Explosion arc pulse power supply control method and system for vacuum ion plating diamond or DLC

Abstract

The invention belongs to the technical field of vacuum ion plating, and provides an explosion arc pulse power supply control method and system for vacuum ion plating diamond or DLC, and the method comprises the steps: initialization, minimum control window setting, data monitoring and feature extraction, optimal regulation and control parameter matching, controller regulation and control, historical regulation and control data collection, and lookup table dynamic updating. The lookup table construction process comprises the steps of original controller setting, feedback-based adjustment, group data integration, regulation and control evaluation, data division, similarity screening, clustering analysis and lookup table updating. According to the method, the lookup table is constructed and the dynamic updating strategy is set, so that the quick matching of the control parameters is realized, the control strategy has the self-learning ability on the basis of keeping the real-time performance of PID control, and the response speed and the control precision are both considered; and through a closed-loop optimization mechanism based on quality evaluation, it is ensured that the control target is the coating quality all the time, and the reliability of the process and the consistency of the coating quality are improved.

Description

Technical Field

[0001] This invention relates to the field of vacuum ion plating technology, and in particular to a method and system for controlling the explosion arc pulse power supply for vacuum ion plating of diamond or DLC. Background Technology

[0002] Vacuum ion plating, as an important surface treatment technology, plays an increasingly important role in modern manufacturing. The core of this technology is the generation of plasma in a vacuum environment, which allows metal or non-metal ions to deposit on the substrate surface to form a thin film, thereby endowing the substrate material with excellent wear resistance, corrosion resistance, and optical properties to meet the needs of various working conditions. Within the vacuum ion plating technology system, explosion arc ion plating is a highly efficient coating method. Its basic principle is to use an electric arc discharge to generate high-temperature plasma on a cathode target. Ions in the plasma are accelerated and bombard the substrate surface under the action of an electric field, achieving thin film deposition. Currently, this technology is widely used in the surface treatment of cutting tools, molds, aerospace components, and other fields, becoming one of the key technologies for improving product performance. In recent years, with the continuous growth of industrial production's demand for high-performance coatings, diamond and DLC coatings have attracted much attention due to their excellent performance. Among them, diamond coatings have extremely high hardness, good thermal conductivity, and chemical stability, and have broad application prospects in cutting tools, abrasives, and other fields; DLC coatings have similar properties to diamond, while also possessing good toughness and a low coefficient of friction, and are widely used in mechanical parts, optical devices, and other fields.

[0003] However, the preparation of diamond and DLC coatings in production is extremely sensitive to process parameters, especially placing very high demands on the power control system. Traditional explosion arc pulse power supply control systems often suffer from low control accuracy, slow response speed, and poor stability when facing complex plasma discharge processes. These problems directly affect the consistency and reliability of coating quality. Currently, the control methods for vacuum ion plating explosion arc pulse power supplies mainly fall into the following categories, each with certain limitations: 1) The traditional PID control method is currently the most commonly used control method, adjusting the output through three stages: proportional (P), integral (I), and derivative (D). However, due to the highly nonlinear and time-varying characteristics of the plasma discharge process, traditional PID controllers struggle to achieve precise control, often resulting in low control accuracy in practical applications; 2) Some studies have attempted to introduce fuzzy logic control to improve control performance by establishing a fuzzy rule base to handle nonlinear problems. However, the formulation of fuzzy control rules often relies on the designer's experience, lacks adaptability, and is difficult to handle complex multivariate coupling problems; 3) Some studies have adopted neural network control methods. Although neural networks have strong nonlinear mapping capabilities, they suffer from long training times and poor real-time performance in practical applications. Especially in applications requiring rapid response, such as plasma discharge, the latency problem of neural networks is particularly prominent. They cannot track the rapid changes in the arc in a timely manner, easily leading to control failure. In summary, the main limitations of existing control methods are concentrated in the following four aspects: Insufficient control accuracy: Traditional control methods struggle to adapt to the complex dynamic characteristics of plasma discharge, resulting in large fluctuations in output parameters, which in turn affects the consistency of coating quality and fails to meet the coating requirements of high-end products; Slow response speed: Faced with rapid changes in the arc, the adjustment speed of existing control methods often cannot keep up with the rate of change, leading to control lag and problems such as arc instability and abnormal discharge; Poor adaptability: Lacking the ability to adaptively adjust to different operating conditions, it is difficult to automatically optimize control parameters according to changes in process conditions, requiring manual adjustments by operators, increasing operational difficulty and the requirements for employees' experience and knowledge; Lack of intelligent decision-making: Existing methods are mainly based on simple feedback adjustment, lacking the ability to analyze and learn from historical data, failing to achieve intelligent control, and struggling to meet the development needs of intelligent manufacturing. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the purpose of this invention is to provide a method and system for controlling the explosion arc pulse power supply for vacuum ion plating of diamond or DLC, which solves the problems of insufficient control accuracy, slow response speed, poor adaptability and lack of intelligent decision-making in the existing methods.

