New energy aluminum alloy precision casting multi-mode quality monitoring device and method

Abstract

The invention relates to the technical field of new energy aluminum alloy precision casting multi-modal quality monitoring, in particular to a new energy aluminum alloy precision casting multi-modal quality monitoring device and method. According to the technical scheme, the new energy aluminum alloy precision casting multi-mode quality monitoring device comprises a quality monitoring device body, an infrared thermal imager probe, a germanium single crystal protection lens, an integrated spectrometer probe debugging port, an acoustic detection microphone array and a honeycomb acoustic shielding cover. An infrared thermal imager probe is arranged on one side of the quality monitoring device main body, and a germanium single crystal protection lens is arranged on the outer side of the infrared thermal imager probe; synchronous acquisition and fusion analysis of multi-physical field data are realized by integrating six modes of spectrum, infrared thermal imaging, vibration, machine vision, acoustics and pressure deformation, and meanwhile, when the shrinkage risk is predicted by the infrared thermal imaging, a die-casting process compensation mechanism is automatically triggered to kill defects in the germination stage, so that the defect rate is greatly reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of multimodal quality monitoring of new energy aluminum alloy precision casting, and in particular to a multimodal quality monitoring device and method for new energy aluminum alloy precision casting. Background Art

[0002] The multimodal quality monitoring device for new energy aluminum alloy precision casting is an intelligent quality control system designed specifically for the high-precision aluminum alloy casting process in the new energy field. Its core lies in achieving defect prevention and quality improvement throughout the entire casting chain through multimodal data fusion and intelligent analysis.

[0003] Existing multimodal quality monitoring methods for new energy aluminum alloy precision castings often rely on a single technology, lack the ability to comprehensively evaluate the multi-dimensional quality of castings, and are prone to missing potential defects. At the same time, most traditional inspections are destructive sampling inspections, or can only be performed after casting is completed, and cannot achieve real-time online monitoring of the casting process and immediate process intervention.

[0004] To address the problems that existing multimodal quality monitoring methods for new energy aluminum alloy precision casting lack the ability to comprehensively evaluate the multi-dimensional quality of castings, are prone to missing potential defects, and are unable to achieve real-time online monitoring and immediate process intervention of the casting process, this solution integrates six major modalities: spectroscopy, infrared thermal imaging, vibration, machine vision, acoustics, and pressure deformation, to achieve synchronous collection and fusion analysis of multi-physical field data. At the same time, when infrared thermal imaging predicts the risk of shrinkage, it automatically triggers the die-casting process compensation mechanism to nip defects in the bud, reducing the defect rate by more than 40%. Summary of the Invention

[0005] In order to overcome the existing multimodal quality monitoring methods for new energy aluminum alloy precision casting, they often rely on a single technology, lack the ability to comprehensively evaluate the multi-dimensional quality of castings, and are prone to missing potential defects. At the same time, most traditional inspections are destructive sampling inspections, or can only be performed after casting is completed, and cannot achieve real-time online monitoring of the casting process and immediate process intervention.

[0006] The technical solution of the present invention is: a multimodal quality monitoring device for precision casting of new energy aluminum alloys, including a quality monitoring device body, an infrared thermal imager probe, a germanium single crystal protective lens, an integrated spectrometer probe debugging port, an acoustic detection microphone array and a honeycomb acoustic shielding cover. An infrared thermal imager probe is provided on one side of the quality monitoring device body, a germanium single crystal protective lens is provided on the outside of the infrared thermal imager probe, an integrated spectrometer probe debugging port is provided on one side of the infrared thermal imager probe, an acoustic detection microphone array is provided on one side of the quality monitoring device body, and a honeycomb acoustic shielding cover is provided on the outside of the acoustic detection microphone array.

[0007] Preferably, infrared radiation data of the mold and casting surface during the die-casting process are collected in real time by an infrared thermal imager probe to generate a temperature field distribution image, and a germanium single crystal protective lens is used to protect the infrared thermal imager probe from high-temperature metal splashing and mechanical impact, while ensuring efficient transmission of infrared radiation. A quick connection and debugging interface for the spectrometer probe is provided through an integrated spectrometer probe debugging port, and sapphire fiber coupling is supported, which facilitates replacement of spectral analysis modules of different wavelengths according to detection requirements. The sound field signal of the casting in the acoustic resonance cavity is collected through an acoustic detection microphone array, and the acoustic detection module is covered by a honeycomb acoustic shielding cover. The environmental noise is absorbed by the honeycomb structure, while ensuring the acoustic permeability of the acoustic detection microphone array.

