Real-time feedback control system and method for laser spot morphology and energy distribution of LDED

Abstract

This invention discloses a real-time laser spot morphology and energy distribution feedback control system and method for LDED (Laser-to-Edge Molten Electrode), relating to the field of additive manufacturing technology. The system includes a laser unit, a multi-dimensional sensor, a real-time laser spot morphology control module, an energy distribution adjustment module, and a controller. The laser unit emits a laser beam; the multi-dimensional sensor acquires molten pool state parameters in real time; the real-time laser spot morphology control module regulates the spatial intensity distribution of the laser beam, dynamically reconstructing the spot morphology; the energy distribution adjustment module regulates the energy characteristics of the laser beam; and the controller incorporates a real-time control and optimization algorithm. This algorithm constructs an objective function based on multi-source feedback data and updates the power density distribution in real time through gradient descent. This invention can continuously reconstruct the spot morphology, spatially partition and adjust the power density, integrate multi-source sensor feedback, and achieve true closed-loop optimization control, improving the uniformity of the formed structure and the consistency of mechanical properties.

Description

Technical Field

[0001] This invention relates to the field of additive manufacturing technology, and in particular to a real-time spot morphology and energy distribution feedback control system and method for LDED. Background Technology

[0002] Laser-directed energy deposition (LDED) is an important metal additive manufacturing technology widely used in aerospace, energy equipment, and high-end mold manufacturing. LDED uses a high-energy laser beam to form a molten pool on the surface of a substrate or existing layer, while simultaneously feeding in metal powder or wire to achieve layer-by-layer deposition. However, current mainstream LDED systems still have certain limitations:

[0003] Existing optical systems primarily rely on fixed optical elements or limited combinations of switching. The spot morphology cannot be continuously varied during deposition; it can only be discretely switched between different optical components, making continuous reconstruction impossible. Traditional systems can only adjust the total laser power as a whole, unable to perform zonal control of the spatial power density distribution (PDD), and cannot achieve radial energy gradient adjustment or angular non-uniform enhancement. Furthermore, due to the highly nonlinear, multi-field coupled, and high-speed dynamic characteristics of the melt pool evolution process, existing systems struggle to achieve real-time fusion of multi-source sensor data and online optimization of control parameters, making closed-loop control difficult to implement in engineering. Therefore, existing technologies are mostly open-loop or semi-closed-loop control, unable to perform real-time objective function optimization and gradient update control based on multi-dimensional sensor data. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a real-time spot morphology and energy distribution feedback control system and method for LDED, which can continuously reconstruct the spot morphology, adjust the power density in spatial partitions, integrate feedback from multiple sources of sensors, and achieve true closed-loop optimization control, thereby improving the uniformity of the formed structure and the consistency of mechanical properties.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide a real-time spot morphology and energy distribution feedback control system for LDED, comprising:

[0007] The laser unit is used to emit a laser beam;

[0008] Multidimensional sensors are used to collect molten pool status parameters in real time;

[0009] The real-time beam spot morphology control module is used to control the spatial intensity distribution of the laser beam and dynamically reconstruct the beam spot morphology.

[0010] An energy distribution adjustment module is used to regulate the energy characteristics of the laser beam;

[0011] The controller incorporates a real-time control and optimization algorithm. This algorithm constructs an objective function based on multi-source feedback data and updates the power density distribution in real time through gradient descent.

[0012] As a further implementation, the real-time beam shape control module includes a beam shaping unit and a multi-level light field reconstruction unit. The beam shaping unit is used to adjust the beam intensity, and the multi-level light field reconstruction unit can form a double-ring beam, an inner high and outer low gradient beam, an edge-enhanced beam, an offset elliptical beam, or a single-sided enhanced beam.

[0013] As a further implementation, the beam shaping unit includes a digital intensity distribution control component and a physical morphology control component for the light field. The digital intensity distribution control component is used to control the spatial distribution of the laser beam's intensity, and the physical morphology control component for the light field is used to physically control the geometric shape, focusing characteristics, and wavefront shape of the laser beam.

