Intelligent and fast arc extinguishing method, device and medium for embedded radio frequency power supply

Abstract

The application discloses an intelligent quick arc extinguishing method and device of an embedded radio frequency power supply and a medium; relates to the technical field of radio frequency power supplies; and solves the problems of low accuracy and easy false triggering of existing arc extinguishing devices. By collecting multi-dimensional operation parameters and extracting a multi-dimensional arc feature set, multi-dimensional feature fusion replaces single feature detection, reduces the false rejection rate and the missed detection rate, and avoids false triggering. An arc recognition model outputs an arc determination signal, determines the arc energy level according to the arc determination signal and the multi-dimensional arc feature set, generates a corresponding differential control instruction according to the energy level, realizes graded processing of the arc, executes a matching power regulation strategy according to different energy levels, and ensures effective arc extinguishing. The number of execution times of the arc extinguishing operation is counted, and when the cumulative number of times reaches a preset threshold, a shutdown signal is output to the radio frequency power supply, the radio frequency power supply is actively cut off in an abnormal state, irreversible damage to the equipment is prevented, and the safety and service life of the equipment are improved.

Description

Technical Field

[0001] This application relates to the field of radio frequency power supply technology, and in particular to an intelligent fast arc extinguishing method, device and medium for embedded radio frequency power supply. Background Technology

[0002] In industrial applications such as semiconductor manufacturing, radio frequency (RF) powered plasma processes are widely used. The RF power supply serves as a core component, providing high-frequency energy to the reaction chamber to form plasma. When there is charge accumulation on the dielectric film on the inner wall of the plasma process chamber, defects on the target surface, gas breakdown, or discharge at the tip of the workpiece surface, an electric arc can easily be generated in the plasma load.

[0003] Existing RF power supply arc extinguishing technology mainly adopts the following methods: it determines whether an arc has occurred by detecting the absolute value of the reflected power or its rate of change. When the detected value exceeds a preset threshold, it extinguishes the arc by directly shutting off the RF power supply or simply reducing the output power. After the arc is extinguished, the power output is restored in a simple ramp manner.

[0004] However, this one-size-fits-all, crude protection strategy has significant drawbacks: on the one hand, it cannot effectively distinguish between micro-arcs with low hazard levels (such as instantaneous burrs) and hard arcs with huge destructive power (such as continuous discharge), often leading to unnecessary deep shutdowns even when there are minor anomalies, interrupting normal process flows; on the other hand, due to the lack of multi-dimensional comprehensive analysis of arc characteristics, traditional circuits are very likely to misjudge normal impedance fluctuations in the process as arcs, causing frequent false triggers and seriously affecting production continuity.

[0005] In addition, existing arc extinguishing devices typically restore power at a fixed rate after performing power reduction operations, lacking differentiated restoration strategies for different arc levels, making it difficult to adapt to complex and ever-changing process environments.

[0006] Therefore, how to quickly and effectively extinguish arcs while ensuring process continuity and equipment safety is a technical problem that urgently needs to be solved by those in this field. Summary of the Invention

[0007] The purpose of this application is to provide an intelligent and rapid arc extinguishing method, device, and medium for embedded radio frequency power supplies, which solves the problems of low accuracy and easy false triggering of existing arc extinguishing devices.

[0008] To address the aforementioned technical problems, this application provides an intelligent fast arc extinguishing method for embedded radio frequency power supplies, comprising: The operating parameters of the RF power supply output terminal are collected, and a multi-dimensional arc feature set is constructed based on the operating parameters; The arc detection feature set is input into a preset arc recognition model for judgment, and the judgment result of whether an arc exists is output. When an electric arc is detected, the electric arc is classified into energy levels according to the electric arc detection feature set, and a corresponding control command is generated to the radio frequency power supply according to the electric arc level to adjust the output power of the radio frequency power supply. The number of times the arc extinguishing operation is executed is counted, and a shutdown signal is output to the radio frequency power supply when the cumulative number reaches a preset threshold.

[0009] Optionally, in the above-mentioned intelligent fast arc extinguishing method for embedded RF power supplies, before the step of acquiring the operating parameters of the RF power supply output and constructing a multi-dimensional arc feature set based on the operating parameters, the method further includes: During the RF power ignition stage, real-time power values, reflection coefficients, and RF on-state data are collected. When the radio frequency is on, determine whether the deviation between the real-time power value and the target value is less than a preset power deviation threshold and whether the reflection coefficient is less than a preset reflection coefficient threshold. If the current state duration meets the first preset duration, then the ignition is determined to be stable and the process proceeds to the step of collecting the operating parameters at the RF power supply output.

[0010] Optionally, in the above-mentioned intelligent fast arc extinguishing method for embedded RF power supplies, the operating parameters of the RF power supply output are collected and a multi-dimensional arc feature set is constructed based on the operating parameters, including: Collect forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage; Based on the collected forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage, the following parameters are calculated: reflected power change rate, impedance phase jump, reflection coefficient, arc energy accumulation, DC voltage abrupt change, DC voltage change rate, reflected power amplitude, and impedance imaginary part change rate. The various feature parameters are combined to form the multi-dimensional arc feature set.

[0011] Optionally, in the above-mentioned intelligent fast arc extinguishing method for embedded RF power supplies, classifying the arc energy level according to the arc detection feature set includes: When the accumulated arc energy exceeds a preset energy threshold, it is determined to be a hard arc level; When the accumulated value of the electric arc energy does not exceed the preset energy threshold, the determination is made based on the amplitude of the reflected power and the rate of change of the imaginary part of the impedance. If the reflected power amplitude is lower than the preset amplitude threshold and the rate of change of the imaginary part of the impedance is lower than the preset rate of change threshold, it is determined to be at the micro-arc level. Otherwise, it is judged as a hard arc level.

[0012] Optionally, in the above-mentioned intelligent fast arc extinguishing method for embedded RF power supplies, the step of generating corresponding control commands to the RF power supply according to the arc level to adjust the output power of the RF power supply includes: In response to the micro-arc level signal, execute the first power reduction operation; In response to the hard arc level signal, perform the second power reduction operation; Wherein, the power reduction magnitude of the second power reduction operation is greater than that of the first power reduction operation, and the response time of the second power reduction operation is shorter than that of the first power reduction operation.

