Adaptive energy regulation method for laser processing

Abstract

The application relates to the technical field of laser processing adaptive control, and discloses a kind of adaptive energy regulation and control methods of laser processing, including whole-link dynamic loss reverse calibration step, multi-physical field progressive coupling energy regulation and control step and online modal identification feedforward correction step, through the cooperation of three steps, realize the accurate closed-loop regulation and control of laser processing energy, solve the core problem of insufficient energy regulation and control precision of prior art, improve the stability and consistency of laser processing.

Description

Technical Field

[0001] This invention belongs to the field of adaptive control technology for laser processing, and specifically relates to an adaptive energy regulation method for laser processing. Background Technology

[0002] Laser processing, as a core technology in intelligent manufacturing, is widely used in industrial scenarios such as metal welding, cutting, and cladding. The core of its processing quality depends on the precise control of laser energy. Existing laser processing energy control schemes all assume that the energy emitted from the laser source equals the actual energy received by the material, ignoring the dynamic losses of the laser throughout the entire process from the resonant cavity to the processing end. Especially during long-term processing, the losses caused by factors such as thermal distortion, aging, and contamination of optical components change in real time, causing a continuous deviation in the laser energy reference. This leads to loss of control over energy control precision during processing, ultimately resulting in processing defects and failing to meet the process requirements of high-precision, long-term continuous laser processing.

[0003] Based on the above problems, there is an urgent need for a technical solution that can solve the energy reference deviation caused by dynamic loss throughout the laser process and achieve precise control of laser energy. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention proposes an adaptive energy control method for laser processing, comprising a full-link dynamic loss reverse calibration step, a multi-physics field progressive coupling energy control step, and an online mode recognition feedforward correction step. The full-link dynamic loss reverse calibration step calculates a full-link loss correction coefficient based on the laser source output parameters and the processing end energy feedback parameters. The multi-physics field progressive coupling energy control step calculates an energy demand coefficient based on the full-link loss correction coefficient. The online mode recognition feedforward correction step calculates a feedforward correction coefficient based on the real-time signal of the processing process. The full-link loss correction coefficient, the energy demand coefficient, and the feedforward correction coefficient are used together to control the laser output energy.

[0005] Preferably, the end-to-end dynamic loss reverse calibration step includes a laser source power acquisition step, an optical component status acquisition step, a processing end energy calculation step, and a loss correction coefficient generation step; the laser source power acquisition step acquires real-time output power data of the laser source, the optical component status acquisition step acquires real-time temperature data of the optical component, the processing end energy calculation step calculates real-time incident power data of the processing end based on the spectral signal of the processing process, and the loss correction coefficient generation step generates an end-to-end loss correction coefficient based on the real-time output power data, the real-time temperature data, the real-time incident power data, and the energy utilization rate data.

[0006] Further preferably, the multiphysics progressive coupling energy control step includes a plasma feature acquisition step, a molten pool morphology acquisition step, a plasma energy transmittance calculation step, a molten pool viscosity correction coefficient calculation step, and a vaporization threshold calibration coefficient calculation step. The plasma feature acquisition step acquires real-time spectral feature data of the plasma plume during processing; the molten pool morphology acquisition step acquires real-time morphology data of the molten pool during processing; the plasma energy transmittance calculation step calculates the plasma energy transmittance based on the end-to-end loss correction coefficient and the real-time spectral feature data; the molten pool viscosity correction coefficient calculation step calculates the molten pool viscosity correction coefficient based on the plasma energy transmittance; the vaporization threshold calibration coefficient calculation step calculates the vaporization threshold calibration coefficient based on the molten pool viscosity correction coefficient; and the energy demand coefficient is generated based on the plasma energy transmittance, the molten pool viscosity correction coefficient, and the vaporization threshold calibration coefficient.

[0007] More preferably, the online modal recognition feedforward correction step includes a multi-source signal acquisition step during the processing, an adaptive modal decomposition step, a disturbance feature recognition step, and a feedforward correction coefficient generation step; the multi-source signal acquisition step acquires real-time light intensity signals, acoustic signals, and molten pool morphology signals during the processing; the adaptive modal decomposition step performs modal decomposition on the real-time light intensity signals, acoustic signals, and molten pool morphology signals to obtain characteristic modes; the disturbance feature recognition step identifies the type and intensity of transient disturbances based on the characteristic modes; and the feedforward correction coefficient generation step generates feedforward correction coefficients based on the type and intensity of the transient disturbances.

[0008] More preferably, the update cycle of the loss correction coefficient generation step matches the change cycle of the thermal distortion of the optical element, and the end-to-end loss correction coefficient generated by the loss correction coefficient generation step is updated in real time to the energy reference parameters of the energy control module.

[0009] More preferably, the plasma energy transmittance calculation step, the molten pool viscosity correction coefficient calculation step, and the vaporization threshold calibration coefficient calculation step are executed sequentially in a fixed order. The output result of the plasma energy transmittance calculation step is used as the input reference for the molten pool viscosity correction coefficient calculation step, and the output result of the molten pool viscosity correction coefficient calculation step is used as the input reference for the vaporization threshold calibration coefficient calculation step.

[0010] More preferably, the laser output energy is calculated using a laser output energy adaptive control formula, which is expressed as: Among them, the For laser output energy, the For the target processing energy, the The energy demand coefficient, the The term refers to the end-to-end loss correction coefficient. This is the feedforward correction coefficient.

[0011] More preferably, the adaptive mode decomposition step decomposes the acquired real-time light intensity signal, acoustic signal, and molten pool morphology signal into low-frequency mode, mid-frequency mode, and high-frequency mode. The low-frequency mode corresponds to the slow thermal accumulation process in the processing, the mid-frequency mode corresponds to the steady-state flow process of the molten pool, and the high-frequency mode corresponds to the transient disturbance process in the processing. The disturbance feature identification step identifies the type and intensity of the transient disturbance based on the high-frequency mode.

[0012] More preferably, the optical element includes a laser resonator output mirror, a beam expander, a galvanometer, and a focusing mirror, and the optical element status acquisition step acquires real-time temperature data of the laser resonator output mirror, the beam expander, the galvanometer, and the focusing mirror.

[0013] More preferably, the full-link dynamic loss reverse calibration step, the multi-physics progressive coupling energy control step, and the online modal recognition feedforward correction step are continuously and cyclically executed throughout the entire laser processing process.

