JPH04322881A - Consumable electrode type arc welding machine - Google Patents

Abstract

PURPOSE:To form and output an optimal output signal by predicting switching of the next short circuit arc by using an inference obtained by a fuzzy theory, based on a feedback signal of an output arc state. CONSTITUTION:A fuzzy inference part 1 inputs feedback signals of an output arc state outputted from a current detecting part 5, a voltage detecting part 4, and a short circuit arc period time counting part 3, and infers the next short circuit start and the arc generation time by a fuzzy theory, and a waveform control part 2 constitutes an output waveform, based on an output from this fuzzy inference part 1, and an optimal arc state is obtained.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a consumable electrode type arc welding machine that performs arc welding by automatically feeding a welding wire, which is a consumable electrode, and particularly to a welding power source thereof.

0002

[Prior Art] Conventional technology for consumable electrode type arc welding machines detects changes from an arc to a short circuit and from a short circuit to an arc based on the fed back arc condition.
After detection, the output waveform was controlled by switching between the output waveform during a short circuit and the output waveform during an arc, thereby controlling the welding output.

0003

[Problems to be Solved by the Invention] However, the welding output was controlled by detecting alternating changes between arc and short circuit based on the arc status fed back as described above, and switching the output waveform after detection. Delays in waveform switching may result in an output waveform that is inappropriate for maintaining an optimal arc, which may cause disturbances in the output state, or it may be impossible to predict when a short circuit will start, so even if a minute short circuit occurs, it can be reliably shorted. There have been problems such as the inability to migrate, which causes spatter, and the uneven repetition period of the short-circuit arc, resulting in uneven short-circuit migration, which impairs the appearance of the bead.

In view of the above-mentioned conventional problems, the present invention measures the repetition period of the short circuit arc based on the feedback signal, predicts the next change of the short circuit arc using inference based on fuzzy theory, and generates an optimal output signal. It is an object of the present invention to provide a consumable electrode type arc welding machine that forms and outputs.

[0005]

[Means for Solving the Problems] The consumable electrode type arc welding machine of the present invention includes a current detection section that measures an output current value and outputs it as a current value signal, and a current detection section that measures an output voltage value and outputs it as a voltage value signal. a short-circuit arc period timer that receives a voltage value signal as input, measures the time from the start of the previous short circuit to arc occurrence, and the time from arc occurrence to short circuit, and outputs it as a short-circuit arc period signal; Fuzzy inference is performed according to predetermined rules using the value signal, voltage value signal, and short circuit arc period signal as input, and the time from the previous change point from arc to short circuit to the next short circuit to arc change point or the previous time A fuzzy inference unit that predicts the time from the point of change from short circuit to arc to the next point of change from arc to short circuit and outputs it as an inference signal, and an output waveform that takes the inference signal as input and realizes the optimal arc condition. The present invention is characterized by comprising a waveform control section that forms a waveform signal and outputs it as a waveform signal, and a drive control section that receives the waveform signal and outputs a drive signal that drives an output drive element that controls the arc output of the welding power source.

[0006]

[Operation] According to the present invention, with the above configuration, the repetition period of the short circuit arc is measured based on the feedback signal, and the switching of the next short circuit arc is performed by inference based on fuzzy theory that also takes into account the output current value and voltage value. can be predicted and the output waveform can be controlled to achieve the optimal output arc condition.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. 1 to 3.

In the block diagram shown in FIG. 1, 1 is a fuzzy inference section, 2 is a waveform control section, 3 is a short-circuit arc period measuring section, 4 is a voltage detection section, 5 is a current detection section, 6 is a drive control section, 7 is a driving element, 8 is an input terminal, 9 is a primary rectifier, 10
is a capacitor, 11 is a main transformer, 12 is a secondary rectifier, 1
3 is a reactor, 14 is a shunt, 15a and 15b are output terminals, 16 is a current contact tip, 17 is a welding wire, and 18 is a base material.

Next, each of these constituent elements will be explained in detail.

The current detection section 5 measures the output arc current value using the shunt 14 and outputs a current value signal AS consisting of a real-time value and an average value. The voltage detection unit 4 measures the output arc voltage value between the output terminals 15a and 15b, and outputs a voltage value signal VS consisting of a real-time value and an average value. The short-circuit arc period timer 3 inputs the voltage value signal VS output from the voltage detector 4, and uses the real-time value in the signal to determine whether the state between the welding wire 17 and the base material 18 is short-circuited. Determine whether an arc is occurring,
The time from the start of a short circuit to the occurrence of an arc and the time from the occurrence of an arc to the start of a short circuit are measured and converted into signals, which are output as a short circuit arc period signal FS.

The fuzzy inference section 1 receives the current value signal AS, the voltage value signal VS, and the short circuit arc period signal FS, and calculates the real-time value and average value of the current value signal AS, the average value of the voltage value signal VS, The time from the start of a short circuit to the occurrence of an arc in the short circuit arc periodic signal FS and the time from the occurrence of an arc to the start of a short circuit are taken in as inference propositions, and the time from the start of the next short circuit to the occurrence of an arc is determined using the multiple fuzzy inference method. , or infers the time from the occurrence of an arc to the start of a short circuit, and outputs the inference result as an inference signal LS. This fuzzy inference section 1 can be easily realized by using a microprocessor or the like.

