A device and method for modifying a diamond synthesis feedstock

Abstract

This invention belongs to the field of diamond synthesis technology, specifically a device and method for modifying diamond synthesis raw materials. The device includes a high-pressure processing chamber, a zoned heating module, a multi-point temperature sensor array, a pressure sensor, and an intelligent controller. The zoned heating module comprises multiple heating units, and the multi-point temperature sensor array is used to collect temperature data from different spatial locations in real time. The intelligent controller is configured to: receive temperature and pressure data; generate a three-dimensional temperature distribution cloud map in real time; compare the cloud map with a target temperature field model to identify hot and cold zones; and, based on the identification results and a preset prediction model, generate independent control commands for each heating unit, dynamically adjusting the output power of each heating unit. This invention, by combining zoned heating with intelligent predictive control, significantly improves the uniformity and control accuracy of the temperature field during the modification process, ensuring the consistency of raw material modification and thus improving the quality of subsequent diamond synthesis.

Description

Technical Field

[0001] This invention relates to the field of diamond synthesis technology, and in particular to a device and method for modifying diamond synthesis raw materials. Background Technology

[0002] In the synthesis of synthetic diamonds, the modification of raw materials (such as graphite powder and catalyst metal powder) is a crucial step determining the final diamond quality and consistency. This process is typically carried out under high temperature and pressure or a specific atmosphere to adjust the physicochemical properties of the raw materials, making them more conducive to diamond nucleation and growth. However, existing raw material modification equipment faces significant technical bottlenecks in practical applications.

[0003] In the prior art, for example, Chinese patent CN119386764A discloses a device and method for modifying diamond synthetic raw materials. This device can monitor the modification process of diamond raw materials, but it mainly focuses on pressure relief and alarm after modification. Another example is Chinese patent CN222019232U, which discloses a device for modifying synthetic diamond raw materials, treating harmful gases through a filter box and mentioning uniform heating of the diamond using a heating tube. However, these devices often employ single or integrated heating methods for temperature field control within the processing chamber, such as simple electric heating wires or annular heating tubes. In actual operation, due to the inherent characteristics of heat conduction, heat radiation, and the chamber structure, this method easily generates edge hot spots and central cold zones within the chamber, resulting in uneven temperature distribution.

[0004] For batch processing of powder raw materials, uneven temperature field results in different heating histories for raw material particles located in different spatial positions, leading to variations in the degree of graphite lattice distortion and catalyst alloying. This inconsistency in raw material modification directly impacts the subsequent diamond synthesis process, resulting in wide diamond crystal size distribution and large quality fluctuations, severely affecting the yield and stability of high-grade diamonds. Therefore, achieving high uniformity and precise control of the temperature field during diamond modification has become a pressing technical challenge for those skilled in the art. Summary of the Invention

[0005] Based on the technical problems existing in the prior art, the present invention proposes a device and method for modifying diamond synthetic raw materials.

[0006] This invention proposes a device and method for modifying diamond synthetic raw materials, comprising a high-pressure treatment chamber, and further comprising: A zoned heating module is disposed on the inner wall or outer periphery of the high-pressure processing cavity. The zoned heating module includes multiple heating units that are independently controlled along the axial and circumferential directions of the cavity. A multi-point temperature sensor array is embedded inside the high-pressure processing cavity to collect temperature data at different spatial locations within the high-pressure processing cavity in real time. Pressure sensors are used to monitor pressure data within the high-pressure processing chamber in real time; And an intelligent controller, which is electrically connected to the partitioned heating module, the multi-point temperature sensor array and the pressure sensor respectively; The intelligent controller is configured to perform the following operations: Receive temperature data collected by the multi-point temperature sensor array and pressure data collected by the pressure sensor; Based on the received temperature data, a three-dimensional temperature distribution cloud map of the high-pressure processing cavity is generated in real time through the built-in finite element temperature field reconstruction model. The three-dimensional temperature distribution cloud map is compared with the preset target temperature field model to identify hot spots and cold areas in the current temperature field; Based on the identification results and the pre-set pressure-temperature coupling prediction model based on long short-term memory neural network, independent control commands are generated for each heating unit to dynamically adjust the output power of each heating unit so that the actual temperature field in the cavity approaches the target temperature field.

