A method and system for gradient heating control of a diamond HPHT synthesis cell

Abstract

This invention provides a gradient heating control method and system for a diamond HPHT synthesis cavity. The method includes real-time acquisition of current, voltage, displacement, and outer wall temperature data during the cavity's heating process; calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate; construction of a compaction discrimination quantity; and division of the heating process into four stages: loading coupling, compaction transition, stable heating, and heat preservation connection. In the loading coupling stage, a first constraint quantity is constructed based on the power fluctuation rate and the outer wall temperature slope; if a threshold is exceeded, the pulse energizing interval and single pulse energy are adjusted. In the compaction transition stage, a second constraint quantity is constructed based on the voltage-current ratio dispersion and displacement change rate; the on / off duration ratio is adjusted according to conditions. In the stable heating stage, the converted internal temperature of the cavity is estimated, and the temperature rise deviation and thermal inertia are combined to match the segmented slope control parameters. In the heat preservation connection stage, multiple factors are integrated to generate a final correction quantity, adjusting the input power range and cycle length, and outputting heating control commands for each stage.

Description

Technical Field

[0001] This application belongs to the field of control, and in particular relates to a gradient heating control method and system for a diamond HPHT synthesis cavity. Background Technology

[0002] High-temperature, high-pressure (HPHT) synthesis is currently the mainstream technology for industrially synthesizing synthetic diamond. The principle involves assembling a carbon source, catalyst, and pressure-transmitting medium within a synthesis chamber, where extremely high temperatures and pressures induce carbon atoms to rearrange and crystallize, forming diamond. Throughout the HPHT process, the temperature field distribution within the synthesis chamber and the heating process play a decisive role in the diamond's nucleation rate, crystal development, purity, and yield. In the initial stages of heating and pressurization, the powder or bulk material inside the chamber undergoes intense compaction and reorganization, causing random and drastic fluctuations in the contact resistance of the heating element. This can easily lead to unstable heating power or even current surges. A single control logic often cannot adapt to this thermodynamic system that changes drastically with the compaction state, easily resulting in localized overheating or thermal stress concentration in the early stages of heating, and severe temperature hysteresis in the later stages, thus disrupting the stable temperature and pressure environment for diamond crystal growth.

[0003] Existing control technologies typically combine external temperature measurements with feedback from basic electrical parameters to attempt to establish a closed-loop heating control system. A common approach involves collecting voltage and current data from the heating equipment, supplemented by thermocouple temperature measurement on the outer wall of the cavity, and employing classic algorithms such as PID control for constant power or multi-segment constant-slope temperature rise control, aiming to ensure the system smoothly heats up according to a preset time curve. However, existing technologies often fail to consider the complex force-thermal-electric coupling characteristics of the cavity during the heating process, transitioning from loose to dense. This makes it impossible to accurately perceive the actual compaction process inside, and thus, to scientifically divide the heating stages based on the compaction state. Consequently, drastic fluctuations and shocks in electrical power are unavoidable during the early transition period. Traditional multi-segment control often severs the connection between stages, neglecting the historical memory effect of the early compaction state on subsequent heat transfer efficiency. This makes the system highly susceptible to overshooting or uncontrolled fluctuations in internal temperature when facing the significant thermal inertia of the later heating and holding stages. This inability to accurately track the target temperature rise trajectory makes it easy for inclusions or lattice defects to form inside the diamond crystal, which restricts the stable preparation and mass production of high-quality single-crystal diamond. Summary of the Invention

[0004] To address the problem that existing technologies cannot adapt to the resistance jumps and thermal inertia memory effects caused by drastic changes in the compaction state during the heating process of the synthesis chamber, which leads to unstable heating power and uncontrolled temperature field, thus affecting the quality of diamond crystals.

[0005] In a first aspect, the present invention provides a gradient heating control method for a diamond HPHT synthesis cavity, comprising: Real-time acquisition of current, voltage, displacement, and outer wall temperature data of the synthesis chamber during the heating process, and calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate using the data. Based on the cumulative input heat and displacement change rate, a compaction discrimination quantity is constructed, and the heating process is divided into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section according to the change of the compaction discrimination quantity. In the loading coupling section, a first constraint is constructed based on the power fluctuation rate and the slope of the outer wall temperature change. When the first constraint exceeds a preset threshold, the pulse energizing interval and single pulse input energy are adjusted within a limited lower limit range. In the compaction transition section, a second constraint is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint meets a preset condition, the ratio of energizing duration to de-energizing duration is adjusted. During the stable heating phase, the equivalent internal temperature of the cavity is estimated based on the input power, the external wall temperature data of the cavity, and the compaction memory factor. The deviation between the equivalent internal temperature and the target temperature rise trajectory is calculated, and the segmented slope control parameters are matched based on the deviation and the thermal inertia indication. During the heat preservation transition phase, the input power range and cycle length of the control cycle are jointly adjusted based on the power fluctuation characteristics, the equivalent internal temperature change, the thermal inertia indication, and the compaction memory factor to generate the final correction amount, and the heating control commands for each stage are output.

[0006] Optionally, the real-time acquisition of current data, voltage data, displacement data, and outer wall temperature data of the synthesis cavity during the heating process, and the calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate using the data, includes: The input power corresponding to the node is obtained by multiplying the current data and voltage data read from the same sampling node. The input power of the sampled nodes is integrated over time according to the set sampling time step to obtain the cumulative input heat. The input power of a set number of consecutive sampling periods before the current moment is extracted to form a data sequence. The standard deviation of the data sequence is calculated and divided by the average value to obtain the power volatility. The displacement change rate is obtained by subtracting the displacement data of the current node from the displacement data of the previous adjacent node and dividing the absolute value of the difference by the sampling time period.

[0007] Optionally, the step of constructing a compaction discrimination quantity based on the cumulative input heat and the displacement change rate, and dividing the heating process into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section according to the change of the compaction discrimination quantity, includes: The cumulative input heat and the displacement change rate are normalized, and the monotonically increasing compaction discrimination quantity is calculated using a preset nonlinear mapping function. When the compaction discrimination value is less than the first boundary line, it is determined that the current stage is the loading coupling section; When the compaction discrimination value is greater than or equal to the first boundary line and less than the second boundary line, it is determined that the current stage is the compaction transition section. When the compaction discrimination value is greater than or equal to the second boundary line and less than the third boundary line, it is determined that the current temperature rise is in the stable heating stage. When the compaction discrimination value is greater than or equal to the third boundary line, it is determined that the current location is the insulation connection section.