[0008] Preferably, a capacitive touch screen is provided on the surface of the main body of the quality monitoring device, and a status indicator light strip is provided on one side of the capacitive touch screen.

[0009] Preferably, a heat dissipation unit is provided on the other side of the main body of the quality monitoring device, and multiple groups of heat dissipation units are provided. An integrated connection circuit board is provided on the other side of the main body of the quality monitoring device. Support pads are provided at the four corners of the bottom surface of the main body of the quality monitoring device. A storage box is provided on the bottom surface of the main body of the quality monitoring device, and an elastic clamp is provided inside the storage box.

[0010] The multimodal quality monitoring method for new energy aluminum alloy precision casting includes the following steps:

[0011] S101: First, the spectral composition, temperature gradient, and gas content of the molten alloy are collected in real time, and the slag morphology is identified through AI;

[0012] S102: Establish a mold temperature field model, perform zoned induction heating and compensate for thermal stress, and combine machine vision to pre-inspect surface quality to provide a stable cavity environment for high-quality casting;

[0013] S103: Coordinated control of filling speed and pressure, and monitoring of acoustic emission signals and infrared temperature fields during the solidification process;

[0014] S104: Adjust heat treatment parameters based on the prediction model, detect hardness gradient and residual stress in real time, and intelligently classify grain size;

[0015] S105: Align the spatiotemporal coordinates of heterogeneous data, mine defect pattern association rules, train quality prediction models, and optimize process parameters in a closed-loop manner;

[0016] S106: Blockchain stores production data, generates three-dimensional quality fingerprints, supports reverse tracing of failure modes, and regularly analyzes traceability data to promote process iteration and upgrades.

[0017] Preferably, when performing dynamic monitoring of multiple parameters during the smelting process, the following steps are included:

[0018] S201: The plasma emission spectrum of the molten alloy is collected every 30 seconds through a sapphire fiber-coupled integrated spectrometer;

[0019] S202: The spectral data is processed by a filtering algorithm to remove electromagnetic interference from the furnace and generate a trend graph of element concentration changes;

[0020] S203: When it is detected that the concentration of the target element deviates from the preset threshold by ±2%, an audible and visual alarm is triggered and the feeding mechanism is automatically adjusted;

[0021] S204: The infrared thermal imager scans the molten pool surface at a frame rate of 200 Hz to construct a three-dimensional temperature field model;

[0022] S205: Calculate the temperature difference inside the melt using a temperature gradient algorithm. If the local temperature difference exceeds 50°C, it is determined to be temperature uneven. The control system activates the electromagnetic stirring device and adjusts the stirring frequency to 80-120 Hz to homogenize the temperature.

[0023] S206: The machine vision system collects images of the molten pool surface, uses the U-Net neural network to segment the slag area, analyzes the slag area ratio and shape factor, and activates the slag removal mechanism when an abnormality occurs;

[0024] S207: The inert gas sensor monitors the oxygen content above the molten pool in real time and controls the argon flow rate to maintain the oxygen content < 0.05%;

[0025] S208: Use differential absorption spectroscopy to detect hydrogen content in the melt, and automatically increase the refining temperature when it exceeds the standard.

[0026] Preferably, when performing intelligent control of mold preheating, the following steps are included:

[0027] S301: Scan the mold surface with an infrared thermal imager, establish a finite element mesh model, predict the temperature distribution in different areas, compare it with the preset optimal preheating curve, and generate a heating plan;

[0028] S302: Divide the mold into six heating zones, independently control the power of the medium-frequency induction heating power supply, and use the PID algorithm to dynamically adjust the heating parameters so that the temperature in each zone reaches the target value synchronously;

[0029] S303: Monitor the thermal expansion of the mold during preheating using a vibration sensor, calculate the thermal stress distribution, and control the hydraulic clamping device to apply a reverse compensation force to avoid mold deformation;

[0030] S304: The machine vision system detects surface defects in the mold cavity, automatically marks the defect location and generates a repair path to guide subsequent manual processing.

