to improve the steel product qualities.

Study on intelligent digital welding machine with a self-learning function

A design idea was proposed that it was about intelligent digital welding machine with self-learning and self- regulation functions. The overall design scheme of software and hardware was provided. It was introduced that a parameter self-learning algorithm was based on large-step calibration and partial Newton interpolation. Furthermore, experimental verification was carried out with different welding technologies. The results show that weld bead is pegrect. Therefore, good welding quality and stability are obtained, and intelligent regulation is realized by parameters self-learning.

Self-learning Fuzzy Controllers Based On a Real-time Reinforcement Genetic Algorithm

This paper presents a novel method for constructing fuzzy controllers based on a real time reinforcement genetic algorithm. This methodology introduces the real-time learning capability of neural networks into globally searching process of genetic algorithm, aiming to enhance the convergence rate and real-time learning ability of genetic algorithm, which is then used to construct fuzzy controllers for complex dynamic systems without any knowledge about system dynamics and prior control experience. The cart-pole system is employed as a test bed to demonstrate the effectiveness of the proposed control scheme, and the robustness of the acquired fuzzy controller with comparable result.

Neuron self-learning PSD control for backside width of weld pool in pulsed GTAW with wire filler

In this paper, the weld pool shape control by intelligent strategy was studied. A neuron self-learning PSD controller for backside width of weld pool in pulsed GTAW with wire filler was designed. The PSD control arithmetic was analyzed, simulating experiment by MATLAB software was done, and the validating experiments on varied heat sink workpiece and varied gap workpiece were successfully implemented. The study results show that the neuron self-learning PSD control method can attain a perfect control effect under different set values and conditions, and is suitable for the welding process with the varied structure and coefficients of control model.

The Self-Learning Gate for Quantum Computing

Self-learning is one of the most important scientific methods that helps develop sciences, as it derives from the desire and interests of the individual. However, self-learning loses importance if it does not follow the scientific methodology for building and organizing information. The case becomes harder if the science is new and few scientific sources are available. Quantum computing is one of the new sciences in computer science and needs the support of specialists to develop it. Quantum computing overlaps with many sciences such as physics, chemistry, and mathematics, so any student in one of the previous disciplines may lose the correct self-learning path to find themselves learning the details of another discipline that does not achieve their goals. This article motivates students and those interested in computer science to begin studying the science of quantum computing and choose the same specialization that suits their interests. The article also provides a roadmap for self-learning steps to protect the learner from losing the correct learning path. I have categorized the stages of learning quantum computing into four steps through which all the essential basics can be learned, provided the goals mentioned in each stage which should be achieved. The learning strategy proposed in this article corresponds with individuals’ self-learning rules. Through my personal experience, the proposed learning strategy has proven its effectiveness in building information in an enjoyable scientific way.

A Self-Learning Diagnosis Algorithm Based on Data Clustering

The article describes an approach to building a self-learning diagnostic algorithm. The self-learning algorithm creates models of the object under consideration. The models are formed periodically through a certain time period. The model includes a set of functions that can describe whole object, or a part of the object, or a specified functionality of the object. Thus, information about fault location can be obtained. During operation of the object the algorithm collects data received from sensors. Then the algorithm creates samples related to steady state operation. Clustering of those samples is used for the functions definition. Values of the functions in the centers of clusters are stored in the computer’s memory. To illustrate the considered approach, its application to the diagnosis of turbomachines is described.

基于PIDNN控制的飞行模拟器人感系统

针对飞行模拟器人感系统的高度非线性和易受干扰性,提出一种基于PIDNN(Proportional Integral Differential Neural Network)的控制方案.首先对飞行模拟器人感系统的模型进行分析研究,对它所受到的外界干扰作理论分析,整理出系统的数学模型,再利用PIDNN控制器优良的在线训练、学习和调整功能对该模型进行仿真控制.与传统PID(Propor-tional Integral Differential)控制器相比,PIDNN结构简单、自适应性强、收敛速度快、不会陷入局部极小.仿真结果表明:PIDNN控制系统响应速度快、稳态精度高、具有良好的动静态特性和鲁棒性,满足实时控制的要求.

