Hardware rotary casting forming process parameter optimization method and system

Abstract

The invention belongs to the technical field of rotary casting forming, and discloses a hardware rotary casting forming technological parameter optimization method and system.The real-time solidification state of molten metal in a mold is deduced by obtaining multi-source real-time data and conducting mode analysis, and technological parameters are dynamically adjusted according to the deviation between the real-time solidification state and the preset state; the solidification state of molten metal in the mold can be sensed in real time, and the limitation that the internal state is difficult to directly monitor is overcome; process parameters can be dynamically adjusted according to the solidification state sensed in real time, and the problem that an existing method lacks the dynamic adjustment capability is solved; and therefore, the internal process deviation occurring in the production process can be corrected, and the stability of the casting quality and the production efficiency are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of spin casting forming, in particular to a hardware spin casting forming process parameter optimization method and system. BACKGROUND

[0002] Hardware spin casting forming process is an important metal forming technology. Its principle is to inject liquid metal into a high-speed rotating mold, and use strong centrifugal force to make the metal liquid spread uniformly on the inner wall of the mold and realize self-upward solidification forming. This process is widely used in the manufacture of tubular or ring-shaped parts with high requirements for dimensional accuracy and surface quality. In the spin casting process, the accurate control of a series of process parameters such as the rotation speed of the mold, the pouring speed of the metal liquid, and the cooling rate of the mold is crucial. There is a complex mutual coupling relationship between these parameters, which together determines the flow behavior of the metal liquid in the mold, the uniformity of the spread, the moving speed and shape of the solidification front, and the dimensional accuracy, surface integrity, and internal organization uniformity of the final casting, and directly affects the generation of casting defects such as pores, shrinkage, and cracks.

[0003] In actual industrial production environment, spin casting equipment usually needs to run continuously to produce hardware in batches. The standard production process usually sets fixed process parameters: after the metal is melted to a specified temperature, the liquid metal is injected into a high-speed rotating mold through a pre-set pouring system at a set speed; the mold rotates at a constant speed, and the centrifugal force generated drives the metal liquid to spread along the inner wall of the mold; the external cooling system usually cools the mold at a set flow rate and temperature to accelerate the solidification process of the metal liquid.

[0004] However, even if the same batch of melted metal, there may still be slight temperature inhomogeneity or flow difference during pouring. When these metal liquids with internal differences enter the high-speed rotating mold, the centrifugal force will amplify these differences, causing uneven initial spreading of the metal liquid on the inner wall of the mold. For example, the metal liquid with slightly lower local temperature may form an initial layer with uneven thickness on the inner wall of the mold due to slightly higher viscosity or faster solidification speed.

[0005] Once the initial spreading is uneven, the subsequent solidification process will be carried out on a non-ideal basis. If the process parameters such as mold rotation speed, pouring speed, and cooling rate remain at fixed preset values at this time, the system lacks dynamic adjustment capability, making it difficult to effectively correct this initial unevenness. The centrifugal force distribution is relatively fixed, which may not be able to redistribute the metal liquid to compensate for the thickness difference; the constant pouring speed may not match the uneven solidification front, causing metal liquid to accumulate or be insufficient in some areas; the fixed cooling rate also cannot adapt to the dynamically changing solidification front, which may cause the solidification speed difference in different areas to further widen, thereby affecting the final wall thickness uniformity, or forming stress concentration points at the solidification front, increasing the risk of cracks.

[0006] More challenging is that the internal solidification state in the spin casting process, such as the real-time flow field of the metal liquid, the precise position and shape of the solidification front, and the internal stress distribution, are usually difficult to measure directly and in real time in an industrial production environment.

[0007] Therefore, in the spin casting production process of hardware, due to the inherent small differences of the metal liquid itself, the complex nonlinear interaction between the coupled process parameters, and the limitation that the internal solidification state is difficult to directly monitor, even if the externally set process parameters remain constant, the actual internal solidification process may still change dynamically and appear uneven. The existing process control method often relies on pre-set fixed parameters or lag adjustment based on final product quality detection results, and lacks the ability to perceive, quickly respond and cooperatively correct the dynamically occurring internal process deviations in the production process, which greatly limits the stability of the casting quality and the improvement of the production efficiency.

[0008] In view of the above problems, the prior art needs to be improved. SUMMARY

[0009] The purpose of the present application is to provide a hardware spin casting forming process parameter optimization method and system, which can perceive the solidification state of the metal liquid inside the mold in real time, adjust the process parameters accordingly, help to correct the internal process deviations occurring in the production process, and improve the stability of the casting quality and the production efficiency.

[0010] In a first aspect, the present application provides a hardware spin casting forming process parameter optimization method for adjusting the process parameters in the hardware spin casting forming process, characterized in that the steps of the method comprise:

[0011] A1. Obtain temperature distribution data of a mold, power data of a driving motor, and vibration data of key parts of a spin casting equipment in a hardware spin casting forming process;

[0012] A2. Perform pattern analysis on the temperature distribution data, the power data, and the vibration data to obtain comprehensive pattern features reflecting the solidification state of the metal liquid inside the mold;

[0013] A3. According to the comprehensive pattern features, infer the real-time internal solidification state of the metal liquid inside the mold, and identify the deviation between the real-time internal solidification state and a preset state;

[0014] A4. According to the deviation between the real-time internal solidification state and the preset state, determine the adjustment amount of the mold rotation speed, the pouring speed, and the cooling rate;

[0015] A5. Adjust the mold rotation speed, the pouring speed, and the cooling rate according to the adjustment amount.

[0016] Preferably, step A1 comprises:

[0017] A101. Obtain temperature measurement values of the mold at different positions along the height direction, to form the temperature distribution data;

[0018] A102. Obtain real-time power consumption data of the driving motor driving the rotation of the mold as the power data;

[0019] A103. Obtain vibration measurement data of key parts of the spin casting equipment, to obtain the vibration data; the key parts of the spin casting equipment include at least one of the pouring gate, the bottom of the mold, and the driving motor.

[0020] Preferably, step A2 comprises:

[0021] A201. Perform first pattern analysis on the temperature distribution data to obtain temperature pattern features reflecting dynamic changes of heat transfer and solidification front of the mold along the height direction;

[0022] A202. Perform second pattern analysis on the power data to obtain power pattern features reflecting changes of metal liquid flow resistance, spreading uniformity, and stress during solidification;

[0023] A203. Perform third pattern analysis on the vibration data to obtain vibration pattern features reflecting vibrations related to metal liquid impact, uneven flow, or potential defects;

[0024] A204. Perform correlation analysis on the temperature pattern features, the power pattern features, and the vibration pattern features to obtain the comprehensive pattern features.

[0025] Preferably, the temperature pattern features include at least one of temperature gradient distribution along the height direction of the mold, temperature change rate at different height positions, position of the highest temperature point, and shape features of the temperature distribution curve;

[0026] The power pattern features include at least one of mean value, standard deviation, fluctuation frequency, and amplitude of specific frequency components of the power data;

[0027] The vibration pattern features include at least one of total energy, main frequency and its amplitude, peak value, and energy in a specific frequency range of the vibration data.

