故障诊断技术的研究现状与发展趋势

回顾了状态监测与故障诊断技术的发展历史，归纳和总结了基于系统数据模型，基于系统输入输出信号处理以及基于人工智能等途径的故障诊断方法，从混合智能诊断技术，BIT技术，远程协作诊断技术三个方面，对现代故障诊断技术的发展趋势和有待解决的问题进行了分析与探讨，提出了故障诊断领域目前和将来的研究方向。     

超声无损检测的发展趋势

介绍了人工智能技术、自适应技术、机器人技术和相关技术在超声无损检测中的理论分析和应用。这些技术的应用,使得无损检测的定位、定性和定量的可靠性和完备性大幅度提高,为无损评价奠定了良好的判定基础。指出现代超声无损检测技术正向着智能化、自动化、图像化、数字化、小型化、系列化、多功能化、信息化和交叉领域的前沿方向发展,实现了复杂形面复合构件的超声扫描成像无损检测,满足现代质量对无损检测的要求。     

检测时用声信号分析可靠解决工程问题

近来，我们看到了在无损评价领域中已有突飞猛进的发展，对新材料的鉴定和检测，对材料和零件进行严格的检测及评价要求，以及无损评价在材料寿命预测和智能处理方面应用的需要促进了ＮＤＥ的飞速发展。伴随着有助于理解检测方法和材料及零件之间相互关系的模拟技术的发展，计算机、机器人自动化和传感技术是帮助ＮＤＥ专业人员迎接这些挑战的最主要手段。信号分析和成象是两个技术上相关的领域，业已被广泛地用来在价格效果这两个基     

人工智能在轧制领域中的应用进展

人工智能在轧制领域中的应用已有30多年的发展历史,20世纪90年代我国轧钢界曾经出现过一波在学术研究、技术开发和现场应用中使用人工智能技术的高潮。本文介绍了神经网络、专家系统、模糊控制、智能优化算法等人工智能方法在轧制过程中应用的一些实例,涉及轧制力等参数计算、负荷分配、温度控制、宽度厚度控制、板形控制等具体内容,以及涉及原料管理、加热、冷轧和热轧、卷取、产品组织性能预报等,并对今后轧制智能化的发展方向进行了展望。     

连续油管无损检测技术的应用发展

介绍了国内外连续油管无损检测技术,详细介绍了电磁检测技术和超声检测技术的工作原理、发展现状及不足之处。对国外在役无损检测的三种检测装置：CTDOG,CT Scope和Argus作了详细的工作原理分析。结合实际检测中出现的各种问题,提出了目前连续油管无损检测技术方面存在的问题。最后分析现有的技术和存在的问题,得出连续油管无损检测技术的发展方向,即智能化和综合化。     

机器学习在无损检测中的应用案例浅析

近几年来,以机器学习为核心的人工智能快速发展并渗透到很多领域,也受到无损检测同行越来越多的关注。本文在简要介绍机器学习定义及主要算法的基础上,概述了人工神经网络等三种典型机器学习算法在无损检测中的应用情况及适用性,分析了机器学习如何助力无损检测,提高检测结果的准确性以及检测效率。最后,对人工智能时代无损检测的发展提出几点思考和建议。     

中走丝线切割机床智能远程监控系统设计

中走丝线切割机床是一种精密的特种加工机床,加工精度高,能加工各种特殊硬质材料,但是比常规加工方式耗时,因而对其进行实时的状态监控就很有必要。针对现有国内主流的中走丝数控系统,提出2种集成监控方式,并结合嵌入式技术、人工智能、网络及移动终端编程和数据库技术研究了中走丝机床智能远程监控系统。首先,对智能远程监控系统的总体框架进行了阐述;其次,对智能远程监控系统的核心功能及其关键技术进行了详细的说明;最后,对中走丝的2种集成监控方式进行了重点说明,特别是基于人工智能技术的图像识别方法。所研究的中走丝线切割机床智能远程监控系统不仅能为机床加工人员提供方便,也能为机床维护人员解决机床故障提供支撑。     

