CN115302044A - System and method for torch weaving - Google Patents

Abstract

A robotic arc welding system comprising: a welding torch; a welding robot configured to manipulate the welding torch during a welding operation; a robot controller operatively connected to the welding robot to control the welding torch to make an oscillating movement along a weld and at an oscillating frequency and an oscillating period; and a welding power supply operatively connected to the welding torch to control a welding waveform and operatively connected to the robot controller to communicate therewith. The welding power source is configured to sample a plurality of welding parameters and form an analysis package during a sampling period of the welding operation, and process the analysis package to generate a weld quality score, wherein the welding power source obtains the weaving frequency or the weaving period, and automatically adjusts the sampling period for forming the analysis package based on the weaving frequency or the weaving period.

Description

System and method for torch oscillation

Cross Reference to Related Applications

This application claims priority from U.S. patent application 17/307,040, filed on 4/5/2021, the disclosure of which is incorporated herein by reference.

Background

Technical Field

The present invention relates to robotic arc welding, and in particular to weaving welding performed by a welding robot.

Background

It is known to calculate a quality index during welding to provide information about the acceptability of the resulting weld. It is also known to perform arc-through seam tracking during robotic weaving welding. However, the accuracy of arc through weld tracking may be negatively affected by using welding waveforms with pulsed portions because the welding current level is frequently changed.

Disclosure of Invention

The following summary presents a simplified summary in order to provide a basic understanding of some aspects of the devices, systems, and/or methods discussed herein. This summary is not an extensive overview of the devices, systems, and/or methods discussed herein. It is not intended to identify key or critical elements or to delineate the scope of such devices, systems, and/or methods. Its sole purpose is to present some concepts in a simplified form as a prelude to the more detailed description that is presented later.

In accordance with one aspect of the present invention, a robotic arc welding system is provided. The system comprises: a welding torch; a welding robot configured to manipulate the welding torch during a welding operation; a robot controller operatively connected to the welding robot to control the welding torch to make an oscillating movement along a weld and at an oscillating frequency and an oscillating period; and a welding power supply operatively connected to the welding torch to control a welding waveform and operatively connected to the robot controller to communicate therewith. The welding power supply is configured to sample a plurality of welding parameters and form an analysis package during a sampling period of the welding operation, and process the analysis package to generate a weld quality score, wherein the welding power supply obtains the weaving frequency or the weaving period, and automatically adjusts the sampling period used to form the analysis package based on the weaving frequency or the weaving period.

In some embodiments, the sampling period used to form the analysis packets is equal to the wobble period. In certain embodiments, the welding power supply receives the swing frequency from the robot controller. In certain embodiments, the welding power supply receives the weaving cycle from the robot controller. In certain embodiments, the welding power supply receives torch position information from the robot controller during the welding operation. In a further embodiment, the welding power supply automatically adjusts the sampling period used to form the analysis package based on the torch position information. In certain embodiments, the welding power supply records arc penetration weld tracking information corresponding to the welding quality score. In certain embodiments, the system includes a through arc weld tracking logic that tracks the weld and calculates a correction to a weld path from welding current data classified as corresponding to one of a pulsed current portion and a low current portion of a welding waveform.

According to another aspect of the present invention, a robotic arc welding system is provided. The system comprises: a welding torch; a welding robot configured to manipulate the welding torch during a welding operation; a robot controller operatively connected to the welding robot to control the welding torch to make an oscillating movement along a weld and at an oscillating frequency and an oscillating period; and a welding power supply operatively connected to the welding torch to control a welding waveform and operatively connected to the robot controller to communicate therewith. The welding power supply is configured to sample a plurality of welding parameters and form an analysis package during a sampling period of the welding operation, and process the analysis package to generate a weld quality score, wherein the welding power supply receives torch position information and automatically adjusts the sampling period used to form the analysis package based on the torch position information.

In certain embodiments, the sampling period used to form the analysis packets is equal to the wobble period. In certain embodiments, the welding power supply receives the swing frequency from the robot controller. In certain embodiments, the welding power supply receives the weaving cycle from the robot controller. In certain embodiments, the welding power supply records arc penetration weld tracking information corresponding to the welding quality score. In certain embodiments, the system includes a through arc weld tracking logic that tracks the weld and calculates a correction to a weld path from weld current data classified as corresponding to one of a pulsed current portion and a low current portion of a welding waveform.

According to another aspect of the present invention, a robotic arc welding system is provided. The system comprises: a welding torch; a welding robot configured to manipulate the welding torch during a welding operation; a robot controller operatively connected to the welding robot to control the welding torch to make an oscillatory movement along a welding path that follows a weld seam; and a welding power supply operatively connected to the welding torch to control a welding waveform and operatively connected to the robot controller to communicate therewith. The welding waveform includes a pulsed current portion and a low current portion having a lower current level than the pulsed current portion. The robotic arc welding system includes a through-arc weld tracking logic that tracks the weld and calculates a correction to the welding path. The arc penetration weld tracking logic calculates a correction to the welding path based on welding current data classified as corresponding to one of a pulsed current portion and a low current portion of the welding waveform and provided by the welding power source.

In certain embodiments, the welding current data is classified by the welding power source as corresponding to one of a pulsed current portion and a low current portion of the welding waveform. In a further embodiment, the welding current data is obtained by filtering welding current measurements from corresponding portions of the welding waveform. In certain embodiments, the welding current data is classified by the welding power source as corresponding to either of a pulsed current portion and a low current portion of the welding waveform. In a further embodiment, the welding current data is segmented according to either of a pulsed current portion and a low current portion of the welding waveform. In a further embodiment, the welding current data is labeled as corresponding to either of a pulsed current portion and a low current portion of the welding waveform. In certain embodiments, the welding power source is configured to sample a plurality of welding parameters during a sampling period of the welding operation and form an analysis package, and process the analysis package to generate a weld quality score, wherein the welding power source obtains a swing frequency or swing period of a swing movement and automatically adjusts the sampling period for forming the analysis package based on the swing frequency or swing period.

Drawings

The above and other aspects of the present invention will become apparent to those skilled in the art to which the present invention relates upon reading the following description with reference to the accompanying drawings, in which:

FIG. 1 is a combined block diagram and computer flow diagram or program illustrating a monitor for an electric arc welder, according to one exemplary embodiment;

FIG. 2 is a current command graph from a wave shape generator showing a command wave shape divided into time segments or states having both fixed and variable durations, according to an exemplary embodiment;

FIG. 3 is a current graph of an actual command signal for arc current with an actual arc current parameter superimposed in dashed lines in accordance with an exemplary embodiment;

FIG. 4 is a block diagram for monitoring signals internal to a welder other than the welding parameters as illustrated in FIGS. 2 and 3, in accordance with an exemplary embodiment;

FIG. 5 is a time-based graph illustrating the wave shape, wire feeder command signals, and the actual wire feeder command signals experienced in the exemplary embodiment shown in FIG. 4;

FIG. 6 is a portion of a parametric curve that demonstrates level monitoring characteristics in accordance with an example embodiment;

FIG. 7 is a block diagram and computer flow diagram or program illustrating a process for stability during selected states of the wave shape shown in FIGS. 2 and 3, according to an exemplary embodiment;

FIG. 8 is a block diagram and computer flow diagram or program for processing information from the level monitor stage of the exemplary embodiment shown in FIG. 1;

FIG. 9 is a flowchart illustrating a weighting method for weighting sampled weld data parameters, in accordance with an exemplary embodiment;

FIG. 10 is a schematic view of a robotic welding system;

FIG. 11 illustrates a weaving welding pattern;

FIG. 12 shows measurement positions of through-arc weld tracking;

FIG. 13 illustrates an example pulsed output current welding waveform;

FIG. 14 illustrates an example pulsed output current welding waveform; and

fig. 15 illustrates an embodiment of an example controller of a welding power supply.