[0031] Preferably, when performing multimodal coupling monitoring of the die casting process, the following steps are included:

[0032] S401: The die-casting machine's built-in pressure sensor and displacement sensor synchronously collect data to construct a PV curve for the filling process;

[0033] S402: Adjust the injection speed in real time through the model predictive control algorithm to ensure smooth flow of molten metal;

[0034] S403: Acoustic detection microphone array collects acoustic emission signals during metal solidification in the frequency range of 20-200kHz, and uses wavelet packet transform to analyze the signal energy distribution and identify shrinkage or crack defect characteristics;

[0035] S404: The infrared thermal imager continuously monitors the temperature gradient changes on the casting surface to determine the position of the solidification front. When the solidification progress deviates from the process window, the mold cooling water flow rate is adjusted;

[0036] S405: Deploy a line laser sensor at the die casting exit to scan the casting contour and generate a 3D point cloud, which is then compared with the CAD model to detect key dimensions.

[0037] Preferably, when performing intelligent optimization of the heat treatment process, the following steps are included:

[0038] S501: Establishing a neural network model of heat treatment process and metallographic structure based on historical data;

[0039] S502: Use a portable Leeb hardness tester to scan different areas of the casting, generate a hardness distribution heat map, compare it with the target hardness range, and mark the areas that require secondary heat treatment;

[0040] S503: Use X-ray diffractometer to measure the residual stress on the casting surface and combine it with finite element analysis to predict the internal stress distribution;

[0041] S504: When the residual stress exceeds 30% of the material yield strength, the aging treatment parameters are automatically adjusted;

[0042] S505: The machine vision system collects metallographic photos, calculates the grain size distribution through image segmentation algorithm, and automatically grades according to ASTM E112 standard to ensure that the grain size meets the -3 grade requirement.

[0043] Preferably, when performing multimodal data fusion analysis, the following steps are included:

[0044] S601: Align the spectrum, temperature, vibration and acoustic time series data using a dynamic time warping algorithm to establish a unified space-time coordinate system;

[0045] S602: Use the Apriori algorithm to analyze the association rules between defect characteristics and process parameters, generate a defect causal network diagram, and identify key influencing factors;

[0046] S603: Use long short-term memory networks to fuse multimodal data and build a casting quality prediction model, which outputs defect probability;

[0047] S604: Based on the quality prediction results, the genetic algorithm generates an optimized process plan, which is automatically deployed to the production end after virtual verification by the digital twin system.

[0048] As a preference, the following steps are included when conducting full life cycle quality traceability:

[0049] S701: Packaging the multimodal inspection data, process parameters, and equipment status information of each casting into a blockchain node;

[0050] S702: Use SHA-256 hash algorithm to encrypt and ensure that data cannot be tampered with;

[0051] S703: Integrates CT scan data and metallographic analysis results to build a 3D quality fingerprint database for castings, supporting quick retrieval by batch, furnace number, and mold number;

[0052] S704: When a casting fails during its service life, the original production data is queried through the blockchain, combined with failure analysis, to reversely locate the defect formation link;

[0053] S705: Regularly analyze quality traceability data, generate process improvement reports, and push improvement plans to the PLM system.

[0054] Beneficial effects of the present invention:

[0055] 1. Compared with traditional multimodal quality monitoring methods for new energy aluminum alloy precision casting, which often rely on a single technology and lack the ability to comprehensively evaluate the multi-dimensional quality of castings, and are prone to missing potential defects, this solution integrates six major modalities: spectroscopy, infrared thermal imaging, vibration, machine vision, acoustics, and pressure deformation, to achieve simultaneous collection and fusion analysis of multi-physics field data. During the casting process, it simultaneously monitors melt composition, mold temperature field, casting structural vibration characteristics, surface defects, internal structure uniformity, and material mechanical properties, forming a comprehensive assessment of casting quality.