基于GA-PIDNN的液压弯辊控制系统设计

针对液压弯辊控制系统的时变性、非线性和不确定性等特点,设计利用G A(遗传算法)优化的P I D神经网络(P I D N N)液压弯辊控制系统。P I D N N控制器不仅具有不依赖被控对象数学模型的优点,而且有很好的动态性能,结构简单易于设计。利用G A代替B P算法对P I D N N权值进行优化,克服了B P算法易陷于局部极小的不足。2种优化方法的仿真结果对比表明:G A-P I D N N控制器能够使液压弯辊力快速达到目标值,并且具有较强的抗干扰能力。

三级倒立摆的GA-PIDNN系统辨识

针对典型的不稳定、多变量、非线性、强耦合的三级倒立摆系统,建立了基于GA优化的PID神经网络(GA-PIDNN)辨识结构,完成了GA与BP两种算法的简单对比,并给出了MATLAB仿真结果.结果表明,GA-PIDNN对于非线性三级倒立摆的辨识是有效的,且GA优于BP算法.

基于PIDNN的污水处理系统参数辨识研究

对污水处理系统进行参数辨识,获取合理的模型,这是污水处理系统分析、预测和控制器设计的关键。为此,文中构建了污水处理系统的神经网络模型,赋予了神经元相应的比例、积分和微分功能。并在介绍PIDNN特征及算法的基础上,提出了一种基于PIDNN的参数辨识方法。最后对污水处理系统进行了仿真,仿真结果能够拟合污水处理系统各项指标,证明了该方法切实可行。

粒子群与PIDNN控制器在VSC-HVDC中的应用

海上风电场并网柔性直流输电系统中,双闭环PI调节器常用来控制风场侧和电网侧变流器,该方法较为成熟,但存在采用的调节器过多、且参数整定困难等不足。文中以神经网络中间层至输出层的权值作为粒子群寻优参数,采用粒子群(particle swarm optimization,PSO)算法设计PID神经网络(PID neural network,PIDNN)控制器,并将该控制器用于控制海上风电场柔性直流输电变流器。根据PIDNN的结构特点,经过简单改进,即将输入层至中间层的权值设定为定值,这时粒子群只需优化中间层至输出层权值,能够明显减少粒子维数,并提高训练速度。用训练获得的PIDNN控制器代替传统PI调节器,建立变流器控制系统的传递函数,开展仿真研究。结果表明,基于合作粒子群算法的PIDNN控制器与传统PI调节器相比,系统的瞬态和稳态性能有明显提高;与传统PIDNN和PSO方法相比,训练次数明显减少,为实施在线训练奠定了基础,同时,也为海上风电场柔性直流输电变流器提供了一种可行的控制方案。

RBF和PIDNN在伺服电机模型中的应用比较

为了更好地发挥RBF和PIDNN神经网络的优势,通过对伺服电机模型辨识和控制问题的分析,对RBF和PIDNN网络的应用效果进行了仿真实验的对比研究。结果表明,RBF神经网络结构复杂,参数难以调整,但具有最佳一致逼近能力,辨识效果优于PIDNN;PIDNN结构简单,比例元、积分元和微分元具有类似PID的控制作用,控制效果优于RBF。

