Methods, controllers, and media for diagnosing and correcting chirped fiber grating dispersion profiles

Abstract

This application relates to the field of femtosecond laser technology, and discloses a method, apparatus, and medium for diagnosing and correcting the dispersion distribution of chirped fiber gratings. The method involves constructing a chirped pulse amplification system; adjusting the pulse compressor so that the compressed pulse, as displayed by an autocorrelator, exhibits the narrowest pulse width under the current state; inserting a laterally movable narrowband filter into the diffraction path of the pulse compressor, sequentially blocking different wavelength spectral components by moving the filter laterally, and using an autocorrelator to monitor the time-domain shape of the compressed pulse after blocking different spectral components in real time; determining the corresponding spectral component as the error spectral component causing the dispersion distribution mismatch based on the time-domain shape of the compressed pulse; and correcting the dispersion distribution of the chirped fiber grating in the grating region corresponding to the determined error spectral component based on the determined error spectral component. By scanning the spectral components, the problem of "perceptible but difficult to locate" high-order dispersion errors is solved.

Description

Technical Field

[0001] This application relates to the field of femtosecond laser technology, and in particular to a method, apparatus and medium for diagnosing and correcting the dispersion distribution of chirped fiber gratings. Background Technology

[0002] In femtosecond laser chirped pulse amplification (CPA) systems, the chirped fiber grating serves as the core pulse broadening device. Its dispersion characteristics (especially higher-order dispersion) must be precisely matched with the dispersion of the pulse compressor to avoid temporal distortions such as base and sidelobes in the compressed pulse, ensuring a high signal-to-noise ratio output. However, existing technologies face the following key challenges:

[0003] 1. High-order dispersion errors are difficult to pinpoint precisely: During the fabrication process, chirped fiber gratings are affected by factors such as optical path offset, mechanical vibration, and fluctuations in light source power, causing their actual dispersion curves to deviate from the design values, resulting in incomplete compensation for high-order dispersion. These minute errors (time-domain delay differences of only tens of femtoseconds to several picoseconds) are difficult to quantify accurately using traditional measurement methods, and existing quality inspection methods cannot directly pinpoint the spectral region corresponding to specific errors.

[0004] 2. Lack of targeted diagnostic and corrective methods: Traditional debugging methods rely solely on adjusting pulse compressor parameters to optimize pulse shape, but cannot pinpoint whether the error originates from the stretcher or the compressor, let alone pinpoint the dispersion deviation of a specific grating region within the chirped fiber grating. Currently, there is no systematic solution for accurately locating the error spectral region by actively screening spectral components, combined with process investigation or online stress adjustment.

[0005] Therefore, a method is urgently needed to solve at least one of the above problems. Summary of the Invention

[0006] This application provides methods, apparatus, and media for diagnosing and correcting the dispersion distribution of chirped fiber gratings. It aims to address the shortcomings of existing technologies, which primarily focus on measuring spectral reflectance characteristics or overall dispersion parameters for detecting fiber grating dispersion errors, neglecting spectral component scanning analysis based on actual pulse compression effects. Furthermore, optimization of pulse time-domain distortion is limited to passively adjusting the compressor structure, without proposing active filtering to locate the error source and guide device fabrication or real-time control. Therefore, existing technologies fail to recognize the direct correlation between "error spectral components and grating dispersion deviation" and do not provide solutions combining spectral filtering scanning with stress adjustment, resulting in significant technological gaps in the diagnosis and correction of higher-order dispersion errors.

[0007] In a first aspect, embodiments of this application provide a method for diagnosing and correcting the dispersion distribution of chirped fiber gratings, comprising:

[0008] A chirped pulse amplification system is constructed, comprising a seed source, a chirped fiber grating as a pulse stretcher, an fiber amplifier, a pulse compressor composed of spatial light diffraction elements, and an autocorrelator for monitoring the time-domain shape of the compressed pulse. The pulse compressor is adjusted so that the autocorrelator displays the compressed pulse at its narrowest pulse width under the current state. If the compressed pulse exhibits time-domain shape distortion, it is determined that there is a mismatch between the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor. The time-domain shape distortion includes the base and side lobes.

[0009] A laterally movable narrowband filter is inserted into the diffraction optical path of the pulse compressor. The blocking direction of the narrowband filter is perpendicular to the distribution direction of the diffraction spectrum. By laterally moving the narrowband filter, different wavelength spectral components are blocked in sequence, and the time-domain shape of the compressed pulse after blocking different spectral components is monitored in real time using the autocorrelation instrument.

[0010] The corresponding spectral component is determined as the error spectral component that causes the dispersion distribution mismatch based on the time-domain shape of the compressed pulse; based on the determined error spectral component, the chirped fiber grating dispersion distribution is corrected in the grating region corresponding to the error spectral component in the chirped fiber grating.

[0011] In some embodiments, the step of correcting the chirped fiber grating dispersion distribution in the grating region corresponding to the determined error spectral component based on the determined error spectral component includes: checking the accuracy of optical path calibration and the state of optical components corresponding to the error spectral component during the chirped fiber grating writing process; improving the writing process for the identified deviations or anomalies; or applying stress to the grating region through a stress application device and adjusting the dispersion curve of the chirped fiber grating corresponding to the error spectral component in real time until the time-domain shape of the compressed pulse displayed by the autocorrelator meets the preset requirements.

[0012] In some embodiments, the construction of the chirped pulse amplification system includes: connecting the seed source output end to the front port of an optical fiber circulator via optical fiber fusion splicing; connecting the middle port of the optical fiber circulator to a chirped fiber grating via optical fiber; connecting the rear port of the optical fiber circulator to the input end of an optical fiber amplifier via optical fiber; connecting the output end of the optical fiber amplifier to an optical fiber collimator via optical fiber; aligning the output end of the optical fiber collimator with a pulse compressor composed of a pair of gratings; providing a laterally movable narrowband filter in the diffraction path of the pulse compressor; and connecting the output optical path of the pulse compressor to the detection end of an autocorrelator.

[0013] In some embodiments, adjusting the pulse compressor so that the autocorrelator displays the compressed pulse at its narrowest pulse width in the current state includes: adjusting the relative angle or spacing of the two surface gratings in the pulse compressor and monitoring the pulse width of the compressed pulse displayed by the autocorrelator until the pulse width value displayed by the autocorrelator reaches the minimum value in the current state, and the pulse width no longer narrows when the adjustment continues.

[0014] In some embodiments, determining that the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor are mismatched if the compressed pulse exhibits time-domain shape distortions such as base and side lobes includes: comparing the time-domain curve of the compressed pulse measured by the autocorrelator with the ideal undistorted pulse curve; if there is a continuous low-intensity signal below the main peak of the pulse or symmetrical or asymmetrical low-intensity peaks appear on both sides of the main peak, then it is determined that there is a higher-order dispersion mismatch between the two dispersion distributions.

[0015] In some embodiments, the step of sequentially blocking spectral components of different wavelengths by laterally moving the narrowband filter includes: the width direction of the narrowband filter is perpendicular to the wavelength distribution direction of the spectrum in the diffraction optical path of the pulse compressor, and the filter is gradually slid along the width direction with a preset step size, and fixed at the current position after each slide, so that the filter sequentially blocks beams of different wavelength ranges in the diffraction spectrum, and the spectral width of a single block is 1-5 nm; the preset step size includes a displacement amount corresponding to 0.1-5 nm.

[0016] In some embodiments, the step of using the autocorrelator to monitor the time-domain shape of the compressed pulse after blocking different spectral components in real time includes: after the narrowband filter blocks a specific spectral component each time, keeping the system stable for at least one pulse cycle, triggering the autocorrelator to collect the time-domain curve of the current compressed pulse, and recording and storing the characteristic parameters of the main peak width, base intensity and sidelobe amplitude of the time-domain curve.

[0017] In some embodiments, determining the corresponding spectral component as the error spectral component causing the dispersion distribution mismatch based on the time-domain shape of the compressed pulse includes: comparing the time-domain curves after blocking different spectral components; when a certain spectral component is blocked, the base intensity of the compressed pulse is significantly reduced or the side lobes completely disappear, and the main peak width is close to the ideal value, and determining that the corresponding blocked spectral component is the error spectral component causing the dispersion distribution mismatch.

