Intelligent regulation and control lock loss prevention cooperative output ultrafast laser system and control method thereof

Abstract

The invention relates to the technical field of lasers, and provides an intelligent regulation and control lock-losing prevention cooperative output ultrafast laser system and a control method thereof, and the intelligent regulation and control lock-losing prevention cooperative output ultrafast laser system comprises a pumping module which is used for sending pumping light; the first beam splitter is arranged on a light path of the pump light, and a first input end of the first beam splitter is connected with the pump module; the input end of the dispersion filtering assembly is connected with the first output end of the first beam splitter, and the dispersion filtering assembly is used for performing dispersion compensation to output laser; the first pulse width adjusting assembly and the second pulse width adjusting assembly are respectively connected with different output ends of the dispersion filtering assembly; the laser is used for outputting first laser and second laser. According to the intelligent regulation and control anti-lock-losing collaborative output ultrafast laser system, the flexibility, the process development efficiency and the adaptability of multi-pulse-width pulse output are remarkably improved, and the differentiated requirements of different high-precision application scenes are met.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of lasers, in particular to an intelligent control and regulation anti-locking cooperative output ultrafast laser system and a control method thereof. BACKGROUND

[0002] An ultrafast laser refers to a device capable of generating laser pulses with a pulse width in the order of picoseconds to femtoseconds. Due to its extremely short pulse duration and extremely high peak power, the ultrafast laser can realize a "cold processing" effect in the process of interaction with matter, that is, to realize ultra-fine processing of materials without generating a heat affected zone.

[0003] Conventional industrial-grade ultrafast lasers are mostly designed with a single output port, and the adjustable range of the output pulse width and the central wavelength is narrow, which is difficult to adapt to scenarios such as precision micro-processing, biomedical imaging, scientific research, and different diversified processing requirements of processing materials. In order to realize the output of multi-pulse-width lasers, multiple pulse-width output ports can be provided on the laser, but such multi-pulse-width output lasers are usually limited to adjustment within the same time scale, and cannot realize the simultaneous output of femtosecond and picosecond cross-scale pulse widths. Moreover, the pulse width adjustment mode is rough and has low precision. If a user needs to switch between femtosecond and picosecond pulses of different orders of magnitude, the laser seed source or even the entire laser needs to be replaced, which not only is cumbersome to operate and has high equipment cost, but also seriously reduces the process development efficiency and production flexibility. SUMMARY

[0004] The intelligent control and regulation anti-locking cooperative output ultrafast laser system and the control method thereof provided by the application improve the laser process development efficiency and flexibility.

[0005] The intelligent control and regulation anti-locking cooperative output ultrafast laser system provided by the first aspect of the application comprises: a pump module configured to send pump light; a first beam splitter arranged on the light path of the pump light, and a first input end of the first beam splitter being connected to the pump module; a dispersion filter assembly arranged on the side of the first beam splitter away from the pump module, an input end of the dispersion filter assembly being connected to a first output end of the first beam splitter, and the dispersion filter assembly being configured to perform dispersion compensation to output laser; a pulse width adjustment assembly comprising a first pulse width adjustment assembly and a second pulse width adjustment assembly, the first pulse width adjustment assembly and the second pulse width adjustment assembly being connected to different output ends of the dispersion filter assembly; the first pulse width adjustment assembly is configured to use a first pulse width modulation mechanism to perform dispersion compensation on the laser to output first laser; the second pulse width adjustment assembly is configured to use a second pulse width modulation mechanism to perform dispersion compensation on the laser to output second laser; wherein the first pulse width modulation mechanism and the second pulse width modulation mechanism are different, and the pulse widths of the first laser and the second laser are different.

[0006] In some possible implementation manners, the dispersion filtering assembly comprises a second beam splitter and a first grating pair module; a first input end of the second beam splitter is connected with the first output end of the first beam splitter, and a first output end of the second beam splitter is connected with the first grating pair module; the first grating pair module is configured to receive the laser and perform dispersion compensation on the laser; the dispersion filtering assembly further comprises a third beam splitter and a base module; a first input end of the third beam splitter is connected with the first grating pair module; a second output end of the second beam splitter is connected with a second input end of the third beam splitter, and the connected optical fiber between the second output end of the second beam splitter and the second input end of the third beam splitter is fixed on the base module; wherein an optical path between the first output end of the second beam splitter and the first input end of the third beam splitter through which the laser is transmitted is a first optical path, and an optical path between the second output end of the second beam splitter and the second input end of the third beam splitter through which the laser is transmitted via the connected optical fiber between the base module is a second optical path; the first optical path is different from the second optical path, so as to filter the laser.

[0007] In some possible implementation manners, the first pulse width adjusting assembly is connected with the dispersion filtering assembly, and the first pulse width adjusting assembly comprises a third fiber collimator and a second grating pair module; an input end of the third fiber collimator is connected with an output end of the third beam splitter, and the third fiber collimator is configured to collimate the laser and emit the collimated laser into the second grating pair module; the second grating pair module is configured to compensate a dispersion value of a third fiber connection section by using a first pulse width adjusting mechanism, the third fiber connection section comprising the connected optical fiber between the third beam splitter and the third fiber collimator.

[0008] In some possible implementation manners, the dispersion filtering assembly is further connected with the second input end of the first beam splitter, the second pulse width adjusting assembly is connected with the second output end of the first beam splitter, and the second pulse width adjusting assembly comprises a circulator and an adjustable fiber stretcher; a first port of the circulator is connected with the second output end of the first beam splitter; a second port of the circulator is connected with the adjustable fiber stretcher; the intelligent control anti-locking cooperative output ultrafast laser system further comprises a fourth fiber collimator; a third port of the circulator is connected with the fourth fiber collimator; wherein the adjustable fiber stretcher is configured to compensate a dispersion value of a second fiber connection section by using a second pulse width adjusting mechanism, the second fiber connection section comprising the connected optical fiber between the first beam splitter and the fourth fiber collimator.

[0009] In some possible implementation manners, the first grating pair module comprises a first grating pair and a first displacement table; a first grating in the first grating pair is arranged on the first displacement table, and the first displacement table is configured to drive the first grating to move along a normal direction of the first grating, so as to adjust a first grating pair spacing of the first grating pair; the first grating pair module is further configured to perform dispersion compensation on the first grating pair spacing.