[0014] The multi-level optical field reconstruction unit includes a composite diffractive optical element, a phase plate, and a ring phase modulation disk. The composite diffractive optical element is used for primary basic modulation, the phase plate is used for secondary precise shaping, and the ring phase modulation disk is used for tertiary dynamic adaptation.

[0015] As a further implementation, the energy distribution adjustment module controls the temperature gradient and cooling rate of the molten pool by adjusting core parameters;

[0016] The core parameters include pulse width, modulation frequency, duty cycle, and power density.

[0017] As a further implementation, the real-time control and optimization algorithm includes a multi-field coupled model numerical calculation algorithm, a state prediction and defect determination algorithm, a closed-loop objective function optimization algorithm, and a gradient descent iterative update algorithm.

[0018] Secondly, embodiments of the present invention also provide a real-time spot morphology and energy distribution feedback control method for LDED, based on the aforementioned control system, comprising:

[0019] Collect data from multiple sources;

[0020] The defect probability and structure prediction parameters are calculated based on multi-source data to obtain the molten pool state;

[0021] The corresponding spot control mode is matched according to the state of the molten pool; and the energy distribution, spot morphology parameters and pulse parameters are updated in real time.

[0022] As a further implementation, the step of matching the corresponding spot control mode according to the molten pool state includes:

[0023] When the molten pool is overheated, it switches to an annular spot; when the cooling is too fast, it switches to a center-enhanced spot; when there is a risk of cracking, it switches to a flat-top spot; and when there is a risk of porosity, it switches to a multi-peak compensation spot.

[0024] As a further implementation, the energy distribution update satisfies:

[0025] ;

[0026] in, This represents the power density distribution in the nth iteration. This represents the updated power density distribution, where η represents the learning rate. Represents the gradient of the objective function;

[0027] Gradient descent closed-loop control is formed.

[0028] As a further implementation, a gradient descent iterative rule is adopted to correct the power density distribution in real time based on the gradient of the objective function, thereby realizing multi-field coupled closed-loop optimization control.

[0029] As a further implementation, the light spot morphology parameters include radial distribution control parameters, angular non-uniformity control parameters, and multi-peak structure adjustment parameters, and the pulse parameters include pulse width, modulation frequency, and duty cycle.

[0030] The beneficial effects of this invention are as follows:

[0031] This invention constructs a fully closed-loop feedback control system for spot morphology and energy distribution during LDED deposition through the coordinated operation of a laser unit, multi-dimensional sensors, a real-time spot morphology control module, an energy distribution adjustment module, and a controller with built-in real-time control and optimization algorithms. This system enables dynamic and precise adjustment of the spot morphology based on the real-time state of the molten pool, adapting to the dynamic changes in the molten pool during deposition. It also achieves refined control of the laser beam energy characteristics, allowing for targeted optimization of the molten pool heat input. By achieving coordinated and precise control of the LDED deposition molten pool from both spot morphology and energy distribution dimensions, this invention effectively compensates for process deviations during deposition, suppresses abnormal molten pool states, reduces the risk of defect generation, and improves the forming quality and microstructure uniformity of the deposited part. Attached Figure Description

[0032] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0033] Figure 1 This is a block diagram of the control system structure according to one or more embodiments of the present invention;

[0034] Figure 2This is a schematic diagram of the light spot according to one or more embodiments of the present invention;

[0035] Figure 3 This is a schematic diagram of the multi-dimensional sensor composition according to one or more embodiments of the present invention;

[0036] Figure 4 This is a schematic diagram of the real-time beam morphology control module according to one or more embodiments of the present invention;

[0037] Figure 5 This is a flowchart of the control method according to one or more embodiments of the present invention. Detailed Implementation

[0038] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0039] Glossary: ​​LDED, Laser Directed Energy Deposition;

[0040] PDD stands for Power Density Distribution.

