Satellite disk rotating speed detection method and system of MOCVD equipment, computer equipment and storage medium

Abstract

The application provides a satellite disc rotating speed detection method and system of an MOCVD device, computer equipment and a storage medium, relates to the technical field of semiconductor detection, and extracts a weak rotating speed characteristic signal from noise accurately through time domain and frequency domain fusion and a state estimation algorithm, significantly improves the detection capability and peak stability of the rotating speed characteristic in a strong noise background, takes the fusion result as an initial value and introduces state estimation of system dynamics, can realize millisecond-level real-time updating and high-precision rotating speed analysis, and the preset system dynamics model can dynamically fuse the current observation value and the historical state, effectively suppresses random errors caused by air flow disturbance, mechanical vibration and electromagnetic interference, and the output result is smooth and stable, and the reliability is greatly enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of semiconductor detection, and in particular, to a satellite disc rotation speed detection method and system for an MOCVD device, a computer device, and a storage medium. BACKGROUND

[0002] Metal organic chemical vapor deposition (MOCVD) is a key process equipment for preparing compound semiconductor epitaxial materials such as GaN, GaAs, and InP, and is widely used in the manufacturing of optoelectronic and microelectronic devices such as light emitting diodes (LEDs), lasers (LDs), and high electron mobility transistors (HEMTs). With the development of the semiconductor industry towards higher performance and lower cost, the uniformity, consistency, and defect control of MOCVD epitaxial growth have become almost stringent requirements.

[0003] To meet the challenge of epitaxial growth of multiple wafers with uniform temperature, the planetary (or planetary gear) reaction chamber structure has become the mainstream technical solution for high-end MOCVD equipment. In this structure, a planet disc drives multiple satellite discs to revolve, and at the same time, through a precise mechanical transmission system such as gear transmission or friction transmission, these satellite discs rotate around their own axes. The core design purpose of this combination of revolution and rotation is to eliminate the inherent asymmetry of gas flow, temperature, and precursor concentration distribution in the reaction chamber through dynamic averaging effect, so as to achieve ultra-uniform epitaxial growth of atomic-level flatness within a wafer and between multiple wafers.

[0004] The rotation speed of the satellite disc is one of the key process parameters that affect the uniformity, composition, and defect density of the epitaxial layer. Currently, the rotation speed of the satellite disc is usually determined according to the design of the mechanical transmission system, and it is assumed that it maintains a fixed transmission ratio with the rotation speed of the planet disc. However, in the actual high-temperature and high-pressure reaction environment, due to thermal expansion, mechanical wear, particle jamming, and other reasons, the actual rotation speed of the satellite disc may deviate from the theoretical value, and such deviation is difficult to be detected.

[0005] Therefore, it is necessary to provide a new satellite disc rotation speed detection method and system for an MOCVD device, a computer device, and a storage medium to solve the above problems in the prior art. SUMMARY

[0006] The purpose of the present application is to provide a satellite disc rotation speed detection method and system for an MOCVD device, a computer device, and a storage medium to overcome the defect that in the conventional technology, the rotation speed of the satellite disc is usually determined according to the design of the mechanical transmission system, and it is assumed that it maintains a fixed transmission ratio with the rotation speed of the planet disc, which may cause rotation speed deviation in the high-temperature and high-pressure reaction environment, affecting the process production.

[0007] In a first aspect, the application provides a satellite disc rotation speed detection method for a MOCVD device, comprising:

[0008] acquiring a spectrum reflection signal of a wafer surface arranged in a planetary reaction cavity;

[0009] performing time domain analysis on the spectrum reflection signal to acquire a periodic pulse signal corresponding to the satellite disc in the spectrum reflection signal, and obtaining time domain rotation speed information;

[0010] performing frequency domain analysis on the spectrum reflection signal, extracting the spectrum reflection signal within a preset time, performing Fourier transform on the spectrum reflection signal, obtaining a frequency domain signal, identifying peak value information corresponding to the satellite disc in the frequency domain signal, and obtaining frequency domain rotation speed information;

[0011] fusing and analyzing the time domain rotation speed information and the frequency domain rotation speed information to obtain initial rotation speed information;

[0012] based on a preset system dynamics model, performing dynamic fusion and recursive state estimation on the initial rotation speed information to obtain real-time optimal rotation speed information.

[0013] In one embodiment, the time domain analysis on the spectrum reflection signal to acquire a periodic pulse signal corresponding to the satellite disc in the spectrum reflection signal, and obtaining time domain rotation speed information, comprises:

[0014] based on the reflectivity difference between the wafer and the satellite disc base, identifying a periodic pulse sequence from the spectrum reflection signal;

[0015] calculating the time interval between consecutive pulses as a pulse period, and calculating the time domain rotation speed information according to the mapping relationship between the pulse period and the rotation speed.

[0016] In one embodiment, the identification of the peak value information corresponding to the satellite disc in the frequency domain signal to obtain the frequency domain rotation speed information, comprises:

[0017] extracting a plurality of spectral peaks within a range defined based on a priori frequency band to form a candidate set, determining a basic spectral peak corresponding to the satellite disc from the candidate set according to the amplitude and harmonic consistency of the spectral peak, and performing interpolation based on the amplitude relationship of the basic spectral peak and its adjacent frequency points to obtain the rotation frequency; and calculating the frequency domain rotation speed information according to the mapping relationship between the rotation frequency and the rotation speed.

[0018] In one embodiment, the fusion and analysis of the time domain rotation speed information and the frequency domain rotation speed information to obtain the initial rotation speed information, comprises:

[0019] calculating the relative deviation of the time domain rotation speed information and the frequency domain rotation speed information;

[0020] in case that the relative deviation is within a preset threshold range, performing weighted summation on the time-domain rotation speed information and the frequency-domain rotation speed information based on preset time-domain factors and frequency-domain factors to obtain the initial rotation speed information;

[0021] in case that the relative deviation exceeds the preset threshold range, correcting the time-domain rotation speed information based on the frequency-domain rotation speed information;

[0022] if the correction is successful, performing weighted summation on the corrected time-domain rotation speed information and the frequency-domain rotation speed information based on preset time-domain factors and frequency-domain factors to obtain the initial rotation speed information; wherein the correction being successful refers to that the relative deviation of the corrected time-domain rotation speed information and the frequency-domain rotation speed information is within the preset threshold range.

