A cast iron casting method based on multi-stage flaw detection for controlling defects

Abstract

The present application relates to the technical field of metal casting process control, and particularly relates to a cast iron piece defect control casting method based on multi-stage flaw detection. The method comprises the following steps: performing water immersion ultrasonic flaw detection on raw material blocks to grade and screen target raw materials; performing online ultrasonic flaw detection on molten iron melt to regulate and control smelting purity; dynamically adjusting pouring flow field based on fusion of high-speed infrared thermal imaging and immersion ultrasonic echo data; monitoring solidification process by using phased array ultrasonic wave to homogenize sand mold cooling rate; verifying process parameter threshold by means of surface magnetic powder and internal ultrasonic double flaw detection after shakeout; and finally performing multi-dimensional composite flaw detection on the castings after heat treatment to determine the performance grading of the castings. The present application realizes full-process flaw detection feedback control from raw materials, smelting, pouring and solidification to heat treatment, significantly reduces the internal shrinkage, inclusion and crack defect rate of cast iron pieces, and improves the consistency of cast density and mechanical properties.

Description

Technical Field

[0001] This invention relates to the field of metal casting process control technology, and in particular to a casting method for controlling defects in cast iron parts based on multi-stage flaw detection. Background Technology

[0002] Cast iron parts are widely used in key fields such as machine tools, wind power, automobile manufacturing, and rail transportation due to their excellent casting performance, vibration damping performance, wear resistance, and machinability. As high-end equipment increasingly demands higher performance, reliability, and consistency in castings, extremely high requirements are placed on the precision of defect control in cast iron parts, especially for thick-section ductile iron castings and complex-structure vermicular graphite cast iron castings. The control of internal defects such as shrinkage porosity, inclusions, cracks, and gas pores directly determines the service performance and lifespan of the castings.

[0003] Chinese patent CN120028438B discloses a flaw detection device and method based on ductile iron pipe production. The device includes a ductile iron pipe mounted on a detection platform, and further includes: a positioning unit mounted on a processing box (a first processing box), comprising a drive assembly and a positioning assembly; and a flaw detection unit mounted on the detection platform, comprising a synchronization assembly and a detection assembly. The detection platform is equipped with a first processing box and a second processing box. This invention only detects finished ductile iron pipes and does not provide flaw detection feedback throughout the entire process from raw materials, smelting, casting, solidification to heat treatment, resulting in uncontrollable casting defects and a high defect rate in the finished product.

[0004] Therefore, there is an urgent need for a full-process flaw detection feedback control method, from raw materials, smelting, casting, solidification to heat treatment, to improve the consistency of casting density and mechanical properties. Summary of the Invention

[0005] To address this, the present invention provides a casting method for controlling defects in cast iron parts based on multi-stage flaw detection, which overcomes the problem in the prior art that it is impossible to achieve continuous online flaw detection throughout the entire casting process, and that it is impossible to provide real-time flaw detection feedback control from raw materials, smelting, pouring, solidification to heat treatment, resulting in uncontrollable defects in castings and a high defect rate in finished products.

[0006] To achieve the above objectives, the present invention provides a casting method for defect control of cast iron parts based on multi-stage flaw detection, comprising: Water immersion ultrasonic testing was performed on several raw material blocks used for smelting cast iron parts to obtain full-section scanning data of the raw material blocks; The defect volume ratio and graphitization segregation degree of each raw material block are determined based on full-section scanning data to determine the grade of each raw material. Target raw material blocks are selected by matching the raw material block grading results with the target defect rate; Online ultrasonic flaw detection is performed on the molten iron formed from the target raw material block to determine the purity of the molten iron, so as to adjust the smelting process parameters and obtain the castable molten iron; After casting, images of filling flow rate and temperature change rate are generated based on high-speed infrared thermal imaging data of casting. Ultrasonic echo signals of flowing molten iron are acquired based on immersion ultrasonic acquisition, so as to adjust casting process parameters based on temperature gradient and flow field state. The phased array ultrasonic flaw detection results of solidification of several consecutive castings are obtained, so as to adjust the initial temperature of the sand mold or the holding time based on the cooling difference between sand molds and the cooling difference within the same mold. Obtain the surface magnetic particle inspection and internal ultrasonic inspection results of the casting after sand removal to determine the effectiveness of the process correction parameters and output the corresponding process parameter thresholds. The castings that pass the initial flaw detection are subjected to heat treatment. The performance grading data of the castings is determined based on the multi-dimensional composite flaw detection results of the full size of the castings. The castings that meet the requirements are determined based on the performance grading data, and the raw material grading threshold is adjusted based on the grading percentage.

[0007] Furthermore, the step of performing water immersion ultrasonic testing on several raw material blocks for smelting cast iron parts to obtain full-section scanning data of the raw material blocks includes: Using a focused probe, spiral or layer-by-layer C-scan is performed on the raw material block to obtain ultrasonic A-wave signals at different depths within the raw material block; Fourier transform is performed on the A-wave signal to extract the bottom wave attenuation coefficient and the defect echo amplitude, and a three-dimensional full-section defect distribution model of the raw material block is reconstructed.

[0008] Furthermore, the determination of the defect volume ratio and graphitization segregation degree of each raw material block based on full-section scanning data to determine the grade of each raw material includes: The grade of each raw material is determined by comparing the defect volume ratio and graphitization segregation degree of each raw material block with the corresponding threshold.

[0009] Further, the online ultrasonic flaw detection of the molten iron formed from the target raw material block to determine the purity of the molten iron, in order to adjust the smelting process parameters and obtain the castable molten iron, includes: A high-temperature resistant waveguide rod is installed in the ladle or tundish, and a high-frequency ultrasonic pulse is emitted into the flowing molten iron using an electromagnetic ultrasonic transducer. Real-time reception of backscattered signals generated by non-metallic inclusions in molten iron; calculation of inclusion number density and average equivalent diameter per unit volume. When the number density of inclusions exceeds the first preset purity threshold, the inert gas bottom blowing time during the smelting process is increased or the inoculant addition ratio is adjusted until the real-time monitored number density of inclusions drops below the second preset purity threshold.

[0010] Furthermore, the step of generating images of filling flow rate and temperature change rate based on high-speed infrared thermal imaging data of casting, and acquiring ultrasonic echo signals of flowing molten iron based on immersion ultrasonic acquisition, to adjust casting process parameters based on temperature gradient and flow field state, includes: An array of immersion ultrasonic sensors is installed on the inner wall of the pouring cup or sprue to detect the turbulence intensity index and bubble entrainment frequency of the molten iron flow during the pouring process. The turbulence intensity index is fused with the temperature gradient at the filling front as reflected by infrared thermal imaging. If the temperature gradient at the filling front is less than the critical feeding temperature gradient, or the turbulence intensity index is greater than the laminar critical threshold, the pouring head height will be increased in real time or the tilt angle of the pouring cup will be adjusted by the servo control mechanism.

[0011] Furthermore, the step of obtaining the phased array ultrasonic flaw detection results of a series of castings solidified, and adjusting the initial temperature or holding time of the sand mold based on the cooling difference between sand molds and the cooling difference within the same mold, includes: The solidification front advance rate of the thick, hot-spotted parts of the casting is monitored using phased array probes arranged on the side wall of the sand box. Calculate the cooling difference between different castings in the same sand box and the cooling difference between sand molds at corresponding positions between different sand boxes in continuous production; When the cooling difference of the same type is greater than the first speed deviation threshold, adjust the thickness of the local chill arrangement of the corresponding sand mold. When the cooling difference between sand molds exceeds the second speed deviation threshold, adjust the preheating temperature of the subsequent sand molds or adjust the transfer cycle of the insulation box after pouring.

[0012] Furthermore, the acquisition of surface magnetic particle inspection and internal ultrasonic inspection results of the casting after sand removal to determine the effectiveness of process correction parameters and output corresponding process parameter thresholds includes: Automated magnetic particle testing is performed on castings after shot blasting to identify the length, direction, and location coordinates of surface and near-surface cracks. The magnetic particle inspection results were matched with the infrared thermal imaging temperature difference data during the casting stage to verify whether the adjustment parameters of the filling flow field eliminated the tendency of thermal cracking. A-type pulse-echo ultrasonic testing was performed on the castings to record the equivalent flat-bottom hole size and number distribution of defects. The measurement results were compared with the target defect rate to verify the effective boundary value of the smelting purity control parameters.

[0013] Furthermore, the output process parameter thresholds include the lower limit of the mass percentage of residual magnesium in the molten iron, the lower limit of the casting temperature, and the upper limit of the casting temperature.

[0014] Furthermore, the heat treatment of castings that pass the initial flaw detection, and the determination of casting performance grading data based on the multi-dimensional composite flaw detection results of the entire size of the casting, includes: During the heating and holding stages of heat treatment, a high-temperature acoustic emission sensor is used to monitor the count rate and energy release rate of acoustic emission events generated by the release of thermal stress inside the casting in order to determine whether there is microcrack propagation. After heat treatment, the casting was subjected to full-size ultrasonic phased array sector scanning and TOFD diffraction time-of-flight method flaw detection to obtain the diffraction signal at the defect tip; Based on the phase relationship of the diffraction signal and the height position of the defect tip in the wall thickness direction, combined with the attenuation mapping data of the matrix sound velocity, the matrix structure uniformity level and internal density level of the casting are output.

[0015] Furthermore, the raw material grading threshold is adjusted based on the grading percentage, including: The number of products that meet the target defect rate is determined based on the grading results, in order to determine the grading percentage; The percentage classification is compared with the percentage control threshold, where, If the grading percentage is less than the percentage control threshold, it is determined to adjust the defect volume ratio threshold and the graphitization segregation degree threshold in the raw material grading threshold to update the grading index.