[0008] Optionally, in the loading coupling section, a first constraint is constructed based on the power fluctuation rate and the slope of the outer wall temperature change. When the first constraint exceeds a preset threshold, the pulse energizing interval and single-pulse input energy are adjusted within a defined lower limit range, including: Extract a preset number of the latest cavity outer wall temperature data points during the heating process, and use the least squares method to perform linear fitting to extract the coefficient of the first term as the slope of the outer wall temperature change. The first constraint quantity is obtained by multiplying the power fluctuation rate by the slope of the outer wall temperature change; When the first constraint exceeds the preset constant threshold, a preset time extension is added to the current pulse energizing interval setting value, and the current single pulse input energy setting value is reduced according to a preset attenuation ratio until the single pulse input energy reaches the preset safe energy lower limit.

[0009] Optionally, in the compaction transition section, a second constraint is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint meets a preset condition, the ratio of the energizing duration to the de-energizing duration is adjusted, including: The voltage data and the matching current data within the energized range within a continuously set period are extracted and divided sequentially to obtain an equivalent impedance sequence. The mean square error of the sequence is then calculated to obtain the voltage-current ratio dispersion. The second constraint quantity is generated by performing a weighted exponential calculation on the dimensionless voltage-current ratio dispersion and the displacement change rate; When the second constraint value is detected to be higher than the warning extreme value, under the premise that the power-on duration is greater than the preset minimum duration, the power-on duration is reduced by a set step size and the power-off duration is increased proportionally to complete the update and writing of the wave transmission ratio parameter.

[0010] Optionally, in the stable heating phase, the equivalent internal temperature of the cavity is estimated based on the input power, the external wall temperature data of the cavity, and the compaction memory factor; the deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory is calculated; and segmented slope control parameters are matched based on the deviation and the thermal inertia indication, including: The transient characteristic constant of the compression discrimination quantity at the moment it crosses the second boundary line is extracted as the compression memory factor. Based on the preset heat conduction matrix model, the input power, the temperature data of the outer wall of the cavity and the compaction memory factor are weighted and integrally transformed to calculate the equivalent internal temperature of the cavity predicted by the current temperature. The deviation between the target temperature rise trajectory set value and the cavity converted internal temperature is calculated, and the thermal inertia indication is obtained by performing second-order differential calculation on the cavity converted internal temperature of adjacent intervals. A two-parameter coordinate system is constructed using the deviation and the thermal inertia indication. The corresponding segmented slope control parameter is extracted from a preset mapping table using an addressing matching function and issued as a voltage regulation command.

[0011] Optionally, in the insulation connection section, based on power fluctuation characteristics, converted internal temperature change, thermal inertia indication, and compression memory factor generation final correction, the input power range and cycle length of the control cycle are jointly adjusted, and heating control commands for each stage are output, including: The maximum value of the first derivative of the input power within the insulation connection section is extracted as the power fluctuation characteristic; Calculate the difference between the converted internal temperature of the cavity in two adjacent control cycles, and use it as the change in the converted internal temperature; The dimensionless thermal inertia indicator, the compaction memory factor, the power fluctuation characteristic, and the converted internal temperature change are input into a preset multivariable nonlinear characteristic polynomial to calculate the final correction amount. When the final correction value is positive, the allowable upper limit of the power output voltage is reduced according to the proportion of the final correction value. At the same time, the total length of the control cycle is extended according to the preset extension ratio to generate the heating control command, and output to the hardware execution module to respond to the heating action.

[0012] In a second aspect, the present invention also provides a gradient heating control system for a diamond HPHT synthesis chamber, comprising: The acquisition module is used to collect current data, voltage data, displacement data, and outer wall temperature data of the synthesis chamber in real time during the heating process. It uses the data to calculate the input power, cumulative input heat, power fluctuation rate, and displacement change rate. Based on the cumulative input heat and the displacement change rate, a compaction discrimination quantity is constructed. According to the change of the compaction discrimination quantity, the heating process is divided into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section. The adjustment module is used to construct a first constraint quantity based on the power fluctuation rate and the slope of the outer wall temperature change in the loading coupling section. When the first constraint quantity exceeds a preset threshold, the pulse energizing interval and single pulse input energy are adjusted within a limited lower limit range. In the compaction transition section, a second constraint quantity is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint quantity meets a preset condition, the ratio of energizing duration to de-energizing duration is adjusted. The output module is used to estimate the equivalent internal temperature of the cavity based on the input power, the temperature data of the outer wall of the cavity, and the compaction memory factor during the stable heating phase; calculate the deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory; and match the segmented slope control parameters based on the deviation and the thermal inertia indication. During the heat preservation transition phase, based on the power fluctuation characteristics, the change in equivalent internal temperature, the thermal inertia indication, and the compaction memory factor, the module generates a final correction amount and jointly adjusts the input power range and cycle length of the control cycle, and outputs heating control commands for each stage.

[0013] Furthermore, the real-time acquisition of current data, voltage data, displacement data, and outer wall temperature data of the synthesis cavity during the heating process, and the calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate using the data, includes: The input power corresponding to the node is obtained by multiplying the current data and voltage data read from the same sampling node. The input power of the sampled nodes is integrated over time according to the set sampling time step to obtain the cumulative input heat. The input power of a set number of consecutive sampling periods before the current moment is extracted to form a data sequence. The standard deviation of the data sequence is calculated and divided by the average value to obtain the power volatility. The displacement change rate is obtained by subtracting the displacement data of the current node from the displacement data of the previous adjacent node and dividing the absolute value of the difference by the sampling time period.

[0014] Furthermore, the step of constructing a compaction discrimination quantity based on the cumulative input heat and the displacement change rate, and dividing the heating process into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section according to the change of the compaction discrimination quantity, includes: The cumulative input heat and the displacement change rate are normalized, and the monotonically increasing compaction discrimination quantity is calculated using a preset nonlinear mapping function. When the compaction discrimination value is less than the first boundary line, it is determined that the current stage is the loading coupling section; When the compaction discrimination value is greater than or equal to the first boundary line and less than the second boundary line, it is determined that the current stage is the compaction transition section. When the compaction discrimination value is greater than or equal to the second boundary line and less than the third boundary line, it is determined that the current temperature rise is in the stable heating stage. When the compaction discrimination value is greater than or equal to the third boundary line, it is determined that the current location is the insulation connection section.

