CO2 welding inverter power supply and its intelligent fuzzy control

Abstract: Based on the analysis of the control characteristics of CO2 welding process, a constant current IGBT inverter power supply is designed. Under different droplet transfer forms, an intelligent fuzzy control scheme for arc length and short circuit frequency is proposed. Experiments have shown that the use of this technology helps to overcome the shortcomings of CO2 welding and can better achieve arc state control.

Keywords: inverter power supply, fuzzy control, CO2 welding

An Inverter－ type Power Supply and Fuzzy Control for CO2 Arc Welding

Abstract:Based on the foundation of analysis of CO2 welding process,a constant current (CC)IGBT inverter is developed.In case of different droplet－ transfer forms,fuzzy control schemes of arc length and short－ circuit frequency are put forwords.It is indicated by experiments that the existing shortcomings of CO2 welding can be overcome,and are states controlled more easily.

Keywords:Inverter－ type power supply Fuzzy control CO2 welding

CO2 welding is an important welding method with the characteristics of high efficiency and low cost. The quality of traditional CO2 welding is limited by the welding power supply and control method, and there are disadvantages such as large spatter, poor forming and the need to adjust the welding parameters. In recent years, with the progress of arc welding inverters and microprocessor technology, a new CO2 welding control method has been developed to improve the welding quality. At present, there are many schemes, but their effects are still limited and their application is difficult [1-3]. How to reasonably design CO2 welding inverters and develop control technology is the main problem currently faced. This paper discusses this and proposes a technical solution. On the basis of the constant current external characteristic control of the power supply, fuzzy control with self-learning ability is adopted to obtain good experimental results.

1 Disadvantages of simple constant voltage CO2 welding power source

For ordinary CO2 welding, most of them use flat characteristic welding machines with transformer tap adjustment, or constant voltage control and thyristor welding machines, with constant speed wire feeding system for welding. Although a certain arc length self-adjustment ability can be obtained, the spatter is large, the forming is poor, and the process effect is not good.

The reason is related to the physical process of CO2 welding. CO2 welding has two forms: free transition and short-circuit transition, and short-circuit transition is the most common. For free transition, the constant voltage power supply has a strong repulsive effect on the molten droplet transition of the welding wire, causing its deflection and spatter, which is difficult to use. For short-circuit transition, which includes short circuit and arcing, the constant voltage power supply limits the short-circuit current and increases the arcing energy by inserting an inductor in series in the main circuit, but it is difficult to take both into account.

The arc load undergoes changes in no-load, short-circuit and arcing states, and these are all normal working states. In the case of short-circuit, the power supply and load must aim to control the current. In the case of arcing, a more appropriate arc length control method is required for automatic or semi-automatic CO2 welding. Short-circuit transition requires a suitable short-circuit transition frequency to improve process stability. Obviously, the constant voltage power supply does not meet the requirements of the droplet transfer process.

2 Characteristics of constant current inverter power supply

Based on the above analysis, this paper adopts a constant current technology solution to design a CO2 welding inverter power supply. The power supply uses IGBT devices and a single-ended forward circuit with an operating frequency of 20kHz. The circuit principle is shown in Figure 1. The power supply has a constant current closed-loop control system and controllable electronic reactor characteristics to meet the current control of the CO2 welding process. Variable structure control is performed through output state judgment. When the power supply is unloaded and lightly loaded, pulse width control is performed to improve the reliability of the power supply. When a short circuit occurs, the welding current switches to a peak value to ensure that the arc is reignited; when the arc length fluctuates too much, it switches to a small current arc maintenance, and the arc length is restored under the action of a relatively large wire feeding speed.

