Electric arc additive and subtractive manufacturing multi-source information online sensing device and method

Abstract

The invention discloses an electric arc additive and subtractive manufacturing multi-source information online sensing device and method. The device comprises a substrate, double mechanical arms, an additive and subtractive system, a multi-source signal collecting system and a signal processing system. The base plate supports the workbench, the double mechanical arms execute material increasing and decreasing operation, the material increasing system achieves preliminary forming of metal parts, and the material decreasing system improves precision. The multi-source signal acquisition system acquires whole-flow information through a high-speed camera, a high-frequency microphone, a thermal imager and a magneto-optical imager, the whole-flow information is preprocessed by a signal acquisition card and then transmitted to the signal processing system, and online monitoring is realized through data fusion, feature extraction and machine learning modeling; the method comprises the steps of model establishment, information collection, additive and subtractive machining and signal judgment. The problems that in the prior art, signal collection is single, and the whole additive and subtractive process cannot be adapted are solved, efficient and high-precision integrated forming is achieved, the sensing precision is high, and the method is suitable for the high-precision requirement field.

Description

Technical Field

[0001] This invention relates to the field of electric arc additive and subtractive manufacturing technology, specifically to an online sensing device and method for multi-source information in electric arc additive and subtractive manufacturing. Background Technology

[0002] Additive manufacturing, as a novel material processing technology, offers advantages over traditional forming processes such as casting and forging, including higher processing efficiency and lower costs. It is particularly advantageous in manufacturing complex parts. Arc additive manufacturing, in particular, is a highly efficient additive manufacturing technology often used to manufacture complex structural components. However, as researchers delve deeper into additive manufacturing, its drawbacks have become increasingly apparent. Parts manufactured using this technology often exhibit unsatisfactory performance, failing to meet high standards in terms of precision and surface finish. This significantly limits the application of additive manufacturing in fields with high precision requirements, such as aerospace and military manufacturing.

[0003] Compared to additive manufacturing, subtractive manufacturing processes have disadvantages such as low production efficiency, low material utilization, and long processing cycles. However, parts manufactured using traditional subtractive manufacturing processes often possess high precision and performance. Therefore, combining additive and subtractive manufacturing, leveraging the advantages of both, can achieve both high forming efficiency and low cost, as well as high precision and high performance.

[0004] However, there is currently a lack of multi-source online sensing systems applicable to both arc additive and subtractive manufacturing. Patent CN116559196A discloses an arc additive manufacturing defect monitoring system and method, which uses a magneto-optical imager, infrared camera, high-speed camera, and acoustic sensor to detect surface and internal defects in arc additive manufactured components from multiple sources. However, this system is not applicable to subtractive manufacturing and cannot provide online monitoring of the forming process. Patent CN116604218A discloses a device for controlling the forming quality of arc additive manufacturing through acoustic reflection feedback. This device uses an acoustic emission acquisition system and a feedback control system to provide online monitoring of arc additive manufactured components, enabling timely detection and remedial adjustments for defects. However, it cannot collect multi-source information and is not suitable for subtractive manufacturing processes. Summary of the Invention

[0005] To overcome the aforementioned shortcomings, an online sensing device and method for multi-source information in arc additive and subtractive manufacturing is provided. First, an arc additive forming system is used to initially form a complex metal part, and four high-precision sensors collect multi-source information, which is then transmitted to a signal analysis system for further analysis and discrimination. Next, a subtractive manufacturing system operates to further process the metal part, improving its precision and performance. During this process, the multi-source signal acquisition and processing system operates again, collecting and analyzing information. The signal processing system extracts feature parameters and uses machine learning methods to establish a defect model, discriminating the multi-source signals, thereby achieving online sensing of multi-source information and comprehensive online monitoring of the component, solving the problems of current online sensing devices.