[0028] Preferably, step A204 comprises:

[0029] B1. Obtain real-time sequence data of the temperature pattern features, the power pattern features, and the vibration pattern features;

[0030] B2. performing time series analysis on each of the real-time sequence data, to obtain respective real-time trend information and fluctuation characteristic information;

[0031] B3. analyzing the interdependence among each of the real-time sequence data, to obtain real-time correlation information;

[0032] B4. analyzing the variation law of the trend information, the fluctuation characteristic information and the correlation information over time, to obtain comprehensive pattern features reflecting the evolution process of the solidification state of the metal liquid inside the mold.

[0033] Preferably, step B3 comprises:

[0034] B301. calculating the correlation index between each of the real-time sequence data;

[0035] B302. according to the lead-lag relationship between each of the real-time sequence data;

[0036] B303. constructing the correlation index and the lead-lag relationship into real-time correlation information between each pattern feature.

[0037] Preferably, step B4 comprises:

[0038] B401. obtaining historical sequence data of the trend information, the fluctuation characteristic information and the correlation information; the historical sequence data contains real-time trend information, fluctuation characteristic information and correlation information;

[0039] B402. based on the historical sequence data, analyzing the variation law of the trend information, the fluctuation characteristic information and the correlation information within the current time window; the variation law includes the rate of change, acceleration and periodicity;

[0040] B403. based on the historical sequence data and the variation law within the current time window, predicting the variation trend of the trend information, the fluctuation characteristic information and the correlation information within the future time window;

[0041] B404. synthesizing the historical sequence data, the variation law within the current time window and the variation trend within the future time window, to obtain comprehensive pattern features reflecting the evolution process of the solidification state of the metal liquid inside the mold.

[0042] Preferably, step A3 comprises:

[0043] A301. according to the comprehensive pattern features, inferring the real-time solidification front position, real-time solidification front shape parameters and real-time metal liquid spreading uniformity of the metal liquid inside the mold;

[0044] A302. Obtain a preset solidification front position, a preset solidification front shape parameter and a preset metal liquid spreading uniformity corresponding to the current time;

[0045] A303. Compare the real-time solidification front position with the preset solidification front position, the real-time solidification front shape parameter with the preset solidification front shape parameter, and the real-time metal liquid spreading uniformity with the preset metal liquid spreading uniformity, and identify the deviation between the real-time internal solidification state and the preset state.

[0046] Preferably, step A4 comprises:

[0047] A401. Extract a deviation index from the deviation between the real-time internal solidification state and the preset state; the deviation index comprises a solidification front position deviation index, a solidification front shape parameter deviation index and a metal liquid spreading uniformity deviation index;

[0048] A402. Determine a coordinated adjustment strategy of the mold rotating speed, the pouring speed and the cooling rate based on the extracted deviation index and the current process parameters;

[0049] A403. Calculate the adjustment amount of the mold rotating speed, the pouring speed and the cooling rate according to the determined coordinated adjustment strategy.

[0050] In a second aspect, the present application provides a hardware part spin casting process parameter optimization system for adjusting process parameters in a hardware part spin casting process, which comprises:

[0051] A data acquisition module for acquiring temperature distribution data of a mold, power data of a driving motor and vibration data of key parts of a spin casting device in a hardware part spin casting process;

[0052] A pattern analysis module for performing pattern analysis on the temperature distribution data, the power data and the vibration data to obtain comprehensive pattern features reflecting the internal solidification state of the metal liquid in the mold;

[0053] An inference and identification module for inferring the real-time internal solidification state of the metal liquid in the mold according to the comprehensive pattern features, and identifying the deviation between the real-time internal solidification state and a preset state;

[0054] An adjustment amount determination module for determining the adjustment amount of the mold rotating speed, the pouring speed and the cooling rate according to the deviation between the real-time internal solidification state and the preset state;

[0055] An adjustment execution module for adjusting the mold rotating speed, the pouring speed and the cooling rate according to the adjustment amount.

[0056] Beneficial effects: The hardware rotating casting forming process parameter optimization method and system provided by the application can infer the real-time solidification state of the metal liquid inside the mold through acquiring multi-source real-time data and performing mode analysis, and dynamically adjust the process parameters according to the deviation from the preset state, can realize real-time sensing of the solidification state of the metal liquid inside the mold, and overcome the limitation that the internal state is difficult to directly monitor; can dynamically adjust the process parameters according to the real-time sensed solidification state, and solve the problem that the existing method lacks dynamic adjustment capability; thereby helping to correct the internal process deviation occurring in the production process, and improving the stability of the casting quality and the production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0057] Figure 1 The flowchart of the hardware rotating casting forming process parameter optimization method provided by the embodiment of the application.

[0058] Figure 2 The structure schematic diagram of the hardware rotating casting forming process parameter optimization system provided by the embodiment of the application.

[0059] Label explanation: 1, data acquisition module; 2, mode analysis module; 3, inference identification module; 4, adjustment amount determination module; 5, adjustment execution module. DETAILED DESCRIPTION

[0060] The technical solutions in the application will be clearly and completely described below with reference to the drawings in the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. The components of the application described and shown in the drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the application provided in the drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the application. Based on the embodiments of the application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the application.

[0061] It should be noted that: similar labels and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of the application, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0062] Reference Figure 1 The application provides a hardware rotating casting forming process parameter optimization method for adjusting the process parameters of the hardware rotating casting forming process, characterized in that the steps of the method include:

[0063] A1. Obtain temperature distribution data of the mold, power data of the driving motor, and vibration data of key parts of the spin casting equipment during the spin casting process of the hardware;

[0064] A2. Perform pattern analysis on the temperature distribution data, the power data, and the vibration data to obtain comprehensive pattern characteristics reflecting the solidification state of the metal liquid inside the mold;

[0065] A3. According to the comprehensive pattern characteristics, infer the real-time internal solidification state of the metal liquid inside the mold, and identify the deviation between the real-time internal solidification state and the preset state;

[0066] A4. According to the deviation between the real-time internal solidification state and the preset state, determine the adjustment amount of the mold rotation speed, the pouring speed, and the cooling rate;

[0067] A5. Adjust the mold rotation speed, the pouring speed, and the cooling rate according to the adjustment amount.

[0068] Wherein, obtaining the temperature distribution data of the mold, the power data of the driving motor, and the vibration data of the key parts of the spin casting equipment during the spin casting process of the hardware refers to collecting the temperature information of different positions on the surface or inside the mold, the power consumption information of the motor driving the mold rotation, and the mechanical vibration information of the specific position of the spin casting equipment in real time during the spin casting process through sensors or other measuring devices, which can be realized by thermocouple array, infrared thermal imager, current and voltage sensor, accelerometer, etc. The main purpose is to obtain multi-source real-time monitoring information reflecting the physical state of the spin casting process to provide basic data for subsequent analysis.

[0069] Wherein, performing pattern analysis on the temperature distribution data, the power data, and the vibration data to obtain comprehensive pattern characteristics reflecting the solidification state of the metal liquid inside the mold refers to processing and analyzing the collected raw data, extracting the regularity information or characteristic values related to the metal liquid flow, spreading, and solidification process, such as the gradient change of temperature with time or space, the frequency component or fluctuation amplitude of the power signal, the energy distribution or dominant frequency of the vibration signal, etc., and further correlating and integrating these different source characteristic information to form a comprehensive index or model that can more comprehensively and accurately describe the real-time solidification state of the metal liquid inside the mold. It can be realized by signal processing technology, statistical analysis method, machine learning algorithm, etc. The main purpose is to extract high-level and meaningful information from the original monitoring data, overcome the complexity and noise interference of the original data, and provide basis for inferring the internal state.

