A method and system for power control of handheld laser welding heads based on molten pool temperature

Abstract

This invention belongs to the field of power control technology, specifically relating to a power control method and system for handheld laser welding heads based on molten pool temperature. The method includes the following steps: acquiring a real-time molten pool temperature sequence from a coaxial temperature sensor of the handheld laser welding head; dynamically adjusting the polynomial order of a Savitzky-Golay filter based on the second-order time derivative of the molten pool temperature sequence to filter the molten pool temperature sequence and obtain a real-time molten pool temperature measurement; determining, based on the real-time molten pool temperature measurement and its first-order time derivative, that the current welding process is in one of three stages: arc initiation, steady-state welding, or arc termination and crater filling, by comparing with preset arc initiation temperature thresholds, steady-state temperature ranges, and process adjustment gradient thresholds; and calculating the temperature deviation between the target temperature and the real-time molten pool temperature measurement when the process is determined to be in the steady-state welding stage. This invention achieves differentiated power control strategies for different process stages and improves the overall quality and consistency of handheld laser welding.

Description

Technical Field

[0001] This invention belongs to the field of power control technology, specifically relating to a power control method and system for handheld laser welding heads based on molten pool temperature. Background Technology

[0002] Handheld laser welding technology is widely used in metal processing, automotive manufacturing, and hardware products due to its advantages such as flexible operation, high welding speed, and small heat-affected zone. However, unlike automated laser welding, handheld welding is significantly affected by human factors, such as welder hand tremors, real-time changes in welding speed, adjustments to the welding torch posture, and unevenness in the workpiece gap. These uncertainties collectively lead to instability in the welding process. This instability directly affects the energy absorption and heat distribution of the molten pool, easily causing welding defects such as uneven penetration, burn-through, incomplete penetration, or poor weld formation. To ensure the stability and consistency of welding quality, real-time closed-loop control of the laser power is essential. Molten pool temperature is the most direct and important physical quantity reflecting the energy balance and molten pool state during the welding process. Real-time monitoring of the molten pool temperature and using it as a feedback signal is a key technical approach to achieving closed-loop control of the handheld laser welding process.

[0003] In existing technologies, the molten pool temperature signal obtained from coaxial temperature sensors is often mixed with a large amount of noise interference from plasma radiation, metal vapor, and ambient light. Filtering methods such as mean filtering, while removing noise, introduce significant signal delays, affecting the real-time response capability of the control system. Furthermore, the control parameters of conventional PID controllers are usually fixed, making it difficult to adapt to the nonlinear time-varying characteristics caused by speed and angle changes during handheld welding. This can easily lead to power output overshoot or slow response, failing to balance speed and stability. Moreover, a complete welding process includes different stages such as arc initiation, steady-state welding, and arc termination and crater filling, each with different power control requirements and objectives. For example, the arc initiation stage requires rapid establishment of a stable molten pool, the steady-state stage requires maintaining a constant temperature, and the arc termination stage requires a gradual power reduction to avoid crater cracks. Most existing control methods employ a single control logic, lacking effective identification of welding process stages and differentiated, refined control strategies, making it difficult to meet the full-process requirements of high-quality welding. Summary of the Invention

[0004] This invention provides a power control method and system for a handheld laser welding head based on the temperature of the molten pool, in order to solve the technical problem that the existing control methods use a single control logic, which makes it difficult to meet the process requirements of high-quality welding throughout the entire process.

[0005] In a first aspect, the present invention provides a power control method for a handheld laser welding head based on the molten pool temperature, comprising the following steps:

[0006] The molten pool temperature sequence is acquired in real time by the coaxial temperature sensor of the handheld laser welding head. Based on the second time derivative of the molten pool temperature sequence, the polynomial order of the Savitzky-Golay filter is dynamically adjusted to filter the molten pool temperature sequence and obtain the real-time molten pool temperature measurement value.

[0007] Based on the real-time molten pool temperature measurement and its first time derivative, and by comparing with the preset arc initiation temperature threshold, steady-state temperature range and process adjustment gradient threshold, the current welding process is determined to be in one of the three stages: arc initiation, steady-state welding or arc termination and crater filling.

[0008] When the process is determined to be in a steady-state welding stage, the temperature deviation between the target temperature and the real-time molten pool temperature measurement is calculated; the variance of the temperature deviation over a predetermined number of cycles is statistically analyzed, and the process stability factor is calculated using a logarithmic mapping function to update the process adjustment gradient threshold; based on the absolute value of the temperature deviation and the process stability factor, the nonlinear gain coefficient is calculated using a bivariate sigmoid function, and the integral term of the temperature deviation, nonlinear gain coefficient, and laser power adjustment over a predetermined number of cycles is fused to calculate the laser power adjustment for the current cycle;

[0009] When the arc initiation stage is determined, the laser power is output according to the preset power-time curve; when the arc termination and crater filling stage is determined, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage, and the laser power is controlled to decrease slowly according to the exponential decay function; the laser power or laser power adjustment amount calculated for each stage is output to the laser controller.