[0005] To achieve the above objectives, the present invention provides the following solution: A method for controlling the explosion arc pulse power supply for vacuum ion plating of diamond or DLC, comprising: The initial parameters of the PID improved controller are set, and the monitoring device is calibrated; the monitoring device includes: a current and voltage detection module, a capacitor energy storage detection module, and an ion information detection module. Set the minimum control window based on the arc triggering frequency; Based on the minimum control window, the monitoring device is used to collect the current and voltage signals of the discharge circuit, the energy storage status data of the capacitor bank, and the plasma parameters to obtain working monitoring data. Feature extraction is then performed on the working monitoring data to obtain multi-dimensional features. The matching degree is calculated and parameters are filtered by using the multi-dimensional features and PID parameters at the current moment and the pre-constructed PID parameter lookup table to obtain the optimal control parameters, and the optimal control parameters are updated to the PID improved controller. The target explosion arc pulse unit is controlled using the aforementioned improved PID controller; Collect data on the process of the PID improved controller regulating the target explosion arc pulse unit to obtain historical regulation data; The PID parameter lookup table is dynamically updated based on the historical control data using the PID parameter lookup table construction method. The process of constructing the PID parameter lookup table includes: A PID controller is set up; the PID controller is a PID controller based on feedback error. The target explosion arc pulse unit is controlled using the PID raw controller, and the PID parameters and the working monitoring data are collected at each moment during operation. The PID parameters corresponding to the current moment, the PID parameters corresponding to the previous moment, and the work monitoring data are integrated into basic group data; The work monitoring data is evaluated using a set comprehensive evaluation function to obtain the work performance evaluation result; Based on the work performance evaluation results and the set evaluation threshold, the basic group data is divided into a normal control group and an abnormal control group, and the normal control group is updated to the PID parameter lookup table. The similarity function is used to filter the basic group data in all the normal control groups and the target abnormal control group whose similarity exceeds the similarity threshold, to obtain a set of similar groups; Cluster analysis is performed on the PID parameters corresponding to the current time in the set of similar groups to obtain optimized PID control parameters; The optimized PID control parameters are used to replace the PID parameters corresponding to the current time in the target abnormal control group, and the replaced target abnormal control group is updated in the PID parameter lookup table.

[0006] Preferably, it further includes: Data whose current work monitoring data and PID parameters match any item in the PID parameter lookup table with a degree lower than the matching degree threshold are marked as data with no matching item. The set of parameters with the highest matching degree between the PID parameter lookup table and the data with no matching items is updated to the PID improvement controller, and the work performance evaluation result corresponding to the work monitoring data at the next time moment is extracted. When the work performance evaluation result is in a normal state, the data with no matching items and the PID parameters corresponding to the data with no matching items are updated to the PID parameter lookup table. When the work performance evaluation result is in an abnormal state, the data with no matching items is marked as a feedback control group, and the controller corresponding to the feedback control group is switched to the PID original controller. The PID original controller is used to regulate the data of the unmatched items whose matching degree with any of the feedback control groups exceeds the matching degree threshold, and the PID parameters after regulation and the working monitoring data are recorded to obtain the candidate group data; The candidate group data is processed and updated according to the method for constructing the PID parameter lookup table to obtain the updated PID parameter lookup table.

[0007] Preferably, the expression for the minimum control window is: ;in, This refers to the minimum control window; The window coefficient; This is the arc triggering frequency.

[0008] Preferably, the multi-dimensional features include: time-domain features, frequency-domain features, statistical features, and plasma features; the frequency-domain features include: total harmonic distortion; the statistical features include: mean, variance, skewness, and kurtosis; and the plasma features include: electron temperature, electron density, and ion energy distribution.

[0009] Preferably, the similarity function employs a cosine similarity algorithm.

[0010] Preferably, the clustering analysis uses the K-means++ clustering algorithm.

[0011] Preferably, the comprehensive evaluation function is: ;in, The evaluation result of the work performance; , , , These are the first weight, second weight, third weight, and fourth weight, respectively. As a hardness evaluation index; For thickness evaluation; As a uniformity evaluation index; It serves as a stability evaluation index.