[0014] Technical effects: This invention solves the core problem of energy reference deviation caused by dynamic loss across the entire process by implementing a reverse calibration step for dynamic loss across the entire process. By combining a multi-physics progressive coupling energy regulation step with an online modal recognition feedforward correction step, it achieves precise closed-loop regulation of laser processing energy across the entire process and under all operating conditions. This effectively improves the energy regulation accuracy of laser processing and ensures the stability of processing quality under long-term continuous processing and transient conditions. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A diagram showing the overall steps and coefficient relationships for adaptive energy control in laser processing; Figure 2 Here is a flowchart of the reverse calibration steps for full-link dynamic loss; Figure 3 This is a flowchart of the multiphysics progressive coupling energy regulation steps. Figure 4A flowchart of the feedforward correction steps for online modality recognition; Figure 5 A parameter correlation diagram for the laser output energy control formula; Figure 6 A flowchart illustrating the cyclic execution logic of the entire laser processing process. Detailed Implementation

[0017] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0018] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0019] Technical problems with existing technologies: Existing laser processing energy control schemes assume that the energy emitted by the laser source is equal to the actual energy received by the material, ignoring the dynamic loss of the entire laser process. This makes it impossible to cope with the energy reference deviation caused by factors such as thermal distortion and aging of optical components, resulting in loss of control over energy control precision and causing processing defects.

[0020] Based on this, please refer to Figures 1-6 This embodiment provides an adaptive energy control method for laser processing, including a full-link dynamic loss reverse calibration step, a multi-physics field progressive coupling energy control step, and an online mode recognition feedforward correction step. The full-link dynamic loss reverse calibration step calculates a full-link loss correction coefficient based on the laser source output parameters and the processing end energy feedback parameters. The multi-physics field progressive coupling energy control step calculates an energy demand coefficient based on the full-link loss correction coefficient. The online mode recognition feedforward correction step calculates a feedforward correction coefficient based on the real-time signal of the processing process. The full-link loss correction coefficient, the energy demand coefficient, and the feedforward correction coefficient are used together to control the laser output energy.

[0021] The hardware system supporting this method includes a laser source module, a power monitoring module, an optical element temperature acquisition module, a spectrum acquisition module, a high-speed vision acquisition module, an acoustic signal acquisition module, a processing module, and a drive control module. The laser source module uses a fiber laser with an output wavelength of 1064nm, a rated output power range of 0 to 2000W, a power adjustment resolution of 1W, and supports both continuous wave and pulsed wave output modes. The laser source module has a built-in power monitoring unit that can acquire real-time output power data from the laser resonator at a frequency of 10kHz. The acquired data is transmitted to the processing module via an Ethernet bus. The optical element temperature acquisition module uses a surface-mount platinum resistance temperature sensor (PT1000 model), with a temperature range of -40℃ to 200℃, a temperature accuracy of ±0.1℃, and a sampling frequency of 10Hz. The sensor is mounted on the surface of the core optical elements in the laser transmission link to acquire real-time temperature data for each optical element. The acquired data is transmitted to the processing module via an RS485 bus. The spectral acquisition module employs a high-speed fiber optic spectrometer with a spectral response range of 200nm to 1100nm, a spectral resolution of 0.1nm, and a maximum sampling frequency of 100kHz. The spectrometer's acquisition probe is fixed to the side of the processing area via a bracket, 30cm away from the laser processing target. The angle between the probe's optical axis and the laser incident optical axis is 45 degrees. This probe is used to acquire the emission spectral data of the plasma plume during processing. The acquired data is transmitted to the processing module via a USB 3.0 bus. The high-speed vision acquisition module uses a high-speed industrial camera with a resolution of 1280×1024 and a maximum sampling frame rate of 2000fps. The camera is equipped with an 850nm bandpass filter and an auxiliary illumination source. The wavelength of the auxiliary illumination source matches the center wavelength of the filter. The camera is fixed directly above the processing area via a bracket, with its optical axis coinciding with the laser incident optical axis. This camera is used to acquire real-time morphological images of the molten pool during processing. The acquired data is transmitted to the processing module via a CameraLink bus. The acoustic signal acquisition module employs a high-frequency acoustic emission sensor with a response frequency range of 100kHz to 1MHz and a sampling frequency of 5MHz. The sensor is fixed to the surface of the workpiece via a magnetic base and is used to acquire acoustic emission signals generated by the plasma plume during processing. The acquired data is transmitted to the processing module via a coaxial cable. The processing module utilizes an industrial control computer equipped with an Intel Core i9-14900K processor and an NVIDIA RTX 4090 graphics card. The processor handles data scheduling, logical operations, and process control, while the graphics card handles image data processing, spectral data analysis, and parallel algorithm calculations. All input and output data of the processing module are scheduled through a real-time operating system, ensuring a data processing latency of less than 1ms, meeting the requirements for real-time control in laser processing.The drive control module uses an analog output board with a 16-bit digital-to-analog conversion resolution and an output voltage range of 0 to 10V. The output voltage is linearly related to the output power of the laser source. The laser output energy control signal output by the processing module is converted into an analog voltage signal by the board and transmitted to the laser source module to realize real-time control of the laser output power.

[0022] Before processing begins, the system completes initialization operations, including hardware self-test, parameter calibration, and reference data acquisition. During the hardware self-test, the processing module sends self-test commands to all hardware modules to confirm the communication and operational status of each module. If any module malfunctions, the system immediately stops the initialization process and issues an alarm. During parameter calibration, the system completes laser source power calibration, optical element initial parameter calibration, and material processing reference parameter calibration. During laser source power calibration, the system controls the laser source to output different power levels, collects the actual output power using an external standard power meter, establishes a correspondence table between the laser source control signal and the actual output power, and stores it in the processing module's storage unit. During optical element initial parameter calibration, the system collects the initial temperature data of each optical element at room temperature, as well as the change in absorption coefficient per unit temperature rise for each optical element. The change in absorption coefficient per unit temperature rise is obtained through pre-calibration of the optical element's material, coating characteristics, and laser wavelength. The calibration data is stored in the processing module's storage unit. During the material processing reference parameter calibration process, the system sets the target processing energy, standard material vaporization threshold, and standard molten pool viscosity parameters based on the material, thickness, and processing requirements of the material to be processed. These reference parameters are stored in the storage unit of the computing module as the basis for subsequent energy control. During the reference data acquisition process, the system collects ambient light data, ambient electromagnetic interference data, and the initial temperature data of the workpiece to be processed before processing begins, which serve as reference data for subsequent interference suppression and signal correction.