[0018] Secondly, embodiments of this application provide a stress unit controller, which includes a memory and a processor. The memory is used to store a computer program, and the processor is used to execute the computer program and implement the method provided in any embodiment of this application when executing the computer program.

[0019] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to implement the method provided in any embodiment of this application.

[0020] The method, apparatus, and medium provided in this application for diagnosing and correcting the dispersion distribution of chirped fiber gratings directly pinpoint the specific spectral region in the chirped fiber grating that causes dispersion mismatch by scanning spectral components, solving the problem of "perceptible but difficult to locate" high-order dispersion errors. Based on the location results, optical path deviations or optical component abnormalities during the writing process can be specifically investigated, or grating dispersion can be corrected in real time through stress adjustment, significantly improving the dispersion matching degree between the chirped fiber grating and the compressor. By eliminating distortions such as the base and side lobes of the compressed pulse, the time-domain signal-to-noise ratio is improved, ensuring the stability and reliability of the femtosecond laser system.

[0021] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic flowchart illustrating the steps of a method for diagnosing and correcting the dispersion distribution of a chirped fiber grating according to an embodiment of this application;

[0024] Figure 2 This is a schematic diagram of a structure for diagnosing and correcting the dispersion distribution of a chirped fiber grating according to an embodiment of this application;

[0025] Figure 3 This is a dispersion curve diagram of a chirped fiber grating provided in one embodiment of this application;

[0026] Figure 4 This is a schematic diagram of the structure of a device for diagnosing and correcting the dispersion distribution of a chirped fiber grating according to an embodiment of this application;

[0027] Figure 5 This is a schematic block diagram of the structure of a stress unit controller provided in an embodiment of this application.

[0028] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Detailed Implementation

[0029] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0031] It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0032] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0033] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0034] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0035] In femtosecond laser chirped pulse amplification (CPA) systems, the chirped fiber grating serves as the core pulse broadening device. Its dispersion characteristics (especially higher-order dispersion) must be precisely matched with the dispersion of the pulse compressor to avoid temporal distortions such as base and sidelobes in the compressed pulse, ensuring a high signal-to-noise ratio output. However, existing technologies face the following key challenges:

[0036] 1. High-order dispersion errors are difficult to pinpoint precisely: During the fabrication process, chirped fiber gratings are affected by factors such as optical path offset, mechanical vibration, and fluctuations in light source power, causing their actual dispersion curves to deviate from the design values, resulting in incomplete compensation for high-order dispersion. These minute errors (time-domain delay differences of only tens of femtoseconds to several picoseconds) are difficult to quantify accurately using traditional measurement methods, and existing quality inspection methods cannot directly pinpoint the spectral region corresponding to specific errors.

[0037] 2. Lack of targeted diagnostic and corrective methods: Traditional debugging methods rely solely on adjusting pulse compressor parameters to optimize pulse shape, but cannot pinpoint whether the error originates from the stretcher or the compressor, let alone pinpoint the dispersion deviation of a specific grating region within the chirped fiber grating. Currently, there is no systematic solution for accurately locating the error spectral region by actively screening spectral components, combined with process investigation or online stress adjustment.

[0038] In existing technologies, the detection of fiber optic grating dispersion errors mostly focuses on measuring spectral reflectance characteristics or overall dispersion parameters, without addressing spectral component scanning analysis based on actual pulse compression effects. Furthermore, optimization of pulse temporal distortion is limited to passively adjusting the compressor structure, without proposing an approach to actively filter and locate the error source to guide device fabrication or real-time control. Therefore, existing technologies neither recognize the direct correlation between "error spectral components and grating dispersion deviation" nor provide a solution combining spectral filtering scanning and stress adjustment, resulting in a significant technological gap in the diagnosis and correction of higher-order dispersion errors.

[0039] To resolve the above issues, please refer to... Figure 1 This application provides a method for diagnosing and correcting the dispersion distribution of chirped fiber gratings, used for encapsulating such... Figure 2 The structure shown for diagnosing and correcting the dispersion distribution of chirped fiber gratings includes a seed source 1, an optical fiber circulator 2, an optical fiber amplifier 3, an optical fiber collimator 4, a diffraction grating 5, a lifting mirror 6, a mirror 7, a narrowband filter 8, a chirped fiber grating 9, a stress application unit 10, an autocorrelator 11, a spectrometer 12, and a stress unit controller 13.

[0040] Specifically, such as Figure 1 As shown, the method for diagnosing and correcting the dispersion distribution of chirped fiber gratings includes steps S101 to S103. Details are as follows:

[0041] Step S101. Construct a chirped pulse amplification system, which includes a seed source, a chirped fiber grating as a pulse stretcher, an fiber amplifier, a pulse compressor composed of spatial light diffraction elements, and an autocorrelator for monitoring the time-domain shape of the compressed pulse; adjust the pulse compressor so that the autocorrelator displays the compressed pulse at its narrowest pulse width under the current state; if the compressed pulse exhibits time-domain shape distortion, it is determined that there is a mismatch between the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor; the time-domain shape distortion includes the base and side lobes.

[0042] Specifically, the core of this step is to initially determine whether the dispersion of the chirped fiber grating (stretcher) and the pulse compressor is matched by setting up a chirped pulse amplification (CPA) system. Specifically, the compressor is adjusted to compress the pulse to its narrowest pulse width under the current state. If temporal distortion (such as base or sidelobe) appears after compression, it is determined that there is a mismatch in the dispersion distribution of the two, especially high-order dispersion errors that lead to incomplete compensation.

[0043] The system is constructed by connecting various components to form an optical path: the short pulse output from the seed source enters a chirped fiber grating (acting as a stretcher) via an optical fiber circulator. The stretched pulse is amplified by an optical fiber amplifier, then coupled to the spatial optical path via an optical fiber collimator. The spectrum is then expanded by a diffraction grating (the core element of the pulse compressor). The optical path is adjusted by lifting and adjusting mirrors. Finally, an autocorrelator monitors the temporal shape of the compressed pulse, while a spectrometer simultaneously monitors the spectral distribution. The pulse compressor is composed of spatial light diffraction elements such as diffraction gratings. The dispersion parameters of the compressor are adjusted by changing the mirror angle or the grating spacing.

[0044] Preliminary adjustment and distortion assessment involve adjusting compressor parameters (such as grating spacing and mirror angle) to minimize the compressed pulse width displayed by the autocorrelator (i.e., the "narrowest pulse width" in the current state). Observing the autocorrelator output, if the pulse time-domain waveform shows a base (elongated pulse tail) or side lobes (small peaks appearing on both sides of the main peak), it is determined that there is a mismatch between the dispersion (especially higher-order dispersion) of the stretcher (chirped fiber grating) and the compressor, leading to compensation errors.

[0045] Step S102. Insert a laterally movable narrowband filter into the diffraction optical path of the pulse compressor. The blocking direction of the narrowband filter is perpendicular to the distribution direction of the diffraction spectrum. By laterally moving the narrowband filter, different wavelength spectral components are blocked in sequence, and the time-domain shape of the compressed pulse after blocking different spectral components is monitored in real time using the autocorrelation instrument.

[0046] Specifically, this step involves inserting a laterally movable narrowband filter into the diffraction path of the compressor to actively filter spectral components of specific wavelengths. By blocking different wavelengths and monitoring changes in the temporal shape of the compressed pulse, the error spectral region causing dispersion mismatch is located. The blocking direction of the narrowband filter is perpendicular to the distribution direction of the diffraction spectrum, ensuring that only spectral components within a single wavelength range are blocked during lateral movement.

[0047] The filter insertion and optical path matching involves installing a precisely movable narrowband filter (such as a slit or bandpass filter with a width of Δλ) perpendicular to the spectral distribution direction (i.e., laterally) in the spectral unfolding optical path behind the diffraction grating (i.e., the plane where the spectrum is distributed laterally by wavelength). The passband direction of the filter is parallel to the spectral distribution direction, and the blocking direction is perpendicular (i.e., covering different wavelength regions when moving laterally). The spectral width of a single blocking operation is 1-5 nm, which is a contradiction, but ensures that the spectral components blocked each time have high wavelength resolution, facilitating precise location of error regions.