[0013] Optionally, in the above-mentioned intelligent fast arc extinguishing method for embedded RF power supply, after generating corresponding control commands to the RF power supply according to the arc level to adjust the output power of the RF power supply, a power recovery step is further included: Real-time monitoring of reflection power during the recovery process; When it is determined that the arc has been eliminated and the load impedance has returned to the normal range, the power recovery operation is initiated. In response to the micro-arc level, a first power recovery operation is performed, which adopts a power recovery curve using a linear ramp plus step method; In response to the hard arc level, a second power recovery operation is performed, which uses a power recovery curve with an exponential ramp and a slow gradient smoothing method. If abnormal fluctuations in reflected power are detected during the power recovery process, the recovery operation should be stopped immediately and the step of collecting the operating parameters of the RF power supply output should be repeated. The recovery speed of the first power recovery operation is faster than that of the second power recovery operation.

[0014] Optionally, the above-mentioned intelligent fast arc extinguishing method for embedded RF power supplies further includes: Periodically read the entire process data of arc extinguishing operation from the storage unit and filter valid samples; Based on the supervised learning algorithm, the feature weights, judgment thresholds and level classification rules of the electric arc recognition model are optimized using the effective samples as the training set. The optimized model parameters are then updated to the arc recognition model.

[0015] To address the aforementioned technical problems, this application also provides an intelligent fast arc extinguishing device for an embedded radio frequency power supply, comprising: The parameter acquisition and processing module is used to acquire the operating parameters of the RF power supply output and construct a multi-dimensional arc feature set based on the operating parameters. An arc detection module is used to input the arc detection feature set into a preset arc detection model for judgment and output a judgment result on whether an arc exists. The arc extinguishing execution module is used to classify the energy level of the electric arc according to the electric arc detection feature set when the existence of the electric arc is determined, and to generate corresponding control commands to the radio frequency power supply according to the electric arc level to adjust the output power of the radio frequency power supply. The safety protection module is used to count the number of times the arc extinguishing operation is performed, and outputs a shutdown signal to the radio frequency power supply when the cumulative number reaches a preset threshold.

[0016] To address the aforementioned technical problems, this application also provides an intelligent fast arc extinguishing device for an embedded radio frequency power supply, comprising: Memory, used to store computer programs; The processor is used to implement the steps of the above-described intelligent fast arc extinguishing method for embedded radio frequency power supply when executing the computer program.

[0017] To address the aforementioned technical problems, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the aforementioned intelligent fast arc extinguishing method for embedded radio frequency power supplies.

[0018] The intelligent rapid arc extinguishing method for embedded RF power supplies provided in this application collects multi-dimensional operating parameters and extracts multi-dimensional arc feature sets. Based on these multi-dimensional features, a comprehensive judgment is made, with multi-dimensional feature fusion replacing single-feature detection. An intelligent arc recognition model replaces simple threshold comparison, reducing false positive and false negative rates, avoiding false triggering, and improving equipment operational stability. The arc recognition model outputs an arc judgment signal, and the arc energy level is determined based on the arc judgment signal and the multi-dimensional arc feature set. Corresponding differentiated control commands are generated according to the energy level to achieve graded arc processing. Matching power adjustment strategies are executed for different energy levels to avoid direct shutdown in cases of minor anomalies, ensuring effective arc extinguishing. The number of arc extinguishing operations is counted, and when the cumulative number reaches a preset threshold, a shutdown signal is output to the RF power supply. In abnormal conditions, the RF power supply is actively cut off to prevent irreversible damage to the equipment, improving equipment safety and lifespan.

[0019] In addition, this application also provides an apparatus and a medium that correspond to the above-mentioned intelligent fast arc extinguishing method for embedded radio frequency power supplies, with the same effect. Attached Figure Description

[0020] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This application provides a flowchart of an intelligent fast arc extinguishing method for an embedded radio frequency power supply. Figure 2 A schematic diagram illustrating a specific implementation of an intelligent fast arc extinguishing method for an embedded radio frequency power supply provided in this application embodiment; Figure 3 A schematic diagram of power change curves for graded composite arc extinguishing and power recovery provided in an embodiment of this application; Figure 4 A structural diagram of an intelligent fast arc extinguishing device for an embedded radio frequency power supply provided in an embodiment of this application; Figure 5 This is a structural diagram of another intelligent fast arc extinguishing device for embedded radio frequency power supply provided in an embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0023] The core of this application is to provide an intelligent and rapid arc extinguishing method, device, and medium for embedded radio frequency power supplies.

[0024] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] This application provides an intelligent fast arc extinguishing method for embedded radio frequency power supplies, such as... Figure 1 As shown, it includes: S1: Collect the operating parameters of the RF power supply output and construct a multi-dimensional arc feature set based on the operating parameters; S2: Input the arc detection feature set into the preset arc recognition model for judgment, and output the judgment result of whether an arc exists; S3: When an electric arc is detected, the energy level of the electric arc is classified according to the electric arc detection feature set, and a corresponding control command is generated to the radio frequency power supply to adjust the output power of the radio frequency power supply according to the electric arc level. S4: Count the number of times the arc extinguishing operation is executed, and output a shutdown signal to the RF power supply when the cumulative number reaches a preset threshold.

[0026] It should be noted that the method in this embodiment typically operates in an embedded RF power supply system. The hardware architecture of this system mainly includes a digital signal processor (DSP) and a programmable gate array (FPGA), and the arc extinguishing process is executed by setting up various processing modules in the FPGA and DSP.

[0027] First, step S1 collects the operating parameters of the RF power supply output and constructs a multi-dimensional arc feature set based on the operating parameters.

[0028] This step refers to the real-time acquisition of raw operating parameters, including but not limited to forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage, through a parameter acquisition unit located at the output of the RF power generator. It should be noted that the operating parameters mentioned here are not limited to the above-mentioned types; they can also include other parameters that can reflect the characteristics of the electric arc, such as load impedance phase and RF reflectivity. This embodiment does not impose strict limitations, as long as it can be used to construct a multi-dimensional electric arc feature set.