[0010] Furthermore, the polynomial order of the Savitzky-Golay filter is dynamically adjusted based on the second-order time derivative of the molten pool temperature sequence to filter the molten pool temperature sequence, including:

[0011] Calculate the absolute value of the second time derivative of the molten pool temperature sequence at the current moment. ;

[0012] Get the order adjustment threshold ;

[0013] when At that time, the polynomial order of the Savitzky-Golay filter was obtained as 2nd order;

[0014] when At that time, the polynomial order of the Savitzky-Golay filter was obtained as 3rd order;

[0015] The obtained polynomial order is used to filter the molten pool temperature sequence.

[0016] Furthermore, the current welding process is determined to be in one of three stages: arc initiation, steady-state welding, or arc termination and crater filling, including:

[0017] Real-time molten pool temperature measurement value With the preset arc initiation temperature threshold If a comparison is made, If so, it is determined to be the arc initiation stage;

[0018] like Then the real-time molten pool temperature measurement value will be... With the preset steady-state temperature range If a comparison is made, If so, it is determined to be in the steady-state welding stage;

[0019] like Then calculate the first-order time derivative of the real-time molten pool temperature measurement. The first time derivative of the real-time molten pool temperature measurement value Adjusting gradient threshold with process If a comparison is made, If so, it is determined to be the arc-closing and pit-filling stage.

[0020] Furthermore, the variance of temperature deviations over a predetermined number of periods is statistically analyzed, and a process stability factor is calculated using a logarithmic mapping function. The gradient threshold for updating the process is then adjusted, including:

[0021] The number of cycles for obtaining the predetermined quantity is N;

[0022] Obtain the temperature deviation sequence over the past N periods and calculate the variance of the temperature deviation sequence. ;

[0023] The process stability factor S is calculated using the following formula:

[0024]

[0025] in A process stability factor with the dimension of temperature. These are mapping coefficients with the dimension of temperature. The coefficients used to make the variance term dimensionless are dimensionless and are the reciprocal of the square of the temperature.

[0026] The process adjustment gradient threshold is updated based on the process stability factor S using the following formula. :

[0027]

[0028] in Based on the gradient threshold, It is a proportionality constant, with the dimension being the reciprocal of time.

[0029] Furthermore, based on the absolute value of the temperature deviation and the process stability factor, the nonlinear gain coefficient is calculated using a bivariate sigmoid function. By integrating the temperature deviation, the nonlinear gain coefficient, and the integral term of the laser power adjustment over a predetermined number of cycles, the laser power adjustment for the current cycle is calculated, including:

[0030] The nonlinear gain coefficient is calculated using the following formula. :

[0031]

[0032] in The temperature deviation for the current cycle is given by S, where S is the process stability factor. The preset maximum gain coefficient, This is the preset kurtosis coefficient for the S-shaped function, with the dimension being the reciprocal of temperature;

[0033] The laser power adjustment for the current cycle is calculated using the following formula. :

[0034]

[0035] in For proportional gain, Here, M represents the integral gain, and M is the previously predetermined number of cycles used to calculate the integral term. This represents the temperature deviation over the j-th past period.

[0036] Furthermore, when the stage is determined to be the arc-closing and crater-filling stage, based on the process stability factor at the end of the steady-state welding stage, the decay time constant of the exponential decay function is determined, and the laser power is controlled to gradually decrease according to the exponential decay function, including:

[0037] Obtain the process stability factor for the last cycle of the steady-state welding stage. ;

[0038] Through formula Determine the decay time constant ,in Based on the decay time constant, The proportionality constant has the dimension of time / temperature; the laser power in the current cycle. Control according to the following formula:

[0039]

[0040] in This is the final laser power during the steady-state welding stage. This refers to the duration after entering the arc-closing and crater-filling phase.

[0041] Secondly, the present invention provides a handheld laser welding head power control system based on molten pool temperature, comprising the following modules:

[0042] The measurement acquisition module acquires the molten pool temperature sequence in real time from the coaxial temperature sensor of the handheld laser welding head. Based on the second time derivative of the molten pool temperature sequence, the polynomial order of the Savitzky-Golay filter is dynamically adjusted to filter the molten pool temperature sequence and obtain the real-time molten pool temperature measurement value.

[0043] The stage determination module, based on the real-time molten pool temperature measurement value and its first time derivative, compares it with the preset arc initiation temperature threshold, steady-state temperature range and process adjustment gradient threshold to determine whether the current welding process is in one of the three stages: arc initiation, steady-state welding or arc termination and crater filling.

[0044] The power adjustment module calculates the temperature deviation between the target temperature and the real-time molten pool temperature measurement when the steady-state welding stage is determined. It also calculates the variance of the temperature deviation over a predetermined number of cycles, calculates the process stability factor using a logarithmic mapping function, and updates the process adjustment gradient threshold. Based on the absolute value of the temperature deviation and the process stability factor, it calculates the nonlinear gain coefficient using a bivariate sigmoid function, and integrates the temperature deviation, the nonlinear gain coefficient, and the integral term of the laser power adjustment over a predetermined number of cycles to calculate the laser power adjustment for the current cycle.