[0041] Example 1:

[0042] This embodiment provides a real-time spot morphology and energy distribution feedback control system for LDED, which can solve the problems of fixed spot morphology, non-regional controllable energy distribution, single feedback and no true closed-loop control in traditional LDED systems. It can realize real-time dynamic control of spot and energy, and improve the forming quality and microstructure uniformity of deposited parts.

[0043] like Figure 1 As shown, the control system in this embodiment includes a laser unit, a multi-dimensional sensor, a real-time spot morphology control module, an energy distribution adjustment module, and a controller, forming a fully closed-loop control system of "laser emission - morphology and energy control - molten pool action - state detection - algorithm decision - reverse control". Among them, the controller has a built-in real-time control and optimization algorithm, and the other modules are execution and sensing units that receive the control commands from the controller and feed back the real-time state of the molten pool. The modules work together to achieve precise control of the LDED deposition process.

[0044] Multidimensional sensors are deployed around the deposition area to collect the temperature field, geometry, defect signals and process status parameters of the molten pool in real time, and transmit the collected multi-source data to the controller. The controller is connected to the multidimensional sensors and the execution interface unit, respectively, to preprocess the collected multi-source data, extract features and determine the status, and generate corresponding control parameters based on the objective function and optimization algorithm.

[0045] The execution interface unit is connected to the controller and each execution module. It is used to convert the control parameters output by the algorithm unit into executable electrical signals or control commands, and send them to the real-time beam shape control module and the energy distribution adjustment module, respectively.

[0046] The real-time beam morphology control module is coupled to the optical path system of the laser unit. It modulates the spatial intensity distribution of the laser beam based on the received beam morphology parameters, enabling continuous dynamic reconstruction of the beam morphology. The energy distribution adjustment module is connected to the drive circuit of the laser unit. It adjusts the laser output power and pulse characteristics based on the received energy control parameters, achieving coordinated control of laser energy in the time and spatial domains. The modulated laser beam acts on the deposition area, causing changes in the molten pool state. A multi-dimensional sensor re-detects the modulated molten pool state and feeds it back to the controller.

[0047] In this embodiment, the light spot morphology parameters include light spot type, light spot size, etc., which are usually derived from a preset process database and are set in advance based on different materials, path characteristics, and forming targets; the energy control parameters include power, pulse frequency, duty cycle, etc.

[0048] Specifically, in this embodiment, the laser unit uses a fiber laser to emit a continuous laser beam, providing the basic laser energy for LDED deposition. For example... Figure 3 As shown, a multi-dimensional sensor array is deployed in the molten pool monitoring area around the LDED deposition head. It consists of an X-ray online imaging sensor, an acoustic emission array, a spectral sensor, a dual-wavelength infrared sensor, a laser scattering monitoring sensor, and a ranging sensor. Through this multi-dimensional sensor array, the molten pool temperature field, morphological features, internal structure, and dynamic response parameters are acquired, enabling a comprehensive characterization of the multi-physics coupling behavior of the LDED process. This overcomes the limitations of single sensor information and improves the accuracy and reliability of defect identification.

[0049] In this embodiment, the dual-wavelength infrared sensor and the spectral sensor are preferably arranged coaxially or nearly coaxially to obtain the temperature field and plasma spectral information of the molten pool; the laser scattering monitoring sensor and the ranging sensor are arranged to the side and front of the laser incident direction to detect the powder flow distribution and interlayer height changes; the X-ray online imaging sensor is arranged to the side or at an angle to the molten pool to achieve imaging detection of internal defects in the molten pool; the acoustic emission array is arranged on the substrate or deposition substrate to collect acoustic signals during the crack initiation process; all sensors are time-synchronized through a unified clock synchronization module, using a combination of time-division acquisition and parallel acquisition.

[0050] An online X-ray imaging sensor detects the formation of pores inside the molten pool, an acoustic emission array captures the acoustic signals of microcrack initiation, a spectral sensor identifies the concentration of evaporating elements and plasma spectral anomalies (overburning signals), a dual-wavelength infrared sensor detects the temperature gradient on the surface and deep layers of the molten pool, a laser scattering monitoring sensor acquires powder concentration matching data, and a ranging sensor collects interlayer height deviations. All detection data are transmitted to the controller in real time, providing multi-source feedback data for algorithm decision-making and solving the problem of single feedback signals.

