Prediction of corrosion rate for friction stir processed WE43 alloy by combining PSO-based virtual sample generation and machine learning

The corrosion rate is a crucial factor that impacts the longevity of materials in different applications.After undergoing friction stir processing(FSP),the refined grain structure leads to a notable decrease in corrosion rate.However,a better understanding of the correlation between the FSP process parameters and the corrosion rate is still lacking.The current study used machine learning to establish the relationship between the corrosion rate and FSP process parameters(rotational speed,traverse speed,and shoulder diameter)for WE43 alloy.The Taguchi L27 design of experiments was used for the experimental analysis.In addition,synthetic data was generated using particle swarm optimization for virtual sample generation(VSG).The application of VSG has led to an increase in the prediction accuracy of machine learning models.A sensitivity analysis was performed using Shapley Additive Explanations to determine the key factors affecting the corrosion rate.The shoulder diameter had a significant impact in comparison to the traverse speed.A graphical user interface(GUI)has been created to predict the corrosion rate using the identified factors.This study focuses on the WE43 alloy,but its findings can also be used to predict the corrosion rate of other magnesium alloys.

基于多目标PSO混合优化的虚拟样本生成

受限于检测技术难度、高时间与经济成本等原因,难测参数的软测量模型建模样本存在数量少、分布稀疏与不平衡等问题,严重制约了数据驱动模型的泛化性能.针对以上问题,提出一种基于多目标粒子群优化(Multi-objective particle swarm optimization, MOPSO)混合优化的虚拟样本生成(Virtual sample generation, VSG)方法.首先,设计综合学习粒子群优化算法的种群表征机制,使其能够同时编码用于连续变量和离散变量;然后,定义具有多阶段多目标特性的综合学习粒子群优化算法适应度函数,使其能够在确保模型泛化性能的同时最小化虚拟样本数量;最后,提出面向虚拟样本生成的多目标混合优化任务以改进综合学习粒子群优化算法,使其能够适应虚拟样本优选过程的变维特性并提高收敛速度.同时,首次借鉴度量学习提出用于评价虚拟样本质量的综合评价指标和分布相似指标.利用基准数据集和真实工业数据集验证了所提方法的有效性和优越性.

面向复杂工业过程的虚拟样本生成综述

用于复杂工业过程难测运行指标和异常故障建模的样本具有量少稀缺、分布不平衡以及内涵机理知识匮乏等特性.虚拟样本生成(Virtual sample generation,VSG)作为扩充建模样本数量及其涵盖空间的技术,已成为解决上述问题的主要手段之一,但已有研究还存在缺乏理论支撑、分类准则与应用边界模糊等问题.本文在描述复杂工业过程难测运行指标和异常故障建模所存在问题的基础上,梳理虚拟样本定义及其内涵,给出面向工业过程回归与分类问题的VSG实现流程;接着,从样本覆盖区域、实现流程与推广应用等方向进行综述;然后,分析讨论VSG的下一步研究方向;最后,对全文进行总结并给出未来挑战.

Automatic Generation Control with Virtual Synchronous Renewables

As synchronous generators(SGs)are gradually displaced by renewable energy sources(RESs),the frequency stability of power systems deteriorates because RESs,represented by utility-scale solar and wind power sources,do not provide the inertial response,primary frequency response,secondary frequency response,and tertiary frequency regulation.As a result,the remaining SGs may not be sufficient to maintain the power balance and frequency stability.The concept and control strategies of virtual synchronous generators(VSGs)enable the inverter-based wind and solar power sources to emulate the outer characteristics of traditional SGs and participate in the active power and frequency control of power systems.This paper focuses on the automatic generation control(AGC)with virtual synchronous renewables(VSRs).First,the VSR strategy that enables the RESs to participate in AGC is introduced.Second,based on the interval representation of uncertainty,the output of RES is transformed into two portions,i.e.,the dispatchable portion and the stochastic portion.In the dispatchable portion,the RESs can participate in AGC jointly with SGs.Accordingly,a security-constrained economic dispatch(SCED)model is built considering the RESs operating in VSR mode.Third,the solution strategy that employs the slack variables to acquire deterministic constraints is introduced.Finally,the proposed SCED model is solved based on the 6-bus and 39-bus systems.The results show that,compared with the maximum power point tracking(MPPT)mode,VSRs can participate in the active power and frequency control jointly with SGs,increase the maximum penetration level of RESs,and decrease the operating cost.