[0012] Preferably, an explosion arc pulse power supply control system for vacuum ion plating of diamond or DLC includes: a charging module, an energy storage capacitor bank, an arc ignition triggering module, a discharge circuit module, and a main control module; the energy storage capacitor bank is composed of multiple high-voltage capacitors connected in parallel; the arc ignition triggering module embeds a laser arc ignition unit and a high-frequency high-voltage arc ignition unit. The charging module is connected to the energy storage capacitor bank; the energy storage capacitor bank is connected to the arc ignition module; the charging module, the energy storage capacitor bank, the arc ignition module, and the discharge circuit module are respectively connected to the main control module; the discharge circuit module, together with the graphite target material and the anode inside the target vacuum cavity, forms a discharge channel; The charging module is used to charge the energy storage capacitor bank; the main control module is used to adaptively and dynamically regulate the charging module, the energy storage capacitor bank, the arc triggering module, and the discharge circuit module.

[0013] Preferably, the charging module has an adjustment range of 300V to 500V; the high-voltage capacitor has parameters of 1000μF / 600V and a parallel connection range of 4 to 8 capacitors; the discharge circuit module has copper busbars as its conductors; and the discharge circuit module has a circuit impedance of less than 50mΩ.

[0014] Preferably, the laser arc-starting unit is an Nd:YAG laser with a wavelength of 1064nm; the output power of the laser arc-starting unit is in the range of 10W to 20W; and the output voltage of the high-frequency high-voltage arc-starting unit is in the range of 5kV to 8kV and the frequency is in the range of 1MHz to 2MHz.

[0015] The present invention discloses the following technical effects: This invention provides a method that solves the problems of insufficient control precision, slow response speed, poor adaptability, and lack of intelligent decision-making in existing methods by constructing a lookup table and setting a dynamic update strategy, thereby achieving rapid matching of control parameters. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the explosion arc pulse power supply control process for vacuum ion plating of diamond or DLC provided in an embodiment of the present invention; Figure 2 A control flowchart provided for embodiments of the present invention; Figure 3 A flowchart illustrating the lookup table construction process provided in this embodiment of the invention; Figure 4 A flowchart of the auxiliary update process provided for embodiments of the present invention; Figure 5 This is a schematic diagram of the power supply structure provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] The purpose of this invention is to provide a method and system for controlling the explosion arc pulse power supply for vacuum ion plating of diamond or DLC, which solves the problems of insufficient control accuracy, slow response speed, poor adaptability and lack of intelligent decision-making in existing methods.

[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] Figure 1 This is a schematic diagram of the explosion arc pulse power supply control process for vacuum ion plating of diamond or DLC provided in an embodiment of the present invention. Figure 2 This is a control flowchart provided in an embodiment of the present invention. Figure 3 A flowchart illustrating the lookup table construction process provided in this embodiment of the invention is shown below. Figures 1 to 3 As shown, the present invention provides a method for controlling the explosion arc pulse power supply for vacuum ion plating of diamond or DLC, comprising: Step 100: Set the initial parameters of the PID improved controller and calibrate the monitoring device; the monitoring device includes: a current and voltage detection module, a capacitor energy storage detection module, and an ion information detection module; Step 200: Set the minimum control window according to the arc triggering frequency; Step 300: Based on the minimum control window, use the monitoring device to collect the current and voltage signals of the discharge circuit, the energy storage status data of the capacitor bank, and the plasma parameters to obtain working monitoring data, and extract features from the working monitoring data to obtain multi-dimensional features; Step 400: Calculate the matching degree and filter parameters by using the multi-dimensional features and PID parameters at the current moment with the pre-built PID parameter lookup table to obtain the optimal control parameters, and update the optimal control parameters to the PID improved controller; Step 500: Use the PID improved controller to regulate the target explosion arc pulse unit; Step 600: Collect data on the process of the PID improved controller regulating the target explosion arc pulse unit to obtain historical regulation data; Step 700: Dynamically update the PID parameter lookup table based on the historical control data using the PID parameter lookup table construction method; The process of constructing the PID parameter lookup table includes: A PID controller is set up; the PID controller is a PID controller based on feedback error. The target explosion arc pulse unit is controlled using the PID raw controller, and the PID parameters and the working monitoring data are collected at each moment during operation. The PID parameters corresponding to the current moment, the PID parameters corresponding to the previous moment, and the work monitoring data are integrated into basic group data; The work monitoring data is evaluated using a set comprehensive evaluation function to obtain the work performance evaluation result; Based on the work performance evaluation results and the set evaluation threshold, the basic group data is divided into a normal control group and an abnormal control group, and the normal control group is updated to the PID parameter lookup table. The similarity function is used to filter the basic group data in all the normal control groups and the target abnormal control group whose similarity exceeds the similarity threshold, to obtain a set of similar groups; Cluster analysis is performed on the PID parameters corresponding to the current time in the set of similar groups to obtain optimized PID control parameters; The optimized PID control parameters are used to replace the PID parameters corresponding to the current time in the target abnormal control group, and the replaced target abnormal control group is updated in the PID parameter lookup table.