[0051] like Figure 4 As shown, the real-time beam shape control module is used to adjust the beam shape of the laser beam in real time according to the molten pool state judgment result output by the controller, so as to realize the continuous reconstruction of the beam shape rather than discrete switching; the module includes a beam shaping unit and a multi-level optical field reconstruction unit, which work together to complete the dual control of light intensity distribution and beam physical shape.

[0052] Furthermore, the beam shaping unit includes a digital intensity distribution control component and a physical morphology control component for the optical field. The digital intensity distribution control component consists of a reflective spatial light modulator (SLM) and a digital micromirror array (DMD), which achieves precise control of the spatial distribution of laser beam intensity through programmable modulation, directly adjusting the radial, angular, and multi-peak structure parameters of the beam intensity distribution function. The physical morphology control component for the optical field consists of a variable-focus liquid lens and a deformable mirror, which physically controls the geometric shape, focusing characteristics, and wavefront shape of the laser beam, completing dynamic focal length adjustment and wavefront distortion correction, providing a stable physical optical field basis for digital intensity control.

[0053] The light spot intensity distribution function is expressed as:

[0054] ;

[0055] in, This represents the instantaneous light intensity of the light spot at the polar coordinates (r, θ). denoted as the standardized reference light intensity, F represents the light intensity distribution function, v represents the radial distribution control parameter, β represents the angular non-uniformity control parameter, and γ represents the multi-peak structure adjustment parameter.

[0056] In this embodiment, the above parameters can be continuously adjusted in real time during the deposition process rather than being switched discretely.

[0057] The multi-level optical field reconstruction unit comprises a composite diffractive optical element, an axisymmetric / asymmetric phase plate, and a rotatable ring phase modulation disk. These three components operate in a progressive hierarchy: primary basic modulation, secondary precise shaping, and tertiary dynamic adaptation. The composite diffractive optical element performs holistic basic modulation on the original Gaussian laser beam emitted by the laser unit, generating an optical field with basic characteristics such as gradient, ring shape, and multiple peaks. The axisymmetric / asymmetric phase plate performs precise shaping based on the basic optical field. The axisymmetric phase plate optimizes the uniformity of the light spot, while the asymmetric phase plate breaks the axisymmetric property, generating asymmetric light spots such as biased ellipses and unilateral enhancement. The rotatable ring phase modulation disk dynamically fine-tunes the shaped optical field by rotating to adjust the angular position and ring width of the phase modulation, achieving real-time adaptation of edge enhancement degree and unilateral enhancement angular position.

[0058] Depending on the state of the molten pool, various morphological light spots can be flexibly generated, such as double-ring light spots, inner high and outer low gradient light spots, edge-enhanced light spots, offset elliptical light spots, and unilateral enhanced light spots, to meet the light spot adaptation requirements of different molten pool states during the deposition process and effectively compensate for the directional heat accumulation caused by changes in scanning direction and gravity deposition deviation.

[0059] like Figure 2 As shown, the Gaussian beam corresponds to a spot shape where the center is brightest, gradually decreasing in intensity towards the edges; the cross-sectional image is a smooth Gaussian curve; the corresponding molten pool morphology is bowl-shaped or semi-circular, with the deepest center and shallower edges. The flat-top beam corresponds to a spot shape where the intensity is uniform in the central region, steeply decreasing towards the edges, and the cross-sectional image is rectangular; the corresponding molten pool morphology is a U-shaped or rectangular pool with a flat bottom and steep sidewalls, uniform melting depth, and neat edges. The ring beam corresponds to a spot shape where the central area is dark (low intensity), the outer ring is bright, and the cross-sectional image has two peaks (bimodal structure); the corresponding molten pool morphology is characterized by an indistinct central convexity or depression, a bimodal shape, deep melting at the edges, and the center may be unmelted or shallowly melted.