[0056] 2. Compared with traditional multimodal quality monitoring methods for new energy aluminum alloy precision casting, most traditional inspections are destructive spot checks or can only be performed after casting is completed. They cannot achieve real-time online monitoring of the casting process and immediate process intervention. This solution uses edge computing and digital twin technology to achieve real-time online monitoring. When spectral analysis finds fluctuations in the melt composition, parameter adjustment instructions can be immediately sent to the die-casting machine through the CAN bus; when infrared thermal imaging predicts the risk of shrinkage, the die-casting process compensation mechanism is automatically triggered to nip defects in the bud, and the defect rate can be reduced by more than 40%. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 Shown is a first three-dimensional structural schematic diagram of the multi-modal quality monitoring device for new energy aluminum alloy precision casting of the present invention;

[0058] Figure 2 Shown is a second three-dimensional structural schematic diagram of the multi-modal quality monitoring device for precision casting of new energy aluminum alloys according to the present invention;

[0059] Figure 3 Shown is a schematic diagram of the bottom three-dimensional structure of the multi-modal quality monitoring device for precision casting of new energy aluminum alloys of the present invention;

[0060] Figure 4 Shown is a schematic diagram of the workflow of the multimodal quality monitoring method for precision casting of new energy aluminum alloys of the present invention;

[0061] Explanation of the accompanying symbols: 1. Quality monitoring device body; 201. Infrared thermal imager probe; 202. Germanium single crystal protection lens; 203. Integrated spectrometer probe debugging port; 204. Acoustic detection microphone array; 205. Honeycomb acoustic shielding cover; 301. Capacitive touch screen; 302. Status indicator light strip; 401. Heat dissipation unit; 402. Integrated connection circuit board; 403. Support foot pad; 404. Storage box; 405. Elastic clamp. DETAILED DESCRIPTION

[0062] The present invention will be further described below with reference to the accompanying drawings and examples.

[0063] See also Figure 1-3 The present invention provides an embodiment: a multimodal quality monitoring device for precision casting of new energy aluminum alloys, comprising a quality monitoring device body 1, an infrared thermal imager probe 201, a germanium single crystal protective lens 202, an integrated spectrometer probe debugging port 203, an acoustic detection microphone array 204 and a honeycomb acoustic shielding cover 205. The infrared thermal imager probe 201 is provided on one side of the quality monitoring device body 1, a germanium single crystal protective lens 202 is provided on the outer side of the infrared thermal imager probe 201, an integrated spectrometer probe debugging port 203 is provided on one side of the infrared thermal imager probe 201, an acoustic detection microphone array 204 is provided on one side of the quality monitoring device body 1, and a honeycomb acoustic shielding cover 205 is provided on the outer side of the acoustic detection microphone array 204.

[0064] Preferably, the surface of the main body 1 of the quality monitoring device is provided with a capacitive touch screen 301, and a status indicator light strip 302 is provided on one side of the capacitive touch screen 301. When in use, the multimodal detection data is displayed in real time through the capacitive touch screen 301, and the process parameter adjustment and early warning functions are integrated to provide an intuitive operation interface. The status indicator light strip 302 provides red, yellow and green indicator lights to intuitively display the equipment status.

[0065] Preferably, a heat dissipation unit 401 is provided on the other side of the quality monitoring device body 1, and multiple groups of heat dissipation units 401 are provided. An integrated connection circuit board 402 is provided on the other side of the quality monitoring device body 1, and support pads 403 are provided at the four corners of the bottom surface of the quality monitoring device body 1. A storage box 404 is provided on the bottom surface of the quality monitoring device body 1, and an elastic clamp 405 is provided inside the storage box 404. When in use, the heat dissipation unit 401 is used to dissipate heat inside the quality monitoring device body 1 to ensure that the internal temperature of the quality monitoring device body 1 remains stable when the device is running at high load for a long time. High-speed data transmission between each detection module and the central control unit is achieved through the integrated connection circuit board 402, and hot plug function is supported to facilitate module expansion and maintenance. The support pads 403 provide stable support between the equipment and the ground, while absorbing vibration and reducing the impact of external interference on detection accuracy. The detection cables are stored through the storage box 404, and different types of detection cables are fixed through the elastic clamp 405 to prevent the cables from loosening or wearing.