基于主动学习机制GAN的MSWI过程二噁英排放风险预警模型

针对构建城市固废焚烧(municipal solid waste incineration,MSWI)过程剧毒污染物二噁英(dioxin,DXN)排放风险预警模型的样本极为稀少的问题,提出一种基于主动学习机制生成对抗网络(generative adversarial network,GAN)的DXN排放风险预警建模方法.首先,以DXN风险等级作为条件信息使得GAN生成候选虚拟样本;然后,利用基于最大均值差异和多视角可视化分布信息的主动学习机制进行虚拟样本的初筛和评估,以获得期望虚拟样本;最后,基于混合样本构建DXN排放风险预警模型.通过基准数据集和MSWI过程数据集验证了所提方法的有效性.基于主动学习机制GAN的DXN排放风险预警建模方法可以有效解决样本稀少的问题,提高模型精度.

An Improved Virtual Inertia Algorithm of Virtual Synchronous Generator

Virtual synchronous generator(VSG)simulates the first-order motion equation of a synchronous generator(SG)with the algorithm.VSG can improve the system voltage and frequency support capabilities of a microgrid or a weak grid.It is now widely applied at a high penetration level of distributed generation(DG)systems.However,because there is a contradiction between active power steady-state deviation of VSG and dynamic impact regulation,the VSG running in grid-connected mode with existing strategies cannot meet the steady and dynamic control requirements.Thus,an improved virtual inertial control strategy of VSG is proposed in this paper.The active power impact is reduced effectively under the circumstance of damping coefficient Dωequal to 0 and a large inertia,thus the dynamic characteristic of active power is improved and its steady-state characteristic is maintained.Firstly,based on the analysis of the damping coefficient effect on the system dynamic process,two forms of improved virtual inertia algorithms are put forward by cascading a differential link into different positions of the first-order virtual inertia forward channel.Then,by comparing the characteristics of the system with the two improved algorithms,the improved virtual inertial strategy based on differential compensation is proven to be better,and the design of its parameters is analyzed.Finally,simulation and experimental results verify the effectiveness of the proposed algorithm.

基于虚拟样本生成技术的多组分机械信号建模

采用具有多组分、非平稳、非线性等特性的机械振动/振声信号构建数据驱动软测量模型,是目前工业界测量高能耗旋转机械设备内部难以检测过程参数的常用手段.针对机械信号产生机理的复杂性导致模型解释性弱,以及工业过程连续不间断运行和机械设备旋转封闭的特殊性导致获取完备训练样本的经济性差和周期性长等问题,本文提出一种基于虚拟样本生成(Virtual sample generation,VSG)技术的多组分机械信号建模方法.首先,将机械信号自适应分解为具有不同时间尺度的平稳子信号并变换为多尺度谱数据;接着,采用适合于小样本高维数据建模的改进选择性集成核偏最小二乘(Selective ensemble kernel partial least squares,SENKPLS)算法构建面向真实训练样本的基于可行性的规划(Feasibilitybased programming,FBP)模型,提出一种综合先验知识和FBP模型等手段面向高维谱数据的VSG技术,用以弥补真实训练样本的短缺问题;然后,基于互信息(Mutual information,MI)对由真实和虚拟训练样本组成的混合建模数据进行自适应特征选择;最后,基于约简的混合训练样本采用SENKPLS构建软测量模型.以近红外谱数据和磨矿过程实验球磨机的筒体振动/振声信号验证所提VSG技术和面向多组分机械信号建模方法的合理性和有效性.