Detailed Description

While the general inventive concepts may be susceptible to embodiment in many different forms, there is shown in the drawings and will herein be described in detail specific embodiments thereof with the understanding that the present disclosure is to be considered as an exemplification of the principles of the general inventive concepts. Accordingly, the general inventive concepts are not intended to be limited to the specific embodiments shown herein. Further, the disclosure of U.S. patent No. 8,987,628, issued 3/24/2015, is incorporated herein by reference in its entirety.

The following are definitions of exemplary terms used throughout this disclosure. Both singular and plural forms of all terms fall within their respective meanings:

"logic", as used herein synonymous with "circuitry", includes but is not limited to hardware, firmware, software and/or combinations of each for performing one or more functions or one or more actions. For example, based on a desired application or needs, logic may include a software controlled microprocessor, discrete logic (e.g., an Application Specific Integrated Circuit (ASIC)), or other programmed logic device. In some examples, logic may also be fully implemented as software.

As used herein, "software" or "computer program" includes, but is not limited to, one or more computer-readable and/or executable instructions that cause a computer or other electronic device to perform functions, actions, and/or behave in a desired manner. The instructions may be embodied in various forms such as routines, algorithms, modules, or programs, including individual applications or code from dynamically linked libraries. Software may also be implemented in various forms such as a stand-alone program, a function call, a servlet (servlet), an applet, instructions stored in a memory, part of an operating system or other types of executable instructions. It will be appreciated by those of ordinary skill in the art that the form of software depends on, for example, the requirements of the desired application, its operating environment, and/or the desires of the designer/programmer or the like.

As used herein, "computer," "processing unit," and "processor" include, but are not limited to, any programmed or programmable electronic device that can store, retrieve, and process data.

As used herein, "at least one," "one or more," and/or "are open-ended expressions that are both conjunctive and disjunctive in operation. For example, each of the expressions "at least one of A, B and C", "at least one of A, B or C", "one or more of A, B and C", "one or more of A, B or C", and "A, B and/or C" refers to a alone, B alone, C, A and B together, a and C together, B and C together, or A, B and C together. Any disjunctive word and/or phrase presenting two or more alternative terms, whether in the description of the embodiments, claims, or drawings, should be understood to encompass the following possibilities: including one of these terms, any one of these terms, or all of them. For example, the phrase "a or B" should be understood to include the following possibilities: "A", or "B", or "A and B".

Referring now to the drawings, FIG. 1 shows a block diagram and a flowchart or program implemented by a standard on-board controller (e.g., microprocessor, microcontroller, computer, etc.) in an electric arc welder 10. For example, the welder 10 may be a Power Wave, inverter-based Electric arc welder sold by The Lincoln Electric Company of Cleveland, ohio, cleveland. The welder 10 includes three phase electrical inputs L1, L2, L3 for directing current to the power source 12 in accordance with standard techniques. The on-board computerized controller operates the inverter-based power source to produce a positive potential at terminal 14 and a negative potential at terminal 16.

The selected arc welding process is performed by directing the selected, previously determined wave shape to the actual welding circuit, which is shown as having a standard smoothing inductor 18. Welder 10 performs the arc welding process between advancing wire 20 from reel 22, which is driven at a desired rate by feeder 24 operating at the speed of motor 26. The heat of the arc melts the wire 20 and the workpiece 30 to deposit molten metal from the wire onto the workpiece. In order to monitor the actual parameters of the welding process, the splitter 32 provides an output signal I on a line 34a from a block 34 _a . This signal represents the actual arc current at any given moment. In a similar manner, block 36 senses the voltage between the wire 20 and the workpiece 30, and thus the output V on line 36a _a Is the instantaneous arc voltage that constitutes the second welding parameter. Shown in FIG. 1The welding parameter is the actual arc current I _a And the actual arc voltage V _a 。

Another parameter controlled for the practice of the present invention is Wire Feed Speed (WFS) produced by rotation of the motor 26. Thus, the three externally readable welding parameters of the welding process are the arc current I on line 34a _a The arc voltage V on line 36a _a And a readable wire feed speed WFS on line 46b as explained below. The WFS on line 46b is read by a tachometer or encoder 46c connected to the drive roller 24 of the feeder gearbox, or alternatively on a driven wheel attached to the welding wire. In fig. 1, the tachometer is shown as being driven by these feed rollers. The tachometer may also be driven by the output shaft of the motor 26, for example.

The Power Wave electric arc welder includes a Wave shape generator for creating a series of rapidly repeating Wave shapes, each Wave shape (e.g., a single sequence of voltage/current waveforms) constituting a welding cycle having a cycle time. These welding cycles are repeated during the welding process to define the welding time. One embodiment of a Power Wave welder 10 is shown in U.S. Pat. No. 5,278,390 to Blankenship, which controls the independent Wave shapes to be output by the Power source 12 via command line 42 and the speed of the motor 26 via command line 44. The command line 44 has a signal that is recognized by a microprocessor on the wire drive control 46 of the motor 26 to output a motor voltage drive PWM pulse in line 46 a. In practice, the information on line 44 is digital and the command signal on line 46a is analog. The wave shape generator 40 creates digital signals in lines 42, 44 to control the desired welding process to be performed by the welder 10. External parameter I _a 、V _a And WFS can be read by a suitable monitoring device.

The wave shape generator 40 divides or segments each of these output wave shapes into a series of time-segmented portions or states. In an exemplary embodiment, the monitor M is a program loaded into, inter alia, a computer of the welder 10 for monitoring the wavesThe parameters are read during a selected segment of the shape. The monitor M may be implemented using software, hardware, and combinations thereof without departing from the spirit and scope of the general inventive concepts. The monitored portion of the wave shape is determined by a wave shape generator 40. Indeed, the monitor M monitors a plurality of different time segments or states of the wave shape output by the generator 40. In practice, the wave shape generator 40 selects several of these time segments forming the wave shape and outputs these different states to the command interface 70. Thus, the command interface 70 causes the measurement of these parameters to be made during a selected time segment of each wave shape output by the generator. The information or data on the command interface 70 includes the state or states being monitored and the various parameters I _a 、V _a And/or a specific value or level of WFS.

The interface 70 of the monitor M contains data identifying the particular state being processed and the value of the welding parameter being read. The data in the interface 70 is analyzed by a level (level) stage 81 to determine the relationship of the parameters on a level basis. These actual parameters are compared to trained or measured parameters during selected states of the wave shape from the generator 40. During a particular segment or state of the wave shape, the level monitor stage 81 reads these actual parameters in lines 34a, 36a and 46 b. The instantaneous values of these actual parameters are stored in an internal memory (labeled reporting logic 82). Reading these actual parameters occurs quickly as indicated by oscillator 84. In one exemplary embodiment, for pulse welding, reading these actual parameters occurs at a rate of 120kHz. The rate may be adjusted; however, the higher the rate, the better the sensitivity of the level measurement. The level monitor 81 also determines the deviation of these actual welding parameters from a minimum or maximum level. In this way, not only these actual values, but also data representing the deviation of the actual reading of the parameter from the minimum or maximum level for a given state can be stored. The report memory or logic 82 records the deviation from a set level during a given state of the wave shape, and the actual level during a selected state of the wave shape. These readings are accumulated, counted, or otherwise processed for the entire welding cycle to determine the quality of the weld and any trends in weld defects.

