Crack detection method, device and system for precision casting

Abstract

The invention relates to the technical field of crack detection, in particular to a crack detection method, device and system for a precision casting. The method comprises the following steps that all-directional sound wave scanning is conducted on the precision casting, sound wave reflection signals are collected, dynamic sound wave imaging mapping is conducted, and a casting three-dimensional space voiceprint graph is constructed; performing transient abrupt change phase position calculation and casting crack space positioning on the casting three-dimensional space voiceprint graph to obtain three-dimensional space position information of casting cracks; based on the three-dimensional space position information of the casting crack, high-frequency magnetic field excitation processing is conducted on the precision casting, an amplitude change curve, a phase change curve and a frequency change curve are obtained through calculation, and crack three-dimensional shape calculation is conducted according to the amplitude change curve, the phase change curve and the frequency change curve to obtain crack three-dimensional shape characteristics. According to the invention, rapid and accurate precision casting crack detection is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of crack detection, and in particular to a crack detection method, device and system for precision castings. Background Art

[0002] With the continuous advancement of manufacturing and engineering technology, precision castings, as important mechanical components, are widely used in a variety of fields, including aerospace, automotive, electronics, and medical. Due to their ability to withstand high loads, withstand high temperatures, and resist corrosion, precision castings have become an indispensable foundational component in modern industrial production. However, with the increasing complexity of the operating environments of precision castings and the continuous improvement of performance requirements, material defects such as cracks have gradually become apparent, posing a serious threat to production efficiency, product quality, and user safety.

[0003] During the production of precision castings, cracks often develop due to the casting process, material properties, or external manipulation, resulting in tiny defects on the surface or inside the casting. Cracks not only affect the structural strength of the casting and reduce its load-bearing capacity, but can also cause equipment failure, production interruptions, and even serious safety incidents. For example, in the aerospace industry, any tiny crack can lead to a high-risk accident; in automotive manufacturing, cracks can cause unforeseen failures while the vehicle is in motion, seriously impacting driving safety. Therefore, crack detection in precision castings, especially the precise detection of tiny cracks, has become a key technology for improving product quality and ensuring industrial safety.

[0004] Traditional methods for crack detection in precision castings mainly rely on non-destructive testing technologies such as visual inspection, magnetic particle testing, penetrant testing, ultrasonic testing, and X-ray testing. Although these methods can detect cracks or defects to a certain extent, they have some shortcomings in practical applications. First, most traditional methods rely on manual operation, which is inefficient and easily affected by the operator's experience. In addition, the detection effect of complex shapes and deep cracks is limited. Secondly, although some traditional methods such as ultrasonic testing and X-ray testing can detect deeper cracks, they are unable to detect small-scale, high-precision cracks. Furthermore, the existing technology lacks real-time and automation, making it difficult to provide rapid feedback and early warning during the casting production process. Summary of the Invention

[0005] In order to solve the above technical problems, the present invention proposes a crack detection method, device and system for precision castings to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a method for crack detection of precision castings, comprising the following steps: Step S1: Perform omnidirectional acoustic scanning on the precision casting, collect acoustic reflection signals, and perform dynamic acoustic imaging mapping to construct a three-dimensional spatial soundprint map of the casting; Step S2: calculating the transient mutation phase position and spatially locating the casting crack on the casting three-dimensional soundprint image to obtain the three-dimensional spatial position information of the casting crack; Step S3: Based on the three-dimensional spatial position information of the casting crack, the precision casting is subjected to high-frequency magnetic field excitation processing, and the amplitude change curve, the phase change curve and the frequency change curve are calculated: Step S4: performing three-dimensional crack morphology estimation based on the amplitude variation curve, the phase variation curve, and the frequency variation curve to obtain three-dimensional crack morphology characteristics; Step S5: performing multi-time point crack diffusion evolution prediction on the three-dimensional crack morphological characteristics, and performing multi-time point diffusion fitting to construct a crack evolution prediction map; Step S6: Optimize preventive process decisions based on the crack evolution prediction graph and build a preventive crack detection and protection model.

[0007] In this specification, a crack detection device for precision castings is provided, which is used to perform the crack detection method for precision castings as described above, comprising: The acoustic imaging module is used to perform omnidirectional acoustic scanning of precision castings, collect acoustic reflection signals, and perform dynamic acoustic imaging mapping to construct a three-dimensional spatial soundprint map of the castings; The crack spatial positioning module is used to calculate the transient mutation phase position of the casting three-dimensional spatial soundprint and locate the casting crack space to obtain the three-dimensional spatial position information of the casting crack; The high-frequency magnetic field excitation module is used to perform high-frequency magnetic field excitation processing on the precision casting based on the three-dimensional spatial position information of the casting crack, and calculate the amplitude change curve, phase change curve and frequency change curve: A crack morphology estimation module is used to estimate the three-dimensional morphology of the crack based on the amplitude change curve, the phase change curve and the frequency change curve to obtain the three-dimensional morphological characteristics of the crack; The crack evolution prediction module is used to predict the crack diffusion evolution at multiple time points based on the three-dimensional crack morphology characteristics, perform multi-time point diffusion fitting, and construct a crack evolution prediction map; The preventive process decision module is used to optimize preventive process decisions based on the crack evolution prediction diagram and build a preventive crack detection and protection model.

[0008] The present invention also provides a computer system comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above-mentioned methods for detecting cracks in precision castings when executing the computer program.

[0009] The beneficial effects of the present invention are as follows: by performing omnidirectional acoustic scanning on precision castings, subtle reflection signals inside the castings can be obtained, ensuring that cracks on both the surface and the interior can be detected. Compared with traditional methods, omnidirectional scanning can improve detection coverage and ensure comprehensive detection. The acoustic reflection signal can be mapped through dynamic acoustic imaging to generate a three-dimensional spatial soundprint of the casting in real time. This image not only shows the geometric shape of the casting, but also reveals potential cracks or defective areas. The high-resolution imaging capability of this process greatly enhances the ability to detect cracks. Dynamic imaging can capture the changes in cracks over time, further improving the accuracy of crack detection. It is particularly suitable for detecting tiny cracks and breaking through the limitations of traditional methods. By calculating the phase position of transient mutations in the three-dimensional spatial soundprint, the relative position of the crack can be accurately identified. Transient mutations are usually manifestations of cracks, and their phase differences are significant, which can help quickly locate potential crack areas. Using the phase changes of the acoustic reflection signal for positioning can accurately locate the three-dimensional coordinates of the crack in space, which is far superior to traditional positioning methods based on single angles or linear scanning. This process enables dynamic, real-time analysis and location of cracks, facilitating immediate repair or treatment measures to prevent further damage from crack propagation. High-frequency magnetic field excitation, by applying a magnetic field of a specific frequency to the casting, stimulates microscopic reactions in the crack area, enhancing the crack's response to electromagnetic waves and thus improving crack detection sensitivity. This is particularly effective for detecting small or deeply hidden cracks. By collecting multi-dimensional electromagnetic wave signals, including amplitude, phase, and frequency variations, a more comprehensive analysis of the crack's physical characteristics is possible, providing richer information and helping to accurately determine its type and development trend. Combining amplitude, phase, and frequency curves, a three-dimensional crack morphology is estimated, enabling quantitative analysis of crack width, depth, and complexity. Compared to traditional two-dimensional analysis methods, this estimation method provides a more accurate description of crack morphology. Comprehensive signal curve analysis reveals the crack's growth dynamics, understanding its internal structure and potential hazards. This provides a detailed basis for subsequent crack repair. This step enables the three-dimensional crack morphology to be presented as a digital model, providing accurate data support for predicting future crack growth. Multi-point crack propagation and evolution prediction allows real-time tracking of crack growth trends and timely prediction of future crack development, helping decision-makers make more accurate maintenance decisions. Crack propagation and evolution predictions are based not only on current data but also on historical trends, making predictions more forward-looking and accurate. Diffusion fitting at different time points provides dynamic, continuously updated predictions. By constructing preventative crack detection and protection models, process decisions can be intelligently optimized, reducing the likelihood of human misjudgment and improving decision-making efficiency.Through accurate crack prediction and active protection, large-scale crack propagation and structural damage can be effectively avoided, thereby reducing repair costs and downtime, and improving overall production efficiency and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 1. A schematic flow chart of the steps of a crack detection method for precision castings according to the present invention; Figure 2 Detailed implementation flow chart of step S1; Figure 3 Detailed implementation flow chart of step S2; Figure 4 Schematic diagram of the detailed implementation steps of step S3. DETAILED DESCRIPTION

[0011] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0012] This application provides a method, device, and system for detecting cracks in precision castings. The execution entities of the method, device, and system include, but are not limited to, the following: mechanical equipment, data processing platforms, cloud server nodes, network upload devices, etc., which can be considered as general computing nodes of this application. The data processing platform includes, but is not limited to, at least one of an audio and image management system, an information management system, and a cloud data management system.