[0023] In one of the embodiments, the correcting the time-domain rotation speed information based on the frequency-domain rotation speed information comprises:

[0024] calculating a theoretical pulse period according to the frequency-domain rotation speed information;

[0025] determining a preset range of a time-domain search window with the theoretical pulse period as the center;

[0026] re-performing pulse identification and period estimation in the time-domain search window, and eliminating candidate pulses falling outside the time-domain search window to obtain the corrected time-domain rotation speed information.

[0027] In one of the embodiments, based on a preset system dynamics model, the initial rotation speed information is dynamically fused and recursively updated to obtain optimal real-time rotation speed information, comprising:

[0028] adopting the preset system dynamics model and taking the initial rotation speed information as an observation to predict the self-rotation state of the satellite disc to form a prior state and an uncertainty thereof;

[0029] and adopting a Kalman filtering algorithm to observe and update the initial rotation speed information, and dynamically weighting and fusing the prior state and the observation according to their respective uncertainties to suppress process noise and measurement noise, and outputting smoothed real-time optimal rotation speed information.

[0030] In one of the embodiments, the method further comprises:

[0031] acquiring historical running data; wherein the historical running data comprises satellite disc rotation speed estimation values, planet tray rotation speeds and process parameters;

[0032] at intervals of a preset time, learning mapping relationships of the parameters in the historical running data by using a machine learning algorithm;

[0033] updating model parameters of the system dynamics model and the Kalman filtering algorithm based on the mapping relationship;

[0034] and applying the updated system dynamics model and the Kalman filtering algorithm to the next speed state estimation.

[0035] In one of the embodiments, the method further comprises:

[0036] preprocessing the acquired spectral reflection signal; wherein the preprocessing step comprises filtering and smoothing the spectral reflection signal, and the filtering comprises filtering out high-frequency noise and / or filtering out very low-frequency drift.

[0037] In one of the embodiments, the method further comprises:

[0038] displaying real-time optimal speed information of the satellite disc and a spectrum diagram, and generating an alarm signal when the real-time optimal speed information exceeds the process setting range.

[0039] In a second aspect, the application provides a satellite disc speed detection system of an MOCVD device, which comprises:

[0040] an acquisition module configured to acquire a spectral reflection signal of a wafer surface arranged in a planetary reaction cavity;

[0041] an analysis module configured to perform time domain analysis on the spectral reflection signal, acquire a periodic pulse signal corresponding to the satellite disc in the spectral reflection signal, and obtain time domain speed information; perform frequency domain analysis on the spectral reflection signal, extract the spectral reflection signal within a preset time, perform Fourier transform on the extracted spectral reflection signal, obtain a frequency domain signal, identify peak value information corresponding to the satellite disc in the frequency domain signal, and obtain frequency domain speed information; and fuse and analyze the time domain speed information and the frequency domain speed information to obtain speed initial information.

[0042] a calculation module configured to perform dynamic fusion and recursive update of state estimation on the speed initial information based on a preset system dynamics model, and obtain real-time optimal speed information.

[0043] In a third aspect, the application further provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method steps in the first aspect when executing the computer program.

[0044] In a fourth aspect, the application further provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method steps in the first aspect.

[0045] The satellite disc rotation speed detection method, system, computer device and storage medium of the MOCVD equipment have at least the following advantages.

[0046] The application obtains the spectral reflection signal of the wafer surface arranged in the planetary reaction cavity, respectively performs time domain analysis and frequency domain analysis on the spectral reflection signal, obtains time domain rotation speed information and frequency domain rotation speed information, fuses the time domain rotation speed information and the frequency domain rotation speed information to obtain initial rotation speed information, and then performs state estimation on the initial rotation speed information based on a preset system dynamics model to obtain real-time optimal rotation speed information. By using the above scheme, the application can accurately extract a weak rotation speed characteristic signal from noise by time domain and frequency domain fusion and state estimation algorithm, significantly improves the detection capability and peak stability of the rotation speed characteristic in a strong noise background, simultaneously uses the fusion result as an initial value and introduces state estimation of the system dynamics to realize millisecond-level real-time updating and high-precision rotation speed analysis, and further, the preset system dynamics model can dynamically fuse the current observation value and the historical state, effectively suppresses random errors caused by airflow disturbance, mechanical vibration and electromagnetic interference, and outputs smooth and stable results, thereby greatly enhancing the reliability. BRIEF DESCRIPTION OF DRAWINGS

[0047] Figure 1 An application environment diagram of the satellite disc rotation speed detection method of the MOCVD equipment in one embodiment is shown in the figure;

[0048] Figure 2 A flowchart of the satellite disc rotation speed detection method of the MOCVD equipment in one embodiment is shown in the figure;

[0049] Figure 3 A flowchart of the time domain analysis step in one embodiment is shown in the figure;

[0050] Figure 4 A flowchart of the step of obtaining initial rotation speed information by fusion analysis in one embodiment is shown in the figure;

[0051] Figure 5 A structural block diagram of the satellite disc rotation speed detection system of the MOCVD equipment in one embodiment is shown in the figure;

[0052] Figure 6 An internal structure diagram of the computer device in one embodiment is shown in the figure. DETAILED DESCRIPTION

[0053] The above objects, advantages and other features of the present application are explained in more detail in connection with specific embodiments of the present application. Other advantages and pertinently-meritorious modifications of the present application will readily occur to those skilled in the art. The application is not limited to the embodiments disclosed herein, but is only limited by the scope of the claims. The embodiments and features described hereinbelow can be combined with each other, if not incompatible, to provide other embodiments of the present application.

[0054] Some example embodiments of the present application are described for illustrative purposes, and it is to be understood that the present application can be implemented in other ways not specifically shown in the drawings.

[0055] The satellite disc rotation speed detection method of the MOCVD equipment provided by the embodiments of the present application can be applied to the application environment as shown in the accompanying drawings. Figure 1 As shown in the accompanying drawings, the planetary reaction cavity 1 is provided with a rotatable planetary disc, the planetary disc is provided with a plurality of self-rotating satellite discs, each satellite disc is provided with a plurality of wafer supporting discs for placing wafers, the planetary disc is controlled to rotate by a motor, and each satellite disc is self-rotated by air flow.

[0056] The upper portion of the planetary reaction cavity 1 is provided with a signal source and acquisition module 2 and a satellite disc rotation speed detection system 3 of the MOCVD equipment.

[0057] The signal source and acquisition module 2 includes an in-situ monitoring light path unit, a photoelectric detector and a multi-channel synchronous data acquisition card.

