A titanium alloy VAR smelting pool to edge control method and storage medium

Abstract

The present application specifically relates to a titanium alloy VAR smelting pool-to-edge control method and storage medium. The control method includes: S1, determining the edge area of the molten pool; S2, predicting and modeling the boundary perturbation propagation model to obtain the boundary perturbation evolution result; S3, obtaining the multi-scale perturbation energy corresponding to the boundary perturbation evolution result; S4, jointly mapping the boundary perturbation evolution result and the corresponding multi-scale perturbation energy to three-dimensional space, and constructing a heterogeneous reconstruction model; S5, using the heterogeneous reconstruction model to obtain the perturbation state vector of the edge perturbation body; S6, clustering the perturbation state vector of the edge perturbation body in three-dimensional space to divide the stability block of the molten pool boundary area; S7, using the divided stability block, constructing the boundary perturbation dominant direction field in the unstable block, designing a dynamic response control model, and realizing the physical feedback control of the titanium alloy VAR smelting pool-to-edge. The present application can significantly improve the stability of the molten pool morphology and the control efficiency.

Description

Technical Field

[0001] The present application belongs to the field of industrial intelligent technology, and specifically relates to a titanium alloy VAR smelting pool to edge control method and storage medium. Background Art

[0002] The titanium alloy vacuum arc remelting (VAR) process is an important metallurgical means to achieve the preparation of high-purity, high-performance titanium alloys. The morphological evolution of the molten pool boundary largely determines the uniformity and defect rate of the finished product structure.

[0003] In related technologies, the morphological evolution at the molten pool boundary mostly relies on single-point temperature signals, coarse image edge extraction and other methods to control the molten pool. These control strategies generally have technical limitations and have much room for improvement. Summary of the Invention

[0004] The present application provides a titanium alloy VAR smelting pool-to-edge control method and storage medium, aiming to improve the molten pool morphology stability and control efficiency to at least a certain extent.

[0005] In a first aspect of the present application, a titanium alloy VAR smelting pool-to-edge control method is provided, the control method comprising: S1, collecting a photothermal response map of the molten pool area, performing basic positioning of the edge area of the molten pool, and determining the edge area of the molten pool; S2, extracting the boundary curves of the edge area under different time frames, using them as input to construct a time series dynamic model, predicting and modeling a boundary perturbation propagation model, and obtaining a boundary perturbation evolution result; S3, constructing a multi-scale boundary perturbation wavelet energy detection mechanism based on the boundary perturbation evolution result, and obtaining a multi-scale perturbation energy corresponding to the boundary perturbation evolution result; S4, The boundary perturbation evolution results and the corresponding multi-scale perturbation energy are jointly mapped to the three-dimensional space to construct a heterogeneous reconstruction model; S5. The heterogeneous reconstruction model is used to obtain the perturbation state vector of the edge perturbation body; S6. The perturbation state vector of the obtained edge perturbation body is clustered and analyzed in the three-dimensional space to divide the stability blocks of the molten pool boundary area, and the stability blocks include a high stability area, a transition area and an unstable area; S7. The divided stability blocks are used to construct the boundary perturbation dominant direction field in the unstable block, and a dynamic response control model is designed to realize physical feedback control of the titanium alloy VAR smelting molten pool to the edge.

[0006] In some embodiments, step S1 specifically includes: collecting and fusing short-wave infrared, medium-wave infrared, and visible light band signals of the molten pool region to form a photothermal response spectrum of the molten pool region; setting the maximum value of the boundary intensity gradient in the fusion output as a preliminary boundary judgment line to achieve basic positioning of the edge region and determine the edge region of the molten pool;

[0007] In some embodiments, the boundary disturbance propagation model is:

[0008] ;

[0009] in, The angle of the lower boundary of the t-th frame location; The angle of the lower boundary of the t-1 frame is location; is the disturbance propagation coefficient; is the relaxation coefficient; is the Laplace operator, is the partial derivative with respect to t.