[0022] refer to Figure 4 It also includes: Data whose current work monitoring data and PID parameters match any item in the PID parameter lookup table with a degree lower than the matching degree threshold are marked as data with no matching item. The set of parameters with the highest matching degree between the PID parameter lookup table and the data with no matching items is updated to the PID improvement controller, and the work performance evaluation result corresponding to the work monitoring data at the next time moment is extracted. When the work performance evaluation result is in a normal state, the data with no matching items and the PID parameters corresponding to the data with no matching items are updated to the PID parameter lookup table. When the work performance evaluation result is in an abnormal state, the data with no matching items is marked as a feedback control group, and the controller corresponding to the feedback control group is switched to the PID original controller. The PID original controller is used to regulate the data of the unmatched items whose matching degree with any of the feedback control groups exceeds the matching degree threshold, and the PID parameters after regulation and the working monitoring data are recorded to obtain the candidate group data; The candidate group data is processed and updated according to the method for constructing the PID parameter lookup table to obtain the updated PID parameter lookup table.

[0023] refer to Figure 5 An explosive arc pulse power supply for vacuum ion plating of diamond / DLC includes a charging module, an energy storage capacitor bank, an arc triggering module, a discharge circuit module, and a main control module. The output voltage of the charging module is adjustable from 300V to 500V, used for constant voltage charging of the energy storage capacitor bank. The energy storage capacitor bank consists of multiple high-voltage capacitors connected in parallel, providing rapid discharge capability. The arc ignition module integrates a laser arc ignition unit and a high-frequency high-voltage arc ignition unit, supporting dual-mode switching. The discharge circuit module uses low-impedance wires to build the circuit, forming a discharge channel with the graphite target and the anode inside the vacuum chamber. After triggering, the instantaneous discharge current peak reaches over 5000A. The main control module is electrically connected to the charging module, the arc ignition module, and the discharge circuit module, respectively, and is used to control charging start / stop, voltage adjustment, arc ignition mode selection, and discharge process monitoring.

[0024] Furthermore, the laser arc-initiating unit is a 1064nm Nd:YAG laser with an output power of 10W to 20W, which bombards the surface of the graphite target with a high-energy laser beam to form initial plasma. The high-frequency, high-voltage arc-initiating unit is a high-frequency inverter with an output voltage of 5kV to 8kV and a frequency of 1MHz to 2MHz, used to break down the vacuum gap between the target and the anode to form a plasma channel. The energy storage capacitor bank uses 1000μF / 600V high-voltage capacitors, with 4 to 8 capacitors connected in parallel, and an energy storage capacity of 4000μF to 8000μF. The discharge circuit module uses copper busbars for its wiring, and the circuit impedance is controlled below 50mΩ to reduce energy loss during current transmission. The main control module is a PLC controller equipped with current and voltage sensors, with a current monitoring accuracy of ±1A and a voltage monitoring accuracy of ±1V, achieving accurate feedback of discharge parameters.

[0025] Optionally, a basic power control method includes the following steps: Capacitor charging: The main control module controls the charging module to start, charging the energy storage capacitor bank with a voltage of 300V to 500V until the capacitor voltage reaches the set value. The charging current can be adjusted according to process requirements. Based on the cleanliness of the graphite target surface and the gas pressure fluctuation in the vacuum chamber, a laser arc ignition mode or a high-frequency high-voltage arc ignition mode is selected. The main control module sends a trigger signal, and the arc ignition unit acts on the graphite target surface to form initial plasma. At the same time, the discharge circuit module is turned on, and the energy storage capacitor bank discharges instantaneously, so that the peak discharge current reaches more than 5000A, exciting the graphite target to generate high-concentration carbon ions. After a single discharge is completed, the main control module controls the charging module to charge again, repeating the above steps to achieve continuous pulse explosive arc discharge.