[0048] Spectral component scanning and monitoring are achieved by controlling the lateral movement of the filter using a stepper motor or precision displacement stage, sequentially blocking different wavelengths of spectral components (e.g., blocking from short wavelengths to long wavelengths in segments, with each blocking width Δλ and an overlap rate of approximately 50% to avoid missed detections). After each blocking, the compressor parameters are kept constant, and the time-domain shape of the compressed pulse is recorded in real time using an autocorrelation instrument. Simultaneously, the wavelength range of the current blocking is confirmed using a spectrometer. Establish the correspondence between "blocking wavelength - time domain waveform".

[0049] Step S103. Determine the corresponding spectral component as the error spectral component that causes the dispersion distribution mismatch based on the time-domain shape of the compressed pulse; perform chirped fiber grating dispersion distribution correction on the grating region in the chirped fiber grating corresponding to the error spectral component based on the determined error spectral component.

[0050] Specifically, based on the change in temporal shape in step S102, the key spectral components causing distortion (i.e., the spectral segments where temporal distortion is significantly improved or worsened when a certain wavelength region is blocked) are identified and determined as "error spectral components". Since the grating regions of the chirped fiber grating correspond one-to-one with the wavelengths (short wavelengths correspond to the grating ends, and long wavelengths correspond to the beginnings), by locating the grating region corresponding to the error spectrum, local stress (such as stretching or squeezing) is applied to the grating region using a stress application unit to adjust its dispersion characteristics, or process errors (such as optical path offset or abnormal exposure dose) of the corresponding grating region are checked during the writing process to achieve dispersion distribution correction.

[0051] Error spectral component localization is achieved by analyzing the time-domain waveform after blocking different spectral components: if a certain wavelength region (e.g.) is blocked... When the pulse base or sidelobe decreases significantly, the wavelength region is identified as an "error spectral component," and the corresponding dispersion error is not compensated by the compressor. Based on the design parameters of the chirped fiber grating (wavelength-grating region position mapping relationship), the grating region position corresponding to this error spectral component is determined (e.g., grating region x corresponds to wavelength λ0).

[0052] Dispersion correction is achieved through online stress adjustment (for packaged devices). A stress application unit (such as a piezoelectric ceramic (PZT) driven mechanical clamp) applies local strain to the target grating region, altering the grating period (ΔΛ) or effective refractive index (Δn) in that region, thereby adjusting its dispersion characteristics (such as second-order and third-order dispersion). The stress unit controller precisely controls the position and amplitude of the applied stress based on the grating region location corresponding to the error spectrum. Dispersion matching is achieved through closed-loop feedback (repeating steps S101-S102 after stress adjustment until temporal distortion is eliminated).

[0053] Process error investigation (for device fabrication stage): If an error is found during the quality inspection stage, the process parameters (such as optical path alignment accuracy, exposure light source power fluctuation, mechanical vibration signal) of the corresponding time period during the writing process are traced according to the location of the gate area. The writing program or equipment parameters are then corrected to avoid the recurrence of the same error.

[0054] In some embodiments, the step of correcting the chirped fiber grating dispersion distribution in the grating region corresponding to the determined error spectral component based on the determined error spectral component includes: checking the accuracy of optical path calibration and the state of optical components corresponding to the error spectral component during the chirped fiber grating writing process; improving the writing process for the identified deviations or anomalies; or applying stress to the grating region through a stress application device and adjusting the dispersion curve of the chirped fiber grating corresponding to the error spectral component in real time until the time-domain shape of the compressed pulse displayed by the autocorrelator meets the preset requirements.

[0055] For the identified error spectral components, two correction strategies are provided: **Writing process investigation and improvement:** Trace the writing parameters of the corresponding error spectral region during the chirped fiber grating manufacturing process to investigate issues such as optical path alignment deviations and abnormal conditions of optical components (e.g., phase masks, exposure light sources), and eliminate systematic errors by improving the manufacturing process. **Real-time stress adjustment:** For packaged devices, apply local stress to the grating region corresponding to the error spectrum using a stress application device to adjust the dispersion characteristics of that region in real time until the compressed pulse time-domain shape meets the requirements.

[0056] Troubleshooting during the writing process involves checking the wavelength range corresponding to the spectral components of the error (e.g., ...). Retrieve the log data from when the raster was written, including: phase mask position offset ( ), fluctuation in the power of the exposure light source ( ), vibration amplitude of mechanical platform ( ); optical path collimation (such as focal length deviation of focusing lens, beam incident angle error). If a parameter is found to exceed the design tolerance (such as insufficient modulation of the grid refractive index due to a sudden drop in exposure power), the writing equipment parameters are corrected (such as enhancing the stability of the light source and optimizing the optical path alignment algorithm), and targeted quality inspections are performed on subsequent batches of devices.

[0057] Real-time stress adjustment is achieved through a piezoelectric ceramic (PZT) driven miniature clamp in the stress application device, precisely positioning it to the target gate region. The grating period in this region can be changed by applying axial tensile or radial compressive stress. The stress unit controller adjusts PZT in preset steps (e.g., 0.1V voltage corresponds to 10με strain). After each adjustment, wait for the system to stabilize (about 10ms). Repeat step S102 to scan the spectral components until the autocorrelation instrument shows that the base intensity drops to less than 5% of the main peak and the side lobes disappear completely.

[0058] In some embodiments, the construction of the chirped pulse amplification system includes: connecting the seed source output end to the front port of an optical fiber circulator via optical fiber fusion splicing; connecting the middle port of the optical fiber circulator to a chirped fiber grating via optical fiber; connecting the rear port of the optical fiber circulator to the input end of an optical fiber amplifier via optical fiber; connecting the output end of the optical fiber amplifier to an optical fiber collimator via optical fiber; aligning the output end of the optical fiber collimator with a pulse compressor composed of a pair of gratings; providing a laterally movable narrowband filter in the diffraction path of the pulse compressor; and connecting the output optical path of the pulse compressor to the detection end of an autocorrelator.

[0059] Clarify the hardware connection method of the chirped pulse amplification system to ensure the optical coupling accuracy of each component, especially the connection between the optical fiber and the spatial optical path, and the position of the narrowband filter in the compressor diffraction optical path.

[0060] The fiber optic link connection is achieved through a seed source (such as a 100fs pulsed laser with a center wavelength of 1030nm) and an output end connected to a fiber optic fusion splicer and a fiber optic circulator (three-port). The front port (port 1) of the circulator is connected; the middle port (port 2) of the circulator is fused with a chirped fiber grating (5-10cm in length, 50nm in reflection bandwidth), and the rear port (port 3) is connected to the input of a fiber amplifier (such as a ytterbium-doped fiber amplifier with a gain of 30dB) via a single-mode fiber.

[0061] The spatial optical path is constructed by coupling the amplifier output to a pulse compressor (composed of a pair of 1200 lines / mm surface gratings with an adjustable spacing of 5-20cm) via an optical fiber collimator (collimated spot diameter 2mm, divergence angle ≤0.5°). In the diffraction path between the two gratings, a narrowband filter (e.g., 5mm width, 2nm passband bandwidth) is installed perpendicular to the spectral distribution direction (i.e., wavelength dispersion direction). The filter is moved laterally by a precision displacement stage (step accuracy 1μm, corresponding wavelength offset 0.05nm). The compressor output is focused by a mirror and incident on the detection end of an autocorrelator (time resolution ≤50fs), which is simultaneously connected to a spectrometer (resolution 0.1nm) to monitor the spectrum.

[0062] In some embodiments, adjusting the pulse compressor so that the autocorrelator displays the compressed pulse at its narrowest pulse width in the current state includes: adjusting the relative angle or spacing of the two surface gratings in the pulse compressor and monitoring the pulse width of the compressed pulse displayed by the autocorrelator until the pulse width value displayed by the autocorrelator reaches the minimum value in the current state, and the pulse width no longer narrows when the adjustment continues.

[0063] By adjusting the relative angle or spacing of the surface gratings in the pulse compressor, the dispersion compensation is optimized to minimize the pulse width of the compressed pulse, which serves as the initial criterion for determining dispersion mismatch.