[0023] Technical problems with existing technologies: Existing end-to-end loss compensation schemes can only compensate based on fixed formulas according to the temperature of optical components, and cannot distinguish between transmission link loss and processing loss, resulting in insufficient compensation accuracy and inability to cope with loss changes caused by non-temperature factors.

[0024] Based on this, the end-to-end dynamic loss reverse calibration step includes a laser source power acquisition step, an optical component status acquisition step, a processing end energy calculation step, and a loss correction coefficient generation step. The laser source power acquisition step acquires real-time output power data of the laser source. The optical component status acquisition step acquires real-time temperature data of the optical component. The processing end energy calculation step calculates real-time incident power data of the processing end based on the spectral signal of the processing process. The loss correction coefficient generation step generates the end-to-end loss correction coefficient based on the real-time output power data, the real-time temperature data, the real-time incident power data, and the energy utilization rate data.

[0025] The laser source power acquisition step uses the power monitoring unit built into the laser source module to acquire the raw output power data of the laser resonator in real time. The acquisition frequency is consistent with the power adjustment frequency of the laser source, which is 10kHz. Each acquired power data is processed by moving average filtering, and the length of the moving average window is 5 sampling points. The filtered power data is used as the real-time output power data of the laser source and transmitted to the loss correction coefficient generation step. The optical component status acquisition step uses temperature sensors attached to the surface of each optical component to acquire the real-time temperature data of each optical component in real time. The acquisition frequency is 10Hz. The difference between the acquired temperature data and the initial temperature data calibrated during initialization is calculated to obtain the real-time temperature rise data of each optical component. The real-time temperature rise data is transmitted to the loss correction coefficient generation step. The energy calculation step at the processing end involves acquiring real-time emission spectral data of the plasma plume during processing using a high-speed fiber optic spectrometer. The acquired spectral data undergoes preprocessing, including dark current correction, spectral smoothing, and characteristic peak extraction. This extracts the intensity of the characteristic peaks corresponding to the laser wavelength, as well as the intensity and half-width at half-maximum (WHM) of the plasma characteristic emission spectral lines. Based on a pre-calibrated relationship between the intensity of the spectral characteristic peaks and the incident laser power at the processing end, the actual real-time incident power data reaching the processing end is calculated. This real-time incident power data is then transmitted to the loss correction coefficient generation step. The loss correction coefficient generation step receives the real-time output power data, the real-time temperature rise data, and the real-time incident power data. Simultaneously, it receives energy utilization rate data calculated in real-time during processing. The energy utilization rate data is the ratio of the energy actually absorbed by the material to the incident energy at the processing end, calculated using molten pool morphology data and material thermophysical parameters. The loss correction coefficient generation step calculates the full-link dynamic loss correction coefficient using the full-link dynamic loss correction coefficient formula, which is expressed as: The core theoretical basis of this formula is the law of conservation of energy. The total energy of the laser in the transmission link equals the emitted light energy minus the absorption loss, coupling loss, and transmission loss of all optical components. All the addition and subtraction terms in the formula have completely consistent dimensions, and the multiplication and division terms have matched dimensions, perfectly conforming to the principle of dimensional homogeneity. In the formula, This is the end-to-end loss correction coefficient, a dimensionless coefficient. Its physical meaning is the ratio of the effective energy actually reaching the processing end to the energy emitted by the laser source. The value ranges from 0 to 1. The closer the value is to 1, the lower the total loss of the transmission link and the higher the energy transmission efficiency. The real-time incident power at the processing end, measured in W, is calculated from the spectral characteristics collected by a spectral sensor. It represents the actual laser power reaching the processing area. This parameter directly reflects the energy level of the laser acting on the processing target after passing through the complete transmission link. Unlike existing technologies that only use the laser source output power as the calculation benchmark, this eliminates the benchmark deviation caused by transmission link loss from the root. The real-time output power of the laser source, measured in W, is acquired through the power monitoring module inside the laser source. It represents the original power output of the laser resonant cavity and serves as the initial energy reference for the entire energy control system. The change in absorption coefficient per unit temperature rise of the nth optical element is expressed in K⁻¹, representing the change in the absorption coefficient of the optical element for laser energy per 1K increase in temperature. This parameter is obtained by pre-calibrating the material parameters and coating characteristics of the optical element. Optical elements with different materials and coatings have significantly different absorption characteristics for laser energy as a function of temperature. Pre-calibration can ensure the accuracy of this parameter. The real-time temperature rise of the nth optical element is expressed in K, representing the difference between the real-time temperature of the optical element and its initial ambient temperature. It is obtained in real time by a temperature sensor on the surface of the optical element and directly reflects the degree of heat accumulation caused by the long-term operation of the optical element. The coupling efficiency of the focusing optical system is a dimensionless coefficient representing the proportion of laser energy focused onto the processing target after passing through the focusing optical system. This coefficient is pre-calibrated using the design parameters of the optical system and reflects its energy focusing capability, influenced by the assembly precision of optical components and the cleanliness of the lenses. In the formula, the summation term calculates the total absorption loss ratio of all optical components due to temperature rise, multiplying it by the ratio of output power to processing power and the coupling efficiency to obtain the end-to-end loss correction coefficient. This coefficient reflects the total loss ratio of the entire laser transmission chain in real time, providing an accurate benchmark for energy control. The core logic that distinguishes this formula from existing technologies lies in its simultaneous integration of measured power data from the laser source and processing end, thermal loss data due to temperature rise of optical components, and coupling efficiency data of the optical system. Through cross-validation of multi-source data, it achieves accurate calculation of the end-to-end loss, unlike existing technologies that rely solely on fixed open-loop compensation based on temperature. This covers all loss changes caused by non-temperature factors such as optical component aging, contamination, and assembly deviations, fundamentally ensuring the accuracy of the energy control benchmark.