[0026] Preferably, when the surface cleanliness of the graphite target is ≥98%, the laser arc ignition mode is used; when the pressure fluctuation amplitude in the vacuum chamber is >5%, the high-frequency high-voltage arc ignition mode is used. During the continuous pulse explosive arc discharge process, a magnetic filtration system is used in conjunction. The magnetic field strength of the magnetic filtration system is adjustable from 0.5T to 1.0T to filter out droplet impurities in the carbon plasma.

[0027] Furthermore, to achieve intelligent control and precise monitoring, this embodiment designs a complete detection module architecture, specifically including: 1) Current and voltage detection module: using high-precision Hall sensors and voltage divider circuits to detect the current and voltage signals of the discharge circuit in real time, with a sampling frequency of over 100kHz, ensuring that instantaneous parameter changes during the discharge process can be captured, providing accurate real-time data for the control algorithm; 2) Capacitor energy storage detection module: by detecting the voltage across the capacitor in real time, combined with the capacitor capacity parameters, the energy storage state of the capacitor bank is calculated in real time, providing energy state feedback for the control algorithm, ensuring that the charging process can be precisely controlled, and avoiding insufficient energy storage or overcharging; 3) Ion information detection system: specifically used to monitor the plasma state, consisting of three parts: a Faraday cup array, distributed and installed on the workpiece turntable inside the vacuum chamber, used to detect ion beam density and energy distribution; a spectral diagnostic system, using optical emission spectroscopy technology, to monitor the electron temperature and density of the plasma in real time; and a Langmuir probe, used to measure the volt-ampere characteristics of the plasma and obtain key parameters such as plasma density and electron temperature.

[0028] Preferably, this embodiment does not use a neural network for main control (to avoid the problem of slow response), but combines real-time feedback PID control, segmented feature extraction, intelligent lookup table, and historical data clustering analysis. The specific core principles are as follows: The segmented control window mechanism determines the minimum control window T based on the arc triggering frequency f, and the calculation formula is as follows: .

[0029] Where k is the window coefficient. Each control window contains multiple arc pulse cycles, and independent feature extraction and parameter optimization are performed within each window to ensure that the control can keep up with the rapid changes in the arc.

[0030] The feature extraction algorithm performs multi-dimensional feature extraction on the monitoring data within each control window to ensure a comprehensive reflection of the discharge state and plasma quality. Specifically, the extracted features include four categories: 1) Time-domain features: mainly including peak current, peak voltage, rise time, and pulse width; 2) Frequency-domain features: by performing FFT transformation on the current and voltage signals, the amplitude and phase of the fundamental component are extracted, the total harmonic distortion (THD) is calculated, and the energy distribution of each harmonic is analyzed. The formula for calculating the total harmonic distortion is as follows: .

[0031] in, Let n be the amplitude of the nth harmonic. This represents the amplitude of the fundamental component.

[0032] 3) Statistical characteristics: mainly including mean, variance, skewness and kurtosis; 4) Plasma characteristics: mainly including electron temperature, electron density and ion energy distribution, wherein electron temperature is calculated by Boltzmann slope method, electron density is calculated by Stark broadening method, and ion energy distribution is obtained by Faraday cup array measurement.

[0033] Furthermore, regarding the construction method of the intelligent lookup table, the lookup table designed in this embodiment is a multi-dimensional mapping relationship, with the core structure being: {F,U,I,P, , ,...}→{Kp,Ki,Kd}, where F is the arc triggering frequency, U is the discharge voltage, I is the discharge current, and P is the discharge power. For electron temperature, Let Kp be the electron density, and Kp, Ki, and Kd be the proportional, integral, and derivative parameters of the PID controller, respectively. The construction process of the PID parameter lookup table needs to be completed before the system officially starts operating. The core is to build a complete lookup table based on the original PID controller control data, containing both normal control groups and optimized abnormal control groups. This provides the foundation for parameter matching in subsequent PID improvement controllers, as detailed below: A PID controller is set up, which is a traditional PID controller based on feedback error, and a conventional dynamic adjustment method for PID parameters based on feedback error is adopted; a pre-designed comprehensive evaluation function is enabled. The evaluation threshold and data classification criteria are clearly defined: data in groups with an evaluation result E≥0.8 are classified as normal control groups, and data in groups with E<0.8 are classified as abnormal control groups; at the same time, a similarity threshold is set for subsequent screening of similar group data.

[0034] Using a pre-defined PID controller, the target explosion arc pulse unit is pre-operated and controlled. Relevant data is collected in real-time at each moment during the control process, including PID parameters and operational monitoring data (current, voltage signals, capacitor bank energy storage status data, and plasma parameters, collected by various monitoring devices). The collected control data is integrated into basic data sets based on the correlation between the current and previous timeframes. Each basic data set includes: the PID parameters corresponding to the current timeframe, the PID parameters corresponding to the previous timeframe, the operational monitoring data corresponding to the previous timeframe, and the operational monitoring data corresponding to the current timeframe.