CNDT&E“云”检测与评价技术

基于云计算技术和检测集成技术的云检测是一个全新的概念,它包容各种物理与化学的检测方法,实现信息共享和远程控制,也是无损检测集成技术发展的趋势。讨论了无损检测技术集成创新带来的优越性,该技术目前的研究现状以及未来的发展方向,即云检测。简要阐述了无损云检测与评价技术计划方案,云检测通过云计算和无损检测集成技术的结合,将传感器采集的数据收集于云端进行存储、处理。云端的功能包括信号处理、存储、评估、预测、信息反馈等一系列软硬件共享资源。用户能够共享软、硬件资源,享受云检测带来的便捷服务和高效率。讨论了无损检测技术建立云检测的可能性,优点和亟待解决的问题,并介绍了一种无损云检测的雏形样机,指出云检测是未来无损检测集成技术的发展方向。     

涡流无损评价的理论与实验方法

本文通过两个重要实例：１任意形状的线圈产生的涡流场分布的计算和磁场测定；２金属管壁中缺陷的再现，介绍了涡流无损评价技术中的正向问题和逆向问题以及一些有关的理论和实验方法。     

超声无损检测的发展趋势

介绍了人工智能技术、自适应技术、机器人技术和相关技术在超声无损检测技术中一些理论分析和应用。指出现代超声无损检测技术正向着智能化、自动化、图像化、数字化、小型化、系列化、多功能化、信息化和交叉领域前沿方向发展，以实现复杂形面复合构件的超声扫描成像无损检测，满足现代质量对无损检测的要求。     

Signal processing for ultrasonic non-destructive evaluation: two applications

The quality control in industrial plants using ultrasonic non-destructive methods is an emerging field for signal processing applications. In this paper we present a general purpose prototype for non-destructive evaluation of materials by means of ultrasonics. The basic prototype could be used in different applications, changing only the signal processing software. In this paper we will show the flexibility of our equipment by means of two different applications: the detection of microcracks in wet ceramic (tiles) and the evaluation of the moisture in cheeses. We describe the necessary signal processing for each application and the results obtained.     

Influence of steel wrapping on magneto-inductive testing of the main cables of suspension bridges

The number of structurally deficient or functionally obsolete bridges is likely to increase in coming years due to continued bridge aging and deterioration and a decrease in funding needed to maintain and repair these structures. In the case of suspension bridges, problems arise due to corrosion and wire breaks within the main cables of the bridge. The magnetic flux leakage (MFL) method has been known for many years and is used for inspection work in different fields, such as the non-destructive evaluation (NDE) of pipelines, arial tramways or stay cable bridges. A special characteristic of the main cables of suspension bridges is the presence of steel wrapping, which is used to compact the cable and hold it in shape. If a NDE by means of magnetic methods is performed on such a cable, disturbances due to the wrapping can be expected in the measured signal. Furthermore, the very large diameters of these cables (up to 500 mm) poses an unsolved challenge in the application of magnetic non-destructive testing methods.The finite element method (FEM) and other simulation techniques are widely used in any field of engineering. The possibility to simulate different scenarios and the ability to quantify physical properties at any position where real measurements cannot be performed are among the main advantages of such methods. In this work, a FE model is presented to investigate the influence of the steel wrapping on MFL data. After the characterization of the disturbance, a postprocessing algorithm to eliminate the disturbances is discussed. Further, boundary conditions for a successful application of the proposed method in a real application are formulated.     

人工智能医疗器械应用中的伦理问题分析

随着人工智能在医疗领域中的迅猛发展，人工智能医疗器械在大幅度提高医疗诊断和治疗效率，提升患者就医便利的同时也带来了医疗安全、数据安全、算法偏见和个体自主等伦理问题。本文通过深入分析人工智能医疗器械应用中的伦理问题，建议通过实施严格的监管和质量控制、减少算法偏见并不断增加透明度、完善技术适应价值多样性，尊重患者和医生的自主权等措施，以期为人工智能医疗器械的监管注入伦理和管控思路。     

Evaluation of NDT methods for detection of prestressing steel damage at post-tensioned concrete structures

For an assessment of the risk potential of existing structures, where in the case of prestressing steel damage the load-bearing capacity could be significantly affected, non-destructive testing techniques which enable reliable evaluation of the condition of the prestressing steels are of utmost importance. During the demolition of a post-tensioned concrete bridge in Berlin where damage of prestressing wires due to hydrogen-induced stress corrosion cracking were found in former investigations, parts of the girders of the bridge superstructure could be taken out. After non-destructive investigations by means of two testing techniques based on magnetic leakage flux measurement and one technique based on electromagnetic resonance measurement, subsequent removal of the concrete cover up to the ducts, opening the ducts and evaluation of the prestressing steels allowed a verification of the results obtained by the NDT methods. From the results it can be concluded that areas with fractures of several wires can be detected by the two techniques based on magnetic leakage flux measurement.     