In one exemplary embodiment, the readings (e.g., a periodically accumulated set of the readings) are weighted based on a plurality of criteria. These readings may be accumulated, for example, every 250ms. In one exemplary embodiment, the set is weighted based on its deviation value from a desired value (e.g., a predetermined threshold, median) and its time contribution of the time segment to the corresponding wave shape. Such a weighting method (e.g., the weighting method 900 shown in fig. 9 and described below) may be implemented, for example, in the level monitor stage 81 or any similar or related data processing stage.

The stability monitor stage 91 reads these actual weld parameters on lines 34a, 36a and 46b at a fast rate determined by oscillator 94. In one exemplary embodiment, for pulse welding, reading these actual parameters occurs at a rate of 120kHz. The stability monitor stage 91 analyzes the standard or absolute deviation of the actual welding parameters during the outputting state of the wave shapes. The report memory or logic 92 records this deviation during a given state of the wave shape, as well as the actual values during a selected state of the wave shape. These readings are accumulated, counted, or otherwise processed for the entire welding cycle to determine the quality of the weld and any trends in weld defects.

In one exemplary embodiment, the readings (e.g., a periodically accumulated set of the readings) are weighted based on a plurality of criteria. These readings may be accumulated, for example, every 250ms. In one exemplary embodiment, the set is weighted based on its deviation value from a desired value (e.g., a predetermined threshold, median) and its time contribution of the time segment to the corresponding wave shape. Such a weighting method (e.g., weighting method 900 shown in fig. 9 and described below) may be implemented, for example, in the stability monitor stage 91 or any similar or related data processing stage.

Several wave shapes can be skipped when using either monitor stage 81 or monitor stage 91. In one exemplary embodiment, after the start sequence, all wave shapes are monitored in order to analyze the actual welding parameters during the different selected states of the wave shape. Several states of a given wave shape during the welding process are monitored and the results are recorded separately for each state of consistency, trend and stability of the level to be analyzed. In measuring stability, I is evaluated in monitor M using a standard deviation algorithm _a 、V _a And/or WFS. This information can be used to analyze each of the individual segments that form the wave shape for the entire welding cycle with a given cycle time. In practice, certain conditions, such as peak current during a pulse wave shape, are monitored in order to determine the stability and horizontal deviation of the pulse welding process. During STT welding, the monitor M records the short circuit time for each wave shape as these segments change in time according to the external conditions of the welding process. The change in the short-circuit time tells the welding engineer that an adjustment is to be made.

The series of rapidly repeating wave shapes generated by the standard wave shape generator 40 is divided into a plurality of time states, as shown in fig. 2 and 3. The output current command wave shape is a pulse wave shape 100 having a peak current 102 with a fixed duration shown in fig. 3 as time segment a and a background current 104 with a variable duration shown in fig. 3 as segment B. Dividing the wave shape into at time t ₁ -t ₄ Such that the command interface 70 receives the particular state being processed by the generator 40 at any given time. As shown in fig. 3 by dashed line 110, the actual arc current from shunt 33 of fig. 1 deviates from the command current signal for wave shape 100.

Reading the actual arc current I during a selected functional state, such as state A or state B, at a rate determined by oscillator 84 or oscillator 94 _a . In practice, this is a single software oscillator. Level monitor stage 81 records actual parametersThe deviation in the coordinate direction between the number 110 and the command level of the wave shape 100. During the selected state, the stability monitor stage 91 reads the statistical standard deviation of the actual parameter. States a and B are typically monitored for a pulse welding process. However, t can be monitored ₁ -t ₂ During ramp-up state and/or t ₃ -t ₄ During which the state of descent is ramped in order to control or at least read the activity of said actual parameter during these states of the shape of said wave. As shown, the background time segment B has a variable time, e.g., time t ₁ Is shown in the variable time of day position. Thus, the state being monitored may have a fixed duration or a variable duration. When of variable duration, the state is monitored until the end of the duration. The reporting logic 82 treats this as being from one time (i.e., t) ₄ ) To the next moment (i.e., t) ₁ ) Is sensed. With time t ₁ With respect to time t ₄ This moment of each wave shape is recorded as a level to be compared with the known moment obtained by the interface 70 by selecting the welding mode of the generator 40.

Monitor M monitors these actual welding parameters during certain selected states of the wave shape; however, the monitor is also programmed to run a computer to determine the stability and/or level characteristics of the internal signal (e.g., the actual input to the motor 26 on line 46 a). This internal monitoring of the signal on line 46a is illustrated in the flow chart shown in fig. 4 using the signal shown in fig. 5.

The microcontroller in the wire feeder includes a subroutine that is a PID comparison network similar to an error amplifier. This PID comparator is shown schematically in fig. 4 as box 152, having a first input 46b at wire feed speed WFS, and a command signal on line 44. The actual WFS on line 46b is read by a tachometer or encoder for reading WFS on the drive roller 24 connected to the feeder gearbox, or alternatively on a driven wheel attached to the welding wire. The output 156 of the PID is the voltage level at the input of a pulse width modulator 158, which is digitized in the microprocessor of the feeder. The output of the pulse width modulator is a command signal on line 46a to the motor 26 for controlling the wire feed speed of the feeder 24.

According to an exemplary embodiment, the monitor M comprises a process program as schematically illustrated in fig. 4, wherein the signal on line 156 is read by the processing block 160 and the result is output on line 162 to the input of the level monitor stage 81 and/or the stability monitor stage 91, as previously discussed with respect to the embodiment shown in fig. 1. Thus, the internal signal on line 156 is read at a fast rate in excess of 1kHz in order to check the level of this internal signal and/or the stability of this signal.

As shown in fig. 5, the wave shape 100 of the pulse weld extends as a continuation of the wave shape from the generator 40. With respect to wire feed speed, the command signal on line 44 from generator 40 takes the form shown in FIG. 5. It includes a start ramp-up portion 170 and an end ramp-down portion 172. These two portions cause a sharp increase or decrease in the command signal on line 44. Between these abnormal command portions of the signal on line 44, wire feed speed level commands are typically employed for the purpose of testing the stability and/or level deviation of this internal signal on line 156. In fig. 5, the wire accelerating portion 170 is maintained until the speed is stabilized. This time is also monitored. Other internal signals may be monitored using the same concepts as shown in fig. 4 and 5. The level monitor stage determines whether the signal on line 156 exceeds the minimum or maximum value for an extended time. For wire feeders, this is typically indicative of a jam in the feeder system.

Fig. 6 shows the concept of a level monitor level, where threshold 180 is the maximum parameter level and threshold 182 is the minimum parameter level. When the parameter (shown as arc current) exceeds the threshold 180 as indicated by the transient value 184, a logged overcurrent event occurs. In a similar manner, when the current is less than the minimum level 182 as indicated by the transient 186, an undercurrent event is recorded. In addition, the events may be weighted based on a number of criteria. In one exemplary embodiment, each event is weighted with respect to the time contribution of the corresponding wave shape based on its value of the amount of deviation from a desired value (e.g., a predetermined threshold, a median value) and its time segment. Such a weighting method (e.g., weighting method 900 shown in fig. 9 and described below) may be implemented, for example, in level monitor stage 81, stability monitor stage 91, or any similar or related data processing stage.

These weighted events are periodically counted or otherwise accumulated to provide the output of the level monitor stage 81, as shown in FIG. 1. These weighted events may be accumulated, for example, every 250ms. Thus, the level monitor stage 81 detects a deviation 184 above a preset threshold and a deviation 186 below a preset level. These levels are set by specific states in the interface 70. Some states of the wave shape employ a level monitor stage 81 with thresholds and other states of the same wave shape may use a stability monitor stage 91. Preferably, and in practice, both monitor stages are used for a selected state or states of the wave shape that the monitor M is interrogating.