[0026] The update cycle of the loss correction coefficient generation step matches the change cycle of the optical element's thermal distortion, with an update cycle of 100ms. The calculation and update of the end-to-end loss correction coefficient is completed every 100ms. The updated end-to-end loss correction coefficient is transmitted in real time to the multiphysics progressive coupling energy control step as the benchmark for calculating the energy demand coefficient. Simultaneously, the loss correction coefficient generation step stores the calculated end-to-end loss correction coefficient in the storage unit of the computational processing module, forming a loss change trend curve for subsequent assessment and early warning of the optical element's aging status.

[0027] Technical problems with existing technologies: Existing laser processing energy control schemes use plasma absorption, molten pool state, and material vaporization characteristics as parallel and independent parameters for compensation, which easily leads to problems of repeated compensation or insufficient compensation, resulting in energy control logic conflicts and insufficient accuracy.

[0028] Based on this, the multiphysics progressive coupling energy control step includes a plasma feature acquisition step, a molten pool morphology acquisition step, a plasma energy transmittance calculation step, a molten pool viscosity correction coefficient calculation step, and a vaporization threshold calibration coefficient calculation step. The plasma feature acquisition step acquires real-time spectral feature data of the plasma plume during processing. The molten pool morphology acquisition step acquires real-time morphology data of the molten pool during processing. The plasma energy transmittance calculation step calculates the plasma energy transmittance based on the end-to-end loss correction coefficient and the real-time spectral feature data. The molten pool viscosity correction coefficient calculation step calculates the molten pool viscosity correction coefficient based on the plasma energy transmittance. The vaporization threshold calibration coefficient calculation step calculates the vaporization threshold calibration coefficient based on the molten pool viscosity correction coefficient. The energy demand coefficient is generated based on the plasma energy transmittance, the molten pool viscosity correction coefficient, and the vaporization threshold calibration coefficient.

[0029] The plasma feature acquisition step involves acquiring real-time emission spectral data of the plasma plume during processing using a high-speed fiber optic spectrometer. After preprocessing the spectral data, the intensity, full width at half maximum (FWHM), and center wavelength shift of the plasma characteristic spectral lines are extracted. The electron temperature and electron density of the plasma are calculated using the Saha-Boltzmann equation. These electron temperatures and densities are used to calculate the plasma absorption coefficient for laser energy. The absorption coefficient, together with the length of the plasma plume, determines the proportion of incident laser energy absorbed by the plasma. The molten pool morphology acquisition step involves acquiring real-time morphological images of the molten pool during processing using a high-speed industrial camera. The acquired images are preprocessed, including image denoising, thresholding, and edge extraction, to extract the contour region of the molten pool. The area, perimeter, major axis length, minor axis length, and flow velocity of the molten pool are calculated. The flow velocity is calculated using the displacement of feature points within the molten pool in two adjacent frames and the frame interval. These feature parameters are used to calculate the dynamic viscosity of the molten pool. The plasma energy transmittance calculation step receives the end-to-end loss correction coefficient, combines it with the real-time output power data of the laser source, calculates the effective incident energy after end-to-end loss compensation, and then combines it with the plasma absorption coefficient to calculate the proportion of energy that actually penetrates the plasma plume and reaches the material surface. This energy proportion is the plasma energy transmittance, which is a dimensionless coefficient ranging from 0 to 1. A larger value indicates less absorption of laser energy by the plasma and a higher proportion of energy penetrating to the material surface. The molten pool viscosity correction coefficient calculation step receives the plasma energy transmittance, combines it with the thermophysical parameters of the material, calculates the real-time temperature field distribution of the molten pool, and then combines it with the molten pool flow characteristic parameters obtained from the molten pool morphology acquisition step to calculate the dynamic viscosity of the molten pool. The ratio of the dynamic viscosity to the standard viscosity is the molten pool viscosity correction coefficient, which is also a dimensionless coefficient. A larger value indicates a higher molten pool viscosity, lower heat conduction efficiency, and a need for higher input energy to achieve the target processing effect. The vaporization threshold calibration coefficient calculation step receives the molten pool viscosity correction coefficient and, combined with the molten pool's thermal conductivity, calculates the energy threshold required for the material to actually reach the vaporization state. The ratio of the actual energy threshold to the standard vaporization threshold is the vaporization threshold calibration coefficient. This calibration coefficient is a dimensionless coefficient; a larger value indicates higher energy required for the material to vaporize, and thus more additional laser input energy is needed. The energy demand coefficient is calculated using the multiphysics progressive coupling energy demand coefficient formula, which is expressed as: The core theoretical basis of this formula is the heat conduction theory of laser processing. The energy interaction process in laser processing is a strictly progressive physical process. The incident laser first interacts with the plasma plume generated by the vaporization of the material. Some energy is absorbed by the plasma, and the remaining energy penetrates the plasma and acts on the material surface, causing the material to melt and form a molten pool. The dynamic viscosity of the molten pool directly determines the heat conduction efficiency inside the molten pool, and the heat conduction efficiency ultimately determines the actual energy required for the material to reach the vaporization threshold. The formula completely follows the progressive logic of this physical process. All parameters are dimensionless coefficients, and the addition, subtraction, multiplication, and division operations completely conform to the principle of dimensional homogeneity. The final output energy demand coefficient is also a dimensionless coefficient, without introducing any dimensional deviation. In the formula, The energy demand coefficient is a dimensionless coefficient. Physically, it represents the ratio of the actual input energy required to achieve the target processing effect to the standard energy. The larger the value is than 1, the more laser energy actually needs to be supplemented. Plasma energy transmittance is a dimensionless coefficient representing the proportion of laser energy that penetrates the plasma plume and reaches the material surface. It is calculated from the electron density and temperature of the plasma. The smaller the value of this parameter, the more laser energy is absorbed by the plasma. In order to ensure that the energy reaching the material surface meets the processing requirements, more laser energy needs to be supplemented. Therefore, it is used as a denominator in the formula, and the logic of energy compensation is realized through reciprocal operation. This is the molten pool viscosity correction coefficient, a dimensionless coefficient that represents the correction value of thermal conductivity corresponding to the dynamic viscosity of the molten pool. The higher the viscosity of the molten pool, the weaker the convective heat transfer inside the molten pool, the lower the thermal conductivity, the lower the temperature reached by the material under the same input energy, and the higher the input energy required. Therefore, this coefficient is positively correlated with the energy demand coefficient and is used as a product term in the formula to achieve energy compensation for changes in the state of the molten pool. The vaporization threshold calibration coefficient is a dimensionless coefficient representing the ratio of the actual vaporization threshold of the material to the standard threshold. The lower the thermal conductivity of the molten pool, the higher the energy threshold required for the material to transition from a molten state to a vaporized state, and the more laser input energy is needed. Therefore, this coefficient is positively correlated with the energy demand coefficient and is used as a product term in the formula to compensate for changes in the material's vaporization characteristics. The core logic that distinguishes this formula from existing technologies lies in its rigorously progressive multi-physics coupling calculation logic. The output of the previous parameter serves as the sole input benchmark for the next parameter. Plasma energy transmittance determines the thermal input benchmark of the molten pool, the molten pool viscosity correction coefficient determines the thermal conductivity, and ultimately, the vaporization threshold calibration coefficient is determined. These three parameters form a chain-like progressive relationship, rather than the parallel independent parameter calculations of existing technologies. This fundamentally avoids the problems of redundant compensation, insufficient compensation, and logical conflicts caused by parallel multi-parameter control. Furthermore, through the coupled calculation of these three parameters, the entire physical process of laser-material interaction is fully covered, achieving accurate calculation of energy requirements at the processing end and providing an accurate processing end demand benchmark for subsequent energy control.