[0058] The in-situ monitoring light path unit is a standard configuration of the MOCVD equipment, and generally includes an emission end for emitting a monitoring light and a detection end for receiving reflected light from the wafer and the surface of the satellite disc, for real-time monitoring of process parameters such as film thickness, temperature or reflectivity.

[0059] The photoelectric detector is used to receive the reflected light signal and the radiation signal of the in-situ monitoring light path unit, and convert them into an analog electrical signal proportional to the light intensity. In the embodiment, the analog electrical signal is a spectral reflection signal of the wafer surface provided in the planetary reaction cavity.

[0060] The multi-channel synchronous data acquisition card is used to synchronously collect and digitize two signals at a high frequency: one is the analog electrical signal from the photoelectric detector, and the other is a reference trigger pulse signal provided by the MOCVD equipment, which represents one rotation of the planetary disc.

[0061] The satellite disc rotation speed detection system 3 of the MOCVD device is connected with the multi-channel synchronous data acquisition card, and is used for acquiring the spectral reflection signal of the wafer surface arranged in the planetary reaction cavity; the spectral reflection signal is analyzed in time domain to acquire the periodic pulse signal corresponding to the satellite disc in the spectral reflection signal, and time domain rotation speed information is obtained; the spectral reflection signal is analyzed in frequency domain to extract the spectral reflection signal within a preset time, and Fourier transform is performed on the spectral reflection signal to obtain a frequency domain signal, peak information corresponding to the satellite disc in the frequency domain signal is identified, and frequency domain rotation speed information is obtained; the time domain rotation speed information and the frequency domain rotation speed information are fused to obtain rotation speed initial information; based on a preset system dynamics model, the rotation speed initial information is dynamically fused and recursively updated state estimation to obtain real-time optimal rotation speed information.

[0062] The satellite disc rotation speed detection method of the MOCVD device acquires the spectral reflection signal of the wafer surface arranged in the planetary reaction cavity, analyzes the spectral reflection signal in time domain and frequency domain respectively, obtains time domain rotation speed information and frequency domain rotation speed information, fuses the time domain rotation speed information and the frequency domain rotation speed information to obtain rotation speed initial information, and then performs state estimation on the rotation speed initial information based on a preset system dynamics model to obtain real-time optimal rotation speed information. By using the above scheme, the weak rotation speed characteristic signal can be accurately extracted from noise by the time domain and frequency domain fusion and the state estimation algorithm, and the detection capability and peak stability of the rotation speed characteristic in a strong noise background are significantly improved; meanwhile, the fusion result is used as an initial value and the state estimation of the system dynamics is introduced, millisecond-level real-time updating and high-precision rotation speed analysis can be realized; further, the preset system dynamics model can dynamically fuse the current observation value and the historical state, effectively suppresses the random error caused by airflow disturbance, mechanical vibration and electromagnetic interference, and the output result is smooth and stable, and the reliability is greatly enhanced.

[0063] Further, the rotation speed of the satellite disc is a crucial process parameter affecting the quality of the epitaxial layer. The stability and accuracy thereof directly affect the growth rate and thickness uniformity, alloy component uniformity and defect density. For example, the rotation speed affects the thickness of the substrate surface stagnation boundary layer and the transport efficiency of the reaction precursor, and is a key to controlling the epitaxial layer growth rate and thickness uniformity; for a multi-element compound (such as InGaN, AlGaAs), the rotation speed determines the uniformity of the alloy component in the radial distribution of the substrate by affecting the doping efficiency of the temperature-sensitive precursor (such as TMIn); a certain rotation speed helps to timely remove reaction by-products, suppresses gas phase pre-reaction, and reduces the density of crystal defects (such as dislocations and point defects).

[0064] However, in the existing MOCVD equipment, the rotation of the satellite disk completely depends on the open-loop mechanical transmission system. Its design is based on an ideal premise: the rotation speed of the satellite disk and the revolution speed of the main disk maintain a fixed transmission ratio. In actual industrial production, this premise is often difficult to establish. The MOCVD reaction chamber is an extreme environment that is continuously at high temperature (usually higher than 1000℃), normal or low pressure, and has chemical corrosion by-products. In this environment, there are many factors that cause the actual rotation speed of the satellite disk to deviate from the theoretical value, mainly including: thermal deformation, non-uniform thermal expansion of the internal parts of the chamber will change the gear meshing gap or the friction wheel contact pressure, causing the transmission efficiency to drift; mechanical wear, long-term continuous operation causes the key transmission parts such as gears and bearings to wear, introduces transmission errors, and this process is slow and irreversible; particle pollution and jamming, small particles generated during the reaction process settle and invade the transmission gap, which may cause instantaneous jamming or permanent obstruction, causing the satellite disk to suddenly drop or even stop rotating.

[0065] The more urgent problem to be solved is that due to the current technical implementation limitations, the existing semiconductor MOCVD equipment generally lacks effective monitoring means for the real-time rotation speed of each satellite disk during the process. Related engineers cannot know the real running state of the satellite disk during operation, and the abnormal rotation speed can usually only be discovered after the process is completed by detecting the scrapped epitaxial wafer. This mode brings huge production risks, including but not limited to: the entire batch of epitaxial wafers is scrapped, causing the loss of expensive raw materials (such as substrates, MO sources, and high-purity gases) and huge energy consumption; batch-to-batch product quality fluctuation, unstable yield, and serious constraints on the large-scale and consistent manufacturing of high-end semiconductor devices; equipment maintenance lag, unable to provide early warning for potential mechanical failures, increasing the risk of unplanned downtime.

[0066] Therefore, there is an urgent need in the art for a technical solution that can accurately monitor the rotation speed of the satellite disk of the MOCVD equipment in real time and effectively control it on this basis, in order to overcome the inherent defects of the existing open-loop mechanical transmission system and ensure the repeatability of the process and the yield of the product.

[0067] Please refer to Figure 2 In an exemplary embodiment, the embodiment of the present application provides a satellite disk rotation speed detection method of a MOCVD equipment, specifically comprising the following steps:

[0068] Step 202, acquiring a spectrum reflection signal of a wafer surface arranged in a planetary reaction chamber.

[0069] Specifically, the spectral reflectance signal refers to the light intensity signal collected by the photodetector after the monitoring light emitted by the in-situ monitoring light path unit irradiates the wafer surface, and the light intensity signal changes with wavelength and / or time, and is normalized by the reference channel to represent the reflectance curve changing with wavelength and time.