[0007] Preferably, the multi-point temperature sensor array includes multiple sapphire fiber Bragg grating temperature sensors, which are uniformly distributed at different axial and radial positions in the high-pressure processing cavity, and the temperature sensor leads are led out of the high-pressure processing cavity through high-pressure sealing joints.

[0008] Preferably, the partitioned heating module includes at least two heating layers distributed along the axial direction of the high-pressure processing cavity. Each heating layer is composed of multiple arc-shaped heating elements evenly distributed circumferentially, and each arc-shaped heating element is an independent heating unit.

[0009] Preferably, the device further includes a process database electrically connected to the intelligent controller. The process database is used to store target temperature field models, historical temperature data, historical pressure data, and corresponding modification result evaluation data for different raw material batches and different modification processes.

[0010] Preferably, the intelligent controller is further configured to perform self-optimization iteration on the parameters of the long short-term memory neural network pressure-temperature coupling prediction model based on the current modification results.

[0011] Preferably, the step of generating independent control instructions performed by the intelligent controller specifically includes: The three-dimensional temperature distribution cloud map is input into the long short-term memory neural network pressure-temperature coupling prediction model to predict the temperature field change trend within a preset time period. Based on the changing trends, feedforward control is performed in advance on the heating units corresponding to the identified hot spots to reduce their expected power; feedforward control is performed in advance on the heating units corresponding to the identified cold spots to increase their expected power; at the same time, feedback adjustment is performed on each heating unit based on the deviation between the current temperature field and the target temperature field.

[0012] A method for modifying diamond synthetic raw materials, applied to a diamond synthetic raw material modification device, includes the following steps: Step S1: Load the diamond synthetic material to be modified into the high-pressure treatment chamber and seal the high-pressure treatment chamber; Step S2: Set the target modification process curves, including the target pressure curve and the target temperature curve, through the intelligent controller; Step S3: Start the device. The intelligent controller controls the partition heating module and pressurization system to start working. At the same time, the multi-point temperature sensor array and pressure sensor collect the temperature and pressure data in the high-pressure processing chamber in real time and transmit them to the intelligent controller. Step S4: The intelligent controller generates a three-dimensional temperature distribution cloud map based on real-time temperature data and a finite element temperature field reconstruction model; Step S5: The intelligent controller compares the three-dimensional temperature distribution cloud map with the target temperature field model at the current moment to identify temperature field deviations; Step S6: The intelligent controller inputs real-time data and temperature field deviation information into the trained long short-term memory neural network pressure-temperature coupling prediction model, and the model outputs the optimal power adjustment amount for each heating unit in the future period. Step S7: The intelligent controller generates a PWM control signal based on the optimal power adjustment amount, and independently drives each heating unit to achieve dynamic and predictive adjustment of the temperature field; Step S8: Repeat steps S3 to S7 until the modification process is complete.

[0013] 8. The method for modifying diamond synthetic raw materials according to claim 7, characterized in that the long short-term memory neural network pressure-temperature coupled prediction model in step S6 is pre-trained and continuously iteratively optimized based on the temperature, pressure, and power time series data and corresponding modification result evaluation data accumulated in historical modification batches.

[0014] Compared with the prior art, the present invention provides a device and method for modifying diamond synthetic raw materials, which has the following beneficial effects: By transforming the traditional integral heating method into independent heating units with axial layering and circumferential partitioning, and combining them with a high-density fiber optic temperature sensor array, gridded, high-resolution monitoring and control of the temperature field inside the cavity is achieved. The application of three-dimensional temperature field reconstruction technology allows the operator to see the real temperature distribution inside the cavity, thereby enabling precise fine-tuning of hot and cold zones. This ensures that raw material particles at any position inside the cavity undergo a nearly uniform heating history, guaranteeing the consistency of raw material modification from the source.