[0016] Compared with existing technologies, the beneficial effects of this invention are that it breaks through the limitation of traditional casting methods that only conduct quality judgment in the final stage. It integrates multi-dimensional flaw detection technologies, such as water immersion ultrasonics, online electromagnetic ultrasonics, infrared thermal imaging, phased array ultrasonics, TOFD, and acoustic emission, throughout the entire process, from raw material selection, smelting and purification, casting and filling, solidification and cooling, sand removal and cleaning, and heat treatment strengthening. Through real-time feedback of flaw detection data at each stage and dynamic closed-loop adjustment of process parameters, it can suppress defect inheritance at the source, eliminate defect inducing factors during the process, and ultimately accurately evaluate product grade. This significantly reduces the rate of shrinkage porosity, gas porosity, inclusions, and hot cracks inside castings, and improves the density and mechanical property consistency of castings.

[0017] Furthermore, by performing water immersion ultrasonic C-scan on the raw material blocks and classifying them based on the defect volume ratio and the degree of graphitization segregation, the present invention can accurately remove unqualified furnace materials containing original shrinkage cavities or severe graphite distortion, effectively preventing internal defects of the raw materials from being inherited to the molten iron matrix during the remelting process, ensuring the initial purity of the molten iron from the source, and reducing the difficulty of subsequent processing.

[0018] Furthermore, this invention utilizes a high-temperature resistant waveguide and an electromagnetic ultrasonic transducer to achieve online quantitative detection of the purity of flowing high-temperature molten iron. Based on a backscattering signal-driven strategy for dynamically adjusting the inert gas bottom blowing time or inoculant addition amount, the density of inclusions in the molten iron is precisely controlled, ensuring that the quality of the molten iron is within the optimal process window before casting.

[0019] Furthermore, this invention performs spatiotemporal fusion analysis of high-speed infrared thermal imaging data and immersion ultrasonic echo signals to establish a correlation model between filling flow velocity, leading-edge temperature gradient, and turbulence intensity. This allows for the early prediction of the formation trends of cold shuts, incomplete filling, and air entrapment defects. By intervening in the pouring posture and pressure head in real time through a servo mechanism, a stable laminar flow state and a reasonable temperature gradient distribution during the filling process are ensured, effectively suppressing the generation of oxide inclusions and porosity defects.

[0020] Furthermore, this invention employs phased array ultrasonic monitoring to assess the solidification front advancement velocity at thick, hot spots in the casting. Based on dual criteria of cooling differences between the same mold and between sand molds, it intelligently adjusts the arrangement of local chills or the preheating temperature of the sand mold. This overcomes the problem of inconsistent solidification rates caused by heat storage fluctuations during multi-cavity sand boxes and continuous production, significantly reducing the range of microstructure differences and mechanical property fluctuations between castings from the same batch and across batches.

[0021] Furthermore, by retrospectively comparing the results of surface magnetic particle testing and internal ultrasonic testing after sand removal, this invention not only verifies the effectiveness of the previous flow field control and purity control measures, but also locks in the threshold values ​​of core process parameters, including residual magnesium content and casting temperature range. The parameter boundary output method based on measured flaw detection data replaces the traditional method relying on experience-based trial and error, providing precise digital basis for process standardization.

[0022] Furthermore, this invention introduces high-temperature acoustic emission technology to monitor the tendency of microcrack propagation during the heating and holding stages of heat treatment. Combined with subsequent full-size ultrasonic phased array sector scanning, time-of-flight diffraction (TOFD) flaw detection, and matrix sound velocity attenuation mapping, the homogeneity and internal density of the casting's matrix structure are comprehensively evaluated. This not only accurately identifies Grade A castings that meet high-precision load-bearing requirements but also guides the downgrading of non-critical castings, maximizing the rational utilization of resources and avoiding unnecessary scrap and waste. Attached Figure Description

[0023] Figure 1 This is a flowchart of the casting method for defect control of cast iron parts based on multi-stage flaw detection according to the present invention. Figure 2 This is a schematic diagram of the process for water immersion ultrasonic flaw detection and grading of raw material blocks according to the present invention; Figure 3 This is a schematic diagram illustrating the adjustment of the casting process according to the present invention; Figure 4 This is a schematic diagram of the casting performance grading and raw material threshold feedback process of the present invention. Detailed Implementation

[0024] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0025] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0026] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0027] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0028] Cast iron parts are widely used in key fields such as machine tools, wind power, automobile manufacturing, and rail transportation due to their excellent casting performance, vibration damping performance, wear resistance, and machinability. As high-end equipment increasingly demands higher performance, reliability, and consistency in castings, extremely high requirements are placed on the precision of defect control in cast iron parts, especially for thick-section ductile iron castings and complex-structure vermicular graphite cast iron castings. The control of internal defects such as shrinkage porosity, inclusions, cracks, and gas pores directly determines the service performance and lifespan of the castings.

[0029] Cast iron parts (such as ductile iron and gray cast iron) are widely used in high-end equipment manufacturing. Their internal defects (such as shrinkage porosity, gas porosity, inclusions, and graphite distortion) directly affect service life and safety. Traditional casting quality control relies heavily on destructive sampling inspection of the final product or single offline flaw detection, which has the following technical bottlenecks: Insufficient raw material control: If the original shrinkage cavities or graphitization segregation in the raw pig iron or scrap steel are not detected, they will be inherited to the casting matrix after remelting. Process black box: There is a lack of real-time quantitative detection methods for the dynamic changes of inclusions during the melting stage, turbulent gas entrapment during the pouring stage, and temperature field imbalance during the solidification stage, resulting in serious delays in process adjustment. Single grading basis: There is a lack of a multi-dimensional data fusion evaluation system that runs through the entire process, making it impossible to accurately classify the performance grade of castings based on real-time flaw detection data.

[0030] Based on this, please refer to Figure 1 The diagram shows a flowchart of the casting method for defect control of cast iron parts based on multi-stage flaw detection according to the present invention. An embodiment of the present invention provides a casting method for defect control of cast iron parts based on multi-stage flaw detection, comprising: Step S1, Water immersion ultrasonic testing and grading of raw material blocks: Water immersion ultrasonic testing was performed on several raw material blocks for smelting cast iron parts to obtain full-section scanning data of the raw material blocks; based on the full-section scanning data, the defect volume ratio and graphitization segregation degree of each raw material block were determined to determine the grade of each raw material; the raw material block grading results were matched with the target defect rate to screen the target raw material blocks.

[0031] Step S2, Online ultrasonic purification control of molten iron: Online ultrasonic flaw detection is performed on the molten iron formed from the target raw material block to determine the purity of the molten iron, so as to adjust the smelting process parameters and obtain the castable molten iron.

[0032] Step S3, Infrared-ultrasonic coupled flow field control during casting: After casting, images of filling flow rate and temperature change rate are generated based on high-speed infrared thermal imaging data of casting. Ultrasonic echo signals of flowing molten iron are acquired based on immersion ultrasonic acquisition, so as to adjust casting process parameters based on temperature gradient and flow field state.

[0033] Step S4, solidification process phased array ultrasonic homogenization cooling control: The phased array ultrasonic flaw detection results of several consecutive solidified castings are obtained, so as to adjust the initial temperature or holding time of the sand mold based on the cooling difference between sand molds and the cooling difference within the same mold.

[0034] Step S5, Dual-mode flaw detection verification and parameter threshold locking after sand removal: Obtain the surface magnetic particle inspection and internal ultrasonic inspection results of the casting after sand removal to determine the effectiveness of the process correction parameters and output the corresponding process parameter thresholds.

[0035] Step S6: Multi-dimensional composite flaw detection and performance grading after heat treatment, and raw material threshold feedback: The castings that pass the initial flaw detection are subjected to heat treatment. The performance grading data of the castings is determined based on the multi-dimensional composite flaw detection results of the full size of the castings. The castings that meet the requirements are determined based on the performance grading data, and the raw material grading threshold is adjusted based on the grading percentage.

[0036] In this embodiment, step S1: Water immersion ultrasonic testing and grading of raw material blocks: Water immersion ultrasonic testing is performed on several raw material blocks for smelting cast iron parts to obtain full-section scanning data of the raw material blocks; based on the full-section scanning data, the defect volume ratio and graphitization segregation degree of each raw material block are determined to determine the grading of each raw material; the grading results of the raw material blocks are matched with the target defect rate to screen target raw material blocks. Before smelting, the internal quality of the raw material blocks is quantitatively evaluated to screen out raw materials that meet the quality requirements of the target casting, thus preventing the inheritance of inherent defects in the raw materials from the source.

[0037] Specifically, full-section scanning data acquisition involves immersing the raw material block to be inspected (such as pig iron ingots or scrap steel blocks) in a water tank, using water as a coupling agent. A focused ultrasonic probe is used to perform a C-scan on the surface of the raw material block using a helical or layered grid trajectory. C-scan is an important data presentation and imaging method in ultrasonic non-destructive testing, generating a two-dimensional planar image of a cross-section of the inspected object perpendicular to the ultrasonic beam. During the scan, an ultrasonic pulse transmitter / receiver records the A-wave signal at each scanning point, i.e., the amplitude-time waveform of the ultrasonic wave penetrating the thickness of the raw material block. After data acquisition, the A-wave signal at each point is subjected to Fourier transform processing to extract two key feature parameters: Bottom wave attenuation coefficient: The ratio of the amplitude of the echo reflected from the bottom surface of the raw material block after the ultrasonic wave penetrates it to the attenuation of the echo from a standard defect-free test block. This parameter reflects the overall absorption and scattering of ultrasonic energy within the raw material block.

[0038] Defect echo amplitude: The height of the abnormal reflected wave that appears before the bottom wave, representing the presence and size of discontinuous interfaces such as internal shrinkage cavities and inclusions.