[0030] The plasma energy transmittance calculation, molten pool viscosity correction coefficient calculation, and vaporization threshold calibration coefficient calculation steps are executed sequentially in a fixed order. The output of the plasma energy transmittance calculation step serves as the input benchmark for the molten pool viscosity correction coefficient calculation step, and vice versa. This ensures logical consistency in parameter calculations and avoids logical conflicts and redundant compensation issues caused by parallel calculations of multiple parameters. The calculated energy demand coefficient is transmitted in real time to the laser output energy control stage, serving as one of the core bases for laser output energy calculation.

[0031] Technical problems with existing technologies: Existing laser processing energy control schemes are all post-feedback closed loops. Under transient conditions such as variable cross-section processing and sudden gap changes, there is a serious response lag problem, which cannot respond to transient disturbances in time, resulting in processing defects.

[0032] Based on this, the online modal recognition feedforward correction step includes a multi-source signal acquisition step during the processing, an adaptive modal decomposition step, a disturbance feature identification step, and a feedforward correction coefficient generation step. The multi-source signal acquisition step during the processing acquires real-time light intensity signals, acoustic signals, and molten pool morphology signals during the processing. The adaptive modal decomposition step performs modal decomposition on the real-time light intensity signals, acoustic signals, and molten pool morphology signals to obtain characteristic modes. The disturbance feature identification step identifies the type and intensity of transient disturbances based on the characteristic modes. The feedforward correction coefficient generation step generates feedforward correction coefficients based on the type and intensity of the transient disturbances.

[0033] The multi-source signal acquisition step in the processing involves synchronously acquiring reflected light intensity signals, plasma acoustic emission signals, and molten pool morphology change signals through corresponding high-speed sensors. The acquisition of these three signals is synchronized via a hardware trigger signal at a frequency of 10kHz, ensuring that the time synchronization error of the three signals is less than 1μs, thus avoiding mode decomposition errors caused by signal asynchrony. The reflected light intensity signal is acquired through the backlight monitoring unit built into the laser source, with a sampling frequency of 10kHz, reflecting changes in laser energy reflection loss during processing. The plasma acoustic emission signal is acquired through a high-frequency acoustic emission sensor, with a sampling frequency of 5MHz, reflecting dynamic changes in the plasma plume and stress changes within the workpiece. The molten pool morphology change signal is calculated from a sequence of molten pool images acquired by a high-speed industrial camera, specifically the rate of change of molten pool area between adjacent frames, with a sampling frequency of 2000fps, reflecting changes in the dynamic stability of the molten pool. The adaptive mode decomposition (EMD) step adaptively decomposes the synchronously acquired multi-source signals using a variational mode decomposition algorithm. This algorithm breaks down the mixed multi-source signals into characteristic modes corresponding to different physical processes. The decomposition process does not require a pre-established feature library and can be completed online in real time. During the decomposition, an iterative optimization algorithm determines the number of modes and their center frequencies, ensuring that each characteristic mode corresponds to a single physical change process. Specifically, the adaptive mode decomposition step decomposes the acquired real-time optical intensity signal, acoustic signal, and molten pool morphology signal into low-frequency, mid-frequency, and high-frequency modes. The low-frequency modes correspond to the slow-change process of thermal accumulation during processing, the mid-frequency modes correspond to the steady-state flow process of the molten pool, and the high-frequency modes correspond to the transient disturbance process during processing. The disturbance feature identification step extracts features from the decomposed high-frequency modes, calculates the amplitude, frequency, and energy proportion of the high-frequency modes, and identifies the type and intensity of transient disturbances using a pre-established disturbance feature library. Disturbance types include abrupt changes in workpiece gap, changes in cross-sectional thickness, inhomogeneous workpiece material, and violent fluctuations in the plasma plume. Disturbance intensity is quantified by the energy proportion and amplitude of the high-frequency modes, categorized into three levels: mild, moderate, and severe. The feedforward correction coefficient generation step generates corresponding feedforward correction coefficients based on the identified transient disturbance type and intensity. These feedforward correction coefficients are dimensionless, ranging from -0.3 to 0.3. Different correction strategies are employed for different types of disturbances. For increased energy demand caused by abrupt changes in workpiece gap and cross-sectional thickness, a positive feedforward correction coefficient is generated to increase laser output energy in advance. For increased energy absorption caused by violent fluctuations in the plasma plume, a corresponding positive correction coefficient is generated to compensate for the energy absorbed by the plasma. For the risk of molten pool collapse, a negative feedforward correction coefficient is generated to appropriately reduce laser output energy and avoid processing defects. The generated feedforward correction coefficients are transmitted in real time to the laser output energy control stage, serving as the basis for feedforward compensation in the laser output energy calculation.

[0034] Technical problems with existing technologies: Existing laser energy control schemes do not clearly define the core calculation formula for laser output energy, and cannot accurately quantify the impact of different correction coefficients on output energy, resulting in insufficient energy control precision. Furthermore, the technical solutions are not fully disclosed, making it impossible for technical personnel in the relevant field to implement them accurately.