[0015] A predictive model based on a long short-term memory neural network is introduced. This model not only adjusts based on the current deviation, but more importantly, it can learn and predict the dynamic response characteristics of the temperature field as the power changes, and intervene in advance to effectively suppress temperature overshoot and fluctuations. This enables predictive and precise control of complex thermal processes, and is especially suitable for key steps such as catalyst alloying that are sensitive to temperature changes.

[0016] With its built-in process database and model iteration mechanism, the device of this invention has learning capabilities. Data from each modification process is recorded and used to optimize the model, making the device increasingly adept at handling the same type of raw materials. It can adaptively optimize control strategies, effectively reducing the cost and technical threshold of process development, and promoting the transformation of production mode from experience-driven to data-intelligent driven. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the present invention; Figure 2 This is a schematic diagram of the internal structure of the high-pressure processing chamber of the present invention; Figure 3 This is a control logic block diagram of the intelligent controller of the present invention; Figure 4 This is a schematic flowchart of the diamond synthesis raw material modification method of the present invention.

[0018] In the diagram: 100, High-pressure processing chamber; 110, Sealed end cap; 120, Pressurization interface; 130, Pressure sensor; 200, Zoned heating module; 210, Upper heating layer; 220, Middle heating layer; 230, Lower heating layer; 300, Multi-point temperature sensor array; 400, Intelligent controller; 410, Data acquisition and preprocessing module; 420, Finite element temperature field reconstruction module; 425, Temperature field deviation identification module; 430, Intelligent prediction and decision-making module; 440, Multi-channel PID coordinated control module; 450, Process database. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] In the description of this invention, it should be understood that the terms "upper", "lower", "front", "rear", "left", "right", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0021] like Figure 1 As shown in the figure, this embodiment provides a diamond synthetic raw material modification processing device, including a high-pressure processing chamber 100. The high-pressure processing chamber 100 is made of high-strength alloy steel and can withstand the high pressure (e.g., 10-100 MPa) required for diamond raw material modification processing. The top of the high-pressure processing chamber 100 is provided with an openable sealing end cap 110 for loading and unloading the diamond synthetic raw material to be modified (such as a mixture of graphite and catalyst metal powder). The bottom or side wall of the high-pressure processing chamber 100 is provided with a pressurization interface 120 that communicates with an external pressurization system (such as a hydraulic station or a gas booster pump).

[0022] To address the core issue of uneven temperature distribution in traditional devices, this device incorporates zoned heating modules 200 on the inner wall of the high-pressure processing chamber 100 or on its outer perimeter adjacent to the inner wall. Combined with... Figure 2 As shown, the partitioned heating module 200 is not a single heating element, but rather consists of multiple independently controlled heating units. Specifically, in this embodiment, the partitioned heating module 200 includes three heating layers distributed axially along the high-pressure processing chamber 100: an upper heating layer 210, a middle heating layer 220, and a lower heating layer 230. Each heating layer, such as the middle heating layer 220, consists of six arc-shaped heating elements evenly distributed along the circumference. Each arc-shaped heating element is an independent resistance heating unit, for example, made by casting high-temperature alloy resistance wire within an insulating ceramic frame. Thus, the entire partitioned heating module 200 forms a matrix of 3 (layers) × 6 (circumferential) = 18 independent heating units. Each heating unit is electrically connected to the intelligent controller 400 via an independent power cable and is controlled by an independent solid-state relay or thyristor power regulator within the intelligent controller 400 using PWM (Pulse Width Modulation), thereby achieving millisecond-level independent power adjustment.