[0013] See also Figures 1 to 4 The present invention provides a method for detecting cracks in precision castings, the method comprising the following steps: Step S1: Perform omnidirectional acoustic scanning on the precision casting, collect acoustic reflection signals, and perform dynamic acoustic imaging mapping to construct a three-dimensional spatial soundprint map of the casting; Step S2: calculating the transient mutation phase position and spatially locating the casting crack on the casting three-dimensional soundprint image to obtain the three-dimensional spatial position information of the casting crack; Step S3: Based on the three-dimensional spatial position information of the casting crack, the precision casting is subjected to high-frequency magnetic field excitation processing, and the amplitude change curve, the phase change curve and the frequency change curve are calculated: Step S4: performing three-dimensional crack morphology estimation based on the amplitude variation curve, the phase variation curve, and the frequency variation curve to obtain three-dimensional crack morphology characteristics; Step S5: performing multi-time point crack diffusion evolution prediction on the three-dimensional crack morphological characteristics, and performing multi-time point diffusion fitting to construct a crack evolution prediction map; Step S6: Optimize preventive process decisions based on the crack evolution prediction graph and build a preventive crack detection and protection model.

[0014] In the embodiment of the present invention, see Figure 1 , is a schematic flow chart of the steps of a crack detection method for precision castings of the present invention. In this example, the steps of the crack detection method for precision castings include: Step S1: Perform omnidirectional acoustic scanning on the precision casting, collect acoustic reflection signals, and perform dynamic acoustic imaging mapping to construct a three-dimensional spatial soundprint map of the casting; In this embodiment, an acoustic scanning device suitable for casting inspection is selected, typically using high-frequency ultrasonic sensors with high sensitivity and a wide frequency response range (e.g., 1 MHz to 10 MHz). Multiple sensors are deployed based on the size and shape of the casting to ensure full coverage. For example, for a casting weighing 200 kg, at least 12 sensors are deployed, with each sensor no more than 5 cm apart, to obtain detailed acoustic reflection data. The casting is secured to the scanning platform to prevent movement during the scanning process. If necessary, a clamp or fixture can be used to prevent interference. For example, before scanning, ensure that the ambient temperature and humidity are within an appropriate range (e.g., 20°C, 60% humidity) to minimize the impact on acoustic wave propagation. The acoustic scanning device is activated, emitting ultrasonic signals along a pre-set scanning trajectory. Each sensor transmits an acoustic wave at a predetermined interval (e.g., 1 second) and simultaneously records the reflected signal. Each sensor should be able to record the emission time, reception time, and reflection intensity of the acoustic wave. For example, suppose sensor A transmits an acoustic wave at time T1 and receives a reflected signal at time T2. Data is recorded: sensor ID, transmission time, reception time, and reflection intensity. Signals from all sensors are centrally processed by a data acquisition system. This system should have real-time data processing capabilities, capable of capturing signals from multiple sensors in parallel and storing them in a database. For example, the data acquisition system should have a sampling rate of 1 MHz to ensure that subtle signal changes, especially in the presence of cracks or defects, are captured. The collected acoustic reflection signals undergo preliminary processing, including denoising, signal enhancement, and time correction. A filtering algorithm (such as a low-pass filter) is used to remove high-frequency noise to improve signal quality. For example, wavelet transform techniques can be applied to denoise the signal and extract useful acoustic reflection features. Dynamic imaging algorithms (such as backprojection or image reconstruction) are used to convert the processed acoustic signal into a three-dimensional acoustic image. This process should be performed by mapping the acoustic signal from each sensor to its corresponding three-dimensional coordinates. For example, suppose the signal intensity distribution obtained by scanning a casting is mapped into three-dimensional space. The resulting acoustic image shows the acoustic wave propagation within the casting. Visualize the acoustic imaging results as a three-dimensional soundprint, using color coding or intensity mapping to show the sound wave intensity in different areas. The soundprint should clearly show the internal structure of the casting and the location of potential defects. For example, use a heat map to represent the sound wave intensity. The red area indicates high sound wave reflection intensity, which may be a crack or defect, while the blue area indicates low reflection intensity, indicating an intact structure. After the acoustic imaging is completed, the generated soundprint and related data are stored in the database for subsequent analysis and reference. Ensure the integrity and accuracy of the data, and record the location information of each sensor and the collected signal characteristics. For example, create a data log including sensor ID, coordinates, reflection intensity, processed signal characteristics, etc., to facilitate subsequent review and analysis.Perform a preliminary analysis of the generated 3D voiceprint to identify potential cracks or defects. Considering the casting's design requirements and material properties, determine whether any abnormal areas in the voiceprint require further testing. For example, if the sound wave reflection intensity in a specific area of ​​the voiceprint is significantly lower than that in the surrounding area, this area warrants further verification and analysis.

[0015] Step S2: calculating the transient mutation phase position and spatially locating the casting crack on the casting three-dimensional soundprint image to obtain the three-dimensional spatial position information of the casting crack; In this embodiment, a three-dimensional acoustic fingerprint of the casting was constructed in the previous step. This fingerprint contains the intensity and phase information of the acoustic reflection signal. Now, the acoustic fingerprint needs to be analyzed to identify transient phase changes in the acoustic signal. For example, if the acoustic fingerprint shows significant phase changes in certain areas, this may be associated with the presence of cracks or other defects. The phase information recorded in the acoustic fingerprint is processed to identify the points where the phase changes occur. This can be achieved by calculating the rate of phase change. For example, a phase change threshold can be set. When the phase change exceeds this threshold, it is marked as a transient phase change. For example, if the phase changes from 0.2π radians to 0.6π radians at a certain location, a change of 0.4π, and the threshold is set to 0.3π, this change is identified as a transient phase change. The transient phase change is mapped back to the three-dimensional spatial coordinates of the casting. The position information of each sensor (such as the x, y, and z coordinates) should be combined with the phase data to determine the specific location of the phase change. For example, suppose a sudden phase change is detected at sensor A (coordinates (10 cm, 15 cm, 5 cm)). These coordinates are recorded as part of the potential crack location. In the soundprint, the locations of all transient phase changes are marked. This can be achieved by highlighting these points in the 3D plot, facilitating subsequent analysis and identification. For example, all phase points where a sudden transient change is detected can be marked with a different color (such as red) to make them more prominent in the plot. The marked transient phase change locations are integrated with the casting's geometric model. By comparing the locations of the phase changes with the actual casting structure, the likelihood of cracks at these locations is analyzed. For example, if the transient phase change points are concentrated in stress concentration areas of the casting, the presence of cracks in these areas can be inferred. Based on the locations of the phase changes and the 3D casting model, the 3D spatial location of the crack is extracted. The specific crack morphology is determined by combining the number and distribution of transient phase changes. For example, if phase changes are observed at multiple sensor locations, it can be inferred that the crack extends along a specific direction. The 3D location of the crack is recorded, including its initiation point, direction, and depth. The extracted crack's 3D spatial location information is visualized to generate a crack model. 3D modeling software can be used to integrate the crack geometry with the overall casting structure. For example, the generated crack model shows the crack's shape, location, and depth, facilitating further analysis and verification.