[0010] The control method of the intelligent control anti-locking cooperative output ultrafast laser system provided in the second aspect of the application is applied in the intelligent control anti-locking cooperative output ultrafast laser system, and the control method of the intelligent control anti-locking cooperative output ultrafast laser system comprises the following steps: determining the corresponding relationship between the stretching amount of the base module and the center wavelength based on the first grating pair spacing set in the dispersion filter assembly, the first stretching amount set of the base module in the dispersion filter assembly, and the first center wavelength set in the dispersion filter assembly; the first center wavelength in the first center wavelength set is in one-to-one correspondence with the first stretching amount in the first stretching amount set; the optical path difference in the dispersion filter assembly is compensated by using the stretching amount, and the optical path difference is controlled to be constant; the dispersion value of the first optical fiber connection section is compensated based on the second grating pair spacing set in the first pulse width adjusting assembly and the corresponding first pulse width set; the dispersion value of the second optical fiber connection section is compensated based on the second pulse width set of the second pulse width adjusting assembly and the dispersion value set of the adjustable fiber stretcher in the second pulse width adjusting assembly; a data set is established; the data set comprises a multi-dimensional feature vector, and the multi-dimensional feature vector comprises a first target center wavelength, a first target pulse width, a second target pulse width, a first measured center wavelength, a first measured pulse width, a second measured pulse width, a first grating pair spacing measured adjustment amount, a second grating pair spacing measured adjustment amount, a first measured stretching amount of the base module, and a measured dispersion value of the second pulse width adjusting assembly; the data set is used to train a double-pulse-width cooperative prediction model; the input of the trained double-pulse-width cooperative prediction model is the multi-dimensional feature vector, and the output is the first grating pair spacing adjustment amount, the second grating pair spacing adjustment amount, the first stretching amount of the base module, the pump light power adjustment amount, and the dispersion value adjustment amount of the second pulse width adjusting assembly; the double-pulse-width cooperative prediction model is optimized by using a Bayesian optimization algorithm.

[0011] In some feasible implementations, within the stable mode-locked pumping power interval of the pump light, the first grating pair spacing set in the first grating pair, the first stretching amount set of the base module, and the first center wavelength set corresponding to the first stretching amount set are collected; the first grating pair spacing adjustment amount is determined based on the first grating pair spacing set; the corresponding relationship between the stretching amount and the center wavelength is determined based on the first grating pair spacing set, the first center wavelength set, and the first stretching amount set; the absolute value of the maximum value of the first grating pair spacing adjustment amount is within a first preset range, the first stretching amount of the base module is within a second preset range, and the first preset range is different from the second preset range.

[0012] In some feasible implementations, the dispersion value of the first optical fiber connection section is compensated based on the second grating pair spacing set in the first pulse width adjusting assembly and the corresponding first pulse width set, which comprises the following steps: the negative dispersion value of the first pulse width adjusting assembly is determined based on the second grating pair spacing set in the first pulse width adjusting assembly and the corresponding first pulse width set; the dispersion value of the first optical fiber connection section is compensated based on the negative dispersion value; the absolute value of the negative dispersion value is greater than or equal to the absolute value of the dispersion value of the first optical fiber connection section.

[0013] In some possible implementation manners, the dual-pulse-width cooperative prediction model is trained by using a data set, including: constructing the dual-pulse-width cooperative prediction model, the dual-pulse-width cooperative prediction model being a nonlinear prediction model fused with an attention mechanism, the dual-pulse-width cooperative prediction model including an input layer, an attention layer, a hidden layer and an output layer connected in sequence; and training the dual-pulse-width cooperative prediction model in combination with the data set and the anti-lockout priority.

[0014] In some possible implementation manners, the dual-pulse-width cooperative prediction model is optimized by using a Bayesian optimization algorithm, including: selecting a gradient boosting tree as a surrogate model, and constructing a target function; wherein the target function includes a first target function, a second target function, a third target function and a fourth target function, the first target function being used to maximize the mode-locked stability, the second target function being used to minimize the pulse-width error of the first laser, the third target function being used to minimize the pulse-width error of the second laser, and the fourth target function being used to minimize the first center wavelength error; constructing a constraint function; in the constraint function, different weights corresponding to different mode-locked states are different; constructing a multi-objective expected improvement acquisition function fused with a target priority; in the multi-objective expected improvement acquisition function, different target priority coefficients corresponding to different mode-locked states are different; and optimizing the dual-pulse-width cooperative prediction model based on the target function, the constraint function and the multi-objective expected improvement acquisition function.

[0015] The control method for the intelligent control anti-lockout cooperative output ultrafast laser system provided in the second aspect of the application establishes a cooperative control mechanism of dispersion compensation and spectral filtering, binds the parameters of dispersion adjustment and filtering in linkage, guarantees the stability of the filtering characteristics, realizes the accurate compensation of the in-cavity dispersion, and at the same time, matches the intelligent prediction model with the mode-locked priority and the dynamic constraint multi-objective optimization algorithm, constructs the intelligent control logic of “mode-locked stability priority, pulse-width accurate adaptation”, and dynamically balances various performance indicators without manual intervention. The control method effectively solves the problems of extensive pulse-width adjustment, mode-locked instability and insufficient hardware reliability in the prior art, significantly improves the flexibility of multi-pulse-width pulse output and the process development efficiency, and meets the differentiated needs of different high-precision application scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, for those skilled in the art, other drawings can also be obtained without creative labor based on these drawings.

[0017] Figure 1 is a structural schematic diagram of an intelligent control anti-lockout cooperative output ultrafast laser system provided by an embodiment of the present application; Figure 2is a structural schematic diagram of a first grating pair module provided by an embodiment of the present application; Figure 3 is a structural diagram of an intelligent regulation and control anti-lost lock cooperative output ultrafast laser system provided by an embodiment of the present application; Figure 4 is one of flowcharts of a control method of an intelligent regulation and control anti-lost lock cooperative output ultrafast laser system provided by an embodiment of the present application; Figure 5 is the other of flowcharts of a control method of an intelligent regulation and control anti-lost lock cooperative output ultrafast laser system provided by an embodiment of the present application.

