【作者】 李芳；

【导师】 吴毅雄； 华学明；

【作者基本信息】 上海交通大学 ， 材料加工工程， 2008， 博士

【摘要】 脉冲熔化极气体保护焊(GMAW-P)在工业生产中的应用日益广泛。当前,铝合金等材料的广泛应用,由此而带来的新工艺等对焊接设备提出了更高的要求。GMAW-P焊接方法在应用中存在的其中一个工艺问题就是:弧长稳定性。由于GMAW-P焊接过程中,脉冲电流波形不断地在基值与峰值电流之间进行切换,因此弧长也在两个值之间不断变化,很容易出现导电嘴的回烧。如何有效的控制电弧长度是GMAW-P亟待解决的工艺问题之一。GMAW-P熔滴形成、长大及过渡与脉冲参数有着密切的关系,而GMAW-P熔滴过渡过程对GMAW-P焊接工艺性能、焊缝成形和焊接质量有重要影响。如何能有效、精确控制熔滴过渡的形式,实现精确的一脉一滴熔滴过渡过程也是GMAW-P研究的一个重要任务。对于GMAW-P焊接电源,应用环境多为开放式的车间或野外,各种高频信号干扰容易导致数字焊接电源发生故障。目前,GMAW-P数字焊接电源从控制方案设计角度存在以下问题:1)采用单一的CPU,信号高度集中;2)资源独占,程序设计复杂;3)可靠性不高,某一环节的问题可能会导致系统的瘫痪。针对当前GMAW-P数字焊接电源存在的问题,应用模块化理论,计算各零部件之间的相似程度系数,形成模糊相似矩阵,从数字控制角度对GMAW-P焊接电源系统进行了模块划分。系统的主模块分为主回路模块和控制模块。控制模块又包括信息交互模块、过程控制模块、辅助功能模块三大部分。信息交互模块与过程控制模块分别采用DSP和MCU独立控制。过程控制模块以DSP为核心,通过RS-485总线与MCU控制的信息交互模块采用软件握手的方式进行数据通讯,实现对GMAW-P焊接过程中脉冲电流、电弧长度控制。此外,通过CAN总线实现了GMAW-P焊接电源与PC机之间的通讯,方便的进行产品的升级。双CPU设计将管理功能与算法功能合理分配,为优化GMAW-P工艺奠定了基础。采用模块化熵(?)度量系统的模块化程度,其中,p为产品模块化的级数;i为模块的级别。论文分析了GMAW-P焊接电源的故障类型,研究了GMAW-P焊接电源系统的可靠性。采用分层控制技术,构建了GMAW-P焊接电源系统故障树模型。确定了故障传播的逻辑关系,根据传播矩阵来确定各故障节点的排查次序。如果故障一旦发生,从直接导致顶事件发生的第一级节点按照排查次序检查各个节点,快速确定故障源,提高故障的诊断效率。根据故障的关键重要度采用了不同的处理方法,可以有效的节约DSP的软件资源,并且快速的处理异常状况,达到保护焊接电源的目的。理想的GMAW-P一脉一滴熔滴过渡方式可以保证熔滴尺寸的一致性,而脉冲频率与熔滴直径有着密切关系: f = k * v_f/D。一个脉冲周期内焊丝熔化的体积由两部分组成,一部分为发生过渡的熔滴体积φ_d ,另一部分为没有过渡而残留在焊丝端部的熔化金属的体积φ_r。基于对焊丝端部熔化金属体积的考虑,修正了GMAW-P熔滴过渡时熔滴的尺寸与脉冲频率的关系方程: f =εk * v_f/D,(?),ε为考虑熔滴过渡后残余熔滴的体积系数。修正公式为优化选择脉冲参数提供了理论支持。建立了等速送丝GMAW-P系统传递函数,从控制理论角度解释缓降外特性对于GMAW-P电弧弧长控制的必要性。针对GMAW-P焊接的特点,对GMAW-P弧长适应控制展开研究,设计了频率-特性复合弧长控制器。一方面通过脉冲频率控制使弧长保持稳定,同时,控制焊接电源的外特性,以保证脉冲频率在小范围内波动。台阶试验和爬坡试验表明,该控制方法的调节过程快速、稳定。一方面通过电压反馈,调节基值时间;同时,监控每个脉冲周期的平均电流变化情况,根据平均电流的大小来实时优化电压给定值,通过焊接电源外特性控制,以保证维持稳定弧长的条件下脉冲频率的波动范围在±10%以内。在弧长受到扰动的当前脉冲周期内及时调整焊接电压的设定值,对于防止焊丝回烧或抑制短路的发生具有重要的意义。搭建了基于Lab VIEW技术和高速摄像的GMAW-P多信息采集分析系统,将焊接过程中的电信号与图像信号相结合,实现了焊接过程的电流、电压与熔滴过渡形态的量化对应,为分析熔滴过渡过程提供了动态信息。基于此系统平台,进行GMAW-P焊接工艺实验,综合分析熔滴过渡的图像信息及同步对应的电信号,提取了GMAW-P熔滴过渡过程特征信号。在一个脉冲周期内熔滴脱离焊丝过渡是一个电,热,力等相互作用,且伴随着各种干扰,复杂的物理化学过程。其中大量随机的、不确定的影响因素导致了熔滴过渡的随机性,最终体现在电压,电流的具体变化中。实验结果表明,对于多脉一滴GMAW-P熔滴过渡形式,发现在脉冲周期的下降阶段脉冲电压波形上存在一个拐点,定义该点为熔滴过渡电压;熔滴过渡电压的持续时间称为熔滴过渡电压持续时间。熔滴过渡电压及该点的电压变化率与熔滴过渡状态密切相关。熔滴过渡电压持续时间决定了熔滴过渡的状态。熔滴过渡电压持续时间小于1ms,在该脉冲周期内,熔滴发生过渡;如果某脉冲周期内,熔滴过渡电压持续时间大于1ms,则熔滴不过渡。因此通过提取熔滴过渡特征电信号参数,建立特征参数与熔滴过渡预判的数学模型。在一个脉冲周期内,熔滴只有过渡和未过渡两种状态,是典型的二分类变量,常采用对数线性模型。取熔滴过渡状态作为因变量,特征信号参数为自变量,建立了基于Logistic回归判断熔滴过渡预测数学模型: (?)。试验验证结果表明:该预测模型对焊接过程中,脉冲周期内熔滴成功过渡的预测准确率为92.9%;对熔滴未过渡的预测准确率为93.3%,可见所建立的模型能够较好的预测焊接过程中的熔滴过渡情况。更多还原

