کنترل یادگیری تکراری سیستم تغذیه متناسب الکترو هیدرولیک در ماشین سلتنج برای نوار فلزی زراعی

A non-linear method, iterative learning control (ILC), is proposed to control the electro-hydraulic feeding process of a new slotting machine. The method attempts to acquire high precision of feeding length and trapezoid feedrate, whereas the complexity of iterative learning control algorithm does not increase much more than that of industrial PID controller. After an analysis of the two-way proportional feeding system, a non-linear dynamical mathematic model with system delay, saturation and dead-zone is developed. The computer flowrate control and the ILC controller are investigated in detail. PID controller, ILC of displacement and ILC of feedrate are compared with the dynamical model under the same desired trajectory. It is experimentally found that the proposed control scheme is more effective to improve the tracking accuracy of hydraulic feeding system of the slotting machine than that of fuzzy PID controller.

Precision baiting for metal bar is widely used in industrial production, such as micro-motor shaft, pin and bearing. It is well known that high speed cropping, radial holding differential cropping and blanking under high axis load are effective and economical processes of bar cutting [1], [2] and [3]. Unlike these common shear methods, where a cropping die is used so as to cause huge shearing force and defective fracture surface, a new method was proposed where the metal fatigue vibration was considered as a factor to improve breakability and surface finish and decrease cutting force and tool wear of metal bar cropping [4]. In the fatigue baiting process, ‘V’ shape grooves were cut at specific length intervals in the metal bar to engender stress concentration at the bottom of the grooves. Then the slotted bar was fed to crop under low-cycle vibration and would rupture after certain fatigue periods. It is obvious that the degree of stress concentration at the bottom of the ‘V’ shape groove primarily determines the surface quality of rupture section and vibrating time. Although there are several different types of slotting machines commercially available on the principle of turning or milling for metal bar or tube, they are scarcely applied to fabricate the ‘V’ shape groove. One reason is that these slotting devices with only one cutting machine tool are unable to keep large local stress at these grooves for mass production for a long period. Moreover, a feeding device is necessary if many metal bars of several meters long are slotted continually at constant intervals, but most slotting machines do not possess the device. A slotting machine with PLC control has been developed to cut sharp grooves with specific intervals in metal bars shown in Fig. 1[5], where two tools are used for slotting bar or tube in turn.In the slotting machine, a hydraulic position control system was adopted to clamp and feed the metal bar during the cropping process for its large torque and high response. Many researchers have paid attention to non-linear dynamic behaviors of hydraulic position servo (HPS) [6], which is widely applied to position control or track for its excellent performance and generally consists of an oil cylinder and an electro-hydraulic servo valve. The transfer function of HPS is generally linearized at the operating point and treated as a second-order or third-order system. Several researchers have reported the fuzzy logic and neural network based methods [7] and [8]. Much work has been focused on variable structure control [6] and [9] and adaptive control [10]. Recently, an optimal-tuning PID control has also been devoted to a proportional control system including a 4/3-way proportional valve and a differential cylinder with computer data acquisition system [11]. However, these methods are relatively more complex compared with the widely used PID control and are very difficult to be realized in common industrial PLC with PID function. Moreover, a hydraulic servo also requires high investment on accurate electro-hydraulic servo valve and sensors [12]. In this paper a PID-type iterative learning control (ILC) was adopted for the hydraulic feeding system of the slotting machine and attempted to improve the performance of PID instruction in industrial PLC. An electro-hydraulic two-way flowrate proportional regulating valve and a 4/3-way electromagnetic valve were also adopted to replace the function of electro-hydraulic servo valve partially. Because the 4/3-way electromagnetic valve keeps permanent flow direction in one feeding operation, the direction of feeding displacement essentially cannot be regulated reversely. This operation is completely different from general electro-hydraulic servo. In fact, a pair of stiff-stoppers were used to ensure feeding length with continual oil supply in the feeding cylinder. However, due to the inertia of feeding parts including piston, connect rod, holding cylinder and metal bar, there was serious impact among stoppers and feeding parts which influences the position accuracy of stoppers. The impact noise generated also seriously influenced the worker health during the feeding process. The purpose of this paper was to investigate whether the proposed iterative learning controller would improve the precision of feeding length and diminish impact noise.

An application of ILC algorithm for feeding length and feedrate track of hydraulic feeding system in a new type of slotting machine was proposed in this paper. A non-linear mathematical model of the hydraulic feeding system was constructed using a group of non-linear differential equations where time delay, dead zone and saturation were taken into account. The step response of the two-way proportional feeding system validated the mathematical model and numerically solving method. Industrial PID controller, ILC of displacement and ILC of feedrate were compared based on the mathematical model to ensure the validation of the proposed ILC controller. The simulation shows that the ILC controller of feedrate can maintain an efficient dynamic performance on the proportional feeding system though the time delay of system is not eliminated. Then a simple leading compensator was added to diminish the influence of time delay to the distinct effect in feedrate control. In the ILC experiments of the feeding system of the slotting machine, the precision of the feeding length was close to the desired trajectory by 0.7 mm and the velocity curve also exhibited the gradual acceleration in the beginning phase and reverses in the end phase. The results may lead to the application of ILC method in industrial PLC with PID instructions as an alternative controller and further research is expected to focus on the designing and implementing ILC instruction.