[0018] Illustration mark: 1, mode-locked module; 2, fiber polarizer; 3, gain module; 4, pump module; 5, wavelength division multiplexer; 6, first beam splitter; a1, dispersion filter assembly; 7, second beam splitter; 8, first fiber collimator; 9, first grating pair module; 901, first mirror; 902, first grating; 903, second grating; 904, first roof mirror; 905, first displacement table; 906, first photoelectric sensor; 10, second fiber collimator; 11, third beam splitter; 12, fourth beam splitter; 13, fiber isolation beam splitter; 14, first base; 15, second base; 16, circulator; 17, adjustable fiber stretcher; 18, third fiber collimator; 19, first optical beam splitter; 20, second optical beam splitter; 21, first frequency domain data acquisition card; 22, first time domain data acquisition card; 23, fourth fiber collimator; 24, second grating pair module; 25, second time domain data acquisition card; 26, second frequency domain data acquisition card; 27, third optical beam splitter; 28, fourth optical beam splitter; 29, first power data acquisition card; 30, second power data acquisition card; 31, third power data acquisition card; 32, first output window; 33, second output window; 34, algorithm control module; 35, main control module. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be clearly described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments of the present application, other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0020] Hereinafter, the terms "first", "second", and the like are used only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second", and the like can explicitly or implicitly include one or more of the features. In the description of the present application, unless otherwise stated, the meaning of "a plurality of" is two or more.

[0021] In addition, in the present application, the orientation terms such as "upper", "lower", "inner", "outer" and the like are defined with respect to the orientation of the components shown in the drawings, and it should be understood that these directional terms are relative concepts, which are used for relative description and clarification, and can be changed accordingly according to the change of the orientation of the components placed in the drawings.

[0022] Ultrafast lasers, generally refer to devices capable of generating laser pulses with pulse widths in the order of picoseconds to femtoseconds. Due to its extremely short pulse duration and extremely high peak power, ultrafast lasers can achieve "cold processing" effect in the process of interaction with matter, that is, to achieve ultra-fine processing of materials without generating almost thermal affected zone. This unique advantage makes it play an irreplaceable role in the fields of precision micro-machining, frontier scientific research, biomedical imaging and surgery, high-speed optical communication and nonlinear optical frequency comb, etc. In the development history of ultrafast lasers, fiber lasers stand out with their unique advantages. With the increasing maturity of fiber fabrication technology, the performance of mode-locked oscillators with fiber as gain medium has made great progress. Compared with traditional solid-state laser oscillators, fiber lasers exhibit excellent mechanical stability, compact size, excellent beam quality and maintenance-free or low-maintenance cost. These advantages make it more suitable for the harsh environment of industrial sites, thus gradually replacing some traditional solid-state lasers and becoming the mainstream technical solution for current ultrafast laser seed sources.

[0023] However, in the conventional industrial-grade ultrafast laser with a single output port design, only one output port is usually provided, and the adjustable range of the output pulse width and central wavelength is narrow, which is difficult to adapt to diversified application requirements. In the scenes of precision micro-machining, biomedical imaging, scientific research, etc., facing diversified processing materials such as metal, semiconductor, polymer, transparent brittle material, etc., and different processing requirements such as cutting, drilling, surface structuring, etc., as a key parameter affecting the processing precision and thermal affected zone size, the pulse width often needs to be explored and optimized to seek the best processing effect. The pulse width is one of the key parameters, which directly affects the processing precision, thermal affected zone size and nonlinear interaction efficiency. At present, when the user needs to change the pulse width, the conventional method is to replace different seed sources, or even replace the whole laser, which not only is cumbersome, time-consuming and laborious, but also seriously restricts the process development efficiency and production flexibility.

[0024] Figure 1 Figure 1 is a structural schematic diagram of an intelligent control anti-locking cooperative output ultrafast laser system provided by an embodiment of the present application.

[0025] To solve the above technical problems, referring to Figure 1 , the present application provides an intelligent control anti-locking cooperative output ultrafast laser system.

[0026] The intelligent control anti-locking cooperative output ultrafast laser system comprises a mode-locked module 1, a fiber polarizer 2, a gain module 3, a pump module 4, a wavelength division multiplexer 5, a first beam splitter 6, a dispersion filter assembly, and a pulse width adjustment assembly.

[0027] Specifically, in the intelligent control anti-locking cooperative output ultrafast laser system, the mode-locked module 1 is taken as a starting point, and the output end of the mode-locked module 1 is connected with the fiber polarizer 2, the gain module 3, and the wavelength division multiplexer 5 in sequence. The mode-locked module 1 is used to generate mode-locked pulses, so as to realize long-term stability of the intelligent control anti-locking cooperative output ultrafast laser system. The fiber polarizer 2 ensures that the laser always maintains the same polarization state transmission; the gain module 3 converts the pump light into target laser and compensates for the transmission loss, thereby improving the output energy.

[0028] The pump module 4 and the gain module 3 are both connected with the wavelength division multiplexer 5, and the pump module 4 is used to send pump light. The pump light is transmitted to the gain module 3 through the wavelength division multiplexer 5, is absorbed by the gain medium to form population inversion, and then amplifies the optical signal through stimulated radiation to generate laser.

[0029] The first beam splitter 6 is arranged on the light path of the pump light, and the first input end of the first beam splitter 6 is connected with the pump module 4 through the wavelength division multiplexer 5. The first beam splitter 6 can be used for beam splitting and beam combining.

[0030] The dispersion filter assembly is arranged on the side of the first beam splitter 6 away from the pump module 4, the input end of the dispersion filter assembly is connected with the first output end of the first beam splitter 6, and the dispersion filter assembly is configured to perform dispersion compensation to output laser.

[0031] Specifically, the dispersion filter assembly can realize the dual functions of dispersion and filtering. In this way, after the laser enters the dispersion filter assembly, not only can the dispersion value be accurately adjusted, but also the laser center wavelength adjustment can be simultaneously completed.

[0032] The pulse width adjustment assembly comprises a first pulse width adjustment assembly and a second pulse width adjustment assembly, and the first pulse width adjustment assembly and the second pulse width adjustment assembly are respectively connected with the dispersion filter assembly. The first pulse width adjustment assembly is used to perform dispersion compensation on the laser by using a first pulse width modulation mechanism to output first laser; and the second pulse width adjustment assembly is used to perform dispersion compensation on the laser by using a second pulse width adjustment mechanism to output second laser.