[0064] Parameter adjustment methods include: Angle adjustment: using a high-precision rotary stage (angle resolution 0.01°) to adjust the relative tilt angle of the two gratings, changing the difference in spectral diffraction path length, thereby adjusting the second-order dispersion (GDD); Spacing adjustment: changing the grating spacing through a linear displacement stage (accuracy 10μm), which simultaneously affects the second-order and higher-order dispersions (such as the third-order dispersion TOD).

[0065] The closed-loop optimization process involves manually or automatically (e.g., through LabVIEW) scanning and adjusting parameters. After each adjustment, a 20ms waiting period (system stabilization time) is waited, and the pulse width value (full width at half maximum, FWHM) displayed by the autocorrelation instrument is collected. When the pulse width no longer decreases after three consecutive adjustments (e.g., change < 1%), the current parameter is recorded as the "current optimal state". If the pulse still has a base (tail intensity > main peak 10%) or sidelobe (amplitude > main peak 5%) at this time, dispersion mismatch is determined.

[0066] In some embodiments, determining that the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor are mismatched if the compressed pulse exhibits time-domain shape distortions such as base and side lobes includes: comparing the time-domain curve of the compressed pulse measured by the autocorrelator with the ideal undistorted pulse curve; if there is a continuous low-intensity signal below the main peak of the pulse or symmetrical or asymmetrical low-intensity peaks appear on both sides of the main peak, then it is determined that there is a higher-order dispersion mismatch between the two dispersion distributions.

[0067] By comparing the measured pulse time-domain curve with the ideal Gaussian / hyperbolic secant curve, the degree of dispersion mismatch is quantified based on characteristic parameters (base strength, sidelobe amplitude) to avoid human misjudgment.

[0068] Assuming the ideal distortion-free pulse is hyperbolic secant, the expression is: , where τ is the full width at half maximum (FWHM) pulse width, and the theoretical pulse width is estimated by seed source parameters and system design (e.g., the target pulse width after compression is 50fs).

[0069] Distortion feature extraction includes: Base determination: Calculate the time range of intensity 15% below the peak value of the main peak. If the duration of the tail signal exceeds 50% of the pulse width of the main peak, a base is determined to exist; Side lobe determination: Detect secondary peaks within ±2τ on both sides of the main peak. If their amplitude exceeds 5% of the main peak, a side lobe is determined to exist.

[0070] The comparison algorithm uses cross-correlation coefficients (CQCs) The difference between the measured curve and the ideal curve is quantified by the root mean square error (threshold < 10%). If the threshold is not met, the dispersion is judged to be mismatched.

[0071] In some embodiments, the step of sequentially blocking spectral components of different wavelengths by laterally moving the narrowband filter includes: the width direction of the narrowband filter is perpendicular to the wavelength distribution direction of the spectrum in the diffraction optical path of the pulse compressor, and the filter is gradually slid along the width direction with a preset step size, and fixed at the current position after each slide, so that the filter sequentially blocks beams of different wavelength ranges in the diffraction spectrum, and the spectral width of a single block is 1-5 nm; the preset step size includes a displacement amount corresponding to 0.1-5 nm.

[0072] By clearly defining the scanning direction, step size, and blocking width of the narrowband filter, the resolution and coverage of spectral component screening are ensured, enabling precise positioning of the error spectrum.

[0073] The filter uses a transmission bandpass filter with a bandwidth of [missing information]. (Full width at half maximum), passband direction parallel to spectral dispersion direction (i.e., perpendicular to the moving direction), physical width 5mm corresponds to a spectral range of approximately 10nm (calculated based on grating dispersion rate, such as 1nm / mm); moving direction perpendicular to spectral distribution (e.g., horizontal scanning, spectrum unfolded along the horizontal axis), preset step size. The corresponding wavelength shift is 0.1 nm (when the dispersion rate is 10 nm / mm, 10 μm corresponds to 0.1 nm).

[0074] The scanning process involves scanning segment by segment from the short-wavelength end of the spectrum (e.g., 1000 nm) to the long-wavelength end (e.g., 1100 nm), with a single spectral block width of 2 nm and an overlap of 0.5 nm between adjacent scan segments (to avoid missed detections); after each movement, the current passband center wavelength is confirmed using a spectrometer. ), and record the mapping relationship between the displacement stage coordinates and the wavelength.

[0075] In some embodiments, the step of using the autocorrelator to monitor the time-domain shape of the compressed pulse after blocking different spectral components in real time includes: after the narrowband filter blocks a specific spectral component each time, keeping the system stable for at least one pulse cycle, triggering the autocorrelator to collect the time-domain curve of the current compressed pulse, and recording and storing the characteristic parameters of the main peak width, base intensity and sidelobe amplitude of the time-domain curve.

[0076] By ensuring system stability after spectral components are blocked, a triggered acquisition method is used to record time-domain curves, and key feature parameters are extracted for subsequent analysis to avoid noise interference.

[0077] System stability and triggering are achieved by waiting at least one pulse cycle (e.g., 1ms for a repetition frequency of 1kHz) after each filter movement, until the optical path vibration decays (vibration amplitude monitored by an accelerometer is <10μm / s). 2 The seed source synchronous signal triggers the autocorrelator (trigger delay accuracy ±10ns), continuously acquiring 10 sets of pulse signals, and averaging them to reduce noise.

[0078] Feature parameter extraction includes: main peak width: the full width at half maximum (FWHM) of the autocorrelation curve, calculated by Gaussian fitting; base strength: the ratio of the average strength of the region below 10% of the main peak value to the main peak value (unit: %); side lobe amplitude: the ratio of the secondary peak value to the main peak value (unit: %), with the maximum side lobe amplitude recorded.

[0079] Data storage is achieved by establishing a database to store each occlusion wavelength range. The corresponding characteristic parameters are associated with the filter displacement coordinates and spectral position.

[0080] In some embodiments, determining the corresponding spectral component as the error spectral component causing the dispersion distribution mismatch based on the time-domain shape of the compressed pulse includes: comparing the time-domain curves after blocking different spectral components; when a certain spectral component is blocked, the base intensity of the compressed pulse is significantly reduced or the side lobes completely disappear, and the main peak width is close to the ideal value, and determining that the corresponding blocked spectral component is the error spectral component causing the dispersion distribution mismatch.

[0081] By comparing the temporal characteristic changes after different spectral components are blocked, the key wavelength regions that cause distortion are identified, and the determination is based on significant changes in characteristic parameters (such as a decrease in base intensity of >30%).

[0082] The baseline data was established by acquiring the "full spectrum" time-domain curves without obscuring any spectral components, and recording the baseline base intensity (Bref) and sidelobe amplitude (Sref).

[0083] Feature difference calculation involves calculating the rate of change in base intensity based on the curve after a certain segment of the spectral composition is blocked. Sidelobe amplitude variation rate If ΔB > 30% and Smeas < 2% (side lobes basically disappear), or the main peak width is close to the theoretical value (deviation < 5%), the corresponding spectral component is determined to be an error component.

[0084] Multiple verifications are performed by reverse verification of the determined error spectral components: after removing the filter, the spectral components are reintroduced. If the time-domain distortion reappears, the positioning is confirmed to be accurate. Combined with spectrometer data, it is checked whether there are abnormal reflectance in the wavelength region (e.g., the designed reflectance is 95%, but the actual measured reflectance is <90%) or bandwidth broadening, which helps to confirm the grating dispersion deviation.

[0085] In some embodiments, by making the compressor adjustment and distortion determination process in step S101 intelligent, the deep reinforcement learning (DRL) algorithm is used to replace manual trial and error adjustment, and the temporal distortion features are automatically identified by the convolutional neural network (CNN) to construct a "adjustment-identification-determination" closed loop.

[0086] The reinforcement learning environment is constructed as follows: State space: containing the real-time pulse width of the autocorrelator, base strength, sidelobe amplitude, and current compressor parameters (grating angle θ1 / θ2, spacing d), a total of 6-dimensional state vectors; Action space: defining the compressor parameter adjustment step size (e.g., ... , There are a total of 9 discrete actions; the reward function is: based on " "As a negative reward, it guides the agent to search in the direction of minimum pulse width and minimum distortion."

[0087] The CNN distortion recognition model includes: Input: Time-domain curve image output by autocorrelation instrument (resolution 200×200 pixels); Output: Three-class classification result (no distortion / base distortion / side lobe distortion), training data comes from simulation and historical test data (100,000 sets of augmented data); Judgment logic: When the classification result is "base" or "side lobe" for 3 consecutive times, the error localization process is triggered.