[0016] Step S3: Based on the three-dimensional spatial position information of the casting crack, the precision casting is subjected to high-frequency magnetic field excitation processing, and the amplitude change curve, the phase change curve and the frequency change curve are calculated: In this embodiment, a high-frequency magnetic field excitation device is selected, typically a high-frequency electromagnetic exciter (such as an electromagnetic coil or electromagnetic generator). This device should be able to generate a stable and adjustable frequency magnetic field. For example, an electromagnetic generator with a frequency range of 50 kHz to 1 MHz is selected to suit the material properties of the casting and the crack detection requirements. The electromagnetic generator is positioned to ensure uniform coverage of the target area of ​​the casting, particularly the area where the crack is located. The casting is secured to the excitation platform to prevent movement during the excitation process. This can be achieved using a clamp or support frame to ensure the stability of the excitation process. For example, in a laboratory environment, the temperature is maintained at 20°C and the humidity is kept below 60% to minimize the impact of the high-frequency magnetic field and ensure data accuracy. The high-frequency electromagnetic exciter is activated and the casting is excited at the set frequency. The exciter output power is adjusted to ensure that the excitation signal is strong enough to detect a crack response. For example, the excitation frequency is set to 100 kHz, the output power is set to 5 W, and the excitation duration is set to 3 minutes to ensure sufficient propagation of the acoustic wave signal and interaction with the crack. During the excitation process, the signal acquisition system is activated to record the electromagnetic wave signal on the casting surface in real time. The system should have a high sampling rate to capture instantaneous signal changes and record amplitude, phase, and frequency data. For example, setting a sampling rate of 1 MHz ensures the capture of 1 million data points per second, providing high-resolution signal analysis. De-noising is performed on the acquired electromagnetic wave signals to improve signal quality. Filters (such as high-pass or low-pass filters) are used to remove noise and retain valid signals. For example, a low-pass filter with a cutoff frequency of 500 kHz can be applied to filter out high-frequency noise and ensure signal clarity. The processed signals are analyzed to calculate an amplitude variation curve. The amplitude can be obtained by calculating the effective value or peak value of the signal and plotted against time. For example, if the signal amplitude changes from 1.5 mV to 2.0 mV at different time points of excitation, the generated amplitude variation curve will show this change. Signal processing algorithms (such as the Hilbert transform) are used to extract the phase variation curve. Phase variation reflects the propagation characteristics of electromagnetic waves in the crack region. For example, calculate how the phase changes during the excitation process. If the phase is recorded to change from 0.1π radians to 0.3π radians, the phase change curve will show this characteristic. Analyze the signal through Fourier transform to obtain a frequency change curve. Frequency changes can reflect the dynamic response of the crack under excitation and identify potential crack characteristics. For example, if the analysis results show that the frequency changes slightly from 100 kHz to 98 kHz during the excitation process, the frequency change curve will reflect this process. The data of the amplitude change curve, phase change curve, and frequency change curve are fully recorded and stored in a database for subsequent analysis and comparison. For example, create a data record table containing the amplitude, phase, and frequency values ​​at each time point to ensure the systematic and traceable nature of the data.A preliminary analysis of the generated variation curves is performed to identify abnormal variations and correlate them with crack characteristics. This helps to evaluate the presence of cracks and their impact on the signal.

[0017] Step S4: performing three-dimensional crack morphology estimation based on the amplitude variation curve, the phase variation curve, and the frequency variation curve to obtain three-dimensional crack morphology characteristics; In this embodiment, the amplitude, phase, and frequency curves have been obtained in the previous step. These curves provide important basic data for estimating the three-dimensional crack morphology. For example, suppose the amplitude curve shows an amplitude increase from 1.5 mV to 2.0 mV within a specific time period, accompanied by significant fluctuations in phase change. This may indicate the presence of a crack. The amplitude, phase, and frequency change data are integrated to form a comprehensive signal feature dataset. This dataset will be used for subsequent crack morphology estimation. For example, the amplitude, phase, and frequency data at each time point can be stored in a matrix for unified processing and analysis. Based on the propagation characteristics of acoustic and electromagnetic waves in materials, an appropriate mathematical model is selected to estimate the three-dimensional crack morphology. Commonly used models include crack growth models and stress concentration models. For example, a crack growth model based on Fourier transform can be selected, which can convert amplitude and phase change information into crack geometric characteristics. An inversion algorithm (such as the least squares method or an optimization algorithm) is used to analyze the integrated signal features to infer the three-dimensional crack morphology. This process requires fitting signal characteristics with model parameters to obtain the optimal solution. For example, initial crack parameters (such as width, depth, and shape) are set, and then these parameters are adjusted through an iterative optimization algorithm to minimize the error between the calculated signal characteristics and the true signal characteristics. Based on the inferred results, the three-dimensional morphological characteristics of the crack are calculated, including the crack's length, width, depth, and direction. These characteristics can reflect the specific location and morphology of the crack in the casting. For example, suppose the inferred results show a crack length of 5 mm, a width of 1 mm, and a depth of 75 mm, and the crack extends along a certain direction in the casting. The calculated three-dimensional crack morphological characteristics are visualized to intuitively show the crack's geometry and location. 3D modeling software can be used to generate a crack model and compare it with the casting structure. For example, the generated 3D crack model can display the crack's morphological characteristics, including its direction and depth, in the software, assisting engineers in further analysis.

[0018] Step S5: performing multi-time point crack diffusion evolution prediction on the three-dimensional crack morphological characteristics, and performing multi-time point diffusion fitting to construct a crack evolution prediction map; In this embodiment, in the previous step, the three-dimensional morphological characteristics of the crack were obtained, including the crack's length, width, depth, and specific location in the casting. These characteristics provide basic data for predicting the evolution of crack propagation. For example, assume a crack is 5 mm long, 1 mm wide, and 75 mm deep, located in the stress concentration zone of the casting. Based on the crack's material properties and operating conditions, an appropriate propagation prediction model is selected. Commonly used models include crack propagation models based on stress intensity factors and models based on energy release rates. For example, a propagation model based on stress intensity factors is suitable for predicting crack propagation behavior under different stress states. Before predicting crack propagation, it is first necessary to analyze the stress state of the casting under actual operating conditions. This can be achieved using finite element analysis (FEA) to calculate the stress distribution of the casting under operating load. For example, assuming the stress state of the casting under a concentrated force of 500 N, the analysis results show that the maximum principal stress near the crack is 120 MPa. Using the crack propagation prediction model, the crack propagation behavior at different time points is simulated. By calculating the stress intensity factor (K value), the crack growth rate at a specific time is determined. For example, assuming a crack growth rate of 0.1 mm / h under the initial stress state, future crack growth can be predicted based on different stress states and operating times. The crack growth data obtained at each time point during the simulation are organized to form a complete growth data set. This data should include the crack length, width, and growth rate at each time point. For example, at time t = 0 hours, the crack length is 5 mm; at t = 1 hour, the crack length is 5.1 mm; at t = 2 hours, the crack length is 5.3 mm, and so on. The collected growth data is analyzed using curve fitting to develop a mathematical model for crack propagation. Linear regression or polynomial regression methods can be used to determine the crack growth trend. For example, if the crack length increases linearly with time, a linear equation can be used to fit the relationship between crack length and time to obtain a mathematical expression for the crack growth rate. Based on the results of the multi-time point growth fitting, a crack evolution prediction diagram is generated. The diagram should show the predicted morphology of the crack at different time points, including the length, width, and position of the crack. For example, the generated prediction diagram can use different colors or lines to represent the crack expansion at different time stages for easy visual observation. Visualizing the crack evolution prediction diagram facilitates subsequent analysis and decision-making. 3D modeling software can be used to compare the predicted crack morphology with the overall structure of the casting to intuitively demonstrate the crack expansion trend. For example, the generated 3D model shows the process of a crack expanding from 5 mm to 7 mm and identifies potential high-risk areas.

[0019] Step S6: Optimize preventive process decisions based on the crack evolution prediction graph and build a preventive crack detection and protection model.