[0033] Specifically, after the laser is preliminarily dispersed and filtered, the dispersed laser is output by the dispersion filtering assembly and transmitted to the first pulse width adjusting assembly and the second pulse width adjusting assembly respectively, for re-dispersion adjustment. The first pulse width adjusting mechanism is different from the second pulse width adjusting mechanism, and the pulse widths of the first laser and the second laser are different. The double-path pulse width control structure can realize modulation of lasers with different pulse widths, thereby meeting the differentiated requirements of multi-scene applications for pulse width accuracy and flexibility, and realizing pulse width targeted modulation.

[0034] It should be emphasized that the "connection" or "connection" mentioned in the present application can include direct connection and indirect connection. Indirect connection can realize coupling and transmission of signals or light paths through intermediate medium or element, and the interfaces between components need to meet the requirements of optical matching and mechanical stability.

[0035] The intelligent control and loss-proof lock coordination output ultrafast laser system provided by the embodiments of the present application establishes a cooperative control mechanism of dispersion compensation and spectral filtering, links the parameters of dispersion adjustment and filtering, guarantees the stability of the filtering characteristics, realizes accurate compensation of the intracavity dispersion, effectively solves the problems of extensive pulse width adjustment, loss of mode locking stability and insufficient hardware reliability in the prior art, significantly improves the flexibility, process development efficiency and adaptability of multi-pulse width pulse output, and meets the differentiated requirements of different high-precision application scenarios.

[0036] In some possible implementation manners, the dispersion filtering assembly can include a second beam splitter 7 and a first grating pair module 9. The first input end of the second beam splitter 7 is connected to the first output end of the first beam splitter 6, and the first output end of the second beam splitter 7 is connected to the first grating pair module 9. The first grating pair module 9 is configured to receive laser and perform dispersion compensation on the laser.

[0037] Specifically, after the laser enters the dispersion filtering assembly, the laser is split by the second beam splitter 7, and one-way laser enters the first grating pair module 9 for dispersion adjustment.

[0038] In a specific implementation, the dispersion filtering assembly can further include a first optical fiber collimator 8 and a second optical fiber collimator 10. The first optical fiber collimator 8 and the second optical fiber collimator 10 are both used for spatial collimation of the light beam. The first optical fiber collimator 8 is arranged between the second beam splitter 7 and the first grating pair module 9, one end of the first optical fiber collimator 8 is connected with the second beam splitter 7, and the other end is an output end which is arranged opposite to the first grating pair module 9. That is, the first output end of the second beam splitter 7 is indirectly connected with the first grating pair module 9 through the first optical fiber collimator 8. The second optical fiber collimator 10 is arranged on the other side of the first grating pair module 9 opposite to the first optical fiber collimator 8, and the laser is incident to the second optical fiber collimator 10 after passing through the first grating pair module 9.

[0039] Figure 2 FIG. 1 is a structural schematic diagram of a first grating pair module provided by an embodiment of the present application.

[0040] In combination with Figure 1 and Figure 2 As shown in FIG. 1, on the above optical path, the laser is injected into the first grating pair module 9 in the form of spatial light. The first grating pair module 9 can include a first mirror 901, a first grating pair, a first roof mirror 904, a first displacement table 905, and a first photoelectric sensor 906. The first grating 902 in the first grating pair is arranged on the first displacement table 905, the second grating 903 in the first grating pair is arranged opposite to the first grating 902, and the first grating 902 and the second grating 903 are placed at the Brewster angle to separate the components of different wavelengths in the incident pulse in space by using the diffraction effect and introduce positive dispersion. The first mirror 901 is arranged in the optical path between the first grating 902 and the second grating 903, and the first mirror 901 is located between the first photoelectric sensor 906 and the second grating 903.

[0041] Specifically, the first grating pair is a pair of diffraction gratings, and the first displacement table 905 can drive the first grating 902 to move along the normal direction of the first grating 902, so as to adjust the first grating pair spacing in the first grating pair, so that the dispersion is compensated in the case of changing the first grating pair spacing. The light beam produces opposite group delay after twice diffraction, that is, negative dispersion, so as to offset the initial chirp of the pulse and realize time domain compression. The first roof mirror 904 is used to change the propagation direction of the light beam and introduce a height difference, so that the compressed laser is output through the first mirror 901: 99% of the energy is reflected to the subsequent optical path, and the remaining 1% of the energy is transmitted to the first photoelectric sensor 906.

[0042] In some possible implementations, the dispersion filtering assembly provides a negative dispersion value which can be calculated by the following formula (1): (1) wherein, is the first grating pair spacing, is the Brewster angle, is the speed of light in vacuum, is the grating constant, is the center wavelength of the incident light. In formula (1), the dispersion value is directly related to the first grating pair spacing. To this end, the first displacement table 905 is driven by the algorithm to move accurately along the normal direction of the first grating 902, and the dispersion value is continuously controlled by adjusting the first grating pair spacing.

[0043] In some possible implementation manners, the dispersion filtering assembly further includes a third beam splitter 11 and a base module. A first input end of the third beam splitter 11 is connected with the first grating pair module 9 through the second optical fiber collimator 10. A second output end of the second beam splitter 7 is connected with a second input end of the third beam splitter 11, and the connected optical fiber between the second output end of the second beam splitter 7 and the second input end of the third beam splitter 11 is fixed on the base module.

[0044] In the possible implementation manners, the first grating pair module 9 is connected with a first output end of the third beam splitter 11 through the second optical fiber collimator 10. The second beam splitter 7 can be a 1x2 optical fiber beam splitter.

[0045] The other laser beam split by the second beam splitter 7 is transmitted to the third beam splitter 11 after passing through the base module. The base module can include a first base 14 and a second base 15 arranged in sequence. The optical path between the first output end of the second beam splitter 7 and the first input end of the third beam splitter 11 is a first optical path L1. That is, the first optical path L1 is the optical path through which the laser passes in sequence through the second beam splitter 7, the first optical fiber collimator 8, the first grating pair module 9, the second optical fiber collimator 10, and the third beam splitter 11. The optical path between the second output end of the second beam splitter 7 and the second input end of the third beam splitter 11 through the connected optical fiber between the base module is a second optical path L2. That is, the second optical path L2 is the optical path through which the laser passes in sequence through the second beam splitter 7, the first base 14 and the second base 15, and the third beam splitter 11. The first optical path L1 is different from the second optical path L2, so as to filter the laser. The third beam splitter 11 can be a 1x2 optical fiber beam splitter.