[0010] In some embodiments, the multi-scale boundary perturbation wavelet energy detection mechanism is:

[0011] ;

[0012] in, is the angle at time t The multi-scale perturbation energy of is the wavelet transform coefficient of the jth scale; The angle of the lower boundary of the t-th frame location; is the scale weight coefficient; is the total number of scales.

[0013] In some embodiments, in step S5, the disturbance state vector of the edge disturbance body is obtained by the following formula:

[0014] ;

[0015] in, is the angle at time t in the three-dimensional reconstruction space The perturbation state vector, The angle of the lower boundary of the t-th frame The location, is the angle at time t The multi-scale perturbation energy of is the angle at time t The relative perturbation steepness.

[0016] In some embodiments, the Use the following formula to obtain:

[0017] ;

[0018] in, is the angle at time t The second derivative of the boundary space position function with respect to the angle, is the partial derivative with respect to t.

[0019] In some embodiments, step S6 specifically includes:

[0020] Construct a density-weighted aggregation model based on disturbance bodies, and form a stability distribution map by evaluating the density and steepness weight of disturbance bodies within a set time window;

[0021] The aggregation degree model is:

[0022] ;

[0023] in, Angle Boundary stability index at ; are the weighting coefficients of energy perturbation and morphology steepness, respectively; To set the time window length, is the angle at time t The multi-scale perturbation energy of is the angle at time t The multi-scale perturbation energy of , t0 is the starting time of the time window;

[0024] Angle domain for all boundary areas Sliding window scanning, according to The distribution is divided into stability blocks, which include a high stability zone, a transition zone and an unstable zone.

[0025] In some embodiments, the boundary perturbation dominant direction field constructed in the unstable block is:

[0026] ;

[0027] in, : At time t, the angle The disturbance direction vector.

[0028] In some embodiments, the dynamic response regulation model is:

[0029] ;

[0030] in, Angle Energy input adjustment at ; Adjust the scaling factor for damping; is the unit vector of the main direction of the disturbance; is the directional projection component.

[0031] In a second aspect of the present application, the present application further provides a computer-readable storage medium, wherein the storage medium stores a program for causing a computer to execute the above-mentioned control method.

[0032] The titanium alloy VAR smelting pool-to-edge control method and control program provided in this application, by constructing a disturbance identification process based on the angle dynamic parameter field, and introducing disturbance direction potential vector field modeling, disturbance steepness function quantification and disturbance causal chain identification mechanism, can not only realize the pre-identification and classification response of the disturbance starting point, but also construct a local directional control strategy based on the disturbance propagation direction, significantly improving the stability of the molten pool morphology and the control efficiency, and is suitable for titanium alloy VAR metallurgical preparation scenarios with high consistency requirements, and has good industrial applicability and promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0034] Figure 1 A flow chart of a titanium alloy VAR smelting pool-to-edge control method based on machine vision proposed in this application is shown. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to understand the present application more clearly, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of this application.

[0036] As a lightweight, high-strength metal material, titanium alloy has a wide range of applications in aerospace, medical equipment and other fields.

[0037] Application. Vacuum consumable arc furnace (VAR), as the main method for melting titanium alloys, has the advantages of low power consumption, high melting speed and good quality reproducibility. However, during the VAR melting process, the control of the molten pool to the edge often depends on the experience and skills of the operator, resulting in certain fluctuations in the surface quality, elemental composition uniformity and yield rate of titanium alloy ingots. Machine vision is a technology that uses computers to perform image recognition and analysis. It can simulate the human visual system to achieve image understanding and processing. Using machine vision instead of the operator's naked eye to identify and judge the control of the molten pool to the edge is beneficial to the stability of the quality and surface quality of titanium alloy ingots. Therefore, it is of great practical significance to develop a method based on machine vision to control the molten pool to the edge of titanium alloy VAR melting.

[0038] In related technologies, methods for controlling the molten pool to the edge of titanium alloy VAR smelting mostly rely on single-point temperature signals and image edge coarse extraction to control the molten pool. These control strategies generally have the following technical limitations:

[0039] First, there is a lack of spatial-temporal disturbance feature modeling capabilities. Current control models often extract single geometric indicators (such as edge width and average diameter) based on static or near-field image information. These models struggle to capture the angular evolutionary trends and directions of disturbances, and are unable to systematically determine the starting point, propagation path, and intensity of disturbances, resulting in a delayed control response.