[0035] Using a defined comprehensive evaluation function, the current-time work monitoring data in each basic group of data is evaluated, and the work performance evaluation result corresponding to each basic group of data is calculated to quantify the control effect and provide a basis for subsequent data classification. Based on the work performance evaluation result and the set evaluation threshold, all basic group data are divided into normal control group and abnormal control group. The normal control group is directly organized and stored in the PID parameter lookup table according to the mapping relationship of [Key=PID parameter of the previous time + work monitoring data of the previous time, Value=PID parameter of the current time], which serves as the basic data for the lookup table.

[0036] Preferably, a similarity function is used to filter basic group data from all normal control groups that have a similarity exceeding a similarity threshold with the target abnormal control group, resulting in a set of similar groups. K-means++ clustering analysis is performed on the PID parameters corresponding to the current time of all basic group data in the similar group set. Through initializing cluster centers, data allocation, center updates, and iterative optimization, optimized PID control parameters are obtained. These optimized PID control parameters are then used to replace the PID parameters corresponding to the current time in the target abnormal control group, completing the optimization of the abnormal control group. The optimized abnormal control groups are then added to the PID parameter lookup table in the same way as normal control groups, expanding the table's coverage. After insertion, all group data in the lookup table are validated, duplicate and invalid data are deleted, and the similarity calculation threshold is calibrated, completing the pre-construction of the PID parameter lookup table.

[0037] Furthermore, after the PID parameter lookup table is constructed, the system enters the formal workflow, which is as follows: When the system is in formal operation, the initial parameter settings of the PID improved controller and the calibration of each monitoring device are completed first, and the minimum control window is set according to the arc triggering frequency; using the minimum control window as a unit, the monitoring devices collect and preprocess the working monitoring data; the current working monitoring data and the PID parameters of the PID improved controller are matched with the PID parameter lookup table to calculate the matching degree and filter the parameters, obtain the optimal control parameters, and update them to the PID improved controller; the PID improved controller is used to control the target explosion arc pulse unit, and the data during the control process is collected synchronously to form historical control data; finally, based on the historical control data, the lookup table is dynamically updated using the PID parameter lookup table construction method to achieve closed-loop optimization.

[0038] Optionally, the lookup table is dynamically updated: historical control data is integrated into basic group data, and after evaluation, it is divided into normal control group and abnormal control group. The normal control group is directly updated incrementally to the lookup table, and the abnormal control group is filtered by similar groups and optimized by clustering, and then the parameters are replaced and put back into the database to ensure that the lookup table can keep up with the actual process conditions in real time and achieve self-learning and self-optimization.

[0039] Optionally, a supplementary control process (processing data without matching items) is provided: This process supplements the main control process and is used to address scenarios where the lookup table matching degree is insufficient. Data without matching items is marked, and initial control is performed using the parameter with the highest matching degree in the lookup table. The evaluation result determines whether to add the data to the database. If the evaluation is abnormal, it is marked as a feedback control group and switched to the PID original controller. The same type of operating condition without matching items is controlled, and candidate group data is collected. Finally, the candidate group data is processed according to the lookup table construction method, the lookup table is updated, and the coverage is improved. The specific operation is as follows: In the actual matching stage, if the calculated highest matching degree does not reach the preset threshold, i.e., there is no reliable matching item in the current lookup table, the PID parameter corresponding to the data with the highest matching degree is used for control. After control is completed, the evaluation function result of this control is analyzed. If the evaluation result E≥0.8 (meets the requirements), the complete set of data corresponding to this control (constructed according to the lookup table group data standard) is added to the lookup table. If the evaluation result E<0.8 (does not meet the requirements), the state (current monitoring data + current PID parameter) corresponding to this control is marked. In subsequent work, if a state matching this mark occurs... When the matching degree is highest, the parameters matched by the lookup table are no longer used. Instead, the traditional PID control method based on feedback error is used for regulation. At the same time, the complete set of data for this regulation is recorded in strict accordance with the data construction standard of the lookup table group. All regulation data corresponding to such marked states are organized into a separate data subset. This data subset consists entirely of data whose matching degree does not reach the threshold and whose initial regulation is unqualified. Subsequently, following the complete construction process of the lookup table, this data subset is analyzed, evaluated, filtered, and clustered for optimization. The optimized control parameters are then updated into the lookup table to improve the coverage of the lookup table and avoid the problem of matching failure in similar scenarios.