[0070] Wherein, according to the comprehensive mode characteristics, the real-time internal solidification state of the metal liquid inside the mold is inferred, and the deviation between the real-time internal solidification state and the preset state is identified, which means that based on the comprehensive characteristics obtained by mode analysis, the actual solidification situation of the metal liquid inside the mold at the current time is predicted or estimated by using a pre-established model or rule, such as the position, shape of the solidification front, and the spread uniformity of the metal liquid, which are difficult to measure directly, and the inferred real-time state parameters are compared with the ideal, preset target state parameters, and the difference or deviation between the two is quantified, which can be realized by simulation calculation based on a physical model, a data-driven prediction model, an expert system, etc., and the main purpose is to obtain the internal solidification state information which is difficult to measure directly, and to quantify the gap between the current process and the ideal process, so as to provide a clear target and direction for subsequent process parameter adjustment.

[0071] The core innovation of the present application is that by analyzing and synthesizing the mold temperature distribution, driving motor power and equipment vibration and other multi-source external measurable data, the real-time solidification state of the metal liquid inside the mold which is difficult to monitor directly is inferred, and based on the deviation between the inferred real-time solidification state and the preset state, the key process parameters such as mold rotating speed, pouring speed and cooling rate are dynamically and cooperatively adjusted, thereby solving the problem that the dynamic changes and non-uniformity of the internal solidification process are difficult to perceive and correct in real time in the background technology, and achieving the effect of improving the stability of casting quality and production efficiency.

[0072] Specifically, the method continuously acquires mold temperature distribution, driving motor power and equipment vibration data during the spin casting process, which are used as real-time feedback signals of the process. Then, mode analysis is performed on these multi-source data to extract mode characteristics reflecting the characteristics of the process, such as temperature mode reflecting heat transfer and solidification front dynamics, power mode reflecting metal liquid flow and stress changes, and vibration mode reflecting impact and non-uniformity. Further, correlation analysis is performed on these mode characteristics to obtain a comprehensive mode characteristic, which integrates information of different physical quantities and more comprehensively describes the solidification state of the metal liquid inside the mold. Based on this comprehensive mode characteristic, the real-time solidification state of the metal liquid inside the mold is inferred, including the position, shape of the solidification front and the spread uniformity of the metal liquid. The inferred real-time solidification state is compared with the preset ideal solidification state to identify the deviation between the two. Finally, according to the identified deviation, the adjustment amount of the mold rotating speed, pouring speed and cooling rate, which are the three key process parameters, is determined, and the adjustment is performed. The whole process forms a closed-loop control circuit, which dynamically responds to the internal state changes and non-uniformity in the spin casting process by real-time monitoring, state inference, deviation identification and parameter adjustment, so that the actual solidification process tends to the preset ideal state.

[0073] Through the above scheme, the application can perceive the dynamic change and non-uniformity of the solidification state of the metal liquid inside the mold during the rotation casting process of the hardware, which is difficult to directly monitor in real time, and dynamically adjust key process parameters based on the dynamic change, thereby effectively correcting process deviation, improving the size precision, surface quality and internal organization uniformity of the casting, reducing the generation of casting defects, and improving the stability of the casting quality and production efficiency.

[0074] In some embodiments, step A1 comprises:

[0075] A101. Obtain temperature measurement values of the mold at different positions along the height direction to constitute the temperature distribution data;

[0076] A102. Obtain real-time power consumption data of a driving motor driving the mold rotation as the power data;

[0077] A103. Obtain vibration measurement data of key parts of the rotation casting equipment to obtain the vibration data; the key parts of the rotation casting equipment include at least one of the pouring gate, the bottom of the mold and the driving motor.

[0078] In step A101, the temperature measurement values of the mold at different positions along the height direction are obtained to constitute the temperature distribution data, because the metal liquid is solidified from bottom to top on the inner wall of the rotation casting mold, and the solidification front will move along the height direction. Obtaining the temperature of only one point of the mold is not enough to reflect the dynamic change of the entire solidification process. By obtaining the temperature measurement values at different positions along the height direction, the temperature distribution of the mold can be obtained, which directly reflects the heat transfer in the mold and the approximate position and shape of the solidification front. This distributed temperature data can more comprehensively capture the dynamic change of the solidification process than single-point temperature, providing key information for subsequent analysis of the dynamic change of the solidification front.

[0079] In step A102, the real-time power consumption data of the driving motor driving the mold rotation is obtained as the power data, and the power consumption of the driving motor is closely related to the flow state, spreading uniformity of the metal liquid in the mold and the resistance generated during the solidification process. For example, poor flow or uneven spreading of the metal liquid will cause changes in the load of the motor, and changes in stress during the solidification process will also be reflected in the power fluctuation. Obtaining real-time power consumption data can indirectly reflect the dynamic behavior of the metal liquid in the mold and the mechanical state during the solidification process, providing a basis for subsequent analysis of the flow resistance, spreading uniformity and stress change during the solidification process of the metal liquid.

[0080] In step A103, vibration measurement data of key parts of the spin casting equipment is obtained to obtain vibration data, wherein the key parts of the spin casting equipment include at least one of the pouring gate, the mold bottom and the driving motor. The vibration in the spin casting process is often related to the impact of the metal liquid, the formation of internal defects such as uneven flow, bubbles or inclusions. The vibration measurement is performed at the key parts such as the pouring gate, the mold bottom or the driving motor, and the vibration of these parts can more sensitively reflect the abnormal situation or internal state change in the spin casting process. For example, the stress release caused by the impact of the metal liquid on the mold, the uneven spreading or the internal solidification can cause specific vibration modes. Obtaining the vibration data of these key parts can serve as a signal for monitoring the impact of the metal liquid, the uneven flow or the potential defects, and provide supplementary information for subsequent analysis.

[0081] By combining the three types of data, i.e., the temperature distribution reflecting the thermal state, the power consumption reflecting the mechanical state and the vibration reflecting the dynamic process and abnormalities, the present scheme provides a richer and more targeted data basis than the prior art. These data depict the real-time state of the metal liquid inside the mold from different dimensions, lay a foundation for more accurate subsequent pattern analysis, inference of solidification state and identification of deviations, and thus improve the effectiveness of the process parameter optimization method.

[0082] In some embodiments, step A2 includes:

[0083] A201. performing first pattern analysis on the temperature distribution data to obtain temperature pattern features reflecting the dynamic changes of heat transfer and solidification front along the height direction of the mold;

[0084] A202. performing second pattern analysis on the power data to obtain power pattern features reflecting the flow resistance of the metal liquid, the uniformity of spreading and the stress changes in the solidification process;

[0085] A203. performing third pattern analysis on the vibration data to obtain vibration pattern features reflecting the vibrations related to the impact of the metal liquid, the uneven flow or potential defects;

[0086] A204. performing correlation analysis on the temperature pattern features, the power pattern features and the vibration pattern features to obtain the comprehensive pattern features.

[0087] Wherein, the pattern analysis refers to the process of extracting representative, regular or specific meaning feature information from the original data, which can be realized by signal processing techniques (such as Fourier transform, wavelet analysis), statistical analysis methods (such as mean, variance, correlation analysis), machine learning algorithms (such as clustering, classification, dimensionality reduction) or physical model-based analysis methods.