[0066] See also Figure 4 In this embodiment, the multimodal quality monitoring method for new energy aluminum alloy precision casting includes the following steps:

[0067] S101: First, the spectral composition, temperature gradient, and gas content of the molten alloy are collected in real time, and the slag morphology is identified through AI;

[0068] S102: Establish a mold temperature field model, perform zoned induction heating and compensate for thermal stress, and combine machine vision to pre-inspect surface quality to provide a stable cavity environment for high-quality casting;

[0069] S103: Coordinated control of filling speed and pressure, and monitoring of acoustic emission signals and infrared temperature fields during the solidification process;

[0070] S104: Adjust heat treatment parameters based on the prediction model, detect hardness gradient and residual stress in real time, and intelligently classify grain size;

[0071] S105: Align the spatiotemporal coordinates of heterogeneous data, mine defect pattern association rules, train quality prediction models, and optimize process parameters in a closed-loop manner;

[0072] S106: Blockchain stores production data, generates three-dimensional quality fingerprints, supports reverse tracing of failure modes, and regularly analyzes traceability data to promote process iteration and upgrades.

[0073] Preferably, when performing dynamic monitoring of multiple parameters during the smelting process, the following steps are included:

[0074] S201: The plasma emission spectrum of the molten alloy is collected every 30 seconds through a sapphire fiber-coupled integrated spectrometer;

[0075] S202: The spectral data is processed by a filtering algorithm to remove electromagnetic interference from the furnace and generate a trend graph of element concentration changes;

[0076] S203: When it is detected that the concentration of the target element deviates from the preset threshold by ±2%, an audible and visual alarm is triggered and the feeding mechanism is automatically adjusted;

[0077] S204: The infrared thermal imager scans the molten pool surface at a frame rate of 200 Hz to construct a three-dimensional temperature field model;

[0078] S205: Calculate the temperature difference inside the melt using a temperature gradient algorithm. If the local temperature difference exceeds 50°C, it is determined to be temperature uneven. The control system activates the electromagnetic stirring device and adjusts the stirring frequency to 80-120 Hz to homogenize the temperature.

[0079] S206: The machine vision system collects images of the molten pool surface, uses the U-Net neural network to segment the slag area, analyzes the slag area ratio and shape factor, and activates the slag removal mechanism when an abnormality occurs;

[0080] S207: The inert gas sensor monitors the oxygen content above the molten pool in real time and controls the argon flow rate to maintain the oxygen content < 0.05%;

[0081] S208: Use differential absorption spectroscopy to detect hydrogen content in the melt, and automatically increase the refining temperature when it exceeds the standard.

[0082] Preferably, when performing intelligent control of mold preheating, the following steps are included:

[0083] S301: Scan the mold surface with an infrared thermal imager, establish a finite element mesh model, predict the temperature distribution in different areas, compare it with the preset optimal preheating curve, and generate a heating plan;

[0084] S302: Divide the mold into six heating zones, independently control the power of the medium-frequency induction heating power supply, and use the PID algorithm to dynamically adjust the heating parameters so that the temperature in each zone reaches the target value synchronously;

[0085] S303: Monitor the thermal expansion of the mold during preheating using a vibration sensor, calculate the thermal stress distribution, and control the hydraulic clamping device to apply a reverse compensation force to avoid mold deformation;

[0086] S304: The machine vision system detects surface defects in the mold cavity, automatically marks the defect location and generates a repair path to guide subsequent manual processing.

[0087] Preferably, when performing multimodal coupling monitoring of the die casting process, the following steps are included:

[0088] S401: The die-casting machine's built-in pressure sensor and displacement sensor synchronously collect data to construct a PV curve for the filling process;

[0089] S402: Adjust the injection speed in real time through the model predictive control algorithm to ensure smooth flow of molten metal;

[0090] S403: Acoustic detection microphone array collects acoustic emission signals during metal solidification in the frequency range of 20-200kHz, and uses wavelet packet transform to analyze the signal energy distribution and identify shrinkage or crack defect characteristics;

[0091] S404: The infrared thermal imager continuously monitors the temperature gradient changes on the casting surface to determine the position of the solidification front. When the solidification progress deviates from the process window, the mold cooling water flow rate is adjusted;

[0092] S405: Deploy a line laser sensor at the die casting exit to scan the casting contour and generate a 3D point cloud, which is then compared with the CAD model to detect key dimensions.