[0045] The adjustment output module outputs laser power according to the preset power-time curve when the arc initiation stage is determined; when the arc termination and crater filling stage is determined, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage, and the laser power is controlled to decrease slowly according to the exponential decay function; the laser power or laser power adjustment amount calculated for each stage is output to the laser controller.

[0046] Furthermore, the polynomial order of the Savitzky-Golay filter is dynamically adjusted based on the second-order time derivative of the molten pool temperature sequence to filter the molten pool temperature sequence, including:

[0047] Calculate the absolute value of the second time derivative of the molten pool temperature sequence at the current moment. ;

[0048] Get the order adjustment threshold ;

[0049] when At that time, the polynomial order of the Savitzky-Golay filter was obtained as 2nd order;

[0050] when At that time, the polynomial order of the Savitzky-Golay filter was obtained as 3rd order;

[0051] The obtained polynomial order is used to filter the molten pool temperature sequence.

[0052] Furthermore, the current welding process is determined to be in one of three stages: arc initiation, steady-state welding, or arc termination and crater filling, including:

[0053] Real-time molten pool temperature measurement value With the preset arc initiation temperature threshold If a comparison is made, If so, it is determined to be the arc initiation stage;

[0054] like Then the real-time molten pool temperature measurement value will be... With the preset steady-state temperature range If a comparison is made, If so, it is determined to be in the steady-state welding stage;

[0055] like Then calculate the first-order time derivative of the real-time molten pool temperature measurement. The first time derivative of the real-time molten pool temperature measurement value Adjusting gradient threshold with process If a comparison is made, If so, it is determined to be the arc-closing and pit-filling stage.

[0056] Furthermore, the variance of temperature deviations over a predetermined number of periods is statistically analyzed, and a process stability factor is calculated using a logarithmic mapping function. The gradient threshold for updating the process is then adjusted, including:

[0057] The number of cycles for obtaining the predetermined quantity is N;

[0058] Obtain the temperature deviation sequence over the past N periods and calculate the variance of the temperature deviation sequence. ;

[0059] The process stability factor S is calculated using the following formula:

[0060]

[0061] in A process stability factor with the dimension of temperature. These are mapping coefficients with the dimension of temperature. The coefficients used to make the variance term dimensionless are dimensionless and are the reciprocal of the square of the temperature.

[0062] The process adjustment gradient threshold is updated based on the process stability factor S using the following formula. :

[0063]

[0064] in Based on the gradient threshold, It is a proportionality constant, with the dimension being the reciprocal of time.

[0065] The beneficial effects are as follows: By filtering the molten pool temperature signal, high-frequency noise can be suppressed while retaining key temperature characteristics. Furthermore, by dividing the welding process into arc initiation, steady-state welding, and arc termination / filling stages, differentiated power control strategies for different process stages are achieved. In the core steady-state welding stage, the fluctuation of the welding process is measured by a process stability factor, and nonlinear gain is calculated in conjunction with temperature deviation. This ensures that the power adjustment not only responds to the magnitude of temperature deviation but also takes into account the immediate stable state of the welding process, suppressing power over-adjustment or oscillation caused by factors such as hand tremors and changes in welding speed. The arc initiation power curve and arc termination power attenuation function ensure the reliability of the welding start and the defect-free formation of the end point, improving the overall quality and consistency of handheld laser welding. Attached Figure Description

[0066] Figure 1 This is a flowchart of a handheld laser welding head power control method based on molten pool temperature. Detailed Implementation

[0067] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0068] An embodiment of the handheld laser welding head power control method based on molten pool temperature provided by the present invention:

[0069] like Figure 1 As shown, the power control method for a handheld laser welding head based on the molten pool temperature includes the following steps:

[0070] S1: Acquire the molten pool temperature sequence in real time from the coaxial temperature sensor of the handheld laser welding head. Based on the second-order time derivative of the molten pool temperature sequence, dynamically adjust the polynomial order of the Savitzky-Golay filter to filter the molten pool temperature sequence and obtain the real-time molten pool temperature measurement value.

[0071] The handheld laser welding head's chip reads the voltage signal from the coaxial temperature sensor and converts it into temperature values ​​according to the calibration curve, forming a temperature time series. The second-order time derivative of the temperature time series is calculated using, for example, the central difference method. If the absolute value of the second-order time derivative exceeds a preset abrupt change threshold, it indicates that the signal may contain spike noise or is in a stage of rapid process change. In this case, the polynomial order of the Savitzky-Golay filter is set to a lower second order to enhance smoothing and noise reduction capabilities. Conversely, if it is below the abrupt change threshold, the order is set to a higher fourth order to better preserve the details of temperature changes. Using the adjusted order and a fixed window length, such as 21 sampling points, polynomial fitting filtering is performed on the latest temperature data segment to obtain the real-time melt pool temperature measurement value for the current period.