[0039] The A-wave feature data obtained from full-surface scanning is reconstructed by three-dimensional spatial interpolation to generate a three-dimensional full-section defect distribution model of the raw material block, which intuitively displays the location, shape and relative size of internal defects.

[0040] Please see the appendix Figure 2 This is a schematic diagram of the process for water immersion ultrasonic flaw detection and grading of raw material blocks according to the present invention.

[0041] Steps for generating a 3D full-section defect distribution model: Data Acquisition and Spatial Coordinate Binding: The raw material block is fixed on the water immersion scanning platform, and a three-dimensional rectangular coordinate system is established. In this embodiment, a corner of the upper surface of the raw material block is taken as the origin (0,0,0), the X and Y axes define the horizontal scanning plane, and the Z axis defines the depth direction. The scanning device drives the focusing probe to move along a preset trajectory. The trajectory type can be selected from the following two: Spiral Scan: The probe starts from the center or edge of the raw material block and moves along a spiral path at a constant linear velocity, suitable for round or irregularly shaped raw materials. Layer-by-Layer Scan: After the probe completes a horizontal layer by scanning line by line at a fixed row spacing, it enters the next depth layer by adjusting the focusing depth of the probe or moving the height of the raw material block, suitable for square raw materials. At each scanning point (Xi,Yi,Zi), the ultrasonic flaw detector emits an ultrasonic pulse and records the A-wave signal waveform data of that point completely. The information contained in each data point is: spatial coordinates (X,Y,Z) and the corresponding time-amplitude sequence.

[0042] A-wave signal feature extraction: Time-domain analysis was performed on the A-wave signal recorded at each scanning point to extract the following two key feature values: (1) Bottom wave attenuation coefficient (B) First, the reference bottom wave amplitude B0 is measured on a defect-free standard test block. Then, in the measured A-wave signal of the raw material block, the wave packet corresponding to the bottom depth is located, and its amplitude B1 is read. The bottom wave attenuation value at this point is calculated: B = B1 / B0; The B value ranges from 0 to 1. The smaller the value, the greater the loss of ultrasonic energy penetration, reflecting the more severe the sound attenuation inside the material at that point, which means there is a higher possibility of defects or loose tissue.

[0043] (2) Defect echo amplitude (F) On the time axis of the A-wave signal, the segment between the surface wave and the bottom wave is extracted. The maximum echo amplitude Fmax within this segment is searched. If Fmax exceeds a preset noise threshold, Fmax is recorded; otherwise, F is recorded as 0. An F value greater than 0 indicates the presence of an abnormal reflection interface on the sound beam path, i.e., defects such as shrinkage cavities or inclusions exist inside; the larger the F value, the stronger the reflection from the defect interface, and the larger the relative size of the defect.

[0044] 3D Mesh Generation and Data Mapping: The 3D spatial shape of the raw material block is divided into a regular cubic mesh (i.e., a voxel mesh). The side length of the voxel can be set according to the detection accuracy requirements, for example, 1mm×1mm×1mm or 2mm×2mm×2mm. For each voxel unit, a value that comprehensively characterizes its internal quality needs to be determined. Due to the diffusion effect during the penetration of the sound beam, the actual ultrasonic response of a single voxel is affected by the combined influence of multiple voxels along its sound beam path. This embodiment adopts a simplified engineering processing method: Voxel assignment rule: When a defect echo F>0 is detected on a certain sound beam path, all voxels on that path from the depth position corresponding to the defect echo to the bottom depth position are marked as the defect-affected area, and their defect characteristic value is assigned as Fmax. Bottom wave attenuation allocation: For the area where no defect echo is detected (i.e., F=0), the quality characteristic value V is determined according to the bottom wave attenuation value B corresponding to the sound beam path.

[0045] After traversing all scan points, each voxel unit obtains a quality characteristic value V. The higher the V value, the higher the probability of defects or tissue abnormalities within that tiny unit.

[0046] The quality characteristic value V is calculated using the following formula: For defective echoes: V = F ÷ F_ref; For defect-free echoes: V = 1 - B; Where F is the defect echo height; F_ref is the reference height, which is based on a standard-sized artificial defect test block. In this embodiment, a test block with a Φ2mm flat-bottom hole defect is taken, and the same probe is used to measure it once, and its defect echo height is recorded. B represents the degree of backwave attenuation, ranging from 0 to 1, where B = B1 / B0. Wherein, B0 is the measurement of a defect-free standard test block, and its bottom wave height is recorded; B1 is the measurement of an actual raw material block, and the bottom wave height B1 at the same depth is recorded. The smaller B is, the more severe the attenuation of ultrasound in the material, and the more likely there is a problem inside.

[0047] Visualization rendering output: Import the assigned 3D voxel array into the visualization software module and map the color scale according to the size of the feature value V, as shown in Table 1.

[0048] ; The determination of defect volume ratio and graphitization segregation degree is based on the above three-dimensional defect distribution model, and two core indicators are quantitatively calculated: (1)Determination of the proportion of defect volume: Voxels in the three-dimensional model with echo amplitudes exceeding the preset defect determination threshold, that is, exceeding the preset noise determination threshold, are marked and counted. Calculate the percentage of the number of marked voxels in the total number of analyzed voxels of the raw material block, which is the proportion of the defect volume of the raw material block. This index directly reflects the severity of macroscopic defects such as shrinkage cavities and porosity inside the raw material.

[0049] Setting of the defect determination threshold: In the system initialization stage, a standard test block made of the same material as the raw material to be inspected and confirmed to have no internal defects by X-ray flaw detection is used for ultrasonic calibration. Record the maximum value of the noise level in the A-wave signal at each depth position of the standard test block, and take a value within the range of 2 to 3 times of this maximum value as the defect echo determination threshold (F_th). When the echo amplitude F corresponding to a certain voxel satisfies F≥F_th, it is determined that the voxel is a defective voxel.

[0050] Voxel marking and counting: Traverse all voxel units in the three-dimensional full-section defect distribution model, and compare the defect echo characteristic value F of each one with the determination threshold F_th one by one: If F≥F_th, mark the voxel as "1" (defective voxel); If F<F_th, mark the voxel as "0" (normal voxel).

[0051] Count the total number of voxels marked as "1", denoted as N_defect. At the same time, count the total number of voxels included in the raw material block model, denoted as N_total Calculation of the proportion of defect volume: The proportion of defect volume P_defect is calculated by the following formula: P_defect=(N_defect / N_total)×100%; The numerical value of the calculation result is positively correlated with the severity of the defect. For the typical reference value range of P_defect and its corresponding meanings, please refer to Table 2.

[0052] ; (2)Determination of the degree of graphitization segregation: The uneven graphitization morphology of carbon in the cast iron raw material (i.e., graphitization segregation) will cause a slight change in the ultrasonic sound velocity. By measuring the propagation time of ultrasonic waves in different regions inside the raw material block, calculate the sound velocity values of each region. Divide the difference between the maximum and minimum sound velocities in the region by the average sound velocity of the raw material block, and the obtained ratio is the degree of graphitization segregation. This index reflects the uniformity of the existing form of carbon inside the raw material. When the segregation is severe, it will cause abnormal local graphite morphology in the casting after remelting.

[0053] Raw material block zoning and sound velocity measurement: The three-dimensional model of the raw material block is evenly divided into several sub-regions along the X, Y, and Z directions. The number of sub-regions is determined according to the size of the raw material block and the detection accuracy requirements. Typically, it is divided into 3×3×3 or 5×5×5 sub-regions. For each sub-region, use an ultrasonic detection system to measure the longitudinal wave sound velocity of ultrasonic waves in this region. The measurement method is as follows: Select the sound beam path corresponding to the center position of this sub-region; accurately measure the time difference Δt of the ultrasonic wave from the upper surface of the raw material block to the bottom surface and back; given the thickness d of the raw material block at this position (directly obtained from the coordinates of the scanning system), the sound velocity v of this sub-region is: v = 2d / Δt; Repeat the measurement of several sound beam paths within this sub-region, and take the average value as the representative sound velocity value of this sub-region.

[0054] Calculation of the average sound velocity of the raw material block: Statistically analyze the sound velocity measurement values of all sub-regions, calculate their arithmetic mean, denoted as v_avg, which is the overall average sound velocity of this raw material block.

[0055] Calculation of the degree of graphitization segregation: Among the sound velocity values of all sub-regions, find the maximum value v_max and the minimum value v_min. The degree of graphitization segregation S is calculated according to the following formula: S = (v_max - v_min) / v_avg; This ratio is dimensionless. The larger the value, the greater the difference in sound velocity inside the raw material block, that is, the more uneven the distribution of the graphitization morphology.

[0056] Evaluation of the degree of segregation: For the typical reference value range and corresponding meanings of the graphitization segregation degree S, please refer to Table 3.

[0057] ; Comprehensive grading application: In actual grading operations, the proportion of defective volume P_defect is used as the main judgment basis, and the graphitization segregation degree S is used as an auxiliary correction basis. The correction rules are as follows: When the P_defect value of the raw material block is near the boundary values of two grades (for example, 0.45% - 0.5%, 1.9% - 2.0%, or 4.6% - 5.0%), if S > 0.05, then downgrade this raw material block by one level; If the P_defect value clearly falls within a certain interval and S < 0.02, then maintain the original grading unchanged.

[0058] Raw material grading determination: Set the grading thresholds T1 and T2 for the proportion of defective volume, and satisfy T1 < T2. The specific values of the thresholds can be pre-calibrated according to the quality grade requirements of the target casting. In this embodiment, T1 is taken as 0.5% and T2 is taken as 2.0%. For the grading rules, please refer to Table 4.