[0006] The technical solution adopted in this invention is as follows: an online sensing device for multi-source information in arc additive and subtractive manufacturing, comprising a substrate, a robotic arm, an arc additive manufacturing forming system, a subtractive manufacturing system, a multi-source signal acquisition system, and a signal processing system; the substrate has a rotatable base for supporting a worktable to accommodate the additive and subtractive operations of the robotic arm; the robotic arm includes a first robotic arm and a second robotic arm, respectively used to perform subtractive processing and additive forming operations; the arc additive manufacturing forming system cooperates with the second robotic arm to achieve the preliminary forming of complex structural metal parts, and the subtractive manufacturing system cooperates with the first robotic arm to perform high-precision processing on the preliminary formed metal parts; the multi-source signal acquisition system collects multi-source information in real time throughout the entire additive and subtractive manufacturing process and transmits it to the signal processing system, and the signal processing system performs fusion processing, feature parameter extraction, and machine learning discrimination on the multi-source information;

[0007] The arc additive manufacturing forming system includes a first control system, a wire feeding system, an electric welding system, an atmosphere protection system, and a water tank. The wire feeding system includes a wire extrusion device and a wire coil, which work together to deliver the wire to the electric welding system. The electric welding system includes a water-cooled argon arc welding torch mounted at the front end of a second robotic arm. The atmosphere protection system includes a suction system, an air purifier, and an argon protection device. The suction system and air purifier are used to absorb and purify harmful gases generated during processing, and the argon protection device is used to release argon to provide a protective atmosphere. The first control system includes a control host and a teach pendant, used to control the operation of each component of the arc additive manufacturing forming system. The water tank cools various parts of the arc additive manufacturing forming system through water pipes.

[0008] The subtractive manufacturing system includes a second control system, a drive system, a machining system, and a cooling system. The drive system includes a powertrain for driving the first robotic arm and the machining system. The machining system includes a clamp and a milling cutter located at the end of the first robotic arm. The cooling system includes a reservoir and nozzles. The reservoir contains cutting fluid, which is delivered to the machining area through the nozzles to achieve cooling and chip removal. The second control system includes a control panel for unified control of the operation of each component of the subtractive manufacturing system.

[0009] The signal processing system includes a data fusion module, a feature parameter extraction module, and a machine learning module. The data fusion module is used to organize and organically fuse received multi-source signals. The feature parameter extraction module is used to extract relevant feature parameters from the fused multi-source information. The machine learning module is used to receive feature parameters and establish a classification and discrimination model.

[0010] Furthermore, the multi-source signal acquisition system includes a signal acquisition card, a high-speed camera, a magneto-optical imager, a thermal imager, and a high-frequency microphone; the high-speed camera is used to acquire image signals of the component to be processed, the high-frequency microphone is used to acquire sound signals during processing, the thermal imager is used to acquire temperature signals during processing, and the magneto-optical imager is used to acquire magnetic field strength signals and current distribution signals during processing; the signal acquisition card receives analog or digital signals output by each sensor, and transmits them to the signal processing system after noise reduction, noise point removal, and normalization processing.

[0011] This invention also discloses a method for online sensing of multi-source information in arc-assisted additive and subtractive manufacturing based on the above-mentioned device, comprising the following steps:

[0012] S1: Turn on each component of the electric arc additive manufacturing system. After each component is running stably, turn on the high-frequency microphone separately to collect the sound data of the equipment operation as a reference.

[0013] S2: Establish a 3D model of the structural component and perform pre-numerical simulation to generate a machining execution file;

[0014] S3: Activate the high-speed camera, magneto-optical imager, and thermal imager to collect multi-source information on the additive and subtractive manufacturing process in real time;

[0015] S4: Based on the processing execution document, the initial forming of the metal part is completed through the arc additive manufacturing forming system;

[0016] S5: High-precision machining of pre-formed metal parts using a subtractive process;

[0017] S6: The signal acquisition card receives multi-source information and performs filtering and normalization processing, then transmits the processed information to the signal processing system.

[0018] S7: The signal processing system fuses the processed multi-source information and extracts feature parameters. It then establishes a classification and discrimination model through machine learning to achieve online perception of the additive and subtractive manufacturing process.

[0019] Furthermore, step S2 specifically includes:

[0020] S2.1: Use 3D modeling software to create 3D models of structural components and output them as STL format files;

[0021] S2.2: The STL format file is sliced, divided into regions, and path planned using slicing software. After coordinate calibration, an LS format file is generated.

[0022] S2.3: Use virtual simulation software to perform pre-numerical simulation verification of the machining trajectory, and finally output the LP format file to the arc additive manufacturing forming system.

