Welding arc tracking method based on neural network algorithm

Abstract

The invention discloses a welding arc tracking method based on a neural network algorithm, and the method comprises the following steps: 1, collecting an electric signal of a welding arc through a welding power supply, and transmitting the collected electric signal to a signal processing unit for further processing; and 2, a signal processing unit carries out filtering processing on the arc voltage and arc current signals to remove noise and interference, carries out amplification and analog-to-digital conversion on the filtered signals, and converts the filtered signals into digital signals for subsequent processing. The electric arc electric signals are processed and analyzed through the neural network algorithm, the deviation value of the welding seam can be accurately calculated in real time, the motion trail of the welding robot is adjusted through a feedback mechanism, accurate tracking of the welding seam is achieved, and due to the fact that the neural network has the high learning ability and adaptability, the welding seam tracking precision is improved. The method can adapt to different welding environments and workpiece requirements, and the welding quality and the production efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of welding arc tracking, and more specifically to a welding arc tracking method based on a neural network algorithm. Background Art

[0002] During the welding process, due to the influence of factors such as workpiece processing error, assembly error and welding thermal deformation, the actual position of the weld often deviates from the ideal position. In order to ensure the welding quality, the weld needs to be tracked in real time.

[0003] Traditional welding robots usually use mechanical contact sensors or visual sensors for weld tracking, but these methods have some limitations. Mechanical contact sensors are prone to wear and have high requirements for the workpiece surface; visual sensors are greatly affected by factors such as ambient light and smoke, and are expensive. Therefore, a new technical solution is needed to solve this problem. Summary of the invention

[0004] The purpose of the present invention is to provide a welding arc tracking method based on a neural network algorithm, which solves the problem that traditional welding robots usually use mechanical contact sensors or visual sensors to track welds, but these methods have some limitations. Mechanical contact sensors are easy to wear and have high requirements on the surface of the workpiece; visual sensors are greatly affected by factors such as ambient light and smoke, and have high costs.

[0005] To achieve the above object, the present invention provides the following technical solution: a welding arc tracking method based on a neural network algorithm, the welding arc tracking method based on a neural network algorithm comprises the following steps:

[0006] Step 1: Collect the electrical signal of the welding arc through the welding power supply, and transmit the collected electrical signal to the signal processing unit for further processing;

[0007] Step 2: The signal processing unit filters the arc voltage and arc current signals to remove noise and interference, amplifies and performs analog-to-digital conversion on the filtered signals, and converts them into digital signals for subsequent processing;

[0008] Step 3: Extract characteristic parameters related to the weld position, which are voltage and current signals;

[0009] Step 4: Build a neural network model, which is used to process the characteristic parameters extracted from the arc electrical signal;

[0010] Step 5: Input the arc electrical signal characteristic parameters collected in real time into the trained neural network model;

[0011] Step 6: The neural network model calculates the deviation of the weld according to the input characteristic parameters;

[0012] Step 7: Feedback the calculated weld deviation to the control system of the welding robot;

[0013] Step 8: The control system uses appropriate control algorithms to adjust the motion trajectory of the welding robot according to the deviation, realizes real-time tracking of the weld, and ensures welding quality and production efficiency.

[0014] As a preferred embodiment of the present invention, the electrical signals collected in step 1 include arc voltage, arc current, and arc height signals.

[0015] As a preferred embodiment of the present invention, the characteristic parameters related to the weld position extracted in step 3 include weld edge positioning, and the weld edge positioning extracts the gradient strength and gradient direction;

[0016] The gradient strength is calculated using the following formula:

[0017]

[0018] The gradient direction uses the following formula:

[0019]

[0020] G x and G y are the gradients of the image in the x and y directions respectively;

[0021] By extracting the gradient strength and gradient direction of the weld edge, the weld position can be located more accurately, misjudgment and missed judgment can be reduced, and the accuracy and reliability of weld tracking can be improved.

[0022] As a preferred embodiment of the present invention, extracting characteristic parameters related to the weld position in step 3 includes calculating the weld centerline;

[0023] Calculation formula based on the center line of the two weld edges:

[0024]

[0025] ρ1, θ1, ρ2, θ2 are the equation parameters of the two selected weld edges;

[0026] Calculating the center line based on the weld edge can more intuitively reflect the direction and position of the weld, provide a more accurate reference for adjusting the motion trajectory of the welding robot, and ensure welding quality and production efficiency.