[0060] The beam pattern corresponding to a multi-peaked beam is characterized by an intensity distribution with multiple random or regular peaks (such as 3-5 bright spots), and the cross-sectional view shows multiple peaks. The corresponding molten pool morphology is characterized by multiple separated or connected molten pits on the surface of the molten pool, resembling "multiple humps" or "beads," and is uneven. The beam pattern corresponding to an asymmetric beam is characterized by an intensity distribution that is stretched or skewed along a certain direction, such as an ellipse, and the cross-sectional view is asymmetrical. The corresponding molten pool morphology is characterized by an asymmetric teardrop shape, with a steep leading edge and a trailing trailing edge, and uneven distribution of melt depth.

[0061] The energy distribution adjustment module is linked with the real-time spot morphology control module to regulate the energy characteristics of the laser beam. This module achieves precise control of the molten pool temperature gradient G and cooling rate R by dynamically adjusting four core parameters: pulse width τ, modulation frequency f, duty cycle D, and power density PDD.

[0062] Among them, the formation of different structures is related to the cooling rate R: G·R = constant; by adjusting the pulse width τ and the modulation frequency f, the cooling rate R can be controlled, thereby controlling the dendrite spacing, porosity, grain size, etc.

[0063] The controller in this embodiment is equipped with high-speed computing hardware of FPGA + high-speed DSP, and has built-in real-time control and optimization algorithms, serving as the core decision-making unit of the system. The algorithm includes a multi-field coupling model numerical calculation algorithm, a state prediction and defect determination algorithm, a closed-loop objective function optimization algorithm, and a gradient descent iterative update algorithm. Its working logic is as follows: it receives multi-source feedback data from multi-dimensional sensors, first calculates physical field parameters such as molten pool temperature gradient, cooling rate, and thermal stress by using the multi-field coupling model numerical calculation algorithm in combination with the heat conduction equation and thermal stress relationship model; then, it quantifies and calculates the defect probability by using the state prediction and defect determination algorithm.

[0064] Furthermore, multi-field coupling models include:

[0065] Heat conduction equation: ;

[0066] Where ρ represents the material density, and c represents the specific heat capacity. The value represents the rate of temperature change, and k represents thermal conductivity. Represents the temperature Laplace operator. This represents the space heat source term.

[0067] Stress relationship: ;

[0068] Let α represent thermal stress, α represent the coefficient of linear expansion, and E represent the elastic modulus. This represents the temperature difference; the thermal stress σ is controlled by adjusting the temperature difference ΔT.

[0069] Furthermore, the closed-loop objective function is: ;

[0070] in, This represents the objective function to be optimized. , , This represents the weighting coefficient, and W represents the measured weld pool width. H represents the target molten pool width, and H represents the measured molten pool depth. Indicates the target molten pool depth. This represents the probability of a defect.

[0071] The update rules are as follows:

[0072] ;

[0073] in, This represents the power density distribution in the nth iteration. This represents the updated power density distribution, where η represents the learning rate. It represents the gradient of the objective function; and implements gradient descent closed-loop control.

[0074] This embodiment takes minimizing J as the optimization objective. Finally, the power density distribution is updated in real time according to the update rules through the gradient descent iterative update algorithm. The optimized power density distribution is mapped to spot morphology parameters (radial distribution control parameters, angular non-uniformity control parameters, multi-peak structure adjustment parameters) and energy regulation parameters (pulse width, modulation frequency, duty cycle), and sent to the spot morphology real-time regulation module and energy distribution regulation module to achieve accurate output of regulation commands.

[0075] The control system in this embodiment achieves continuous real-time reconstruction of the spot shape through the collaborative operation of the beam shaping unit and the multi-level optical field reconstruction unit of the real-time spot shape control module. It can flexibly generate various spots according to the state of the molten pool, solving the problem of fixed spot shape and discrete switching in traditional systems. It effectively compensates for process deviations such as directional heat accumulation, unilateral overheating, and deposition offset in LDED deposition. The energy distribution adjustment module realizes spatial partition control and dynamic temporal modulation of laser energy, breaking the limitation of traditional systems that can only adjust the total laser power as a whole. It can directionally control the temperature gradient and cooling rate of the molten pool, and achieve precise control of the deposition structure.