[0046] Specifically, the end of the second beam splitter 7 and the third beam splitter 11 are directly connected, considering the intelligent regulation and anti-locking cooperation output ultrafast laser system stability, the coating layer of part of the optical fiber is removed at a position more than 10 cm away from the optical fiber fusion point, and the optical fiber is fixed on the first base 14 and the second base 15 by using ultraviolet curing glue, and the removed coating layer part is kept at the middle position of the first base 14 and the second base 15. Wherein, the first base 14 and the second base 15 can be electric bases with precise adjustment function, and the two can be controlled to move. Since the first optical path L1 is different from the second optical path L2, the laser is split by the second beam splitter 7 and recombined at the third beam splitter 11. Since the optical paths experienced by the two beams of light are different, interference will occur when they are combined, thereby realizing the function of wavelength selective filtering. The free spectral range of the filtering can be calculated by the following formula (2): (2) ; wherein, is the wavelength difference between two adjacent transmission peaks in the free spectral range, is the effective refractive index, .

[0047] is the visibility, also known as the interference fringe contrast, which is the core index for measuring the good or bad of the interference effect. When two coherent lights are superimposed in space, the total light intensity distribution can be calculated by the following formula (3): (3) ; wherein, and are the light intensities of the two lights, is the phase difference of the two lights, when changes, the total light intensity changes between the maximum value and the minimum value , the visibility V can be calculated by the following formula (4): (4) ; when = , =1, the light and dark contrast of the interference fringes is the strongest, the interference effect is the best, and the filtering ability is the strongest. Usually, a beam splitter with a 50:50 splitting ratio is used, and the deformation loss of the grating during stretching and the transmission efficiency of the grating are comprehensively considered. In order to improve the interference effect and overall efficiency of the dispersion filtering assembly, the splitting ratio of the second beam splitter 7 and the third beam splitter 11 is further limited in the embodiment, and 0.9≤ ≤1. The grating single transmission efficiency is about 96%, the overall transmission rate is about 85%, and the corresponding loss is about 0.705 dB; when the grating is stretched to the maximum length, the loss is about 0.09 dB, so the beam splitting ratio is set between 44:56 and 63:37. In the embodiment of the application, the beam splitting ratio of 55:45 is preferred, so that the upper arm is slightly more to compensate for the loss of the grating, while leaving a certain amount.

[0048] In a specific implementation, the intelligent control anti-lost lock cooperative output ultrafast laser system can further include a fourth beam splitter 12 and a fiber isolating beam splitter 13, and the fourth beam splitter 12 is arranged at the output end of the third beam splitter 11. After the laser passes through the dispersion filtering assembly, it can return to the first beam splitter 6 in sequence through the fourth beam splitter 12 and the fiber isolating beam splitter 13, forming a loop structure to ensure that the laser always transmits in the clockwise direction in the loop. It needs to be emphasized that the clockwise direction is the direction of the arrow of the loop structure shown in the figure. Figure 1

[0049] Specifically, in this implementation, the first output end of the fourth beam splitter 12 is connected with the fiber isolating beam splitter 13, and the second output end of the fourth beam splitter 12 is connected with the first pulse width adjusting assembly. The fourth beam splitter 12 can be a 1x2 fiber beam splitter.

[0050] In this implementation, continuing to refer to Figure 1 As shown in the figure, the first pulse width adjusting assembly is indirectly connected with the dispersion filtering assembly through the fourth beam splitter 12, and the first pulse width adjusting assembly can include a third fiber collimator 18 and a second grating pair module 24. The input end of the third fiber collimator 18 is connected with the output end of the third beam splitter 11 through the fourth beam splitter 12, and the third fiber collimator 18 is configured to collimate the laser and then shoot into the second grating pair module 24. The second grating pair module 24 is configured to compensate for the dispersion value of the first fiber connection section by using a first pulse width adjusting mechanism, and the first fiber connection section includes the connecting fiber between the third beam splitter 11 and the third fiber collimator 18. That is, the first fiber connection section can be the connecting fiber between the laser transmission from the dispersion filtering assembly to the third fiber collimator 18.

[0051] Specifically, the structure of the second grating pair module 24 can be the same as that of the first grating pair module 9. The second grating pair module 24 can include a second mirror, a third grating, a fourth grating, a second moving table, a second roof mirror and a second photoelectric detector (not shown in the figure). The dispersion adjustment of the laser by the second grating pair module 24 can refer to the formulas (1) to (4) shown above.

[0052] ​In some possible implementation manners, the dispersion filtering assembly can be further connected to the second input end of the first beam splitter 6 through a fiber isolator beam splitter 13, and the second pulse width adjusting assembly is connected to the second output end of the first beam splitter 6. In this way, after the laser is subjected to the dispersion and filtering operations of the dispersion filtering assembly, the laser can be transmitted to the first pulse width adjusting module through the first output end of the fourth beam splitter 12, and can also re-enter the first beam splitter 6 through the fiber isolator beam splitter 13 and be input into the second pulse width adjusting assembly through the second output end of the first beam splitter 6. The first beam splitter 6 can be a 2*2 fiber beam splitter.

[0053] The second pulse width adjusting assembly can include a circulator 16 and an adjustable fiber stretcher 17, the first port of the circulator 16 is connected to the second output end of the first beam splitter 6, the second port of the circulator 16 is connected to the adjustable fiber stretcher 17, and the intelligent control anti-locking cooperative output ultrafast laser system further includes a fourth fiber collimator 23, the third port of the circulator 16 is connected to the fourth fiber collimator 23, and the adjustable fiber stretcher 17 is configured to compensate for the dispersion value of a second fiber connection section by using a second pulse width adjusting mechanism, and the compensated laser is output through the third port of the circulator 16, the second fiber connection section includes the connected optical fibers between the first beam splitter 6 and the fourth fiber collimator 23.

[0054] In this way, by arranging the first pulse width adjusting assembly and the second pulse width adjusting assembly, after the laser is output from the dispersion filtering assembly, double-path dispersion compensation can be achieved to ensure that the laser of the required pulse is output.