[0072] In an optional embodiment, the polynomial order of the Savitzky-Golay filter is dynamically adjusted based on the second-order time derivative of the molten pool temperature sequence to filter the molten pool temperature sequence, including:

[0073] Calculate the absolute value of the second time derivative of the molten pool temperature sequence at the current moment. ;

[0074] Get the order adjustment threshold ;

[0075] when At that time, the polynomial order of the Savitzky-Golay filter was obtained as 2nd order;

[0076] when At that time, the polynomial order of the Savitzky-Golay filter was found to be 3rd order;

[0077] The obtained polynomial order is used to filter the molten pool temperature sequence.

[0078] The absolute value of the second time derivative of the molten pool temperature represents the curvature or drastic change in the temperature profile. In stable welding processes, temperature fluctuations are small, and the absolute value of the second derivative typically remains at a low level. For example, obtaining the order adjustment threshold... The temperature is 100,000 degrees Celsius per square second. The absolute value of the calculated second-order time derivative when the welding process is stable. The temperature is likely 60,000 degrees Celsius per square second, which is less than the order adjustment threshold. Therefore, a third-order polynomial is used for filtering. The third-order polynomial can better fit the true shape of the signal, retain more high-frequency details, and achieve accurate tracking of the molten pool temperature.

[0079] When sudden disturbances occur during welding, such as spatter, porosity, or unstable wire feeding, the temperature of the molten pool can change drastically and instantaneously. This drastic change will cause the absolute value of the second derivative to spike instantaneously. For example, a large spatter event could cause the absolute value of the second time derivative to increase dramatically. It surged to 250,000 degrees Celsius per square second, a value far exceeding the obtained order adjustment threshold. At this point, the polynomial order of the filter is reduced to second order. Second-order polynomials have stronger smoothing capabilities, effectively suppressing the severe spikes caused by noise and preventing the control system from over-responding due to erroneous signals. When the disturbance disappears and the temperature change returns to a smooth state, the absolute value of the second-order time derivative... Once the filter order falls back below the order adjustment threshold, it returns to the 3rd order, ensuring the adaptive filtering effect throughout the process.

[0080] S2, based on the real-time molten pool temperature measurement value and its first time derivative, compares with the preset arc initiation temperature threshold, steady-state temperature range and process adjustment gradient threshold to determine whether the current welding process is in one of the three stages: arc initiation, steady-state welding or arc termination and pit filling.

[0081] In one embodiment, the pre-set arc ignition temperature threshold is 1400 degrees Celsius, and the steady-state temperature range is 1750 to 1850 degrees Celsius. Those skilled in the art should understand that the above data is merely exemplary, and the corresponding values ​​will differ due to variations in the maximum power and application of different handheld laser welding machines. In each control cycle, the real-time molten pool temperature measurement is checked. If it is below 1400 degrees Celsius, it is determined to be in the arc ignition stage. If the temperature has entered the steady-state range of 1750 to 1850 degrees Celsius, the first-order time derivative of the temperature is further calculated. If the absolute value of the first-order time derivative is less than the current process adjustment gradient threshold, for example, 200 degrees Celsius per second, it is determined to be in the steady-state welding stage, indicating that the molten pool has stabilized. If the controller receives a signal from the operator to release the welding torch switch, it immediately determines to be in the arc termination and crater filling stage.

[0082] In an optional embodiment, determining whether the current welding process is in one of three stages—arc initiation, steady-state welding, or arc termination and crater filling—includes:

[0083] Real-time molten pool temperature measurement value With the preset arc initiation temperature threshold If a comparison is made, If so, it is determined to be the arc initiation stage;

[0084] like Then the real-time molten pool temperature measurement value will be... With the preset steady-state temperature range If a comparison is made, If so, it is determined to be in the steady-state welding stage;

[0085] like Then calculate the first-order time derivative of the real-time molten pool temperature measurement. The first time derivative of the real-time molten pool temperature measurement value Adjusting gradient threshold with process If a comparison is made, If so, it is determined to be the arc-closing and pit-filling stage.

[0086] The welding process can be divided by setting reasonable arc initiation temperature thresholds and process adjustment gradient thresholds. Taking the welding of stainless steel as an example, the melting point of stainless steel is approximately 1450 degrees Celsius, and the desired molten pool temperature during steady-state welding is around 1600 degrees Celsius. Therefore, the arc initiation temperature threshold can be... Set to 1200 degrees Celsius, the lower limit of the steady-state temperature range. Set to 1550 degrees Celsius, upper limit Set to 1650 degrees Celsius. At the start of welding, the workpiece is heated from room temperature, and the temperature of the molten pool is measured in real-time. At 850 degrees Celsius, because it is lower than The system determines that the current stage is the arc initiation stage. As the laser continues to heat the weld, when the temperature rises to 1610 degrees Celsius, the real-time molten pool temperature measurement falls into the steady-state temperature range, and the system switches to the steady-state welding stage.