The embodiment shown in fig. 1 monitors the level and/or stability of the actual parameter during a selected state of the wave shape from the generator 40 or throughout the welding as explained in relation to the disclosure of fig. 4 and 5 to obtain an internal control signal. The monitor M in fig. 1 (as explained so far) provides weighted data for analyzing a welding cycle or an entire run of the welder over a period of time of operation. After the data is determined and stored, the data is processed using a different analysis program. According to an exemplary embodiment, the weighted stability data from the monitor stage 91 is analyzed by two procedures as shown in fig. 7. It is within the skill of the art to analyze the stability data for recording, display and process intervention or evaluation in a wide variety of computer programs.

As shown in fig. 7, analysis program 200 uses the results of monitor stage 91 of monitor M (i.e., the weighted stability values). As an example, at time t ₂ -t ₃ Time state in between (this is a wave-shaped current peak)Partially, as shown in fig. 2 and 3), the routine 200 is run. The analysis routine 200 is shown as a computer flow diagram showing two systems used to analyze the results of the stability stage 91 during peak current conditions, where the statistical standard deviation of the actual current in the line 34a is calculated. In practice there is a slight delay before the monitor stage 91 produces the calculated deviation. For reading the state t is shown as a sample selector or filter 90a ₂ -t ₃ During period I _a But ignore I elsewhere _a The sample selection feature of (1). Incorporated in the filter 90a at time segments t ₂ -t ₃ This programmed delay at the beginning allows the monitor to ignore current fluctuations experienced during each horizontal deviation in the various states of the output wave shape.

In the programmed flow chart shown in FIG. 7, the stability output from the monitor stage 91 is shown as a computer program read by block 210 at a pass time t ₃ At the end of each wave shape determined by the presence of (c), the block is reset as indicated by the logic on line 210 a. Thus, block 210 captures the stability of each wave shape. The captured stability data is processed according to two separate analysis programs.

The first program includes a pass analysis routine 212. If the stability of a given wave shape passes the desired threshold set in block 212, this information is output on line 214. If a particular wave shape has a stability less than a desired threshold, a logic signal is present in line 216. During each of these welding cycles, logic on line 224 enables counters 220, 222. Thus, the stability of each wave shape during the welding cycle is counted by the signal in counter 220 or counter 222. Of course, each state t is ignored ₂ -t ₃ To allow parameter I _a And (4) stabilizing. The results of these two counters are read, stored, or otherwise retained, as indicated by read blocks 220a, 222a, respectively. In one exemplary embodiment, if the instability accumulated by the counter stage 222 exceeds a desired amount, then rejection is performed as indicated by block 226Stopping the welding cycle.

A second analysis implementation of the computer program 200 shown in fig. 7 is shown as block 230. This is the procedure that is enabled during the welding cycle. The total instability of the welding cycle accumulated during all wave shapes was analyzed as a total number, with 100 being the most stable arc. The output of this stability accumulator and analysis stage is read, stored, or otherwise retained, as indicated by block 236. If the read level 234 is below the set stability, the welding cycle is rejected as indicated by block 238. One skilled in the art can devise other programs for analyzing the results from the stability stage 91 of the monitor M. The computer program 200 presents two implementations for analyzing the obtained weighted stability data. Both implementations may be selectively enabled (one or the other or both) depending on the nature of the arc stability or weld quality issues that the monitor is configured to detect. It is advantageous to read the stability in only selected states of the wave shape, since stability on variable pulses is not available.

According to another exemplary embodiment, a computer program for analyzing the results (i.e. the weighted readings) of the level monitor stage 81 of the monitor M is shown in fig. 8. In the illustrated embodiment, the horizontal analysis program 250 processes the output from the monitor level stage 81 in two separate routines, identified as a minimum monitor stage 81a with filter 80c and a maximum monitor stage 81b with filter 80 d. Any of these states may be used alone or in practice they may be combined. Subsection 81a relates to determining a transition 186 as shown in FIG. 6, which is the event that the actual parameter is below threshold minimum 182. When the program step 252 selects stage 81a, the minimum level from generator 40 on line 202a is used. For each of these welding cycles, block 254 counts these events, as shown. The counter is enabled by logic on line 254a during the welding cycle. The counter 254 runs for all wave shapes used in the welding cycle. By time t in the output of generator 40 ₃ Is out ofThe count is now made to obtain the number of wave shapes as shown by line 258. As previously noted, the first part of the state is typically ignored in order to remove normal inconsistencies at the beginning of any particular state. Block 260 is a computer flow diagram subroutine for dividing the cumulative minimum events 186 from the monitor stage 81a by the number N from the counter 256. This provides an average of the minimum transition values during the welding cycle, which is provided to subroutine 262. These average minimum transition values are read, stored, or otherwise output, as indicated at block 262 a. If the average is above some threshold amount provided by the wave generator or by program step 264, then program routine 266 determines that the welding cycle is unacceptable. If acceptable, no action is taken. However, if the acceptable routine 266 determines that the average is only near the quantity 264, then a block 266a provides an alarm signal. The total unacceptable provides a weld rejection signal via routine 266 b. Other computer programs for performing an analysis of the minimum current deviation or transition of the actual parameter may be envisaged by the person skilled in the art, since said actual parameter relates to a set threshold value.

In fig. 8, the maximum monitor stage 81b operates in conjunction with the minimum stage 81 a. The maximum level is used on line 202b, from generator 40, and when stage 81b is selected by routine 270. Similar data information and programming remains the same amount. Counter 272 versus state t ₂ -t ₃ The number of events 184 in the period is counted. Subroutine 280 provides the average value of events 184 during the different wave shapes formed during the welding cycle. This average value in block 282 is read, stored, or otherwise used as indicated by block 282 a. In block 286, the receptivity subroutine is processed, wherein the quantity output from the generator 40 or otherwise implemented by the computer program, as indicated by block 284, is compared to the average value from block 282 to provide an alarm signal, as indicated by block 286a, when the average value approaches the set quantity indicated by block 284. If the number is reached, a rejection subroutine is implemented, as shown at block 286 b.

In practice, stage 81a and stage 81b are implemented together and the average of the two transitions from blocks 262 and 282 is analyzed by the acceptable number of readings to give an alarm and/or rejection for a given welding cycle. Thus, in practice, the minimum horizontal deviation is analyzed, the maximum horizontal deviation is analyzed, and the total horizontal deviation is analyzed. All this is realized by a computer program as schematically shown in fig. 8. The level stages 81a, 81b output level conditions that are stored and/or displayed as discussed with respect to the reporting logic 82. The horizontal conditions output by the horizontal stages 81a, 81b may be weighted as discussed herein.

In view of the above, using magnitude and time contribution weights provides a more accurate measure of parameter stability, and thus overall weld quality. In this way, an easily understandable numerical value or score can be calculated to quantify the overall quality of the weld. In one exemplary embodiment, a weld score between 0-100 or 0% -100% is calculated for a weld based on monitored weld conditions or parameters, such as those monitored by the exemplary embodiment shown in fig. 1. Such a weighting method (e.g., weighting method 900 shown in fig. 9 and described below) may be implemented, for example, in the level monitor stage 81, the stability monitor stage 91, or any similar or related data processing stage.

A weighting method 900 according to an exemplary embodiment is shown in fig. 9. The weighting method may be implemented, for example, in the monitor M. In an initial step 902 of the weighting method 900, the wave shape of the welding cycle is divided into a series of time segmented portions or states. Next, in step 904, welding parameters (e.g., voltage, amperage) corresponding to at least one of the states are sampled at a given rate. In one exemplary embodiment, the sampling rate is 120kHz. In one exemplary embodiment, the sampling rate is greater than or equal to 120kHz. In one exemplary embodiment, the sampling rate may be used to generate interrupts for Interrupt Service Routine (ISR) processing.