[0023] To achieve accurate feedback for these independent heating units, a high-density temperature sensing element must be deployed within the high-pressure processing chamber 100. This device embeds a multi-point temperature sensor array 300 inside the high-pressure processing chamber 100. Considering that diamond synthesis raw material modification processing is typically accompanied by high temperatures (up to thousands of degrees Celsius) and strong electromagnetic interference (from high-power heating currents), this embodiment preferably uses sapphire fiber Bragg grating (FBG) temperature sensors, which possess extremely high temperature resistance, electromagnetic interference resistance, and long-term stability. The multi-point temperature sensor array 300 includes multiple sensor probes, which are carefully arranged within the chamber. To accurately reflect the three-dimensional temperature field, these sensor probes are distributed at different axial layers (corresponding to the upper, middle, and lower layers) and different radial positions (near the inner wall of the high-pressure processing chamber 100 and near the center of the high-pressure processing chamber 100). For example, three sensor probes can be arranged radially in the central plane of each heating layer, located in the near-wall region, the middle region, and the central region, respectively. Temperature sensor leads are routed from the side wall or bottom of the high-pressure processing chamber 100 via a specially designed high-pressure sealed connector and connected to the data acquisition card of the intelligent controller 400, ensuring sealing and reliable signal transmission under high-pressure conditions. In addition, the device is equipped with a conventional pressure sensor 130, which is also electrically connected to the intelligent controller 400 for real-time monitoring of pressure data within the high-pressure processing chamber 100.

[0024] The core of this device's intelligence is the intelligent controller 400. For example... Figure 3 As shown in the control logic block diagram, the Intelligent Controller 400 is essentially an industrial control system integrating data acquisition, model calculation, and logic control. It may consist of a high-performance PLC (Programmable Logic Controller) paired with a host computer (industrial PC), or it may utilize an embedded industrial control motherboard. The Intelligent Controller 400 mainly includes the following functional modules: Data acquisition and preprocessing module 410: This module receives multiple raw optical signals from the multi-point temperature sensor array 300 at a high-speed sampling frequency (e.g., 100Hz) and demodulates them into temperature values. At the same time, it receives pressure signals from the pressure sensor 130. This module filters and denoises these signals and packages the data. On one hand, it sends the data to the human-machine interface (HMI) for real-time display, and on the other hand, it sends the data to the core algorithm module.

[0025] Finite Element Temperature Field Reconstruction Module 420: This module incorporates a pre-established finite element temperature field reconstruction model based on the geometry and material thermophysical parameters of the high-pressure processing cavity 100. Upon receiving discrete point temperature data (e.g., 18 points) from the data acquisition and preprocessing module 410, this module uses the actual measured values ​​of these discrete points as boundary conditions and correction criteria. Through real-time calculation, it quickly solves for the continuous temperature field distribution inside the entire cavity, ultimately outputting a visualized three-dimensional temperature distribution cloud map. This cloud map clearly displays hot spots (overheated areas) and cold spots (underheated areas) at any location within the cavity.

[0026] Temperature Field Deviation Identification Module 425: This module receives the current three-dimensional temperature distribution cloud map from the finite element temperature field reconstruction module 420 and retrieves the target temperature field model corresponding to the current moment from the process database 450. This target temperature field model is an ideal temperature field state derived from a preset target modification process curve (such as a target temperature curve). For example, in the heat preservation stage, an ideal uniform temperature field model is one where the temperature at every point in the entire cavity is completely consistent with the target temperature value. The core function of the temperature field deviation identification module 425 is to perform point-by-point or region-by-region comparison calculations between the real-time generated three-dimensional temperature distribution cloud map and this preset target temperature field model. Through algorithms, this module can accurately identify the difference between the actual temperature field and the ideal temperature field, specifically including: calculating the maximum temperature deviation value and root mean square error, and clearly identifying the specific spatial location, range, and temperature difference amplitude of hot spots (areas where the actual temperature is higher than the upper limit of the target temperature) and cold zones (areas where the actual temperature is lower than the lower limit of the target temperature).

[0027] Intelligent Prediction and Decision Module 430: This module incorporates a pre-trained Long Short-Term Memory Neural Network (LSTM) pressure-temperature coupled prediction model, a deep learning network suitable for time series prediction. Its training data comes from a large amount of data accumulated in historical modification batches, including: historical power time-series data of each heating unit, historical temperature time-series data of the multi-point temperature sensor array 300, historical pressure time-series data of the pressure sensor 130, and the final modification result evaluation data of the corresponding batches. Through training, this LSTM prediction model learns the complex nonlinear dynamic relationship between "power input - temperature and pressure response - final effect". In actual operation, this module receives the current three-dimensional temperature field cloud map, pressure data, and temperature field deviation information output by the temperature field deviation identification module 425. Based on this, the model predicts the temperature field change trend for the next 5-10 seconds. More importantly, based on this prediction, the model directly outputs a control decision: the optimal power adjustment amount required by each of the 18 heating units in the future to achieve and maintain the target temperature field.