[0088] Among them, the first mode analysis refers to the mode analysis specially for temperature distribution data, aiming to extract mode features related to heat transfer and solidification front, which can be achieved by analyzing temperature gradient, temperature change rate, temperature distribution curve shape, etc.

[0089] Among them, the second mode analysis refers to the mode analysis specially for power data, aiming to extract mode features related to metal liquid flow, spreading and stress, which can be achieved by analyzing the average value, fluctuation amplitude, frequency component, etc.

[0090] Among them, the third mode analysis refers to the mode analysis specially for vibration data, aiming to extract mode features related to equipment vibration, metal liquid impact or internal defects, which can be achieved by analyzing the energy, main frequency, peak value, spectral characteristics, etc.

[0091] Among them, the correlation analysis refers to the process of analyzing the relationship between different types of mode features, aiming to find their dependence, time sequence relationship or collaborative change law, which can be achieved by cross-correlation analysis, Granger causality analysis, multivariate time series analysis or graph model-based analysis method.

[0092] Specifically, the scheme proposes a more detailed and comprehensive pattern analysis method to address the problem that the prior art fails to fully consider the complex correlation between different data types when analyzing multi-source monitoring data, resulting in an inability to accurately reflect the internal solidification state. The method first separately analyzes the temperature distribution data, power data, and vibration data, which are three key data types, to extract their respective pattern features reflecting specific physical processes. Step A201 extracts temperature pattern features reflecting the heat transfer path, rate, and dynamic changes in the position and shape of the solidification front within the mold by performing a first pattern analysis on the temperature distribution data. This is because temperature is a physical quantity that directly reflects the thermal state, and its distribution and changes are closely related to the cooling and solidification process of the metal liquid. Step A202 extracts power pattern features reflecting the flow resistance of the metal liquid in the mold, whether the spread is uniform, and the stress changes during the solidification process by performing a second pattern analysis on the power data. The power consumption of the driving motor is directly related to the motion state and stress of the metal liquid, so the power data can indirectly reflect the flow behavior and internal stress state of the metal liquid. Step A203 extracts vibration pattern features reflecting vibrations related to disturbances or potential defects (such as bubble rupture, solidification shrinkage, etc.) caused by the metal liquid impacting the mold and uneven flow by performing a third pattern analysis on the vibration data. Vibration is the response of the equipment during the dynamic process and can sensitively capture transient events and abnormal situations occurring inside. Then, the scheme further analyzes the correlation between these different types of pattern features. This is because various physical phenomena in the spin casting process are coupled, for example, uneven metal liquid spread (which may be reflected in power and vibration data) can affect heat transfer and the solidification front (which is reflected in temperature data), and stress changes during the solidification process (which are reflected in power data) can also cause vibrations. By analyzing the interdependence, timing correlation, etc. between these pattern features, a more in-depth understanding of the overall state and evolution of the internal solidification process can be obtained, resulting in a truly "comprehensive" and accurate pattern feature that accurately reflects the internal metal liquid solidification state of the mold.

[0093] This way of separately extracting pattern features from different types of data and then performing correlation analysis can more finely capture complex physical process information that cannot be revealed by a single data type, and by analyzing their interactions, it can more comprehensively and accurately depict the real-time solidification state of the metal liquid inside the mold and its evolution, laying a solid foundation for subsequent accurate inference of the real-time solidification state and process parameter adjustment.

[0094] In some optional embodiments, the temperature pattern features include at least one of the temperature gradient distribution along the height direction of the mold, the temperature change rate at different height positions, the position of the highest temperature point, and the shape features of the temperature distribution curve.

[0095] The power mode features include at least one of the average value, the standard deviation, the fluctuation frequency, and the amplitude of specific frequency components of the power data;

[0096] The vibration mode features include at least one of the total energy, the main frequency and its amplitude, the peak value, and the energy of a specific frequency range of the vibration data.

[0097] Among them, the temperature gradient distribution in the temperature mode features along the mold height direction reflects the change of temperature along the vertical direction of the mold, which is related to the direction and rate of heat transfer; the temperature change rate at different height positions reflects the change speed of temperature at a specific position with time, which is related to the local cooling speed and solidification process; the position of the highest temperature point indicates the position of the highest temperature along the height direction of the mold, which is related to the heat concentration area; the shape characteristics of the temperature distribution curve comprehensively reflect the overall thermal state of the mold and the morphology of the solidification front.

[0098] Among them, the average value of the power data in the power mode features reflects the average energy required to drive the mold to rotate, which is related to the overall flow resistance of the metal liquid; the standard deviation reflects the dispersion degree of the power data, which is related to the uniformity of the metal liquid spreading or the internal stress change; the fluctuation frequency reflects the periodic change of the power data, which is related to the dynamic behavior of the metal liquid; the amplitude of the specific frequency component reflects the intensity of the specific frequency in the power data, and the specific frequency component is a specific frequency component related to the known defect mode (which can be determined in advance through historical data analysis).

[0099] Among them, the total energy of the vibration data in the vibration mode features reflects the overall vibration level of the device, which is related to the impact or uneven flow of the metal liquid; the main frequency and its amplitude indicate the main vibration mode and intensity, which are related to physical phenomena; the peak value reflects the maximum instantaneous intensity of the vibration signal, which is related to the impact; the energy of a specific frequency range is used to identify the vibration signal related to defects, and the specific frequency range refers to a frequency interval associated with certain specific physical states, behaviors or potential problems in the hardware rotational casting process. By monitoring and analyzing the signal features in these specific frequency ranges, information about the solidification state of the metal liquid inside the mold can be indirectly obtained, such as judging whether the metal liquid spreading is smooth or identifying whether there is a vibration mode related to the formation of a specific defect. The specific frequency range can be determined in advance through historical data analysis.

[0100] Specifically, after obtaining the mold temperature distribution data, the motor power data, and the vibration data of the key parts of the spin casting equipment, pattern analysis is performed on these data. By extracting the specific temperature, power, and vibration pattern features, key information reflecting the internal thermal field of the mold, the metal liquid flow, and the dynamic response of the equipment can be obtained from the original monitoring data. For example, the temperature gradient distribution can finely depict the dynamic changes of the internal thermal field of the mold and the solidification front; the changes in power data can indirectly perceive the flow state of the metal liquid and the stress changes in the solidification process; the vibration signal can capture dynamic information related to the metal liquid flow, solidification, and defect formation. These specific pattern features are more representative and distinguishable than general pattern analysis. Correlation analysis of these specific temperature, power, and vibration pattern features can obtain comprehensive pattern features reflecting the solidification state of the metal liquid inside the mold. This comprehensive pattern feature based on multiple sources and specific features can accurately capture the solidification state of the metal liquid inside the mold, which is difficult to directly observe, such as the position, shape of the solidification front, and the uniformity of the metal liquid spreading. Accurate solidification state information is the basis for subsequent deviation inference between the solidification state and the preset state, and then the adjustment amount of the mold rotating speed, the pouring speed, and the cooling rate can be determined to realize dynamic adjustment of the process parameters of the spin casting process.

[0101] By providing these specific, physically meaningful pattern features, the present scheme makes it possible to accurately infer the internal solidification state from external measurable data, thereby overcoming the limitation that the internal state is difficult to monitor in the prior art, and providing key support for realizing process parameter optimization based on real-time state feedback.