[0070] Optionally, the embodiments of the present application also preprocess the acquired spectral reflectance signal; wherein the preprocessing step includes filtering and smoothing the spectral reflectance signal, and the filtering includes filtering out high-frequency noise and / or filtering out very low-frequency drift.

[0071] Step 204, time domain analysis is performed on the spectral reflectance signal to obtain the periodic pulse signal corresponding to the satellite disk in the spectral reflectance signal, and time domain speed information is obtained.

[0072] Specifically, the spectral reflectance signal usually includes periodic high-frequency pulses caused by the rotation of each satellite disk, and periodic low-frequency pulse signals caused by the revolution of the planet disk, and the frequencies and shapes of the two are different.

[0073] The time domain analysis refers to the time domain processing of the spectral reflectance signal changing with time, and the time domain speed information of the satellite disk rotation is obtained by identifying the periodic high-frequency pulse sequence formed by the satellite disk rotating under the action of the satellite disk rotating through the monitoring point alternately between the wafer with high reflectivity and the satellite disk with low reflectivity, and calculating the pulse interval. The periodic low-frequency pulse signal caused by the revolution of the planet disk is used for gating and checking, and does not participate in the pulse interval calculation.

[0074] Step 206, frequency domain analysis is performed on the spectral reflectance signal, the spectral reflectance signal within a preset time is extracted, and Fourier transform is performed thereon to obtain a frequency domain signal, peak information corresponding to the satellite disk in the frequency domain signal is identified, and frequency domain speed information is obtained.

[0075] Specifically, the frequency domain analysis refers to window function processing and fast Fourier transform (FFT) on the spectral reflectance signal collected within a preset time, which is converted from the time domain to the frequency domain, the frequency information related to the speed is obtained by identifying the characteristic peak corresponding to the satellite disk in the frequency spectrum, and the frequency domain speed information is obtained by conversion.

[0076] Step 208, the time domain speed information and the frequency domain speed information are fused to obtain initial speed information.

[0077] Specifically, the purpose of fusion is to obtain higher precision, continuous and stable speed observation under the conditions of noise and shielding, to provide high credible input for subsequent state estimation, and to significantly improve the real-time performance and reliability of detection.

[0078] At step 210, the initial rotation speed information is dynamically fused and recursively updated state estimation based on a preset system dynamics model to obtain real-time optimal rotation speed information.

[0079] Specifically, the system dynamics model is a mathematical model used to describe the real rotation speed state of the satellite disc evolving over time. Generally, the rotation speed state of the satellite disc includes angular velocity and rotation speed. In the embodiment, the initial rotation speed information is taken as the observation of the system dynamics model, and the rotation speed state of the satellite disc is taken as the state quantity of the system dynamics model. Therefore, the system dynamics model is used to describe the corresponding relationship between the observation and the state quantity.

[0080] Further, in the embodiment, state prediction is performed based on the preset system dynamics model, and a filter is used to observe and update the initial rotation speed information and dynamically fuse the initial rotation speed information to obtain real-time optimal rotation speed estimation. The filter includes Kalman filtering and / or extended Kalman filtering.

[0081] The satellite disc rotation speed detection method of the MOCVD device described above obtains the spectral reflection signal of the wafer surface arranged in the planetary reaction cavity, respectively performs time domain analysis and frequency domain analysis on the spectral reflection signal to obtain time domain rotation speed information and frequency domain rotation speed information, fuses the time domain rotation speed information and the frequency domain rotation speed information to obtain initial rotation speed information, and then performs state estimation on the initial rotation speed information based on a preset system dynamics model to obtain real-time optimal rotation speed information. By using the above scheme, the weak rotation speed characteristic signal can be accurately extracted from noise by the time domain and frequency domain fusion and the state estimation algorithm, and the detection capability and peak stability of the rotation speed characteristic in a strong noise background are significantly improved. Meanwhile, the fusion result is taken as the initial value and introduced into the state estimation of the system dynamics to realize millisecond-level real-time updating and high-precision rotation speed analysis. Further, the preset system dynamics model can dynamically fuse the current observation value and the historical state to effectively suppress the random error caused by air flow disturbance, mechanical vibration and electromagnetic interference, and the output result is smooth and stable, which greatly enhances the reliability. Further, the frequency domain analysis is not sensitive to signal amplitude fluctuation, so that the system has stronger tolerance to common engineering problems such as light path window pollution and light source aging, and the stability of long-term measurement is ensured.

[0082] Please refer to Figure 3 Optionally, the time domain analysis is performed on the spectral reflection signal to obtain the periodic pulse signal corresponding to the satellite disc in the spectral reflection signal to obtain the time domain rotation speed information, including:

[0083] At step 302, the periodic pulse sequence is identified from the spectral reflection signal based on the reflectivity difference between the wafer and the satellite disc substrate.

[0084] In step 304, the time interval between the continuous pulses is calculated as a pulse period, and the time-domain rotation speed information is calculated according to a mapping relationship between the pulse period and the rotation speed.

[0085] Specifically, the reflectivity of the wafer area is relatively high, and the reflectivity of the satellite disc area is relatively low. The two alternately appear at the monitoring point with the rotation of the satellite disc, so that the waveform of the spectral reflection signal presents periodic high-low changes. Based on the above high-low changes and equal intervals, the pulse sequence generated by the wafer and the satellite disc edge transition can be identified by a peak detection algorithm.

[0086] Further, the time interval between the continuous pulses is calculated, and the average value thereof is taken as a pulse period. Then, the time-domain rotation speed information is calculated according to a mapping relationship. The expression of the mapping relationship is RPM = 60 / T, where T is the pulse period, and RPM is the rotation speed of the satellite disc. If the time-domain rotation speed information also includes an angular velocity, the expression of the mapping relationship is angular velocity = 2π / T.

[0087] By using the above scheme, since the pulse signal is generated by the physical contrast between the high-reflectivity area and the low-reflectivity area, the pulse signal has natural high signal-to-noise ratio and strong discriminability, and can effectively suppress slow-changing interference such as light source drift and window pollution. By performing robust statistics on the pulse interval, the output time-domain rotation speed information has small jitter and high accuracy.