[0015] Further, in the loading coupling section, a first constraint is constructed based on the power fluctuation rate and the slope of the outer wall temperature change. When the first constraint exceeds a preset threshold, the pulse energizing interval and single-pulse input energy are adjusted within a defined lower limit range, including: Extract a preset number of the latest cavity outer wall temperature data points during the heating process, and use the least squares method to perform linear fitting to extract the coefficient of the first term as the slope of the outer wall temperature change. The first constraint quantity is obtained by multiplying the power fluctuation rate by the slope of the outer wall temperature change; When the first constraint exceeds the preset constant threshold, a preset time extension is added to the current pulse energizing interval setting value, and the current single pulse input energy setting value is reduced according to a preset attenuation ratio until the single pulse input energy reaches the preset safe energy lower limit.

[0016] Further, in the compaction transition section, a second constraint is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint meets a preset condition, the ratio of the energizing duration to the de-energizing duration is adjusted, including: The voltage data and the matching current data within the energized range within a continuously set period are extracted and divided sequentially to obtain an equivalent impedance sequence. The mean square error of the sequence is then calculated to obtain the voltage-current ratio dispersion. The second constraint quantity is generated by performing a weighted exponential calculation on the dimensionless voltage-current ratio dispersion and the displacement change rate; When the second constraint value is detected to be higher than the warning extreme value, under the premise that the power-on duration is greater than the preset minimum duration, the power-on duration is reduced by a set step size and the power-off duration is increased proportionally to complete the update and writing of the wave transmission ratio parameter.

[0017] Further, in the stable heating phase, the equivalent internal temperature of the cavity is estimated based on the input power, the external wall temperature data of the cavity, and the compaction memory factor; the deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory is calculated; and segmented slope control parameters are matched based on the deviation and the thermal inertia indication, including: The transient characteristic constant of the compression discrimination quantity at the moment it crosses the second boundary line is extracted as the compression memory factor. Based on the preset heat conduction matrix model, the input power, the temperature data of the outer wall of the cavity and the compaction memory factor are weighted and integrally transformed to calculate the equivalent internal temperature of the cavity predicted by the current temperature. The deviation between the target temperature rise trajectory set value and the cavity converted internal temperature is calculated, and the thermal inertia indication is obtained by performing second-order differential calculation on the cavity converted internal temperature of adjacent intervals. A two-parameter coordinate system is constructed using the deviation and the thermal inertia indication. The corresponding segmented slope control parameter is extracted from a preset mapping table using an addressing matching function and issued as a voltage regulation command.

[0018] Furthermore, in the insulation connection section, based on power fluctuation characteristics, converted internal temperature change, thermal inertia indication, and compression memory factor generation final correction, the input power range and cycle length of the control cycle are jointly adjusted, and heating control commands for each stage are output, including: The maximum value of the first derivative of the input power within the insulation connection section is extracted as the power fluctuation characteristic; Calculate the difference between the converted internal temperature of the cavity in two adjacent control cycles, and use it as the change in the converted internal temperature; The dimensionless thermal inertia indicator, the compaction memory factor, the power fluctuation characteristic, and the converted internal temperature change are input into a preset multivariable nonlinear characteristic polynomial to calculate the final correction amount. When the final correction value is positive, the allowable upper limit of the power output voltage is reduced according to the proportion of the final correction value. At the same time, the total length of the control cycle is extended according to the preset extension ratio to generate the heating control command, and output to the hardware execution module to respond to the heating action.

[0019] This invention constructs a compaction discrimination quantity through multi-dimensional operational data fusion, dividing the heating process of the diamond synthesis chamber into four independent stages with specific physical characteristics, thereby achieving refined segmented heat control. In the loading coupling stage and compaction transition stage, a dedicated constraint quantity is constructed based on multiple parameters. Within a limited range, the pulse parameters and the ratio of on / off power duration are directionally adjusted to avoid arc impact and uneven heating problems in the initial heating stage, ensuring that the chamber completes physical compaction smoothly and uniformly. After entering the stable heating stage, the internal temperature is estimated using the compaction memory factor and thermal inertia indicator, compensating for the hysteresis effect of heat transfer in the HPHT chamber in advance, making the actual temperature rise trajectory fit the target curve, fundamentally suppressing temperature overshoot or hysteresis. In the critical heat preservation transition stage, a final correction quantity is generated by combining multiple features to finely intervene in the input power and control cycle, eliminating severe temperature oscillations when entering the isothermal period. It improves the uniformity of the temperature field distribution inside the cavity and the smoothness of the heating curve, reduces the risk of cracking in the synthesis cavity and temperature runaway, and effectively ensures the growth quality and synthesis yield of high-grade diamond crystals. Attached Figure Description

[0020] Figure 1 A flowchart of a gradient heating control method for a diamond HPHT synthesis chamber; Figure 2 A schematic diagram of the closed-loop adjustment curve of pulse energy and energizing interval in the charging coupling section; Figure 3 A schematic diagram of the equivalent PWM waveform ratio contraction adjustment for the compaction transition section; Figure 4 A schematic diagram of the converted internal temperature servo follow-up response curve after correction of the final segment of the multi-dimensional parameters. Detailed Implementation

[0021] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this specification as detailed in the appended claims.

[0022] It should be understood that the terms “comprising” and “having” and any variations thereof in the embodiments of this specification are intended to cover but not exclude inclusion. For example, a product or device that includes a series of components is not necessarily limited to those components that are explicitly listed, but may include other components that are not explicitly listed or that are inherent to such product or device.

[0023] Example 1 In Embodiment 1 of the present invention, as Figure 1As shown, a gradient heating control method for a diamond HPHT synthesis cavity includes: S1. Real-time acquisition of current data, voltage data, displacement data, and outer wall temperature data of the synthesis chamber during the heating process. The input power, cumulative input heat, power fluctuation rate, and displacement change rate are calculated using the data. Based on the cumulative input heat and the displacement change rate, a compaction discrimination quantity is constructed. The heating process is divided into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section according to the change of the compaction discrimination quantity.

[0024] The AC current and AC voltage data of the synthesis chamber are collected in real time using Hall current sensors and voltage transformers, respectively. Displacement data of the synthesis chamber cylinders is collected using linear displacement sensors, and temperature data of the outer wall of the synthesis chamber is collected using an infrared thermometer. The sampling frequency of the analog-to-digital converter is set to 50Hz. The input power at the current moment is calculated by multiplying the collected equivalent current value by the equivalent voltage value. The cumulative input heat is calculated by summing and integrating the products of all input power from the start moment to the current moment with the sampling period. The standard deviation of the input power within a time sliding window consisting of the most recent 100 sampling points is calculated using the `std` function of the NumPy library in Python, and then divided by the average input power within that window to obtain the power fluctuation rate. The displacement change rate is obtained by subtracting the displacement data from the previous moment from the current displacement data and dividing by the sampling time interval.