[0102] As a preferred embodiment, the present scheme is implemented as follows: the extraction of temperature pattern features can be achieved by arranging multiple temperature sensors, such as thermocouples, along the height direction of the mold to measure temperature values at different positions, and obtaining temperature gradient distribution by calculating the ratio of temperature difference to distance between adjacent sensors; the temperature change rate is obtained by differentiating or fitting the readings of a single temperature sensor over time; the position of the highest temperature point is determined by comparing the readings of all temperature sensors; the shape parameters of the curve are extracted by curve fitting or using specific algorithms on the temperature readings along the height direction. The extraction of power pattern features can be achieved by connecting power sensors to the power supply lines of the driving motor to obtain real-time power data, and calculating the mean and standard deviation of the power data within a set time window; the fluctuation frequency and amplitude of specific frequency components are obtained by performing Fourier transform or other spectral analysis on the power data. The extraction of vibration pattern features can be achieved by installing acceleration sensors on key parts of the spin casting equipment, such as the bottom of the mold or the driving motor, to obtain vibration data; the root mean square value of the vibration signal is calculated as the total energy; the main frequency and its amplitude of the vibration signal are identified by spectral analysis; the peak value of the original waveform of the vibration signal is recorded; the energy of a specific frequency range is obtained by band-pass filtering the vibration data and then calculating the energy of the filtered signal.

[0103] Preferably, step A204 can include:

[0104] B1. Obtain real-time sequence data of the temperature pattern features, the power pattern features, and the vibration pattern features;

[0105] B2. Perform time series analysis on each of the real-time sequence data to obtain real-time trend information and fluctuation characteristic information respectively;

[0106] B3. Analyze the mutual dependence relationship between each of the real-time sequence data to obtain real-time correlation information;

[0107] B4. Analyze the variation law of the trend information, the fluctuation characteristic information, and the correlation information over time to obtain comprehensive pattern features reflecting the evolution process of the solidification state of the metal liquid inside the mold.

[0108] Wherein, the real-time sequence data refers to a collection of data of temperature pattern features, power pattern features, and vibration pattern features changing over time continuously collected or recorded during the spin casting process, which can be realized by time stamp data stream, fixed interval sampling sequence, or event triggered recording sequence. The end time of the real-time sequence data is the current time.

[0109] Among them, time series analysis refers to the analysis of data arranged in chronological order to identify patterns, trends, periodicity or randomness in the data, which can be achieved by using moving average, exponential smoothing, Fourier transform or wavelet analysis, etc.

[0110] Among them, trend information refers to the overall upward, downward or flat direction of time series data over a long time span, which can be extracted by using linear regression, polynomial fitting or non-parametric trend estimation method.

[0111] Among them, fluctuation characteristic information refers to the characteristics of the fluctuation of time series data around the trend line, including the amplitude, frequency, periodicity or randomness of the fluctuation, which can be quantified by using standard deviation, variance, spectral analysis or autocorrelation analysis, etc.

[0112] Among them, the interdependence relationship is the linear or nonlinear correlation strength, lag relationship or causality between different mode characteristic sequences, which can be quantified by using Pearson correlation coefficient, cross-correlation function or Granger causality, etc. to obtain real-time correlation information. For example, calculating the cross-correlation function between the real-time temperature gradient sequence and the real-time power standard deviation sequence can obtain the real-time correlation between them and the lag time of power fluctuation relative to temperature gradient change. Analyzing the Granger causality between the real-time power standard deviation sequence and the real-time vibration total energy sequence can determine whether power fluctuation has a predictive effect on vibration energy change. This correlation information reveals the coupling effect of different physical quantities in the solidification process.

[0113] Among them, the change rule over time refers to how the trend information, fluctuation characteristic information and correlation information themselves change over time as the spin casting process proceeds, such as whether the trend is accelerating, the fluctuation is enhancing or the correlation is weakening, which can be achieved by using secondary time series analysis, rate calculation or pattern recognition on the time series of these information.

[0114] Among them, the comprehensive mode characteristic is a feature set that integrates the dynamic changes (trend and fluctuation) of temperature, power, vibration mode characteristics and their dynamic correlations, and further analyzes the evolution rule of these dynamic information itself over time, to more comprehensively and dynamically reflect the solidification state of the metal liquid inside the mold and its evolution process.

[0115] Specifically, the present scheme lays a foundation for dynamic analysis by acquiring real-time sequence data of temperature, power, and vibration mode features. Then, time series analysis is performed on the real-time sequence data of each mode feature to extract their respective trends and fluctuation characteristics over time, thereby understanding the dynamics of the solidification process reflected by each physical quantity itself. On this basis, further analysis of the time-varying interdependence between these real-time sequence data captures the associated information generated by the dynamic coupling between different mode features. Finally, by analyzing the time-varying laws of these trend information, fluctuation characteristic information, and associated information themselves, such as their change rates, accelerations, or periodicities, a higher-level, more informative comprehensive mode feature is obtained. This comprehensive mode feature not only reflects the current snapshot of the solidification state but also contains historical trajectory and future trend information of its dynamic evolution. This multi-level, dynamic analysis method enables more accurate and real-time capture of the complex dynamic evolution process of the metal liquid solidification state inside the mold. The comprehensive mode feature obtained in this way can be more effectively used to infer the real-time internal solidification state of the metal liquid and identify deviations from the preset state, thereby supporting more accurate and faster process parameter adjustment.

[0116] Preferably, step B3 can include:

[0117] B301. Calculate the correlation index between each of the real-time sequence data;

[0118] B302. According to the lead-lag relationship between each of the real-time sequence data;

[0119] B303. The correlation index and the lead-lag relationship constitute the real-time association information between each mode feature.

[0120] Wherein, the correlation index refers to a statistical quantity used to quantify the linear association strength and direction between two real-time sequence data, which can be implemented by Pearson correlation coefficient, Spearman rank correlation coefficient or the value of cross-correlation function at zero lag.

[0121] Wherein, the lead-lag relationship refers to the time offset relationship existing between two real-time sequence data in the time dimension, i.e. the change of one sequence leading or lagging behind the change of another sequence, which can be implemented by cross-correlation analysis, Granger causality test or time series regression analysis.

[0122] The correlation index and the lead-lag relationship constitute real-time association information between each mode feature. The association information is obtained by combining the quantified correlation value and the determined lead-lag time, forming a comprehensive information describing the mutual influence between different mode features. For example, the correlation coefficient between temperature and power and the lead-lag time can be taken as an association information pair. In this way, the mutual dependence relationship between real-time sequence data is not only quantified, but also the dynamic time sequence characteristics are captured. This analysis method combining association strength and time sequence relationship can provide more comprehensive information than analyzing only a single dependence relationship, and help to more accurately reflect the complex dynamic interaction of the metal liquid solidification process inside the mold.