[0055] In the embodiment of the application, the first laser can be a femtosecond laser, and the second laser can be a picosecond laser, so that the stable operation of the mode-locked state can be maintained while the picosecond or femtosecond pulse is stably output by the ultrafast laser. The establishment and maintenance of the mode-locked state require that the net dispersion in the cavity be controlled within a suitable range, and under this condition, the nonlinear effect and the dispersion can achieve dynamic balance, so that the narrow pulse output under the transform limit is formed and maintained. Because the pulse laser is easily affected by the group velocity dispersion when transmitted in the optical fiber, the propagation speeds of different frequency components are different, which causes the pulse to be widened, and therefore the first grating pair module 9 provides adjustable negative dispersion to compensate for the positive dispersion introduced by the optical fiber in the resonant cavity.

[0056] Figure 3 FIG. 1 is a structural diagram of an intelligent control anti-locking cooperative output ultrafast laser system provided by an embodiment of the application.

[0057] Referring to FIG. 1, Figure 3 As shown in FIG. 1, the first laser is a femtosecond laser, and the second laser is a picosecond laser, and for the purpose of achieving stable mode locking and outputting the target pulse width, the following conditions need to be met wherein D represents the dispersion value, D ABCDEThe dispersion value provided for all the optical fibers in the resonant cavity of the dispersion filtering assembly, D1 is the maximum dispersion value that the first grating pair module 9 can provide, that is, the dispersion provided by the first grating pair module 9 substantially offsets the total dispersion of the optical fibers in the cavity, so that the net dispersion of the system approaches zero. Specifically, the first grating pair spacing Lg of the first grating pair when providing the required amount of negative dispersion can be determined by formula (1) i When the first grating pair spacing is adjusted to the appropriate value, the net dispersion of the entire intelligent control anti-loss collaborative output ultrafast laser system will tend to zero (i.e., near-zero dispersion state), at which time the mode-locked operation is most stable, laying the foundation for achieving the target pulse output.

[0058] Since the laser passes through the CDF segment optical fiber and is additionally provided with a certain amount of dispersion, in order to obtain the target femtosecond pulse output, it is required that , wherein D CDF is the dispersion value provided by the CDF segment optical fiber, and D2 is the maximum dispersion value that the second grating pair module 24 can provide. The laser passes through the first port to the second port of the circulator 16 to reach the adjustable optical fiber stretcher 17, is stretched after being provided with a certain dispersion value, and is output by the third port of the circulator 16. D BGH is the dispersion value provided by the BGH segment optical fiber, D3 is the dispersion value provided by the adjustable optical fiber stretcher 17, The second-order dispersion amount introduced can be calculated by the following formula: (5) , wherein is the pulse width of the directly output picosecond in the cavity, is the target picosecond pulse width, and finally, the target picosecond pulse is output through the second output window 33.

[0059] In some possible implementation manners, continuing to refer to Figure 1 , the intelligent control anti-loss collaborative output ultrafast laser system can further include: a first output window 32 and a second output window 33, the first output window 32 is arranged on the output light path of the second grating pair module 24, and is used to output the first laser; and the second output window 33 is arranged on the output light path of the fourth optical collimator 23, and is used to output the second laser.

[0060] In some possible implementation manners, the intelligent control anti-loss collaborative output ultrafast laser system can further include: a first optical beam splitter 19, a second optical beam splitter 20, a first frequency domain data acquisition card 21, and a first time domain data acquisition card 22.

[0061] Specifically, the first optical beam splitter 19 is arranged between the second grating pair module 24 and the first output window 32, and the first laser can be split into two paths by the first optical beam splitter 19, one of which is transmitted to the first output window 32 through the first optical beam splitter 19, and the other is reflected to the second optical beam splitter 20 by the first optical beam splitter 19 and transmitted to the first frequency domain data acquisition card 21 and the first time domain data acquisition card 22 respectively by the second optical beam splitter 20, for real-time monitoring of spectral characteristics.

[0062] In some possible implementation manners, the intelligent regulation and control anti-lost lock cooperative output ultrafast laser system can further include a third optical beam splitter 27, a fourth optical beam splitter 28, a second frequency domain data acquisition card 26 and a second time domain data acquisition card 25.

[0063] Specifically, the third optical beam splitter 27 is arranged between the fourth fiber collimator 23 and the second output window 33, and the second laser can be split into two paths by the third optical beam splitter 27, one of which is transmitted to the second output window 33 through the third optical beam splitter 27, and the other is reflected to the fourth optical beam splitter 28 by the third optical beam splitter 27 and transmitted to the second frequency domain data acquisition card 26 and the second time domain data acquisition card 25 respectively by the fourth optical beam splitter 28, for real-time monitoring of spectral characteristics.

[0064] In some possible implementation manners, the intelligent regulation and control anti-lost lock cooperative output ultrafast laser system can further include a first power data acquisition card 29, a second power data acquisition card 30 and a third power data acquisition card 31.

[0065] Specifically, the first power data acquisition card 29 is connected with the fiber isolation beam splitter 13, for detecting the power value of the laser output by the laser. The second power data acquisition card 30 is connected with the first grating pair module 9, for detecting the power value of the laser inside the dispersion filter assembly. The third power data acquisition card 31 is connected with the second grating pair module 24, for detecting the power value of the laser inside the second pulse width assembly.

[0066] In some possible implementation manners, the intelligent regulation and control anti-lost lock cooperative output ultrafast laser system can further include an algorithm control module 34 and a main control module 35.

[0067] The algorithm control module 34 is connected with the first frequency domain data acquisition card 21, the first time domain data acquisition card 22, the second frequency domain data acquisition card 26, the second time domain data acquisition card 25, the first power data acquisition card 29, the second power data acquisition card 30, the third power data acquisition card 31, the first base 14, the second base 15 and the adjustable optical fiber stretcher 17, the first grating pair module 9 and the second grating pair module 24 respectively, and the algorithm control module 34 is used for receiving data provided by the acquisition cards and executing an optimization algorithm and control logic. The main control module 35 is connected with the algorithm control module 34 and the pump module 4 respectively, and the main control module 35 is used for receiving instructions of an upper computer and coordinating the modules to work cooperatively.