[0025] The cumulative input heat and the displacement change rate are normalized respectively to obtain dimensionless heat parameters. and dimensionless rate of change of displacement The compaction discrimination quantity D, which is monotonically increasing and dimensionless, is calculated using a nonlinear mapping function. Based on the numerical range of D, the stages are divided as follows: when D is less than the first boundary line, the material loading coupling stage is determined; when D is greater than or equal to the first boundary line and less than the second boundary line, the compaction transition stage is determined; when D is greater than or equal to the second boundary line and less than the third boundary line, the stable heating stage is determined; and when D is greater than or equal to the third boundary line, the heat preservation transition stage is determined. The first, second, and third boundary lines correspond to the initial closure of the pressure-transmitting medium pores, the critical point of the solid-phase reaction between the graphite tube and the catalyst, and the saturation state of the stable growth temperature and pressure conditions for diamond, respectively.

[0026] In one possible embodiment, the real-time acquisition of current data, voltage data, displacement data, and outer wall temperature data of the synthesis cavity during the heating process, and the calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate using the data, includes: The input power corresponding to the node is obtained by multiplying the current data and voltage data read from the same sampling node. The input power of the sampled nodes is integrated over time according to the set sampling time step to obtain the cumulative input heat. The input power of a set number of consecutive sampling periods before the current moment is extracted to form a data sequence. The standard deviation of the data sequence is calculated and divided by the average value to obtain the power volatility. The displacement change rate is obtained by subtracting the displacement data of the current node from the displacement data of the previous adjacent node and dividing the absolute value of the difference by the sampling time period.

[0027] In the data acquisition system of the diamond HPHT synthesis press, electrical signals are synchronously acquired using a high-precision Hall current sensor and a differential voltage probe, while mechanical and temperature signals are acquired using a magnetostrictive displacement sensor and a patch-type K-thermocouple. The hardware sampling frequency is preferably set to 50Hz to 100Hz, corresponding to a sampling time step of 0.01s to 0.02s. Within each control cycle, instantaneous current and voltage are extracted and multiplied to obtain the instantaneous input power, for example, 13.208kW. Using a rectangular or trapezoidal integral algorithm, the instantaneous power is multiplied by the sampling step and accumulated in a buffer register to obtain the continuous cumulative input heat. To calculate the power fluctuation rate, a sliding data window with a depth of 100 is constructed. The sample mean and standard deviation of the power sequence within the window are calculated in real time, and the two are divided to obtain the fluctuation rate, for example, 0.05. The absolute value of the difference between the current sampling point's displacement reading and the previous node's reading is divided by the 0.02s sampling period to calculate the displacement change rate, for example, 0.1mm / s.

[0028] In one possible embodiment, the step of constructing a compaction discrimination quantity based on the cumulative input heat and the displacement change rate, and dividing the heating process into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section according to the change of the compaction discrimination quantity, includes: The cumulative input heat and the displacement change rate are normalized, and the monotonically increasing compaction discrimination quantity is calculated using a preset nonlinear mapping function. When the compaction discrimination value is less than the first boundary line, it is determined that the current stage is the loading coupling section; When the compaction discrimination value is greater than or equal to the first boundary line and less than the second boundary line, it is determined that the current stage is the compaction transition section. When the compaction discrimination value is greater than or equal to the second boundary line and less than the third boundary line, it is determined that the current temperature rise is in the stable heating stage. When the compaction discrimination value is greater than or equal to the third boundary line, it is determined that the current location is the insulation connection section.

[0029] The extreme value mapping method is used to normalize the cumulative input heat and displacement change rate. For example, a theoretical limit value for the cumulative input heat is set. for J、 The peak value of the displacement change rate is 0. 1.5mm / s If the value is 0, the normalized heat parameter and normalized displacement parameter in the interval [0,1] are calculated respectively. These two parameters are then input into a preset nonlinear mapping function, which takes the form of a compaction discrimination factor. The weighting constant The preferred value is 0.6, the exponential decay factor. The preferred value is 2.5, for displacement weight. The preferred value is 0.4. This combination of exponent and quadratic ensures that the discriminant quantity increases strictly monotonically between 0 and 1, while amplifying the characteristic weights of rapid initial displacement changes and heat accumulation in the middle and later stages. In the state machine of the control strategy, three sets of threshold parameters with clear physical boundaries are configured: the first boundary is preferably 0.15, representing the critical point when the pores of the pressure-transmitting medium in the synthetic block initially close; the second boundary is preferably 0.45, corresponding to the thermodynamic equilibrium point where the graphite tube and the catalyst begin to produce a solid-phase reaction; the third boundary is preferably 0.85, marking the saturation state of the temperature and pressure conditions required for the stable growth of diamond nuclei in the core reaction zone. The compaction discriminant quantity D, calculated in real time with a period of 10ms, is polled. Based on the different numerical ranges of the discriminant quantity, the global state pointer is safely switched to the corresponding compaction transition segment, and the temperature and pressure coupling adjustment algorithm for subsequent stages is activated simultaneously.

[0030] S2, in the loading coupling section, a first constraint quantity is constructed based on the power fluctuation rate and the slope of the outer wall temperature change. When the first constraint quantity exceeds a preset threshold, the pulse energizing interval and single pulse input energy are adjusted within a limited lower limit range. In the compaction transition section, a second constraint quantity is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint quantity meets the preset conditions, the ratio of energizing duration to de-energizing duration is adjusted.

[0031] In the loading coupling section, the power fluctuation rate is multiplied by the slope of the outer wall temperature change to obtain the first constraint value. The temperature change slope is obtained by linear regression of the outer wall temperature data over time using the least squares method. When the first constraint value exceeds a preset safety alarm threshold, it indicates a drastic change in the contact resistance of the conductive components within the synthesis cavity and a risk of localized transient overheating. At this point, the advanced timer module built into the control system's main chip is invoked. Based on a pre-defined 50ms interval, the pulse energizing interval is increased in 5ms increments, and the single-pulse input energy is gradually reduced according to a preset attenuation ratio until the first constraint value steadily falls below the safety alarm threshold. Figure 2 As shown in the figure, the graph clearly illustrates the reverse trend of the single-pulse input energy gradually decreasing from 4000J to 1800J and the pulse energizing interval gradually increasing from 50ms to 200ms during the process of adjusting the number of steps from 0 to 15. At the same time, the safe energy lower limit of 1600J is marked, which intuitively reflects the dual role of closed-loop control in preventing local overheating and avoiding the single-pulse energy from falling below the safe threshold, and helps to understand the adjustment logic of flexible heating in the loading coupling section.