[0088] Optionally, peak information corresponding to the satellite disc in the frequency-domain signal is identified to obtain frequency-domain rotation speed information, including:

[0089] A plurality of spectral peaks are extracted in a range defined based on a priori frequency band to form a candidate set. A basic spectral peak corresponding to the satellite disc is determined from the candidate set according to the amplitude and harmonic consistency of the spectral peak. The amplitude relationship of the basic spectral peak and adjacent frequency points is interpolated to obtain the rotation frequency. The frequency-domain rotation speed information is calculated according to a mapping relationship between the rotation frequency and the rotation speed.

[0090] Specifically, according to process setting or estimation at the last moment, a reasonable search interval is set, and a plurality of spectral peaks with an amplitude greater than a preset threshold are extracted in the search interval to form a candidate set. A spectral peak that satisfies both amplitude and harmonic consistency conditions is selected from the candidate set as a basic spectral peak, which is the most significant peak. The two conditions are met at the same time, that is, the spectral peak with the highest signal-to-noise ratio and the corresponding filter near the integer multiple of the spectral peak exist. Since the frequency-domain signal is a discrete frequency point, the true peak usually falls between two points. The amplitude asymmetry of the basic spectral peak and the left and right adjacent frequency points is used for interpolation, which can greatly improve the frequency resolution and obtain a more accurate rotation frequency.

[0091] Further, the frequency domain rotation speed information is calculated according to a mapping relationship between the rotation frequency and the rotation speed, wherein the expression of the mapping relationship is RPM_freq = 60 x f_freq, and RPM_freq is the frequency domain rotation speed and f_freq is the rotation frequency.

[0092] By using the above scheme, the prior band constraint limits the search within a reasonable range, and the amplitude and harmonic consistency screening of the basic spectral peak can effectively distinguish the rotation base peak from the harmonic, sideband and environmental pseudo-peak, thereby reducing the false detection rate. Meanwhile, the interpolation is performed on the basic spectral peak and its adjacent frequency points, which can improve the accuracy of the frequency resolution and significantly reduce the rotation speed calculation deviation.

[0093] Please refer to Figure 4 Optionally, the time domain rotation speed information and the frequency domain rotation speed information are fused to obtain the rotation speed initial information, including:

[0094] In step 402, the relative deviation between the time domain rotation speed information and the frequency domain rotation speed information is calculated.

[0095] In step 404, when the relative deviation is within a preset threshold range, the time domain rotation speed information and the frequency domain rotation speed information are weighted and summed based on preset time domain factors and frequency domain factors to obtain the rotation speed initial information.

[0096] In step 406, when the relative deviation is out of the preset threshold range, the time domain rotation speed information is corrected based on the frequency domain rotation speed information.

[0097] In step 408, if the correction is successful, the corrected time domain rotation speed information and the frequency domain rotation speed information are weighted and summed based on the preset time domain factors and the frequency domain factors to obtain the rotation speed initial information. The correction success means that the relative deviation between the corrected time domain rotation speed information and the frequency domain rotation speed information is within the preset threshold range.

[0098] Specifically, the rotation speed initial information is the consistency verification and adaptive weighted fusion of the time domain rotation speed information (RPM_time) and the frequency domain rotation speed information (RPM_freq), and the output is the optimal joint analysis rotation speed value.

[0099] Further, the expression of the relative deviation is δ = |RPM_time - RPM_freq| / RPM_freq.

[0100] When the relative deviation is within the preset threshold range, it is considered that the two results confirm each other, the verification is passed, and the adaptive weighted fusion process is entered, that is, the weighted fusion algorithm based on the confidence is performed on the two to output the optimal joint analysis value, and the rotation speed initial information is obtained.

[0101] Exemplarily, the calculation formula of the weighted fusion is represented as: RPM_joint = a x RPM_time + b x RPM_freq, wherein a and b are time domain factor and frequency domain factor respectively, and satisfy a + b = 1.

[0102] Further, the determination of the above-mentioned time domain factor and frequency domain factor is jointly determined by one or more confidence indicators evaluated in real time. The time domain confidence indicator, the higher the signal-to-noise ratio (SNR) of the pre-processed reflectivity signal, the better the waveform regularity of the identified pulse sequence, and the a value increases accordingly. The frequency domain confidence indicator, the higher the peak noise ratio of the spectrum peak corresponding to the satellite disc rotation frequency, the higher the spectrum peak sharpness, and the better the harmonic consistency with the revolution frequency, and the b value increases accordingly. The system state indicator, at very high or very low speed, or poor system motion stability, that is, the estimated speed variance is large, automatically reduces a and increases b, and relies more on the frequency domain analysis with strong anti-noise performance.

[0103] Optionally, in another embodiment, since the frequency domain analysis has stronger anti-random noise capability, the frequency domain speed information obtained by the frequency domain analysis can also be preferentially selected as the initial speed information, and at this time, the time domain speed information can be used as a credibility reference.

[0104] Further, in the case where the relative deviation exceeds the preset threshold range, the error correction mode is started. Since the frequency domain analysis has stronger anti-random noise capability, the time domain analysis is corrected based on the frequency domain speed information. After the correction is successful, the above-mentioned adaptive weighted fusion process is executed based on the frequency domain speed information and the corrected time domain speed information. If the correction is still unsuccessful, the time domain speed information at this time can be directly output and a warning is triggered, indicating that the time domain channel may have a systematic failure.

[0105] Optionally, the above-mentioned correction of the time domain speed information based on the frequency domain speed information comprises:

[0106] The theoretical pulse period is calculated according to the frequency domain speed information; a preset range of time domain search window is determined with the theoretical pulse period as the center; pulse recognition and period estimation are re-performed in the time domain search window, and candidate pulses falling outside the time domain search window are eliminated to obtain the corrected time domain speed information.

[0107] Specifically, according to the frequency domain speed information (RPM_freq) obtained by the frequency domain analysis, a more accurate pulse period theoretical value can be calculated, and the expression is: T_theoretical = 60 / RPM_freq. A narrower and dynamic time search window is defined with the pulse period theoretical value as the center, for example, the window length is T_theoretical ± 10%. In this narrow window constrained by the frequency domain result, the pulse is searched and identified again. If a stable pulse sequence is found in this narrow window, the corrected time domain speed information is calculated using the new pulse period. The setting of the above-mentioned narrow window constrained by the frequency domain result can effectively avoid the false detection and missed detection of false pulses caused by noise.