[0088] The implementation process includes: initializing the DRL agent (using the PPO algorithm), loading the pre-trained model for rapid convergence; the agent automatically adjusts the compressor parameters, waits 50ms for the system to stabilize after each action, and the CNN judges the distortion type in real time until the reward value tends to stabilize (fluctuation <1%), which is determined as the current optimal state. If distortion still exists, it enters the S102 scan.

[0089] In some embodiments, for the spectral scanning in step S102, the traditional point-by-point scanning mode is broken through. Bayesian optimization (BO) is used to dynamically select the most informative spectral region for masking, and the mapping relationship between "spectral composition and temporal distortion" is modeled by Gaussian process regression (GPR), reducing the number of scans by more than 70%.

[0090] The surrogate model is established as follows: Input: center wavelength λ of the occlusion spectrum (a continuous variable from 1000-1100 nm); Output: pedestal intensity B(λ), sidelobe amplitude S(λ), constructing a GPR model to fit the distortion response in the unknown region; Initial sample collection: 10 wavelength points are randomly selected for scanning as prior data. Acquisition function optimization includes: using the expected improvement (EI) as the acquisition function, prioritizing scanning the wavelength points with the greatest predicted distortion improvement potential (i.e., λ with the highest EI value); updating the GPR model after each scan, iterating until the EI value is below a threshold (e.g., 0.01) or the number of scans reaches 50.

[0091] The hardware co-design includes: the narrowband filter is driven by a high-precision electronically controlled displacement stage (positioning accuracy 0.01nm), and the wavelength-displacement mapping relationship is calibrated in real time by an integrated spectrometer; a BO algorithm control module is developed, which communicates with the displacement stage and autocorrelator in real time through an API interface to realize dynamic adjustment of the scanning strategy.

[0092] In some embodiments, in step S103, a fully connected neural network (FCN) is used to directly map the error spectral components from the time-domain feature parameters. The (PID) controller realizes closed-loop feedback of stress application and constructs an "identification-control" intelligent unit.

[0093] The error localization neural network consists of: an input layer (3D features: peak width, base intensity, and sidelobe amplitude of the temporal curve after occlusion); a hidden layer (2 layers with 128 neurons, ReLU activation function); and an output layer for spectral region classification (50 wavelength ranges, 1000-1100nm per 2nm class). Training data is generated by simulating different grating dispersion errors to produce corresponding temporal features, constructing 100,000 sets of labeled data with a test accuracy ≥95%.

[0094] The stress PID control loop includes: the controlled object: the stress application device for the target gate region (PZT voltage V, range 0-10V); the feedback quantity: the base strength B after the current error spectral component is blocked, and the target value. (Without base); PID parameters: proportional coefficient Kp=0.5, integral coefficient Ki=0.1, derivative coefficient Kd=0.2, tuned using the Ziegler-Nichols method; control logic: real-time error calculation. Output voltage regulation until e < 1%.

[0095] The implementation steps include: the FCN model receives the feature parameters collected by S102 and outputs the error spectral range with the highest probability (e.g., Locate the corresponding grating region (calculated based on the grating chirp rate; the grating region position corresponding to 1050nm is x=7.2cm), start the PID controller to adjust the stress, and update the control quantity every 10ms until the autocorrelator shows that the distortion has been eliminated.

[0096] In some embodiments, to address the problem of lacking ideal samples in steps S101-S103, an ideal distortion-free pulse generated by GAN is used as a benchmark, and an unsupervised anomaly detection algorithm (such as isolated forest) is combined to automatically identify the spectral components of the error, which is suitable for the diagnosis of unknown types of dispersion errors.

[0097] The ideal pulse generation of GAN includes: Generator G: input random noise vector, output simulated ideal autocorrelation curve (Gaussian / hyperbolic secant); Discriminator D: distinguishes between real curve and generated curve, and through adversarial training, makes G generate an ideal curve close to the actual system; Training data: collected compressed pulse data of defect-free chirped fiber gratings of the same type as real samples.

[0098] Unsupervised anomaly detection calculates the Wasserstein distance (WD) between the temporal curve of each occluded spectrum acquired by S102 and the ideal curve generated by GAN; anomaly detection of WD values ​​is performed using the isolated forest algorithm. When the WD value of a certain wavelength region exceeds 3 times the standard deviation, it is determined to be an error spectral component; combined with the grating dispersion model, the abnormal wavelength is mapped to a specific grating region (e.g., the grating region with the highest WD value λ=1060nm corresponds to x=8cm, which is due to insufficient exposure dose during writing, resulting in dispersion deviation).

[0099] For detected abnormal gate regions, the adaptive correction strategy calls the optimal stress adjustment parameters for similar errors in the historical database (e.g., λ=1060nm corresponds to a PZT voltage of 4.2V) to achieve rapid compensation; if historical data is missing, the gradient descent algorithm is started to optimize the stress parameters with the goal of minimizing the WD value, with a step size α=0.1 and the number of iterations ≤20.

[0100] In some embodiments, a multimodal diagnostic system based on transfer learning is constructed to address the individualized errors of different batches of chirped fiber Bragg gratings. This system integrates data from spectrometers, autocorrelators, and stress sensors to quickly adapt to the dispersion error diagnosis of new devices and reduce the cost of repeated training.

[0101] Multimodal feature extraction includes: spectral modes: extracting the peak wavelength, 3dB bandwidth, and ripple (high-frequency components after Fourier transform) of the reflection spectrum of chirped fiber gratings; temporal modes: the main peak width, base energy ratio, and sidelobe symmetry of the autocorrelation curve (quantized by Hu moment); stress modes: the correlation matrix between the historical adjustment of each stress-applying unit and the distortion improvement.

[0102] The transfer learning architecture includes: Pre-training stage: Training a general feature extractor using multiple sets of historical device data (based on...) It captures the cross-modal mapping of "gate location - dispersion error - temporal characteristics"; fine-tuning stage: only 20 sets of scan data are needed for new devices, and the fully connected layer is adjusted through domain adaptation to quickly adapt to the error mode of specific devices.

[0103] After the intelligent diagnostic process is connected to the system through the new device, it first collects 5 sets of full-spectrum compressed pulse data, inputs them into the pre-trained model to obtain the initial error region prediction (such as the first 3 wavelength intervals with confidence > 90%); it prioritizes scanning the prediction region, inputs the data into the fine-tuned model, and combines Bayesian optimization to correct the prediction, and finally locates the error spectral components, which is more efficient than the traditional method.

[0104] In some embodiments, when a femtosecond laser system employs chirped pulse amplification, its pulse time-domain stretching device is a dispersive medium or grating device capable of providing a large amount of dispersion. For fiber femtosecond lasers, chirped fiber gratings are commonly used as pulse stretchers. Chirped fiber gratings possess a wide range of designable and manufacturable dispersion characteristics, including center wavelength, dispersion coefficient, reflection spectrum shape, and reflection spectrum width, and maintain an all-fiber structure, facilitating fusion splicing with fiber amplifiers. Therefore, they are commonly used pulse stretching devices in fiber femtosecond lasers. The pulse compressor of a femtosecond laser system generally uses a surface grating or a volume Bragg grating. While these spatial light diffraction elements provide second-order dispersion compensation, they also introduce higher-order dispersion. If this higher-order dispersion cannot compensate for the higher-order dispersion of the front-end chirped fiber grating, it will adversely affect the shape of the compressed femtosecond pulse, causing phenomena such as pulse sacs and sidelobes that reduce the pulse time-domain signal-to-noise ratio. Therefore, when designing chirped fiber grating pulse stretchers, the higher-order dispersion introduced by the pulse compressor is taken into account. The goal is to ensure that the chirped fiber grating possesses a higher-order dispersion that is opposite in magnitude and equal in amount to the higher-order dispersion introduced by the pulse compressor, thus canceling each other out and achieving the purpose of compensating for higher-order dispersion. However, chirped fiber gratings, especially those with large dispersion coefficients and wideband chirped fiber gratings, are inevitably affected by external accidental factors during the manufacturing process. For example, optical path misalignment during fiber grating etching, mechanical vibration disturbances, and fluctuations in the power of the photolithography source can all cause the etched dispersion curve to deviate slightly from the design curve. This error can lead to incomplete higher-order dispersion compensation, affecting the signal-to-noise ratio of the compressed pulse. While this small higher-order dispersion error is difficult to measure precisely during or after the manufacturing process of the chirped fiber grating, the resulting change in the shape of the compressed pulse is quite significant. It is a quality-related parameter that is difficult to inspect in advance but has a significant impact on applications.