When the given current and wire feeding speed are within a certain range, the arc is stable without short circuit process, that is, free transition. When the wire feeding speed is high, continuous arc burning and short circuit process will occur, that is, short circuit transition. By adjusting the wire feeding speed, different forms of droplet transition can be obtained, and the arc voltage working point and short circuit transition frequency can be changed. Due to the strong threshold control and constraint of the step characteristic, it has certain adaptive characteristics. Compared with the flat characteristic power supply, it avoids the repulsive effect of the flat characteristic on the droplet transition, making the arc smooth and spattering small. The current and wire feeding speed are adjusted independently, and the combination of the two can control the arc state and weld formation. However, this method still has obvious shortcomings: when the wire feeding speed and the height of the welding gun fluctuate, the effect is not ideal, which is manifested in two aspects, namely, the arc length of the free transition changes, and the short circuit frequency changes during the short circuit transition, thereby affecting the process stability and welding quality.

Figure 1 Constant current welding inverter power supply

(a) 20A, 100μs/grid

(b) 20A, 5ms/grid
Figure 2 Output current dynamic process

3 Intelligent fuzzy control system

In order to solve the shortcomings of constant current power supply, microcomputer control and fuzzy control technology are introduced. The system can carry out different control schemes of free or short circuit accordingly. The control of free transition is relatively simple, and its goal is to maintain a suitable arc voltage to ensure a stable arc length. In addition to controlling the current and arcing voltage of the short circuit process, the short circuit transition also needs to control the short circuit frequency.

The low-cost single-chip microcomputer system in Figure 3 can replace the simple switching of analog electronic circuits, that is, the "microcomputer + simulation" method is adopted to adjust the welding current and dynamic characteristics in real time. The design interface is easy, simple and reliable, and has better controllability. Through process tests, under different working conditions, the applicable range relationship between wire feeding speed and welding current under open-loop conditions, as well as the relationship with dynamics, is established. On this basis, an intelligent control scheme is implemented.

Figure 3 Microcomputer controlled welding inverter power supply

3.1 Arc length control of free transition

According to the roughly selected welding wire feeding speed and current from the open-loop test, the system has unstable arc length due to disturbance during welding. Through welding voltage feedback and fuzzy control adjustment, the controller is constructed and implemented with a single chip.

Assume that the arc voltage deviation is E, the arc voltage deviation change rate is EC, and the welding current adjustment is △I. Select its domain and fuzzy language variables, and determine the membership function according to the normal distribution. The arc voltage deviation takes into account the fluctuation caused by the droplet size, which helps to overcome the arc irritability. Determine the rules based on experience, and then calculate the fuzzy relationship matrix according to the FUZZY reasoning synthesis rule, select the weighted average method to defuzzify, and form the control system query table. It has a fast real-time processing speed, reflecting the characteristics of fuzzy control parallel processing and strong adaptability.

E={NB、NM、MS、NO、PO、PS、PM、PB}

EC={NB、NM、NS、0、PS、PM、PB}

△I={NB、NM、NS、0、PS、PM、PB}

The basic fuzzy design is subjective and should be determined based on experiments. It is not unique and aims to achieve a satisfactory control effect. Since the voltage deviation takes into account the instantaneous droplet transition, it is in line with the actual characteristics. At the same time, the software sets the voltage threshold. When the deviation is too large, the current drops to the arc maintenance current. When it is close to a short circuit, short circuit control is performed.

Figure 4 Arc length fuzzy control principle

3.2 Process control of short-circuit transition

The control of short-circuit transition can be divided into two levels: one is the control of the short-circuit process itself, and the other is to ensure stability through short-circuit frequency control. Here we first discuss the short-circuit process control.

Short-circuit transition is an alternating process of short-circuiting and arcing. During the short-circuit period, the current rise is actually controlled, and during the arcing period, the arc length is controlled, that is, the arc voltage is controlled. Compared with the traditional control method, this is a time-sharing control. The current control principle during the short-circuit period can be carried out according to the physical state of the arc, but in fact, due to the large short-circuit time constant (L/Rsc) of the inverter, it is still difficult to obtain the ideal effect. In order to simplify the design, the current remains unchanged after the short circuit, and the current rise rate is controlled after a delay (1ms). During the arcing period, the arc length is controlled by adjusting the current through arc voltage feedback and fuzzy control, and the scheme is the same as above.

However, due to the randomness of the welding process, it is not enough to control the short-circuit process itself. Under a given arc voltage control, disturbances will cause short-circuit frequency changes. Therefore, it is also necessary to introduce frequency control of short-circuit transition.