[0035] Based on this, the laser output energy is calculated using the laser output energy adaptive control formula, which is expressed as follows: ; The core theoretical basis of this formula is adaptive closed-loop control theory. Precise control of laser output energy requires simultaneously achieving three core objectives: first, compensating for dynamic losses in the laser transmission link to ensure the accuracy of the energy reference; second, matching the energy demand changes caused by variations in material and molten pool states during processing to ensure the processing effect meets expectations; and third, proactively addressing transient disturbances during processing to avoid the response lag problem of traditional feedback closed loops. The formula fully follows the core logic of adaptive closed-loop control, integrating the three core objectives into a single calculation formula to form a complete three-level coordinated control logic. The dimensions on both sides of the equal sign are completely consistent, the dimensions of addition and subtraction operations are unified, and the combination of dimensions in multiplication and division operations fully conforms to physical laws, strictly adhering to the principle of dimensional homogeneity. In the formula... Let J be the laser energy that the laser source needs to output. J is the final output of the formula and is directly used to control the output parameters of the laser source, thus determining the actual energy input for laser processing. The target processing energy, measured in J, is the standard energy that the material surface needs to receive to achieve the target processing effect. It is preset by the processing requirements and serves as the target benchmark for the entire energy control system. It can be flexibly adjusted according to different processing processes, material types, and processing requirements. The energy demand coefficient is a dimensionless coefficient obtained by progressive coupling calculation of multiple physics fields. It represents the correction ratio of the energy demand in the processing process and fully reflects the influence of plasma absorption, changes in the molten pool state, and changes in material vaporization characteristics on the energy demand. It is a precise quantitative result of the energy demand at the processing end. This is the end-to-end loss correction coefficient, a dimensionless coefficient obtained from the reverse calibration of the end-to-end dynamic loss. It represents the loss ratio of the entire laser transmission link and reflects the energy transmission efficiency of the laser from the resonant cavity to the processing target. Therefore, it is used as a denominator in the formula and the energy loss of the transmission link is compensated through division to ensure that the energy reaching the processing end meets the target requirements. The feedforward correction coefficient is a dimensionless coefficient obtained from online mode identification. It represents the energy correction ratio required for transient disturbances and is used to proactively address energy demand changes in transient conditions, solving the response lag problem of traditional feedback closed loops. The formula uses... The formula employs a computational form to achieve feedforward compensation for transient disturbances. Positive feedforward correction coefficients increase laser output energy, while negative ones decrease it, perfectly matching the control requirements of different types of disturbances. The core logic that distinguishes this formula from existing technologies lies in integrating correction coefficients from three core dimensions—end-to-end loss calibration, multi-physics coupling requirement calculation, and transient disturbance feedforward—into a single closed-loop control formula. This forms a complete, progressively linked closed loop. The three correction coefficients correspond to three different physical processes: laser transmission, material interaction, and transient disturbance, with clear division of labor and logical coherence. This achieves precise energy control across all scenarios while avoiding logical conflicts in multi-parameter control, fundamentally solving the core bottleneck of insufficient energy control precision in existing technologies. Furthermore, the formula's structural design fully conforms to the core logic of adaptive control theory. Feedforward control addresses the response lag problem of transient disturbances, while feedback closed-loop control solves the dynamic compensation problem of slowly varying parameters, achieving a deep integration of feedforward and feedback, ensuring the energy control precision and response speed throughout the laser processing process.

[0036] The calculated laser output energy data is converted into corresponding control signals by the processing module. These control signals are then converted into analog voltage signals by a digital-to-analog converter and transmitted to the laser source module. Based on the received analog voltage signals, the laser source module adjusts its output power in real time, achieving precise control of the laser output energy. The laser output energy adjustment frequency is 10kHz, consistent with the laser source's power adjustment frequency, ensuring that each power adjustment is based on the latest correction coefficient calculation results, thus achieving real-time closed-loop control throughout the entire processing process.

[0037] Technical problems with existing technologies: Existing optical component status acquisition schemes do not clearly define the range of optical components to be acquired, resulting in incomplete coverage of loss calculations and inability to accurately calculate the total loss of the entire link, thus affecting compensation accuracy.

[0038] Based on this, the optical components include a laser resonator output mirror, a beam expander, a galvanometer, and a focusing mirror. The optical component status acquisition step collects real-time temperature data of the laser resonator output mirror, the beam expander, the galvanometer, and the focusing mirror. These four optical components are the core components in the laser transmission link and generate the main energy loss. After the laser is output from the resonator, it passes sequentially through the output mirror, the beam expander, the galvanometer, and the focusing mirror before finally focusing on the processing target. The loss of these four components accounts for more than 95% of the total loss of the entire transmission link. Collecting the temperature data of these four components can completely cover the main sources of loss in the laser transmission link, ensuring the completeness and accuracy of the entire link loss calculation. Two temperature sensors are mounted on the surface of each optical component, one on the edge of the incident surface and the other on the edge of the exit surface, to avoid obstructing the laser transmission path. The average of the temperature data collected by the two sensors is taken as the real-time temperature data of the optical component, ensuring the accuracy of temperature acquisition.

[0039] Technical problems with existing technologies: Existing energy control schemes only perform parameter calibration once before processing, and cannot continuously track parameter changes during processing, resulting in a continuous decline in energy control accuracy during long-term processing.

[0040] Based on this, the full-link dynamic loss reverse calibration step, the multi-physics progressive coupling energy control step, and the online modal recognition feedforward correction step are continuously and cyclically executed throughout the laser processing. During processing, the system updates the full-link loss correction coefficient once every 100ms (large cycle); calculates the multi-physics energy demand coefficient and updates the feedforward correction coefficient once every 1ms (small cycle); and calculates and controls the laser output energy once every 100μs (smallest cycle). The calculation processes of different cycles are independent of each other, and the data is exchanged in real time through shared memory, ensuring that high-priority control commands can be executed in real time, and that low-priority parameter updates do not affect the response speed of real-time control. Throughout the processing, the three core steps are continuously and cyclically executed, tracking changes in laser transmission link loss, changes in multi-physics state during processing, and changes in transient disturbances in real time, continuously updating the corresponding correction coefficients, achieving precise control of laser output energy at all times and under all working conditions, ensuring that the energy control accuracy remains within a stable range during long-term continuous processing, and that the accuracy does not decrease due to the extension of processing time.