[0023] Furthermore, step S4 specifically includes:

[0024] S4.1: Perform coordinate calibration on the second robotic arm and the worktable, and verify the actual printing trajectory by referring to the numerical simulation results;

[0025] S4.2: Activate the water tank, welding wire extrusion device, metal wire ring, argon gas protection device, air purifier and suction system in sequence, and start printing after each component is working stably;

[0026] S4.3: The second robotic arm and water-cooled argon arc welding gun are started through the teach pendant. An electric arc is formed between the water-cooled argon arc welding gun and the substrate, and the molten metal wire is piled up to form a shape.

[0027] S4.4: After printing one layer, the second robotic arm lifts the water-cooled argon arc welding gun away from the substrate to pause printing. After the molten pool cools down, it moves to the arc-starting position to continue printing, repeating the cycle until the initial forming is completed.

[0028] Furthermore, step S5 specifically includes:

[0029] S5.1: Using the control panel, bring the milling cutter close to the workpiece to be machined, observe the contact situation, and then complete the tool setting using the tool setting function on the control panel;

[0030] S5.2: Simulate the machining program in advance and start the subtractive machining process after confirming that there are no errors.

[0031] Furthermore, the feature extraction processing of multi-source information in step S6 includes: using an edge detection algorithm to extract weld edge and part contour features from dynamic image information acquired by a high-speed camera; extracting peak, valley, and current density distribution features from magnetic field and current distribution data acquired by a magneto-optical imager; extracting mean, variance, hot spot location, and temperature change rate features from temperature field data acquired by a thermal imager; and converting acoustic data acquired by a high-frequency microphone into frequency domain features through Fourier transform to extract amplitude and frequency features.

[0032] Furthermore, the dynamic image data acquired by the high-speed camera is filtered, and the median value is taken as the new pixel value. Then, the pixel value is mapped to the [0, 1] interval for normalization, as shown in the formula:

[0033]

[0034] Where X is the original pixel value.

[0035] Furthermore, the feature parameters extracted in step S7 include weld edge features E, amplitude A, frequency F, mean temperature distribution M, variance V, hot spot location H, temperature change rate R, peak magnetic field strength P, valley magnetic field strength G, and current distribution density D.

[0036] 10. The online sensing method for multi-source information in arc-assisted manufacturing according to claim 3, characterized in that the machine learning adopts linear discriminant analysis, and the established discriminant function is:

[0037] LDA1=w1·E+w2·A+w3·F+w4·M+w5·V+w6·H+w7·R+w8·P+w9·G+w 10 ·D

[0038] LDA2=w1·E+w2·A+w3·F+w4·M+w5·V+w6·H+w7·R+w8·P+w9·G+w 10 ·D

[0039] Where w1, w2, w3... are weighting coefficients.

[0040] The beneficial effects of this invention are:

[0041] 1. This invention effectively combines the arc additive manufacturing process and the subtractive manufacturing process, overcoming the problems of high efficiency but insufficient precision in arc additive manufacturing and low efficiency in subtractive manufacturing, and realizing high-efficiency and high-precision integrated forming of complex structural parts.

[0042] 2. This invention effectively combines a high-speed camera, a high-frequency microphone, a thermal imager, and a magneto-optical imager, overcoming the problem that existing online monitoring systems only collect single signals and cannot be simultaneously applied to the manufacturing process of adding or subtracting materials. It achieves online sensing of multi-source signals for the integrated forming of complex structural parts, with extremely high sensing accuracy. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the structure of the present invention;

[0044] In the diagram: 1-Control panel; 2-Reservoir tank; 3-Powertrain; 4-First robotic arm; 5-Clamping fixture; 6-Milling cutter; 7-Nozzle; 8-Component to be processed; 9-Workbench; 10-Baseboard; 11-Water-cooled argon arc welding torch; 12-Second robotic arm; 13-Water tank; 14-Wire extrusion device; 15-Metal wire coil; 16-Argon gas protection device; 17-Teach pendant; 18-Control host; 19-Signal processing system; 20-Signal acquisition card; 21-High-speed camera; 22-Magneto-optical imager; 23-Thermal imager; 24-High-frequency microphone; 25-Air purifier; 26-Suction system. Detailed Implementation

[0045] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort. To facilitate understanding of the present invention, the present invention will be described in more detail below with reference to the accompanying drawings and specific embodiments.

[0046] like Figure 1 As shown, the present invention is a multi-source information online sensing device for arc additive and subtractive manufacturing, including a robotic arm, an arc additive manufacturing forming system, a subtractive manufacturing system, a multi-source signal acquisition and signal processing system, and a substrate 10; the substrate 10 has a rotatable base that can provide support for a worktable 9 so that the robotic arm can perform additive and subtractive operations on the worktable 9; the robotic arm includes a first robotic arm 4 and a second robotic arm 12, used to operate the component 8 to be processed on the worktable 9.