[0029] Based on the above principle, the raw operating parameters collected are processed by the feature extraction unit of the FPGA to extract multi-dimensional feature parameters such as the rate of change of reflected power, impedance phase jump, reflection coefficient, cumulative value of arc energy, DC voltage mutation value, DC voltage change rate, reflected power amplitude and rate of change of the imaginary part of impedance, and these feature parameters are combined to form a multi-dimensional arc feature set.

[0030] This step refers to using high-speed hardware circuits to capture physical quantities in real time. Specifically, the high-speed analog-to-digital converter (ADC) in the parameter acquisition module acquires the fundamental components of forward power, reflected power, voltage, and current in real time at a sampling rate of not less than 100 MSPS; at the same time, the DC voltage acquisition unit acquires the DC voltage at the load terminal at a frequency of not less than 1 kHz.

[0031] Because it uses multi-dimensional feature acquisition instead of a single reflection power feature, this embodiment can more comprehensively capture various physical phenomena when an electric arc occurs, thus providing richer judgment basis for subsequent arc identification.

[0032] Secondly, step S2 inputs the arc detection feature set into a preset arc recognition model for judgment, and outputs a judgment result indicating whether an arc exists. The arc recognition model can be a deep learning model, such as a CNN-LSTM hybrid model (a model combining convolutional neural networks and long short-term memory networks), or other machine learning models with classification capabilities. This model is pre-trained on historical arc data and can output the probability or judgment result of whether an arc exists based on the input multi-dimensional arc feature set. That is, this embodiment uses intelligent model judgment to replace the traditional fixed threshold comparison method.

[0033] At this point, since the model makes a comprehensive judgment based on multi-dimensional features, it can effectively distinguish between normal impedance fluctuations in the process and real arc events, thereby significantly reducing the misjudgment rate.

[0034] Step S3: When an electric arc is detected, the arc is classified into energy levels based on the arc detection feature set. Corresponding control commands are then generated for the RF power supply to adjust its output power. The energy level classification does not necessarily mean only two levels; it can be two levels: micro-arc and hard arc. More levels can also be defined as needed.

[0035] Specifically, when an electric arc is detected, the system further combines the characteristics of the electric arc energy accumulation value, reflected power amplitude, and impedance imaginary part change rate from the multi-dimensional electric arc feature set to determine whether the current electric arc is a micro-arc (low energy, transient) or a hard arc (high energy, continuous). Then, differentiated power adjustment commands are generated according to different levels.

[0036] Because of the tiered processing strategy, micro-arcs can be eliminated with only a slight power reduction intervention, while hard arcs require a deep power reduction or even shutdown. Therefore, this embodiment can effectively extinguish arcs while avoiding unnecessary deep shutdowns during minor anomalies, thus ensuring process continuity.

[0037] Step S4 counts the number of arc-extinguishing operations performed and outputs a shutdown signal to the RF power supply when the cumulative number reaches a preset threshold. This step refers to the real-time counting of the number of arc-extinguishing operations performed. When the cumulative number of arc-extinguishing operations reaches the preset threshold (e.g., 8 times, which can be configured according to actual process requirements), it indicates that the equipment may be in an abnormal state, and repeated arcing may cause irreversible damage to the cavity equipment and RF circuits. At this time, a shutdown signal is actively output to cut off the RF power supply, and an audible and visual alarm message is generated.

[0038] Steps S2-S4 are implemented sequentially through the arc recognition module, arc extinguishing execution module, and safety protection module set in the DSP. The arc recognition module has a built-in arc recognition model.

[0039] The intelligent rapid arc extinguishing method for embedded RF power supplies provided in this application collects multi-dimensional operating parameters and extracts multi-dimensional arc feature sets. Based on these multi-dimensional features, a comprehensive judgment is made, with multi-dimensional feature fusion replacing single-feature detection, and intelligent arc recognition model judgment replacing simple threshold comparison. This reduces false positive and false negative rates, avoids false triggering, and improves equipment operational stability. The arc recognition model outputs an arc judgment signal, and the arc energy level is determined based on the arc judgment signal and the multi-dimensional arc feature set. Corresponding differentiated control commands are generated according to the energy level to achieve graded arc processing. Matching power adjustment strategies are executed for different energy levels to avoid direct shutdown in cases of minor anomalies, ensuring effective arc extinguishing. The number of arc extinguishing operations is counted, and when the cumulative number reaches a preset threshold, a shutdown signal is output to the RF power supply. In abnormal conditions, the RF power supply is actively cut off to prevent irreversible damage to the equipment, improving equipment safety and lifespan.

[0040] According to the above embodiment, specifically, before the step of collecting the operating parameters of the RF power supply output terminal and constructing a multi-dimensional arc feature set based on the operating parameters, the method further includes: During the RF power ignition stage, real-time power values, reflection coefficients, and RF on-state data are collected. When the radio frequency is on, determine whether the deviation between the real-time power value and the target value is less than a preset power deviation threshold and whether the reflection coefficient is less than a preset reflection coefficient threshold. If the current state duration meets the first preset duration, then the ignition is determined to be stable and the process proceeds to the step of collecting the operating parameters at the RF power supply output.

[0041] During the RF power ignition stage, real-time power values, reflection coefficients, and RF on-state data are collected.

[0042] It should be noted that during the ignition phase of an RF power supply, the power and reflection coefficient naturally fluctuate during frequency modulation. If arc suppression detection is activated directly, these normal fluctuations can easily be misinterpreted as an electric arc, leading to false triggering. In other words, the ignition phase itself is an unstable transition process, and any abnormal signals detected at this time are not actual electric arcs.

[0043] Then, this embodiment executes the determination logic: when the radio frequency is on, it determines whether the deviation between the real-time power value and the target value is less than a preset power deviation threshold (e.g., 50W) and whether the reflection coefficient is less than a preset reflection coefficient threshold (e.g., 0.2).

[0044] This step refers to the fact that the deviation between the real-time power value and the target value reflects whether the power has reached the set value, and the reflection coefficient reflects the impedance matching status. Only when both conditions are met can it be said that the RF power supply has entered a stable operating state.