[0032] In the compaction transition section, voltage data and matching current data within the energized range over a continuously set period are extracted and sequentially divided to obtain an equivalent impedance sequence. The root mean square error of this sequence is calculated to obtain the voltage-current ratio dispersion. The dimensionless voltage-current ratio dispersion is weighted and exponentially calculated with the displacement change rate to generate a second constraint quantity. When the second constraint quantity is greater than the preset severe compaction judgment dispersion threshold, it indicates that the internal pyrophyllite pressure transmission medium and graphite carbon source powder are undergoing severe displacement compaction and local arc discharge. The PID control algorithm in the Python control system library, namely the python-control library, is called to generate a duty cycle adjustment signal. Under the premise that the energization duration is greater than the preset minimum duration, the energization duration is reduced by a set step size and the de-energization duration is increased proportionally to complete the update and writing of the wave generation ratio parameter, thereby suppressing the sudden short-circuit current surge during the cavity compaction process.

[0033] In one possible embodiment, in the loading coupling section, a first constraint is constructed based on the power fluctuation rate and the slope of the outer wall temperature change. When the first constraint exceeds a preset threshold, the pulse energizing interval and single-pulse input energy are adjusted within a defined lower limit range, including: Extract a preset number of the latest cavity outer wall temperature data points during the heating process, and use the least squares method to perform linear fitting to extract the coefficient of the first term as the slope of the outer wall temperature change. The first constraint quantity is obtained by multiplying the power fluctuation rate by the slope of the outer wall temperature change; When the first constraint exceeds the preset constant threshold, a preset time extension is added to the current pulse energizing interval setting value, and the current single pulse input energy setting value is reduced according to a preset attenuation ratio until the single pulse input energy reaches the preset safe energy lower limit.

[0034] In the charging coupling section, to prevent tube burn-out accidents caused by local current concentration, an outer wall temperature queue with 20 sampling points is configured. The controller retrieves the queue data in real time and performs first-order linear regression fitting using the least squares method, extracting the slope term of the straight line as the outer wall temperature change slope. This parameter filters out temperature sampling jumps caused by ambient air cooling. Multiplying this temperature slope by the real-time calculated power fluctuation rate yields a first constraint quantity representing the severity of thermal shock. The system memory sets a constant threshold of 0.035 for this first constraint quantity. When the real-time first constraint quantity exceeds this threshold, an anti-shock flexible heating mechanism is triggered: the controller increases the time extension by 10ms each time based on the originally set 50ms pulse energizing interval, extending the pulse energizing interval to 60ms, increasing the buffer time for heat diffusion; and the current single-pulse input energy setting is reduced according to a preset attenuation ratio of 5%. The reduction logic is executed iteratively and is equipped with safety protection limits. When the continuous decay causes the single pulse energy to reach the preset safety energy lower limit, the lock will no longer be lowered to avoid the input energy being too low and causing the pressure transmission material in the synthesis chamber to fail to reach the minimum enthalpy value required for glass transition.

[0035] In one possible embodiment, during the compaction transition section, a second constraint is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint satisfies a preset condition, the ratio of the energizing duration to the de-energizing duration is adjusted, including: The voltage data and the matching current data within the energized range within a continuously set period are extracted and divided sequentially to obtain an equivalent impedance sequence. The mean square error of the sequence is then calculated to obtain the voltage-current ratio dispersion. The second constraint quantity is generated by performing a weighted exponential calculation on the dimensionless voltage-current ratio dispersion and the displacement change rate; When the second constraint value is detected to be higher than the warning extreme value, under the premise that the power-on duration is greater than the preset minimum duration, the power-on duration is reduced by a set step size and the power-off duration is increased proportionally to complete the update and writing of the wave transmission ratio parameter.

[0036] The compaction transition section is a sensitive area for the thermal deformation and phase transition of graphite and catalyst. The control system sets an observation window containing 50 pulse energizing intervals. Within this window, synchronous transient voltage and current values ​​are extracted, and equivalent impedance values ​​are obtained by performing division operations one by one, forming an equivalent impedance sequence of length 50. By calculating the root mean square deviation of all elements in this sequence from the average impedance, the dispersion of the voltage-current ratio, reflecting the roughness of the contact surface and the inhomogeneity of the conductive phase within the cavity, can be expressed. This dispersion is subjected to maximum-minimum dimensionless processing, and the dispersion and the normalized displacement change rate are substituted into the weighted exponential model, with the second constraint quantity... For example, C2 = 0.362. To prevent localized overheating of the inner wall or molten metal splashing, the warning extreme value is set to 0.30. When the current second constraint value is detected to exceed this warning extreme value, a pulse width modulation-based wave ratio shaping program is immediately started. Under the premise that the current power-on duration is greater than the preset minimum power-on duration, the power-on duration is deducted in fixed steps of 5ms, and the deducted 5ms is proportionally added to the power-off duration. This duty cycle adjustment method reduces power density while maintaining the total wave control cycle unchanged. The updated wave ratio parameters are directly written into the control register of the underlying PLC or DSP trigger, allowing the conductive microfilament network inside the cavity to be safely reconstructed under a mild thermal environment. Figure 3 As shown in the figure, the bar chart clearly illustrates the changes in the power-on duration from 40ms to the preset minimum power-on duration of 15ms and the power-off duration from 60ms to 85ms within 7 adjustment trigger cycles. It intuitively demonstrates the adjustment law that the ratio of power-on to power-off duration decreases according to the preset step size, as well as the safety limit of the minimum power-on duration of 15ms. This helps to understand the control logic of suppressing arc discharge by reducing power density in the compaction transition section.

[0037] S3, in the stable heating stage, the equivalent internal temperature of the cavity is estimated based on the input power, the temperature data of the outer wall of the cavity, and the compaction memory factor. The deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory is calculated. Based on the deviation and the thermal inertia indication, the segmented slope control parameters are matched. In the heat preservation transition stage, based on the power fluctuation characteristics, the change in equivalent internal temperature, the thermal inertia indication, and the compaction memory factor, the final correction amount is generated. The input power range and cycle length of the control cycle are jointly adjusted, and heating control commands for each stage are output.