[0068] In a specific implementation, when the pump light power reaches a mode locking threshold P1, for example, 53 mW, the output pulse laser is generated. As the pump power continues to increase to P n , for example, 98 mW, the pulse presents a split state, and thus the working interval of the stable mode locking corresponds to the pump power . Generally, the pump power point slightly higher than the mode locking threshold is selected for mode locking, for example, 58 mW, but when the central wavelength or the pulse width is tuned, the degree of intracavity loss changes, the first power data acquisition card 29 records the laser output power in real time, when the output power fluctuation exceeds 3% of the average output power, the main control module 35 starts to adjust the current of the pump module 4 to ensure that the output power stability rms<0.5%, for example, the output power is 2.4 mW, and the output power stability rms≈0.32%.

[0069] The intelligent regulation and control anti-lost-lock cooperative output ultrafast laser system provided by the embodiment of the application adopts an integrated cooperative architecture, and production costs are effectively reduced. Through linkage of the dispersion filtering assembly and the pulse width adjusting assembly, multiple independent laser generating units do not need to be arranged, multiple pulse widths in a single system are flexibly regulated and controlled, the structure is more compact, and no additional optical path switching component is needed. In addition, the stability fluctuation caused by hardware switching can be avoided through an intelligent algorithm, and the accuracy and adaptability of pulse width tuning are better.

[0070] Corresponding to the foregoing embodiment of the intelligent regulation and control anti-lost-lock cooperative output ultrafast laser system, the application further provides a control method of the intelligent regulation and control anti-lost-lock cooperative output ultrafast laser system.

[0071] Figure 4 is one of flowcharts of the control method of the intelligent regulation and control anti-lost-lock cooperative output ultrafast laser system provided by the embodiment of the application.

[0072] Referring to Figure 4 , the control method of the intelligent regulation and control anti-lost-lock cooperative output ultrafast laser system provided by the embodiment of the application can include the following steps S1 to S7.

[0073] Step S1: determining the correspondence between the stretch amount and the center wavelength of the base module based on the first set of grating pair spacings in the dispersion filter assembly, the first set of stretch amounts of the base module in the dispersion filter assembly, and the first set of center wavelengths in the dispersion filter assembly; the first center wavelength in the first set of center wavelengths corresponds to the first stretch amount in the first set of stretch amounts one by one. Wherein the stretch amount of the base module refers to the stretch amount generated during the pulling of the optical fiber between the first base and the second base in the base module.

[0074] Step S1 can include steps S11 to S13.

[0075] Step S11: within the stable mode-locked pumping power interval of the pump light, collecting the first set of grating pair spacings in the first grating pair, the first set of stretch amounts of the base module in the dispersion filter assembly, and the first set of center wavelengths corresponding to the first set of stretch amounts.

[0076] In this step, within the stable mode-locked pumping power interval , the system historical operation parameters are collected, including the first set of grating pair spacings in the first grating pair , the first set of stretch amounts of the base module , and the first set of center wavelengths corresponding to the first stretch amounts .

[0077] Step S12: determining the first grating pair spacing adjustment amount based on the first set of grating pair spacings.

[0078] The first grating pair spacing step value , and the first grating pair spacing adjustment amount .

[0079] Step S13: determining the correspondence between the stretch amount and the center wavelength based on the first set of grating pair spacings, the first set of center wavelengths, and the first set of stretch amounts; wherein the absolute value of the maximum value of the first grating pair spacing adjustment amount is within the first preset range, the first stretch amount is within the second preset range, and the first preset range is different from the second preset range.

[0080] Determine the near-zero dispersion point by the dispersion compensation principle, satisfy , calculate the deviation | of the current first grating pair spacing and the near-zero dispersion point. Based on the collected parameters, calculate the average change amount of the center wavelength corresponding to the unit stretch amount , to establish the quantitative correspondence between the stretch amount and the center wavelength: (6) wherein, and It is the change in two adjacent center wavelengths, the optical path difference between the two arms of the dispersive filter component. Combining with formula (2), the free spectral range (FSR) can be adjusted by changing the stretching amount ΔL. This step aims to establish a quantitative correspondence between the spacing, stretching amount, center wavelength, and FSR of the first grating pair using historical parameters, clarify hardware safety constraints, and define the first preset range: Second preset range: The subsequent coordinated adjustment provides data support, while verifying the matching between dispersion compensation and filtering characteristics.

[0081] Step S2: Use stretching to compensate for the optical path difference in the dispersive filter component and control the optical path difference to be constant.

[0082] Because the filtering function in the dispersive filter component needs to ensure the optical path difference between the two arms... Constant, first grating pair spacing varies This will directly cause the first optical path L1 to change synchronously, that is To maintain a constant optical path difference between the two arms and ensure filtering stability, the stretching amount needs to be adjusted. Compensate for optical path changes, so that The change in optical path due to grating stretching is Therefore, the synergistic regulation formula (7) is derived: (7) Set constraints: First grating spacing adjustment amount and ; stretching amount ≤3mm, when calculated At that time, take Synchronously correct the spacing adjustment of the first grating The first center wavelength satisfies This step establishes a forced linkage mechanism between the spacing and stretching of the first grating pair to avoid optical path difference imbalance caused by independent adjustment, such as filter failure, or hardware damage such as fiber stretching breakage or overtravel of the first grating pair spacing, thus ensuring the coordinated optimization of dispersion compensation and spectral filtering.

[0083] In the traditional system, the independent adjustment of the grating pair spacing and the fiber stretching amount is easy to cause the fiber fracture or the grating over-range. The embodiments of the application guarantee the hardware safety through double constraints: one is to explicitly define the physical boundary in the cooperative adjustment model, and to automatically correct the adjustment amount through the linkage formula, such as synchronously reducing the grating pair spacing adjustment amount when the stretching amount is over the limit; the other is to add a secondary check in the adjustment execution, to judge the boundary of the adjustment amount output by the model, and to preferentially return to the safe state of the near-zero dispersion point. At the same time, the fiber beam splitter with a splitting ratio of 55:45 is adopted to ensure that the interference visibility is 0.9≤V≤1, reduce the fiber stretching loss, and significantly improve the long-term operation reliability of the system.

[0084] Step S3: compensating the dispersion value of the first fiber connection section based on the second grating pair spacing set in the first pulse width adjustment component and the corresponding first pulse width set.

[0085] Step S3 can include step S31 and step S32.