[0059] ; Meanwhile, the degree of graphitization segregation serves as an auxiliary criterion for judgment: when the defect volume ratio of a raw material block is at the grade boundary, if the degree of graphitization segregation exceeds the preset segregation threshold, the raw material block will be downgraded by one grade to fully ensure the uniformity of graphite morphology in molten iron.

[0060] Target raw material block screening: Based on the target defect rate requirements of the castings to be produced, raw material blocks of the corresponding grade are matched from the raw material library. If the production of precision load-bearing structural parts has strict target defect rate requirements, only Grade 1 raw materials are selected for the furnace. If the production of ordinary non-critical castings is carried out, a certain proportion of Grade 2 raw materials can be appropriately added. Through the above-mentioned graded screening, it is ensured that the quality of the raw material blocks entering the melting process matches the quality requirements of the final product, reducing the risk of casting defects from the source. For example, if the target defect rate is set at a porosity / shrinkage cavity ratio of less than or equal to 3%, the corresponding matching is to select only Grade 1 raw materials for the furnace; if the target defect rate is set at a porosity / shrinkage cavity ratio of less than or equal to 5%, the corresponding matching is to select Grade 1 raw materials for the furnace and combine them with Grade 2 raw materials with a mass of 20% Grade 1 raw materials.

[0061] In this embodiment, step S2: Online ultrasonic purity control of molten iron: Online ultrasonic flaw detection is performed on the molten iron formed from the target raw material block to determine the purity of the molten iron, so as to adjust the smelting process parameters to obtain the castable molten iron. During the transfer or settling stage from the completion of molten iron smelting to before casting, the content of non-metallic inclusions inside the high-temperature molten iron is monitored in real time online, and the smelting or ladle treatment process parameters are dynamically adjusted according to the monitoring results to ensure that the purity of the molten iron entering the mold cavity meets the quality requirements of the target casting.

[0062] Specifically, the hardware configuration of the online ultrasonic testing system: The online ultrasonic testing system is deployed at key workstations where molten iron is transferred or left to settle, and includes the following components: (1) The high-temperature resistant waveguide is made of silicon nitride ceramic or corundum. One end is immersed to a specific depth (usually 50-150 mm) below the surface of the molten iron, and the other end is exposed above the surface of the molten iron and connected to the ultrasonic transducer. The function of the waveguide is to guide the ultrasonic waves into the high-temperature molten iron while protecting the transducer from high-temperature corrosion. The immersed end of the waveguide is designed to be conical or wedge-shaped to optimize the directivity of the sound beam.

[0063] (2) The electromagnetic ultrasonic transducer (EMAT) is installed on the top of the exposed end of the waveguide rod and uses the non-contact electromagnetic induction principle to excite and receive ultrasonic waves. Compared with traditional piezoelectric transducers, EMAT does not require coupling agent, can withstand the high temperature conducted by the waveguide rod, and is suitable for continuous operation environment in industrial sites.

[0064] (3) The signal transmission and acquisition unit includes a high-frequency pulse generator, a preamplifier, an A / D conversion module, and an industrial control computer. The transmitting unit generates high-frequency electrical pulses with a frequency range of 2 to 5 MHz, which are converted into ultrasonic waves by EMAT and transmitted to the molten iron along the waveguide rod; the receiving unit acquires the ultrasonic backscattered signal returned from the molten iron in real time.

[0065] Inclusion detection principle: When ultrasound propagates in molten iron, it encounters non-metallic inclusion particles (such as oxides). Inclusions (such as sulfides, MnS, and silicates) cause significant acoustic impedance differences between the inclusions and the molten iron matrix, resulting in some ultrasonic energy being scattered in all directions. The signal that is backscattered back towards the waveguide is called the backscatter signal. The intensity of the backscatter signal is proportional to the square or higher power of the inclusion particle size; the number of backscattered signals received per unit time is positively correlated with the inclusion number density per unit volume of molten iron. Therefore, by analyzing the amplitude distribution and temporal occurrence rate of the backscatter signal, the purity of the molten iron can be quantitatively assessed.

[0066] Real-time calculation of purity index: The industrial control computer continuously samples and analyzes the received backscattered signal. The sampling period can be set to update the purity index every 1 second or every 5 seconds.

[0067] (1) Identification of effective scattering events A backscattering amplitude threshold (A_th) is set. This threshold is determined through a calibration experiment: the amplitude of the backscattered signal is measured in molten iron with a known purity that meets the requirements, and three times the average amplitude is taken as A_th. When the amplitude of the received backscattered signal exceeds A_th, it is counted as a valid scattering event.

[0068] (2) Calculation of inclusion number density (N_d) The number of effective scattering events occurring within a unit of time (e.g., 1 minute) is denoted as n. Combined with the effective detection volume V_s of the waveguide beam in molten iron (this volume depends on the beam spread angle and the waveguide immersion depth, and can be obtained through acoustic field simulation or standard test block calibration), the inclusion number density N_d is: N_d = n / V_s; The unit is cells / cm³ or cells / mL.

[0069] (3) Calculation of average equivalent diameter (D_avg) The signal amplitude A_i of each effective scattering event is recorded. According to ultrasonic scattering theory, in the Rayleigh scattering region (where the inclusion size is much smaller than the wavelength), the backscattering amplitude is proportional to the inclusion volume. An amplitude-equivalent diameter calibration curve is pre-established using experimental data, and the equivalent diameter d_i is obtained by mapping each amplitude A_i to the curve. Then, the average diameter of all effective scattering events within this statistical period is calculated as the average equivalent diameter. D_avg=(d_1+d_2+...+d_n) / n Where i takes values ​​from 1, 2, ..., n, and n is the total number of effective scattering events. Purity threshold setting and process parameter linkage control: Based on the quality grade requirements of the target casting, two levels of purity thresholds are preset: First preset purity threshold (high limit threshold): The first threshold for the number density of impurities is 50 particles / cm³, or the first threshold for the average equivalent diameter is 20μm. When the purity deteriorates to this threshold, process intervention is triggered. In this embodiment, the process intervention is triggered when the number density of impurities N_d > 50 particles / cm³, or the average equivalent diameter D_avg > 20μm.

[0070] The second preset purity threshold (target threshold) is: the second threshold for the number density of impurities is 10 particles / cm³, or the second threshold for the average equivalent diameter is 10μm. The goal of the process intervention is to improve the purity to below this threshold, for example, N_d≤10 particles / cm³ and D_avg≤10μm.

[0071] When the online monitoring system detects that the current purity index exceeds the first preset purity threshold, the PLC control system automatically triggers one or more of the following adjustment actions until the purity index falls back below the second threshold. Please refer to Table 5.

[0072] ; Pouring Permission Determination: The system continuously monitors the purity of the molten iron. When the following conditions are simultaneously met, the molten iron is deemed to be of acceptable quality, and a "pouring permission" signal is sent to the pouring control system: The number density of inclusions N_d ≤ the corresponding second preset purity threshold, that is, N_d ≤ 10 inclusions / cm³; The average equivalent diameter D_avg ≤ the corresponding second preset purity threshold, that is, D_avg ≤ 10μm; The above state must remain stable for at least 30 seconds.

[0073] If the purity index still fails to meet the standard after multiple process adjustments, the system will issue an alarm and suggest downgrading the ladle of molten iron or returning it to the electric furnace for refining.

[0074] Please see Figure 3This is a schematic diagram illustrating the adjustment of the casting process according to the present invention.

[0075] In this embodiment, step S3: Infrared-ultrasonic coupled flow field control during the pouring process: After pouring, images of the filling flow rate and temperature change rate are generated based on high-speed infrared thermal imaging data of the pouring process. Ultrasonic echo signals of the flowing molten iron are acquired using immersion ultrasonic technology to adjust the pouring process parameters based on the temperature gradient and flow field state. During the dynamic process of molten iron filling the mold cavity, the temperature field distribution and flow turbulence of the molten iron flow are simultaneously sensed, and the two types of information are fused and analyzed to determine in real time whether the filling state deviates from the ideal process window. Then, the pouring parameters are dynamically intervened through a servo actuator to prevent filling defects such as cold shuts, insufficient pouring, air entrapment, and oxide inclusions.

[0076] Specifically, step S3 includes hardware system configuration and data acquisition, image and signal processing and feature parameter extraction, data fusion and process criterion definition, and dynamic adjustment of casting process parameters.

[0077] 1. Hardware system configuration and data acquisition (1) High-speed infrared thermal imaging subsystem A high-speed infrared thermal imager is fixedly installed directly above or at an angle to the pouring station, with its lens vertically or at an angle aimed at the visible areas of the pouring cup, sprue, and cavity of the mold. The thermal imager parameters are as follows: frame rate ≥ 50fps, temperature measurement range covering 1200℃~1450℃, spatial resolution ≤ 2mm / pixel. After pouring begins, the thermal imager continuously records a sequence of infrared thermal images of the entire filling process and transmits them in real time to a host industrial computer via industrial Ethernet.

[0078] (2) Immersive ultrasonic flow field sensing subsystem An immersion ultrasonic sensor array is pre-embedded at a specific location on the inner wall of the pouring cup or the side wall of the sprue. This array consists of at least one pair of high-temperature resistant piezoelectric ceramic wafers (a transmitting probe and a receiving probe), with the sensing surface directly contacting the molten iron being tested. The transmitting probe continuously emits ultrasonic waves at a frequency of 1MHz to 2MHz into the flowing molten iron, while the receiving probe collects the ultrasonic signals penetrating the molten iron flow in real time. These signals are then conditioned by a preamplifier and synchronously transmitted to a host industrial control computer.