[0045] In addition, it is necessary to determine whether the duration of the above state meets the first preset duration (e.g., 3-10 seconds, or 5 seconds). The instantaneous fulfillment of the condition may be accidental; only a sustained period of stability can confirm true ignition stability.

[0046] It should be noted that the above threshold parameters (power deviation threshold, reflection coefficient threshold, and first preset duration) can all be configured according to the specific process scenario and RF power supply model, and this embodiment does not impose strict limitations.

[0047] Based on the above principle, ignition is determined to be stable only when the radio frequency is turned on, the absolute value of the deviation between the real-time power value and the target value is less than or equal to the preset power deviation threshold and the variance is stable, and the reflection coefficient is continuously lower than the preset threshold for a first preset time. Then, the step of collecting the operating parameters of the radio frequency power supply output terminal (i.e., turning on the arc extinguishing detection) is entered.

[0048] Therefore, this embodiment solves the problem of false arc extinction caused by signal fluctuations during the ignition stage by using a pre-ignition stability determination step, thereby improving the ignition success rate and operational stability of the equipment.

[0049] According to the above embodiments, specifically, the operation parameters of the RF power supply output are collected, and a multi-dimensional arc feature set is constructed based on the operation parameters, including: Collect forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage; Based on the collected forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage, the following parameters are calculated: rate of change of reflected power, impedance phase jump, reflection coefficient, cumulative arc energy, DC voltage abrupt change, DC voltage rate of change, reflected power amplitude, and rate of change of the imaginary part of impedance. The various feature parameters are combined to form a multi-dimensional arc feature set.

[0050] The system collects forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage. The forward power and reflected power can be separated and sampled in real time using a bidirectional coupler; the fundamental voltage and fundamental current components can be acquired using a high-speed analog-to-digital converter (ADC) with a sampling rate of not less than 100 MSPS; and the DC voltage can be acquired using a DC voltage acquisition unit (with an acquisition frequency of not less than 1 kHz).

[0051] It should be noted that the above data acquisition methods are merely exemplary and do not imply that these specific devices must be used. For example, forward power and reflected power can also be obtained through other power detection methods. This embodiment does not impose strict limitations.

[0052] Based on the collected forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage, the following parameters are calculated: rate of change of reflected power, impedance phase jump, reflection coefficient, cumulative arc energy, DC voltage abrupt change, DC voltage rate of change, reflected power amplitude, and rate of change of the imaginary part of impedance.

[0053] This step refers to the feature extraction unit performing calculations on the original sampled data to extract multi-dimensional parameters that can reflect the characteristics of the electric arc.

[0054] The reflected power change rate reflects the rate of change of reflected power; the reflected power increases sharply when an arc occurs. The impedance phase jump reflects the amount of change in the load impedance phase; an arc causes a sudden change in the plasma state. The reflection coefficient reflects the degree of impedance matching. The arc energy accumulation value is calculated based on the fundamental components of voltage and current, reflecting the total energy of the arc discharge. The DC voltage jump value reflects the magnitude of the DC voltage drop. The DC voltage change rate reflects the rate of drop of the DC voltage. The reflected power amplitude reflects the absolute magnitude of the reflected power. The impedance imaginary part change rate reflects the rate of change of the load reactance component.

[0055] It should be noted that not all of the above feature parameters must be used simultaneously. A selection of features can be chosen to construct the feature set based on the actual process scenario. This embodiment lists eight feature parameters as preferred options, but this does not represent a strict limitation on the composition of the feature set.

[0056] The various feature parameters are combined to form a multi-dimensional arc feature set. These feature parameters can be organized into feature vectors for use in subsequent arc recognition models.

[0057] Based on the above principles, the multi-dimensional arc feature set constructed in this embodiment covers information from multiple dimensions such as power, impedance, and voltage, and can comprehensively reflect various physical phenomena during arc occurrence. Compared with the existing technology that relies only on a single reflected power feature, the multi-dimensional feature set provided in this embodiment offers richer and more reliable judgment criteria for subsequent intelligent identification.

[0058] Furthermore, the aforementioned characteristic parameters are all calculated in real time in the FPGA, with a sampling granularity reaching the microsecond level, thus enabling the capture of the rapid changes in transient electric arcs.

[0059] Specifically, the energy level of the electric arc is classified based on the arc detection feature set, including: When the accumulated arc energy exceeds the preset energy threshold, it is determined to be a hard arc level; When the cumulative value of the electric arc energy does not exceed the preset energy threshold, the determination is made based on the amplitude of the reflected power and the rate of change of the imaginary part of the impedance. If the reflected power amplitude is lower than the preset amplitude threshold and the rate of change of the imaginary part of the impedance is lower than the preset rate of change threshold, it is determined to be at the micro-arc level. Otherwise, it is judged as a hard arc level.

[0060] When the accumulated arc energy exceeds a preset energy threshold, it is classified as a hard arc. The accumulated arc energy is calculated based on the fundamental components of voltage and current and reflects the total energy of the arc discharge. It should be noted that the accumulated energy of a continuous arc (hard arc) will continuously increase, while the accumulated energy of a transient arc (micro-arc) will be lower. Therefore, when the accumulated arc energy exceeds the preset energy threshold, it can be directly classified as a hard arc without further judgment.

[0061] When the accumulated arc energy does not exceed the preset energy threshold, a determination is made based on the reflected power amplitude and the rate of change of the imaginary part of the impedance. This step refers to the need to further distinguish between a true micro-arc and the initial stage of a hard arc where the accumulated energy has not yet reached the threshold for an arc with a low energy accumulation value. At this point, the reflected power amplitude and the rate of change of the imaginary part of the impedance become the key distinguishing indicators.

[0062] If the reflected power amplitude is lower than the preset amplitude threshold and the rate of change of the imaginary part of the impedance is lower than the preset rate of change threshold, it is determined to be a micro-arc level; otherwise, it is determined to be a hard arc level.

[0063] Micro-arcs are typically transient, low-energy arcs with low reflected power amplitudes and minimal impact on load impedance (lower rate of change of the imaginary part of impedance); while hard arcs, even in their initial stages, may have higher reflected power amplitudes or a more significant impact on load impedance (higher rate of change of the imaginary part of impedance).