[0041] The implementation process of this solution is fully described below through a specific embodiment. The application scenario of this embodiment is laser welding of stainless steel plates. The workpiece to be processed is a 304 stainless steel plate with a thickness of 2mm. The welding process is butt welding, the welding speed is 50mm / s, and the target weld penetration is 1.5mm. Before processing begins, the system completes initialization operations, all hardware self-tests are normal, and the laser source power calibration is completed. The calibrated laser source output power range is 0 to 1500W, and the linear correspondence between the control signal and output power is 1V to 150W. The initial parameter calibration of the optical components is completed. The initial temperature of the four core optical components is 25℃, and the changes in absorption coefficient per unit temperature rise are 0.0002K⁻¹, 0.00015K⁻¹, 0.0001K⁻¹, and 0.00025K⁻¹, respectively. The material processing reference parameters are calibrated, with the target processing energy set to 120J and the standard material vaporization threshold set to 1.5×10⁻¹. 6 The standard molten pool viscosity is 0.005 Pa·s. Baseline data acquisition was completed at an ambient temperature of 25℃ and an ambient light intensity of 500 lux, with no significant electromagnetic interference. During processing, the system continuously executes three core steps in a loop. The update cycle of the end-to-end loss correction coefficient is 100ms. During processing, the temperature of the optical components gradually increases from 25℃ to 45℃, and the end-to-end loss correction coefficient gradually changes from the initial 0.92 to 0.85. The system compensates for the changes in transmission link loss in real time. The multi-physics energy demand coefficient is updated in real time according to the plasma state and molten pool state to ensure matching of energy demand at the processing end. The online modal recognition step monitors transient disturbances during processing in real time. When a sudden change in workpiece gap of 0.1mm occurs, the system identifies a severe disturbance, generates a feedforward correction coefficient of 0.2, and increases the laser output energy in advance to avoid insufficient penetration. After processing, the welded joint was inspected. The weld penetration depth was stable between 1.45mm and 1.55mm. The weld was uniformly formed and free from defects such as porosity, cracks, and incomplete penetration, achieving high-precision laser welding processing.

[0042] This solution can also employ several alternative implementation methods. The adaptive mode decomposition step can use an ensemble empirical mode decomposition algorithm instead of a variational mode decomposition algorithm, selecting the optimal decomposition algorithm for different signal characteristics to ensure the accuracy of mode decomposition. In the calculation of the end-link loss correction coefficient, a correction term for the degree of optical component contamination can be added. By periodically collecting surface images of optical components, the degree of contamination can be identified, and additional compensation can be applied to the loss correction coefficient to further improve the accuracy of loss calculation. In the calculation of molten pool viscosity, a correction term for the real-time workpiece temperature can be added. By collecting real-time temperature data of the workpiece using an infrared thermometer, the calculated molten pool viscosity can be corrected, improving the accuracy of energy demand coefficient calculation.

[0043] Example 1 This embodiment applies to the laser cutting of aluminum alloy sheets. The workpiece to be processed is an aluminum alloy sheet, and the processing technology is laser cutting. Before processing begins, the system completes initialization operations, all hardware devices perform self-checks normally, the laser source power is calibrated, the initial parameters of the optical components are calibrated, the initial temperature data of each core optical component is collected and calibrated in relation to the change in absorption coefficient per unit temperature rise, the material processing reference parameters are calibrated, the target processing energy is set according to the laser cutting process requirements of the aluminum alloy sheet, the standard material vaporization threshold and standard molten pool viscosity are calibrated, reference data is collected, and the ambient temperature and ambient light intensity before processing are recorded to confirm that there is no significant electromagnetic interference.

[0044] During processing, the system continuously and iteratively executes the full-link dynamic loss reverse calibration step, the multi-physics progressive coupling energy control step, and the online modal recognition feedforward correction step. The full-link loss correction coefficient is updated in real time according to a set update cycle. As the optical components experience temperature increases due to continuous operation, the full-link loss correction coefficient is dynamically adjusted according to the changes in the optical component's state, and the system compensates for the loss changes in the laser transmission link in real time. The multi-physics energy demand coefficient is continuously updated based on the real-time spectral characteristics of the plasma plume and the real-time morphology data of the molten pool during processing, accurately matching the actual energy demand at the processing end during aluminum alloy laser cutting. The online modal recognition step monitors transient disturbances during processing in real time. When a transient disturbance with localized non-uniformity in the material is detected, the system accurately identifies the type and intensity of the disturbance, generates the corresponding feedforward correction coefficient, and adjusts the laser output energy in advance to avoid cutting effect deviations caused by material differences.

[0045] After processing, the cut surface of the aluminum alloy sheet is inspected. The cut surface is flat and smooth, without any processing defects such as slag or burrs. The cut is uniform and the cutting contour is highly consistent with the preset processing trajectory, thus realizing high-precision laser cutting of aluminum alloy sheets. Example 2

[0046] This embodiment applies to laser cladding of carbon steel substrates. The workpiece to be processed is a carbon steel substrate, and laser cladding is performed using cladding alloy powder. The processing technology is laser cladding. Before processing begins, the system completes initialization, all hardware self-tests are normal, laser source power calibration is completed, the correspondence between laser source control signals and actual output power is established, initial parameter calibration of optical components is completed, initial temperature and related characteristic calibration data of each core optical component are collected, material processing benchmark parameter calibration is completed, target processing energy is set according to the laser cladding process requirements of carbon steel substrates, standard material vaporization threshold and standard molten pool viscosity and other parameters adapted to the cladding process are calibrated, benchmark data is collected, environmental data before processing is recorded, and it is confirmed that there is no significant external interference.

[0047] During processing, the system continuously executes three core steps in a loop. The end-to-end loss correction coefficient is updated in real time according to the working state of the optical components, accurately compensating for the dynamic loss of the laser transmission link to ensure stable transmission of effective laser energy. The multi-physics progressive coupling energy control step continuously calculates and updates the energy demand coefficient based on the dynamic characteristics of the plasma plume and the real-time state of the molten pool during cladding, adapting to the actual energy requirements of molten pool formation and powder melting during laser cladding, ensuring the foundation for cladding layer formation. The online modal recognition feedforward correction step collects multi-source signals in real time during processing. When transient disturbances such as fluctuations in the cladding powder feed rate are detected, the system quickly identifies the type and intensity of the disturbance, generates corresponding feedforward correction coefficients, and adjusts the laser output energy in advance to compensate for energy demand changes caused by powder feed rate fluctuations, avoiding forming defects in the cladding layer.