[0038] The equivalent heat capacity coefficient at the moment when the compaction discrimination value crosses the second boundary line is extracted as the compaction memory factor. The equivalent heat capacity coefficient is obtained by least squares identification based on the input power, the rate of change of the cavity outer wall temperature, and the rate of change of displacement at that moment. Based on the current input power multiplied by the cavity equivalent thermal resistance model parameters fitted using historical experimental data, the cavity outer wall temperature data at the current moment, and the product of the compaction memory factor and the heat conduction attenuation compensation coefficient, a ternary linear regression equation is constructed. The Ridge regression algorithm model of the scikit-learn library is used in Python to estimate the converted internal temperature of the internal heating element cavity at the current moment. The actual temperature difference deviation is obtained by subtracting the estimated cavity internal temperature from the theoretical ideal temperature value at the current moment in the system's built-in target temperature time series trajectory function. The second-order time difference data of the cavity internal temperature over the past five consecutive sampling periods is calculated using the `diff` function from the NumPy library and used as a thermal inertia indicator. A control parameter matching rule library based on a fuzzy logic inference system is established. The input and output membership functions are established using the skfuzzy fuzzy logic processing library in Python. The actual temperature difference deviation and the thermal inertia indicator are used as input variables for fuzzy inference calculations. After defuzzification using the maximum membership method, the piecewise slope control parameters of the control system's PID algorithm are directly matched to obtain the new proportional gain coefficient and integral time coefficient.

[0039] In the insulation connection section, the FFT (Fast Fourier Transform) function of the NumPy library is used in Python to perform frequency domain transformation on the power time series within the most recent 30-second time window and extract the highest amplitude of power fluctuation in the low-frequency band from 1Hz to 5Hz as the power fluctuation feature. The cavity equivalent internal temperature of the current cycle is calculated by subtracting the cavity equivalent internal temperature of the previous cycle to obtain the equivalent internal temperature change. The power fluctuation feature, the equivalent internal temperature change, the thermal inertia indicator, and the compaction memory factor are combined into a four-dimensional input vector and fed into a pre-trained and converged radial basis neural network. The RBFLayer radial basis model of the PyTorch artificial intelligence framework is called to perform a forward propagation operation and output a single scalar final correction value. Based on the sign and absolute magnitude of the final correction amount, the upper and lower limit ranges of the input power and the total cycle length of the PWM signal driving the voltage regulation device are jointly fine-tuned and corrected for the next hardware control cycle. If the final correction amount is greater than zero, the upper limit of the input power is linearly reduced according to the corresponding ratio, and the cycle length is extended proportionally at the same time. The newly calculated input power range and cycle length parameters are converted into hexadecimal digital control instructions for the thyristor phase-shift voltage regulator through the industrial-grade 485 bus protocol and Modbus communication standard. These instructions are then sequentially queued and executed by the system's underlying main process to completely complete the closed-loop heating control task of each physical stage.

[0040] The radial basis function neural network (RBFNN) is a three-layer feedforward network. The input layer receives multi-dimensional feature vectors; the hidden layers use radial basis functions, typically Gaussian functions, as activation functions. Each neuron corresponds to a data center, and the network outputs a local response by calculating the Euclidean distance between the input and the center and performing a nonlinear transformation. The output layer is a linearly weighted sum, yielding a single scalar, i.e., the final correction value. The training process typically involves two steps: first, clustering, such as K-means, is used to determine the data centers and expansion constants of the hidden layer neurons; then, the output layer weights are trained using least squares or gradient descent to make the network output approximate the target value. Before training, a large amount of process data needs to be collected as a sample set. After convergence, the network parameters can be fixed for real-time inference.

[0041] In one possible embodiment, during the stable heating phase, the equivalent internal temperature of the cavity is estimated based on the input power, the external wall temperature data of the cavity, and the compaction memory factor; the deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory is calculated; and segmented slope control parameters are matched based on the deviation and the thermal inertia indication, including: The transient characteristic constant of the compression discrimination quantity at the moment it crosses the second boundary line is extracted as the compression memory factor. Based on the preset heat conduction matrix model, the input power, the temperature data of the outer wall of the cavity and the compaction memory factor are weighted and integrally transformed to calculate the equivalent internal temperature of the cavity predicted by the current temperature. The deviation between the target temperature rise trajectory set value and the cavity converted internal temperature is calculated, and the thermal inertia indication is obtained by performing second-order differential calculation on the cavity converted internal temperature of adjacent intervals. A two-parameter coordinate system is constructed using the deviation and the thermal inertia indication. The corresponding segmented slope control parameter is extracted from a preset mapping table using an addressing matching function and issued as a voltage regulation command.

[0042] When the system detects that the compaction discrimination value from the previous stage crosses the second boundary, the microcontroller immediately latches the equivalent system heat capacity and static thermal damping coefficient at that time as a compaction memory factor. In the current control cycle, a 3×3 heat conduction matrix model pre-calibrated in finite element thermodynamic simulation is used, incorporating the real-time sampled input power P and the Kalman-filtered outer wall temperature of the cavity. and extracted compacted memory factors The vector is constructed as a column vector. By performing a time-domain weighted integral transformation on this column vector, the thermal hysteresis effect caused by pressure-transmitting media such as dolomite or pyrophyllite can be compensated, and the estimated internal temperature of the cavity at the center of the synthesis reaction zone can be calculated. After obtaining the converted internal temperature, the converted internal temperature is subtracted from the pre-stored ideal target temperature rise trajectory curve to obtain the instantaneous temperature difference deviation. Using the converted internal temperatures of the past three time points, the central difference method is used to perform discrete second-order differential operations to obtain the thermal inertia indicator representing the acceleration of temperature rise and fall. The deviation and the thermal inertia indicator are combined to form a two-dimensional lookup table coordinate (x,y)=(-4.5,0.2), which is substituted into the two-dimensional fuzzy mapping table of the control system to perform bilinear interpolation addressing. This two-dimensional fuzzy mapping table is obtained by cluster analysis of temperature difference deviation and thermal inertia indicator in historical process experimental data, combined with an offline calibration of a fuzzy rule base formulated by expert experience. This mapping table extracts the segmented slope control parameters applicable to the current operating conditions. These parameters are immediately converted into voltage regulation increment commands, which are then sent to the thyristor rectifier control unit via real-time industrial networks such as the EtherCAT bus for execution, thereby achieving high-precision servo tracking of the nonlinear thermal field.

[0043] In one possible embodiment, at the insulation connection section, based on power fluctuation characteristics, converted internal temperature change, thermal inertia indication, and compression memory factor generation final correction, the input power range and cycle length of the control cycle are jointly adjusted, and heating control commands for each stage are output, including: The maximum value of the first derivative of the input power within the insulation connection section is extracted as the power fluctuation characteristic; Calculate the difference between the converted internal temperature of the cavity in two adjacent control cycles, and use it as the change in the converted internal temperature; The dimensionless thermal inertia indicator, the compaction memory factor, the power fluctuation characteristic, and the converted internal temperature change are input into a preset multivariable nonlinear characteristic polynomial to calculate the final correction amount. When the final correction value is positive, the allowable upper limit of the power output voltage is reduced according to the proportion of the final correction value. At the same time, the total length of the control cycle is extended according to the preset extension ratio to generate the heating control command, and output to the hardware execution module to respond to the heating action.