[0064] Furthermore, the preset amplitude threshold and preset rate of change threshold can be configured according to specific process scenarios. The optimal threshold can be determined based on historical data statistics. This embodiment does not impose strict limitations on specific values.

[0065] Based on the above principles, this embodiment achieves a refined classification of electric arc levels. First, obvious hard arcs are quickly screened out using the cumulative arc energy value; then, for boundary cases, precise distinctions are made using the reflected power amplitude and the rate of change of the imaginary part of the impedance. Thus, this hierarchical judgment logic ensures both judgment efficiency and accuracy.

[0066] Specifically, based on the arc level, corresponding control commands are generated and sent to the RF power supply to adjust the output power of the RF power supply, including: In response to the micro-arc level signal, execute the first power reduction operation; In response to the hard arc level signal, perform the second power reduction operation; The power reduction magnitude of the second power reduction operation is greater than that of the first power reduction operation, and the response time of the second power reduction operation is shorter than that of the first power reduction operation.

[0067] The system responds to a micro-arc level signal and performs the first power reduction operation. When the signal is determined to be at the micro-arc level, a corresponding control command is generated, and the power reduction operation is performed within a relatively slow time (e.g., 1-10 μs) with a small power reduction magnitude (e.g., reducing the RF power to 30%-70% of the original power).

[0068] In response to a hard arc level signal, a second power reduction operation is executed. When a hard arc level is detected, a corresponding control command is generated, and the power reduction operation is executed within a very short time (e.g., less than 1 μs) with a large power reduction magnitude (e.g., reducing the RF power to less than 10% of the original power or shutting it off directly).

[0069] It should be noted that the parameters mentioned above (response time, power reduction amplitude) are merely illustrative and do not represent strict numerical limitations. For example, the micro-arc power reduction time can be 1μs, 5μs, or 10μs, and the power reduction amplitude can be 30%, 50%, or 70% of the original power. The specific values ​​can be configured according to process requirements. This embodiment does not impose strict limitations.

[0070] The power reduction amplitude of the second power reduction operation is greater than that of the first power reduction operation, and the response time of the second power reduction operation is shorter than that of the first power reduction operation. Since micro-arcs are low-energy, transient arcs with a lower degree of hazard, a gentler power reduction operation can effectively extinguish them while avoiding excessive disturbance to the process. Hard arcs, on the other hand, are high-energy, continuous arcs with a high degree of hazard, requiring a rapid and deep power reduction operation to effectively extinguish them and prevent damage to equipment and wafers.

[0071] Based on the above principles, this embodiment implements a graded composite arc extinguishing strategy. Different power reduction rates and amplitudes are adopted for different levels of arcs; after the power reduction is completed, the arc is maintained for a preset duration (e.g., 20 μs for micro-arcs and 100 μs for hard arcs) to ensure that the arc is completely extinguished.

[0072] Therefore, this embodiment ensures the arc extinguishing effect while maximizing process continuity. Compared to the single strategies in existing technologies that either directly shut off the power supply (leading to process interruption) or simply reduce power (resulting in low arc extinguishing efficiency), the graded composite arc extinguishing strategy is more scientific and efficient.

[0073] Specifically, after generating corresponding control commands to the RF power supply based on the arc level to adjust the output power of the RF power supply, a power recovery step is also included: Real-time monitoring of reflection power during the recovery process; When it is determined that the arc has been eliminated and the load impedance has returned to the normal range, the power recovery operation is initiated. In response to the micro-arc level, the first power recovery operation is executed, which adopts a power recovery curve with a linear ramp and step method. In response to the hard arc level, a second power recovery operation is performed, which uses a power recovery curve with an exponential ramp and a slow gradient smoothing method. If abnormal fluctuations in reflected power are detected during the power recovery process, the recovery operation should be stopped immediately and the steps of collecting the operating parameters of the RF power supply output should be repeated. The recovery speed of the first power recovery operation is faster than that of the second power recovery operation.

[0074] It should be noted that power recovery is a critical step in the arc extinguishing process. An inappropriate recovery strategy can easily trigger a secondary arc, leading to repeated arcing and even equipment damage. Therefore, this embodiment designs an intelligent power recovery mechanism that matches the arc level.

[0075] Real-time monitoring of reflected power during the recovery process. This refers to continuously monitoring the reflected power data output by the parameter acquisition module after arc extinguishing, providing a basis for subsequent recovery initiation determination and anomaly monitoring.

[0076] Power restoration is initiated when the arc is determined to be extinguished and the load impedance returns to normal. Power restoration can only be initiated if the arc is truly extinguished and the load has returned to normal. Forcibly restoring power before the load has returned to normal may immediately trigger a secondary arc. In other words, the prerequisite for power restoration is that the load condition is safe.

[0077] In response to the micro-arc level, the first power recovery operation is performed, employing a power recovery curve with a linear ramp and step pattern. For the micro-arc level, due to the minimal disturbance to the process, a faster recovery speed can be used. The linear ramp and step pattern means that the power increases at a linear rate, possibly with step-like plateaus, forming a smooth but rapid recovery curve.

[0078] In response to a hard arc level, a second power recovery operation is executed, employing an exponential ramp followed by a slow gradient smoothing method for power recovery. For hard arc levels, due to the deep power reduction operation, load recovery requires greater caution. The exponential ramp followed by a slow gradient smoothing method means that power increases slowly along an exponential curve, with an extremely slow recovery rate initially, gradually accelerating as the load condition improves, resulting in a smooth, slow recovery curve.

[0079] During power recovery, if abnormal fluctuations in reflected power are detected, the recovery operation is immediately stopped, and the step of collecting operating parameters from the RF power supply output is repeated. This indicates that the load may not have fully recovered, or there are signs of arc recurrence. If power recovery continues at this time, it is very likely to trigger a secondary arc. Therefore, this embodiment is designed with a mechanism to stop recovery and re-detect, eliminating the generation of secondary arcs at the source.

[0080] The first power recovery operation has a faster recovery speed than the second power recovery operation. This difference in recovery speed is designed based on the different characteristics of micro-arc and hard-arc processes: the faster micro-arc recovery is to restore the process as quickly as possible and minimize the impact on production capacity.