[0087] When the welding command ends, the laser power begins to decrease, and the molten pool enters the cooling process. At this time, the temperature of the molten pool may briefly exceed [a certain value] due to thermal inertia. However, the more crucial criterion is the cooling rate. For example, adjusting the gradient threshold of the process. Set to 500 degrees Celsius per second. At the instant welding ends, the molten pool temperature begins to drop rapidly; this is the first-order time derivative of the real-time molten pool temperature measurement. It could reach -800 degrees Celsius per second. Since -800 is less than -500, this satisfies the condition. The conditions were met, and the judgment process entered the arc-closing and crater-filling stage. The judgment triggered a specific power reduction procedure to ensure good weld end formation and avoid defects such as shrinkage cavities.

[0088] S3, when the steady-state welding stage is determined, calculate the temperature deviation between the target temperature and the real-time molten pool temperature measurement; statistically analyze the variance of the temperature deviation within a predetermined number of cycles, calculate the process stability factor using a logarithmic mapping function, and update the process adjustment gradient threshold; based on the absolute value of the temperature deviation and the process stability factor, calculate the nonlinear gain coefficient using a bivariate sigmoid function, and integrate the temperature deviation, the nonlinear gain coefficient, and the integral term of the laser power adjustment within a predetermined number of cycles to calculate the laser power adjustment for the current cycle.

[0089] Using the median of the steady-state temperature range, 1800 degrees Celsius, as the target temperature, the deviation between the target temperature and the real-time measured molten pool temperature is calculated. The controller maintains a 50-cycle FIFO queue, storing the temperature deviation values ​​for the most recent 50 cycles, and calculates the variance of the data in the FIFO queue in real time. The process stability factor is expressed by a formula, for example... The calculation involves k being the adjustment coefficient; a larger variance corresponds to a smaller S, indicating greater process instability. The updated process adjustment gradient threshold is equal to a base threshold multiplied by the process stability factor S. The nonlinear gain coefficient is calculated using a bivariate sigmoid function based on the absolute value of the temperature deviation and the process stability factor S. This bivariate sigmoid function increases the gain as the process becomes more stable or the temperature deviation increases. The laser power adjustment consists of three parts: the nonlinear gain coefficient multiplied by the current temperature deviation, an integral coefficient multiplied by the sum of temperature deviations over the past 100 cycles, and a differential coefficient multiplied by the difference between the current temperature deviation and the temperature deviation of the previous cycle.

[0090] In an optional embodiment, the variance of temperature deviations over a predetermined number of past periods is statistically analyzed, a process stability factor is calculated using a logarithmic mapping function, and the process adjustment gradient threshold is updated, including:

[0091] The number of cycles for obtaining the predetermined quantity is N;

[0092] Obtain the temperature deviation sequence over the past N periods and calculate the variance of the temperature deviation sequence. ;

[0093] The process stability factor S is calculated using the following formula:

[0094]

[0095] in A process stability factor with the dimension of temperature. These are mapping coefficients with the dimension of temperature. The coefficients used to make the variance term dimensionless are dimensionless and are the reciprocal of the square of the temperature.

[0096] The process adjustment gradient threshold is updated based on the process stability factor S using the following formula. :

[0097]

[0098] in Based on the gradient threshold, It is a proportionality constant, with the dimension being the reciprocal of time.

[0099] For example, if the number of statistical periods N is 100, and each period is 5 milliseconds, then the stability over the past 500 milliseconds is examined. Under very stable welding conditions, the variance of the temperature deviation sequence is... It could be 25 degrees Celsius squared. The mapping coefficients A and B are set to 10 degrees Celsius and 0.01 degrees Celsius squared, respectively. The calculated process stability factor S is approximately 2.2 degrees Celsius, a very small value that intuitively reflects the high stability of the process.

[0100] The process stability factor S is used to adaptively adjust the judgment sensitivity during the arc-closing and crater-filling stage. The basic gradient threshold is obtained. 500 degrees Celsius per second, scaling factor The threshold is 20 per second. Under the above stable conditions, the updated process adjusts the gradient threshold. Approximately 544 degrees Celsius per second. Conversely, if the welding process is unstable, for example, due to drastic temperature fluctuations caused by variations in shielding gas flow, the variance... This could increase to 400 squared degrees Celsius. At this point, the calculated process stability factor S will increase to approximately 16.1 degrees Celsius. The updated process adjustment gradient threshold... This is approximately equal to 822 degrees Celsius per second. This adjustment is beneficial because it increases the process adjustment gradient threshold when the process is unstable, avoids misjudging normal, violent fluctuations as the arc-ending phase, and enhances the system's adaptability and robustness to different operating conditions.

[0101] In an optional embodiment, based on the absolute value of the temperature deviation and the process stability factor, a nonlinear gain coefficient is calculated using a bivariate sigmoid function. The temperature deviation, the nonlinear gain coefficient, and the integral term of the laser power adjustment over a previously predetermined number of cycles are then fused to calculate the laser power adjustment for the current cycle, including:

[0102] The nonlinear gain coefficient is calculated using the following formula. :

[0103]

[0104] in The temperature deviation for the current cycle is given by S, where S is the process stability factor. The preset maximum gain coefficient, This is the preset kurtosis coefficient for the S-shaped function, with the dimension being the reciprocal of temperature;

[0105] The laser power adjustment for the current cycle is calculated using the following formula. :

[0106]

[0107] in For proportional gain, Here, M represents the integral gain, and M is the previously predetermined number of cycles used to calculate the integral term. This represents the temperature deviation over the j-th past period.