[0044] The controller continuously calculates the time-varying rate of change of the input power within a 10-second historical sliding window and detects the absolute peak value of the rate of change as a power fluctuation characteristic. The difference between the cavity-converted internal temperature measurements of two adjacent long control cycles is calculated to obtain the converted internal temperature change. These two newly acquired parameters, along with the previously inherited thermal inertia indication and compaction memory factor, are scaled and normalized, and then input into a calibrated quaternary nonlinear polynomial: [Final correction value]. . To compress the memory factor, the preferred constant term is configured as follows: =0.2, =0.5, =0.4, =0.15, and the final correction value is obtained after calculation. When the control logic determines that the final correction value is positive, it indicates that the system still has residual heat accumulation and a tendency for temperature rise overshoot. Therefore, the controller directly limits the upper limit of the allowable voltage of the primary side power output of the main transformer according to this value ratio, thereby clamping the peak input power. Based on a constant preset extension coefficient, the period extension is calculated, extending the original 1.00 second total PWM control period to 1.05 seconds, further diluting the energy injection density in the time dimension. The adjustment parameters are encapsulated into heating control commands in the form of standard industrial control messages and sent to the underlying thyristor triggers and PLC execution modules to drive the hardware to generate response actions, so as to smoothly send the cavity system into the steady-state heat preservation period, such as Figure 4 As shown in the figure, the curve clearly illustrates the following relationship between the calculated internal temperature of the cavity and the target temperature rise trajectory setting during the entire heating process: the two curves almost overlap during the stable heating segment, and the heat preservation connection segment achieves a smooth transition without overshoot through the power limiting at the end, which intuitively demonstrates the effect of the multi-dimensional parameter correction mechanism on suppressing overshoot during heating.

[0045] Example 2 Embodiment 2 of the present invention provides a gradient heating control system for a diamond HPHT synthesis chamber, comprising: The acquisition module is used to collect current data, voltage data, displacement data, and outer wall temperature data of the synthesis chamber in real time during the heating process. It uses the data to calculate the input power, cumulative input heat, power fluctuation rate, and displacement change rate. Based on the cumulative input heat and the displacement change rate, a compaction discrimination quantity is constructed. According to the change of the compaction discrimination quantity, the heating process is divided into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section. The adjustment module is used to construct a first constraint quantity based on the power fluctuation rate and the slope of the outer wall temperature change in the loading coupling section. When the first constraint quantity exceeds a preset threshold, the pulse energizing interval and single pulse input energy are adjusted within a limited lower limit range. In the compaction transition section, a second constraint quantity is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint quantity meets a preset condition, the ratio of energizing duration to de-energizing duration is adjusted. The output module is used to estimate the equivalent internal temperature of the cavity based on the input power, the temperature data of the outer wall of the cavity, and the compaction memory factor during the stable heating phase; calculate the deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory; and match the segmented slope control parameters based on the deviation and the thermal inertia indication. During the heat preservation transition phase, based on the power fluctuation characteristics, the change in equivalent internal temperature, the thermal inertia indication, and the compaction memory factor, the module generates a final correction amount and jointly adjusts the input power range and cycle length of the control cycle, and outputs heating control commands for each stage.

[0046] It should be understood that in the foregoing description of the embodiments in this specification, various features are combined in a single embodiment, drawing, or description for the purpose of simplifying the description and to aid in understanding a feature. However, this does not mean that the combination of these features is necessary, and those skilled in the art, upon reading this specification, may readily identify some of the devices as separate embodiments. That is, the embodiments in this specification can also be understood as an integration of multiple secondary embodiments. And the content of each secondary embodiment is valid even if it contains fewer than all the features of a single foregoing disclosed embodiment.

Claims

1. A gradient heating control method for a diamond HPHT synthesis cavity, characterized in that, include: Real-time acquisition of current, voltage, displacement, and outer wall temperature data of the synthesis chamber during the heating process, and calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate using the data. Based on the cumulative input heat and displacement change rate, a compaction discrimination quantity is constructed, and the heating process is divided into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section according to the change of the compaction discrimination quantity. In the loading coupling section, a first constraint is constructed based on the power fluctuation rate and the slope of the outer wall temperature change. When the first constraint exceeds a preset threshold, the pulse energizing interval and single pulse input energy are adjusted within a limited lower limit range. In the compaction transition section, a second constraint is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint meets a preset condition, the ratio of energizing duration to de-energizing duration is adjusted. During the stable heating phase, the equivalent internal temperature of the cavity is estimated based on the input power, the external wall temperature data of the cavity, and the compaction memory factor. The deviation between the equivalent internal temperature and the target temperature rise trajectory is calculated, and the segmented slope control parameters are matched based on the deviation and the thermal inertia indication. During the heat preservation transition phase, the input power range and cycle length of the control cycle are jointly adjusted based on the power fluctuation characteristics, the equivalent internal temperature change, the thermal inertia indication, and the compaction memory factor to generate the final correction amount, and the heating control commands for each stage are output.

2. The method according to claim 1, characterized in that, The real-time acquisition of current, voltage, displacement, and outer wall temperature data during the heating process of the synthesis cavity, and the calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate using the data, includes: The input power corresponding to the node is obtained by multiplying the current data and voltage data read from the same sampling node. The input power of the sampled nodes is integrated over time according to the set sampling time step to obtain the cumulative input heat. The input power of a set number of consecutive sampling periods before the current moment is extracted to form a data sequence. The standard deviation of the data sequence is calculated and divided by the average value to obtain the power volatility. The displacement change rate is obtained by subtracting the displacement data of the current node from the displacement data of the previous adjacent node and dividing the absolute value of the difference by the sampling time period.

3. The method according to claim 1, characterized in that, The method involves constructing a compaction discrimination quantity based on the cumulative input heat and the displacement change rate, and dividing the heating process into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section based on the change of the compaction discrimination quantity. The cumulative input heat and the displacement change rate are normalized, and the monotonically increasing compaction discrimination quantity is calculated using a preset nonlinear mapping function. When the compaction discrimination value is less than the first boundary line, it is determined that the current stage is the loading coupling section; When the compaction discrimination value is greater than or equal to the first boundary line and less than the second boundary line, it is determined that the current stage is the compaction transition section. When the compaction discrimination value is greater than or equal to the second boundary line and less than the third boundary line, it is determined that the current temperature rise is in the stable heating stage. When the compaction discrimination value is greater than or equal to the third boundary line, it is determined that the current location is the insulation connection section.