[0081] Based on the above principles, this embodiment achieves intelligent power recovery that matches the arc level. First, recovery initiation requires load status confirmation; second, the recovery speed matches the arc level; furthermore, anomalies are monitored in real time during the recovery process, and the process is immediately terminated upon detection of an anomaly. Therefore, this embodiment fundamentally eliminates the generation of secondary arcs. Compared to the simple recovery mechanism in existing technologies with fixed-rate recovery and no real-time monitoring, the intelligent power recovery method of this embodiment is safer and more reliable.

[0082] Specifically, it also includes: Periodically read the entire process data of arc extinguishing operation from the storage unit and filter valid samples; Based on supervised learning algorithms, the feature weights, judgment thresholds, and classification rules of the electric arc recognition model are optimized using effective samples as the training set. The optimized model parameters are then updated in the arc recognition model.

[0083] The entire arc extinguishing process data is periodically read from the storage unit to filter valid samples. The storage unit (such as flash memory / memory card) records all data for each arc extinguishing detection in real time, including feature sets, judgment results, arc extinguishing strategies, execution effects, power recovery curves, etc. The data storage granularity reaches the microsecond level, ensuring the integrity and traceability of historical data. This includes normal operation samples (feature data when there is no arc), micro-arc samples (feature data when a micro-arc occurs), hard arc samples (feature data when a hard arc occurs), as well as misjudged and missed samples (data where the model made incorrect judgments).

[0084] Based on supervised learning algorithms, the feature weights, decision thresholds, and classification rules of the electric arc recognition model are optimized using effective samples as the training set. The supervised learning algorithm can be a CNN-LSTM hybrid model or other machine learning algorithms with classification capabilities. The DSP processor uses historical data as the training set to retrain or incrementally train the electric arc recognition model, optimizing its feature weights, decision thresholds, and classification rules, thereby reducing the false positive and false negative rates.

[0085] The optimized model parameters are then updated to the arc recognition model. After the model iteration is complete, the DSP updates the optimized model parameters (including weight matrix, bias term, threshold parameters, etc.) to the arc recognition model, enabling the model to use the optimized parameters for judgment in actual operation.

[0086] Based on the above principles, this embodiment achieves self-learning and self-optimization of the arc extinguishing detection model. First, training data is accumulated through the historical data storage module; second, the DSP periodically reads the data and iterates the model; furthermore, the optimized parameters are updated to the arc recognition model in real time. Thus, this embodiment forms a complete data-driven model iteration closed loop.

[0087] Furthermore, because the model can be continuously optimized, the system can adapt to changes in arc characteristics under different process scenarios, and the system's adaptability is continuously improved. For example, when changes in process conditions (such as changing the target material or adjusting the gas flow rate) cause changes in arc characteristics, the model can automatically adapt to the new feature distribution through iterative optimization without manual intervention.

[0088] Figure 2 This is a schematic diagram illustrating a specific implementation of an intelligent rapid arc extinguishing method for an embedded radio frequency (RF) power supply provided in this application. Initially, the RF power supply is turned on. Relevant parameters are collected in real-time, including forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage. Determining whether the RF is working properly refers to determining whether the RF power supply has entered a stable operating state. If not, the process returns to continue collecting data; if yes, it proceeds to the next step. Here, whether the RF is working properly corresponds to the ignition stability determination.

[0089] Once ignition is determined to be stable, the arc extinguishing detection function is activated, and the cumulative number of arc extinguishing attempts is increased by 1.

[0090] Data processing generates a multi-dimensional arc detection feature set. The arc identification model determines whether an arc exists. If no arc is detected, monitoring continues; if an arc is detected, the process proceeds to the next step. Arc level classification increments the arc count by 1. Based on the arc detection feature set, the arcs are classified into energy levels, and the arc count is increased by 1.

[0091] The safety protection module determines whether the cumulative number of arc extinguishing attempts has reached a preset threshold. If the threshold is reached, a warning is generated and the RF power supply is turned off; if not, when the cumulative number of arc extinguishing attempts reaches the preset threshold, an audible and visual alarm is generated and a shutdown signal is output to the RF power supply to cut off the RF output.

[0092] The graded arc extinguishing strategy is implemented, and corresponding control commands are generated according to the arc level.

[0093] Monitor whether the arc has been extinguished. If it has not been extinguished, return to enhanced arc extinguishing measures (such as extending the duration or increasing the power reduction). If it has been extinguished, implement a power recovery strategy.

[0094] Perform the corresponding power restoration operation based on the arc level.

[0095] Determine if the power has reached the preset value. If not, return to continue recovery; if it has, save the data and continue monitoring the arc.

[0096] Figure 3 This is a schematic diagram of the power change curve for a graded composite arc extinguishing and power recovery method provided in an embodiment of this application. Figure 3 As shown, t0: arc extinguishing start time; t1: hard arc power reduction completion time; t2: micro-arc power reduction completion time; t3: power recovery start time; t4: micro-arc recovery completion time; t5: hard arc recovery completion time; k3: micro-arc power reduction slope; k4: micro-arc power increase slope; k1: hard arc power reduction slope; k2: hard arc power increase slope; Parc_thd: arc grading threshold. The micro-arc curve is the upper curve; the hard arc curve is the lower curve. It can be seen that the micro-arc uses a gentler power reduction and a faster recovery, while the hard arc uses an extremely fast power reduction and a slower recovery. This ensures the arc extinguishing effect while maintaining process continuity and preventing the generation of secondary arcs.

[0097] The above embodiments have described in detail the intelligent fast arc extinguishing method for embedded RF power supplies. This application also provides embodiments corresponding to the intelligent fast arc extinguishing device for embedded RF power supplies. It should be noted that this application describes the embodiments of the device part from two perspectives: one is based on the functional modules, and the other is based on the hardware.