[0076] Meanwhile, the multi-dimensional sensor array provides full-dimensional feedback data on the state of the molten pool. Combined with the multi-algorithm fusion real-time control and optimization algorithm built into the controller, a true multi-field coupled closed-loop gradient optimization control system is constructed to achieve full closed-loop regulation.

[0077] Example 2:

[0078] This embodiment provides a real-time spot morphology and energy distribution feedback control method for LDED, implemented based on the control system described in Embodiment 1, as follows: Figure 5 As shown, the specific steps are as follows:

[0079] Step 1: Multi-source data acquisition.

[0080] Multidimensional sensor arrays are used to synchronously collect real-time data of the molten pool during the deposition process. The multi-source data includes: measured molten pool width W, measured molten pool depth H, molten pool temperature T, plasma spectral intensity Is, acoustic emission signal Ae, and interlayer height deviation Δh.

[0081] The collected raw data is transmitted to the controller in real time. The controller performs filtering, noise reduction, normalization, and time synchronization preprocessing on the data to eliminate errors caused by equipment interference, dimensional differences, and asynchronous acquisition, resulting in a high-quality standardized dataset.

[0082] Step 2: Prediction and determination of molten pool state.

[0083] The controller's built-in real-time control and optimization algorithms perform calculations and analyses on the preprocessed multi-source data to obtain the molten pool state:

[0084] First, using a multi-field coupled model numerical calculation algorithm, combined with material properties (material density ρ, specific heat capacity c, thermal conductivity k, etc.), and based on the heat conduction equation and thermal stress relationship model, core physical field parameters such as the molten pool temperature gradient G, cooling rate R, and thermal stress σ are calculated. Then, using a state prediction and defect determination algorithm, based on the above physical field parameters and sensor-collected data, the defect probability P of the molten pool is quantitatively calculated. defect (Including the probability of occurrence of cracks, porosity, and overheating defects) and microstructure prediction parameters λ (including dendrite spacing, grain size, etc.); at the same time, the measured molten pool width W and measured molten pool depth H are compared with the target molten pool width W preset by the process. target Target molten pool depth H target By comparing and combining temperature gradients, spectral anomalies, etc., the current specific state of the molten pool can be accurately determined, including five types: overheating, rapid cooling, risk of cracking, risk of porosity, and normal state.

[0085] Step 3: Matching the light spot adjustment mode.

[0086] Based on the specific state of the molten pool obtained in step 2, the controller matches the corresponding spot control mode according to the preset process rules to achieve directional switching of the spot shape. The specific state-spot mode matching relationship is as follows:

[0087] If the molten pool is determined to be in an overheated state (judgment criteria: dual-wavelength infrared detection of excessively high deep temperature gradient and spectral sensor detection of abnormal plasma spectrum), a ring-shaped spot mode is matched to disperse the laser center energy and avoid local overheating of the molten pool;

[0088] If the molten pool is determined to be in a state of excessively rapid cooling (judgment criteria: sudden drop in molten pool surface temperature and cooling rate R exceeding the process threshold), a center-enhanced spot mode is matched to concentrate the laser core energy, increase the molten pool temperature, and slow down the cooling rate.

[0089] If the molten pool is determined to have a risk of cracking (judgment criteria: the acoustic emission array detects microcrack initiation signals and the calculated value of thermal stress σ exceeds the standard), a flat-top laser spot mode is matched to reduce the temperature gradient of the molten pool by uniformly distributing laser energy, thereby reducing thermal stress and inhibiting crack initiation;

[0090] If the molten pool is determined to have a risk of porosity (judgment criteria: X-ray online fluoroscopy detects porosity formation, and laser scattering monitoring shows powder concentration mismatch), a multi-peak compensation spot mode is matched. By optimizing the multi-peak distribution of laser energy, the flow of the molten pool and the melting and fusion effect of the powder are improved, and porosity is eliminated.