[0040] Second, it is impossible to identify the causal mechanism of disturbances in advance and implement source control intervention. In actual VAR processes, disturbances often spread from local sources in a chain-like manner. Existing methods often rely on post-hoc judgment or boundary threshold triggering mechanisms. They lack a mechanism to identify the disturbance's "source-diffusion" causal chain, making it impossible to focus control precision on the disturbance's starting area. This results in dispersed control energy consumption and large fluctuations in boundary stability.

[0041] Based on the above technical problems, this application provides a method for controlling the edge of the melt pool in titanium alloy VAR smelting. By constructing a disturbance identification process based on the angular dynamic parameter field, and for the first time introducing the disturbance direction potential vector field modeling, disturbance steepness function quantification, and disturbance causal chain identification mechanism. This method not only realizes the pre-identification and classification response of the disturbance starting point, but also constructs a local directional control strategy based on the disturbance propagation direction, significantly improving the stability of the melt pool morphology and the control efficiency. It is suitable for titanium alloy VAR metallurgical preparation scenarios with high consistency requirements and has good industrial applicability and promotion value.

[0042] Figure 1 The following is a flow chart of a titanium alloy VAR smelting pool-to-edge control method based on machine vision proposed in this application. Figure 1In a first aspect, the present application provides a method for controlling the molten pool to the edge of a titanium alloy VAR smelting process based on machine vision, comprising the following steps:

[0043] S1. Collect the photothermal response map of the molten pool area, perform basic positioning of the edge area of the molten pool, and determine the edge area of the molten pool.

[0044] During the VAR melting process, strong arc interference, surface tension disturbances, and thermal conduction coupling can make identifying the melt pool boundary difficult. By fusing the acquired shortwave infrared (SWIR), mediumwave infrared (MWIR) and visible light signals, a photothermal response map of the melt pool region is generated. This map accurately separates the concentrated radiation energy region at the melt pool edge (representing the thermal energy boundary) from the visible brightness boundary (representing the liquid surface fluctuation boundary). By setting the maximum boundary intensity gradient in the fused output as the initial boundary judgment line, basic edge location is achieved.

[0045] S2. Extract the boundary curves of the edge area under different time frames and use them as input to build a time series dynamic model, predict and model the boundary disturbance propagation model, and obtain the boundary disturbance evolution results.

[0046] The molten pool boundary in VAR smelting is not static, but exhibits quasi-periodic deformation under arc disturbance. To quantify this deformation process, a time series dynamic model is constructed and a family of time series boundary functions is introduced. , where the boundary contour of each time point is expressed in polar coordinates;

[0047] In order to model this deformation behavior, a boundary disturbance propagation model is constructed to describe the impact intensity of the previous moment's disturbance on the current boundary position.

[0048] Traditional methods rely on boundary locations extracted from single or intermittent frames, using instantaneous geometric parameters of the contour edge (such as diameter, area, and maximum-minimum radius difference) to assess boundary status. These methods lack continuous time series modeling. Furthermore, they fail to model the temporal evolution of disturbances. This means that the impact of previous disturbances on current boundary behavior cannot be quantified, resulting in a lack of boundary trend prediction and limited passive control based on empirical thresholds.

[0049] The boundary disturbance propagation model is as follows:

[0050] ;

[0051] in, The angle of the lower boundary of the t-th frame The position (polar coordinate radius) at , i.e. the boundary perturbation evolution result; The angle of the lower boundary of the t-1 frame is location; is the disturbance propagation coefficient, which describes the boundary inertia and controls the inertia inheritance strength of the previous frame disturbance trend on the current boundary; : relaxation coefficient, which describes the spatial diffusion of disturbances, suppresses local peak morphology, and is used to smooth the curve (disturbance diffusion); The Laplace operator is used to model the smooth diffusion trend of the boundary curve and to describe the boundary at an angle. The second-order spatial variation on reflects the steepness of the morphology. is the partial derivative with respect to t, which is used to describe the instantaneous rate of change of the boundary position over time.