[0108] With the above scheme, in the case where the relative deviation exceeds the preset threshold range, the theoretical period is recalculated based on the frequency domain result, and the time domain pulse is rechecked in the narrow time window, which can effectively eliminate false pulses and reduce the deviation and jump caused by window counting. At the same time, the confidence weighted output joint value combines the low delay advantage of the time domain and the high resolution advantage of the frequency domain, and also significantly reduces the speed estimation variance and jitter. Further, the frequency domain analysis is not sensitive to signal amplitude fluctuations, and the correction of the time domain analysis based on the frequency domain analysis can make the system have stronger tolerance to common engineering problems such as light path window pollution and light source aging, and ensure the stability of long-term measurement.

[0109] Optionally, based on a preset system dynamics model, the initial speed information is dynamically fused and recursively updated to obtain optimal real-time speed information, including:

[0110] The preset system dynamics model is used to predict the self-rotation state of the satellite disc with the initial speed information as the observation, to form the prior state and its uncertainty; and the Kalman filtering algorithm is used to observe and update the initial speed information, to dynamically weight and fuse the prior state and the observation according to their respective uncertainties, to suppress the process noise and the measurement noise, and to output the smoothed real-time optimal speed information.

[0111] Specifically, the system dynamics model is used to predict the self-rotation state of the satellite disc from the previous time to the next time to obtain the current prior state and its covariance, which represents the uncertainty of the prior prediction. The greater the value is, the lower the expected accuracy of the prior prediction is. Further, the Kalman filtering algorithm is also used to construct innovation and its variance based on the observation to calculate the Kalman gain, and the observation is optimally weighted according to the uncertainty to inject into the prior state according to the Kalman gain, to obtain the posterior state and the posterior covariance, so as to obtain the real-time optimal estimation in the sense of minimum variance.

[0112] With the above scheme, the time domain and frequency domain fusion and the state estimation algorithm can accurately extract the weak rotation speed characteristic signal from the noise, and significantly improve the detection capability and peak stability of the rotation speed characteristic in a strong noise background; taking the fusion result as the initial value and introducing the state estimation of the system dynamics can realize millisecond-level real-time updating and high-precision rotation speed analysis. The dynamic fusion mechanism of the system dynamics model and the Kalman recursive estimation can output smooth real-time rotation speed in the sense of minimum variance, significantly reduce the fluctuations caused by random noise and intermittent false detection, and effectively suppress the random errors caused by airflow disturbance, mechanical vibration and electromagnetic interference; when the observation quality decreases or is lost for a short time, the model prediction is relied on to maintain continuous and uninterrupted output, and has strong robustness; the recursive update is performed according to the sampling rhythm, without long window accumulation, and meets the real-time requirement of online control.

[0113] Optionally, the satellite disc rotation speed detection method of the MOCVD equipment further includes:

[0114] The historical operation data is acquired, and the historical operation data includes the satellite disc rotation speed estimation value, the planetary tray rotation speed and the process parameters; the mapping relationship between the parameters in the historical operation data is learned by using a machine learning algorithm at intervals of a preset time; the model parameters of the system dynamics model and the Kalman filtering algorithm are updated based on the mapping relationship; and the updated system dynamics model and the Kalman filtering algorithm are applied to the next rotation speed state estimation.

[0115] Specifically, the historical operation data is continuously acquired, and a model training process is started regularly or triggered, a linear regression (Linear Regression) or a time series analysis (Time Series Analysis) machine learning algorithm is used to learn the internal mapping relationship between the parameters, and a regression model of the satellite disc rotation speed and the gas flow of the process parameters is obtained.

[0116] After the new model obtained by training is verified, the model parameters of the system dynamics model and the Kalman filtering algorithm are updated in real time. The next state estimation is performed by using the updated model, so that the online automatic compensation and calibration of the slow-varying system errors such as mechanical wear and thermal expansion are realized, and a closed-loop system for continuous self-optimization is formed.

[0117] With the above scheme, the system dynamics model and the Kalman filtering parameters are updated periodically based on the historical data, so that the rotation speed estimation can always maintain high precision and robustness in long-term operation and multiple working conditions.

[0118] Optionally, the satellite disc rotation speed detection method of the MOCVD equipment further includes:

[0119] The real-time optimal rotation speed information and the spectrum diagram of the satellite disc are displayed, and an alarm signal is generated when the real-time optimal rotation speed information exceeds the process setting range.

[0120] Specifically, the application also provides real-time optimal rotation speed information, spectrum diagrams and other analysis results for users through a human-machine interface, so as to facilitate users to quickly identify abnormalities, trigger rectification or interlocking in advance, and shorten the processing delay.

[0121] The satellite disc rotation speed detection method of the MOCVD device obtains the spectrum reflection signal of the wafer surface arranged in the planetary reaction cavity, respectively performs time domain analysis and frequency domain analysis on the spectrum reflection signal, obtains time domain rotation speed information and frequency domain rotation speed information, fuses the time domain rotation speed information and the frequency domain rotation speed information to obtain rotation speed initial information, and then performs state estimation on the rotation speed initial information based on a preset system dynamics model to obtain real-time optimal rotation speed information. By using the above scheme, the application can accurately extract a weak rotation speed characteristic signal from noise by time domain and frequency domain fusion and state estimation algorithm, significantly improves the detection capability and peak stability of the rotation speed characteristic in a strong noise background, simultaneously takes the fusion result as an initial value and introduces state estimation of the system dynamics to realize millisecond-level real-time updating and high-precision rotation speed analysis, further, the preset system dynamics model can dynamically fuse the current observation value and the historical state, effectively suppresses random errors caused by air flow disturbance, mechanical vibration and electromagnetic interference, and the output result is smooth and stable, and the reliability is greatly enhanced, further, the frequency domain analysis is not sensitive to signal amplitude fluctuation, so that the system has stronger tolerance to common engineering problems such as light path window pollution and light source aging, and the stability of long-term measurement is ensured.

[0122] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and these steps can be executed in other sequences. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0123] Based on the same inventive concept, the embodiment of the present application also provides a satellite disc rotation speed detection system of a MOCVD device, which is suitable for the satellite disc rotation speed detection method of the MOCVD device. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme described in the above method. Therefore, the specific limitations in one or more device embodiments provided below can refer to the limitations of the method described above, and will not be described here.

[0124] Please refer to Figure 5 In one embodiment, the satellite disc rotation speed detection system of the MOCVD device comprises an acquisition module, an analysis module and a calculation module.

