铝电解关键指标预测方法的研究与应用

为维持铝电解生产的持续性，保证电解槽的物耗稳定和能耗稳定，通过对铝厂数据的挖掘与建模，提出一整套维持电解槽稳定的策略方法，并用于指导实际生产。首先为数据去噪，针对铝厂数据分布特征未知的特点，提出一种无参自适应的模糊聚类方法，通过迭代自适应得到类簇个数和簇中心；根据聚类结果，将铝厂数据按实际意义标签化，提出一种基于距离的连续属性朴素贝叶斯算法，对分类器使用增量思想，使算法动态分类准确率得到提高；应用单槽测试集数据，通过累积法完成当天各指标等级趋势的预测，确定各指标下变量相对于前一天的变化量，完成预测。实验发现，预测模型可完成铝电解关键指标的预测；提出的聚类、分类算法在UCI数据及铝厂数据上表现良好。     

基于功能参数的渐进成形件表面质量表征研究

目的 以圆锥台成形件为例，基于粗糙表面的区域支承率曲线，进行了用功能参数表征表面几何形貌的研究，以全面、准确、定量地表征渐进成形件的表面质量，预测表面性能。方法 用渐进成形工艺加工成形角分别为35°、45°和55°的圆锥台件，将各圆锥台的成形面分为顶部、中部和底部三个区域，并使用三维表面轮廓仪，在上述三个区域中随机测量各4个样本。根据ISO 25178-2:2012，将各样本表面的区域支承率曲线分为峰区、核心区和谷区，选用部分功能参数——核心区高度Sk、峰区材料体积Vmp、核心区材料体积Vmc和谷区空隙体积Vvv表征表面，并与常用的高度表征参数——Sa、Sq、Ssk和Sku进行比较。结果 成形件表面的Ssk和Sku的计算值极不稳定，变异系数（标准偏差与均值之比）最高值分别可达127.67%和39.11%，Sa、Sq的最大变异系数分别为4.41%和4.26%，虽然它们的参数计算值较为稳定，但常无法体现表面形貌的实际差异。功能参数独立表征不同功能区，Sk和Vmc的最大变异系数分别为9.32%和7.73%，说明同一表面各处，在长期工作阶段具有较为一致的表面性能，Vmp和Vvv的最大变异系数分别为60.53%和31.25%，说明表面各处峰、谷区的形貌有所不同，因而在磨合时间、磨粒存储等方面会有不同的表现。结论 粗糙表面的三维本质要求用三维表征参数才能全面表征其几何形貌。高度参数虽然计算简单，但常难以解读它们的物理意义，无法定量评价表面性能；功能参数具有明确的物理意义，可为表征表面质量、预测表面性能提供定量、实用的方法。     

Studying Formability Limits By Combining Conventional and Incremental Sheet Forming Process

In this work it is assessed the potential of combining conventional and incremental sheet forming processes in a same sheet of metal.This so-called hybrid forming approach is performed through the manufacture of a pre-forming by conventional forming,followed by incremental sheet forming.The main objective is analyzing strain evolution.The pre-forming induced in the conventional forming stage will determine the strain paths,directly influencing the strains produced by the incremental process.To conduct the study,in the conventional processes,strains were imposed in three different ways with distinct true strains.At the incremental stage,the pyramid strategy was adopted with differ-ent wall slopes.From the experiments,the true strains and the final geometries were analyzed.Numerical simulation was also employed for the sake of comparison and correlation with the measured data.It could be observed that single-stretch pre-strain was directly proportional to the maximum incremental strains achieved,whereas samples subjected to biaxial pre-strain influenced the formability according to the degree of pre-strain applied.Pre-strain driven by the prior deep-drawing operation did not result,in this particular geometry,in increased formability.     