[0052] The boundary disturbance propagation model extracts the boundary curves of the edge areas of multiple consecutive frames and is used to predict the boundary trend at the next moment.

[0053] S3. Based on the boundary perturbation evolution result, a multi-scale boundary perturbation wavelet energy detection mechanism is constructed to obtain the multi-scale perturbation energy corresponding to the boundary perturbation evolution result.

[0054] During the VAR process of titanium alloys, when local energy accumulation at the boundary exceeds the limit, it can cause upthrust at the edge of the molten pool or abnormal cooling fluctuations. In the multi-scale boundary perturbation wavelet energy detection mechanism, the boundary perturbation curve of each frame is projected into the wavelet domain, and the high-frequency energy mutation area is detected at multiple frequency scales to obtain the multi-scale perturbation energy. The mechanism uses the following formula:

[0055] ;

[0056] in, is the angle at time t The multi-scale perturbation energy of is the wavelet transform coefficient of the jth scale; is the scale weight coefficient, which reflects the sensitivity of different frequency bands to the disturbance energy; is the total number of scales, which is selected based on the frequency range of the boundary disturbance. Usually 5 to 10 scales are selected to cover multi-band responses.

[0057] High recognition The value area is used as the key point of potential edge distortion for the next step of three-dimensional imaging reconstruction.

[0058] Traditional methods for detecting anomalies at the edge of a melt pool often rely on image grayscale histogram analysis, edge sharpness gradient changes, or fixed contour offset thresholds. These methods are unable to identify local high-frequency disturbances or nonlinear energy superposition effects. Compared to traditional methods based on single-domain signal processing in time or space, this method performs a wavelet expansion on the boundary disturbance sequence in the angle-time domain, enabling energy anomalies to be clearly captured at multiple frequency scales.

[0059] S4. Construct a heterogeneous reconstruction model to jointly map the boundary perturbation evolution results and the corresponding multi-scale perturbation energy in the two-dimensional image into the three-dimensional space to form an edge perturbation body.

[0060] Since only the boundary perturbation energy anomaly is identified in step S3, VAR process control requires a clear understanding of its actual manifestation in the three-dimensional physical morphology. To this end, a heterogeneous reconstruction model is established to jointly map the polar coordinate perturbation trajectory in the two-dimensional image and its corresponding energy spectrum into three-dimensional space to form the edge perturbation volume.

[0061] S5. Using the heterogeneous reconstruction model, we can obtain the perturbation state vector of the edge perturbation body. :

[0062] ;

[0063] in, is the angle at time t in the three-dimensional reconstruction space The perturbation state vector of The angle of the lower boundary of the t-th frame The location, is the angle at time t The multi-scale perturbation energy of is the angle at time t The relative perturbation steepness.

[0064] in, Use the following formula to obtain:

[0065] ;

[0066] in, is the angle at time t The second-order derivative of the boundary space position function with respect to the angle reflects the curvature change trend of the boundary in the angle dimension. is the partial derivative with respect to t, which is used to describe the instantaneous rate of change of the boundary position over time.

[0067] In practice, if Represents the radius function of the boundary contour in polar coordinates, then the second-order derivative represents the degree of drastic change of the contour curvature with angle, that is, the "turning" speed.

[0068] S6. Perform cluster analysis on the disturbance state vector of the obtained edge disturbance body in three-dimensional space to divide the stability block of the molten pool boundary area, wherein the stability block includes a high stability area, a transition area and an unstable area.

[0069] The stability of the titanium alloy VAR melt pool boundary may vary significantly at different angles, with overheating zones and boundary depression zones coexisting. Therefore, a stability classification in the spatial domain is necessary. To this end, a polymerization degree model is proposed.

[0070] When performing cluster analysis, a density-weighted aggregation model based on disturbance bodies is constructed. By evaluating the density and steepness weights of disturbance bodies within a certain time window, a stability distribution map is formed.