Welding data is calculated using the sampled welding parameters. In the exemplary weighted method 900, the weld data includes an execution count, a voltage sum squared, an amperage sum summed, and an amperage sum squared. The execution count starts at zero and increments by one for each sampling period (e.g., every 120 kHz). The voltage sum and the amperage sum start with zero and increase the sampled voltage and sampled amperage, respectively, in each sampling period. Similarly, the voltage sum of squares and the amperage sum of squares start with zero and are increased by the square of the sampled voltage and the square of the sampled amperage, respectively, for each sampling period.

After a predefined sampling period, in step 906, the sampled weld data is passed on for further processing (as described below), the weld data values are reset to zero, and the sampling process is repeated (i.e., step 904). In one exemplary embodiment, the sampling period is 250ms. Each set of sampled weld data forms an analysis package. After further processing of the analysis package (e.g., every 250 ms), additional weld data representing the current weld quality level for the corresponding state is available. This additional weld data may be patterned and/or averaged. The average of these grades over the length of the weld (i.e., the weld cycle) provides an indication of the overall quality of the weld.

Further processing of the weld data for each analysis package for each sampled state in step 906 results in the calculation of additional weld data. The additional weld data includes a performance count, a voltage average, a voltage Root Mean Square (RMS), a voltage variation, an amperage average, an amperage RMS, and an amperage variation. The value of the execution count of the additional welding data is a value copied from the execution count of the welding data. The voltage average is calculated as the voltage sum (from the weld data) divided by the execution count. The voltage RMS is calculated as the square root of the quotient obtained by dividing the sum of the voltage squares (from the weld data) by the execution count. The voltage change is calculated as the voltage RMS minus the voltage average. The amperage average is calculated as the sum of the amperages (from the weld data) divided by the execution count. The amperage RMS is calculated as the square root of the quotient obtained by dividing the sum of the amperage squared (from the weld data) by the execution count. The amperage change was calculated as the amperage RMS minus the average amperage.

Following step 906, subsequent processing depends on whether the current weld is a training weld to be used to determine weld quality parameters or a normal weld to be evaluated for such weld quality parameters. Thus, in step 908, it is determined whether the current weld is a training weld or a normal weld. In one exemplary embodiment, the default condition is that the weld is a normal weld unless otherwise specified (e.g., by user input).

If the current weld is determined to be a training weld in step 908, the following additional weld data values are saved for a significant portion (e.g., 20 to 30 seconds) of the training weld: the performance count, voltage average, voltage change, amperage average, and amperage change may be performed while other weld data values and additional weld data values may be discarded. A significant portion of the training weld is the training period. In one exemplary embodiment, the training period corresponds to at least 80 consecutive analysis packets (i.e., sampling periods).

These additional weld data values saved during the training period are then used to calculate a weld quality parameter in step 910. For example, the following weld quality parameters are calculated for each of the sampled states: a quality execution count average, a quality execution count standard deviation, a quality voltage average, a quality voltage standard deviation, a quality amperage average, a quality amperage standard deviation, a quality voltage variation average, a quality voltage variation standard deviation, a quality amperage variation average, and a quality amperage variation standard deviation.

The quality execution count average is calculated as the average of the execution counts from all analysis packets processed during the training period. These execution counts may be rounded to integers. A quality execution count standard deviation is calculated as a standard deviation of the execution count from each analysis packet processed during the training period relative to the quality execution count average. The quality voltage average is calculated as the average of the voltage averages from all analysis packets processed during the training period. A quality voltage standard deviation is calculated as a standard deviation of a voltage mean from each analysis packet processed during the training period relative to the quality voltage mean. The quality amperage average is calculated as the average of the amperage averages from all analysis packets processed during the training period. A quality amperage standard deviation is calculated as the standard deviation of the amperage average from each analysis packet processed during the training period relative to the quality amperage average. The quality voltage change average is calculated as the average of the voltage changes from all analysis packets processed during the training period. A quality voltage change standard deviation is calculated as a standard deviation of a voltage change from each analysis packet processed during the training period relative to the quality voltage change. The quality amperage change average is calculated as the average of the amperage changes from all analysis packets processed during the training period. A quality amperage change standard deviation is calculated as a standard deviation of the amperage change from each analysis packet processed during the training period relative to the quality amperage change. As noted above, subsequent welds may be measured or rated using these quality parameters based on the quality of the good or acceptable weld identified by the delivery as a benchmark.

If it is determined in step 908 that the current weld is an evaluation weld other than a training weld (i.e., a weld for which quality is required to be evaluated), then no such weld data or additional weld data need be saved. Instead, results of a plurality of different quality calculations are obtained and saved. These quality calculations include initially detecting the presence of a plurality of different outliers in step 914. Outliers are the following data points or values: the data point or value is more than a threshold distance from its resulting median value. In one exemplary embodiment, an outlier is a value that falls outside of the limits with three standard deviations from the median.

In the weighting method 900, the outliers searched for at step 914 include a performance outlier, a voltage change outlier, an amperage outlier, and an amperage change outlier. For each of the monitored states, each of the analysis packages is evaluated to detect the presence of any of the outliers.

An execution outlier is considered if the analysis package satisfies the following relationship: absolute value of (execution count-quality execution count average >) (3 x quality execution count standard deviation). A voltage outlier is considered if the analysis package satisfies the following relationship: absolute value of (voltage average-quality voltage average >) (3 x quality voltage standard deviation). A voltage change outlier is considered if the analysis package satisfies the following relationship: absolute value of (mean of voltage change-quality voltage change) > (3 x standard deviation of quality voltage change). An amperage outlier is considered if the analysis package satisfies the following relationship: absolute value of (amp mean-quality amp mean) > (3 x quality amp standard deviation). An amperage change anomaly is considered if the analysis package satisfies the following relationship: absolute value of (amperage change-quality amperage change average >) (3 x quality amperage change standard deviation).

After these outliers are detected, the quality index for the corresponding analysis packet is calculated using the two-step weighted sum of each outlier (i.e., from steps 916 and 918).

The first step of weighting each of these outliers (i.e., step 916) is determined by the magnitude of the outlier relative to the triple standard deviation limit. Generally, approximately 0.3% of these data points or values may fall outside the three standard deviation limits and are therefore considered outliers. The weighting of the outliers increases as their value increases above the three times standard deviation limit. The outliers are weighted at full 100% at four times the standard deviation value and at maximum 200% at five times the standard deviation. In general, the probability of a fully (i.e., 100%) weighted outlier in the normal dataset is 1 for 15,787.

Thus, in step 916, each of these outliers is weighted in this way. The weight applied to each execution outlier is calculated as the absolute value of (the quantity/quality execution count standard deviation three times the standard deviation limit or more), where the maximum weight value is 2.0. The weight applied to each voltage outlier is calculated as the absolute value of (magnitude/quality voltage standard deviation three times greater than the standard deviation limit) with the maximum weight value of 2.0. The weight applied to each voltage change outlier is calculated as the absolute value of (magnitude/quality voltage change standard deviation three times greater than the standard deviation limit) with the maximum weight value being 2.0. The weight applied to each amperage anomaly is calculated as the absolute value of (amount/quality amperage standard deviation three times above the standard deviation limit) with the maximum weight value of 2.0. The weight applied to each amperage change anomaly is calculated as the absolute value of (amount/quality amperage change standard deviation above three standard deviation limits) with a maximum weight value of 2.0.