[0108] Obtain the maximum gain coefficient of the parameter The kurtosis coefficient of the S-shaped function is 5. The proportional gain is 0.2 per degree Celsius. The integral gain is 0.5 watts per degree Celsius. The cycle length is 0.05 watts per degree Celsius. Assume the current process is very stable, with a process stability factor S of 3 degrees Celsius. If the temperature deviation e(t) is small, for example, 2 degrees Celsius, the calculated nonlinear gain coefficient will be less than the absolute value of the temperature deviation compared to the process stability factor. It will approach the maximum gain coefficient. Half of that, 2.5, results in a mild controller response, avoiding overshooting for minor disturbances.

[0109] Under the same stable conditions, i.e., a process stability factor S of 3 degrees Celsius, if a large temperature deviation occurs, such as e(t) of 20 degrees Celsius, the absolute value of the temperature deviation is much greater than the stability factor S, and the calculated nonlinear gain coefficient... It will rapidly increase and approach its maximum value of 5. This makes the effective gain of the proportional term... This significantly enhances the ability to rapidly correct large temperature deviations. On the other hand, if the process itself is unstable and the process stability factor S is large, for example, 15 degrees Celsius, then even if a temperature deviation of 10 degrees Celsius occurs, the absolute value of the temperature deviation will still be smaller than the process stability factor S, resulting in a lower nonlinear gain coefficient. It will also be kept at a low level to prevent the controller from making overly aggressive adjustments in an already fluctuating system, thus exacerbating oscillations. Laser power adjustment amount. It combines proportional regulation with an integral term to eliminate steady-state error, thus achieving closed-loop temperature control.

[0110] S4, when the arc initiation stage is determined, the laser power is output according to the preset power-time curve; when the arc termination and crater filling stage is determined, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage, and the laser power is controlled to decrease slowly according to the exponential decay function; the laser power or laser power adjustment amount calculated for each stage is output to the laser controller.

[0111] During the arc initiation phase, the laser power linearly increases from 0 to 120% of the rated power within the first 0.1 seconds to rapidly establish the molten pool, and then linearly decreases to the rated power within the next 0.2 seconds, after which closed-loop control is entered. During the arc termination and crater filling phase, the process stability factor S of the last cycle before entering this phase is obtained. For example, the decay time constant τ = basic constant + scaling factor × S; that is, the more stable the steady-state process, the larger the value of S, and the longer the power decay time. The laser power P is calculated from the power value at the start of the arc termination phase. Begin, according to The power decreases exponentially until it reaches zero, where t is the time after entering the arc-ending phase.

[0112] During the arc initiation and arc termination filling stages, the target laser power value, calculated directly based on a preset curve or exponential decay function, is sent to the laser controller via the CAN bus. During the steady-state welding stage, the calculated laser power adjustment is added to the laser power value of the previous cycle to obtain the target laser power value for the current cycle, which is then sent to the laser controller for execution.

[0113] In an optional embodiment, when the welding stage is determined to be the arc-closing and crater-filling stage, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage, and the laser power is controlled to gradually decrease according to the exponential decay function, including:

[0114] Obtain the process stability factor for the last cycle of the steady-state welding stage. ;

[0115] Through formula Determine the decay time constant ,in Based on the decay time constant, The proportionality constant has the dimension of time / temperature; the laser power in the current cycle. Control according to the following formula:

[0116]

[0117] in This is the final laser power during the steady-state welding stage. This refers to the duration after entering the arc-closing and crater-filling phase.

[0118] For example, obtaining the basic decay time constant. For 80 milliseconds, the scaling factor The stability factor is 6 milliseconds per degree Celsius. Assuming welding ends after a very smooth welding process, the process stability factor for the last cycle of the steady-state welding phase is... The temperature is only 2 degrees Celsius. The calculated decay time constant at this temperature... It is 92 milliseconds. If the final laser power during the steady-state welding stage... At 1800 watts, the power during arc termination and crater filling will decrease exponentially according to a relatively fast time constant of 92 milliseconds. This corresponds to a small and stable molten pool, which requires a rapid cooling rate to form a dense weld end.

[0119] Conversely, if the welding process experiences an unstable state near its end, such as significant fluctuations in heat input, leading to a decrease in the process stability factor... The temperature reached as high as 20 degrees Celsius. This indicates that the molten pool before the end may have been too large or overheated. In this case, the calculated decay time constant... It will be 200 milliseconds. The laser power will decay slowly exponentially according to this longer time constant. This slower energy withdrawal process allows sufficient solidification time for larger or more unstable molten pools, which helps gas escape and metal feeding, effectively preventing defects such as crater cracks and porosity commonly seen during the arc-closing and crater-filling stages, and improving the overall quality of the weld joint.