[0071] The aggregation degree model is as follows:

[0072] ;

[0073] in, Angle Boundary stability index at (higher values indicate more instability); are the weighting coefficients of energy perturbation and morphology steepness, respectively; To set the time window length, is the angle at time t The multi-scale perturbation energy of is the angle at time t The multi-scale perturbation energy of t0 is the starting time of the time window.

[0074] Angle domain for all boundary areas Sliding window scanning, according to The distribution is divided into stability blocks: high stability zone, transition zone and unstable zone, which serve as the basis for selecting control and regulation areas.

[0075] In this application, stability binning is identified based on a statistically weighted model of disturbance energy and steepness. However, this approach is inherently result-oriented, classifying disturbances based on the disturbance phenomenon. To further enhance the control foresight and response efficiency of this step, a "source-diffusion" causal chain identification mechanism for melt pool disturbances is introduced as a priori layer before stability binning. This mechanism predicts the possible evolution path of the disturbance from its causal chain, effectively improving the time sensitivity and control accuracy of boundary disturbance identification.

[0076] The mechanism includes: Disturbance source candidate screening: Through continuous multi-frame visual image streams, the changes in the melt pool temperature distribution gradient, the changes in the boundary morphology curvature, and the spatial distribution density of visual interference information are analyzed, and short-term high-frequency change points are clustered and marked to identify a set of possible disturbance source candidate points;

[0077] Disturbance chain construction and classification: Perform outward expansion analysis on the candidate disturbance source in the time domain and track its impact on the surrounding area and The dynamic impact path of the disturbance is constructed to construct the causal propagation path of the disturbance. The disturbance chain with obvious outward expansion behavior is judged as the "primary source-diffusion chain", and the remaining isolated disturbances are judged as "isolated non-causal disturbances";

[0078] Disturbance priority weight assignment: For the identified disturbance chain structure, a high-weighted priority intervention level is assigned. In particular, when a disturbance source point has diffused multiple times within a historical period, and the propagation range covers multiple boundary angle segments, the disturbance block corresponding to the source point will be marked as a "predicted unstable block". Even if the current disturbance intensity value has not yet reached the traditional judgment threshold, it must be prioritized for subsequent control links.

[0079] Causal chain-guided control nesting: During the subsequent steps of disturbance stability partitioning, directional vector field construction, and control input parameter allocation, all angular regions marked as "source-driven blocks" are automatically designated as high-weight control zones. Control resources (such as energy fine-tuning amplitude and arc waveform frequency) are then allocated to these blocks, effectively blocking the source of disturbance propagation.

[0080] This application effectively breaks the lag of "adjustment after disturbance occurs" in traditional VAR melting boundary stability control, realizes early prediction of disturbance, source identification and differentiated control, and significantly improves the overall boundary stability of titanium alloy VAR melt pool and the consistency of finished product crystal.

[0081] S7. Utilize the divided stability blocks to construct the boundary disturbance dominant direction field within the unstable blocks, design a dynamic response control model, and realize physical feedback control from the molten pool to the edge of the titanium alloy VAR smelting process.

[0082] The dynamic perturbation at the boundary has not only amplitude characteristics but also directionality, that is, the perturbation is more significant in certain directions, which can easily lead to melt pool upthrust or boundary collapse. An evolution vector field constructed based on the perturbation gradient is proposed to characterize the main direction of perturbation development:

[0083] ;

[0084] in, is the angle at time t The disturbance direction vector of

[0085] The first dimension describes the direction of the energy gradient, and the second dimension describes the direction of rapid morphological changes. Current boundary control in titanium alloy VAR smelting processes is primarily based on static identification of disturbance amplitude or temperature distribution, lacking dynamic modeling of disturbance evolution trends and directions. This is especially true in non-steady-state regions, where traditional methods are unable to identify the dominant propagation direction of disturbances, resulting in a delayed or non-targeted regulatory response. This disturbance direction gradient vector field, by jointly analyzing the energy density change gradient and the boundary morphology change gradient, constructs the dominant direction field of disturbance propagation in the angular domain for the first time. This allows for precise identification of disturbance evolution paths at the microscopic level, providing a direction-oriented response basis for subsequent regulatory strategies.