[0091] If the molten pool is determined to be in normal condition, the current spot shape and energy parameters are kept unchanged, and real-time monitoring continues.

[0092] It should be noted that when the molten pool temperature T>T m +ΔT oh When T is in an overheated state, the molten pool is determined to be in an overheated state. m ΔT represents the melting point temperature of the metal or alloy used. oh The allowable temperature overshoot is typically set between 200℃ and 300℃, but can be adjusted according to the specific material.

[0093] When the cooling rate of the molten pool R > R c When the molten pool is judged to be cooling too quickly, R is among the factors. c This is the critical cooling rate corresponding to the material, which can be set according to the material properties and grain control requirements, and is generally taken in the range of 10. 3 ~10 5 K / s. The cooling rate can be calculated from the temperature gradient G and the liquid phase cooling time t, i.e., R≈G / t.

[0094] Through the above quantification standards, the controller can map real-time multi-source data into molten pool state parameters, thereby matching the corresponding spot control mode and energy distribution parameters to achieve closed-loop optimized control.

[0095] Step 4: Real-time updates of energy distribution and parameters.

[0096] Based on the molten pool state calculation results from step 2, the controller updates the energy distribution and synchronously adjusts the spot morphology parameters and pulse parameters through a closed-loop objective function optimization algorithm and a gradient descent iterative update algorithm.

[0097] Energy distribution update: Controller constructs objective function

[0098] ;

[0099] And according to the gradient descent iteration rule:

[0100] ;

[0101] Update frequency ≥ 1kHz.

[0102] The controller accurately maps the updated power density distribution to spot shape parameters and pulse parameters, and sends the control commands to the spot shape real-time control module and the energy distribution adjustment module. The spot shape parameters include radial distribution control parameters, angular non-uniformity control parameters, and multi-peak structure adjustment parameters, which are used to guide the spot shape real-time control module to complete the accurate reconstruction of the spot shape. The pulse parameters include pulse width, modulation frequency, and duty cycle, which are used to guide the energy distribution adjustment module to complete the dynamic control of the laser beam energy characteristics.

[0103] Step 5: High-speed closed-loop execution and feedback.

[0104] The real-time beam morphology control module receives beam morphology parameter commands from the controller and adjusts the optical components of the beam shaping unit and the multi-level optical field reconstruction unit in real time to complete the rapid reconstruction of the beam morphology. The energy distribution adjustment module receives pulse parameter commands from the controller and dynamically adjusts the pulse width, modulation frequency, and duty cycle to achieve spatiotemporal distribution control of laser energy. The reconstructed beam and the controlled laser energy work together to change the state of the molten pool. The multi-dimensional sensor array detects the new state of the molten pool after control in real time and transmits the new multi-source data back to the controller, returning to step 1 to start the next round of control cycle, forming a continuous multi-field coupled closed-loop optimization control.

[0105] The control method in this embodiment is based on multi-source sensor data acquisition in all dimensions, which realizes precise quantitative judgment of the molten pool state. Compared with the traditional method that only judges the molten pool state through temperature and visible light images, the judgment result is more comprehensive and accurate, providing a reliable basis for subsequent control. A precise matching rule between molten pool state and spot control mode is established, which can switch the spot shape in a directional manner according to different abnormal states of the molten pool, realizing dynamic adaptive control of spot shape, effectively solving common process problems such as molten pool overheating, excessive cooling, crack risk, and porosity risk. The power density distribution is updated in real time by using gradient descent iteration rule, and the spot shape parameters and pulse parameters are adjusted simultaneously, realizing the coordinated control of spot shape and energy distribution. This breaks the limitation of traditional methods that control spot and energy separately, and realizes precise control of the molten pool from two dimensions. It can effectively control the temperature gradient and cooling rate of the molten pool and realize the directional optimization of the deposition structure.