Intelligent Kano classification of product features based on customer reviews

Product features can be classified into different categories based on customer opinions. The rapid development of artificial intelligence and machine learning paves the way toward computational analysis of customer reviews for opinion mining. This paper presents an Intelligent Kano framework to extract, quantify, and classify different product features based on customer reviews. The framework is enabled by a novel integration of multiple artificial intelligence and machine learning techniques such as sentiment analysis and anomaly detection. A case study is conducted to validate the framework’s effectiveness. Over 12,000 customer reviews on two coffee machines are analyzed for the classification.     

基于群体智能算法的排球高鲁棒性目标识别研究

基于小波变换和人工神经网络的目标识别是图像处理的一个重要研究方向。但是,此类方法采用的梯度下降规则容易产生局部极小值。为了解决该问题,提出了一种基于群体智能算法的高鲁棒性目标识别算法,可有效应用于各种图像识别任务,如排球目标识别等。首先对图像进行预处理并变换成HSV空间进行背景分割,并通过小波不变矩对图像进行特征提取。然后采用新兴的群智能算法-狼群算法,对基于小波神经网络的目标图像识别进行优化,以便提升全局收敛性和鲁棒性。仿真实验结果显示:相比原有的方法,提出优化方法具有更高的识别精度和稳定性。     

An effective genetic algorithm for circularity error unified evaluation

There are four methods commonly used to evaluate the circularity error. They are: minimum zone circle (MZC) method, minimum circumscribed circle (MCC) method, maximum inscribed circle (MIC) method and least square circle (LSC) method. However, so far there is no a robust and effective approach to implement a unified evaluation of these four methods. In this paper, an effective genetic algorithm is presented for searching the above four circularity error evaluation methods simultaneously. The algorithm is implemented in real-code and only blend crossover operators are applied to two randomly selected individuals from the existing population. The algorithm does not require genetic parameters such as crossover and mutation probabilities to be set in advance as does a canonical GA; therefore it is very convenient to use in engineering metrology. The objective function calculation approaches of four circularity errors are developed and the initial population generation methods are given in order to save optimization time. Finally, the experimental results evaluated by different methods confirm that the proposed method can find the optimal solutions of these four methods. In comparison with existing evaluation methods, the algorithm is not only simple and robust, but also it unifies these four kinds of circularity evaluation. The algorithm can also be used for solving difficult form error minimization and profile evaluation problems of various geometric parts in engineering metrology.     

工业人工智能的现状与发展趋势

人工智能技术在生活、网络、服务和金融等领域应用效果显著，而工业人工智能需要实时解决不同生产对象和场景的各种难点问题，还没有广泛深入应用。简述工业人工智能技术的含义、背景、要素和现状，分析其发展历程和典型应用，指出工业智能化在促进工业生产安全、降本、增效、提质等方面的系统框架和要点，探讨其中技术难点和应用所面临的问题，并预测工业人工智能技术的一些发展方向。     

Definition of quality in cast aluminum alloys and its characterization with appropriate indices

In this paper, the issue of “quality” of cast aluminum alloys from various viewpoints is interpreted. Many methods to characterize the quality of materials are available; the methods used currently for the quality evaluation of cast aluminum alloys include nondestructive testing, characterization of the microstructure, and mechanical testing. With regard to mechanical testing, a number of quality indices have been devised to evaluate and characterize the quality of cast aluminum alloys. As these quality indices use different mechanical properties for the quality evaluation, they are expected to lead to different results. In this work, the application of proposed quality indices and their suitability is discussed for a number of situations, including minor variations in chemical composition, different solidification rate, solid solution and artificial aging heat treatments.     

人工智能应用于临床试验的伦理问题分析及对策

人工智能的发展日臻成熟,已经渗透到临床试验的各个领域,人工智能为临床试验带来了新的发展机遇。然而,人工智能应用于临床试验仍然处于探索阶段,面临很多伦理问题,包括数据质量导致的试验风险、数据监管带来的隐私保护以及数据授权与知情同意矛盾等。我们应该精准定位人工智能于临床试验的可实现应用并理解其现实的伦理问题,制定相应的应对策略以保障临床试验全过程性能的最大化改进,包括加强数据质量管理,降低临床试验风险;优化数据监管机制,保障数据安全和隐私;搭建数据授权平台,健全知情同意司法保护等。