[0079] 2. Image and signal processing and feature parameter extraction (1) Infrared image processing and thermodynamic parameter extraction The industrial control computer processes each frame of the infrared thermal image as follows: Temperature field reconstruction: Based on the preset emissivity of molten iron (0.3~0.4), the grayscale values ​​are converted into absolute temperature values ​​to generate a two-dimensional temperature distribution map.

[0080] Filling front identification: A temperature threshold (more than 200℃ higher than the background sand mold temperature) is set to extract the contour edge of the molten iron surface. By comparing the displacement of the same feature edge point in two adjacent image frames and dividing by the inter-frame time interval, the local filling velocity at that point is calculated. The average velocity of all filling front points is obtained by statistically averaging the velocities.

[0081] Temperature change rate calculation: First-order difference operation is performed on the time-series temperature values ​​of each pixel location to generate a pseudo-color map of temperature change rate. Along the filling flow direction, the temperature difference and distance between two spatially adjacent points are extracted to calculate the temperature gradient at the filling front (G_front), in °C / cm. G_front directly reflects the rate of heat loss of molten iron during the filling process. The critical feeding temperature gradient is preferably set to 5 °C / cm.

[0082] (2) Ultrasonic signal processing and flow field parameter extraction The industrial control computer performs real-time analysis on the continuous ultrasonic signals acquired by the receiving probe. Turbulence intensity index (I_turb) calculation: The received ultrasonic signal amplitude sequence is extracted within a certain time window (e.g., 0.5 seconds), and the ratio of its standard deviation to the average amplitude is calculated; this is I_turb. Under steady laminar flow conditions, the I_turb value is small (typically <0.3); when turbulent eddies occur in the liquid flow, the acoustic impedance fluctuations along the sound beam path intensify, the received amplitude jitter is severe, and the I_turb value increases significantly.

[0083] Bubble entrapment frequency (f_bubble) detection: When the transmit-receive mode is switched to pulse echo mode (or an independent probe is used), the presence of short-duration strong reflection pulses exceeding a preset threshold in the echo signal is analyzed. In this embodiment, the preset threshold is set as follows: the pulse width of the echo signal is set to ≤3μs (2MHz probe) or ≤1.5μs (5MHz probe), and the amplitude is set to exceed the background noise amplitude by 8 to 15 times, preferably 10 times. When the set pulse and amplitude are simultaneously satisfied, it is recorded as a bubble entrapment event. The number of such pulses detected per unit time is recorded as f_bubble, in units of times / second. The laminar flow critical threshold is set to 0 times / second. f_bubble>0 indicates that gas is being entrapped into the center of the molten iron flow within the sprue or pouring cup.

[0084] 3. Definition of Data Fusion and Process Criteria The G_front extracted from infrared thermal imaging at the same time point is fused with the I_turb and f_bubble extracted from ultrasonic signals, aligning them along the time axis. The system presets the following process criterion thresholds, see Table 6: ; The logical relationships and decision priorities among the criteria are as follows: First priority (thermodynamic safety): G_front < 5℃ / cm. When this condition is triggered, regardless of the flow field state, measures to increase the pressure head height will be implemented first to accelerate the temperature rise.

[0085] Second priority (kinetic cleanliness): G_front ≥ 5℃ / cm and (I_turb ≥ 0.3 or f_bubble > 0). When this condition is triggered, it indicates that there is sufficient heat but the flow state is poor, and turbulence needs to be suppressed.

[0086] Third priority (equipment protection): V_front > 2.5 m / s. Even if the above indicators are all normal, if the flow rate exceeds the erosion resistance limit of the molding sand, the flow rate must be limited to prevent sand erosion defects.

[0087] 4. Dynamic Adjustment of Casting Process Parameters: Based on the output of the fusion criterion, the PLC controller sends specific adjustment commands to the servo actuator of the casting machine. The adjustment objects are the pouring head height (H) and the pouring cup tilt angle (θ), and the adjustment logic is shown in Table 7 below: ; In this embodiment, step S4: Phased array ultrasonic homogenization cooling control during solidification: Phased array ultrasonic testing results of several consecutive castings during solidification are acquired to adjust the initial sand mold temperature or holding time based on the cooling difference between sand molds and the same-mold cooling difference. During the solidification stage after casting is completed, phased array ultrasonic technology is used to monitor the solidification front advancement process of thick, hot sections of the casting in real time. The consistency of cooling rate between different castings within the same sand mold (same-mold cooling difference) and between different sand molds in continuous production (inter-mold cooling difference) is quantitatively evaluated. The initial sand mold temperature or holding time is automatically adjusted according to the type and magnitude of the difference, thereby homogenizing the cooling conditions under batch production conditions, suppressing the generation of shrinkage porosity, shrinkage cavities, and thermal stress cracks, and ensuring the batch stability of the internal quality and mechanical properties of the castings.

[0088] Specifically, during the solidification process of castings, the liquid metal gradually transforms into a solid state, accompanied by a significant change in acoustic impedance. When ultrasound propagates in the solid phase, the backwave is clear and stable, while in the liquid phase or the mushy region of the solid-liquid two-phase system, the height of the backwave is significantly reduced or disappears due to strong scattering and absorption.

[0089] Monitoring principle: By continuously transmitting ultrasonic pulses and receiving the reflected echoes from the bottom surface along the wall thickness direction of the casting, the variation pattern of the bottom wave height can be analyzed to indirectly infer the migration process of the solidification front (solid line). The specific steps are as follows: A-scan waveform acquisition: The phased array flaw detector focuses the sound beam on the designated hot spot at a repetition frequency of 10 to 50 times per second and records the complete A-scan waveform (time-amplitude data).

[0090] Bottom wave identification and height calculation: In the A-scan waveform, the bottom reflected echo wave packet is identified according to the current wall thickness sound path range, and the maximum amplitude of the wave packet is calculated and recorded as the bottom wave height (H_bottom).

[0091] Solidity correlation: The correspondence between the bottom wave height and the solidity (or fully solidified state) of the casting at that location is established through preliminary calibration experiments. When the bottom wave height first reaches a certain proportion of the reference height in the fully solidified state (80% in this embodiment), it is determined that the solidification front of the hot spot has reached the bottom surface, that is, the cross section has achieved through solidification.

[0092] Calculation of solidification front advance velocity (V_solid): Record the time interval Δt_solid from the end of casting to the moment when the bottom wave height reaches the threshold. Given that the equivalent wall thickness (or thickness corresponding to the thermal modulus) of this hot spot is d, the average solidification front advance velocity is: V_solid=d / Δt_solid; This velocity value reflects the cooling intensity at the hot spot. A larger V_solid indicates faster cooling; a smaller V_solid indicates slower cooling and a higher risk of shrinkage.

[0093] The system compares and analyzes monitoring data from the same casting batch and across batches, calculating two types of cooling difference indicators: (1) Cooling difference of the same type (ΔV_intra) Applicable to multi-cavity sand boxes. Calculate the difference between the maximum and minimum solidification front advance velocities at the same hot spot location for different castings within the same sand box: ΔV_intra=V_solid_max-V_solid_min; In the formula, V_solid_max is the maximum value of the solidification front movement velocity of all monitored casting hot spots in the same sand box. The V_solid values ​​measured by all phased array probes in the sand box are sorted and the maximum value is taken. The unit is mm / s. V_solid_min is the minimum value of the solidification front movement velocity of all monitored casting hot spots in the same sand box. The V_solid values ​​measured by all phased array probes in the sand box are sorted and the minimum value is taken. The unit is mm / s.

[0094] This indicator reflects the degree of uneven cooling within the same sand box due to differences in gating layout, sand mold wall thickness, or uneven local heat storage.

[0095] (2) Cooling difference between sand molds (ΔV_inter) Applicable to continuous production scenarios. Record the V_solid values ​​at the same hot spot at the corresponding position in the Nth and N+1th boxes of continuous production, and calculate the absolute value of the difference between the two values: ΔV_inter=|V_solid(N+1)-V_solid(N)|; In the formula, V_solid(N) is the solidification front movement velocity of the hot spot at a specific location in the Nth production sand box, in mm / s; V_solid(N+1) is the solidification front movement speed of the hot spot at the same position in the N+1th continuous production sand box, ensuring that the monitoring position corresponds completely with the Nth box (such as the same cavity and the same hot spot), with the unit being mm / s.

[0096] This indicator reflects the consistency of cooling conditions between batches due to fluctuations in sand box turnover temperature, changes in initial sand mold temperature, or drift in environmental cooling conditions.

[0097] Based on process test data regarding casting material (e.g., ductile iron, gray cast iron) and structural complexity, allowable thresholds for two types of discrepancies are preset, as shown in the table below: ; The V_solid baseline value is obtained by calculating the average value after removing the maximum and minimum values ​​from the simultaneously collected data.

[0098] Scenario 1: When the cooling difference of the same type is greater than the first speed deviation threshold, it is determined that the sand mold has uneven local cooling, and a local chill adjustment strategy is implemented: Identify the location of the cavity containing the casting with the lowest V_solid value (slowest cooling); In the next molding cycle, graphite or metal chills are added to the inner wall of the sand mold near the hot spot area corresponding to the cavity. When the deviation reaches 0.5 mm / s, the chills are added with a thickness of 5 mm. When the deviation reaches 2.0 mm / s, the chills are thickened to the upper limit of 15 mm. If the deviation is between 0.5 mm / s and 2.0 mm / s, the thickness of the chills = 5 + 10 × (deviation - 0.5) ÷ 1.5, in mm. Alternatively, a cooling channel is added inside the sand core corresponding to this position, and compressed air is introduced after casting to accelerate cooling.