[0027] As a preferred embodiment of the present invention, the standard mathematical expression of the neural network model in step 5 is:

[0028] H=Kst·I+C

[0029] H is the torch height or arc length, which is the variable to be predicted or controlled.

[0030] K st It is the proportionality factor between torch height and welding current.

[0031] I is the welding current, which is a variable measured in real time and can be used as the input feature of the neural network.

[0032] C is a constant term representing the theoretical value or offset at maximum weld height.

[0033] As a preferred embodiment of the present invention, the mathematical expression of the control algorithm in step 8 is:

[0034]

[0035] u(t) is the control input; e(t) is the error; K p , K i , K d They are proportional, integral and derivative gains respectively.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] The present invention uses a neural network algorithm to process and analyze arc electrical signals, can accurately calculate the deviation of the weld in real time, and adjust the motion trajectory of the welding robot through a feedback mechanism to achieve accurate tracking of the weld. At the same time, since the neural network has strong learning ability and adaptability, the method can adapt to different welding environments and workpiece requirements, and improve welding quality and production efficiency. DETAILED DESCRIPTION

[0038] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0039] The present invention provides a technical solution: a welding arc tracking method based on a neural network algorithm, the welding arc tracking method based on a neural network algorithm comprises the following steps:

[0040] Step 1: Collect the electrical signal of the welding arc through the welding power supply, and transmit the collected electrical signal to the signal processing unit for further processing;

[0041] Step 2: The signal processing unit filters the arc voltage and arc current signals to remove noise and interference, amplifies and performs analog-to-digital conversion on the filtered signals, and converts them into digital signals for subsequent processing;

[0042] Step 3: Extract characteristic parameters related to the weld position, which are voltage and current signals;

[0043] Step 4: Build a neural network model, which is used to process the characteristic parameters extracted from the arc electrical signal;

[0044] Step 5: Input the arc electrical signal characteristic parameters collected in real time into the trained neural network model;

[0045] Step 6: The neural network model calculates the deviation of the weld according to the input characteristic parameters;

[0046] Step 7: Feedback the calculated weld deviation to the control system of the welding robot;

[0047] Step 8: The control system uses appropriate control algorithms to adjust the motion trajectory of the welding robot according to the deviation, realizes real-time tracking of the weld, and ensures welding quality and production efficiency.

[0048] Further improved, the electrical signals collected in step 1 include arc voltage, arc current, and arc height signal. Adding the collection of arc height signal can more comprehensively reflect the state of the welding arc and provide a richer data basis for subsequent signal processing and feature extraction, thereby improving the accuracy and stability of weld tracking.

[0049] Further improved, the characteristic parameters related to the weld position extracted in step 3 include weld edge positioning, and the weld edge positioning extracts the gradient strength and gradient direction;

[0050] The gradient strength is calculated using the following formula:

[0051]

[0052] The gradient direction uses the following formula:

[0053]

[0054] G x and G y are the gradients of the image in the x and y directions respectively;

[0055] By extracting the gradient strength and gradient direction of the weld edge, the weld position can be located more accurately, misjudgment and missed judgment can be reduced, and the accuracy and reliability of weld tracking can be improved.

[0056] Further improved, the extraction of characteristic parameters related to the weld position in step 3 includes calculation of the weld centerline;

[0057] Calculation formula based on the center line of the two weld edges:

[0058]

[0059] ρ1, θ1, ρ2, θ2 are the equation parameters of the two selected weld edges;

[0060] Calculating the center line based on the weld edge can more intuitively reflect the direction and position of the weld, provide a more accurate reference for adjusting the motion trajectory of the welding robot, and ensure welding quality and production efficiency.

[0061] Further improved, the standard mathematical expression of the neural network model in step 5 is:

[0062] H=Kst·I+C

[0063] H is the torch height or arc length, which is the variable to be predicted or controlled.

[0064] K st It is the proportionality factor between torch height and welding current.

[0065] I is the welding current, which is a variable measured in real time and can be used as the input feature of the neural network.

[0066] C is a constant term, which represents the theoretical value or offset at the maximum welding height;

[0067] Clarifying the standard mathematical expression of the neural network model helps to understand the input, output and internal relationship of the model, provides theoretical support for the training and optimization of the model, and ensures that the model can accurately calculate the weld deviation.