The second step of weighting each of these outliers (i.e., step 918) is determined by the execution count of the states of the outliers. In particular, the value of each outlier is multiplied by the execution count of the state of said outlier, thereby taking into account the time contribution of said state with respect to the total wave shape. In this way, states with larger execution counts (i.e., execution times) produce outliers with relatively heavier weights. Therefore, as the number of executions of a particular outlier increases, the weight of that outlier will also increase.

The weighting of these outliers in steps 916 and 918 produces a set of final weighted outliers including a final weighted execution outlier, a final weighted voltage change outlier, a final weighted amperage outlier, and a final weighted amperage change outlier. These final weighted outliers are summed in step 920 to produce a final weighted outlier sum for each analysis packet. The quality index for each of these analysis packages is then calculated in step 922 as the quotient obtained by dividing the final weighted sum of outliers subtracted from the perfect quality value by the perfect quality value. The perfect quality value is equal to the execution count of the analysis packet multiplied by the number of outlier classes (i.e., five in this case).

Thus, an instantaneous quality index (i.e., a weld score or weld quality score from a currently completed analysis package) may be determined during the welding process and may be communicated to the welder or otherwise utilized. In this way, potential problems may be detected when they occur, i.e. during the welding process, as opposed to only after the welding is completed, when it is likely too late to take any corrective action.

Furthermore, the average of the quality indicators accumulated up to any point in time during the welding process may be averaged in order to determine a quality indicator (e.g., a weld score or a weld quality score) of the weld up to that point in time. For example, after the welding process is completed, all of the individual quality indicators may be averaged to obtain a total quality indicator, score, grade, rating, etc. of the completed weld. The overall quality index for the weld may be compared to a predetermined quality index (e.g., obtained from a training weld) that reflects a minimum quality index value for an acceptable weld.

In this manner, the quality of the weld may be determined accurately, efficiently, consistently, and/or automatically, in real-time or near real-time. This is particularly advantageous because visual inspection of the weld is not always sufficient to gauge its quality, and because the operator may not detect or know deviations or other problems that may affect the overall weld quality during the welding process.

In some exemplary embodiments, a quality index of a weld (i.e., a weld score, or a weld quality score) is an efficient tool to evaluate multiple welds that are repeatedly produced under substantially the same conditions and according to substantially the same arc welding process, such as in an automated (e.g., robotic) welding process. An automated quality control process may be adapted for an arc welding process by calculating an instantaneous, periodic, and/or total weld score for each weld. Specifically, a minimum acceptable welding score or range of acceptable welding scores is initially identified as a threshold value as a function of the welding conditions and the arc welding process. Thereafter, each weld compares its (instantaneous, periodic and/or total) weld score to the threshold in order to quickly and accurately determine whether the weld should be accepted or rejected. In addition, by evaluating trends in the weld score for a production pass or set of passes, problems in the production process may be more easily identified and/or the production process may be more easily optimized. The quality index or weld score discussed above is based on a weighted statistical measure reflecting quality. These weld scores may be evaluated over time to determine whether there is any trend moving away from an acceptable weld score (e.g., as evidenced by a continuing decrease in the weld score).

During a welding process, such as a robotic welding process, a welding score is periodically calculated (based on one or more sampled or otherwise measured parameters) to reflect the current state of the weld. The weld score may be calculated as an instantaneous measurement reflecting the current state of the weld, or as an average of several measurements reflecting the state of the weld over a certain period of time during the welding process (corresponding to these measurements). In one exemplary embodiment, a weld score is calculated by averaging all measurements taken from the beginning of the welding process, which reflects the current overall state of the weld. The weld score may be compared to a predetermined threshold weld score. The threshold weld score is the minimum weld score for a good or acceptable weld condition. Determining the current status of the weld as good if the weld score is greater than or equal to the threshold weld score. Otherwise, the current state of the weld is determined to be poor.

Fig. 10 illustrates a schematic diagram of a robotic welding system 300. Robotic welding system 300 includes a welding robot 302 (e.g., a 6-axis robotic arm) for manipulating a welding torch 304 to perform arc welding on a workpiece 306. The arc welding may be Gas Metal Arc Welding (GMAW), flux Cored Arc Welding (FCAW), gas Tungsten Arc Welding (GTAW), submerged Arc Welding (SAW), or the like. The welding robot 302 receives positioning commands from the robot controller 308 to control the movement of the robot during welding. In certain embodiments, the robot controller 308 may include through arc weld tracking logic 309 to correct the weld path. The through arc weld tracking is discussed in further detail below. The robot controller 308 communicates with the welding power supply 310 to exchange information during welding. Such exchanged information may include welding parameters such as welding voltage and current, welding status information, and torch position information. The welding power supply 310 includes a processor-based controller 80 for controlling welding parameters such as welding waveform, welding voltage, welding current, wire feed speed, and the like. The robotic welding system may include a wire feeder 312 for feeding a consumable welding electrode to the welding torch 304, and may include a shielding gas supply 314 that delivers shielding gas to the welding torch 304. The robotic welding system 300 is but one example of an automated arc welding system. Other example types of automated arc welding systems include orbital welders for welding pipes, tanks, and the like, linear orbital welders that move linearly along a weld joint, and column and boom carriage systems. Each automated arc welding system comprises: suitable torch manipulators, such as robotic arm 302, torch carriage, and the like; and a corresponding motion controller operatively connected to the torch manipulator for moving the welding torch along the torch path as needed.

Fig. 11 illustrates an example wiggle pattern that may be performed by a robotic welding system while welding a first workpiece 306a to a second workpiece 306 b. As the robot moves the welding torch 304 in the direction 316 of the weld during welding, the robot swings the welding torch side-to-side across the weld joint. The right side R, left side L, and center C swing positions are indicated in fig. 11. The robot may oscillate the welding torch 304 back and forth across the weld at a generally constant frequency, which may depend on variables such as the travel speed of the welding torch 304 in the direction 316 of the weld and the wire feed speed of the welding electrode. An example wobble frequency is in the range of 1 to 5Hz, but other wobble frequencies are possible.

When calculating a quality index (e.g., a weld score or weld quality score) for a weaving welding process, the accuracy of the weld score may be improved by calculating at consistent locations along the weaving pattern (e.g., L, R and/or C). This may allow the analysis package used to calculate the weld score to contain similar data (e.g., data over one period of oscillation). The robot controller may provide the welding power supply with torch position information and/or information regarding weaving speed (e.g., a weaving frequency or period) such that the welding power supply may make a welding score calculation at a consistent location along the weaving pattern. For example, the robot controller informs the welding power supply of the position of the torch within the welding pattern, and the welding power supply may calculate a welding score based on the position of the torch. Each time the torch reaches the left position L (or some other predetermined position) of the wobble pattern, the welding power supply may process the analysis package as described above to calculate a welding score for one wobble period. In this scenario, instead of a sampling period having a predetermined or fixed duration for collecting the analysis package, the sampling period is determined by a wobble period or frequency based on torch position information provided to the welding power supply by the robot controller. Thus, the welding power supply may use the torch position information as synchronization information for calculating the welding score. The robot controller may also inform the welding power source of the swing frequency or swing period, and the welding power source may adjust the sampling period of the analysis package based on the swing frequency/period. For example, the welding power source may set the sampling period of the analysis package equal to the period of the wobble. For a 2Hz swing frequency, the swing period is 500ms, and the sampling period of the welding score analysis package may be set or adjusted by the welding power source to 500ms (thereby allowing the welding power source to calculate the welding score at a consistent location along the swing pattern). In the case where the sampling period of the welding score analysis package is set to the weaving period, the welding power supply may also automatically adjust the sampling period, if necessary, with reference to torch position information provided by the robot controller to obtain a welding score at a consistent weaving position. If the robot controller is performing through arc weld tracking (TAST) to automatically follow the change in direction of the weld, the welding power supply may use the TAST information to enhance the welding score. For example, deviations in the weld may be recorded along with the weld score, or used in a weighting process when calculating the weld score. The TAST information recorded with the welding score may correspond to the welding score.