[0098] From the perspective of functional modules Figure 4 A structural diagram of an intelligent fast arc extinguishing device for an embedded radio frequency power supply provided in this application embodiment is shown below. Figure 4 As shown, an intelligent fast arc extinguishing device for an embedded radio frequency power supply includes: The parameter acquisition and processing module 11 is used to acquire the operating parameters of the RF power supply output and construct a multi-dimensional arc feature set based on the operating parameters; The arc recognition module 12 is used to input the arc detection feature set into the preset arc recognition model for judgment and output the judgment result of whether an arc exists. The arc extinguishing execution module 13 is used to classify the energy level of the electric arc according to the electric arc detection feature set when the existence of the electric arc is determined, and to generate corresponding control commands to the radio frequency power supply according to the electric arc level to adjust the output power of the radio frequency power supply. The safety protection module 14 is used to count the number of times the arc extinguishing operation is performed, and outputs a shutdown signal to the radio frequency power supply when the cumulative number reaches a preset threshold.

[0099] The output of the RF power generator is connected to the input of an impedance matching network, and the output of the impedance matching network is connected to the input of the plasma process chamber load. The impedance matching network transforms the impedance of the plasma process chamber load to a value that matches the output impedance of the RF power generator, thereby reducing reflected power and improving power transmission efficiency. The plasma process chamber load is an area prone to arc generation. The parameter acquisition and processing module connects its input between the output of the RF power generator and the impedance matching network. It is used to acquire the operating parameters of the RF power output and construct a multi-dimensional arc feature set based on these parameters.

[0100] In this embodiment, the parameter acquisition and processing module corresponds to the high-speed ADC and DSP data processing unit in the hardware, responsible for constructing the feature set. The arc recognition module corresponds to the processor core running the CNN-LSTM model, responsible for outputting the judgment result. The arc extinguishing execution module corresponds to the FPGA logic control and RF power generator drive circuit, responsible for performing graded power reduction. The safety protection module corresponds to the counting logic and relay control circuit in the DSP, responsible for counting the number of times and emergency shutdown.

[0101] The parameter acquisition and processing module includes: a parameter acquisition unit and a feature extraction unit; The parameter acquisition unit is used to acquire forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage. The feature extraction unit is used to calculate the rate of change of reflected power, impedance phase jump, reflection coefficient, arc energy accumulation, DC voltage mutation, DC voltage change rate, reflected power amplitude, and impedance imaginary part change rate based on the collected forward power, reflected power, voltage fundamental component, current fundamental component, and DC voltage; and to combine the feature parameters to form a multi-dimensional arc feature set.

[0102] Also includes: The startup module is used to collect real-time power values, reflection coefficients, and RF on-state data during the RF power ignition phase. The judgment module is used to determine whether the deviation between the real-time power value and the target value is less than a preset power deviation threshold and whether the reflection coefficient is less than a preset reflection coefficient threshold when the radio frequency is turned on. The stable trigger module is used to determine that ignition is stable and proceed to the step of collecting the operating parameters of the RF power supply output if the duration of the current state meets the first preset duration.

[0103] Also includes: The recovery monitoring module is used to monitor the reflected power in real time during the recovery process; The recovery operation module is used to initiate power recovery operation when it is determined that the arc has been eliminated and the load impedance has returned to the normal range; The first recovery module is used to respond to the micro-arc level and perform the first power recovery operation. The first power recovery operation adopts a power recovery curve with a linear ramp and step method. The second recovery module is used to respond to the hard arc level and perform the second power recovery operation. The second power recovery operation adopts a power recovery curve with an exponential ramp and a slow gradient smoothing method. The stop recovery module is used to immediately stop the recovery operation and re-execute the step of collecting the operating parameters of the RF power supply output if abnormal fluctuations in reflected power are detected during the power recovery process. The recovery speed of the first power recovery operation is faster than that of the second power recovery operation.

[0104] Also includes: The update module is used to periodically read the running data of the entire arc extinguishing operation from the storage unit and filter valid samples; The learning module is used to optimize the feature weights, decision thresholds, and classification rules of the electric arc recognition model based on supervised learning algorithms and using effective samples as the training set. The iteration module is used to update the optimized model parameters to the arc recognition model.

[0105] Since the embodiments of the apparatus and the embodiments of the method correspond to each other, please refer to the description of the embodiments of the method for the embodiments of the apparatus, which will not be repeated here.

[0106] Figure 5 A structural diagram of another intelligent fast arc extinguishing device for embedded radio frequency power supply provided in this application embodiment is shown below. Figure 5 As shown, the intelligent fast arc extinguishing device for embedded RF power supply includes: a memory 20 for storing computer programs; The processor 21 is used to execute a computer program to implement the steps of the method for obtaining user operation habit information as described in the above embodiment (intelligent fast arc extinguishing method for embedded RF power supply).

[0107] The intelligent fast arc extinguishing device for embedded RF power supply provided in this embodiment can include, but is not limited to, mobile terminals, personal computers, workstations, etc.

[0108] The processor 21 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 21 may be implemented using at least one of the following hardware forms: Digital Signal Processor (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 21 may also include a main processor and a coprocessor. The main processor, also known as the Central Processing Unit (CPU), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 21 may integrate a Graphics Processing Unit (GPU), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 21 may also include an Artificial Intelligence (AI) processor, which handles computational operations related to machine learning.

[0109] The memory 20 may include one or more computer-readable storage media, which may be non-transitory. The memory 20 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In this embodiment, the memory 20 is used to store at least the following computer program 201, which, after being loaded and executed by the processor 21, is capable of implementing the relevant steps of the intelligent fast arc extinguishing method for embedded RF power supplies disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 20 may also include an operating system 202 and data 203, and the storage method may be temporary or permanent storage. The operating system 202 may include Windows, Unix, Linux, etc. The data 203 may include, but is not limited to, data involved in implementing the intelligent fast arc extinguishing method for embedded RF power supplies.

[0110] In some embodiments, the intelligent fast arc extinguishing device of the embedded RF power supply may further include a display screen 22, an input / output interface 23, a communication interface 24, a power supply 25, and a communication bus 26.

[0111] Those skilled in the art will understand that Figure 5 The structure shown does not constitute a limitation on the intelligent fast arc extinguishing device for embedded RF power supplies and may include more or fewer components than shown.