[0068] Further improved, the mathematical expression of the control algorithm in step 8 is:

[0069]

[0070] u(t) is the control input; e(t) is the error; K p , K i , K d They are proportional, integral and derivative gains respectively;

[0071] Clarifying the mathematical expression of the control algorithm helps to understand the input, output and regulation mechanism of the control system, provides theoretical support for the design and optimization of the control system, and ensures that the welding robot can be accurately adjusted according to the weld deviation.

[0072] Working principle: First, the electrical signals of the welding arc including arc voltage, arc current and arc height signals are collected through the welding power supply, and these signals are transmitted to the signal processing unit for further processing. The signal processing unit filters the arc voltage and arc current signals to remove noise and interference, and then amplifies and converts the filtered signals into digital signals for subsequent processing. Next, the characteristic parameters related to the weld position, such as weld edge positioning and weld centerline calculation, are extracted. These parameters are calculated based on the voltage and current signals to construct a neural network model, which is used to process the characteristic parameters extracted from the arc electrical signal and calculate the deviation of the weld. The characteristic parameters of the arc electrical signal collected in real time are input into the trained neural network model to obtain the calculation result of the weld deviation, and the calculated weld deviation is fed back to the control system of the welding robot. The control system uses an appropriate control algorithm to adjust the motion trajectory of the welding robot according to the deviation, so as to realize real-time tracking of the weld and ensure welding quality and production efficiency.

[0073] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes that fall within the meaning and scope of the equivalent elements of the claims are included in the present invention.

[0074] Finally, a few points should be explained: First, in the description of this application, it should be noted that, unless otherwise specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, and can be mechanical or electrical connections, or internal connectivity between two components, or direct connections. "Up", "down", "left", "right", etc. are only used to indicate relative position relationships. When the absolute position of the object being described changes, the relative position relationship may change.

[0075] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A welding arc tracking method based on a neural network algorithm, characterized in that: The welding arc tracking method based on the neural network algorithm includes the following steps: Step 1: Collect the electrical signal of the welding arc through the welding power supply, and transmit the collected electrical signal to the signal processing unit for further processing; Step 2: The signal processing unit filters the arc voltage and arc current signals to remove noise and interference, amplifies and performs analog-to-digital conversion on the filtered signals, and converts them into digital signals for subsequent processing; Step 3: Extract characteristic parameters related to the weld position, which are voltage and current signals; Step 4: Build a neural network model, which is used to process the characteristic parameters extracted from the arc electrical signal; Step 5: Input the arc electrical signal characteristic parameters collected in real time into the trained neural network model; Step 6: The neural network model calculates the deviation of the weld according to the input characteristic parameters; Step 7: Feedback the calculated weld deviation to the control system of the welding robot; Step 8: The control system uses appropriate control algorithms to adjust the motion trajectory of the welding robot according to the deviation, realizes real-time tracking of the weld, and ensures welding quality and production efficiency.

2. A welding arc tracking method based on a neural network algorithm according to claim 1, characterized in that: The electrical signals collected in step 1 include arc voltage, arc current, and arc height signals.

3. A welding arc tracking method based on a neural network algorithm according to claim 1, characterized in that: The characteristic parameters related to the weld position extracted in step 3 include weld edge positioning, and the weld edge positioning extracts the gradient strength and gradient direction; The gradient strength is calculated using the following formula: The gradient direction uses the following formula: G x and G y are the gradients of the image in the x and y directions, respectively.

4. A welding arc tracking method based on a neural network algorithm according to claim 1, characterized in that: Extracting characteristic parameters related to the weld position in step 3 includes calculating the weld centerline; Calculation formula based on the center line of the two weld edges: ρ1, θ1, ρ2, θ2 are the equation parameters of the two selected weld edges.

5. The welding arc tracking method based on a neural network algorithm according to claim 1, characterized in that: The standard mathematical expression of the neural network model in step 5 is: H=Kst·I+C H is the torch height or arc length, which is the variable to be predicted or controlled. K st It is the proportionality factor between torch height and welding current. I is the welding current, which is a variable measured in real time and can be used as the input feature of the neural network. C is a constant term representing the theoretical value or offset at maximum weld height.

6. A welding arc tracking method based on a neural network algorithm according to claim 1, characterized in that: The mathematical expression of the control algorithm in step 8 is: u(t) is the control input; e(t) is the error; K p , K i , K d They are proportional, integral and derivative gains respectively.

Images