4. The method according to claim 2, characterized in that, In the loading coupling section, a first constraint is constructed based on the power fluctuation rate and the slope of the outer wall temperature change. When the first constraint exceeds a preset threshold, the pulse energizing interval and single-pulse input energy are adjusted within a defined lower limit range, including: Extract a preset number of the latest cavity outer wall temperature data points during the heating process, and use the least squares method to perform linear fitting to extract the coefficient of the first term as the slope of the outer wall temperature change. The first constraint quantity is obtained by multiplying the power fluctuation rate by the slope of the outer wall temperature change; When the first constraint exceeds the preset constant threshold, a preset time extension is added to the current pulse energizing interval setting value, and the current single pulse input energy setting value is reduced according to a preset attenuation ratio until the single pulse input energy reaches the preset safe energy lower limit.

5. The method according to claim 1, characterized in that, In the compaction transition section, a second constraint is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint meets a preset condition, the ratio of the energizing duration to the de-energizing duration is adjusted, including: The voltage data and the matching current data within the energized range within a continuously set period are extracted and divided sequentially to obtain an equivalent impedance sequence. The mean square error of the sequence is then calculated to obtain the voltage-current ratio dispersion. The second constraint quantity is generated by performing a weighted exponential calculation on the dimensionless voltage-current ratio dispersion and the displacement change rate; When the second constraint value is detected to be higher than the warning extreme value, under the premise that the power-on duration is greater than the preset minimum duration, the power-on duration is reduced by a set step size and the power-off duration is increased proportionally to complete the update and writing of the wave transmission ratio parameter.

6. The method according to claim 3, characterized in that, In the stable heating phase, the equivalent internal temperature of the cavity is estimated based on the input power, the external wall temperature data of the cavity, and the compaction memory factor. The deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory is calculated, and segmented slope control parameters are matched based on the deviation and the thermal inertia indication, including: The transient characteristic constant of the compression discrimination quantity at the moment it crosses the second boundary line is extracted as the compression memory factor. Based on the preset heat conduction matrix model, the input power, the temperature data of the outer wall of the cavity and the compaction memory factor are weighted and integrally transformed to calculate the equivalent internal temperature of the cavity predicted by the current temperature. The deviation between the target temperature rise trajectory set value and the cavity converted internal temperature is calculated, and the thermal inertia indication is obtained by performing second-order differential calculation on the cavity converted internal temperature of adjacent intervals. A two-parameter coordinate system is constructed using the deviation and the thermal inertia indication. The corresponding segmented slope control parameter is extracted from a preset mapping table using an addressing matching function and issued as a voltage regulation command.

7. The method according to claim 1, characterized in that, In the insulation connection section, based on power fluctuation characteristics, converted internal temperature change, thermal inertia indication, and compression memory factor generation final correction, the input power range and cycle length of the control cycle are jointly adjusted, and heating control commands for each stage are output, including: The maximum value of the first derivative of the input power within the insulation connection section is extracted as the power fluctuation characteristic; Calculate the difference between the converted internal temperature of the cavity in two adjacent control cycles, and use it as the change in the converted internal temperature; The dimensionless thermal inertia indicator, the compaction memory factor, the power fluctuation characteristic, and the converted internal temperature change are input into a preset multivariable nonlinear characteristic polynomial to calculate the final correction amount. When the final correction value is positive, the allowable upper limit of the power output voltage is reduced according to the proportion of the final correction value. At the same time, the total length of the control cycle is extended according to the preset extension ratio to generate the heating control command, and output to the hardware execution module to respond to the heating action.

8. A gradient heating control system for a diamond HPHT synthesis chamber, characterized in that, include: The acquisition module is used to collect current data, voltage data, displacement data, and outer wall temperature data of the synthesis chamber in real time during the heating process. It uses the data to calculate the input power, cumulative input heat, power fluctuation rate, and displacement change rate. Based on the cumulative input heat and the displacement change rate, a compaction discrimination quantity is constructed. According to the change of the compaction discrimination quantity, the heating process is divided into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section. The adjustment module is used to construct a first constraint quantity based on the power fluctuation rate and the slope of the outer wall temperature change in the loading coupling section. When the first constraint quantity exceeds a preset threshold, the pulse energizing interval and single pulse input energy are adjusted within a limited lower limit range. In the compaction transition section, a second constraint quantity is constructed based on the voltage-current ratio dispersion and the displacement change rate. When the second constraint quantity meets a preset condition, the ratio of energizing duration to de-energizing duration is adjusted. The output module is used to estimate the equivalent internal temperature of the cavity based on the input power, the temperature data of the outer wall of the cavity, and the compaction memory factor during the stable heating phase; calculate the deviation between the equivalent internal temperature of the cavity and the target temperature rise trajectory; and match the segmented slope control parameters based on the deviation and the thermal inertia indication. During the heat preservation transition phase, based on the power fluctuation characteristics, the change in equivalent internal temperature, the thermal inertia indication, and the compaction memory factor, the module generates a final correction amount and jointly adjusts the input power range and cycle length of the control cycle, and outputs heating control commands for each stage.

9. The system according to claim 8, characterized in that, The real-time acquisition of current, voltage, displacement, and outer wall temperature data during the heating process of the synthesis cavity, and the calculation of input power, cumulative input heat, power fluctuation rate, and displacement change rate using the data, includes: The input power corresponding to the node is obtained by multiplying the current data and voltage data read from the same sampling node. The input power of the sampled nodes is integrated over time according to the set sampling time step to obtain the cumulative input heat. The input power of a set number of consecutive sampling periods before the current moment is extracted to form a data sequence. The standard deviation of the data sequence is calculated and divided by the average value to obtain the power volatility. The displacement change rate is obtained by subtracting the displacement data of the current node from the displacement data of the previous adjacent node and dividing the absolute value of the difference by the sampling time period.

10. The system according to claim 8, characterized in that, The method involves constructing a compaction discrimination quantity based on the cumulative input heat and the displacement change rate, and dividing the heating process into a loading coupling section, a compaction transition section, a stable heating section, and a heat preservation connection section based on the change of the compaction discrimination quantity. The cumulative input heat and the displacement change rate are normalized, and the monotonically increasing compaction discrimination quantity is calculated using a preset nonlinear mapping function. When the compaction discrimination value is less than the first boundary line, it is determined that the current stage is the loading coupling section; When the compaction discrimination value is greater than or equal to the first boundary line and less than the second boundary line, it is determined that the current stage is the compaction transition section. When the compaction discrimination value is greater than or equal to the second boundary line and less than the third boundary line, it is determined that the current temperature rise is in the stable heating stage. When the compaction discrimination value is greater than or equal to the third boundary line, it is determined that the current location is the insulation connection section.

Images