[0120] An embodiment of the handheld laser welding head power control system based on molten pool temperature provided by the present invention includes the following modules:

[0121] The measurement acquisition module acquires the molten pool temperature sequence in real time from the coaxial temperature sensor of the handheld laser welding head. Based on the second time derivative of the molten pool temperature sequence, the polynomial order of the Savitzky-Golay filter is dynamically adjusted to filter the molten pool temperature sequence and obtain the real-time molten pool temperature measurement value.

[0122] The stage determination module, based on the real-time molten pool temperature measurement value and its first time derivative, compares it with the preset arc initiation temperature threshold, steady-state temperature range and process adjustment gradient threshold to determine whether the current welding process is in one of the three stages: arc initiation, steady-state welding or arc termination and crater filling.

[0123] The power adjustment module calculates the temperature deviation between the target temperature and the real-time molten pool temperature measurement when the steady-state welding stage is determined. It also calculates the variance of the temperature deviation over a predetermined number of cycles, calculates the process stability factor using a logarithmic mapping function, and updates the process adjustment gradient threshold. Based on the absolute value of the temperature deviation and the process stability factor, it calculates the nonlinear gain coefficient using a bivariate sigmoid function, and integrates the temperature deviation, the nonlinear gain coefficient, and the integral term of the laser power adjustment over a predetermined number of cycles to calculate the laser power adjustment for the current cycle.

[0124] The adjustment output module outputs laser power according to the preset power-time curve when the arc initiation stage is determined; when the arc termination and crater filling stage is determined, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage, and the laser power is controlled to decrease slowly according to the exponential decay function; the laser power or laser power adjustment amount calculated for each stage is output to the laser controller.

[0125] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A power control method for a handheld laser welding head based on molten pool temperature, characterized in that, Includes the following steps: The process involves acquiring the molten pool temperature sequence in real time from the coaxial temperature sensor of the handheld laser welding head, dynamically adjusting the polynomial order of the Savitzky-Golay filter based on the second-order time derivative of the molten pool temperature sequence, and filtering the molten pool temperature sequence. This includes calculating the absolute value of the second-order time derivative of the molten pool temperature sequence at the current moment. ; Obtain the order adjustment threshold ;when When the order of the polynomial of the Savitzky-Golay filter is 2, then... At that time, the polynomial order of the Savitzky-Golay filter was obtained as 3rd order; the obtained polynomial order was used to filter the molten pool temperature sequence to obtain the real-time molten pool temperature measurement value. Based on the real-time molten pool temperature measurement and its first time derivative, and by comparing with the preset arc initiation temperature threshold, steady-state temperature range and process adjustment gradient threshold, the current welding process is determined to be in one of the three stages: arc initiation, steady-state welding or arc termination and crater filling. When the process is determined to be in a steady-state welding stage, the temperature deviation between the target temperature and the real-time molten pool temperature measurement is calculated. The variance of the temperature deviation over a predetermined number of periods is statistically analyzed, and a process stability factor is calculated using a logarithmic mapping function. The process adjustment gradient threshold is then updated, including: obtaining a predetermined number of periods (N); obtaining the temperature deviation sequence over the past N periods; and calculating the variance of the temperature deviation sequence. The process stability factor S is calculated using the following formula: ,in A process stability factor with the dimension of temperature. These are mapping coefficients with the dimension of temperature. The coefficients used to make the variance term dimensionless are dimensionless and are the reciprocal of the square of the temperature. The process adjustment gradient threshold is updated based on the process stability factor S using the following formula. : ,in Based on the gradient threshold, The proportionality coefficient is the reciprocal of time. Based on the absolute value of the temperature deviation and the process stability factor, the nonlinear gain coefficient is calculated using a bivariate sigmoid function. The temperature deviation, the nonlinear gain coefficient, and the integral term of the laser power adjustment over a predetermined number of cycles are then fused to calculate the laser power adjustment for the current cycle. This includes calculating the nonlinear gain coefficient using the following formula. : ,in This represents the temperature deviation for the current cycle. The preset maximum gain coefficient, This is the preset kurtosis coefficient for the S-shaped function, with the dimension being the reciprocal of temperature; The laser power adjustment for the current cycle is calculated using the following formula. : ,in For proportional gain, Here, M represents the integral gain, and M is the previously predetermined number of cycles used to calculate the integral term. This represents the temperature deviation over the j-th past period. When the arc initiation stage is determined, the laser power is output according to the preset power-time curve; when the arc termination and crater filling stage is determined, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage, and the laser power is controlled to decrease slowly according to the exponential decay function; the laser power or laser power adjustment amount calculated for each stage is output to the laser controller.