[0020] In this embodiment, in the previous step, a crack evolution prediction graph has been generated, which shows the expansion of the crack at a future time point. This graph provides important data support for subsequent preventive process decisions. For example, suppose the prediction graph shows that the length of the crack will expand from 5 mm to 10 mm in the next six months, and the speed of crack expansion will significantly accelerate in a specific area. According to the crack evolution prediction graph, potential high-risk areas in the casting are identified. These areas are usually places where cracks expand faster or stress is concentrated. For example, on the crack evolution prediction graph, if the crack expansion rate in a certain area exceeds 0.5 mm / h, the area is identified as a high-risk area and requires special attention. Based on the crack evolution prediction results, a corresponding monitoring strategy is formulated. Focus on high-risk areas and determine the monitoring frequency and method to detect potential crack expansion in a timely manner. For example, a detailed inspection is set for high-risk areas once a month, while a quarterly inspection is set for low-risk areas. This differentiated monitoring strategy can optimize resource allocation. Select appropriate detection technology to implement the monitoring strategy. Consider using methods such as ultrasonic testing, magnetic particle testing, or X-ray testing, choosing the right method based on the characteristics of the casting material and crack type. For example, ultrasonic testing is considered an effective inspection method for aluminum alloy castings, accurately identifying internal cracks. Based on the monitoring strategy, construct a preventive crack detection and protection model. This model should include inspection frequency, methods, responsible individuals, and contingency plans to ensure a prompt response when cracks are discovered. For example, a table or database could be created to record the monitoring plan for each area, including inspection methods, scheduled dates, and responsible individuals, to ensure that all work is carried out in an orderly manner. Within the protection model, develop contingency plans to address potential crack growth or emergencies. These plans should include procedures for handling cracks after detection, division of responsibilities, and resource allocation. For example, if monitoring reveals crack growth exceeding expectations, the contingency plan should be immediately activated, requiring more detailed inspections and considering shutdowns for maintenance. After implementing the preventive crack detection and protection model, regularly evaluate its effectiveness. By analyzing inspection data, determine whether the model is able to detect cracks promptly and prevent their growth. For example, if multiple potential cracks are successfully identified and repaired within six months of implementation, the model is considered effective. Based on implementation results and monitoring data, the preventive crack detection and protection model is continuously optimized. Monitoring strategies and emergency response plans are regularly reviewed to ensure their adaptability and effectiveness. For example, if certain monitoring frequencies are found to be excessive and wasteful of resources, they can be adjusted to more appropriate frequencies based on actual conditions.

[0021] In this embodiment, refer to Figure 2 , is a flowchart of the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Based on the high-density three-dimensional acoustic wave scanning array, the precision casting is scanned in all directions to collect the acoustic wave reflection signal; Calculating the timestamp of the acoustic wave reflection signal; calculating the time delay deviation of different signals on the acoustic wave reflection signal to obtain the signal time delay deviation; Performing time synchronization processing on the signal time delay deviation based on the timestamp to obtain a time-synchronized acoustic wave reflection signal; Performing phase correction on the time-synchronized acoustic wave reflection signal to obtain a phase-corrected optimized signal; Dynamic acoustic imaging mapping is performed based on the phase correction optimization signal to construct a three-dimensional spatial soundprint map of the casting.

[0022] In this embodiment, a high-density three-dimensional acoustic wave scanning array is used, equipped with multiple acoustic wave sensors to cover the entire surface of the casting. These sensors should have high sensitivity and a wide frequency response to ensure that they can capture acoustic wave reflection signals of various frequencies. For example, ultrasonic sensors with a frequency range of 1 MHz to 10 MHz are selected to ensure that they can detect subtle cracks and defects. The casting is placed on a scanning platform, and the acoustic wave scanning array is activated to transmit acoustic waves in all directions. Each sensor transmits acoustic waves at regular intervals, recording the propagation time of the acoustic waves in the casting and the reflected signals. Assume that during the scanning process, the sensor spacing is set to 5 cm and the scanning angle is set to 360 degrees, ensuring that each sensor covers different parts of the casting and obtains comprehensive reflected signals. During the scanning process, the acoustic wave reflection signals are captured in real time and stored in the data acquisition system. Each signal should include information such as the transmission time, reception time, and reflection intensity. For example, a sensor transmitting an acoustic wave at time T1 and receiving a reflected signal at time T2 is recorded and stored as "sensor ID, transmission time, reception time, and reflection intensity." For each acoustic wave reflection signal, a timestamp is calculated, recording the specific time of acoustic wave transmission and reception. This timestamp is used in subsequent signal analysis. For example, if at a certain point in time, the sensor transmits the sound wave at time T1 = 0.001 seconds and receives the reflected signal at time T2 = 0.005 seconds, the timestamp is T2 - T1 = 0.004 seconds. Time delay deviation is calculated for all collected sound wave reflection signals. By comparing the signal time delays between different sensors, waveform deviations can be identified. This process can help identify variations in sound wave propagation within the casting. For example, if the reflection time from sensor A is 0.004 seconds and the reflection time from sensor B is 0.003 seconds, the time delay deviation is calculated as 0.004 seconds - 0.003 seconds = 0.001 seconds, indicating that the signal from sensor A is delayed by 1 millisecond relative to the signal from sensor B. The calculated timestamp and time delay deviation are recorded in the data storage system for subsequent data processing and analysis. Each record should include information such as the sensor ID, timestamp, and signal delay deviation. For example, the record format could be "sensor ID, timestamp, time delay deviation" to ensure systematic and traceable data. Based on the calculated timestamps, signal time delay deviations are synchronized. Before signal merging, all signals are aligned to the same time base to avoid waveform distortion caused by time differences. For example, a reference sensor (such as sensor B) is selected, and the signals from other sensors are aligned to coincide with sensor B, ensuring that all signals are analyzed within the same time frame. Phase correction is then performed on the time-synchronized acoustic reflection signals, using a phase compensation algorithm to eliminate phase deviations caused by time delays. This process is crucial for improving imaging accuracy.For example, using the least squares method or phase matching algorithm, the phase deviation of each signal is calculated based on the phase of the reflected signals from different sensors, and the corresponding correction is performed. Phase correction generates optimized acoustic reflection signals. These optimized signals should have a higher signal-to-noise ratio and accuracy, enabling more effective use in subsequent dynamic acoustic imaging. Dynamic acoustic imaging is performed using the phase-corrected optimized signals. The optimized signals are input into the imaging algorithm to generate a three-dimensional acoustic image of the casting. An appropriate imaging algorithm, such as an inversion algorithm or image reconstruction algorithm, is selected to achieve high-resolution imaging. For example, using the backprojection method, the signals captured by each sensor are used as imaging input, and image synthesis techniques are used to generate a three-dimensional acoustic image of the casting. Based on the dynamic acoustic imaging results, a three-dimensional acoustic fingerprint of the casting is constructed. The acoustic fingerprint should clearly demonstrate the acoustic propagation within the casting, including the location of defects such as cracks and bubbles. For example, the generated acoustic fingerprint image shows the location and size of fine cracks within the casting, and the severity of the defect is indicated by color based on the intensity of the acoustic reflection. The generated three-dimensional acoustic fingerprint image is analyzed to verify its accuracy and reliability. The imaging effect and detection sensitivity can be evaluated by comparing it with known defects. For example, by comparing it with an actual physically cut sample, the consistency between the crack location shown in the voiceprint and the actual location can be confirmed to verify the effectiveness of the detection method.

[0023] In this embodiment, the specific steps of performing dynamic acoustic imaging mapping based on the phase correction optimization signal to construct a three-dimensional voiceprint map of the casting are as follows: Calculating the signal strength of the phase correction optimization signal; performing a three-dimensional spatial distribution analysis on the signal intensity to obtain three-dimensional spatial intensity distribution data; Perform multi-time sequence coherent superposition on the phase correction optimization signal to construct a time sequence superposition acoustic wave sequence; Identifying scanning points of the high-density three-dimensional acoustic wave scanning array; Calculating the spatial position coordinates of the scanning point; Signal intensity matching is performed on the three-dimensional spatial intensity distribution data based on the spatial position coordinates, and dynamic three-dimensional acoustic wave imaging mapping is performed according to the time-series superposition of acoustic wave sequences to construct a three-dimensional spatial soundprint map of the casting.