[0123] Specifically, the scheme improves the accuracy and dynamics of the association information acquisition by refining the analysis process of the mutual dependence relationship between real-time sequence data of each mode feature, thereby more accurately reflecting the real-time solidification state of the metal liquid inside the mold. First, the correlation index between each real-time sequence data is calculated, quantifying the strength and direction of the linear association between different mode features. This helps to identify which mode features have significant synchronous change trends. Then, the lead-lag relationship between each real-time sequence data is further considered. In the dynamic rotational casting process, the change of one physical quantity often leads to the subsequent change of another physical quantity, with a time delay or lead. Analyzing the lead-lag relationship can reveal deeper causal or influence chains between different mode features and understand their dynamic association in the time dimension. Finally, the correlation index and the determined lead-lag relationship are comprehensively utilized to obtain the real-time association information between each mode feature. This comprehensive analysis method not only considers the degree of synchronous change between mode features, but also considers their dynamic response relationship in time. By combining these two aspects of information, the mutual dependence relationship between different mode features can be more comprehensively and accurately characterized, thereby obtaining real-time association information that more accurately reflects the complex and dynamic solidification process of the metal liquid inside the mold. These more accurate association information is a key component of constructing comprehensive mode features, providing a basis for accurately inferring the real-time internal solidification state and identifying deviations from the preset state. In this way, when analyzing the mutual dependence relationship between mode features, the scheme considers their dynamics and time sequence characteristics, enabling the obtained association information to more accurately capture the subtle changes in the solidification state of the metal liquid during the rotational casting process, thereby improving the perception of internal process deviations.

[0124] Through the above scheme, the present application can more accurately capture the complex and dynamic interaction between different mode features, and the obtained association information can accurately reflect the real-time solidification state of the metal liquid inside the mold, thereby improving the accuracy and timeliness of subsequent inference of the solidification state and adjustment of process parameters, and enhancing the effective perception and correction ability of internal deviations in the rotational casting process.

[0125] Preferably, step B4 can comprise:

[0126] B401. obtaining historical sequence data of the trend information, the fluctuation characteristic information, and the correlation information; the historical sequence data contains real-time trend information, fluctuation characteristic information, and correlation information;

[0127] B402. based on the historical sequence data, analyzing the change law of the trend information, the fluctuation characteristic information, and the correlation information within a current time window; the change law includes a change rate, an acceleration, and a periodicity;

[0128] B403. based on the historical sequence data and the change law within the current time window, predicting the change trend of the trend information, the fluctuation characteristic information, and the correlation information within a future time window;

[0129] B404. synthesizing the historical sequence data, the change law within the current time window, and the change trend within the future time window to obtain a comprehensive mode feature reflecting the evolution process of the solidification state of the metal liquid inside the mold.

[0130] Wherein, the historical sequence data refers to a data set of the trend information, the fluctuation characteristic information, and the correlation information continuously collected or recorded within a period of time (which can be the period of time from the beginning of the rotation casting of the hardware to the current time, if the length of the period of time from the beginning of the rotation casting of the hardware to the current time exceeds the preset observation window length, the period of time can also be taken as the preset observation window length, which can be set according to actual needs), which provides a time dimension background for subsequent analysis and prediction.

[0131] Wherein, the change law within the current time window refers to the dynamic change characteristics of the trend information, the fluctuation characteristic information, and the correlation information within a specific length of time range (i.e. the length of the time window, which can be set according to actual needs), which can be obtained by calculating the first derivative (change rate), the second derivative (acceleration), or performing periodicity detection. The change rate refers to the speed of information change over time. The acceleration refers to the speed of change rate change over time. The periodicity refers to the pattern that information appears repeatedly over time. For example, a sliding time window technique can be used to define a fixed length current time window, and numerical differentiation method can be used to calculate the change rate and acceleration of the data within the window, such as difference method. Periodicity analysis can use fast Fourier transform (FFT) or autocorrelation function to detect whether there is a significant periodic component in the data.

[0132] The change trend in the future time window refers to the prediction of the possible development direction of the trend information, the fluctuation characteristic information, and the correlation information in a future period of time (i.e., the future time window length, which can be set according to actual needs) based on historical data and current rules. It can be realized by using a time series prediction model. For example, a machine learning model such as an autoregressive integrated moving average model (ARIMA), a long short-term memory network (LSTM), or a gated recurrent unit (GRU) can be used to predict the change trend of the trend information, the fluctuation characteristic information, and the correlation information in a future period of time based on historical sequence data and the change rule in the current time window.

[0133] The comprehensive mode feature refers to an abstract representation obtained by fusing historical sequence data, change rules in the current time window, and change trends in the future time window, which can more comprehensively depict the evolution process of the solidification state of the molten metal. For example, historical sequence data, change rule parameters in the current time window, and predicted trend data in the future time window can be used as inputs, and a comprehensive feature vector that can comprehensively represent the dynamic evolution process of the solidification state can be output by weighted averaging, feature splicing, or using another machine learning model (such as a support vector machine or a neural network). This comprehensive feature vector contains past, present, and future information, and is more descriptive and predictive than features based only on real-time information.

[0134] Specifically, the present scheme proposes a method of comprehensively and accurately obtaining the comprehensive mode feature by utilizing historical data, current change law and future prediction. Firstly, step B401 obtains historical sequence data of trend information, fluctuation characteristic information and correlation information, which provides information accumulation in time dimension for subsequent analysis and prediction, ensures the timeliness of analysis, and provides past information background, so that the understanding of the solidification state evolution process is more in-depth and comprehensive. Then, step B402 analyzes the change law of trend information, fluctuation characteristic information and correlation information in the current time window based on the historical sequence data, such as change rate, acceleration and periodicity, which enables to capture the dynamic characteristics of the current solidification process and understand whether it is accelerating, decelerating or showing certain periodic fluctuations, providing a refined description for understanding the current solidification state and its change trend. Further, step B403 predicts the change trend of trend information, fluctuation characteristic information and correlation information in the future time window based on the historical sequence data and the change law in the current time window. By combining the long-term background provided by the historical data and the short-term dynamics revealed by the current change law, the future solidification state evolution trend can be predicted. Finally, step B404 integrates the historical sequence data, the change law in the current time window and the change trend in the future time window to obtain the comprehensive mode feature reflecting the solidification state evolution process of the metal liquid inside the mold. This comprehensive consideration of past, present and future information enables the obtained comprehensive mode feature to not only reflect the current solidification state, but also contain its evolution history and future development trend.

[0135] Through the above scheme, the present application can overcome the deficiencies and lag that may be caused by real-time or short-term data analysis, and can more comprehensively and dynamically capture the evolution process of the metal liquid solidification state, providing a more reliable basis for subsequent solidification state inference and process parameter adjustment. This prediction capability enables the system to predict possible deviations in advance, thereby gaining time for subsequent process parameter adjustment, achieving more proactive and timely control, and improving the understanding and control ability of complex solidification process.

[0136] In some embodiments, step A3 comprises:

[0137] A301. According to the comprehensive mode feature, the real-time solidification front position, real-time solidification front shape parameter and real-time metal liquid spreading uniformity of the metal liquid inside the mold are inferred;

[0138] A302. Obtain a preset solidification front position, a preset solidification front shape parameter and a preset metal liquid spreading uniformity corresponding to the current time;

[0139] A303. Comparing the real-time solidification front position with the preset solidification front position, the real-time solidification front shape parameter with the preset solidification front shape parameter, and the real-time metal liquid spreading uniformity with the preset metal liquid spreading uniformity, to identify the deviation between the real-time internal solidification state and the preset state.