2. The handheld laser welding head power control method based on molten pool temperature according to claim 1, characterized in that, The current welding process is determined to be in one of three stages: arc initiation, steady-state welding, or arc termination and crater filling. Real-time molten pool temperature measurement value With the preset arc initiation temperature threshold If a comparison is made, If so, it is determined to be the arc initiation stage; like Then the real-time molten pool temperature measurement value will be... With the preset steady-state temperature range If a comparison is made, If so, it is determined to be in the steady-state welding stage; like Then calculate the first-order time derivative of the real-time molten pool temperature measurement. The first time derivative of the real-time molten pool temperature measurement value Adjusting gradient threshold with process If a comparison is made, If so, it is determined to be the arc-closing and pit-filling stage.

3. The handheld laser welding head power control method based on molten pool temperature according to any one of claims 1-2, characterized in that, When the welding stage is determined to be the arc-closing and crater-filling stage, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage. The laser power is then controlled to gradually decrease according to the exponential decay function, including: Obtain the process stability factor for the last cycle of the steady-state welding stage. ; Through formula Determine the decay time constant ,in Based on the decay time constant, The proportionality constant has the dimension of time / temperature; the laser power in the current cycle. Control according to the following formula: in This is the final laser power during the steady-state welding stage. This refers to the duration after entering the arc-closing and crater-filling phase.

4. A system for implementing the handheld laser welding head power control method based on molten pool temperature as described in any one of claims 1-3, characterized in that, Includes the following modules: The measurement acquisition module acquires the molten pool temperature sequence in real time from the coaxial temperature sensor of the handheld laser welding head. Based on the second time derivative of the molten pool temperature sequence, the polynomial order of the Savitzky-Golay filter is dynamically adjusted to filter the molten pool temperature sequence and obtain the real-time molten pool temperature measurement value. The stage determination module, based on the real-time molten pool temperature measurement value and its first time derivative, compares it with the preset arc initiation temperature threshold, steady-state temperature range and process adjustment gradient threshold to determine whether the current welding process is in one of the three stages: arc initiation, steady-state welding or arc termination and crater filling. When the power adjustment module determines that the welding stage is in a steady state, it calculates the temperature deviation between the target temperature and the real-time molten pool temperature measurement; it statistically analyzes the variance of the temperature deviation over a predetermined number of periods in the past, calculates the process stability factor using a logarithmic mapping function, and updates the process adjustment gradient threshold. Based on the absolute value of the temperature deviation and the process stability factor, the nonlinear gain coefficient is calculated using a bivariate sigmoid function. The temperature deviation, the nonlinear gain coefficient, and the integral term of the laser power adjustment within a previously predetermined number of cycles are then fused to calculate the laser power adjustment for the current cycle. The adjustment output module outputs laser power according to the preset power-time curve when the arc initiation stage is determined; when the arc termination and crater filling stage is determined, the decay time constant of the exponential decay function is determined based on the process stability factor at the end of the steady-state welding stage, and the laser power is controlled to decrease slowly according to the exponential decay function; the laser power or laser power adjustment amount calculated for each stage is output to the laser controller.

5. The handheld laser welding head power control system based on molten pool temperature according to claim 4, characterized in that, The polynomial order of the Savitzky-Golay filter is dynamically adjusted based on the second-order time derivative of the molten pool temperature sequence to filter the molten pool temperature sequence, including: Calculate the absolute value of the second time derivative of the molten pool temperature sequence at the current moment. ; Get the order adjustment threshold ; when At that time, the polynomial order of the Savitzky-Golay filter was obtained as 2nd order; when At that time, the polynomial order of the Savitzky-Golay filter was obtained as 3rd order; The obtained polynomial order is used to filter the molten pool temperature sequence.

6. The handheld laser welding head power control system based on molten pool temperature according to claim 5, characterized in that, The current welding process is determined to be in one of three stages: arc initiation, steady-state welding, or arc termination and crater filling. Real-time molten pool temperature measurement value With the preset arc initiation temperature threshold If a comparison is made, If so, it is determined to be the arc initiation stage; like Then the real-time molten pool temperature measurement value will be... With the preset steady-state temperature range If a comparison is made, If so, it is determined to be in the steady-state welding stage; like Then calculate the first-order time derivative of the real-time molten pool temperature measurement. The first time derivative of the real-time molten pool temperature measurement value Adjusting gradient threshold with process If a comparison is made, If so, it is determined to be the arc-closing and pit-filling stage.

7. The handheld laser welding head power control system based on molten pool temperature according to claim 4, characterized in that, The variance of temperature deviations over a predetermined number of periods is statistically analyzed, and a process stability factor is calculated using a logarithmic mapping function. The gradient threshold is then adjusted during the update process, including: The number of cycles for obtaining the predetermined quantity is N; Obtain the temperature deviation sequence over the past N periods and calculate the variance of the temperature deviation sequence. ; The process stability factor S is calculated using the following formula: in A process stability factor with the dimension of temperature. These are mapping coefficients with the dimension of temperature. The coefficients used to make the variance term dimensionless are dimensionless and are the reciprocal of the square of the temperature. The process adjustment gradient threshold is updated based on the process stability factor S using the following formula. : in Based on the gradient threshold, It is a proportionality constant, with the dimension being the reciprocal of time.

Images