[0047] The arc additive manufacturing forming system includes a first control system, a wire feeding system, an electric welding system, an atmosphere protection system, and a water tank 13. The wire feeding system includes a wire extrusion device 14 and a wire coil 15. The wire coil 15 is powered by a motor and, in conjunction with the wire extrusion device 14, transports the prepared wire along the second robotic arm 12 to the water-cooled argon arc welding torch 11. The electric welding system includes the water-cooled argon arc welding torch 11, which is mounted at the front end of the second robotic arm 12. The atmosphere protection system consists of a suction system 26, an air purifier 25, and an argon protection device 16. The suction system 26 and the air purifier 25 are located near the worktable 9 to absorb and purify harmful gases generated during the forming process, preventing environmental pollution. The argon protection device 16 releases argon gas to displace other gases, and the argon gas does not react with or dissolve in the metal, providing a protective atmosphere for the arc additive manufacturing forming process. The first control system includes a control host 18 and a teach pendant 17. The control host 18 controls the water-cooled argon arc welding torch 11, the second robotic arm 12, the water tank 13, the wire extrusion device 14, the wire coil 15, the argon gas protection device 16, the air purifier 25, and the suction system 26, respectively, and is directly controlled by the teach pendant 17. The water tank 13 provides cooling for the arc additive manufacturing forming system. The water in the water tank 13 flows along the water pipes throughout the arc additive manufacturing forming system to reduce the temperature.

[0048] The subtractive manufacturing system includes a drive system, a machining system, a cooling system, and a second control system. The drive system includes a powertrain 3, which uses a motor and reducer to drive the first robotic arm 4 and the machining system, providing energy for the machining system. The machining system includes a fixture 5 and a milling cutter 6, both located at the end of the first robotic arm 4. The cooling system includes a coolant tank 2 and nozzles 7. The coolant tank 2 contains cutting fluid and delivers it along the first robotic arm 4 to the nozzles 7 at its end, cooling the subtractive manufacturing process and simultaneously removing chips. The second control system includes a control panel 1, which is connected to the coolant tank 2, the powertrain 3, the first robotic arm 4, the fixture 5, the milling cutter 6, and the nozzles 7, and is uniformly controlled by the control panel 1.

[0049] The multi-source signal acquisition system includes a signal acquisition card 20, a high-speed camera 21, a magneto-optical imager 22, a thermal imager 23, and a high-frequency microphone 24. The high-speed camera 21 uses a high-speed CMOS camera to acquire image signals of the component 8 to be processed. It has extremely high resolution and frame rate; at a resolution of 1280×800 pixels, the frame rate can reach 40,000 fps. In standard mode, at a resolution of 1280×400 pixels, the frame rate can reach 75,000 fps, and the electronic shutter can reach 1 / 10 to 1 / 1666 seconds. The high-frequency microphone 24 uses a diaphragmless optical ultrasonic microphone to acquire acoustic signals during the processing of the component 8. It has an extremely wide detectable frequency range, reaching 50kHz to 2MHz. Its diaphragmless design eliminates moving parts, greatly reducing interference from machine resonance. The thermal imager 23 uses a short-wave infrared thermal imager to acquire... The temperature signal during the processing of component 8 has a high resolution of 640H×512V pixels and a detectable temperature range of 350℃~1800℃, meeting the monitoring requirements for arc-assisted manufacturing. The magneto-optical imager 22 uses a magnetic field camera based on the Faraday effect to collect the magnetic field strength signal and current distribution signal around component 8 during processing. Its resolution is extremely high, up to 15μm. The signal acquisition card 20 receives analog or digital signals from the high-speed camera 21, magneto-optical imager 22, thermal imager 23 and high-frequency microphone 24, performs noise reduction, noise point removal and normalization processing on the data, and then sends it to the signal processing system 19 to provide data accuracy.

[0050] The signal processing system 19 includes a data fusion module, a feature parameter extraction module, and a machine learning module. The data fusion module is used to process and organically fuse the received multi-source signals; the feature parameter extraction module is responsible for extracting relevant feature parameters from the processed multi-source information; the machine learning module receives the feature parameters, classifies them, and establishes a classification and discrimination model.

