【作者】 吴仕鹏；

【导师】 王平； 陈继明；

【作者基本信息】 中国石油大学 ， 检测技术与自动化装置， 2008， 硕士

【摘要】 电火花加工技术在机械制造业中占有十分重要的地位,在很多情况下(如高硬度、高韧性、高脆性、高弹性材料的零件)是唯一有效的加工手段,可以实现常规加工不可能实现或需花费昂贵代价才能实现的加工。随着机械制造业的发展,人们对电火花加工技术的加工效率和加工质量提出了更高的要求。神经网络模糊技术的发展和其独特的控制特性,为电火花加工技术的发展带来了契机。本文深入的研究了现有脉冲电源拓扑结构的优缺点,结合前人的研究成果,提出了一种改进的电流型节能电火花加工脉冲电源的拓扑结构。该电源不仅解决了传统电源电流爬升慢、拖尾长的问题,而且仅需一套直流电源,电能利用率更高、投资更少。然后使用Matlab/Simulink,将本文设计的电源与哈尔滨工业大学设计的PWM控制节能式电火花加工脉冲电源进行了性能对比,从而验证了该电源拓扑结构设计的可行性。针对目前电火花加工过程控制难以建立起精确数学模型的特点,分析实验数据,研究了在线参数对工艺指标的影响,确定采用以电火花加工过程中统计的一段时间内的空载率、正常放电率、不正常放电率(包括电弧及短路放电)、相对于脉冲周期和抬刀周期的正常放电率为输入,在线参数调节量为输出的模糊RBF(radial basis function)神经网络控制器对在线参数进行实时控制,以达到提高电火花加工过程稳定性的目的。利用模糊控制的解耦性,将五输入三输出的控制器解耦为三个多输入单输出的模糊RBF神经网络控制器,简化了设计过程。以电火花加工伺服参考电压模糊RBF神经网络控制器为例,在Simulink仿真试验的基础上,完成了该智能控制器的软硬件设计,该控制器采用双MPU结构,双机间通过双口RAM IDT7007来进行通讯。更多还原

【Abstract】 The Electrical Discharge Machining (EDM) technology plays an important role in the machining manufacturing. Under many circumstances (such as high rigidity, high tenacity, high brittleness material), it is the only effective method which can achieve the machining that general methods cost a lot or can not accomplish. With the development of machining manufacturing, the EDM technology is required to improve. The progress of the Fuzzy Neural Network (FNN) and its unique control characters give the EDM chances to develop.This paper proposes an improved pulse generator based on the analysis of advantages and disadvantages of the existed ones. The pulse generator not only eliminates the rising slope and machining current tail, but also uses only one set of DC power supply. Based on the platform of Simulink, the pulse generator from Haerbin Institute of technology and this newly proposed one are analyzed and compared to validate the newly designed pulse generator’s performance.As for the EDM can not be mathematically modeled and after studying the influences of online parameters to the craft index, this paper designs a real-time adjusting intelligent controller based on the fuzzy RBF Neural-Network theory. The controller takes the rate of correct spark, error spark(including short and electric arc), idle spark and correct spark with regard to the period of raising machining tool and pulse as inputs, and outputs the change of tool-raising period, pulse period and reference voltage so as to level off the EDM process. Using decoupling, the MIMO controller is decoupled into 3 MISO controllers.In order to validate the intelligent controller, this paper takes the gap reference voltage controller for example. The gap reference voltage FNN controller is designed in detail. Using the Simulink, the controller is simulated firstly and accomplished its hardware and software design. The controller is established on the double MPU, and the conmmunation between the 2 MPU is achieved by IDT7007.更多还原