[0099] Scenario 2: When the cooling difference between sand molds exceeds the second speed deviation threshold, it is determined to be a fluctuation in the overall thermal state of the sand mold, and a sand mold temperature or heat preservation cycle adjustment strategy is implemented: If the V_solid of the (N+1)th mold is greater than that of the Nth mold (cooling is too fast), it indicates that the initial temperature of the sand mold is too low. The preheating temperature of the next sand mold should be increased by 10℃ to 30℃ using an online temperature control system to slow down the cooling rate.

[0100] If the V_solid of the (N+1)th mold is less than that of the Nth mold (cooling is too slow), it indicates that the initial temperature of the sand mold is too high or the heat preservation is insufficient. The preheating temperature of the next sand mold can be reduced by 5℃ to 15℃ using an online temperature control system, or the transfer rhythm of the mold on the heat preservation box / line after pouring can be adjusted to extend the heat preservation time by 5 to 15 minutes, thus delaying the opening time and utilizing residual heat to achieve uniform graphitization or stress release.

[0101] In this embodiment, step S5: Dual-mode flaw detection verification and parameter threshold locking after sand removal: Obtain the surface magnetic particle inspection and internal ultrasonic flaw detection results of the casting after sand removal to determine the effectiveness of the process correction parameters and output the corresponding process parameter thresholds. After the casting has completed sand removal and cleaning, two non-destructive testing methods, surface magnetic particle inspection and internal ultrasonic flaw detection, are used to conduct comprehensive defect detection on the casting. The detection results are spatially correlated with the previous process monitoring data (infrared thermal imaging temperature distribution data in step S3) to confirm the effectiveness of the pouring flow field control and smelting purity control measures. Based on the actual operating parameters of qualified batches, the lower limit of residual magnesium content and the upper and lower limits of pouring temperature are scientifically locked to form standardized process control thresholds.

[0102] Surface magnetic particle inspection and crack feature extraction: After sand removal and shot blasting to remove surface oxide scale and adhering sand, the castings enter the automated magnetic particle inspection station. Fluorescent wet continuous magnetic particle inspection is used, and allowable crack limits are set according to product technical requirements. For cold shut cracks and hot cracks, which are the focus of this embodiment, the judgment criterion is: if the crack length (L, in mm) is greater than the maximum allowable crack length (L_allow, in mm), which is 3 mm in this embodiment, it is marked as an excessive crack; the number of excessive cracks and the total length ΣL_exceed for each casting are counted.

[0103] Internal ultrasonic testing and defect quantification: An A-type pulse-echo ultrasonic flaw detector, equipped with a single-crystal straight probe, was used. The testing area covered the critical stress areas and corresponding geometric thermal points specified in the casting drawings. The equivalent flat-bottom hole method was used to quantitatively evaluate the excessive cracks in the internal defects, and the parameters in Table 9 below were recorded.

[0104] ; Defect distribution statistics: Summarizing and statistically analyzing the flaw detection results of multiple castings from the same batch. Average defect equivalent size: Φ_avg=(Φ_eq1+Φ_eq2+…+Φ_eqk) / k; k is the total amount of data for the equivalent flat-bottom hole diameter, Φ_eqi is the equivalent flat-bottom hole diameter for each defect, i=1,2,……,k.

[0105] Defect number density: ρ_defect=k / V, where V is the inspected volume of a single casting (dm³), and ρ_defect is in units of number / dm³.

[0106] The effectiveness of the process correction parameters was verified in step S3, where adjustments were made to the pouring head height or pouring cup tilt angle to address insufficient temperature gradient or excessive turbulence intensity at the filling front. The effectiveness of these adjustments should be reflected in a significant reduction in cold shut cracks and subcutaneous porosity on the casting surface. The crack location coordinates (X, Y, Z) identified by magnetic particle inspection in step S5 were matched with the point cloud of the casting's 3D CAD model using camera extrinsic parameter calibration to determine the precise 3D coordinates of the crack on the casting surface. These coordinates were then spatially matched with the temperature distribution data recorded by infrared thermal imaging during the filling stage in step S3. The crack coordinates were projected onto the corresponding location on the surface of the casting's 3D model; it was then determined whether this location belonged to a low-temperature zone (temperature more than 50°C below the liquidus) or a sluggish flow zone (local flow velocity less than 50% of the average flow velocity) recorded by the infrared thermal imager during the filling stage.

[0107] Validity judgment criteria: If the overlap between the crack location and the low temperature / stagnation zone during the filling stage (R_overlap) is less than 20%, and the average total crack length ΣL_exceed of this batch of castings is reduced by ≥50% compared with the batch before adjustment, then the pouring flow field control parameters are considered valid.

[0108] Understandably, R_overlap (crack-low temperature zone spatial overlap ratio) is a dimensionless ratio parameter that measures the spatial correspondence between the location of a crack on the casting surface and the low temperature / stagnation zone during the pouring and filling stage. On the three-dimensional surface of the casting, it represents the percentage of the area where the crack is located and the area of ​​the crack recorded by the infrared thermal imager that lies between the low temperature / stagnation zone during the filling stage, relative to the total area of ​​the crack distribution.

[0109] The R_overlap value ranges from 0% to 100%. A larger value indicates a stronger correlation between the crack location and the temperature anomaly region during the filling stage, meaning the crack is more likely to be caused by problems in the filling flow field (such as cold shuts or local stagnation). A smaller value indicates that the crack's appearance is not directly causally related to the control of the filling flow field and may be caused by other factors (such as shrinkage stress or insufficient sand mold collapsibility).

[0110] The calculation method for R_overlap is as follows: 1. Data Preparation (1) Crack distribution data (from step S5 magnetic particle inspection) The location of each crack is represented by a set of three-dimensional coordinate points on the surface of the casting. For longer cracks, they are discretized along their length into a sequence of sampling points spaced no more than 2 mm apart. The sampling points of all cracks are merged to form a crack point set P_crack={(X1,Y1,Z1),(X2,Y2,Z2),…(Xn,Yn,Zn)}.

[0111] (2) Low temperature / sluggish region data (from step S3 infrared thermal imaging) In step S3, the infrared thermal imager records the temperature distribution for each frame during the filling stage. The temperature distribution map at the end of filling (when the cavity is completely filled) is texture-mapped to the 3D CAD model of the casting, marking pixel areas with temperatures below critical values ​​(e.g., 50°C below the liquidus temperature) or stagnant areas with flow rates below 50% of the average flow rate. These areas are projected onto the 3D surface of the casting, forming a set of low-temperature / stagnant region patches, A_cold. This region has clearly defined spatial coordinate boundaries on the casting surface.

[0112] If the overlap is still high, it indicates that the adjustment range is insufficient or the direction is wrong. The adjustment strategy of step S3 needs to be corrected in the next round of production, and the corrected parameters should be recorded as new process parameters.

[0113] 2. Criteria for determining overlap For each sampling point P_i in the crack point set, determine whether it is located within the low-temperature / hysteresis region A_cold. The determination method is as follows: Calculate the shortest distance from point P_i to the boundary of region A_cold. If this distance is less than or equal to the tolerance threshold ε (in this embodiment, ε = 2 mm, considering the spatial resolution and projection error of the infrared thermal imager), then the crack point is determined to be located in the low temperature / hysteresis region.

[0114] 3. R_overlap calculation formula The number of crack sampling points located in the low-temperature / hysteresis region is counted and denoted as N_overlap. The total number of crack sampling points is N_total. Then: R_overlap=(N_overlap / N_total)×100%; If the cracks are represented by length data instead of discrete point sets, a length-weighted calculation method can also be used: For each crack, calculate the length L_i_overlap of its portion within A_cold; calculate the total length L_total of all cracks; then: R_overlap=(ΣL_i_overlap / L_total)×100%.

[0115] In this embodiment, setting R_overlap to be less than 20% is not a fixed constant, but a reference threshold determined based on process validation experience. The setting logic is as follows: If the pouring flow field control measures are completely ineffective and the low temperature zone of the mold filling is not improved, cold shut cracks are very likely to occur in this area. In this case, R_overlap is usually high (>50%). If the control measures are effective, the low-temperature zone will shrink or disappear. At this time, the cracks generated are mostly caused by shrinkage stress and other factors, and their location has a low degree of overlap with the original low-temperature zone (usually <20%). Understandably, 20% is used as an empirical threshold, allowing for a certain degree of accidental coincidence (such as measurement errors or coincidences in local structures), but it is sufficient to distinguish whether flow field issues are still the dominant factor. In actual production, this threshold can be adjusted according to the specific casting structure and quality requirements.

[0116] In this embodiment, the effectiveness of the smelting purity control parameters is verified: In step S2, to address the issue of excessive inclusion density in the molten iron, the inert gas bottom blowing time was extended or the inoculant addition ratio was adjusted. The effectiveness of the adjustment should be reflected in a significant reduction in the size and number of inclusion defects inside the casting. The defect equivalent size Φ_eq and number density ρ_defect obtained from ultrasonic testing in step S5 are compared with the preset target defect rate.

[0117] The target defect rate is preset based on the casting quality grade, and in this embodiment, it is set as follows: Grade 1 castings: Φ_eq≤1.5mm, and ρ_defect≤0.5 pieces / dm³; Secondary castings: Φ_eq≤2.5mm, and ρ_defect≤1.0 pieces / dm³.

[0118] Validity judgment criteria: If Φ_avg and ρ_defect of this batch of castings both meet the target grade requirements, then the smelting purity control parameters are deemed valid.

[0119] If large inclusion defects are still detected, the bottom blowing time needs to be extended or the inoculation process parameters need to be optimized, or the purity threshold needs to be adjusted (the standard needs to be raised), and the current actual parameters should be recorded as a boundary reference.

[0120] When the surface magnetic particle inspection and internal ultrasonic inspection results of several consecutive batches (≥5 batches in this embodiment) of castings consistently meet the quality requirements, the threshold values ​​of each process parameter under the production conditions corresponding to the batches that continuously meet the quality requirements are locked and output as standardized process specifications.