[0024] In this embodiment, it is ensured that the phase correction process has been completed to obtain the optimized acoustic wave reflection signal. The signal of each sensor should contain reflection intensity, phase information and timestamp, which will be used for subsequent intensity calculations. For example, suppose that the received optimization signal intensity of a sensor is 0.8 V (peak voltage) when the excitation frequency is 5 MHz. The effective value (RMS) or peak intensity method is used to calculate the signal intensity. The effective value is the average level of signal energy and can more accurately reflect the true intensity of the signal. If the amplitude range of the waveform of the optimized signal within one cycle is -0.8 V to 0.8 V, the effective value is calculated as: The calculated signal intensity is associated with the spatial position of the sensor to generate a three-dimensional spatial distribution of signal intensity. This process can be achieved by mapping the signal intensity into a three-dimensional coordinate system. Using 3D visualization software, the coordinates of each sensor and the corresponding signal intensity data are input to generate a three-dimensional spatial intensity distribution map. Color or size can be used to represent high and low intensity levels for intuitive analysis. For example, using a heat map or contour map to display intensity changes, ensure that areas of high signal intensity are highlighted in the map, as these may indicate cracks or defects. Multi-time series coherent superposition is the superposition of signals from multiple time points at the same location to improve signal strength and signal-to-noise ratio. This process can enhance the detection of subtle defects. For example, signal data from the same location is selected, aligned in time, and then the signals from each time point are added together to form a new superimposed signal. Using the previously calculated signal intensity and timestamp, the signals from each sensor are time-aligned to form a time-series superimposed acoustic wave sequence. By comparing signals from different time periods, persistent signal features can be identified. For example, if the signals from sensor A at multiple time points are 0.5 V, 0.6 V, and 0.7 V, respectively, the resulting time series signal after superposition is: S(A) = 0.5 + 0.6 + 0.7 = 1.8 V. Based on the configuration of the high-density 3D acoustic scanning array, all scanning points are identified. Each scanning point should correspond to a specific sensor location for subsequent signal analysis and mapping. For example, suppose 100 sensors are arranged on the casting surface. Each sensor's unique ID and spatial coordinates (x, y, z) are recorded. For each scanning point, its specific coordinates in 3D space are calculated. This process should take into account the sensor's installation angle and position to ensure coordinate accuracy. For example, the coordinates of sensor A are (10 cm, 15 cm, 5 cm), the coordinates of sensor B are (12 cm, 17 cm, 5 cm), and so on. Based on the calculated spatial position coordinates, signal intensity matching is performed on the 3D spatial intensity distribution data. The signal intensity of each sensor is associated with its corresponding spatial position to generate a more accurate intensity distribution model. For example, if the coordinates of sensor A are (10, 15, 5) and its signal intensity is 1.8 V, then that point in the 3D model is marked with an intensity of 1.8 V. Dynamic 3D acoustic imaging mapping is performed based on the time-series superposition of acoustic wave sequences and intensity matching results. Using an imaging algorithm, the signal is converted into a 3D image, visually displaying the acoustic wave propagation characteristics and possible cracks within the casting. For example, using the backprojection method, the intensity distribution generated by the acoustic wave signal is used to reconstruct a 3D acoustic print of the casting, showing the internal structure and potential defects. The resulting acoustic print should display the intensity distribution within the casting, including the location of defects such as cracks and bubbles. Using different colors or depths to represent different intensities facilitates subsequent analysis and detection.For example, the voiceprint image shows that the intensity of the crack area is significantly lower than that of the surrounding area, and the changes in color depth visually reflect the severity of the defect.

[0025] In this embodiment, refer to Figure 3 , is a flowchart of the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Perform acoustic texture recognition on the three-dimensional acoustic print of the casting to obtain acoustic texture features; Perform spatial filtering analysis on the acoustic wave texture features to extract the acoustic wave texture intensity and texture direction; Calculate the texture gradient based on the acoustic wave texture features to obtain the density of the texture gradient; Detect transient texture mutations based on texture gradient density, acoustic texture intensity, and texture direction, and mark transient texture mutation areas. Calculating the texture phase difference and amplitude difference between the transient mutation texture region and adjacent regions; The transient mutation phase position is calculated based on the texture phase difference and amplitude difference, thereby obtaining the relative position coordinates of the mutation texture in the voiceprint image; The precision casting is spatially positioned for the casting crack based on the relative position coordinates to obtain three-dimensional spatial position information of the casting crack.

[0026] In this embodiment, a three-dimensional voiceprint image of the casting was generated in the previous step. This image contains the intensity distribution of the acoustic wave reflection signal, reflecting the internal structure and potential defects of the casting. The goal of acoustic texture recognition is to extract features from this image and analyze its texture characteristics. For example, areas in the voiceprint image that exhibit dramatic intensity variations may correspond to cracks or other defects. Texture feature recognition is performed on the voiceprint image using image processing techniques. Common methods include gray-level co-occurrence matrix (GLCM) and local binary pattern (LBP). These methods can extract texture features such as contrast, uniformity, and energy. For example, GLCM can be used to calculate the texture features of each pixel, identifying the texture pattern and intensity variations of the voiceprint image. These features can be used to derive the overall characteristics of the acoustic texture. Spatial filtering analysis is performed on the extracted acoustic texture features to remove noise and enhance signal characteristics. Common filtering methods include high-pass and low-pass filters. The appropriate filter should be selected as needed. For example, a high-pass filter can be used to enhance acoustic texture details, making small features such as cracks more visible. After filtering, the intensity and texture direction of the acoustic texture are extracted. Image gradient operators (such as the Sobel operator or Canny edge detection) are used to calculate the gradient at each point and obtain texture directional information. For example, if the texture intensity of a region is calculated to be 0.75 V and the texture direction is 45 degrees, this information will provide the basis for subsequent texture gradient calculation. Texture gradient calculation is performed based on the extracted acoustic wave texture features. Texture gradient reflects the rate and direction of texture change and can indicate possible defect areas. For example, if the texture intensity at a certain location changes from 0.5 V to 1.0 V, the calculated gradient is 2.0 V / cm, indicating significant texture change in that area. Statistical analysis of the entire voiceprint image assesses the density of texture gradients. A threshold can be set to identify areas with texture gradients exceeding the threshold and mark them as possible defect areas. For example, if a threshold of 1.5 V / cm is set and a gradient of 2.0 V / cm is detected in a region, this region is marked as a high-density texture area. Transient texture refers to texture features that undergo significant changes within a short period of time. These areas are identified and marked based on the density and intensity of the texture gradients. For example, if the texture intensity of a certain area jumps from 0.4 V to 1.2 V within a short period of time, this area can be marked as a transient mutation texture area. The identified transient mutation texture area is marked, and its location and characteristics in the voiceprint are recorded. This process provides a basis for subsequent phase difference and amplitude difference calculations. For example, the marked area contains coordinates (x, y, z) as well as intensity and direction information to ensure data integrity. For transient mutation texture areas, the texture phase difference and amplitude difference between them and adjacent areas are calculated. The phase difference can be obtained by comparing the phase data of the transient area with the adjacent areas, while the amplitude difference is calculated based on the difference in signal intensity.For example, assuming the phase of the transient region is θ1 and the phase of the adjacent region is θ2, the phase difference is calculated as |θ1 - θ2|. The calculation results are recorded for subsequent analysis and location. Phase difference and amplitude difference data are crucial for determining the location and nature of the crack. This information is mapped into three-dimensional space to determine the relative position of the crack. For example, if the phase position of the transient region is determined to be (x0, y0, z0), the relative position coordinates are calculated using the phase difference. The obtained relative position coordinates are used to spatially locate the crack in the precision casting. Ultimately, the three-dimensional spatial position information of the casting crack is generated for subsequent analysis and processing. For example, assuming the crack location is (10 cm, 15 cm, 5 cm), this information is recorded to facilitate subsequent repair and quality control.

[0027] In this embodiment, refer to Figure 4 , is a flowchart of the detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include: Obtaining a preset precision casting safety magnetic field frequency, and defining a maximum safety magnetic field frequency according to the safety magnetic field frequency; Based on the maximum safe magnetic field frequency and the three-dimensional spatial position information of the casting crack, the precision casting is subjected to high-frequency magnetic field excitation processing, and electromagnetic wave signals are collected in real time; Performing adaptive signal range gain on the collected electromagnetic wave signal to obtain an adaptive gain electromagnetic wave signal; Perform wavelet transform decomposition on the adaptive gain electromagnetic wave signal to extract sub-signals of different time scales; Multi-scale time-frequency feature analysis is performed on sub-signals of different time scales to obtain amplitude change curves, phase change curves and frequency change curves.