The waveform control section 2 selects and combines the output waveforms of the next short circuit and arc occurrence from an internal memory data bank based on the inference signal LS given from the fuzzy inference section 1, and forms the waveform. Output as a signal. The drive control unit 6 inputs the waveform signal WS, converts it into a drive signal DS, and supplies the drive signal DS to the output drive element 7.

FIG. 2 is an explanatory diagram of the operation of the fuzzy inference section 1.
FIG. 3 is a waveform diagram of a voltage value signal and a current value signal. The inference performed inside the fuzzy inference unit 1 is based on the membership function μ(
x), inference is made using multiple fuzzy inference format,
A conclusion is drawn and output using the AX-MIN composite centroid method. There are six inputs to the inference unit as shown in FIGS. 2(a) and 2(b), and inference is made according to predetermined rules.
As a conclusion, ts1 or ta1 is output. That is, FIG.
In Figure 3, as shown in Figure 2(a), when inferring the current short circuit time Ts1, the previous short circuit time Ts0, the previous arc time Ta0, and the current value at the change point from the previous arc to the short circuit are used. Membership function μ
(x), inference is made according to a predetermined rule, and the current short circuit time inference value ts1 is output as a conclusion. In addition, as shown in FIG. 2(b), when inferring the current arc time Ta1, the previous arc time Ta0, the current short circuit time Ts1, and the current value Ip at the change point from the previous short circuit to the arc
0, current change point current value Ib1 from arc to short circuit, output current average value Ia' and output voltage average value Va' of the previous arc/short circuit period are incorporated into the membership function μ(x), and a predetermined The arc time inferred value ta1 is output as a conclusion.

[0014] The above rule is IF ~ (antecedent part) THEN...
(consequent part), and each input value Ia of the antecedent part
, Ia', Va, Va', Ts0, Ts1, Ta0, I
The membership functions for p0, Ib0, and Ib1 are
For example, as shown in FIG.
edium) and B (Big), and the membership functions of the output values ts1 and ta1 of the consequent part are also the same three variables as above.

In that case, the membership function μ(x) of the fuzzy variable S is x≦b1 :μ(x)=1
b1 ≦x≦b2 :μ(x)=1−(x−b1)
/(b2 - b1) b2 ≦x
:μ(x)=0 Also, the membership function μ(x) of fuzzy variable M is μ(x)=max(0,1−|x−
b2 |/(b3 - b2 )) Also, the membership function μ(x) of fuzzy variable B is x≦b2
: μ(x)=0 b2 ≦x≦b3 : μ
(x)=(x-b2)/(b3-b2)
b3≦x: Each is given by μ(x)=1.

[0016] The above rule is based on the empirical rule of a skilled welding engineer, for example, IF Ia=S and Va=M and T
s0=M and Ta0=
M and Ip0=M and Ib0=M
THEN ts1=MIF Ia'=S and
Va'=M and Ts1=M and
Ta0=M and Ip0=
M and Ib1=M THEN ta1=M
・・・・・・・・・
・
・・・ ・・・・・・・・・

...IF Ia=B and Va=S a
nd Ts0=S and
Ta0=B and Ip0=M and Ib
0=S THEN ts1=B...
・・・
・・・ ・・・
・・・・・・
..., inference is made using these multiple fuzzy inference formats, and a conclusion is drawn using the MAX-MIN composite center of gravity method.
Output.

[0017]

[Effects of the Invention] As explained above, according to the consumable electrode type arc welding machine of the present invention, the start of the next short circuit and the occurrence of the next arc can be predicted in advance based on the current state of the arc being output. In addition to eliminating time delays in waveform control, the timing at which short circuits start can be controlled, eliminating micro short circuits.
Since it is possible to shift to reliable short circuit generation, it is possible to suppress the occurrence of spatter, and it is also possible to achieve a uniform arc with a smooth bead appearance, allowing the operator to easily achieve a stable and optimal arc. This results in great industrial effects.

[Brief explanation of drawings]

FIG. 1 is a block diagram of one embodiment of the present invention.

FIG. 2 is an explanatory diagram of the operation of a fuzzy inference section.

FIG. 3 is a waveform diagram of a voltage value signal and a current value signal.

FIG. 4 is an explanatory diagram of membership functions.

[Explanation of symbols]

1 Fuzzy reasoning part 2 Waveform control section 3 Short circuit arc period timer 4 Voltage detection section 5 Current detection section 6 Drive control section 7 Output drive element

Claims (1)

[Claims] Claim 1: A current detection section that measures an output current value and outputs it as a current value signal; a voltage detection section that measures an output voltage value and outputs it as a voltage value signal; and a voltage detection section that measures an output voltage value and outputs it as a voltage value signal; A short-circuit arc period clock section that measures the time from the start to the arc occurrence and the time from the arc occurrence to the short circuit and outputs it as a short-circuit arc period signal, and a current value signal, voltage value signal, and short-circuit arc period signal are inputted in advance. Perform fuzzy inference according to established rules and calculate the time from the previous arc to short circuit change point to the next short circuit to arc change point, or the time from the previous short circuit to arc change point to the next arc to short circuit change point. A fuzzy inference unit that predicts the time to the change point and outputs it as an inference signal, a waveform control unit that takes the inference signal as input, forms an output waveform to realize an optimal arc state, and outputs it as a waveform signal, and a waveform signal. 1. A consumable electrode type arc welding machine comprising: a drive control unit that outputs a drive signal to drive an output drive element that controls the arc output of a welding power source.