[0040] Optionally, the parametric clustering analysis method, in this embodiment, uses the K-means++ clustering algorithm to perform clustering analysis on the labeled data. The specific steps are as follows: Initialize cluster centers: Select data points that are far apart as initial cluster centers; Data allocation: Calculate the similarity between each data point and each cluster center (using cosine similarity), and assign each data point to the nearest cluster center to form multiple clusters; Center update: Calculate the feature mean and PID parameter mean of all data points in each cluster, and use them as new cluster centers; Iterative optimization: Repeat the above steps until the cluster centers no longer change or the maximum number of iterations is reached, then stop clustering to obtain stable clustering results. After clustering is completed, we calculate the optimal PID parameters for each cluster. The calculation formula is as follows: .

[0041] .

[0042] .

[0043] Where mean is the mean function, std is the standard deviation function, and i is each data point in the cluster. The optimal parameters calculated by this formula can take into account both the universality and adaptability within the cluster, ensuring that precise control can be achieved under similar working conditions.

[0044] The beneficial effects of this invention are as follows: (1) The present invention proposes a method for dynamically adjusting the control window according to the arc triggering frequency, which improves the control accuracy and response speed and can track the dynamic changes of the arc in a timely manner.

[0045] (2) This invention comprehensively extracts features from four dimensions: time domain, frequency domain, statistics and plasma, and establishes a complete feature vector space. In particular, it introduces plasma diagnostic technology, realizing all-round monitoring from macroscopic electrical parameters to microscopic plasma state, providing rich data support for precise control, and can more comprehensively reflect the discharge state and ion quality.

[0046] (3) This invention avoids the latency problem of simply using neural networks. It designs a composite control strategy that combines a PID improved controller with an intelligent lookup table. The original PID controller is only used for feedback control in the pre-construction of the lookup table and the abnormal scenario of no matching item. The PID improved controller, as the core control unit for formal operation, is responsible for the precise control of the target explosion arc pulse unit. The intelligent lookup table enables the rapid matching of control parameters while retaining the real-time advantage of PID control, forming a dynamic update closed-loop control, enabling the system to have self-learning ability and continuously optimize control performance through historical data accumulation, taking into account both response speed and control accuracy.

[0047] (4) A complete closed-loop system from process control to quality evaluation was established. The work effect was quantified through a comprehensive evaluation function, and the coating quality was directly linked to the control parameters. Parameter optimization and lookup table updates were performed based on the evaluation results. This mechanism ensures that the control objective always revolves around the coating quality, effectively improving the reliability of the process and the consistency of the coating quality.

[0048] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0049] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for controlling the explosive arc pulse power supply for vacuum ion plating of diamond or DLC, characterized in that, include: Set the initial parameters of the PID improved controller and calibrate the monitoring device; The monitoring device includes: a current and voltage detection module, a capacitor energy storage detection module, and an ion information detection module; Set the minimum control window based on the arc triggering frequency; Based on the minimum control window, the monitoring device is used to collect the current and voltage signals of the discharge circuit, the energy storage status data of the capacitor bank, and the plasma parameters to obtain working monitoring data. Feature extraction is then performed on the working monitoring data to obtain multi-dimensional features. The matching degree is calculated and parameters are filtered by using the multi-dimensional features and PID parameters at the current moment and the pre-constructed PID parameter lookup table to obtain the optimal control parameters, and the optimal control parameters are updated to the PID improved controller. The target explosion arc pulse unit is controlled using the aforementioned improved PID controller; Collect data on the process of the PID improved controller regulating the target explosion arc pulse unit to obtain historical regulation data; The PID parameter lookup table is dynamically updated based on the historical control data using the PID parameter lookup table construction method. The process of constructing the PID parameter lookup table includes: A PID controller is set up; the PID controller is a PID controller based on feedback error. The target explosion arc pulse unit is controlled using the PID raw controller, and the PID parameters and the working monitoring data are collected at each moment during operation. The PID parameters corresponding to the current moment, the PID parameters corresponding to the previous moment, and the work monitoring data are integrated into basic group data; The work monitoring data is evaluated using a set comprehensive evaluation function to obtain the work performance evaluation result; Based on the work performance evaluation results and the set evaluation threshold, the basic group data is divided into a normal control group and an abnormal control group, and the normal control group is updated to the PID parameter lookup table. The similarity function is used to filter the basic group data in all the normal control groups and the target abnormal control group whose similarity exceeds the similarity threshold, to obtain a set of similar groups; Cluster analysis is performed on the PID parameters corresponding to the current time in the set of similar groups to obtain optimized PID control parameters; The optimized PID control parameters are used to replace the PID parameters corresponding to the current time in the target abnormal control group, and the replaced target abnormal control group is updated in the PID parameter lookup table.