[0086] When designing the dynamic response control model, the molten pool disturbance in the VAR process is locally energy-guided by fine-tuning the arc current pulse and changing the electrode height;

[0087] By using the perturbation direction vector The damping adjustment factor is used to construct a dynamic response control model so that the disturbance energy can be effectively absorbed or balanced in the unstable block.

[0088] Existing VAR melting control strategies often rely on regional uniform regulation or preset fixed power distributions, which are unable to precisely intervene based on the real-time location and propagation direction of disturbances. This can easily lead to regulation lag, over-response, or boundary imbalance, especially when edge disturbances asymmetrically increase. This model constructs an excitation function based on the dominant disturbance direction and the local disturbance intensity gradient, dynamically adjusting the distribution shape and intensity of the control response within the angular domain. This enables "directional repair" of the melt pool edge and local convergence control, significantly improving boundary stability and energy response efficiency.

[0089] The dynamic response control model is as follows:

[0090] ;

[0091] in, Angle Energy input adjustment at ; Adjust the scaling factor for damping; is the unit vector of the main direction of the disturbance; is the directional projection component, representing the growth trend of the disturbance along the main direction.

[0092] Damping adjustment proportional factor The acquisition is dynamically adjusted based on the change rate of the dominant direction of the disturbance and the intensity of the local energy gradient. Specifically, by calculating the time change rate of the disturbance direction vector modulus, it is judged whether the disturbance expansion is aggravated, and at the same time, the maximum energy gradient in the boundary area is combined to identify the concentration of the disturbance. On this basis, considering the response inertia of the system itself to the control signal, a dynamic adjustment strategy of the adjustment factor is formed comprehensively, so that It automatically increases when the disturbance increases, improving the regulation concentration, and decreases when the disturbance stabilizes, avoiding excessive response of the system, thereby achieving real-time optimization control of the regulation intensity.

[0093] The arc energy distribution in the local area is adjusted according to the model results (such as fine-tuning the pulse peak, reducing the local arc residence time, etc.) to achieve physical feedback control of boundary disturbances.

[0094] Based on the above control method, in a second aspect, the present application further provides a storage medium storing a program for causing a computer to execute the above control method.

[0095] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0096] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the existing technology, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0097] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0098] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0099] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0100] The titanium alloy VAR smelting pool-to-edge control method and its control program provided in this application, by constructing a disturbance identification process based on the angle dynamic parameter field, can accurately perceive the evolution characteristics of the disturbance at the pool boundary in multi-dimensional space; further combined with the disturbance direction potential vector field modeling, it can realize the dynamic prediction of the disturbance development trend. On this basis, the disturbance steepness function is used to quantitatively evaluate the local abnormal changes to ensure that the system has the ability to finely distinguish the strength and mutation degree of the disturbance; at the same time, through the disturbance causal chain identification mechanism, the system can identify the potential diffusion pattern and possible affected area in the early stage of the disturbance, significantly improving the timeliness of the pre-warning and response decision. Compared with the existing control technology that only relies on global parameters or static feature analysis, this method can actively generate a local directional control strategy according to the disturbance propagation direction, realize dynamic adaptive correction of the pool morphology, and greatly improve the system's ability to maintain the stability of the pool boundary under complex disturbance environments and the control accuracy. In actual VAR smelting production, the method of this application can effectively reduce the frequency of boundary distortion, improve the consistency of the molten pool morphology, and shorten the control response time. It is suitable for titanium alloy metallurgical preparation scenarios with extremely high requirements for composition uniformity and organizational continuity, and has good industrial applicability, scalability and promotion and application value.

[0101] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not intended to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application, and all of these should be included in the scope of the claims of the present application.