[0028] In this embodiment, a preset safe magnetic field frequency for precision castings is first determined based on the properties of the casting material and safety standards. This frequency should be within the tolerance range of the casting material to ensure no damage to the casting. For example, assume the preset safe magnetic field frequency is 60 Hz, determined based on the electromagnetic properties of the casting material and relevant standards. Based on the preset safe magnetic field frequency, a maximum safe magnetic field frequency is defined. Generally, this frequency should be below the critical failure frequency of the material to avoid potential material fatigue or damage. In this example, assume that analysis has determined a maximum safe magnetic field frequency of 50 Hz to ensure that the casting will not be damaged when excitation is performed at this frequency. Based on the defined maximum safe magnetic field frequency, the precision casting is subjected to high-frequency magnetic field excitation. This excitation process can be performed using electromagnetic excitation equipment (such as a high-frequency electromagnetic coil), ensuring a stable and controllable excitation frequency. For example, the excitation equipment is set to perform magnetic field excitation at a frequency of 50 Hz for a duration of 5 minutes to ensure sufficient stimulation of the material response within the casting. During the high-frequency magnetic field excitation process, electromagnetic wave signals from the surface and interior of the casting are collected in real time. These signals provide basic data for subsequent signal processing. Using a high-frequency oscilloscope and sensor, we can accurately capture changes in electromagnetic wave signals and record information such as signal amplitude, frequency, and phase. Because actual electromagnetic wave signals may contain significant noise or have low amplitudes, adaptive signal range gain is necessary to improve signal usability and accuracy. For example, suppose the initial signal amplitude is 0.1 mV, and the goal is to increase it to a usable level above 1 mV. Signal processing algorithms (such as adaptive gain control) are used to perform gain processing on the acquired electromagnetic wave signals. This algorithm dynamically adjusts the gain based on the real-time signal characteristics to optimize signal quality. During implementation, a dynamic gain factor is set and automatically adjusted based on signal strength to ensure that the final output signal quality meets the requirements of subsequent processing. Wavelet transform is an effective signal processing technique that can analyze signals simultaneously in the time and frequency domains and is suitable for processing non-stationary signals (such as electromagnetic wave signals). Selecting an appropriate wavelet basis (such as Haar wavelet or Daubechies wavelet) for the transform is crucial. For example, the Daubechies wavelet can be used as a basis function for signal decomposition to extract multi-level features. The adaptively gained electromagnetic wave signal is decomposed using a wavelet transform to extract sub-signals at different time scales. The wavelet transform decomposes the signal into components of different frequencies, facilitating subsequent analysis. For example, if the wavelet transform decomposition layer is set to four, the resulting sub-signals include low-frequency components (stationary components) and high-frequency components (transient components). Based on the sub-signals at different time scales, multi-scale time-frequency feature analysis is performed to extract amplitude, phase, and frequency curves. These features can reflect the internal state of the casting and possible crack information.For example, analyzing the amplitude variation of a sub-signal may reveal significant amplitude fluctuations at a specific time point, indicating a potential crack location. The amplitude, phase, and frequency of each sub-signal are calculated to generate a corresponding variation curve. Methods such as Fourier transforms or Hilbert transforms can be used to extract phase and frequency features. For example, key points of the amplitude variation curve, including maximum and minimum values ​​and their corresponding time points, are recorded to form complete characteristic data.

[0029] In this embodiment, step S4 includes the following steps: Calculating the amplitude attenuation rate of the amplitude change curve; The crack width is obtained by analyzing the dynamics of electromagnetic wave propagation based on the amplitude attenuation rate; Performing propagation delay effect analysis based on the phase change curve to extract the electromagnetic wave propagation delay value; identifying signal diffusion characteristics according to the frequency variation curve; Performing a crack depth quantitative analysis based on the electromagnetic wave propagation delay value and the signal diffusion characteristics to obtain the crack depth; The three-dimensional crack morphology is estimated based on the crack width and crack depth to obtain the three-dimensional crack morphology characteristics.

[0030] In this embodiment, in the previous step, the wavelet transform of the electromagnetic wave signal has been completed, and an amplitude change curve has been extracted from it. This curve shows the amplitude change of the electromagnetic wave signal over time. For example, suppose that the amplitude value drops from 1.5 mV to 0.5 mV over a period of time, and the time points of this change are recorded. The amplitude decay rate can be calculated by the slope of the amplitude change curve. A specified time interval Δt is selected, and the decay rate is calculated by the ratio of the amplitude change ΔA to the time change Δt. For example, if the amplitude drops from 1.5 mV to 0.5 mV within Δt = 2 seconds, the amplitude decay rate is calculated as: decay rate = ΔA / Δt = (0.5mV-1.5mV) / 2s = -0.5mV / s. Based on the calculated amplitude decay rate, an electromagnetic wave propagation dynamics analysis is performed. This analysis can help deduce the width of the crack, because the crack affects the propagation characteristics of the electromagnetic wave in the material. For example, an increase in the amplitude decay rate may indicate the presence of a crack, resulting in rapid attenuation of the electromagnetic wave. Based on propagation dynamics models (such as the wave equation or attenuation model), the crack width can be derived by correlating the amplitude decay rate with material properties (such as dielectric constant and conductivity). For example, assuming a decay rate of -0.5 mV / s at a specific frequency, the crack width can be calculated using the formula: Crack Width = k·Decay Rate, where k is a material constant. Assuming a k value of 2, the crack width is calculated to be 1 mm. The phase change curve obtained in the previous step is used to analyze the phase shift caused by the crack during electromagnetic wave propagation. The delay effect reflected by the phase change is crucial for crack feature extraction. For example, if the signal phase increases from π / 4 radians to π / 2 radians during a certain time period, the slope of the phase change curve can be calculated to determine the propagation delay of the electromagnetic wave. The relationship between phase change and time can be used to solve for the delay. For example, if the phase change is Δφ = π / 4 - π / 2 within a time period Δt = 1 second, the propagation delay can be calculated as: Propagation delay = Δφ / 2πf, where f is the signal frequency. Assuming f is 50 Hz, the propagation delay is calculated to be 0.025 seconds. Frequency analysis of the electromagnetic wave signal yields a frequency curve. This curve reflects the frequency variation characteristics of the signal during propagation and can indicate the diffusion effect of a crack. For example, a frequency drop from 50 Hz to 45 Hz is recorded over a certain period of time. The frequency curve can be used to identify the diffusion characteristics of the signal. The degree of signal diffusion can be determined by observing the frequency change rate and amplitude changes. For example, a frequency change rate of -5 Hz / s indicates that the signal is diffusing during propagation, which may be related to the presence of a crack. Based on the electromagnetic wave propagation delay and signal diffusion characteristics, the crack depth can be quantified. Crack depth generally affects the propagation characteristics of the electromagnetic wave and the signal attenuation. For example, deeper cracks result in more significant signal delay and attenuation.By establishing a model that relates crack depth to electromagnetic wave propagation characteristics and combining propagation delay with signal diffusion characteristics, the crack depth is calculated. For example, assuming the model derives that crack depth is proportional to propagation delay and the proportionality factor is 3, the crack depth can be calculated as: crack depth = 3 × propagation delay. Substituting a propagation delay of 0.025 seconds, the crack depth is 0.075 meters. Based on the acquired crack width and depth information, the three-dimensional morphological characteristics of the crack are estimated. This step can be achieved using a geometric model, combining the crack width and depth for a three-dimensional reconstruction. For example, assuming a crack width of 1 mm and a depth of 75 mm, the crack morphology is estimated based on the geometric model. The estimated three-dimensional crack morphological characteristics are recorded and visualized. Using 3D modeling software, the crack geometric characteristics are input to generate a complete three-dimensional crack model. For example, the generated 3D model can display the crack's specific morphology, including its direction, width, and depth, providing a basis for subsequent analysis and processing.

[0031] In this embodiment, the specific steps of step S5 are: Calculating signal strength fluctuations of the adaptive gain electromagnetic wave signal to obtain signal strength fluctuation characteristics; Mining the signal time series fluctuation trend based on the signal strength fluctuation characteristics to obtain the time series fluctuation trend law; The crack propagation evolution of the three-dimensional crack morphology is predicted at multiple time points based on the time series fluctuation trend to generate crack evolution prediction data. Calculate the potential crack propagation direction and speed based on crack evolution prediction data; Perform diffusion evolution path analysis on crack evolution prediction data and extract multiple diffusion evolution paths; Multi-time point diffusion fitting is performed based on the multiple diffusion evolution paths, potential crack diffusion directions and speeds to construct a crack evolution prediction map.