[0125] The acquisition module is configured to acquire a spectrum reflection signal of a wafer surface arranged in a planetary reaction cavity.

[0126] The analysis module is configured to perform time domain analysis on the spectrum reflection signal to acquire a periodic pulse signal corresponding to the satellite disc in the spectrum reflection signal, and obtain time domain rotation speed information; perform frequency domain analysis on the spectrum reflection signal to extract the spectrum reflection signal within a preset time, perform Fourier transform on the spectrum reflection signal, obtain a frequency domain signal, identify peak value information corresponding to the satellite disc in the frequency domain signal, and obtain frequency domain rotation speed information; and fuse and analyze the time domain rotation speed information and the frequency domain rotation speed information to obtain initial rotation speed information.

[0127] The calculation module is configured to perform dynamic fusion and recursive state estimation on the initial rotation speed information based on a preset system dynamics model, and obtain real-time optimal rotation speed information.

[0128] Optionally, the analysis module performs time domain analysis on the spectrum reflection signal to acquire a periodic pulse signal corresponding to the satellite disc in the spectrum reflection signal, and obtain time domain rotation speed information, including: identifying a periodic pulse sequence from the spectrum reflection signal based on reflectivity difference between the wafer and the satellite disc base; calculating a time interval between consecutive pulses as a pulse period, and calculating the time domain rotation speed information according to a mapping relationship between the pulse period and the rotation speed.

[0129] Optionally, the analysis module identifies peak value information corresponding to the satellite disc in the frequency domain signal to obtain frequency domain rotation speed information, including:

[0130] extracting a plurality of spectral peaks within a range defined based on a priori frequency band to form a candidate set, determining a basic spectral peak corresponding to the satellite disc from the candidate set according to amplitude and harmonic consistency of the spectral peak, performing interpolation based on amplitude relationship of the basic spectral peak and adjacent frequency points to obtain a rotation frequency, and calculating the frequency domain rotation speed information according to a mapping relationship between the rotation frequency and the rotation speed.

[0131] Optionally, the analysis module fuses the time-domain rotation speed information and the frequency-domain rotation speed information to obtain rotation speed initial information, including: calculating a relative deviation of the time-domain rotation speed information and the frequency-domain rotation speed information; in a case where the relative deviation is within a preset threshold range, performing weighted summation on the time-domain rotation speed information and the frequency-domain rotation speed information based on preset time-domain factors and frequency-domain factors to obtain the rotation speed initial information; in a case where the relative deviation is out of the preset threshold range, correcting the time-domain rotation speed information based on the frequency-domain rotation speed information; if the correction is successful, performing weighted summation on the corrected time-domain rotation speed information and the frequency-domain rotation speed information based on the preset time-domain factors and the frequency-domain factors to obtain the rotation speed initial information; wherein the correction success refers to that the relative deviation of the corrected time-domain rotation speed information and the frequency-domain rotation speed information is within the preset threshold range.

[0132] Optionally, the analysis module corrects the time-domain rotation speed information based on the frequency-domain rotation speed information, including: calculating a theoretical pulse period according to the frequency-domain rotation speed information; determining a preset range of a time-domain search window with the theoretical pulse period as the center; re-performing pulse identification and period estimation in the time-domain search window, and eliminating candidate pulses falling outside the time-domain search window to obtain the corrected time-domain rotation speed information.

[0133] Optionally, the calculation module performs dynamic fusion and recursive update of state estimation on the rotation speed initial information based on a preset system dynamics model to obtain optimal real-time rotation speed information, including:

[0134] The preset system dynamics model is used to predict the self-rotation state of the satellite disc with the rotation speed initial information as the observation quantity, to form a prior state and an uncertainty thereof; and a Kalman filtering algorithm is used to perform observation update on the rotation speed initial information, to dynamically weight and fuse the prior state and the observation quantity according to their respective uncertainties, to suppress process noise and measurement noise, and to output smoothed real-time optimal rotation speed information.

[0135] Optionally, the satellite disc rotation speed detection system of the MOCVD equipment further includes a model updating module.

[0136] The model updating module is configured to acquire historical running data, wherein the historical running data includes satellite disc rotation speed estimation values, planetary tray rotation speeds and process parameters; learn mapping relationships of the parameters in the historical running data at intervals of a preset time by using a machine learning algorithm; update model parameters of the system dynamics model and the Kalman filtering algorithm based on the mapping relationships; and apply the updated system dynamics model and the Kalman filtering algorithm to the next rotation speed state estimation.

[0137] Optionally, the satellite disc rotation speed detection system of the MOCVD equipment further includes a display module.

[0138] The display module is used for displaying the real-time optimal rotation speed information and the spectrum diagram of the satellite disc, and generating an alarm signal when the real-time optimal rotation speed information exceeds the process setting range.

[0139] The satellite disc rotation speed detection system of the MOCVD device obtains the spectrum reflection signal of the wafer surface arranged in the planetary reaction cavity, respectively performs time domain analysis and frequency domain analysis on the spectrum reflection signal, obtains time domain rotation speed information and frequency domain rotation speed information, fuses the time domain rotation speed information and the frequency domain rotation speed information to obtain rotation speed initial information, and then performs state estimation on the rotation speed initial information based on a preset system dynamics model to obtain real-time optimal rotation speed information. By using the above scheme, the weak rotation speed characteristic signal can be accurately extracted from the noise by the time domain and frequency domain fusion and the state estimation algorithm, and the detection capability and the peak position stability of the rotation speed characteristic in the strong noise background are significantly improved. Meanwhile, the fusion result is used as the initial value and the state estimation of the system dynamics is introduced, millisecond-level real-time updating and high-precision rotation speed analysis can be realized, the preset system dynamics model can dynamically fuse the current observation value and the historical state, the random error caused by the air flow disturbance, the mechanical vibration and the electromagnetic interference is effectively inhibited, the output result is smooth and stable, and the reliability is greatly enhanced. Further, the frequency domain analysis is not sensitive to the signal amplitude fluctuation, the system has stronger tolerance to common engineering problems such as light path window pollution and light source aging, and the stability of long-term measurement is ensured.

[0140] Each module in the satellite disc rotation speed detection system of the MOCVD device can be realized by software, hardware or a combination thereof. Each module can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so as to be called and executed by the processor to perform the operation corresponding to each module.