Part accuracy improvement in two point incremental forming with a partial die using a model predictive control algorithm

As a flexible forming technology, Incremental Sheet Forming (ISF) is a promising alternative to traditional sheet forming processes in small-batch or customised production but suffers from low part accuracy in terms of its application in the industry. The ISF toolpath has direct influences on the geometric accuracy of the formed part since the part is formed by a simple tool following the toolpath. Based on the basic structure of a simple Model Predictive Control (MPC) algorithm designed for Single Point Incremental Forming (SPIF) in our previous work Lu et al. (2015) [1] that only dealt with the toolpath correction in the vertical direction, an enhanced MPC algorithm has been developed specially for Two Point Incremental Forming (TPIF) with a partial die in this work. The enhanced control algorithm is able to correct the toolpath in both the vertical and horizontal directions. In the newly-added horizontal control module, intensive profile points in the evenly distributed radial directions of the horizontal section were used to estimate the horizontal error distribution along the horizontal sectional profile during the forming process. The toolpath correction was performed through properly adjusting the toolpath in two directions based on the optimised toolpath parameters at each step. A case study for forming a non-axisymmetric shape was conducted to experimentally validate the developed toolpath correction strategy. Experiment results indicate that the two-directional toolpath correction approach contributes to part accuracy improvement in TPIF compared with the typical TPIF process that is without toolpath correction.     

五轴数控渐进成形中挤压工具姿态的确定

在基于STL模型的五轴数控渐进成形中,为了避免三角面片边界处挤压工具的姿态发生突变现象,提出一种挤压工具姿态确定方法。首先,根据挤压工具头中心点与三角面片的位置关系,将工具头中心点分为位于三角面片顶点、三角面片边和三角面片内部3种类型;然后,对不同类型的工具头中心点,分别采用不同的距离加权算法,计算工具头中心点的法向量;最后,根据工具头中心点的法向量和挤压工具轴相对于工具头中心点法向量的引导角和倾斜角,确定挤压工具姿态。采用VC++和OPENGL完成算法的系统实现。算法应用实例表明,所确定的挤压工具姿态变化平缓,在三角面片的边界处,没有发生挤压工具姿态突变现象,软件运行稳定可靠。     

An incremental intrusion detection system using a new semi‐supervised stream classification method

In recent years, the utilization of machine learning and data mining techniques for intrusion detection has received great attention by both security research communities and intrusion detection system (IDS) developers. In intrusion detection, the most important constraints are the imbalanced class distribution, the scarcity of the labeled data, and the massive amounts of network flows. Moreover, because of the dynamic nature of the network flows, applying static learned models degrades the detection performance significantly over time. In this article, we propose a new semi‐supervised stream classification method for intrusion detection, which is capable of incremental updating using limited labeled data. The proposed method, called the incremental semi‐supervised flow network‐based IDS (ISF‐NIDS), relies on an incremental mixed‐data clustering, a new supervised cluster adjustment method, and an instance‐based learning. The ISF‐NIDS operates in real time and learns new intrusions quickly using limited storage and processing power. The experimental results on the KDD99, Moore, and Sperotto benchmark datasets indicate the superiority of the proposed method compared with the existing state‐of‐the‐art incremental IDSs.     

A time integration algorithm for a 3D constitutive model for SMAs including permanent inelasticity and degradation effects

Components based on shape‐memory alloys are often subjected to several loading cycles that result in substantial alteration of material behavior. In such a framework, accurate models, as well as robust and efficient numerical approaches, become essential to allow for the simulation of complex devices. The present paper focuses on the numerical simulation of quasi‐static problems involving shape‐memory alloy structures or components subjected to multiple loading‐unloading cycles. A novel state‐update procedure for a three‐dimensional phenomenological model able to describe the saturation of permanent inelasticity, including degradation effects, is proposed here. The algorithm, being of the predictor‐corrector type and relying on an incremental energy minimization approach, is based on elastic checks, closed‐form solutions of polynomial equations, and nonlinear scalar equations solved through a combination of Newton‐Raphson and bisection methods. This allows for an easy implementation of model equations and to avoid the use of regularization parameters for the treatment of nonsmooth functions. Numerical results assess the good performances of the proposed approach in predicting both pseudoelastic and shape‐memory material behavior under cyclic loading as well as algorithm robustness.