Claims

1. A titanium alloy VAR smelting pool edge control method, characterized in that: The control method includes: S1. Collecting a photothermal response map of the molten pool area, performing basic positioning of the edge area of the molten pool, and determining the edge area of the molten pool; S2. Extract the boundary curves of the edge area under different time frames and use them as input to build a time series dynamic model, predict and model the boundary disturbance propagation model, and obtain the boundary disturbance evolution result; S3. Based on the boundary perturbation evolution result, construct a multi-scale boundary perturbation wavelet energy detection mechanism to obtain the multi-scale perturbation energy corresponding to the boundary perturbation evolution result; S4, jointly mapping the boundary perturbation evolution result and the corresponding multi-scale perturbation energy into a three-dimensional space, constructing a heterogeneous reconstruction model, and forming an edge perturbation body; S5. Using the heterogeneous reconstruction model, the perturbation state vector of the edge perturbation body is obtained; S6. Performing cluster analysis on the disturbance state vector of the obtained edge disturbance body in three-dimensional space to divide the stability block of the molten pool boundary area, wherein the stability block includes a high stability zone, a transition zone, and an unstable zone; S7. Utilize the divided stability blocks to construct the boundary disturbance dominant direction field within the unstable blocks, design a dynamic response control model, and realize physical feedback control from the molten pool to the edge of the titanium alloy VAR smelting process.

2. The titanium alloy VAR smelting pool edge control method according to claim 1, characterized in that: The step S1 specifically includes: Collect and fuse the short-wave infrared, medium-wave infrared and visible light band signals of the molten pool area to form a photothermal response map of the molten pool area; By setting the maximum value of the boundary intensity gradient in the fusion output as the preliminary boundary judgment line, basic positioning of the edge area is achieved, and the edge area of the molten pool is determined.

3. The titanium alloy VAR smelting pool edge control method according to claim 1, characterized in that: The boundary disturbance propagation model is: ； in, The angle of the lower boundary of the t-th frame location; The angle of the lower boundary of the t-1 frame is location; is the disturbance propagation coefficient; is the relaxation coefficient; is the Laplace operator, and ∂ is the partial derivative with respect to t.

4. The titanium alloy VAR smelting pool edge control method according to claim 1, characterized in that: The multi-scale boundary perturbation wavelet energy detection mechanism is: ； in, is the angle at time t The multi-scale perturbation energy of is the wavelet transform coefficient of the jth scale; The angle of the lower boundary of the t-th frame location; is the scale weight coefficient; is the total number of scales.

5. The titanium alloy VAR smelting pool edge control method according to claim 1, characterized in that: In step S5, the disturbance state vector of the edge disturbance body is obtained by the following formula: ； in, is the angle at time t in the three-dimensional reconstruction space The perturbation state vector, The angle of the lower boundary of the t-th frame The location, is the angle at time t The multi-scale perturbation energy of is the angle at time t The relative perturbation steepness.

6. The titanium alloy VAR smelting pool edge control method according to claim 1, characterized in that: The step S6 specifically includes: Construct a density-weighted aggregation model based on disturbance bodies, and form a stability distribution map by evaluating the density and steepness weight of disturbance bodies within a set time window; The polymerization degree model is: ； in, Angle Boundary stability index at ; are the weighting coefficients of energy perturbation and morphology steepness, respectively; To set the time window length, is the angle at time t The multi-scale perturbation energy of , t0 is the starting time of the time window, is the angle at time t The relative perturbation steepness; Angle domain for all boundary areas Sliding window scanning, according to The distribution is divided into stability blocks, which include a high stability zone, a transition zone and an unstable zone.

7. The titanium alloy VAR smelting pool edge control method according to claim 6, characterized in that: The dominant direction field of the boundary disturbance constructed in the unstable block is: ； in, : At time t, the angle The disturbance direction vector.

8. The titanium alloy VAR smelting pool edge control method according to claim 7, characterized in that: The dynamic response control model is: ； in, Angle Energy input adjustment at ; Adjust the scaling factor for damping; is the unit vector of the main direction of the disturbance; is the directional projection component.

9. A computer-readable storage medium, characterized in that The storage medium stores a program for causing a computer to execute the control method according to any one of claims 1 to 8.

Images