Fig. 12 shows TAST during weaving welding. The TAST uses the welding current and voltage feedback of the robot and the swing function to determine the lateral position of the torch in the weld joint. At the center of the joint, the arc current is at a minimum and the arc length or voltage is at its maximum. When the torch reaches the edges of its swing cycle (the L and R positions), the arc current peaks and the voltage drops. If the value of the peak current increases at the edge of the weaving cycle, the torch is moving away from the weld joint and TAST makes the necessary corrections to the robot welding path to follow the weld. The TAST may also determine the vertical position of the torch. As the contact tip-to-workpiece distance increases, the current becomes smaller, and as the contact tip-to-workpiece distance decreases, the amount of current increases. The TAST can correct for variations in the vertical position of the torch to maintain a constant dry elongation. The TAST calculation can utilize both the welding voltage and current to determine the edge of the weld (e.g., TAST can calculate the impedance or change in impedance from the voltage and current measurements). Fig. 12 shows the left L, center C and right R swing positions for performing the TAST calculation. In addition to calculating the motion correction vectors for weld tracking and overhang control, a TAST calculation may be performed to adjust the weaving or oscillation width, oscillation speed, travel speed, and/or sidewall dwell time of the welding torch based on the change in size of the weld groove. For example, if the weld groove widens, the TAST calculation can increase the oscillation width of the torch and slow its travel speed along the weld joint to maintain the deposit of weld metal as needed.

Fig. 13 illustrates an example pulsed output current welding waveform that may be used during weaving welding and when performing TAST calculations. The waveforms in fig. 13 are merely exemplary and are provided for ease of explanation. It should be appreciated that many different welding waveforms (e.g., surface tension transition STT) may be suitable for use with the present disclosure. Waveform 700 in fig. 13 includes a peak pulse 710. After the peak pulse 710 is issued, a short may occur, beginning at time 720, for example, and continuing until time 730, for example, at which time the short is cleared. Times 720 and 730 define a shorting interval 740. The peak pulses 710 are emitted at regular intervals during multiple pulse periods or cycles of the welding process. Waveform 700 also includes a plasma boost pulse 750 to help prevent another short from occurring immediately after the just cleared short. An example frequency range for waveform 700 is 100 to 200Hz, but other frequencies are possible.

Fig. 14 illustrates another example pulsed output current welding waveform that may be used during weaving welding and when performing TAST calculations. Waveform 760 in fig. 14 is an STT weld waveform. Background current 762 (e.g., between 50 and 100 amps) maintains an arc and facilitates heating of the matrix metal. After the welding electrode is initially shorted to the weld pool, the current is rapidly reduced to a minimum level 764 to ensure a solid state short. A ramped-up pinch current is then applied to squeeze molten metal on the electrode tip down into the weld pool while monitoring the neck-like contraction of the liquid bridging portion of the electrode. When the liquid bridge is about to break, the welding power supply reacts by reducing the welding current to a minimum level. Immediately after arc re-establishment, a peak current 766 or plasma boost pulse is applied to generate a plasma force that pushes the weld pool downward to prevent accidental shorting between the electrode and the weld pool and to heat the weld puddle and joint. The exponential tail-out segment 768 is then adjusted to adjust the total heat input. The tail-out segment 768 returns the welding current to the background current 762 level. An example frequency range for the STT waveform is 100 to 200Hz, but other frequencies are possible.

In conventional TAST, a robot controller, which may include a dedicated TAST controller (e.g., a TAST control board or TAST logic), monitors the welding current and voltage associated with the position of the torch in the weld joint. The robot controller knows when the torch is near the edge or center of the weaving cycle and captures data based on the position of the torch and performs a TAST calculation. If a dedicated TAST controller is used, the main robot control processor may communicate torch position information to the TAST controller. TAST data capture and calculation can be done very close to the edge or center of the swing, for example within 1/100 inch to 1/10,000 inch of the edge or center of the swing.

According to the pulse welding waveforms discussed above, the current and voltage are constantly changing between peak values, low current levels, background levels, and the like. When making a TAST current/voltage measurement and performing a TAST calculation, successive TAST calculations may be affected by which portion of the welding waveform is dominant. For example, one TAST calculation may be based primarily on current pulses, while another is based primarily on background current portions. This may lead to inaccuracies in the TAST calculation, which may result in over-or under-correction of the weld tracking adjustment. To address this issue, in one embodiment, the welding power supply may make welding voltage and current measurements during welding and provide filtered or labeled welding voltage and current data to the robot controller or the TAST controller. The data is filtered or marked according to the portion of the welding waveform to which the data corresponds (e.g., peak or background). The welding power supply may receive torch position information from the robot controller and transmit filtered or marked welding voltage and current data as the torch approaches the edge or center of the swing. Examples of filtered data are voltage and current data obtained from measurements made only during the peak pulse or only during the background portion of the waveform (or some other portion of the welding waveform, as desired). The test calculation can then be performed using the filtered data such that consistent portions of the welding waveform are used in successive test calculations. For example, only data corresponding to the peak pulse or background portion of the welding waveform is used to calculate the correction vector. The robot controller or the TAST controller may also be provided with voltage and current data for a plurality of different portions of the welding waveform (if the data is marked or segmented) so that data from one portion of the welding waveform can be distinguished from data corresponding to another portion of the welding waveform (e.g., segmented into differentiable data blocks). The marked data need not be segmented into differentiable data blocks, as the marked data will include the appropriate identifier recognized by the TAST controller. Then, a TAST calculation may be performed using the selected portion of the welding waveform. The filtered or labeled data is classified or identified according to the portion of the welding waveform to which they correspond, such as current measurements made during the pulse current portion of the welding waveform, and current measurements made during the low current portion (e.g., background portion) of the waveform. The TAST calculation may be performed using voltage and current data associated with one or both of a pulsed portion of the welding waveform and a low current portion (e.g., background portion) of the welding waveform. In certain embodiments, the TAST calculation is performed using voltage and current data corresponding to two or more different portions of the welding waveform (e.g., a pulse portion and a background portion), and the TAST logic weights the data differently depending on which portion of the welding waveform the data corresponds to. For example, the pulse portion of the welding waveform may be more consistent than the background portion and provide more useful information in making the TAST calculation. Therefore, the voltage and current data corresponding to the pulse portion may be weighted more (given more weight) in the TAST calculation than the data corresponding to the background portion. The data corresponding to the impulse portion may be given a smaller weight than the data corresponding to the background portion, or the data may be given the same weight. In other embodiments, the TAST calculation is performed using data corresponding to a single portion (e.g., a pulse portion) of the welding waveform.

In certain embodiments, the welding power supply may perform a TAST calculation and provide path correction instructions to the robot controller. In such an embodiment, the welding power supply would include TAST logic.

In another example embodiment, the welding power supply receives torch position information from the robot controller. When the torch is near the swing edge, the welding power supply adjusts the welding waveform to a predefined level for the TAST measurement. For example, a background current level or pulse may be applied when the torch is near the weaving edge. The robot controller or the TAST controller may then make a TAST measurement based on the consistent portion of the welding waveform.

Fig. 15 illustrates an embodiment of an example controller 80 of a welding power supply. Controller 80 includes at least one processor 814 that communicates with a number of peripheral devices via bus subsystem 812. These peripheral devices may include a storage subsystem 824, which includes, for example, a memory subsystem 828 and a file storage subsystem 826, user interface input devices 822, user interface output devices 820, and a network interface subsystem 816. The input and output devices allow a user to interact with the controller 80. Network interface subsystem 816 provides an interface to an extranet and couples to corresponding interface devices in other computer systems.