[0112] The intelligent fast arc extinguishing device for embedded radio frequency power supply provided in this application includes a memory and a processor. When the processor executes the program stored in the memory, it can implement the following method: intelligent fast arc extinguishing method for embedded radio frequency power supply.

[0113] Finally, this application also provides an embodiment corresponding to a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps described in the above embodiment of the intelligent fast arc extinguishing method for embedded RF power supplies.

[0114] It is understood that if the methods in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and executes all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0115] The computer-readable storage medium provided in this embodiment stores a computer program. When the processor executes the program, the following method can be implemented: an intelligent and fast arc extinguishing method for embedded radio frequency power supplies.

[0116] The above provides a detailed description of the intelligent fast arc extinguishing method, apparatus, and medium for embedded radio frequency power supplies provided in this application. The various embodiments in the specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

[0117] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A smart and fast arc extinguishing method for an embedded radio frequency power supply, characterized in that, include: The operating parameters of the RF power supply output terminal are collected, and a multi-dimensional arc feature set is constructed based on the operating parameters; The arc detection feature set is input into a preset arc recognition model for judgment, and the judgment result of whether an arc exists is output. When an electric arc is detected, the electric arc is classified into energy levels according to the electric arc detection feature set, and a corresponding control command is generated to the radio frequency power supply according to the electric arc level to adjust the output power of the radio frequency power supply. The number of times the arc extinguishing operation is executed is counted, and a shutdown signal is output to the radio frequency power supply when the cumulative number reaches a preset threshold.

2. The intelligent fast arc extinguishing method for embedded RF power supply according to claim 1, characterized in that, Before the step of acquiring the operating parameters of the radio frequency power supply output and constructing a multi-dimensional arc feature set based on the operating parameters, the method further includes: During the RF power ignition stage, real-time power values, reflection coefficients, and RF on-state data are collected. When the radio frequency is on, determine whether the deviation between the real-time power value and the target value is less than a preset power deviation threshold and whether the reflection coefficient is less than a preset reflection coefficient threshold. If the current state duration meets the first preset duration, then the ignition is determined to be stable and the process proceeds to the step of collecting the operating parameters at the RF power supply output.

3. The intelligent fast arc extinguishing method for embedded RF power supply according to claim 1, characterized in that, The operating parameters of the RF power supply output are collected, and a multi-dimensional arc feature set is constructed based on the operating parameters, including: Collect forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage; Based on the collected forward power, reflected power, fundamental voltage component, fundamental current component, and DC voltage, the following parameters are calculated: reflected power change rate, impedance phase jump, reflection coefficient, arc energy accumulation, DC voltage abrupt change, DC voltage change rate, reflected power amplitude, and impedance imaginary part change rate. The various feature parameters are combined to form the multi-dimensional arc feature set.

4. The intelligent fast arc extinguishing method for embedded radio frequency power supply according to claim 3, characterized in that, The electric arc is classified into energy levels based on the electric arc detection feature set, including: When the accumulated arc energy exceeds a preset energy threshold, it is determined to be a hard arc level; When the accumulated value of the electric arc energy does not exceed the preset energy threshold, the determination is made based on the amplitude of the reflected power and the rate of change of the imaginary part of the impedance. If the reflected power amplitude is lower than the preset amplitude threshold and the rate of change of the imaginary part of the impedance is lower than the preset rate of change threshold, it is determined to be at the micro-arc level. Otherwise, it is judged as a hard arc level.

5. The intelligent fast arc extinguishing method for embedded RF power supply according to claim 4, characterized in that, The step of generating corresponding control commands to the radio frequency power supply based on the arc level to adjust the output power of the radio frequency power supply includes: In response to the micro-arc level signal, execute the first power reduction operation; In response to the hard arc level signal, perform the second power reduction operation; Wherein, the power reduction magnitude of the second power reduction operation is greater than that of the first power reduction operation, and the response time of the second power reduction operation is shorter than that of the first power reduction operation.

6. The intelligent fast arc extinguishing method for embedded RF power supply according to claim 5, characterized in that, After generating corresponding control commands to the radio frequency power supply based on the arc level to adjust the output power of the radio frequency power supply, the method further includes a power recovery step: Real-time monitoring of reflection power during the recovery process; When it is determined that the arc has been eliminated and the load impedance has returned to the normal range, the power recovery operation is initiated. In response to the micro-arc level, a first power recovery operation is performed, which adopts a power recovery curve using a linear ramp plus step method; In response to the hard arc level, a second power recovery operation is performed, which uses a power recovery curve with an exponential ramp and a slow gradient smoothing method. If abnormal fluctuations in reflected power are detected during the power recovery process, the recovery operation should be stopped immediately and the step of collecting the operating parameters of the RF power supply output should be repeated. The recovery speed of the first power recovery operation is faster than that of the second power recovery operation.

7. The intelligent fast arc extinguishing method for embedded radio frequency power supply according to claim 1, characterized in that, Also includes: Periodically read the entire process data of arc extinguishing operation from the storage unit and filter valid samples; Based on the supervised learning algorithm, the feature weights, judgment thresholds and level classification rules of the electric arc recognition model are optimized using the effective samples as the training set. The optimized model parameters are then updated to the arc recognition model.

8. An intelligent fast arc extinguishing device for an embedded radio frequency power supply, characterized in that, include: The parameter acquisition and processing module is used to acquire the operating parameters of the RF power supply output and construct a multi-dimensional arc feature set based on the operating parameters. An arc detection module is used to input the arc detection feature set into a preset arc detection model for judgment and output a judgment result on whether an arc exists. The arc extinguishing execution module is used to classify the energy level of the electric arc according to the electric arc detection feature set when the existence of the electric arc is determined, and to generate corresponding control commands to the radio frequency power supply according to the electric arc level to adjust the output power of the radio frequency power supply. The safety protection module is used to count the number of times the arc extinguishing operation is performed, and outputs a shutdown signal to the radio frequency power supply when the cumulative number reaches a preset threshold.

9. An intelligent fast arc extinguishing device for an embedded radio frequency power supply, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the intelligent fast arc extinguishing method for an embedded radio frequency power supply as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the intelligent fast arc extinguishing method for an embedded radio frequency power supply as described in any one of claims 1 to 7.

Images