[0140] In step A301, the physical state parameters of the metal liquid inside the mold are calculated or predicted according to known comprehensive mode characteristics through established models or algorithms, which can be achieved by physical model-based calculation, statistical regression analysis, machine learning prediction model or rule-based reasoning, etc. The real-time solidification front position refers to the instantaneous position of the interface where the metal liquid changes from liquid to solid along the height direction of the mold during the spin casting process, which can be expressed as the vertical height from the bottom of the mold. The real-time solidification front shape parameter refers to the instantaneous geometric shape characteristics of the solidification front on the inner wall of the mold, such as its flatness, inclination angle, whether there are local protrusions or depressions, etc., which can be expressed in the form of mathematical functions, discrete point sets or shape descriptors, etc. The real-time metal liquid spreading uniformity refers to the uniformity of the metal liquid distribution on the inner wall of the mold, such as the thickness consistency along the circumferential direction or the smooth transition along the height direction, which can be measured by the standard deviation of thickness deviation, uniformity index or image analysis results, etc.

[0141] In step A301, the physical state parameters of the metal liquid inside the mold are calculated or predicted according to known comprehensive mode characteristics through established models or algorithms, which can be achieved by physical model-based calculation, statistical regression analysis, machine learning prediction model or rule-based reasoning, etc. The real-time solidification front position refers to the instantaneous position of the interface where the metal liquid changes from liquid to solid along the height direction of the mold during the spin casting process, which can be expressed as the vertical height from the bottom of the mold. The real-time solidification front shape parameter refers to the instantaneous geometric shape characteristics of the solidification front on the inner wall of the mold, such as its flatness, inclination angle, whether there are local protrusions or depressions, etc., which can be expressed in the form of mathematical functions, discrete point sets or shape descriptors, etc. The real-time metal liquid spreading uniformity refers to the uniformity of the metal liquid distribution on the inner wall of the mold, such as the thickness consistency along the circumferential direction or the smooth transition along the height direction, which can be measured by the standard deviation of thickness deviation, uniformity index or image analysis results, etc.

[0142] Specifically, the scheme converts the abstract comprehensive pattern features obtained from sensor data analysis into specific descriptions of the key physical states (solidification front position, shape, metal liquid spreading uniformity) of the mold internal metal liquid solidification process. First, the comprehensive pattern features are used to infer these real-time physical state parameters, which is a key step in mapping data analysis results to actual physical processes. Subsequently, the preset physical state parameters corresponding to the current time are obtained as a reference. Finally, by comparing the real-time parameters and the preset parameters item by item, the deviation between them is quantitatively identified. For example, the difference between the real-time solidification front position and the preset position can be calculated to evaluate whether the shape of the solidification front deviates from the preset ideal shape, and whether the spreading of the metal liquid meets the preset uniformity requirement. This method of converting abstract features into specific physical quantities and performing quantitative comparison makes the perception of internal solidification state more intuitive and accurate, clearly indicating the gap between the real-time process and the ideal process, thereby providing a direct and clear basis for determining the adjustment direction and amplitude of the process parameters.

[0143] The scheme, in combination with the aforementioned steps of obtaining sensor data and performing pattern analysis, realizes indirect perception and deviation identification of the internal solidification state that is difficult to directly monitor during spin casting, overcoming the problem of lack of real-time, quantitative evaluation of internal solidification state and its deviation in the prior art, and laying a foundation for realizing dynamic process parameter adjustment based on process state.

[0144] In some embodiments, step A4 comprises:

[0145] A401. From the deviation between the real-time internal solidification state and the preset state, extract a deviation indicator; the deviation indicator includes a solidification front position deviation indicator, a solidification front shape parameter deviation indicator, and a metal liquid spreading uniformity deviation indicator;

[0146] A402. Based on the extracted deviation indicators and the current process parameters, determine a coordinated adjustment strategy for the mold rotation speed, the pouring speed, and the cooling rate;

[0147] A403. According to the determined coordinated adjustment strategy, calculate the adjustment amount of the mold rotation speed, the pouring speed, and the cooling rate.

[0148] Wherein, the deviation indicator refers to the value extracted from the deviation between the real-time internal solidification state and the preset state.

[0149] Wherein, based on the extracted deviation indicators and the current process parameters, determining the coordinated adjustment strategy of the mold rotation speed, pouring speed and cooling rate refers to formulating a scheme to link the adjustment of the three parameters after obtaining the quantitative deviation information, combined with the current mold rotation speed, pouring speed and cooling rate values used in the spin casting process. Specifically, a rule-based method can be used to achieve this, for example, a series of "if-then" rules are preset, when a specific combination of deviation indicators and current process parameters conditions are met, the corresponding adjustment strategy is triggered (for example, "if the position deviation is too large and the spread is uneven, then appropriately increase the rotation speed and fine-tune the pouring speed"). A model-based method can also be used to achieve this, for example, a mathematical model describing the relationship between deviation indicators and process parameter adjustment amounts is established, and the optimal adjustment strategy is determined by solving the model. Further, a machine learning-based method can also be used to achieve this, for example, a neural network or support vector machine model is trained, the input is the deviation indicator and the current process parameter, and the output is the type or direction of the adjustment strategy. It is necessary to determine the coordinated adjustment strategy because the mold rotation speed, pouring speed and cooling rate interact with each other, and the adjustment of a single parameter may not effectively solve multiple deviations, and may even introduce new problems. By comprehensively analyzing the complex relationship between different deviation indicators and the three process parameters, a linkage adjustment scheme that balances the role of each parameter can be developed.

[0150] According to the determined coordinated adjustment strategy, calculating the adjustment amount of the mold rotation speed, pouring speed and cooling rate refers to converting the previously determined adjustment strategy into specific numerical values, i.e. how much each parameter needs to be increased or decreased. Specifically, a proportional-integral-derivative (PID) control method can be used to achieve this, for example, the deviation indicator is used as the error signal, and the corresponding adjustment amount is calculated by the PID controller, but the coupling between the three parameters needs to be considered, and a multivariable PID or other form of decoupling control may be required. A lookup table-based method can also be used to achieve this, for example, a table is pre-established, and the corresponding adjustment amount is looked up according to the deviation indicator and the current process parameter. Further, an optimization algorithm-based method can also be used to achieve this, for example, a target function (such as minimizing deviation) is defined, and the parameter adjustment amount that optimizes the target function is calculated by the optimization algorithm. Calculating the specific adjustment amount is a key link to convert the control strategy into actual operation, and directly guides the parameter adjustment execution mechanism of the spin casting equipment.

[0151] Thus, by extracting specific deviation indicators and determining a collaborative adjustment strategy based on these indicators and current process parameters, the final parameter adjustment amount is calculated, making the response to real-time solidification state deviation more accurate and effective. This method closely links the internal solidification state information inferred from multi-source data (including solidification front position, shape, and spreading uniformity) with process parameter adjustment, forming a closed-loop control mechanism based on real-time state feedback. Compared to solutions that only identify deviations but do not provide specific adjustment methods, this solution provides detailed steps to convert deviations into executable adjustment instructions, enabling more effective response to complex dynamic changes and non-uniformity issues in the spin casting process.