[0141] In a feasible embodiment, a computer device, which can be a terminal, has an internal structure as shown in Figure 6The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner. The wireless manner can be achieved through WIFI, mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by the processor to implement the satellite disc rotating speed detection method of the MOCVD device. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0142] Those skilled in the art can understand that, Figure 6 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0143] In a feasible embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method steps in the satellite disc rotating speed detection method of the MOCVD device.

[0144] In a feasible embodiment, a computer readable storage medium is provided, storing a computer program, and the computer program is executed by the processor to implement the method steps in the satellite disc rotating speed detection method of the MOCVD device.

[0145] In a feasible embodiment, a computer program product is provided, including a computer program, and the computer program is executed by the processor to implement the method steps in the satellite disc rotating speed detection method of the MOCVD device.

[0146] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0147] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0148] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for detecting the rotation speed of a satellite disc of an MOCVD apparatus, characterized in that, The method comprises: acquiring a spectral reflection signal of a wafer surface arranged in a planetary reaction cavity; performing time domain analysis on the spectral reflection signal to acquire a periodic pulse signal corresponding to the satellite disc in the spectral reflection signal, and obtaining time domain rotation speed information; performing frequency domain analysis on the spectral reflection signal, extracting the spectral reflection signal within a preset time, and performing Fourier transform on the spectral reflection signal to obtain a frequency domain signal, identifying peak information corresponding to the satellite disc in the frequency domain signal, and obtaining frequency domain rotation speed information; fusing and analyzing the time domain rotation speed information and the frequency domain rotation speed information to obtain initial rotation speed information; based on a preset system dynamics model, taking the initial rotation speed information as an observation, predicting the rotation state of the satellite disc to form a prior state and its uncertainty; and using a Kalman filtering algorithm to observe and update the initial rotation speed information, dynamically weighting the prior state and the observation according to their respective uncertainties to obtain real-time optimal rotation speed information.

2. The method of claim 1, wherein, The time domain analysis on the spectral reflection signal to acquire a periodic pulse signal corresponding to the satellite disc in the spectral reflection signal, and obtaining time domain rotation speed information, comprises: based on the reflectivity difference between the wafer and the satellite disc base, identifying a periodic pulse sequence from the spectral reflection signal; calculating the time interval between consecutive pulses as the pulse period, and calculating the time domain rotation speed information according to the mapping relationship between the pulse period and the rotation speed.

3. The method of claim 1, wherein, The identification of the peak information corresponding to the satellite disc in the frequency domain signal to obtain the frequency domain rotation speed information, comprises: extracting a plurality of spectral peaks within the range defined based on the prior frequency band to form a candidate set, and determining the basic spectral peak corresponding to the satellite disc from the candidate set according to the amplitude and harmonic consistency of the spectral peak, and interpolating based on the amplitude relationship of the basic spectral peak and its adjacent frequency points to obtain the rotation frequency; and calculating the frequency domain rotation speed information according to the mapping relationship between the rotation frequency and the rotation speed.

4. The method of claim 1, wherein, The fusion and analysis of the time domain rotation speed information and the frequency domain rotation speed information to obtain the initial rotation speed information, comprises: calculating the relative deviation of the time domain rotation speed information and the frequency domain rotation speed information; in the case where the relative deviation is within a preset threshold range, weighting and summing the time domain rotation speed information and the frequency domain rotation speed information based on preset time domain factors and frequency domain factors to obtain the initial rotation speed information; in the case where the relative deviation exceeds the preset threshold range, correcting the time domain rotation speed information based on the frequency domain rotation speed information; if the correction is successful, weighting and summing the corrected time domain rotation speed information and the frequency domain rotation speed information based on preset time domain factors and frequency domain factors to obtain the initial rotation speed information; wherein the correction success refers to that the relative deviation of the corrected time domain rotation speed information and the frequency domain rotation speed information is within the preset threshold range.

5. The method of claim 4, wherein, The correction of the time domain rotation speed information based on the frequency domain rotation speed information, comprises: calculating a theoretical pulse period according to the frequency domain rotation speed information; determining a preset range of time domain search window centered on the theoretical pulse period; The pulse recognition and period estimation are performed again in the time domain search window, and candidate pulses falling outside the time domain search window are eliminated, to obtain modified time domain rotation speed information.

6. The method of claim 1, wherein, The method further comprises: acquiring historical operation data; wherein the historical operation data comprises satellite disc rotation speed estimation values, planetary tray rotation speeds and process parameters; at intervals of a preset time, learning mapping relationships of the parameters in the historical operation data by using a machine learning algorithm; updating model parameters of the system dynamics model and the Kalman filtering algorithm based on the mapping relationships; and applying the updated system dynamics model and the Kalman filtering algorithm to the next rotation speed state estimation.

7. The method of claim 1, wherein, The method further comprises: preprocessing the acquired spectral reflection signal; wherein the preprocessing step comprises filtering and smoothing the spectral reflection signal, and the filtering comprises filtering out high-frequency noise and / or filtering out very low-frequency drift.

8. The method of claim 1, wherein, The method further comprises: displaying real-time optimal rotation speed information and a spectrum diagram of the satellite disc, and generating an alarm signal when the real-time optimal rotation speed information exceeds a process setting range.

9. A satellite disk rotation speed detection system of an MOCVD apparatus, characterized by, The system comprises: an acquisition module configured to acquire a spectral reflection signal of a wafer surface arranged in a planetary reaction chamber; an analysis module configured to perform time domain analysis on the spectral reflection signal, to acquire a periodic pulse signal corresponding to a satellite disc in the spectral reflection signal, to obtain time domain rotation speed information; to perform frequency domain analysis on the spectral reflection signal, to extract the spectral reflection signal within a preset time, to perform Fourier transform on the extracted spectral reflection signal, to obtain a frequency domain signal, to identify peak value information corresponding to the satellite disc in the frequency domain signal, to obtain frequency domain rotation speed information; and to fuse and analyze the time domain rotation speed information and the frequency domain rotation speed information, to obtain rotation speed initial information; a calculation module configured to predict a self-rotation state of the satellite disc based on a preset system dynamics model and the rotation speed initial information as an observation, to form a prior state and an uncertainty thereof; and to perform observation update on the rotation speed initial information by using a Kalman filtering algorithm, to dynamically weight and fuse the prior state and the observation according to their respective uncertainties, to obtain real-time optimal rotation speed information. 10.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-9. The processor implements the steps of the method of any one of claims 1-8 when executing the computer program.

11. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method of any one of claims 1-8.

Images