2. The explosion arc pulse power supply control method for vacuum ion plating of diamond or DLC according to claim 1, characterized in that, Also includes: Data whose current work monitoring data and PID parameters match any item in the PID parameter lookup table with a degree lower than the matching degree threshold are marked as data with no matching item. The set of parameters with the highest matching degree between the PID parameter lookup table and the data with no matching items is updated to the PID improvement controller, and the work performance evaluation result corresponding to the work monitoring data at the next time moment is extracted. When the work performance evaluation result is in a normal state, the data with no matching items and the PID parameters corresponding to the data with no matching items are updated to the PID parameter lookup table. When the work performance evaluation result is in an abnormal state, the data with no matching items is marked as a feedback control group, and the controller corresponding to the feedback control group is switched to the PID original controller. The PID original controller is used to regulate the data of the unmatched items whose matching degree with any of the feedback control groups exceeds the matching degree threshold, and the PID parameters after regulation and the working monitoring data are recorded to obtain the candidate group data; The candidate group data is processed and updated according to the method for constructing the PID parameter lookup table to obtain the updated PID parameter lookup table.

3. The explosion arc pulse power supply control method for vacuum ion plating of diamond or DLC according to claim 1, characterized in that, The expression for the minimum control window is: ;in, This refers to the minimum control window; The window coefficient; This is the arc triggering frequency.

4. The explosion arc pulse power supply control method for vacuum ion plating of diamond or DLC according to claim 1, characterized in that, The multi-dimensional features include: time-domain features, frequency-domain features, statistical features, and plasma features; the frequency-domain features include: total harmonic distortion; the statistical features include: mean, variance, skewness, and kurtosis; the plasma features include: electron temperature, electron density, and ion energy distribution.

5. The explosion arc pulse power supply control method for vacuum ion plating of diamond or DLC according to claim 1, characterized in that, The similarity function uses the cosine similarity algorithm.

6. The explosion arc pulse power supply control method for vacuum ion plating of diamond or DLC according to claim 1, characterized in that, The clustering analysis used the K-means++ clustering algorithm.

7. The explosion arc pulse power supply control method for vacuum ion plating of diamond or DLC according to claim 1, characterized in that, The comprehensive evaluation function is: ;in, The evaluation result of the work performance; , , , These are the first weight, second weight, third weight, and fourth weight, respectively. As a hardness evaluation index; For thickness evaluation; As a uniformity evaluation index; It serves as a stability evaluation index.

8. A burst arc pulse power supply control system for vacuum ion plating of diamond or DLC, used to implement the burst arc pulse power supply control method for vacuum ion plating of diamond or DLC as described in claim 1, comprising: The system comprises a charging module, an energy storage capacitor bank, an arc ignition triggering module, a discharge circuit module, and a main control module; the energy storage capacitor bank is composed of multiple high-voltage capacitors connected in parallel; the arc ignition triggering module embeds a laser arc ignition unit and a high-frequency high-voltage arc ignition unit. The charging module is connected to the energy storage capacitor bank; the energy storage capacitor bank is connected to the arc ignition module; the charging module, the energy storage capacitor bank, the arc ignition module, and the discharge circuit module are respectively connected to the main control module; the discharge circuit module, together with the graphite target material and the anode inside the target vacuum cavity, forms a discharge channel; The charging module is used to charge the energy storage capacitor bank; the main control module is used to adaptively and dynamically regulate the charging module, the energy storage capacitor bank, the arc triggering module, and the discharge circuit module.

9. A burst arc pulse power supply control system for vacuum ion plating of diamond or DLC according to claim 8, characterized in that, The charging module has an adjustment range of 300V to 500V; the high-voltage capacitor has parameters of 1000μF / 600V and a parallel range of 4 to 8 capacitors; the discharge circuit module has copper busbars as its conductors; and the discharge circuit module has a circuit impedance of less than 50mΩ.

10. A burst arc pulse power supply control system for vacuum ion plating of diamond or DLC according to claim 8, characterized in that, The laser arc-initiating unit is an Nd:YAG laser with a wavelength of 1064nm; the output power of the laser arc-initiating unit ranges from 10W to 20W; the output voltage of the high-frequency high-voltage arc-initiating unit ranges from 5kV to 8kV and the frequency ranges from 1MHz to 2MHz.

Images