[0093] Preferably, when performing intelligent optimization of the heat treatment process, the following steps are included:

[0094] S501: Establishing a neural network model of heat treatment process and metallographic structure based on historical data;

[0095] S502: Use a portable Leeb hardness tester to scan different areas of the casting, generate a hardness distribution heat map, compare it with the target hardness range, and mark the areas that require secondary heat treatment;

[0096] S503: Use X-ray diffractometer to measure the residual stress on the casting surface and combine it with finite element analysis to predict the internal stress distribution;

[0097] S504: When the residual stress exceeds 30% of the material yield strength, the aging treatment parameters are automatically adjusted;

[0098] S505: The machine vision system collects metallographic photos, calculates the grain size distribution through image segmentation algorithm, and automatically grades according to ASTM E112 standard to ensure that the grain size meets the -3 grade requirement.

[0099] Preferably, when performing multimodal data fusion analysis, the following steps are included:

[0100] S601: Align the spectrum, temperature, vibration and acoustic time series data using a dynamic time warping algorithm to establish a unified space-time coordinate system;

[0101] S602: Use the Apriori algorithm to analyze the association rules between defect characteristics and process parameters, generate a defect causal network diagram, and identify key influencing factors;

[0102] S603: Use long short-term memory networks to fuse multimodal data and build a casting quality prediction model, which outputs defect probability;

[0103] S604: Based on the quality prediction results, the genetic algorithm generates an optimized process plan, which is automatically deployed to the production end after virtual verification by the digital twin system.

[0104] As a preference, the following steps are included when conducting full life cycle quality traceability:

[0105] S701: Packaging the multimodal inspection data, process parameters, and equipment status information of each casting into a blockchain node;

[0106] S702: Use SHA-256 hash algorithm to encrypt and ensure that data cannot be tampered with;

[0107] S703: Integrates CT scan data and metallographic analysis results to build a 3D quality fingerprint database for castings, supporting quick retrieval by batch, furnace number, and mold number;

[0108] S704: When a casting fails during its service life, the original production data is queried through the blockchain, combined with failure analysis, to reversely locate the defect formation link;

[0109] S705: Regularly analyze quality traceability data, generate process improvement reports, and push improvement plans to the PLM system.

[0110] Example 1

[0111] Implementation Background: This method was implemented on the aluminum alloy engine cylinder head production line of an automotive parts manufacturer. The traditional manual sampling and single-point temperature monitoring solution was compared with the multimodal quality monitoring device and method described in this article. The production cycle was 3 months, and the sample size was 1,000 pieces each.

[0112] Implementation equipment and methods:

[0113] Control group:

[0114] 1. Handheld infrared thermometer.

[0115] 2. Portable hardness tester.

[0116] 3. Manual visual inspection.

[0117] 4. Inspection frequency: 10 pieces per batch.

[0118] Experimental group:

[0119] 1. Multimodal quality monitoring device for new energy aluminum alloy precision casting.

[0120] 2. Real-time monitoring of the entire process, automatic storage and analysis of data.

[0121] 3. Automatically generate quality reports and process optimization suggestions.

[0122] Comparison data table:

[0123]

[0124] Implementation effect:

[0125] 1. The experimental group effectively identified shrinkage-prone areas through online detection of hydrogen content during the smelting process and acoustic emission monitoring during the solidification process, reducing porosity by 61.9%.

[0126] 2. Multimodal coupling monitoring of the die-casting process enables precise coordinated control of filling speed and pressure, reducing dimensional deviation from 0.12mm to 0.03mm.

[0127] 3. Intelligent control of mold preheating (zoned induction heating + thermal stress compensation) makes the residual stress distribution of castings more uniform, reducing the maximum value by 34.6%, and extending the fatigue life of parts.

[0128] 4. The quality prediction model constructed by multimodal data fusion analysis has a defect detection rate of 98.7%, avoiding the risk of missed detections in manual sampling.

[0129] Conclusion: The multimodal quality monitoring device and method for new energy aluminum alloy precision casting significantly improves the casting quality stability and production efficiency through multi-physical field coupling monitoring, intelligent data analysis and closed-loop optimization, providing key technical support for the manufacturing of lightweight components for new energy vehicles.