[0028] It should be noted that the modification result evaluation data are quantitative indicators used to measure the quality of modification of this batch of raw materials, including at least one or more of the following: diamond yield in subsequent diamond synthesis processes, average grain size of synthesized diamond crystals, crystal grain size distribution uniformity index, and the proportion of high-grade diamond produced. These result evaluation data are entered into the process database 450 and correlated with the aforementioned time-series data, serving together as training samples.

[0029] Multi-channel PID coordinated control module 440: This module transforms the optimal power adjustment output from the intelligent prediction and decision module 430 into specific control commands. Essentially, it is a feedforward-feedback composite controller. The adjustment from the LSTM prediction model serves as the feedforward control quantity, used to suppress predicted temperature field fluctuations in advance. Simultaneously, this module also retains PID feedback calculations for the current temperature field deviation, used to correct any minor errors that may exist in the model prediction. The combination of these two components generates the final PWM duty cycle signal, which is sent to the solid-state relays (SSRs) driving the 18 heating units, enabling independent, precise, and dynamic control of each heating unit.

[0030] Process Database 450: This database stores all process-related data, including target modification process curves (temperature, pressure) for various raw materials, target temperature field models, historical data for each treatment, and corresponding modification result evaluation data. This data serves two purposes: firstly, it allows for historical querying and tracing via the human-computer interface; secondly, it enables periodic offline or online retraining of the LSTM prediction model, achieving continuous model iteration and process self-optimization.

[0031] Based on the aforementioned diamond synthesis raw material modification and treatment apparatus, this embodiment also provides a method for modifying diamond synthesis raw materials. Combined with... Figure 4 As shown, the method includes the following steps: Step S1: Loading and sealing. The operator loads the diamond synthesis raw material to be modified into the high-pressure processing chamber 100. For example, after loading the uniformly mixed graphite-nickel-manganese-cobalt catalyst powder into a specific crucible or mold, it is placed into the high-pressure processing chamber 100, and then the sealing end cap 110 is closed and locked.

[0032] Step S2: Parameter Setting. The operator selects or inputs the target modification process on the human-machine interface of the intelligent controller 400. This includes setting the target pressure curve (e.g., first pressurizing to 50 MPa at a certain rate and holding at that pressure) and the target temperature curve (e.g., heating to 1200°C at a rate of 10°C / min and holding at that temperature for 30 minutes). The intelligent controller 400 automatically generates a target temperature field model for the corresponding time based on the target curve (e.g., the ideal state is a completely uniform temperature field, i.e., a maximum temperature difference of 0).

[0033] Step S3: Real-time data acquisition. After the device is started, the intelligent controller 400 controls the pressurization system to start working, and the partitioned heating module 200 starts heating simultaneously. Throughout the modification process, the multi-point temperature sensor array 300 collects temperature data at 18 spatial points within the high-pressure processing chamber 100 at a high frequency (e.g., 100Hz), and the pressure sensor 130 simultaneously collects pressure data. All data is transmitted to the intelligent controller 400 in real time.

[0034] Step S4: Three-dimensional temperature field reconstruction. The finite element temperature field reconstruction module 420 within the intelligent controller 400 calculates the three-dimensional temperature distribution within the entire high-pressure processing chamber 100 in real time based on the received temperature data from 18 discrete points, and generates a visualized three-dimensional temperature distribution cloud map. For example, during the heating phase, the cloud map may show that a certain area of ​​the upper heating layer 210 near the wall has a higher temperature, while the central area of ​​the middle heating layer 220 has a lower temperature.