[0106] Each step is executed cyclically to form a fully closed-loop control, realizing real-time adaptive control of the LDED deposition process. This solves the problem that traditional open-loop / semi-closed-loop methods cannot compensate for dynamic process deviations, thus improving the stability of the deposition process.

[0107] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A real-time spot morphology and energy distribution feedback control system for LDED, characterized in that, include: The laser unit is used to emit a laser beam; Multidimensional sensors are used to collect molten pool status parameters in real time; The real-time beam shape control module is used to control the spatial intensity distribution of the laser beam and dynamically reconstruct the beam shape; the energy distribution adjustment module is used to control the energy characteristics of the laser beam. The controller has a built-in real-time control and optimization algorithm. The real-time control and optimization algorithm constructs an objective function based on multi-source feedback data and updates the power density distribution in real time through gradient descent. The real-time beam morphology control module includes a beam shaping unit and a multi-level optical field reconstruction unit. The beam shaping unit is used to adjust the beam intensity, and the multi-level optical field reconstruction unit can form a double-ring beam, an inner high-outer low gradient beam, an edge-enhanced beam, an offset elliptical beam, or a single-sided enhanced beam. The beam shaping unit includes a digital intensity distribution control component and a physical beam morphology control component. The digital intensity distribution control component is used to control the spatial distribution of the laser beam intensity, and the physical beam morphology control component is used to physically control the geometric shape, focusing characteristics, and wavefront shape of the laser beam. The multi-level optical field reconstruction unit includes a composite diffractive optical element, a phase plate, and a ring phase modulation disk. The composite diffractive optical element is used for primary basic modulation, the phase plate is used for secondary precise shaping, and the ring phase modulation disk is used for tertiary dynamic adaptation.

2. The real-time spot morphology and energy distribution feedback control system for LDED according to claim 1, characterized in that, The energy distribution adjustment module controls the temperature gradient and cooling rate of the molten pool by adjusting core parameters; The core parameters include pulse width, modulation frequency, duty cycle, and power density.

3. The real-time spot morphology and energy distribution feedback control system for LDED according to claim 1, characterized in that, The real-time control and optimization algorithms include a multi-field coupled model numerical calculation algorithm, a state prediction and defect determination algorithm, a closed-loop objective function optimization algorithm, and a gradient descent iterative update algorithm.

4. A method for real-time beam morphology and energy distribution feedback control for LDED, characterized in that, Based on the control system as described in any one of claims 1-3, it includes: Collect data from multiple sources; The defect probability and structure prediction parameters are calculated based on multi-source data to obtain the molten pool state; The corresponding spot control mode is matched according to the state of the molten pool; and the energy distribution, spot morphology parameters and pulse parameters are updated in real time.

5. The method for real-time spot morphology and energy distribution feedback control for LDED according to claim 4, characterized in that, The step of matching the corresponding spot control mode according to the molten pool state includes: When the molten pool is overheated, it switches to an annular spot; when the cooling is too fast, it switches to a center-enhanced spot; when there is a risk of cracking, it switches to a flat-top spot; and when there is a risk of porosity, it switches to a multi-peak compensation spot.

6. The method for real-time spot morphology and energy distribution feedback control for LDED according to claim 4, characterized in that, The energy distribution update satisfies: ； in, This represents the power density distribution in the nth iteration. This represents the updated power density distribution, where η represents the learning rate. Represents the gradient of the objective function; Gradient descent closed-loop control is formed.

7. The method for real-time spot morphology and energy distribution feedback control for LDED according to claim 6, characterized in that, By employing gradient descent iterative rules, the power density distribution is corrected in real time based on the gradient of the objective function, thereby achieving multi-field coupled closed-loop optimization control.

8. The method for real-time spot morphology and energy distribution feedback control for LDED according to claim 4, characterized in that, The light spot morphology parameters include radial distribution control parameters, angular non-uniformity control parameters, and multi-peak structure adjustment parameters. The pulse parameters include pulse width, modulation frequency, and duty cycle.

Images