[0130] The embodiments of the present invention are described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Various changes can be made within the scope of knowledge of those skilled in the art without departing from the spirit of the present invention.

Claims

1. New energy aluminum alloy precision casting multi-modal quality monitoring device; its characteristics are: The device comprises a quality monitoring device body (1), an infrared thermal imager probe (201), a germanium single crystal protective lens (202), an integrated spectrometer probe debugging port (203), an acoustic detection microphone array (204), and a honeycomb acoustic shielding cover (205). The infrared thermal imager probe (201) is provided on one side of the quality monitoring device body (1), the germanium single crystal protective lens (202) is provided on the outer side of the infrared thermal imager probe (201), the integrated spectrometer probe debugging port (203) is provided on one side of the infrared thermal imager probe (201), the acoustic detection microphone array (204) is provided on one side of the quality monitoring device body (1), and the honeycomb acoustic shielding cover (205) is provided on the outer side of the acoustic detection microphone array (204).

2. The multi-modal quality monitoring device for new energy aluminum alloy precision casting according to claim 1 is characterized in that: The surface of the quality monitoring device body (1) is provided with a capacitive touch screen (301), and a status indicator light strip (302) is provided on one side of the capacitive touch screen (301).

3. The multi-modal quality monitoring device for new energy aluminum alloy precision casting according to claim 1 is characterized in that: A heat dissipation unit (401) is provided on the other side of the quality monitoring device body (1), and a plurality of heat dissipation units (401) are provided. An integrated connection circuit board (402) is provided on the other side of the quality monitoring device body (1). Supporting foot pads (403) are provided at the four corners of the bottom surface of the quality monitoring device body (1). A storage box (404) is provided on the bottom surface of the quality monitoring device body (1), and an elastic clamping block (405) is provided inside the storage box (404).

4. A multimodal quality monitoring method for new energy aluminum alloy precision casting includes the following steps: S101: First, the spectral composition, temperature gradient, and gas content of the molten alloy are collected in real time, and the slag morphology is identified through AI; S102: Establish a mold temperature field model, perform zoned induction heating and compensate for thermal stress, and combine machine vision to pre-inspect surface quality to provide a stable cavity environment for high-quality casting; S103: Coordinated control of filling speed and pressure, and monitoring of acoustic emission signals and infrared temperature fields during the solidification process; S104: Adjust heat treatment parameters based on the prediction model, detect hardness gradient and residual stress in real time, and intelligently classify grain size; S105: Align the spatiotemporal coordinates of heterogeneous data, mine defect pattern association rules, train quality prediction models, and optimize process parameters in a closed-loop manner; S106: Blockchain stores production data, generates three-dimensional quality fingerprints, supports reverse tracing of failure modes, and regularly analyzes traceability data to promote process iteration and upgrades.

5. The multimodal quality monitoring method for new energy aluminum alloy precision casting according to claim 4 is characterized in that: When conducting dynamic monitoring of multiple parameters during the smelting process, the following steps are included: S201: The plasma emission spectrum of the molten alloy is collected every 30 seconds through a sapphire fiber-coupled integrated spectrometer; S202: The spectral data is processed by a filtering algorithm to remove electromagnetic interference from the furnace and generate a trend graph of element concentration changes; S203: When it is detected that the concentration of the target element deviates from the preset threshold by ±2%, an audible and visual alarm is triggered and the feeding mechanism is automatically adjusted; S204: The infrared thermal imager scans the molten pool surface at a frame rate of 200 Hz to construct a three-dimensional temperature field model; S205: Calculate the temperature difference inside the melt using a temperature gradient algorithm. If the local temperature difference exceeds 50°C, it is determined to be temperature uneven. The control system activates the electromagnetic stirring device and adjusts the stirring frequency to 80-120 Hz to homogenize the temperature. S206: The machine vision system collects images of the molten pool surface, uses the U-Net neural network to segment the slag area, analyzes the slag area ratio and shape factor, and activates the slag removal mechanism when an abnormality occurs; S207: The inert gas sensor monitors the oxygen content above the molten pool in real time and controls the argon flow rate to maintain the oxygen content < 0.05%; S208: Use differential absorption spectroscopy to detect hydrogen content in the melt, and automatically increase the refining temperature when it exceeds the standard.