The user interface input devices 822 may include a keyboard, a pointing device (such as a mouse, trackball, touchpad, or tablet), a scanner, a touch screen incorporated into the display, audio input devices (such as voice recognition systems, microphones), and/or other types of input devices. In general, use of the term "input device" is intended to include all possible types of devices and ways to input information into the controller 80 or over a communication network.

User interface output devices 820 may include a display subsystem, a printer, a facsimile machine, or a non-visual display such as an audio output device. The display subsystem may include a Cathode Ray Tube (CRT), a flat panel device such as a Liquid Crystal Display (LCD), a projection device, or some other mechanism for creating a visible image. The display subsystem may also provide non-visual displays, such as via audio output devices. In general, use of the term "output device" is intended to include all possible types of devices and ways to output information from controller 80 to a user or to another machine or computer system.

Storage subsystem 824 stores programming and data constructs that provide the functionality of some or all of the modules described herein. These software modules are generally executed by processor 814, either alone or in combination with other processors. The memory 828 used in the storage subsystem may include a plurality of memories including: a main Random Access Memory (RAM) 830 for storing instructions and data during program execution; and a Read Only Memory (ROM) 832 in which the fixed instructions are stored. The file storage subsystem 826 may provide persistent storage for program and data files, and may include a hard disk drive, a floppy disk drive along with associated removable media, a CD-ROM drive, an optical drive, or removable media cartridges. Modules that implement the functionality of certain embodiments may be stored in storage subsystem 824, or in other machines accessible to processor(s) 814, by file storage subsystem 826.

Bus subsystem 812 provides a mechanism for multiple different components and subsystems of controller 80 to communicate with one another as intended. Although bus subsystem 812 is shown schematically as a single bus, alternative embodiments of the bus subsystem may use multiple buses.

The controller 80 may be of a variety of different types including a workstation, a server, a computing cluster, a blade server, a server farm, or any other data processing system or computing device. Due to the ever-changing nature of computing devices and networks, the description of controller 80 depicted in FIG. 15 is intended only as a specific example for purposes of illustrating some embodiments. Many other configurations of the controller 80 (with more or fewer components than the controller depicted in fig. 15) are possible.

It should be clear that the present disclosure is by way of example and that various changes may be made by adding, modifying or removing details without departing from the fair scope of the teaching contained in the present disclosure. Therefore, the invention is not to be limited to the specific details of the disclosure unless the appended claims are necessarily so limited.

Claims (20)

1. An automated arc welding system, comprising:

a torch manipulator;

a welding torch extending from the torch manipulator;

a motion controller operatively connected to the torch manipulator to control the torch to move along a welding path in a direction of a weld during a welding operation and simultaneously control the torch to oscillate back and forth across the weld at an oscillation frequency and an oscillation period; and

a welding power supply operatively connected to the welding torch to control a welding waveform and operatively connected to the motion controller to communicate therewith, wherein the welding power supply is configured to sample a plurality of welding parameters and form an analysis package and process the analysis package to generate a weld quality score during a sampling period of the welding operation,

wherein the welding power source obtains the wobble frequency or the wobble period and automatically adjusts the sampling period for forming the analysis package based on the wobble frequency or the wobble period.

2. The automated arc welding system of claim 1, wherein the sampling period used to form the analysis package is equal to the wobble period.

3. The automated arc welding system of claim 1, wherein the welding power source receives the wobble frequency from the motion controller.

4. The automated arc welding system of claim 1, wherein the welding power supply receives the weaving cycle from the motion controller.

5. The automated arc welding system of claim 1, wherein the welding power supply receives torch position information from the motion controller during the welding operation.

6. The automated arc welding system of claim 5, wherein the welding power supply automatically adjusts the sampling period used to form the analysis package based on the torch position information.

7. The automated arc welding system of claim 1, wherein the welding power source records through arc weld tracking information corresponding to the weld quality score.

8. The automated arc welding system of claim 1, further comprising a cross arc weld tracking logic that tracks the weld and calculates a correction to the welding path from welding current data classified as corresponding to at least one of a pulsed current portion and a low current portion of the welding waveform.

9. An automated arc welding system, comprising:

a torch manipulator;

a welding torch extending from the torch manipulator;

a motion controller operatively connected to the torch manipulator to control the torch to move along a welding path in a direction of a weld during a welding operation and simultaneously control the torch to oscillate back and forth across the weld at an oscillation frequency and an oscillation period; and

a welding power supply operatively connected to the welding torch to control a welding waveform and operatively connected to the motion controller to communicate therewith, wherein the welding power supply is configured to sample a plurality of welding parameters and form an analysis package and process the analysis package to generate a weld quality score during a sampling period of the welding operation,

wherein the welding power supply receives torch position information and automatically adjusts the sampling period for forming the analysis package based on the torch position information.

10. The automated arc welding system of claim 9, wherein the sampling period used to form the analysis package is equal to the wobble period.

11. The automated arc welding system of claim 9, wherein the welding power source receives the wobble frequency from the motion controller.

12. The automated arc welding system of claim 9, wherein the welding power source receives the weaving cycle from the motion controller.

13. The automated arc welding system of claim 9, wherein the welding power source records through-arc seam tracking information corresponding to the weld quality score.

14. The automated arc welding system of claim 9, further comprising arc-through weld tracking logic that tracks the weld and calculates a correction to the welding path from welding current data classified as corresponding to at least one of a pulsed current portion and a low current portion of the welding waveform.

15. An automated arc welding system, comprising:

a torch manipulator;

a welding torch extending from the torch manipulator;

a motion controller operatively connected to the welding torch manipulator to control movement of the welding torch along a welding path in a direction of a weld during a welding operation and simultaneously control a back and forth oscillatory movement of the welding torch across the weld; and

a welding power supply operatively connected to the welding torch to control a welding waveform and operatively connected to the motion controller to communicate therewith, wherein the welding waveform includes a pulsed current portion and a low current portion having a lower current level than the pulsed current portion,

wherein the automated arc welding system includes through-arc weld tracking logic that tracks the weld and calculates a correction to the weld path,

wherein the through arc weld tracking logic calculates a correction to the welding path based on welding current data provided by the welding power supply classified as corresponding to a pulsed current portion of the welding waveform and welding current data classified as corresponding to a low current portion of the welding waveform, and

wherein the through arc weld tracking logic weights the welding current data classified as corresponding to a pulsed current portion of the welding waveform as being different from the welding current data classified as corresponding to a low current portion of the welding waveform when calculating the correction to the welding path.

16. The automated arc welding system of claim 15, wherein the through arc weld tracking logic weights more heavily the welding current data classified as corresponding to a pulsed current portion of the welding waveform than the welding current data classified as corresponding to a low current portion of the welding waveform when calculating the correction to the welding path.

17. The automated arc welding system of claim 15, wherein the motion controller comprises the through-arc weld tracking logic.

18. The automated arc welding system of claim 15, wherein the welding power supply comprises the through arc weld tracking logic.

19. The automated arc welding system of claim 15, wherein the welding power source is configured to sample a plurality of welding parameters and form an analysis package during a sampling period of the welding operation and process the analysis package to generate a weld quality score, wherein the welding power source obtains a swing frequency or swing period of a swing movement and automatically adjusts the sampling period used to form the analysis package based on the swing frequency or swing period.

20. The automated arc welding system of claim 18, wherein the welding power source records through arc weld tracking information corresponding to the weld quality score.

Images