[0035] Step S5: Temperature Field Deviation Identification. The intelligent controller 400 compares the current three-dimensional temperature distribution cloud map generated in step S4 with the target temperature field model (i.e., the ideal uniform temperature field) at the current moment. The algorithm automatically calculates the temperature deviation at each spatial point and clearly identifies the specific location, range, and temperature difference amplitude of hot spots (such as the aforementioned upper near-wall area) and cold spots (such as the aforementioned middle-layer central area).

[0036] Step S6: Intelligent Prediction and Decision-Making. The intelligent controller 400 inputs the real-time data from step S3, the temperature field deviation information identified in step S5, and the current power status of each heating unit into the pre-trained long short-term memory neural network pressure-temperature coupled prediction model. After rapid inference, the model outputs how each of the 18 heating units needs to adjust its power to eliminate the impending temperature field unevenness within a certain period of time (e.g., 5 seconds). For example, the model predicts that if no intervention is taken, the current hot spot in the upper near-wall area will overheat by more than 2°C in 3 seconds. Therefore, it makes a decision in advance: reduce the power of 1-2 arc-shaped heating elements in the corresponding area of ​​the upper heating layer 210. At the same time, in order to balance the heat, it fine-tunes the power of several heating elements around the middle heating layer 220 and appropriately increases the power of the corresponding heating unit in the central area of ​​the middle layer.

[0037] Step S7: Dynamic Command Output and Execution. The multi-channel PID coordinated control module 440 of the intelligent controller 400 receives the optimal power adjustment amount from step S6 as a feedforward command, and combines it with PID calculations of the current temperature field deviation to finally generate precise PWM control signals. These signals are sent to the power drivers of the 18 heating units, independently and dynamically adjusting the output power of each heating unit, thereby achieving fine-grained and predictive regulation of the temperature field inside the cavity.

[0038] Step S8: Repeat until completion. Steps S3 to S7 form a high-speed control closed loop, continuously cycling at millisecond intervals throughout the entire heating, holding, and cooling phases of the modification treatment. When the treatment process (e.g., holding time) ends, the intelligent controller 400 automatically controls the heating system to stop heating and controls the pressurization system to depressurize. Once the temperature drops to a safe threshold, the system prompts the operator to remove the processed modified raw material.

[0039] Step S9: Data Recording and Model Iteration. All data from this processing (temperature, pressure, power, final temperature field uniformity, etc.) are stored in the process database 450. When this batch of raw materials completes subsequent diamond synthesis and produces the final result, this result data can also be entered into the process database 450. The intelligent controller 400 can use this new data to periodically retrain or fine-tune the LSTM prediction model, continuously optimizing the model's control strategy. This will result in increasingly better temperature field control when processing similar raw materials in the future.

[0040] Through the above-mentioned device and method, the present invention improves the temperature field control of diamond synthesis raw material modification treatment from the traditional single-point, delayed, and extensive management to a three-dimensional, predictive, and intelligent level, which significantly improves the consistency of raw material modification and process stability, and has extremely high industrial application value.

[0041] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A diamond synthesis raw material modification treatment apparatus comprising a high pressure treatment chamber (100), characterized by, Also includes: A partitioned heating module (200) is disposed on the inner wall or periphery of the high-pressure processing cavity (100). The partitioned heating module (200) includes multiple heating units that are independently controlled along the axial and circumferential directions of the cavity. A multi-point temperature sensor array (300) is embedded inside the high-pressure processing cavity (100) to collect temperature data at different spatial locations within the high-pressure processing cavity (100) in real time. A pressure sensor (130) is used to monitor the pressure data inside the high-pressure processing chamber (100) in real time; And an intelligent controller (400), which is electrically connected to the partitioned heating module (200), the multi-point temperature sensor array (300) and the pressure sensor (130), respectively; The intelligent controller (400) is configured to perform the following operations: Receive temperature data collected by the multi-point temperature sensor array (300) and pressure data collected by the pressure sensor (130); Based on the received temperature data, a three-dimensional temperature distribution cloud map inside the high-pressure processing cavity (100) is generated in real time through the built-in finite element temperature field reconstruction model. The three-dimensional temperature distribution cloud map is compared with the preset target temperature field model to identify hot spots and cold areas in the current temperature field; Based on the identification results and the pre-set pressure-temperature coupling prediction model based on long short-term memory neural network, independent control commands are generated for each heating unit to dynamically adjust the output power of each heating unit so that the actual temperature field in the cavity approaches the target temperature field.