[0032] In this embodiment, in the previous steps, the electromagnetic wave signal has been adaptively gain processed. At this time, the signal after gain is used to calculate the signal intensity fluctuation characteristics. The signal should contain amplitude, phase and time information. For example, suppose that after excitation, the recorded signal intensities are 1.0 mV, 1.2 mV, 0.8 mV, 1.5 mV, and 1.1 mV, respectively. Calculate the standard deviation and mean of the signal intensity to quantify the signal fluctuation. The standard deviation can reflect the degree of discreteness of the signal intensity. The larger the standard deviation, the more obvious the fluctuation. For example, calculate the mean and standard deviation of the above signal intensity data: Mean: Mean = (1.0+1.2+0.8+1.5+1.1) / 5 =1.12mV, Standard Deviation = Based on the signal intensity fluctuation characteristics, the temporal fluctuation trend of the signal is analyzed. By mining the temporal trend of signal intensity, potential crack propagation information can be identified. For example, the temporal behavior of recorded signal intensity changes may show periodic fluctuations or linear decay. Signal intensity is modeled using time series analysis methods (such as the autoregressive moving average (ARMA) model) to identify the trend pattern of intensity fluctuations. For example, by fitting the ARMA model to the time series data, a trend line is generated, which allows identification of the rising and falling phases of the fluctuations and further analysis of their relationship with crack propagation. In the previous step, the three-dimensional morphological characteristics of the crack were obtained. Based on these characteristics and the temporal fluctuation trend of the signal, crack propagation evolution is predicted. For example, given a crack width of 1 mm and a depth of 75 mm, its future expansion can be predicted by combining the fluctuation trend. Numerical simulations (such as finite element analysis) can be combined with the signal intensity fluctuation trend to generate crack evolution prediction data. Crack propagation behavior can be simulated by considering different material properties and stress states. For example, assuming a crack could grow to 1.2 mm within an hour, simulations can be performed to obtain data on the crack's state at several future time points. Based on the crack evolution prediction data, the potential crack propagation direction can be inferred. The potential crack propagation path can be determined by analyzing the crack's 3D morphological characteristics and signal intensity fluctuation trends. For example, if the crack propagates primarily along the material's fiber direction, its potential propagation direction should also align with this direction. The crack propagation rate can be calculated by comparing the crack width changes at different time points. Assuming the crack width grows from 1 mm to 1.2 mm within an hour, the propagation rate is: Propagation rate = Δwidth / Δt = 1.2 mm - 1.0 mm / 1 hour = 0.2 mm / hour. Based on the crack evolution prediction data, the crack propagation path can be analyzed. Visualization tools can be used to convert the predicted data into an understandable path diagram, demonstrating the crack propagation trend. For example, using 3D modeling software, the crack's starting position and predicted propagation direction can be plotted as path lines. In path analysis, multiple diffusion evolution paths are extracted, and the characteristics of each path, including crack width, depth, and propagation rate, are recorded. These paths provide a basis for crack assessment and subsequent treatment. For example, the coordinate information of different paths is recorded, such as (1.0 mm, 0.0 mm, 75 mm) to (1.2 mm, 0.0 mm, 76 mm), representing the crack propagation. Based on the multiple diffusion evolution paths extracted, the potential crack propagation direction and velocity, a multi-point diffusion fitting is performed. Curve fitting methods can be used to establish a crack propagation model. For example, polynomial regression or spline interpolation methods can be used to generate a mathematical model of crack propagation to describe the dynamic changes of the crack. Finally, the fitting results are visualized to construct a crack evolution prediction diagram. This diagram should show the crack morphology and propagation path at different time points to facilitate subsequent analysis and decision-making.For example, different colors or lines can be used in the generated graph to represent the crack extension at different time stages, providing an intuitive reference for decision makers.

[0033] In this embodiment, the specific steps of step S6 are: Based on the three-dimensional spatial position information of the casting crack, the stress state at the crack is calculated by finite element method to obtain the stress state in the crack area; The stress distribution evolution is carried out according to the stress state of the crack area, and a crack stress distribution map is constructed; The crack evolution prediction diagram is optimized based on the crack stress distribution diagram to construct a preventive crack detection and protection model.

[0034] In this embodiment, in the previous steps, the three-dimensional spatial location information of the casting crack was obtained, including the crack width, depth, and specific coordinates within the casting. This information provides basic data for subsequent stress state calculations. For example, assume that the crack is located on the casting surface, has a depth of 75 mm, a width of 1 mm, and coordinates of (10 cm, 15 cm, 5 cm). A finite element model is established based on the casting's geometry and material properties. The model should include the casting's overall structure and the actual location of the crack. Finite element software (such as ANSYS, ABAQUS, etc.) can be used for modeling. For example, the casting's material properties are set to aluminum alloy, with an elastic modulus of 70 GPa and a Poisson's ratio of 0.33, and the crack's geometry is accurately reflected in the model. Appropriate boundary conditions and external loads are applied to the finite element model. Boundary conditions should take into account the way the casting is fixed in actual application, while loads should be set based on actual operating conditions (such as internal and external pressure, temperature fluctuations, etc.). For example, a concentrated force of 500 N is applied to the surface, while the bottom of the model is fixed to simulate actual conditions. Run a finite element analysis to calculate the stress state in the crack region. The analysis results will display the stress distribution around the crack, including principal stresses and shear stresses. For example, the calculation results show a maximum principal stress of 120 MPa and a minimum principal stress of 20 MPa in the crack region. These data will provide a basis for subsequent stress distribution evolution. Based on the stress state in the crack region, analyze the stress evolution over time or with changing conditions. This analysis helps understand the behavior and growth trends of cracks in real applications. For example, the stress around the crack may change over time or under varying loads, requiring dynamic analysis to capture these changes. Visualize the finite element analysis results to generate a crack stress distribution map. This stress distribution map provides a visual representation of the stress field around the crack, including areas of high and low stress concentrations. For example, color-coding can be used to represent different stress values, with red indicating high stress (>100 MPa) and blue indicating low stress (<30 MPa), making it easier to identify potential crack growth areas. Consider dynamic monitoring of the stress distribution during actual operation to capture stress trends over time and under varying environmental conditions. This can be achieved by installing strain gauges or other sensors. For example, strain gauges are installed at various locations on the casting to record stress changes in real time for long-term monitoring and analysis. In the previous step, a crack evolution prediction map was generated. This stress distribution map and crack evolution prediction map are combined and analyzed to optimize preventive process decisions. For example, the overlap between stress concentration areas and crack propagation paths is observed to identify high-risk areas and develop appropriate monitoring and maintenance plans. Based on the stress distribution and crack evolution prediction data, a preventive crack detection and protection model is constructed.

[0035] The preventive process decisions specifically include: Localized heating: Exploiting thermal stress relief mechanisms: For areas of high stress concentration, localized heating can be used to reduce mechanical stress. This approach utilizes a localized heat treatment to offset and weaken the existing mechanical stresses through a thermal stress relief mechanism. Localized heating devices, such as electric heating elements or infrared heaters, are installed around the high-stress areas. By adjusting the heating power and duration, the localized temperature rise can be precisely controlled, thereby inducing and relieving thermal stress.

[0036] Local cooling: using shrinkage stress to suppress cracks: For areas where cracks and other defects are prone to growth, localized cooling can be used to inhibit further crack expansion. This approach utilizes the contraction stress generated during the cooling process to inhibit crack opening and growth. Localized cooling devices, such as water cooling pipes or blast cooling devices, are installed in the area surrounding the defect. By controlling the cooling intensity and duration, the magnitude and distribution of contraction stress can be fine-tuned, thereby inhibiting crack growth.

[0037] External pressure application: using compressive stress to suppress cracks For some severe crack defects, external pressure can be applied to the casting surface to inhibit further expansion of the defect through compressive stress. A pressure-applying device, such as a hydraulic jack or compressed gas device, is placed on the surface of the defect area. By precisely controlling the magnitude and duration of the applied pressure, a certain compressive stress field is generated in the area surrounding the defect, thereby inhibiting the initiation and expansion of the crack.

[0038] Process parameter optimization: In addition to the aforementioned local process measures, overall stress levels can also be reduced by optimizing casting process parameters from a global perspective. For example, adjusting parameters such as pouring temperature, mold temperature, and pouring speed can alleviate stress concentrations caused by thermal and shrinkage stresses. Through multiple tests and simulation analyses, the optimal combination of process parameters that minimizes stress distribution is identified.

[0039] In this embodiment, a crack detection device for precision castings is further provided, which is used to perform the crack detection method for precision castings as described above, comprising: The acoustic imaging module is used to perform omnidirectional acoustic scanning of precision castings, collect acoustic reflection signals, and perform dynamic acoustic imaging mapping to construct a three-dimensional spatial soundprint map of the castings; The crack spatial positioning module is used to calculate the transient mutation phase position of the casting three-dimensional spatial soundprint and locate the casting crack space to obtain the three-dimensional spatial position information of the casting crack; The high-frequency magnetic field excitation module is used to perform high-frequency magnetic field excitation processing on the precision casting based on the three-dimensional spatial position information of the casting crack, and calculate the amplitude change curve, phase change curve and frequency change curve: A crack morphology estimation module is used to estimate the three-dimensional morphology of the crack based on the amplitude change curve, the phase change curve and the frequency change curve to obtain the three-dimensional morphological characteristics of the crack; The crack evolution prediction module is used to predict the crack diffusion evolution at multiple time points based on the three-dimensional crack morphology characteristics, perform multi-time point diffusion fitting, and construct a crack evolution prediction map; The preventive process decision module is used to optimize preventive process decisions based on the crack evolution prediction diagram and build a preventive crack detection and protection model.