[0121] The output process parameter thresholds also include: the lower limit of the mass percentage of residual magnesium in molten iron, the lower limit of the casting temperature, and the upper limit of the casting temperature.

[0122] The lower limit of residual magnesium content (Mg_min) is determined by analyzing the residual magnesium content (mass percentage) using a direct-reading spectrometer before pouring the molten iron, and denoted as Mg_i. The relationship between the Mg_i value of each batch and the corresponding ultrasonic flaw detection pass rate and metallographic spheroidization rate is statistically analyzed. The minimum measured Mg_i value of all qualified batches is taken as Mg_qualified_min. Considering the spectral analysis error (±0.002%) and furnace fluctuations, a safety margin ΔMg is added (0.002% to 0.005% in this embodiment), then: Mg_min=Mg_qualified_min+ΔMg; The lower limit of the pouring temperature (T_pour_min) is determined by the reading of the immersion thermocouple or infrared thermometer at the start of pouring in step S3, denoted as T_pour_i. The T_pour_i values ​​for each batch are statistically analyzed in relation to the cold shut crack detection rate in magnetic particle testing step S5. The minimum measured T_pour_i value among all batches without cold shut cracks is taken and denoted as T_no_cold_min. Considering the temperature drop loss during pouring (typically 2℃~5℃ / min) and temperature measurement error, a safety margin ΔT_cold (5℃~10℃ in this embodiment) is added, then: T_pour_min=T_no_cold_min+ΔT_cold; The upper limit of the pouring temperature (T_pour_max) is determined by the reading of the immersion thermocouple or infrared thermometer at the start of pouring in step S3, denoted as T_pour_i. The relationship between the T_pour_i value of each batch and the rate of exceeding the ultrasonic shrinkage defect limit in step S5 is statistically analyzed, while also considering the refractoriness limit of the molding sand. The maximum measured T_pour_i value is taken from all batches where internal shrinkage defects do not exceed the limit and there is no sand adhering to the surface, and denoted as T_no_shrink_max. A safety upper limit T_sand_limit is set according to the type of molding sand (usually ≤1420℃ for wet molding sand, and slightly higher for resin sand). Therefore: T_pour_max=min(T_no_shrink_max,T_sand_limit-ΔT_sand); Where ΔT_sand is the safety margin for preventing sand from sticking (in this embodiment, it is taken as 5℃~10℃).

[0123] The determined Mg_min, T_pour_min, and T_pour_max values ​​are written into the casting process specification for this product and input into the process parameter control module of the MES system. During subsequent mass production, the system automatically monitors whether the actual parameters fall within the threshold range. If they exceed the threshold, an alarm is triggered and the anomaly is recorded, forming a continuous and stable quality control closed loop.

[0124] Please see Figure 4 This is a schematic diagram of the process for performance grading of castings and raw material threshold feedback in this invention.

[0125] In this embodiment, step S6: multi-dimensional composite flaw detection and performance grading after heat treatment and raw material threshold feedback: heat treatment is performed on castings that pass the initial flaw detection, and the performance grading data of the castings is determined based on the multi-dimensional composite flaw detection results of the full size of the castings. Based on the performance grading data of the castings, castings that meet the requirements are determined, and the raw material grading threshold is adjusted based on the grading percentage.

[0126] This step involves two aspects: First, the castings that have passed the aforementioned processes and initial flaw detection undergo final heat treatment. During and after heat treatment, acoustic emission monitoring and multi-dimensional ultrasonic composite flaw detection technology are used to finely evaluate the uniformity of the casting's matrix structure and its internal density, outputting casting performance grading data to determine the casting's delivery grade or downgrade usage plan. Second, based on this grading result, the percentage of products meeting the target defect rate (i.e., the grading percentage) is calculated. This percentage is compared with a preset control threshold. When the pass rate does not meet the production target, the raw material grading threshold in step S1 is adjusted in reverse, further tightening the raw material grading thresholds (defect volume percentage thresholds T1 and T2, and graphitization segregation degree threshold). Preferably, each adjustment tightens the original threshold by 10% to 20%, forming a closed-loop optimization mechanism from finished product quality feedback to raw material selection standards.

[0127] Specifically, the principle of online acoustic emission monitoring during heat treatment is that during the heating and holding stages of heat treatment, the casting undergoes phase transformations in the matrix (such as the transformation from pearlite to austenite) and the accumulation and release of thermal stress. If micro-discontinuous regions exist inside (such as microcracks, inclusion tips, and shrinkage cavities), localized plastic deformation or microcrack propagation will occur due to stress concentration. This process is accompanied by the release of high-frequency elastic stress waves, i.e., the acoustic emission phenomenon. By monitoring the frequency and energy of acoustic emission events, it is possible to determine whether harmful micro-damage propagation has occurred during the heat treatment process without damaging the casting.

[0128] Sensor Placement: High-temperature acoustic emission sensors are mounted on the surfaces of critical stress points or geometrically hot spots of the casting using high-temperature coupling agents or mechanical clamps. The sensor frequency response range is 100kHz to 1MHz, and the temperature resistance is not lower than the maximum heat treatment holding temperature (usually ≥600℃). For large castings, a multi-channel sensor array can be used to improve positioning accuracy. Signal Acquisition and Processing: The acoustic emission sensor signals are amplified by a preamplifier and then transmitted to a multi-channel acoustic emission acquisition card. The acquisition card continuously records the acoustic emission event waveforms and extracts characteristic parameters in real time; please refer to Table 10.

[0129] ; Microcrack propagation assessment, under normal conditions: During the initial heating and holding stages of heat treatment, a certain number of low-energy acoustic emission signals will appear due to thermal expansion and the release of phase transformation stress. As the temperature homogenizes, N_AE and E_AE should gradually decrease and tend to stabilize.

[0130] Anomaly Criteria: If any of the following situations occur during the mid-to-late stage of heat preservation or the cooling stage, it is determined that there is a risk of microcrack propagation: N_AE shows a continuous upward trend and is >50 times / minute; E_AE shows a sudden spike, with the energy value per minute exceeding 3 times the average value of the stable stage.

[0131] Handling method: Castings that trigger abnormal criteria are automatically marked as suspected acoustic emission components and will be subject to key review or direct downgrading in subsequent classification.

[0132] After heat treatment, the casting is subjected to full-size, multi-dimensional composite flaw detection. After the heat treatment is completed and cooled to room temperature, the casting is subjected to full-size, multi-dimensional ultrasonic composite flaw detection to comprehensively evaluate its internal quality.

[0133] (1) Ultrasonic phased array sector scan detection Inspection objective: To obtain a two-dimensional cross-sectional defect distribution image of the casting across its entire wall thickness, and to visually identify the location and extent of macroscopic defects such as shrinkage porosity, gas porosity, and slag inclusions.

[0134] Inspection method: An ultrasonic phased array flaw detector, equipped with linear or area array probes, is used to perform a full-coverage scan along a preset scanning trajectory on the surface of the casting. Phased array electronic scanning and dynamic focusing functions are used to generate sector scan images of each scan section.

[0135] Output parameters: Record the planar projected area of ​​each defect within the detection area and the equivalent size of the largest defect.

[0136] (2) Time-of-flight diffraction (TOFD) method for detection Detection objective: To accurately measure the height of suspected defects detected by phased array sector scanning in order to quantitatively evaluate the extent of defect extension in the wall thickness direction.

[0137] Detection method: A pair of TOFD probes (one transmitting and one receiving) are symmetrically arranged on both sides of the defect location. The height of the defect is calculated by utilizing the arrival time difference of the diffracted waves generated at the upper and lower tips of the defect. Please refer to Table 11.

[0138] ; (3) Matrix sound velocity attenuation mapping Purpose of testing: To evaluate the uniformity of the matrix structure of castings. Graphite morphology, pearlite / ferrite ratio, and microporousness all affect the propagation speed and attenuation coefficient of ultrasonic waves in the matrix.

[0139] Detection method: Using an ultrasonic phased array probe or a single-crystal straight probe, measure the bottom echo time and echo amplitude at multiple measuring points evenly distributed on the casting surface. Given the actual wall thickness d at each measuring point (obtained from a CAD model or calipers), calculate the actual sound velocity v_i at each measuring point: v_i = 2d / t_i; Where t_i is the time (μs) from the transmission of the ultrasonic pulse to the reception of the bottom wave at the measuring point.

[0140] Matrix homogeneity evaluation parameters: Statistical deviation of sound velocity σ_v at all measuring points: ; Where m is the total number of measuring points, and v_avg is the average sound velocity at all measuring points. The smaller σ_v is, the smaller the fluctuation in sound velocity in the matrix and the more uniform the structure. √ is the square root symbol.

[0141] Casting performance grading data output: Based on the comprehensive acoustic emission monitoring results, TOFD defect height measurement results, and matrix sound velocity attenuation mapping results, the performance grading data of the casting is output according to the preset grading standard. Please refer to Table 12.

[0142] ; In this embodiment, the recommendations for casting delivery judgment and downgraded use are as follows: Delivery conditions met: When the matrix uniformity grade of the casting reaches Grade A and the internal density grade reaches Grade 1 or 2, the casting is deemed to meet the delivery requirements for precision load-bearing structural components and is approved for release.

[0143] Downgrade usage recommendation: When the internal density level is lower than level 2 but higher than level 4 (i.e., level 3), the system will automatically generate a downgrade usage recommendation for the casting, such as for non-critical support components, counterweights, or general structural components.

[0144] Scrap determination: If the matrix uniformity grade is C or the internal density grade is 4, the product is deemed unqualified and scrapping and remelting are recommended.