[0105] The present invention addresses the problem of accurately locating and finding higher-order dispersion errors in chirped fiber gratings. Using a chirped pulse amplification system as a test platform, it identifies the spectral components affecting pulse compression quality through active spectral filtering, thereby achieving precise location of the source of higher-order dispersion errors. This provides accurate positioning guidance for improving the chirped fiber grating writing process and also provides targeted feedback data for actively controlling the higher-order dispersion of chirped fiber gratings online.

[0106] Once the pulse broadening and compression of the femtosecond laser chirped pulse amplification system are designed, the device parameters of the pulse compressor can be determined, and the corresponding compressor dispersion curve, including the spectral distributions of second-order and higher-order dispersions, is determined. To match the second-order and higher-order dispersions of the pulse compressor, the dispersion characteristics of the pulse broadener, especially its dispersion curve, are accordingly determined. The dispersion curve of the chirped fiber grating, as the dispersion medium, is determined by the positions and spacing of Bragg reflection points corresponding to different wavelengths etched in the fiber. A combination of equally spaced Bragg reflection points constitutes a linear chirped fiber grating, where the time delay corresponding to each adjacent reflection wavelength is consistent, i.e., it has linear group velocity delay characteristics, and the dispersion curve is a straight line with a fixed slope. However, considering the existence of higher-order dispersion, and to compensate for the higher-order dispersion introduced by the pulse compressor, the dispersion curve of the chirped fiber grating cannot be a straight line but should be a curve, and the slope of the curve may differ at different wavelength positions, meaning the change in second-order dispersion differs at different wavelength positions. The difference in the rate of change of second-order dispersion at different wavelengths, i.e. higher-order dispersion, is reflected in the pulse compression effect as the inability of certain spectral components of the pulse to achieve precise time delay with other spectral components. This results in the pulse energy of different wavelengths not being synchronized in the time domain, making it impossible for the pulse to be completely dechirped. At this time, pulse widening or pulse shape deformation can be measured.

[0107] In practical applications, the deviation of a chirped fiber grating from its design dispersion curve at a single point will not be excessive. The resulting pulse time-domain delay mismatch will only cause pulse time-domain compression errors of tens of femtoseconds or at most picoseconds, which can be clearly measured on an autocorrelator. Finely adjusting the pulse compressor of a femtosecond laser system—that is, a grating-pair compressor composed of two surface gratings—will reveal a compressed pulse with a base or side lobes when the autocorrelator displays the narrowest achievable pulse. This is caused by the mismatch in the dispersion curves of the pulse stretcher and the pulse compressor. Inserting a narrowband filter into the optical path of the grating-pair pulse compressor at this point can block diffraction spectral components of several nanometers, while other spectral components continue to propagate along the original optical path without interference. This narrowband filter is an opaque medium that can be moved laterally along the diffracted light waves propagating side-by-side. Moving it to different positions blocks different spectral components, and the lateral width of the filter determines the spectral width that can be blocked each time. By gradually moving the filter to block different diffraction spectral components, the shape of the compressed pulse measured by the autocorrelation instrument will change accordingly. When those spectral components that deviate from the designed dispersion curve are blocked, the resulting pulse shape distortion is eliminated, and the pulse shape is improved. After identifying these spectral components that degrade the compressed pulse shape, the accuracy of the optical path calibration corresponding to that spectral component and the surface quality of optical components in the chirped fiber grating manufacturing process can be checked. This allows for targeted investigation of the causes of the fiber grating dispersion curve deviating from the design value, thus helping to quickly locate the fault point and improve the fiber grating writing process. This measurement feedback data can also be applied to real-time tuning of chirped fiber gratings. After determining the spectral component corresponding to the imperfect dispersion point, stress is applied to the grating region where that spectral component is located on the chirped fiber grating, changing the dispersion curve of the chirped fiber grating online in real time to correct it and bring it closer to the design value. This achieves fine-tuning of higher-order dispersion and matching with the higher-order dispersion of the compressor.

[0108] This invention uses a chirped pulse amplification system as a test platform, in which a chirped fiber grating (GFG) is installed as a pulse stretcher. By introducing an adjustable active filter element into the pulse compression optical path, the spectral components of the pulse are scanned and filtered. Simultaneously, the corresponding changes in the time-domain shape of the compressed pulse are observed, identifying the spectral components affecting the pulse shape or signal-to-noise ratio (SNR). This identifies the spectral reflection region on the chirped fiber grating where the higher-order dispersion deviates from the design value. Therefore, it is a test method for accurately locating error points in the higher-order dispersion curve of the chirped fiber grating. After identifying the spectral components causing the error, the etching process of that spectral region in the fiber grating can be reviewed, quickly pinpointing the cause of the error and providing accurate guidance for improving the photolithography process. It can also assist in online real-time feedback stress modulation of that spectral reflection grating region in the chirped fiber grating, correcting the higher-order dispersion error of the pulse stretcher, achieving good matching compensation between the higher-order dispersion of the chirped fiber grating pulse stretcher and the higher-order dispersion of the pulse compressor, and outputting a compressed femtosecond pulse with a high SNR.

[0109] The experimental setup is a standard chirped pulse amplification system, including a seed source, pulse stretcher, laser amplifier, and pulse compressor. Testing instruments include an autocorrelation analyzer and a spectrometer. The auxiliary measurement device is a narrowband filter with sliding translation capability. If real-time feedback of the chirped fiber grating in the stretcher is required, a feedback signal reading control circuit and a stress application device are also needed.

[0110] First, the standard optical path of the chirped pulse amplification system was constructed. The laser pulse output from the seed source has a center wavelength of 1030nm, a spectral width of 22nm, and a pulse width of 500fs. It enters the chirped fiber grating pulse stretcher via fiber optic fusion splicing. The pulse stretcher includes a chirped fiber grating, an optical circulator, and a fiber grating stress adjustment device. The chirped fiber grating has a center wavelength of 1030nm, a reflection bandwidth of 20nm, a reflectivity of 70%, a second-order dispersion coefficient of 56ps², and a third-order dispersion coefficient of -0.85ps³. Its dispersion characteristics are shown by the dashed line in the dispersion curve diagram, while the solid curve is the design curve that matches the dispersion characteristics of the compressor. The middle port of the optical circulator is fused to the chirped fiber grating, the front port is connected to the seed source, and the rear port outputs the stretched pulse. After the stretcher, it is fused to an optical fiber amplifier, which can be a single-stage or multi-stage optical fiber amplifier. The output fiber of the final-stage fiber amplifier is connected to a collimator, allowing the amplified laser beam to be collimated and incident into a pulse compressor consisting of two face gratings. The face gratings are reflective gratings, 150×30mm in size, with a diffraction center wavelength of 1030nm and a groove density of 1700L / mm. The pulse width of the compressor output is monitored using an autocorrelator, and the narrowest possible pulse is carefully adjusted. If the dispersion of the chirped fiber grating does not perfectly match the dispersion of the pulse compressor, the compressed pulse will exhibit a base or side lobes. In this case, an adjustable narrowband filter is inserted laterally into the diffraction path of the pulse compressor. This narrowband filter is a thin aluminum sheet, 1-5mm wide and >15mm high, with a natural-colored oxidized surface and a finely frosted finish. The aluminum sheet is inserted vertically, with its surface approximately perpendicular to the diffraction path, blocking a portion of the diffracted light. It can be moved laterally to block diffracted light at different positions, i.e., blocking diffracted beams of different wavelengths. The filter achieves the filtering of different spectral components of the pulse by blocking beams of different wavelengths. The filtered spectral components are no longer displayed on the autocorrelator used to measure the pulse width, and their influence on the compressed pulse is thus eliminated. During the process of moving the filter to block different spectral components, the autocorrelator monitors the shape of the compressed pulse in real time. When a certain spectral component (e.g., the wavelength 1020~1025nm) is blocked, the measured pulse shape improves, such as a smaller base and the disappearance of side lobes. This confirms that the dispersion curve at the Bragg reflection point of the chirped fiber grating in the stretcher corresponding to the blocked spectral component differs significantly from the design value (corresponding to...). Figure 3The two curves in the dispersion curve diagram shown are in a mismatched region, which fails to perfectly compensate for the higher-order dispersion introduced by the compressor.