[0051] The sensing method using the above-mentioned device in this invention includes the following steps:

[0052] Step 1: Turn on each component of the electric arc additive manufacturing system in sequence. After each component has stabilized and reached normal operating status, turn on the high-frequency microphone 24 separately to collect the sound data of the equipment operation, which will provide a reference for subsequent judgment on whether the equipment is operating normally and whether there are any abnormalities.

[0053] Step 2, Model Building and Preliminary Numerical Simulation, includes the following steps:

[0054] Step 2.1: Use 3D modeling software to create a 3D model of the structural component and output it as an STL format file;

[0055] Step 2.2: Perform slicing using professional slicing software, divide the area and plan the path, and generate an LS format file after coordinate calibration;

[0056] Step 2.3: Use virtual simulation software to perform pre-numerical simulation, verify the processing trajectory, and finally output the LP format file to the arc additive manufacturing control system to start the printing operation.

[0057] Step 3: Collect multi-source information on the additive and subtractive manufacturing process; turn on the high-speed camera 21, magneto-optical imager 22 and thermal imager 23 respectively to collect multi-source information on the additive and subtractive manufacturing process.

[0058] Step 4, the printing process, includes the following steps:

[0059] Step 4.1: Perform coordinate calibration on the second robotic arm 12 and the worktable 9, and simulate the actual printing operation with reference to the numerical simulation results to confirm whether the printing trajectory is correct.

[0060] Step 4.2: With the printing trajectory correct, activate the water tank 13, welding wire extrusion device 14, metal wire ring 15, argon gas protection device 16, air purifier 25 and suction system 26 in sequence to ensure that each part works stably before officially starting the printing process.

[0061] Step 4.3: Start the second robotic arm 12 and the water-cooled argon arc welding gun 11 through the teaching box 17. The water-cooled argon arc welding gun 11 will form an arc between the substrates 10, and melt the metal wire to build up.

[0062] Step 4.4: After printing one layer according to the set path, the second robotic arm 12 lifts the water-cooled argon arc welding gun 11 away from the substrate 10 to pause printing. After the molten pool has cooled down, the second robotic arm 12 moves the water-cooled argon arc welding gun 11 to the arc-starting position according to the predetermined trajectory. This process is repeated until printing is completed.

[0063] Step 5, the subtractive processing, includes the following steps:

[0064] Step 5.1: Using the control panel 1, bring the milling cutter 6 close to the workpiece 8 to be machined, observe the contact between the milling cutter 6 and the workpiece 8, and then use the tool setting function on the control panel 1 to set the tool.

[0065] Step 5.2: After tool setting is completed, simulate the machining program in control panel 1 beforehand. Once confirmed to be correct, begin machining.

[0066] Step 6, Analysis and processing of multi-source information, including the following steps:

[0067] Step 6.1: After the signal acquisition card 20 receives multi-source information, it will use an edge detection algorithm to extract the weld edge and part contour features from the dynamic image information acquired by the high-speed camera 21; extract the peak value, valley value and current density distribution features of the magnetic field intensity from the magnetic field and current distribution data acquired by the magneto-optical imager 22; extract the mean, variance, hot spot location and temperature change rate features of the temperature distribution from the temperature field monitoring data acquired by the thermal imager 23; and use Fourier transform to convert the time domain features to frequency domain features from the acoustic monitoring data acquired by the high-frequency microphone 24, and extract the amplitude and frequency features.

[0068] Step 6.2: Signal acquisition card 20 filters and normalizes the raw data acquired in step 6.1; specifically, for the dynamic image data acquired by high-speed camera 21, filtering is performed, the median value is taken as the new value of the pixel, and normalization is performed, mapping the pixel value to [0, 1], using the formula: , where X is the original pixel value, and other data are also filtered and normalized to make the data smoother and easier to integrate with other data.

[0069] Step 7: After receiving the processed multi-source information, the signal processing system 19 uses linear discriminant analysis to perform machine learning classification and modeling on this information. The final extracted feature parameters of the modeling data are: weld edge features E, amplitude A, frequency F, mean M and variance V of temperature distribution, hot spot location H, temperature change rate R, peak magnetic field strength P, valley magnetic field strength G, and current distribution density D. Based on the above feature parameters, two discriminant functions are listed:

[0070] LDA1=w1·E+w2·A+w3·F+w4·M+w5·V+w6·H+w7·R+w8·P+w9·G+w 10 ·D

[0071] LDA2=w1·E+w2·A+w3·F+w4·M+w5·V+w6·H+w7·R+w8·P+w9·G+w 10 ·D

[0072] Among them, w1, w2, w3... are weight coefficients, which are determined by the training dataset.