[0048] After processing, the cladding layer of the carbon steel substrate is inspected. The cladding layer and the carbon steel substrate achieve metallurgical bonding. The bonding interface is tight and free from defects such as pores and cracks. The overall thickness of the cladding layer is uniform, the surface is flat, and there are no problems such as segregation or lack of fusion. The appearance and performance of the cladding layer meet the high-precision process requirements of laser cladding.

[0049] Unless otherwise specified, the equipment components involved in the above embodiments are all conventional equipment components, and the connection methods and control methods involved are all conventional connection methods and control methods unless otherwise specified.

[0050] The present invention has been described in detail above with reference to the embodiments. However, those skilled in the art will understand that, without departing from the spirit of the present invention, various specific parameters in the above embodiments can be changed to form multiple specific embodiments, all of which are common variations of the present invention, and will not be described in detail here.

Claims

1. An adaptive energy control method for laser processing, characterized in that, This includes a full-link dynamic loss reverse calibration step, a multi-physics progressive coupling energy regulation step, and an online mode identification feedforward correction step; The full-link dynamic loss reverse calibration step calculates the full-link loss correction coefficient based on the laser source output parameters and the processing end energy feedback parameters. The multi-physics progressive coupling energy control step calculates the energy demand coefficient based on the full-link loss correction coefficient. The online mode recognition feedforward correction step calculates the feedforward correction coefficient based on the real-time signal of the processing process. The full-link loss correction coefficient, the energy demand coefficient, and the feedforward correction coefficient are used together to control the laser output energy.

2. The adaptive energy control method for laser processing according to claim 1, characterized in that, The end-to-end dynamic loss reverse calibration step includes a laser source power acquisition step, an optical component status acquisition step, a processing end energy calculation step, and a loss correction coefficient generation step. The laser source power acquisition step acquires real-time output power data of the laser source, the optical component status acquisition step acquires real-time temperature data of the optical component, the processing end energy calculation step calculates real-time incident power data of the processing end based on the spectral signal of the processing process, and the loss correction coefficient generation step generates the end-to-end loss correction coefficient based on the real-time output power data, the real-time temperature data, the real-time incident power data, and the energy utilization rate data.

3. The adaptive energy control method for laser processing according to claim 1, characterized in that, The multiphysics progressive coupling energy control step includes a plasma feature acquisition step, a molten pool morphology acquisition step, a plasma energy transmittance calculation step, a molten pool viscosity correction coefficient calculation step, and a vaporization threshold calibration coefficient calculation step. The plasma feature acquisition step acquires real-time spectral feature data of the plasma plume during processing. The molten pool morphology acquisition step acquires real-time morphology data of the molten pool during processing. The plasma energy transmittance calculation step calculates the plasma energy transmittance based on the end-to-end loss correction coefficient and the real-time spectral feature data. The molten pool viscosity correction coefficient calculation step calculates the molten pool viscosity correction coefficient based on the plasma energy transmittance. The vaporization threshold calibration coefficient calculation step calculates the vaporization threshold calibration coefficient based on the molten pool viscosity correction coefficient. The energy demand coefficient is generated based on the plasma energy transmittance, the molten pool viscosity correction coefficient, and the vaporization threshold calibration coefficient.

4. The adaptive energy control method for laser processing according to claim 1, characterized in that, The online modal recognition feedforward correction step includes a multi-source signal acquisition step during the processing, an adaptive modal decomposition step, a disturbance feature recognition step, and a feedforward correction coefficient generation step. The multi-source signal acquisition step of the processing process acquires real-time light intensity signals, acoustic signals, and molten pool morphology signals during the processing. The adaptive mode decomposition step performs mode decomposition on the real-time light intensity signals, acoustic signals, and molten pool morphology signals to obtain characteristic modes. The disturbance feature identification step identifies the type and intensity of transient disturbances based on the characteristic modes. The feedforward correction coefficient generation step generates feedforward correction coefficients based on the type and intensity of the transient disturbances.

5. The adaptive energy control method for laser processing according to claim 2, characterized in that, The update cycle of the loss correction coefficient generation step matches the change cycle of the thermal distortion of the optical element, and the end-to-end loss correction coefficient generated by the loss correction coefficient generation step is updated in real time to the energy reference parameters of the energy control module.

6. The adaptive energy control method for laser processing according to claim 3, characterized in that, The plasma energy transmittance calculation step, the molten pool viscosity correction coefficient calculation step, and the vaporization threshold calibration coefficient calculation step are executed in a fixed order. The output result of the plasma energy transmittance calculation step is used as the input reference for the molten pool viscosity correction coefficient calculation step, and the output result of the molten pool viscosity correction coefficient calculation step is used as the input reference for the vaporization threshold calibration coefficient calculation step.

7. The adaptive energy control method for laser processing according to claim 1, characterized in that, The laser output energy is calculated using an adaptive control formula for laser output energy, which is expressed as follows: ； Among them, the For laser output energy, the For the target processing energy, the The energy demand coefficient, the The end-to-end loss correction coefficient is the... This is the feedforward correction coefficient.

8. The adaptive energy control method for laser processing according to claim 4, characterized in that, The adaptive mode decomposition step decomposes the acquired real-time light intensity signal, acoustic signal, and molten pool morphology signal into low-frequency mode, mid-frequency mode, and high-frequency mode. The low-frequency mode corresponds to the slow thermal accumulation process in the processing, the mid-frequency mode corresponds to the steady-state flow process of the molten pool, and the high-frequency mode corresponds to the transient disturbance process in the processing. The disturbance feature identification step identifies the type and intensity of the transient disturbance based on the high-frequency mode.

9. The adaptive energy control method for laser processing according to claim 2, characterized in that, The optical components include a laser resonator output mirror, a beam expander, a galvanometer, and a focusing mirror. The optical component status acquisition step acquires real-time temperature data of the laser resonator output mirror, the beam expander, the galvanometer, and the focusing mirror.

10. The adaptive energy control method for laser processing according to claim 1, characterized in that, The full-link dynamic loss reverse calibration step, the multi-physics progressive coupling energy control step, and the online modal recognition feedforward correction step are continuously and cyclically executed throughout the entire laser processing process.

Images