2. The diamond synthesis raw material modification and treatment device according to claim 1, characterized in that, The multi-point temperature sensor array (300) includes multiple sapphire fiber Bragg grating temperature sensors, which are uniformly distributed in different axial layers and radial positions of the high-pressure processing cavity (100). The temperature sensor leads are led out of the high-pressure processing cavity (100) through high-pressure sealing joints.

3. The diamond synthesis raw material modification and treatment device according to claim 1, characterized in that, The partitioned heating module (200) includes at least two heating layers distributed along the axial direction of the high-pressure processing cavity (100). Each heating layer is composed of multiple arc-shaped heating elements evenly distributed along the circumference, and each arc-shaped heating element is an independent heating unit.

4. The diamond synthesis raw material modification and treatment device according to claim 1, characterized in that, The device also includes a process database (450) electrically connected to the intelligent controller (400), the process database (450) being used to store target temperature field models, historical temperature data, historical pressure data and corresponding modification result evaluation data for different raw material batches and different modification processes.

5. The diamond synthesis raw material modification and treatment device according to claim 4, characterized in that, The intelligent controller (400) is also used to perform self-optimization iteration on the parameters of the long short-term memory neural network pressure-temperature coupling prediction model based on the current modification results.

6. The diamond synthesis raw material modification and treatment device according to claim 1, characterized in that, The steps for generating independent control instructions executed by the intelligent controller (400) specifically include: The three-dimensional temperature distribution cloud map is input into the long short-term memory neural network pressure-temperature coupling prediction model to predict the temperature field change trend within a preset time period. Based on the changing trends, feedforward control is performed in advance on the heating units corresponding to the identified hot spots to reduce their expected power; feedforward control is performed in advance on the heating units corresponding to the identified cold spots to increase their expected power; at the same time, feedback adjustment is performed on each heating unit based on the deviation between the current temperature field and the target temperature field.

7. A method for modifying diamond synthetic raw materials, applied to the diamond synthetic raw material modification apparatus as described in any one of claims 1 to 6, characterized in that, Includes the following steps: Step S1: Load the diamond synthetic material to be modified into the high-pressure treatment chamber (100) and seal the high-pressure treatment chamber (100). Step S2: Set the target modification process curve, including the target pressure curve and the target temperature curve, through the intelligent controller (400); Step S3: Start the device. The intelligent controller (400) controls the partition heating module (200) and pressurization system to start working. At the same time, the multi-point temperature sensor array (300) and pressure sensor (130) collect the temperature and pressure data in the high-pressure processing chamber (100) in real time and transmit them to the intelligent controller (400). Step S4: The intelligent controller (400) generates a three-dimensional temperature distribution cloud map based on real-time temperature data and a finite element temperature field reconstruction model; Step S5: The intelligent controller (400) compares the three-dimensional temperature distribution cloud map with the target temperature field model at the current moment to identify the temperature field deviation; Step S6: The intelligent controller (400) inputs real-time data and temperature field deviation information into the trained long short-term memory neural network pressure-temperature coupling prediction model, and the model outputs the optimal power adjustment amount of each heating unit in the future period. Step S7: The intelligent controller (400) generates a PWM control signal according to the optimal power adjustment amount, and independently drives each heating unit to achieve dynamic and predictive adjustment of the temperature field; Step S8: Repeat steps S3 to S7 until the modification process is complete.

8. The method for modifying diamond synthetic raw materials according to claim 7, characterized in that, The long short-term memory neural network pressure-temperature coupled prediction model in step S6 is pre-trained and continuously iteratively optimized based on the time-series data of temperature, pressure, and power accumulated in historical modification batches and the corresponding modification result evaluation data.

Images