铸造现场监测数据的网络化管理

针对大量铸造检测设备的使用使各种检测信息处理流程复杂化的现状，利用最新的电子技术，提出铸造企业现场检测、监测数据的网络化管理方案，可以实现铸造企业现场检测数据的分散采集，自动记录，实时传输，集中管理，统一分析，多部门信息共享，极大地减少了操作和管理人员的工作量，提高了效率和检查效果。     

钢带冲压件表面硬度不均匀原因分析

采用化学成分分析、金相组织检测及热处理等手段，对我公司生产的Q345热轧钢带产品硬度偏低及分布不均匀的原因进行了分析。结果表明：热轧钢带表面脱碳太深、脱碳量大是造成工件硬度偏低且不均匀的主要原因。     

多传感器信息融合在液压系统智能故障诊断中的应用

针对采矿工程机械液压系统故障诊断方法存在的局限性,提出了一种基于多传感器信息融合的智能故障诊断方尊。该方法采用模糊神经网络融合诊断中心作为故障诊断的执行机构,算法上采用BP算法。通过一实例论证了在液压系统故障诊断中采用多传感器信息融合故障诊断方法比采用单传感器信息故障诊断方法更具有准确性和可靠性。     

电液伺服系统模型跟随自适应控制

提出一种改进的模型跟随自适应控制算法, 该改进算法把系统当前输出误差引入控制信号,同时利用线性神经网络估计系统新的输出误差.利用该改进算法对电液伺服系统进行控制,实践表明该算法是可行的.     

磨削智能预测控制系统研究

建立了一种智能预测控制系统。该系统根据各磨削阶段的特点,在不同阶段分别采用不同的优化控制策略：粗磨阶段,采用在烧伤极限内大进给和变速磨削优化策略;精磨阶段,采用由神经网络预测、模糊逻辑控制的工件尺寸智能优化方法;光磨阶段,采用工件表面粗糙度模糊神经网络预测辨识控制方法。基于神经网络的专家系统,提供各阶段初始磨削加工参数。实验结果表明,该系统在外圆磨削加工中适应性强,可极大地提高磨削质量和效率。     

小样本条件下某机载设备挂架结构的可靠性研究

某机载设备挂架结构的工作应力属于小样本事件，本文针对该结构利用有限元进行应力分析，以分析结果为样本，分别建立了结构工作应力与外载荷及结构尺寸的非线性映射关系。在此基础上形成了应力-强度干涉模型，并对结构进行了可靠性分析，得出了载荷波动对可靠性的影响大于尺寸波动的影响，为挂架结构的可靠性优化设计奠定了基础。     

电液速度伺服RBF模糊自适应控制

由于电液速度伺服系统的非线性和参数的不确定性,难以建立精确的数学模型,本文引入RBF(径向基函数)模糊自适应控制,利用RBF神经网络进行自学习、修改和完善模糊规则,改善其动态性能。仿真结果表明该方法具有较强的自适应和自学习能力,即使对复杂的非线性系统也能取得良好的控制效果。     

Exploring thermal spray gray alumina coating pore network architecture by combining stereological protocols and impedance electrochemical spectroscopy

Complex multiscale pore network architecture characterized by multimodal pore size distribution and connectivity develops during the manufacture of ceramic thermal spray coatings from intra- and interlamellar cracks generated when each lamella spreads and solidifies to globular pores resulting from lamella stacking defects. This network significantly affects the coating properties and their in-service behaviors. De Hoff stereological analysis permits quantification of the three-dimensional (3D) distribution of spheroids (i.e., pores) from the determination of their two-dimensional (2D) distribution estimated by image analysis when analyzing the coating structure from a polished plane. Electrochemical impedance spectroscopy electrochemically examines a material surface by frequency variable current and potential and analyzes the complex impedance. When a coating covers the material surface, the electrolyte percolates through the more or less connected pore network to locally passivate the substrate. The resistive and capacitive characteristics of the equivalent electrical circuit will depend upon the connected pore network architecture. Both protocols were implemented to quantify thermal spray coating structures. Al₂O₃-13TiO₂ coatings were atmospherically plasma sprayed using several sets of power parameters, are current intensity, plasma gas total flow rate, and plasma gas composition in order to determine their effects on pore network architecture. Particle characteristics upon impact, especially their related dimensionless numbers, such as Reynolds, Weber, and Sommerfeld criteria, were also determined. Analyses permitted identification of (a) the major effects of power parameters upon pore architecture and (b) the related formation mechanisms. This article was originally published inBuilding on 100 Years of Success, Proceedings of the 2006 International Thermal Spray Conference (Seattle, WA), May 15–18, 2006, B.R. Marple, M.M. Hyland, Y.-Ch. Lau, R.S. Lima, and J. Voyer, Ed., ASM International, Materials Park, OH, 2006.     

Approach into the use of probabilistic neural networks for automated classification of tool malfunctions in broaching

The condition of broaching tools has crucial importance for the surface quality of the machined components. If undetected, tool malfunctions such as wear, chipping and breakage of cutting teeth can result in severe damage or even scrapping expensive components, with direct implications on increasing the overall manufacturing costs. In contrast with other machining operations, broaching is characterised by non-symmetric distributions of cutting forces vs. time, making more difficult the task of recognising tool malfunctions. The paper reports on a methodology to automatically detect and classify tool malfunctions in broaching. The method was demonstrated through the use of time domain distribution of the push-off cutting force as a key sensory signal to monitor broaching tool condition when machining a nickel-based aerospace alloy. The characteristic features of the sensory signals have been extracted using in-house-developed programs and, afterwards, used to train and test a probabilistic neural network that enables automated classification of tools with fresh, worn, chipped and broken teeth. Inputting new pattern characteristics to the main categories of tool malfunctions, the system successfully classified them even when variations of signal amplitude and ranking of malfunctioned teeth occurred.     

Learn more by training less: systematicity in sentence processing by recurrent networks

Connectionist models of sentence processing must learn to behave systematically by generalizing from a small training set. To what extent recurrent neural networks manage this generalization task is investigated. In contrast to Van der Velde et al. (Connection Sci., 16, pp. 21–46, 2004), it is found that simple recurrent networks do show so-called weak combinatorial systematicity, although their performance remains limited. It is argued that these limitations arise from overfitting in large networks. Generalization can be improved by increasing the size of the recurrent layer without training its connections, thereby combining a large short-term memory with a small long-term memory capacity. Performance can be improved further by increasing the number of word types in the training set.