[0073] The various embodiments described in this specification are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-source online sensing device for arc-assisted additive and subtractive manufacturing, characterized in that, The system includes a substrate (10), a robotic arm, an arc additive manufacturing forming system, a subtractive manufacturing system, a multi-source signal acquisition system, and a signal processing system (19). The substrate (10) has a rotatable base for supporting a worktable (9) to accommodate the additive and subtractive operations of the robotic arm. The robotic arm includes a first robotic arm (4) and a second robotic arm (12), which are used to perform subtractive processing and additive forming operations, respectively. The arc additive manufacturing forming system cooperates with the second robotic arm (12) to achieve the initial forming of complex structure metal parts. The subtractive manufacturing system cooperates with the first robotic arm (4) to perform high-precision processing on the initially formed metal parts. The multi-source signal acquisition system collects multi-source information in real time throughout the additive and subtractive manufacturing process and transmits it to the signal processing system (19). The signal processing system (19) performs fusion processing, feature parameter extraction, and machine learning discrimination on the multi-source information. The arc additive manufacturing forming system includes a first control system, a wire feeding system, an electric welding system, an atmosphere protection system, and a water tank (13); the wire feeding system includes a wire extrusion device (14) and a wire coil (15), which work together to deliver the wire to the electric welding system; the electric welding system includes a water-cooled argon arc welding gun (11) installed at the front end of the second robotic arm (12); the atmosphere protection system includes a suction system (26), an air purifier (25), and an argon protection device (16), the suction system (26) and the air purifier (25) are used to absorb and purify harmful gases generated during the processing, and the argon protection device (16) is used to release argon to provide a protective atmosphere; the first control system includes a control host (18) and a teach pendant (17), which are used to control the operation of each component of the arc additive manufacturing forming system; the water tank (13) cools each part of the arc additive manufacturing forming system through water pipes; The subtractive manufacturing system includes a second control system, a drive system, a machining system, and a cooling system; the drive system includes a powertrain (3) for driving the first robotic arm (4) and the machining system; the machining system includes a clamp (5) and a milling cutter (6) located at the end of the first robotic arm (4); the cooling system includes a reservoir (2) and a nozzle (7), the reservoir (2) contains cutting fluid, and the cutting fluid is delivered to the machining area through the nozzle (7) to achieve cooling and chip removal; the second control system includes a control panel (1) for uniformly controlling the operation of each component of the subtractive manufacturing system; The signal processing system (19) includes a data fusion module, a feature parameter extraction module, and a machine learning module; the data fusion module is used to organize and organically fuse the received multi-source signals; the feature parameter extraction module is used to extract relevant feature parameters from the fused multi-source information; The machine learning module is used to receive feature parameters and build a classification and discrimination model.

2. The multi-source online sensing device for arc-assisted additive and subtractive manufacturing according to claim 1, characterized in that, The multi-source signal acquisition system includes a signal acquisition card (20), a high-speed camera (21), a magneto-optical imager (22), a thermal imager (23), and a high-frequency microphone (24). The high-speed camera (21) is used to acquire image signals of the component to be processed (8), the high-frequency microphone (24) is used to acquire sound signals during the processing, the thermal imager (23) is used to acquire temperature signals during the processing, and the magneto-optical imager (22) is used to acquire magnetic field strength signals and current distribution signals during the processing. The signal acquisition card (20) receives analog or digital signals output by each sensor, and transmits them to the signal processing system (19) after noise reduction, noise point removal, and normalization.