[0040] The present invention also provides a computer system comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above-mentioned methods for detecting cracks in precision castings when executing the computer program.

[0041] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0042] Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0043] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it is stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution is embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media for storing program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0044] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present 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 present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting cracks in precision castings, characterized in that: The following steps are involved: Step S1: Perform omnidirectional acoustic scanning on the precision casting, collect acoustic reflection signals, and perform dynamic acoustic imaging mapping to construct a three-dimensional spatial soundprint map of the casting; Step S2: calculating the transient mutation phase position and spatially locating the casting crack on the casting three-dimensional soundprint image to obtain the three-dimensional spatial position information of the casting crack; Step S3: Based on the three-dimensional spatial position information of the casting crack, the precision casting is subjected to high-frequency magnetic field excitation processing, and the amplitude change curve, the phase change curve and the frequency change curve are calculated: Step S4: performing three-dimensional crack morphology estimation based on the amplitude variation curve, the phase variation curve, and the frequency variation curve to obtain three-dimensional crack morphology characteristics; Step S5: performing multi-time point crack diffusion evolution prediction on the three-dimensional crack morphological characteristics, and performing multi-time point diffusion fitting to construct a crack evolution prediction map; Step S6: Optimize preventive process decisions based on the crack evolution prediction graph and build a preventive crack detection and protection model.

2. The crack detection method for precision castings according to claim 1, characterized in that: The specific steps of step S1 are: Based on the high-density three-dimensional acoustic wave scanning array, the precision casting is scanned in all directions to collect the acoustic wave reflection signal; Calculating the timestamp of the acoustic wave reflection signal; calculating the time delay deviation of different signals on the acoustic wave reflection signal to obtain the signal time delay deviation; Performing time synchronization processing on the signal time delay deviation based on the timestamp to obtain a time-synchronized acoustic wave reflection signal; Performing phase correction on the time-synchronized acoustic wave reflection signal to obtain a phase-corrected optimized signal; Dynamic acoustic imaging mapping is performed based on the phase correction optimization signal to construct a three-dimensional spatial soundprint map of the casting.

3. The method according to claim 2, characterized in that The specific steps of performing dynamic acoustic imaging mapping based on the phase correction optimization signal to construct a three-dimensional spatial voiceprint map of the casting are as follows: Calculating the signal strength of the phase correction optimization signal; performing a three-dimensional spatial distribution analysis on the signal intensity to obtain three-dimensional spatial intensity distribution data; Perform multi-time sequence coherent superposition on the phase correction optimization signal to construct a time sequence superposition acoustic wave sequence; Identifying scanning points of the high-density three-dimensional acoustic wave scanning array; Calculating the spatial position coordinates of the scanning point; Signal intensity matching is performed on the three-dimensional spatial intensity distribution data based on the spatial position coordinates, and dynamic three-dimensional acoustic wave imaging mapping is performed according to the time-series superposition of acoustic wave sequences to construct a three-dimensional spatial soundprint map of the casting.

4. The method according to claim 1, wherein The specific steps of step S2 are: Perform acoustic texture recognition on the three-dimensional acoustic print of the casting to obtain acoustic texture features; Perform spatial filtering analysis on the acoustic wave texture features to extract the acoustic wave texture intensity and texture direction; Calculate the texture gradient based on the acoustic wave texture features to obtain the density of the texture gradient; Detect transient texture mutations based on texture gradient density, acoustic texture intensity, and texture direction, and mark transient texture mutation areas. Calculating the texture phase difference and amplitude difference between the transient mutation texture region and adjacent regions; Calculating the transient mutation phase position based on the texture phase difference and amplitude difference, thereby obtaining the relative position coordinates of the mutation texture in the voiceprint image; The precision casting is spatially positioned for the casting crack based on the relative position coordinates to obtain three-dimensional spatial position information of the casting crack.

5. The crack detection method for precision castings according to claim 1, characterized in that: The specific steps of step S3 are: Obtaining a preset precision casting safety magnetic field frequency, and defining a maximum safety magnetic field frequency according to the safety magnetic field frequency; Based on the maximum safe magnetic field frequency and the three-dimensional spatial position information of the casting crack, high-frequency magnetic field excitation processing is performed on the precision casting, and electromagnetic wave signals are collected in real time; Performing adaptive signal range gain on the collected electromagnetic wave signal to obtain an adaptive gain electromagnetic wave signal; Perform wavelet transform decomposition on the adaptive gain electromagnetic wave signal to extract sub-signals of different time scales; Multi-scale time-frequency feature analysis is performed on sub-signals of different time scales to obtain amplitude change curves, phase change curves and frequency change curves.

6. The crack detection method for precision castings according to claim 1, characterized in that: The specific steps of step S4 are: Calculating the amplitude attenuation rate of the amplitude change curve; The crack width is obtained by analyzing the dynamics of electromagnetic wave propagation based on the amplitude attenuation rate; Performing propagation delay effect analysis based on the phase change curve to extract the electromagnetic wave propagation delay value; identifying signal diffusion characteristics according to the frequency variation curve; Performing a crack depth quantitative analysis based on the electromagnetic wave propagation delay value and the signal diffusion characteristics to obtain the crack depth; The three-dimensional crack morphology is estimated based on the crack width and crack depth to obtain the three-dimensional crack morphology characteristics.

7. The crack detection method for precision castings according to claim 1, characterized in that: The specific steps of step S5 are: Calculating signal strength fluctuations of the adaptive gain electromagnetic wave signal to obtain signal strength fluctuation characteristics; Mining the signal time series fluctuation trend based on the signal strength fluctuation characteristics to obtain the time series fluctuation trend law; The crack propagation evolution of the three-dimensional crack morphology is predicted at multiple time points based on the time series fluctuation trend to generate crack evolution prediction data. Calculate the potential crack propagation direction and speed based on crack evolution prediction data; Perform diffusion evolution path analysis on crack evolution prediction data and extract multiple diffusion evolution paths; Multi-time point diffusion fitting is performed based on the multiple diffusion evolution paths, potential crack diffusion directions and speeds to construct a crack evolution prediction map.

8. The crack detection method for precision castings according to claim 1, characterized in that: The specific steps of step S6 are: Based on the three-dimensional spatial position information of the casting crack, the stress state at the crack is calculated by finite element method to obtain the stress state in the crack area; The stress distribution evolution is carried out according to the stress state of the crack area, and a crack stress distribution map is constructed; The crack evolution prediction diagram is optimized based on the crack stress distribution diagram to construct a preventive crack detection and protection model.

9. A crack detection device for precision castings, characterized in that: The method for detecting cracks in a precision casting according to claim 1 comprises: The acoustic imaging module is used to perform omnidirectional acoustic scanning of precision castings, collect acoustic reflection signals, and perform dynamic acoustic imaging mapping to construct a three-dimensional spatial soundprint map of the castings; The crack spatial positioning module is used to calculate the transient mutation phase position of the casting three-dimensional spatial soundprint and locate the casting crack space to obtain the three-dimensional spatial position information of the casting crack; The high-frequency magnetic field excitation module is used to perform high-frequency magnetic field excitation processing on the precision casting based on the three-dimensional spatial position information of the casting crack, and calculate the amplitude change curve, phase change curve and frequency change curve: A crack morphology estimation module is used to estimate the three-dimensional morphology of the crack based on the amplitude change curve, the phase change curve and the frequency change curve to obtain the three-dimensional morphological characteristics of the crack; The crack evolution prediction module is used to predict the crack diffusion evolution at multiple time points based on the three-dimensional crack morphology characteristics, perform multi-time point diffusion fitting, and construct a crack evolution prediction map; The preventive process decision module is used to optimize preventive process decisions based on the crack evolution prediction diagram and build a preventive crack detection and protection model.

10. A computer system comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the precision casting crack detection method according to any one of claims 1 to 7 are implemented.

Images