3.3 Frequency control of short-circuit transition

With the short-circuit frequency given as the target, the actual short-circuit frequency deviation E, the deviation change rate EC, the fuzzy controller is realized by a single-chip microcomputer, and the control quantity is the arc voltage given value △U. Then, in the above arc voltage control process, the current adjustment of the inverter is realized. This is an indirect control. Considering the low-frequency characteristics of this control, a correction is made about 0.25s, and the calculation amount is small. In order to facilitate flexible adjustment, a fixed fuzzy control table is not used, but a controller with a correction factor is used.

△Ug=αEi＋(1－α)ECi

Where 0<α <1, i covers all quantization levels.

This method is not only simple and flexible, but also has obvious physical significance. It imitates the thinking characteristics of manual control by human operators. The size of the correction factor indicates the weight added to the deviation and the rate of change of the deviation. At the same time, it reflects the continuity and instantaneous single value characteristics of the artificial thinking process, and can avoid the possible non-smoothness of the control rule. The size of the correction factor can be selected through multiple tests and offline adjustments.

Considering the complexity and time-varying characteristics of the influencing factors of the arc process, the fuzzy control system established entirely based on the experiment should introduce a certain degree of adaptability. When the error is large, the system control quantity should increase the proportional component and reduce the differential component to improve the response speed, that is, a larger α should be taken. Otherwise, the opposite is true. Based on this idea, under the condition of certain constraints on the correction factor, the linear interpolation method is used to adjust the control factor.

α=[(αH－αL)Ei/m]＋αL

αH and αL are the maximum and minimum values ​​of α respectively, and m is the number of quantization levels.

It can be seen that this method avoids the complicated calculation process and only needs simple comparison and addition and subtraction, which is convenient for single chip microcomputer implementation. In the above formula, the self-learning purpose is achieved by correcting the control factor in real time.

3.4 Dual-knob adaptive adjustment method

On the welding machine panel, two potentiometers are set. One is the nominal average current, which sets the wire feeding speed. The other is the nominal average voltage, which sets the arc state. When it is free transition, it is the arc length (arc voltage), and when it is short circuit transition, it is the frequency. Its size determines the free transition or short circuit transition welding mode, so as to select different control parameters and strategy methods. The initial value of the welding current is determined by looking up the table, and the welding current is adjusted through intelligent control to achieve control of the arc and droplet transition.

Different from the unified control of ordinary CO2, this system can be selected by the operator, making the arc state controllable instead of being limited to a certain state, while also realizing intelligent adaptive control. It can be called "dual-knob adaptive adjustment".

4 Test results

CO2 gas shielded welding was carried out using this self-developed IGBT inverter power supply. The welding wire was H08Mn2Si copper-clad welding wire with a diameter of 0.8mm. A 200A wire drawing welding gun and a S86A wire push feeder were used for welding tests.

The free transition of fuzzy control improves the stability of arc length, and the arc is smooth, with almost no spatter and good forming during the welding process. It has the best effect in the flat welding position. The short-circuit transition of intelligent fuzzy control has slightly more spatter than the free transition, but the process is stable. The controllable short-circuit frequency is about 80Hz. At lower frequencies, semi-automatic welding can achieve the best results, and fine welds can be obtained without coarse ripples. At high frequencies, the weld penetration is small, the pile height is large, and the welding speed needs to be increased accordingly.

As shown in Figure 6, it is the short-circuit transition current waveform of the constant current CO2 inverter power supply, where (a) and (b) are the waveforms before and after intelligent fuzzy control, respectively.

Figure 5 Short-circuit frequency fuzzy control

(a)

(b)
Figure 6 CO2 inverter power supply welding current waveform

5 Conclusion

Based on the analysis of the control characteristics of CO2 welding process, a constant current IGBT inverter power supply was designed, and then an intelligent fuzzy control scheme for arc length and short circuit frequency was proposed. Experiments show that the use of this technology can help overcome the problems of large spatter, poor forming and parameter adjustment in CO2 welding, and can better achieve arc state control.