3. A method for online sensing of multi-source information in arc-assisted manufacturing based on the device described in claim 1, characterized in that, Includes the following steps: S1: Turn on each component of the electric arc additive manufacturing forming system. After each component is running stably, turn on the high-frequency microphone (24) separately to collect the sound data of the equipment operation as a reference. S2: Establish a 3D model of the structural component and perform pre-numerical simulation to generate a machining execution file; S3: Turn on the high-speed camera (21), magneto-optical imager (22) and thermal imager (23) to collect multi-source information on the additive and subtractive manufacturing process in real time; S4: Based on the processing execution document, the initial forming of the metal part is completed through the arc additive manufacturing forming system; S5: High-precision machining of pre-formed metal parts using a subtractive process; S6: The signal acquisition card (20) receives multi-source information and performs filtering and normalization processing, and then transmits the processed information to the signal processing system (19). S7: The signal processing system (19) fuses the processed multi-source information, extracts feature parameters, establishes a classification and discrimination model through machine learning, and realizes online perception of the material addition and subtraction manufacturing process.

4. The method for online sensing of multi-source information in arc-assisted additive and subtractive manufacturing according to claim 3, characterized in that, Step S2 specifically includes: S2.1: Use 3D modeling software to create 3D models of structural components and output them as STL format files; S2.2: The STL format file is sliced, divided into regions, and path planned using slicing software. After coordinate calibration, an LS format file is generated. S2.3: Use virtual simulation software to perform pre-numerical simulation verification of the machining trajectory, and finally output the LP format file to the arc additive manufacturing forming system.

5. The method for online sensing of multi-source information in arc-assisted additive and subtractive manufacturing according to claim 3, characterized in that, Step S4 specifically includes: S4.1: Perform coordinate calibration on the second robotic arm (12) and the worktable (9), and verify the actual printing trajectory by referring to the numerical simulation results; S4.2: Activate the water tank (13), welding wire extrusion device (14), metal wire ring (15), argon gas protection device (16), air purifier (25) and suction system (26) in sequence, and start printing after each component is working stably; S4.3: The second robotic arm (12) and the water-cooled argon arc welding gun (11) are started by the teaching box (17). An electric arc is formed between the water-cooled argon arc welding gun (11) and the substrate (10), and the molten metal wire is stacked into shape. S4.4: After printing one layer, the second robotic arm (12) drives the water-cooled argon arc welding gun (11) to lift away from the substrate (10) and pause printing. After the molten pool cools down, it moves to the arc-starting position to continue printing. The cycle continues until the initial forming is completed.

6. The method for online sensing of multi-source information in arc-assisted additive and subtractive manufacturing according to claim 3, characterized in that, Step S5 specifically includes: S5.1: Using the control panel (1), bring the milling cutter (6) close to the workpiece (8) to be processed, observe the contact situation, and then use the tool setting function of the control panel (1) to complete the tool setting; S5.2: Simulate the machining program in advance and start the subtractive machining process after confirming that there are no errors.

7. The online sensing method for multi-source information in arc-assisted additive and subtractive manufacturing according to claim 3, characterized in that, The feature extraction processing of multi-source information in step S6 includes: using an edge detection algorithm to extract weld edge and part contour features from the dynamic image information acquired by the high-speed camera (21); extracting the peak value, valley value and current density distribution features from the magnetic field and current distribution data acquired by the magneto-optical imager (22); extracting the mean, variance, hot spot location and temperature change rate features from the temperature field data acquired by the thermal imager (23); and converting the acoustic data acquired by the high-frequency microphone (24) into frequency domain features through Fourier transform to extract amplitude and frequency features.

8. The method for online sensing of multi-source information in arc-assisted additive and subtractive manufacturing according to claim 3, characterized in that, The dynamic image data acquired by the high-speed camera (21) is filtered, and the median value is taken as the new value of the pixel. Then, the pixel value is mapped to the interval [0, 1] for normalization. The formula is as follows: Where X is the original pixel value.

9. The online sensing method for multi-source information in arc-assisted additive and subtractive manufacturing according to claim 3, characterized in that, The feature parameters extracted in step S7 include weld edge features E, amplitude A, frequency F, mean temperature distribution M, variance V, hot spot location H, temperature change rate R, peak magnetic field strength P, valley magnetic field strength G, and current distribution density D.

10. The online sensing method for multi-source information in arc-assisted additive and subtractive manufacturing according to claim 3, characterized in that, The machine learning method employs linear discriminant analysis, and the established discriminant function is as follows: LDA1=w1·E+w2·A+w3·F+w4·M+w5·V+w6·H+w7·R+w8·P+w9·G+w 10 ·D LDA2 = w1·E+w2·A+w3·F+w4·M+w5·V+w6·H+w7·R+w8·P+w9·G+w 10 ·D Where w1, w2, w3... are weighting coefficients.

Images