6. The multimodal quality monitoring method for new energy aluminum alloy precision casting according to claim 4 is characterized in that: When performing intelligent control of mold preheating, the following steps are included: S301: Scan the mold surface with an infrared thermal imager, establish a finite element mesh model, predict the temperature distribution in different areas, compare it with the preset optimal preheating curve, and generate a heating plan; S302: Divide the mold into six heating zones, independently control the power of the medium-frequency induction heating power supply, and use the PID algorithm to dynamically adjust the heating parameters so that the temperature in each zone reaches the target value synchronously; S303: Monitor the thermal expansion of the mold during preheating using a vibration sensor, calculate the thermal stress distribution, and control the hydraulic clamping device to apply a reverse compensation force to avoid mold deformation; S304: The machine vision system detects surface defects in the mold cavity, automatically marks the defect location and generates a repair path to guide subsequent manual processing.

7. The multimodal quality monitoring method for new energy aluminum alloy precision casting according to claim 4 is characterized in that: The following steps are included in the multimodal coupling monitoring of the die casting process: S401: The die-casting machine's built-in pressure sensor and displacement sensor synchronously collect data to construct a PV curve for the filling process; S402: Adjust the injection speed in real time through the model predictive control algorithm to ensure smooth flow of molten metal; S403: Acoustic detection microphone array collects acoustic emission signals during metal solidification in the frequency range of 20-200kHz, and uses wavelet packet transform to analyze the signal energy distribution and identify shrinkage or crack defect characteristics; S404: The infrared thermal imager continuously monitors the temperature gradient changes on the casting surface to determine the position of the solidification front. When the solidification progress deviates from the process window, the mold cooling water flow rate is adjusted; S405: Deploy a line laser sensor at the die casting exit to scan the casting contour and generate a 3D point cloud, which is then compared with the CAD model to detect key dimensions.

8. The multimodal quality monitoring method for new energy aluminum alloy precision casting according to claim 4 is characterized in that: When performing intelligent optimization of heat treatment process, the following steps are included: S501: Establishing a neural network model of heat treatment process and metallographic structure based on historical data; S502: Use a portable Leeb hardness tester to scan different areas of the casting, generate a hardness distribution heat map, compare it with the target hardness range, and mark the areas that require secondary heat treatment; S503: Use X-ray diffractometer to measure the residual stress on the casting surface and combine it with finite element analysis to predict the internal stress distribution; S504: When the residual stress exceeds 30% of the material yield strength, the aging treatment parameters are automatically adjusted; S505: The machine vision system collects metallographic photos, calculates the grain size distribution through image segmentation algorithm, and automatically grades according to ASTM E112 standard to ensure that the grain size meets the -3 grade requirement.

9. The multimodal quality monitoring method for new energy aluminum alloy precision casting according to claim 4, characterized in that: When performing multimodal data fusion analysis, the following steps are included: S601: Align the spectrum, temperature, vibration and acoustic time series data using a dynamic time warping algorithm to establish a unified space-time coordinate system; S602: Use the Apriori algorithm to analyze the association rules between defect characteristics and process parameters, generate a defect causal network diagram, and identify key influencing factors; S603: Use long short-term memory networks to fuse multimodal data and build a casting quality prediction model, which outputs defect probability; S604: Based on the quality prediction results, the genetic algorithm generates an optimized process plan, which is automatically deployed to the production end after virtual verification by the digital twin system.

10. The multimodal quality monitoring method for new energy aluminum alloy precision casting according to claim 4, characterized in that: When conducting full life cycle quality traceability, the following steps are included: S701: Packaging the multimodal inspection data, process parameters, and equipment status information of each casting into a blockchain node; S702: Use SHA-256 hash algorithm to encrypt and ensure that data cannot be tampered with; S703: Integrates CT scan data and metallographic analysis results to build a 3D quality fingerprint database for castings, supporting quick retrieval by batch, furnace number, and mold number; S704: When a casting fails during its service life, the original production data is queried through the blockchain, combined with failure analysis, to reversely locate the defect formation link; S705: Regularly analyze quality traceability data, generate process improvement reports, and push improvement plans to the PLM system.

Images