【Abstract】 Pulsed gas metal arc welding (GMAW-P) becomes more and more important with the application and development of semi-automatic and automatic industrial production. GMAW-P has high efficiency and good quality with wide process scope and good environmental protection effect. Now the wider use of aluminium and aluminium alloy is brought the new process of welding equipment, such as a higher demand. It is important to further improve GMAW-P welding control performance and corresponding process performance by digital control design in welder power source research.One of the problems of GMAW-P application is the arc stability. Since the switch of current waveform from base to pulse, the arc length is fluctuated, too. Metal transfer process is affected by control performance of welding power source, so it presents higher requirement for output characteristic of welding power source. Moreover droplet formation, dimension and transfer form are affected by pulse parameters. The arc length stability and the metal transfer are two urgent problems for GMAW-P.The application environment of GMAW-P welding power source is bad, which are located mostly in workshop or field. The failure of welding power source based on microcontroller is resulted easily from all kinds of disturbances of high frequency signal. There some problems with digital welding power source: 1) A single CPU: the signal is highly concentrated; 2) Resource-exclusive: the program design becomes complicated and low reliable; 3) Lower reliability: one part’s error may lead to the paralysis of the whole system.Considering all above problems that exist in current GMAW-P welding power source, it is a good solution to use the module method. Module division of GMAW-P welding power source system is separated to main circuit module and control module. There are three parts of control module which are information interactive module, process control module and accessorial module. Two CPUs respectively for two modules, one is MCU for information interactive module and the other is DSP for process control module. The communication between DSP and MCU is based on RS485, and the communication between MCU and pc is based on can bus. Dual CPU design makes more rational allocation of the algorithm and management for optimization GMAW-P technology. The module entropy (?) IS discussed to value the system modular degree.The fault tree of GMAW-P welding power source system is built by the hierarchical control technology. Modules of fault trees can be used to reduce the computational cost of basic operation on fault trees. Graph theory (directed graph) which reflects the relation of fault-dissemination obviously divides the fault-nodes to each level use the matrix to mean the relation of the fault-nodes and top-fault node, and calculate the length of the directed side from one node to the top-fault node, we can assurance the weight of each thoroughfare with the key-importance calculated with the inefficiency of the bottom affairs, then we can know which node to be first checked and which one is last. This test matrix is easy to be processing for the computer, avoiding the difficulty of finding the minimums combination and improve the efficiency of diagnosis. The ideal metal transfer mode in GMAW-P is ODOP (one-droplet-one-pulse) which is related to pulse frequency as the function: f = k * v_f/D. But in one pulse period, there two parts of the melton metal, one is the droplet and the other is the melton string at the end of the electrode. So the function is corrected: f =εk * v_f/D, (?). A new method of arc length control is discussed which has double control parameters: static characteristic of welding power source and frequency. The results show that the frequency fluctuation is within 10%.Metal droplet transfer mode which are influenced by the pulse parameters such as pulse current ( I_p), pulse time ( t_p), base current ( I_b) and pulse frequency ( f ) etc. has play an important role in the weld quality. An experimental system has been developed to sample the transient electrical parameters and record the images of droplet transfer process simultaneously which is based on the Lab VIEW virtual instrument and high speed photography technology. Use one trigger signal to start the electrical parameters sample and at the same time the high speed camera records the droplet transfer process. After the welding, all the data can copy into the PC. By this system, how pulse welding parameters affect on metal transfer form was studied on. The system established in this paper provided found support for controlling the droplet transfer form and optimizing the welding parameters.A transition voltage U d in the pulse voltage drop edge of electrical waveform which has been discovered according to the analysis of synchronous electric and image signals. The relationship between the metal transfer mode and the slope of the transition voltage has been proposed. The logistic regression model is designed: (?)The result of experiments is shown that the model can be used in closed-loop controlling of one droplet per pulse transfer mode and also can help to optimize the welding parameters.更多还原