[0111] There are two ways to improve this. One is to examine the optical path and lenses on the chirped fiber grating writing platform corresponding to the spectral component, checking for deviations in the lithography optical path at the Bragg reflection point, and for damage or contamination on the surface of the optical components in the optical path. After identifying the fault, the lithography process is improved accordingly to eliminate errors in the fiber grating writing process. The second method is to feed the spectral component information back to the stress application device of the chirped fiber grating, and adjust the stress at the grating region position corresponding to the spectral component accordingly, thereby changing the grating dispersion curve with stress. This method first relies on the known distribution of the pulse spectrum along the fiber grating region. After measuring the pulse distortion caused by a certain spectral component, the precise position of the grating region corresponding to that spectral component is retrieved. The stress device (such as piezoelectric ceramic or TEC thermoelectric controller) on that grating region is adjusted to apply stress to change the dispersion curve of the grating region at that position. At the same time, an autocorrelation instrument is used to monitor the changing trend of the compressed pulse shape. The stress direction and intensity are repeatedly adjusted according to the trend of pulse shape improvement or deterioration until the optimal compressed pulse shape is obtained.

[0112] Please see Figure 4 As shown, Figure 4 This is a schematic diagram of the structure of the apparatus 200 for diagnosing and correcting the dispersion distribution of chirped fiber Bragg gratings provided in this application embodiment. The apparatus 200 is used to perform the steps of the methods for diagnosing and correcting the dispersion distribution of chirped fiber Bragg gratings shown in the above embodiments. The apparatus 200 can be a single server or a server cluster, or it can be a terminal, such as a handheld terminal, a laptop computer, a wearable device, or a robot.

[0113] like Figure 4 As shown, the apparatus 200 for diagnosing and correcting the dispersion distribution of chirped fiber gratings includes:

[0114] System building unit 201 is used to build a chirped pulse amplification system. The chirped pulse amplification system includes a seed source, a chirped fiber grating as a pulse stretcher, an fiber amplifier, a pulse compressor composed of spatial light diffraction elements, and an autocorrelator for monitoring the time-domain shape of the compressed pulse. The pulse compressor is adjusted so that the autocorrelator displays the compressed pulse at its narrowest pulse width under the current state. If the compressed pulse exhibits time-domain shape distortion, it is determined that there is a mismatch between the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor. The time-domain shape distortion includes the base and side lobes.

[0115] The waveplate insertion unit 202 is used to insert a laterally movable narrowband filter into the diffraction optical path of the pulse compressor. The blocking direction of the narrowband filter is perpendicular to the distribution direction of the diffraction spectrum. By laterally moving the narrowband filter, different wavelength spectral components are blocked in sequence, and the time-domain shape of the compressed pulse after blocking different spectral components is monitored in real time by the autocorrelation instrument.

[0116] The distribution correction unit 203 is used to determine the corresponding spectral component as the error spectral component that causes the dispersion distribution mismatch based on the time-domain shape of the compressed pulse; and to perform chirped fiber grating dispersion distribution correction on the grating region in the chirped fiber grating corresponding to the error spectral component based on the determined error spectral component.

[0117] It should be noted that those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the apparatus and modules for diagnosing and correcting the dispersion distribution of chirped fiber gratings described above can be referred to the corresponding process in the method embodiments for diagnosing and correcting the dispersion distribution of chirped fiber gratings described above, and will not be repeated here.

[0118] The aforementioned method for diagnosing and correcting the dispersion distribution of chirped fiber gratings can be implemented as a computer program, which can be used in, for example... Figure 4 It runs on the device shown.

[0119] Please see Figure 5 , Figure 5 This is a schematic block diagram of the stress unit controller provided in an embodiment of this application. The stress unit controller includes a processor, a memory, and a network interface connected via a device bus, wherein the memory may include a storage medium and internal memory.

[0120] The storage medium may store operating devices and computer programs. The computer program includes program instructions that, when executed, cause the processor to perform any method for diagnosing and correcting the dispersion distribution of chirped fiber gratings.

[0121] The processor provides computing and control capabilities to support the operation of the entire stress unit controller.

[0122] Internal memory provides an environment for the execution of computer programs in non-volatile storage media. When executed by a processor, the computer program enables the processor to perform any method for diagnosing and correcting the dispersion distribution of chirped fiber gratings.

[0123] This network interface is used for network communication, such as sending assigned tasks. Those skilled in the art will understand that... Figure 5The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the terminal to which the present application is applied. The specific stress unit controller may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0124] It should be understood that the processor can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, a general-purpose processor can be a microprocessor or any conventional processor.

[0125] In one embodiment, the processor is configured to run a computer program stored in memory to perform the following steps:

[0126] A chirped pulse amplification system is constructed, comprising a seed source, a chirped fiber grating as a pulse stretcher, an fiber amplifier, a pulse compressor composed of spatial light diffraction elements, and an autocorrelator for monitoring the time-domain shape of the compressed pulse. The pulse compressor is adjusted so that the autocorrelator displays the compressed pulse at its narrowest pulse width under the current state. If the compressed pulse exhibits time-domain shape distortion, it is determined that there is a mismatch between the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor. The time-domain shape distortion includes the base and side lobes.

[0127] A laterally movable narrowband filter is inserted into the diffraction optical path of the pulse compressor. The blocking direction of the narrowband filter is perpendicular to the distribution direction of the diffraction spectrum. By laterally moving the narrowband filter, different wavelength spectral components are blocked in sequence, and the time-domain shape of the compressed pulse after blocking different spectral components is monitored in real time using the autocorrelation instrument.

[0128] The corresponding spectral component is determined as the error spectral component that causes the dispersion distribution mismatch based on the time-domain shape of the compressed pulse; based on the determined error spectral component, the chirped fiber grating dispersion distribution is corrected in the grating region corresponding to the error spectral component in the chirped fiber grating.

[0129] In some embodiments, the step of correcting the chirped fiber grating dispersion distribution in the grating region corresponding to the determined error spectral component based on the determined error spectral component includes: checking the accuracy of optical path calibration and the state of optical components corresponding to the error spectral component during the chirped fiber grating writing process; improving the writing process for the identified deviations or anomalies; or applying stress to the grating region through a stress application device and adjusting the dispersion curve of the chirped fiber grating corresponding to the error spectral component in real time until the time-domain shape of the compressed pulse displayed by the autocorrelator meets the preset requirements.

[0130] In some embodiments, the construction of the chirped pulse amplification system includes: connecting the seed source output end to the front port of an optical fiber circulator via optical fiber fusion splicing; connecting the middle port of the optical fiber circulator to a chirped fiber grating via optical fiber; connecting the rear port of the optical fiber circulator to the input end of an optical fiber amplifier via optical fiber; connecting the output end of the optical fiber amplifier to an optical fiber collimator via optical fiber; aligning the output end of the optical fiber collimator with a pulse compressor composed of a pair of gratings; providing a laterally movable narrowband filter in the diffraction path of the pulse compressor; and connecting the output optical path of the pulse compressor to the detection end of an autocorrelator.

[0131] In some embodiments, adjusting the pulse compressor so that the autocorrelator displays the compressed pulse at its narrowest pulse width in the current state includes: adjusting the relative angle or spacing of the two surface gratings in the pulse compressor and monitoring the pulse width of the compressed pulse displayed by the autocorrelator until the pulse width value displayed by the autocorrelator reaches the minimum value in the current state, and the pulse width no longer narrows when the adjustment continues.

[0132] In some embodiments, determining that the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor are mismatched if the compressed pulse exhibits time-domain shape distortions such as base and side lobes includes: comparing the time-domain curve of the compressed pulse measured by the autocorrelator with the ideal undistorted pulse curve; if there is a continuous low-intensity signal below the main peak of the pulse or symmetrical or asymmetrical low-intensity peaks appear on both sides of the main peak, then it is determined that there is a higher-order dispersion mismatch between the two dispersion distributions.