[0145] Adjusting the raw material classification threshold based on the hierarchical percentage, including: 1. Determination of the hierarchical percentage After continuously producing a certain number of batches (in this embodiment, 20 batches or 200 castings are taken), count the above performance classification results and calculate the hierarchical percentage P_pass: P_pass=(N_pass / N_total)×100%; Among them, N_pass is the number of qualified castings that meet the target defect rate, unit: piece, and it is the total number of castings whose performance classification results meet Grade A of matrix structure uniformity and internal density levels 1 or 2 during the statistical period; N_total is the total number of castings produced during the statistical period, unit: piece, and it is the total number of castings fully inspected in step S6 during the same statistical period.

[0146] 2. Comparison with the percentage control threshold Compare the calculated P_pass with the preset percentage control threshold P_target. P_target is the target qualification rate set according to the production plan and customer order requirements. In this embodiment, it is set to 85%. For the comparison results and corresponding adjustment strategies, please refer to Table 13.

[0147] ; 3. Adjustment method for the raw material classification threshold When P_pass<P_target, the system judges that it is necessary to tighten the raw material screening criteria. The adjustment objects are the defect volume ratio thresholds T1, T2 in step S1 and the graphitization segregation degree threshold S_th. The adjustment rules are as follows: (1) For the adjustment of the defect volume ratio threshold, please refer to Table 14.

[0148] ; (2) Adjustment of the graphitization segregation degree threshold The adjustment direction of the graphitization segregation degree threshold S_th is to decrease, that is, the requirement for the raw material graphitization uniformity is more strict: S_th_new=S_th×(0.8~0.9) 4. Verification and固化 after adjustment Verification period: After the new threshold takes effect, continue to count the P_pass value in the next cycle.

[0149] If P_pass rises above P_target, it means the adjustment is effective, and the new threshold is固化 as the standard classification index under the current production conditions.

[0150] Iterative optimization: If P_pass is still lower than P_target after adjustment, the threshold is tightened according to the above rules, but a lower limit protection value needs to be set to prevent the raw material utilization rate from being too low due to an overly strict threshold. In this embodiment, T1≥0.2%, T2≥1.0%, and S_th≥0.02 are used.

[0151] Reverse adjustment: If P_pass remains higher than P_target+10% (e.g., >95%) for multiple consecutive statistical periods, it indicates that the current threshold may be too strict, resulting in the waste of high-quality raw materials. In this case, the threshold can be appropriately relaxed (in this embodiment, T1 and T2 are adjusted to 1.05 to 1.1 times the original value) to improve the utilization rate of raw materials.

[0152] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A casting method for defect control of cast iron parts based on multi-stage flaw detection, characterized in that, include: Water immersion ultrasonic testing was performed on several raw material blocks used for smelting cast iron parts to obtain full-section scanning data of the raw material blocks; The defect volume ratio and graphitization segregation degree of each raw material block are determined based on full-section scanning data to determine the grade of each raw material. Target raw material blocks are selected by matching the raw material block grading results with the target defect rate; Online ultrasonic flaw detection is performed on the molten iron formed from the target raw material block to determine the purity of the molten iron, so as to adjust the smelting process parameters and obtain the castable molten iron; After casting, images of filling flow rate and temperature change rate are generated based on high-speed infrared thermal imaging data of casting. Ultrasonic echo signals of flowing molten iron are acquired based on immersion ultrasonic acquisition, so as to adjust casting process parameters based on temperature gradient and flow field state. The phased array ultrasonic flaw detection results of solidification of several consecutive castings are obtained, so as to adjust the initial temperature of the sand mold or the holding time based on the cooling difference between sand molds and the cooling difference within the same mold. Obtain the surface magnetic particle inspection and internal ultrasonic inspection results of the casting after sand removal to determine the effectiveness of the process correction parameters and output the corresponding process parameter thresholds. The castings that pass the initial flaw detection are subjected to heat treatment. The performance grading data of the castings is determined based on the multi-dimensional composite flaw detection results of the full size of the castings. The castings that meet the requirements are determined based on the performance grading data, and the raw material grading threshold is adjusted based on the grading percentage.

2. The casting method for controlling defects in cast iron parts according to claim 1, characterized in that, The process of performing water immersion ultrasonic testing on several raw material blocks used for smelting cast iron parts to obtain full-section scanning data of the raw material blocks includes: Using a focused probe, spiral or layer-by-layer C-scan is performed on the raw material block to obtain ultrasonic A-wave signals at different depths within the raw material block; Fourier transform is performed on the A-wave signal to extract the bottom wave attenuation coefficient and the defect echo amplitude, and a three-dimensional full-section defect distribution model of the raw material block is reconstructed.

3. The casting method for controlling defects in cast iron parts according to claim 2, characterized in that, The determination of the defect volume ratio and graphitization segregation degree of each raw material block based on full-section scanning data, in order to determine the grade of each raw material, includes: The grade of each raw material is determined by comparing the defect volume ratio and graphitization segregation degree of each raw material block with the corresponding threshold.

4. The casting method for controlling defects in cast iron parts according to claim 1, characterized in that, The step of performing online ultrasonic flaw detection on the molten iron formed from the target raw material block to determine the purity of the molten iron, in order to adjust the smelting process parameters and obtain the castable molten iron, includes: A high-temperature resistant waveguide rod is installed in the ladle or tundish, and a high-frequency ultrasonic pulse is emitted into the flowing molten iron using an electromagnetic ultrasonic transducer. Real-time reception of backscattered signals generated by non-metallic inclusions in molten iron; calculation of inclusion number density and average equivalent diameter per unit volume. When the inclusion density exceeds the first preset purity threshold, the inert gas bottom blowing time during the smelting process is increased or the inoculant addition ratio is adjusted until the real-time monitored inclusion density drops below the second preset purity threshold, thus obtaining the cast iron.

5. The casting method for controlling defects in cast iron parts according to claim 1, characterized in that, The process involves generating images of filling flow rate and temperature change rate based on high-speed infrared thermal imaging data of the casting process, acquiring ultrasonic echo signals of flowing molten iron using immersion ultrasonic technology, and adjusting casting process parameters based on temperature gradient and flow field conditions, including: An array of immersion ultrasonic sensors is installed on the inner wall of the pouring cup or sprue to detect the turbulence intensity index and bubble entrainment frequency of the molten iron flow during the pouring process. The turbulence intensity index is fused with the temperature gradient at the filling front as reflected by infrared thermal imaging. If the temperature gradient at the filling front is less than the critical feeding temperature gradient, or the turbulence intensity index is greater than the laminar critical threshold, the pouring head height or the tilt angle of the pouring cup will be adjusted accordingly by the servo control mechanism.

6. The casting method for controlling defects in cast iron parts according to claim 1, characterized in that, The acquisition of phased array ultrasonic flaw detection results of solidified multiple castings, and the adjustment of the initial temperature or holding time of the sand mold based on the cooling difference between sand molds and the cooling difference within the same mold, includes: The solidification front advance rate of the thick, hot-spotted parts of the casting is monitored using phased array probes arranged on the side wall of the sand box. Calculate the cooling difference between different castings in the same sand box and the cooling difference between sand molds at corresponding positions between different sand boxes in continuous production; When the cooling difference of the same type is greater than the first speed deviation threshold, adjust the thickness of the local chill arrangement of the corresponding sand mold. When the cooling difference between sand molds exceeds the second speed deviation threshold, adjust the preheating temperature of the subsequent sand molds or adjust the transfer cycle of the insulation box after pouring.

7. The casting method for controlling defects in cast iron parts according to claim 1, characterized in that, The process involves obtaining the surface magnetic particle inspection and internal ultrasonic inspection results of the casting after sand removal to determine the effectiveness of the process correction parameters and outputting the corresponding process parameter thresholds, including: Automated magnetic particle testing is performed on castings after shot blasting to identify the length, direction, and location coordinates of surface and near-surface cracks. The magnetic particle inspection results were matched with the infrared thermal imaging temperature difference data during the casting stage to verify whether the adjustment parameters of the filling flow field eliminated the tendency of thermal cracking. A-type pulse-echo ultrasonic testing was performed on the castings to record the equivalent flat-bottom hole size and number distribution of defects. The measurement results were compared with the target defect rate to verify the effective boundary value of the smelting purity control parameters.

8. The casting method for controlling defects in cast iron parts according to claim 7, characterized in that, The output process parameter thresholds include the lower limit of the mass percentage of residual magnesium in molten iron, the lower limit of the casting temperature, and the upper limit of the casting temperature.

9. The casting method for controlling defects in cast iron parts according to claim 1, characterized in that, The process of heat-treating castings that pass the initial flaw detection, and determining casting performance grading data based on multi-dimensional composite flaw detection results across the entire size of the casting, includes: During the heating and holding stages of heat treatment, a high-temperature acoustic emission sensor is used to monitor the count rate and energy release rate of acoustic emission events generated by the release of thermal stress inside the casting in order to determine whether there is microcrack propagation. After heat treatment, the casting was subjected to full-size ultrasonic phased array sector scanning and TOFD diffraction time-of-flight method flaw detection to obtain the diffraction signal at the defect tip; Based on the phase relationship of the diffraction signal and the height position of the defect tip in the wall thickness direction, combined with the attenuation mapping data of the matrix sound velocity, the matrix structure uniformity level and internal density level of the casting are output.

10. The casting method for controlling defects in cast iron parts according to claim 9, characterized in that, Adjusting the raw material grading threshold based on the grading percentage includes: The number of products that meet the target defect rate is determined based on the grading results, in order to determine the grading percentage; The percentage classification is compared with the percentage control threshold, where, If the grading percentage is less than the percentage control threshold, it is determined to adjust the defect volume ratio threshold and the graphitization segregation degree threshold in the raw material grading threshold to update the grading index.

Images