基于改进粒子群算法的PIDNN控制器在VSC-HVDC中的应用

针对海上风电场并网柔性直流输电(voltage sourceconverter based high-voltage direct-current,VSC-HVDC)系统比例–积分–微分神经网络(PID neural network,PIDNN)控制器参数寻优过程中存在的问题,提出一种基于限制竞争小生境混沌变异的改进粒子群算法(improved niche chaoticparticle swarm optimization,INCPSO)。该算法中小生境技术引入限制竞争淘汰机制,使其具有良好的全局寻优能力(探索),配合改进的帐篷映射混沌变异算法,可获得局部精细遍历性能(发现)。在解决粒子群算法早熟收敛和搜索精度低等问题的同时,最大程度地平衡了粒子群算法在解空间内的探索和发现能力。给出了VSC-HVDC系统中PIDNN控制器参数寻优INCPSO算法步骤,并进行算例分析验证。仿真结果表明,该算法寻优效率和搜索精度高,鲁棒性好,INCPSO-PIDNN控制器可用于海上风电场柔性直流输电变流器。

Physical neural networks with self-learning capabilities

Physical neural networks are artificial neural networks that mimic synapses and neurons using physical systems or materials.These networks harness the distinctive characteristics of physical systems to carry out computations effectively,potentially surpassing the constraints of conventional digital neural networks.A recent advancement known as“physical self-learning”aims to achieve learning through intrinsic physical processes rather than relying on external computations.This article offers a comprehensive review of the progress made in implementing physical self-learning across various physical systems.Prevailing learning strategies that contribute to the realization of physical self-learning are discussed.Despite challenges in understanding the fundamental mechanism of learning,this work highlights the progress towards constructing intelligent hardware from the ground up,incorporating embedded self-organizing and self-adaptive dynamics in physical systems.

Traffic-Aware Fuzzy Classification Model to Perform IoT Data Traffic Sourcing with the Edge Computing

The Internet of Things(IoT)has revolutionized how we interact with and gather data from our surrounding environment.IoT devices with various sensors and actuators generate vast amounts of data that can be harnessed to derive valuable insights.The rapid proliferation of Internet of Things(IoT)devices has ushered in an era of unprecedented data generation and connectivity.These IoT devices,equipped with many sensors and actuators,continuously produce vast volumes of data.However,the conventional approach of transmitting all this data to centralized cloud infrastructures for processing and analysis poses significant challenges.However,transmitting all this data to a centralized cloud infrastructure for processing and analysis can be inefficient and impractical due to bandwidth limitations,network latency,and scalability issues.This paper proposed a Self-Learning Internet Traffic Fuzzy Classifier(SLItFC)for traffic data analysis.The proposed techniques effectively utilize clustering and classification procedures to improve classification accuracy in analyzing network traffic data.SLItFC addresses the intricate task of efficiently managing and analyzing IoT data traffic at the edge.It employs a sophisticated combination of fuzzy clustering and self-learning techniques,allowing it to adapt and improve its classification accuracy over time.This adaptability is a crucial feature,given the dynamic nature of IoT environments where data patterns and traffic characteristics can evolve rapidly.With the implementation of the fuzzy classifier,the accuracy of the clustering process is improvised with the reduction of the computational time.SLItFC can reduce computational time while maintaining high classification accuracy.This efficiency is paramount in edge computing,where resource constraints demand streamlined data processing.Additionally,SLItFC’s performance advantages make it a compelling choice for organizations seeking to harness the potential of IoT data for real-time insights and decision-making.With the Self-Learning process,the SLItFC model monitors the network traffic data acquired from the IoT Devices.The Sugeno fuzzy model is implemented within the edge computing environment for improved classification accuracy.Simulation analysis stated that the proposed SLItFC achieves 94.5%classification accuracy with reduced classification time.

四旋翼飞行器自适应PIDNN控制研究

针对传统四旋翼PID控制器参数整定困难和控制效果较难达到最优的问题,综合了传统PID控制器工程意义明确、参数整定简单以及神经网络的非线性映射和自学习的优点,构造了四旋翼飞行器神经网络PID(PIDNN)控制器。利用神经网络的非线性映射特点和自学习能力优化了传统PID控制器的控制效果,借助PID控制器的结构,解决了神经网络层数、节点数和连接权重初值选取困难的问题。同时利用自适应调整比例神经元加权系数,增加了系统的响应速度。最后,通过非线性全数值仿真验证了算法的合理性和有效性。