[0133] In some embodiments, the step of sequentially blocking spectral components of different wavelengths by laterally moving the narrowband filter includes: the width direction of the narrowband filter is perpendicular to the wavelength distribution direction of the spectrum in the diffraction optical path of the pulse compressor, and the filter is gradually slid along the width direction with a preset step size, and fixed at the current position after each slide, so that the filter sequentially blocks beams of different wavelength ranges in the diffraction spectrum, and the spectral width of a single block is 1-5 nm; the preset step size includes a displacement amount corresponding to 0.1-5 nm.

[0134] In some embodiments, the step of using the autocorrelator to monitor the time-domain shape of the compressed pulse after blocking different spectral components in real time includes: after the narrowband filter blocks a specific spectral component each time, keeping the system stable for at least one pulse cycle, triggering the autocorrelator to collect the time-domain curve of the current compressed pulse, and recording and storing the characteristic parameters of the main peak width, base intensity and sidelobe amplitude of the time-domain curve.

[0135] In some embodiments, determining the corresponding spectral component as the error spectral component causing the dispersion distribution mismatch based on the time-domain shape of the compressed pulse includes: comparing the time-domain curves after blocking different spectral components; when a certain spectral component is blocked, the base intensity of the compressed pulse is significantly reduced or the side lobes completely disappear, and the main peak width is close to the ideal value, and determining that the corresponding blocked spectral component is the error spectral component causing the dispersion distribution mismatch.

[0136] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the steps of the method for diagnosing and correcting the dispersion distribution of chirped fiber gratings as provided in any embodiment of this application.

[0137] The computer-readable storage medium can be an internal storage unit of the stress unit controller described in the foregoing embodiments, such as a hard disk or memory of the stress unit controller. Alternatively, the computer-readable storage medium can be an external storage device of the stress unit controller, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the stress unit controller.

[0138] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for diagnosing and correcting the dispersion distribution of chirped fiber gratings, characterized in that, include: A chirped pulse amplification system is constructed, comprising a seed source, a chirped fiber grating as a pulse stretcher, an optical fiber amplifier, a pulse compressor composed of spatial light diffraction elements, and an autocorrelator for monitoring the time-domain shape of the compressed pulse. The pulse compressor is adjusted so that the autocorrelator displays the narrowest pulse width of the compressed pulse in the current state. If the compressed pulse exhibits temporal shape distortion, it is determined that there is a mismatch between the dispersion distribution of the chirped fiber grating and the dispersion distribution of the pulse compressor. The temporal shape distortion includes the base and side lobes. A laterally movable narrowband filter is inserted into the diffraction optical path of the pulse compressor. The blocking direction of the narrowband filter is perpendicular to the distribution direction of the diffraction spectrum. By laterally moving the narrowband filter, different wavelength spectral components are blocked sequentially. This includes: the width direction of the narrowband filter is perpendicular to the wavelength distribution direction of the spectrum in the diffraction optical path of the pulse compressor; the filter is slid gradually along this width direction with a preset step size; after each slide, it is fixed at the current position, so that the filter blocks beams of different wavelength ranges in the diffraction spectrum sequentially. The spectral width blocked in a single block is 1-5 nm; the preset step size includes a displacement of 0.1-5 nm; and the time-domain shape of the compressed pulse after blocking different spectral components is monitored in real time using an autocorrelation instrument. Based on the time-domain shape of the compressed pulse, the corresponding spectral component is determined as the error spectral component causing the dispersion distribution mismatch. Based on the determined error spectral component, the chirped fiber grating dispersion distribution is corrected in the grating region corresponding to the error spectral component. This includes: checking the accuracy of the optical path calibration and the state of the optical components corresponding to the error spectral component during the chirped fiber grating writing process; improving the writing process for the identified deviations or anomalies; or applying stress to the grating region through a stress application device and adjusting the dispersion curve of the chirped fiber grating corresponding to the error spectral component in real time until the time-domain shape of the compressed pulse displayed by the autocorrelator meets the preset requirements. This method utilizes GAN-generated ideal, distortion-free pulses as a benchmark, combined with an unsupervised anomaly detection algorithm to automatically identify error spectral components, suitable for diagnosing unknown types of dispersion errors. The GAN ideal pulse generation includes: a generator G that takes a random noise vector as input and outputs a simulated ideal autocorrelation curve; a discriminator D that distinguishes between the real and generated curves, using adversarial training to make G generate an ideal curve close to the actual system; training data collected from compressed pulse data of defect-free chirped fiber gratings of the same type as real samples; and unsupervised anomaly detection by calculating the Wasserstein distance between the time-domain curve of each acquired occluded spectrum and the ideal curve generated by the GAN. An isolated forest algorithm is used to refine the Wasserstein distance. Anomaly detection is performed based on the distance to the corresponding WD value. If the WD value of a wavelength region exceeds 3 times the standard deviation, it is identified as an error spectral component. Combining the grating dispersion model, the abnormal wavelength is mapped to a specific grating region, including the grating region with the highest WD value of λ=1060nm corresponding to x=8cm. For the detected abnormal grating region, the adaptive correction strategy calls the optimal stress adjustment parameters for similar errors in the historical database, including the PZT voltage of 4.2V corresponding to λ=1060nm, to achieve rapid compensation. If historical data is missing, the gradient descent algorithm is started to optimize the stress parameters with the goal of minimizing the WD value, with a step size α=0.1nm and ≤20 iterations.

2. The method according to claim 1, characterized in that, The construction of the chirped pulse amplification system includes: The seed source output is connected to the front port of the fiber optic circulator via fiber optic fusion splicing. The middle port of the fiber optic circulator is fused to a chirped fiber grating, and the rear port is connected to the input of the fiber optic amplifier via fiber optic cable. The output end of the fiber amplifier is connected to the fiber collimator via an optical fiber. The output end of the fiber collimator is aligned with a pulse compressor consisting of a pair of gratings. A narrow-band filter that can move laterally is provided in the diffraction optical path of the pulse compressor. The output optical path of the pulse compressor is connected to the detection end of the autocorrelator.

3. The method according to claim 1, characterized in that, Adjusting the pulse compressor so that the autocorrelator displays the compressed pulse at its narrowest pulse width under the current state includes: By adjusting the relative angle or spacing of the two surface gratings in the pulse compressor and monitoring the pulse width of the compressed pulse displayed by the autocorrelation meter, the pulse width value displayed by the autocorrelation meter reaches the minimum value of the current state, and the pulse width no longer narrows when the adjustment continues.

4. The method according to claim 1, characterized in that, If the compressed pulse exhibits temporal shape distortion, determining that the dispersion distribution of the chirped fiber grating does not match the dispersion distribution of the pulse compressor includes: Compare the compressed pulse time-domain curve measured by the autocorrelation instrument with the ideal undistorted pulse curve. If there is a continuous low-intensity signal below the main peak of the pulse or symmetrical or asymmetrical low-intensity peaks appear on both sides of the main peak, it is determined that there is a high-order dispersion mismatch in the dispersion distribution of the two.

5. The method according to claim 1, characterized in that, The method of using the autocorrelator to monitor the time-domain shape of the compressed pulse after blocking different spectral components in real time includes: After the narrowband filter blocks a specific spectral component each time, the system is kept stable for at least one pulse cycle. The autocorrelator is then triggered to acquire the time-domain curve of the current compressed pulse, and the characteristic parameters of the main peak width, base intensity, and sidelobe amplitude of the time-domain curve are recorded and stored.

6. The method according to claim 1, characterized in that, The step of determining the corresponding spectral component as the error spectral component causing the dispersion distribution mismatch based on the time-domain shape of the compressed pulse includes: By comparing the time-domain curves after blocking different spectral components, when a certain spectral component is blocked, the base intensity of the compressed pulse is significantly reduced or the side lobes disappear completely, and the width of the main peak is close to the ideal value. The blocked spectral component is determined to be the error spectral component that causes the mismatch in dispersion distribution.

7. A stress unit controller, characterized in that, The stress unit controller includes a memory and a processor; The memory is used to store computer programs; The processor is configured to execute the computer program and, in executing the computer program, implement the method as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, causes the processor to implement the method as described in any one of claims 1 to 6.

Images