[0152] Reference Figure 2 The present application provides a hardware spin casting process parameter optimization system for adjusting process parameters in the hardware spin casting process, which comprises:

[0153] A data acquisition module 1 is used to acquire temperature distribution data of the mold, power data of the driving motor, and vibration data of key parts of the spin casting equipment during the hardware spin casting process (the specific process can refer to step A1 in the foregoing description) ;

[0154] A pattern analysis module 2 is used to perform pattern analysis on the temperature distribution data, the power data, and the vibration data to obtain comprehensive pattern characteristics reflecting the solidification state of the molten metal inside the mold (the specific process can refer to step A2 in the foregoing description) ;

[0155] An inference and identification module 3 is used to infer the real-time internal solidification state of the molten metal inside the mold according to the comprehensive pattern characteristics, and identify the deviation between the real-time internal solidification state and the preset state (the specific process can refer to step A3 in the foregoing description) ;

[0156] An adjustment amount determination module 4 is used to determine the adjustment amount of the mold rotation speed, the pouring speed, and the cooling rate according to the deviation between the real-time internal solidification state and the preset state (the specific process can refer to step A4 in the foregoing description) ;

[0157] An adjustment execution module 5 is used to adjust the mold rotation speed, the pouring speed, and the cooling rate according to the adjustment amount (the specific process can refer to step A5 in the foregoing description).

[0158] The above only describes the embodiments of the present application and is not used to limit the protection scope of the present application. For those skilled in the art, the present application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for optimizing process parameters of a hardware spin-cast forming process for adjusting process parameters of a hardware spin-cast forming process, characterized in that, The method comprises the following steps: A1. Obtain the temperature distribution data of the mold, the power data of the driving motor and the vibration data of the key parts of the spin casting equipment during the spin casting process of the hardware; A2. Perform pattern analysis on the temperature distribution data, the power data and the vibration data to obtain comprehensive pattern features reflecting the solidification state of the molten metal inside the mold; A3. According to the comprehensive pattern features, infer the real-time internal solidification state of the molten metal inside the mold, and identify the deviation between the real-time internal solidification state and the preset state; A4. According to the deviation between the real-time internal solidification state and the preset state, determine the adjustment amount of the mold rotation speed, the pouring speed and the cooling rate; A5. According to the adjustment amount, adjust the mold rotation speed, the pouring speed and the cooling rate.

2. The method of claim 1, wherein, Step A1 comprises: A101. Obtain temperature measurement values of the mold at different positions along the height direction to form the temperature distribution data; A102. Obtain the real-time power consumption data of the driving motor driving the mold rotation as the power data; A103. Obtain vibration measurement data of the key parts of the spin casting equipment to obtain the vibration data; the key parts of the spin casting equipment include at least one of the pouring gate, the bottom of the mold and the driving motor.

3. The method of claim 1, wherein the method is characterized by: Step A2 comprises: A201. Perform first pattern analysis on the temperature distribution data to obtain temperature pattern features reflecting the dynamic changes of heat transfer and solidification front along the height direction of the mold; A202. Perform second pattern analysis on the power data to obtain power pattern features reflecting the flow resistance of the molten metal, the uniformity of spreading and the stress changes during the solidification process; A203. Perform third pattern analysis on the vibration data to obtain vibration pattern features reflecting the vibrations related to the impact of the molten metal, uneven flow or potential defects; A204. Perform correlation analysis on the temperature pattern features, the power pattern features and the vibration pattern features to obtain the comprehensive pattern features.

4. The method of claim 3, wherein the method is characterized by: The temperature pattern features include at least one of the temperature gradient distribution along the height direction of the mold, the temperature change rate at different height positions, the position of the highest temperature point and the shape features of the temperature distribution curve; The power pattern features include at least one of the average value, the standard deviation, the fluctuation frequency and the amplitude of the specific frequency component of the power data; The vibration pattern features include at least one of the total energy, the main frequency and its amplitude, the peak value and the energy of a specific frequency range of the vibration data.

5. The method of claim 3, wherein the method is characterized by: Step A204 comprises: B1. Obtain real-time sequence data of the temperature pattern features, the power pattern features and the vibration pattern features; B2. Perform time series analysis on each of the real-time sequence data to obtain respective real-time trend information and fluctuation characteristic information; B3. Analyze the mutual dependence relationship between each of the real-time sequence data to obtain real-time correlation information; B4. Analyze the change law of the trend information, the fluctuation characteristic information and the correlation information over time to obtain comprehensive pattern features reflecting the evolution process of the solidification state of the molten metal inside the mold.

6. The method of claim 5, wherein the method is characterized by: Step B3 comprises: B301. Calculate the correlation index between each of the real-time sequence data; B302. According to the lead-lag relationship between each of the real-time sequence data; B303. The correlation index and the lead-lag relationship constitute the real-time association information between each mode feature.

7. The method of claim 5, wherein the method is characterized by: Step B4 includes: B401. Obtain the historical sequence data of the trend information, the fluctuation characteristic information and the association information; the historical sequence data contains real-time trend information, fluctuation characteristic information and association information; B402. Based on the historical sequence data, analyze the change rule of the trend information, the fluctuation characteristic information and the association information in the current time window; the change rule includes change rate, acceleration and periodicity; B403. Based on the historical sequence data and the change rule in the current time window, predict the change trend of the trend information, the fluctuation characteristic information and the association information in the future time window; B404. Comprehensive the historical sequence data, the change rule in the current time window and the change trend in the future time window, get the comprehensive mode feature reflecting the evolution process of the solidification state of the metal liquid in the mold.

8. The method of claim 1, wherein, Step A3 includes: A301. According to the comprehensive mode feature, deduce the real-time solidification front position, real-time solidification front shape parameter and real-time metal liquid spreading uniformity of the metal liquid in the mold; A302. Obtain the preset solidification front position, preset solidification front shape parameter and preset metal liquid spreading uniformity corresponding to the current time; A303. Compare the real-time solidification front position with the preset solidification front position, the real-time solidification front shape parameter with the preset solidification front shape parameter, and the real-time metal liquid spreading uniformity with the preset metal liquid spreading uniformity, identify the deviation between the real-time internal solidification state and the preset state.

9. The method of claim 8, wherein the method is characterized by: Step A4 includes: A401. Extract the deviation index from the deviation between the real-time internal solidification state and the preset state; the deviation index includes solidification front position deviation index, solidification front shape parameter deviation index and metal liquid spreading uniformity deviation index; A402. Based on the extracted deviation index and the current process parameter, determine the cooperative adjustment strategy of the mold rotating speed, the pouring speed and the cooling rate; A403. According to the determined cooperative adjustment strategy, calculate the adjustment amount of the mold rotating speed, the pouring speed and the cooling rate.

10. A hardware spin-cast forming process parameter optimization system for process parameter adjustment of a hardware spin-cast forming process, characterized by, The system comprises: A data acquisition module for acquiring temperature distribution data of a mold, power data of a driving motor and vibration data of a key part of a spin casting equipment during a hardware spin casting forming process; A mode analysis module for performing mode analysis on the temperature distribution data, the power data and the vibration data to obtain comprehensive mode features reflecting the solidification state of the metal liquid in the mold; a deduction and identification module for deducing the real-time internal solidification state of the metal liquid in the mold according to the comprehensive mode features, and identifying the deviation between the real-time internal solidification state and a preset state; An adjustment amount determination module is configured to determine adjustment amounts of the mold rotating speed, the pouring speed and the cooling rate according to a deviation between the real-time internal solidification state and the preset state. An adjustment execution module is configured to adjust the mold rotating speed, the pouring speed and the cooling rate according to the adjustment amounts.

Images