[0086] Step S31: determining the negative dispersion value of the first pulse width adjustment component based on the second grating pair spacing set in the first pulse width adjustment component and the corresponding first pulse width set.

[0087] Collecting the historical adjustment parameters of the first pulse width adjustment module, including the second grating pair spacing set , the second grating pair spacing step value , and the corresponding first pulse width set .

[0088] Step S32: compensating the dispersion value of the first fiber connection section based on the negative dispersion value; the absolute value of the negative dispersion value is greater than or equal to the absolute value of the dispersion value of the first fiber connection section.

[0089] The negative dispersion amount provided by the second grating pair , and the fiber dispersion value D of the CDF of the first fiber connection section is compensated by using the negative dispersion amount CDF , to ensure that the fiber dispersion of the CDF section is completely compensated.

[0090] Step S4: compensating the dispersion value of the second fiber connection section based on the second pulse width set of the second pulse width adjustment component and the dispersion value set of the tunable fiber stretcher in the second pulse width adjustment component.

[0091] Collecting the historical adjustment parameters of the second pulse width adjustment component, including the dispersion value set D3 of the tunable fiber stretcher , the step value , and the corresponding second pulse width set . Based on formula (5), the D3 constraint formula (8) is obtained as: (8) In this step, by establishing the parameter quantization relationship of the pulse width adjustment module, the dispersion compensation constraint boundary is determined, which provides a theoretical basis for the precise control of different pulse lasers such as femtosecond and picosecond double pulses, and ensures that the pulse width adjustment matches the overall performance of the system.

[0092] The control method of the intelligent control anti-locking cooperative output ultrafast laser system provided by the application directly realizes the cooperative control of dispersion, filtering and pulse width parameters at the optical level without complex electrical signal modulation, and the structure is more compact and has no additional electrical driving redundancy. At the same time, combined with the intelligent algorithm with a priority of mode locking, the continuous adjustment of femtosecond and picosecond pulse widths is realized at the optical level, and the tuning process does not need electrical signal conversion, which balances the mode locking stability and pulse width accuracy, and avoids the response delay and error caused by electrical-optical conversion.

[0093] Step S5: establishing a data set; the data set includes a multi-dimensional feature vector, and the multi-dimensional feature vector includes a first target center wavelength, a first target pulse width, a second target pulse width, a first measured center wavelength, a first measured pulse width, a second measured pulse width, a first measured adjustment amount of a grating pair interval, a second measured adjustment amount of a grating pair interval, a measured stretching amount of a base module, and a measured dispersion value of a second pulse width adjustment component.

[0094] In this step, the first frequency domain data acquisition card, the first time domain data acquisition card, the second frequency domain data acquisition card, and the second time domain data acquisition card can synchronously collect two pulse parameter and mode locking state data at a sampling frequency of 1 kHz, and the effective sample set is reserved through a clock synchronization signal of the main control module : When the pump power is less than 50 dB or the radio frequency spectrum has no stable repetition frequency peak, it is determined that the mode locking is lost, the current first grating pair interval is recorded, and the mode locking backoff mechanism is triggered. An input 10-dimensional feature vector is constructed, and target parameters, measured parameters, and component state parameters are integrated: . Among them , , , the first target center wavelength, the first target pulse width, and the second target pulse width, the first measured center wavelength, the first measured pulse width, the second measured pulse width, the first measured adjustment amount of the grating pair interval, the second measured adjustment amount of the grating pair interval, the first measured stretching amount of the base module, and the measured dispersion value of the second pulse width adjustment component. The feature vector is normalized to map the feature value to the interval [0, 1], and the dimensionless vector can be calculated by the following formula (9): (9) wherein, is the minimum value, is the maximum value. After processing the data to obtain the data set, the data set can be split into a training set and a test set in a ratio of 8:2. In this way, high-quality model input samples can be constructed through data screening, feature integration and standardization, and the generalization ability and adjustment accuracy of the subsequent prediction model can be improved, thereby providing a reliable data foundation for intelligent regulation.

[0095] Step S6: training the dual pulse width collaborative prediction model using the data set; the input of the trained dual pulse width collaborative prediction model is a multi-dimensional feature vector, and the output is the first grating pair spacing adjustment amount, the second grating pair spacing adjustment amount, the first stretching amount of the base module, the pump light power adjustment amount and the dispersion value adjustment amount of the second pulse width adjustment component.

[0096] Step S6 can include step S61 and step S62.

[0097] Step S61: constructing a dual pulse width collaborative prediction model, the dual pulse width collaborative prediction model being a nonlinear prediction model fused with an attention mechanism, the dual pulse width collaborative prediction model including an input layer, an attention layer, a hidden layer and an output layer connected in sequence.

[0098] Step S62: training the dual pulse width collaborative prediction model in combination with the data set and the anti-lockout priority.

[0099] A nonlinear prediction model fused with an attention mechanism is constructed, the anti-lockout priority is embedded in the training of the model, the collaborative mapping of the stable priority of mode locking and the accurate matching of pulse parameters is realized, the dual pulse width collaborative prediction model includes an input layer x, an attention layer, two hidden layers and an output layer connected in sequence, each hidden layer includes 180 neurons and uses a GELU activation function.

[0100] The input layer and the output layer include, respectively: ; The attention layer ; the first hidden layer: ; The output feature vector of the first hidden layer is The weight of the first hidden layer is The bias vector of the first hidden layer is The second hidden layer: ; The output feature vector of the first hidden layer is The weight of the first hidden layer is The bias vector of the first hidden layer is The output layer: . a weight for the output layer, a bias vector for the output layer, and an attention weight matrix and bias, a feature dimension, =8e-5 is a L2 norm regularization coefficient. A weighted loss function is adopted to strengthen the training priority of the mode-locked state, which can be calculated by the following formula (10): (10) ; wherein, is a mode-locked state identifier, is a stable mode-locked, is a loss of lock, is a model predicted mode-locked state, =5 is a mode-locked state weight, that is, higher than the adjustment error weight, 、 、 、 is an actual adjustment amount of the i th sample; 、 、 、 is a corresponding adjustment amount predicted by the model.

[0101] In the training parameter setting, the optimizer can be Ranger, which converges faster, trains faster, and has more accurate training results. The initial learning rate =1.8e-5, mini-batch=48, adaptive early stopping is enabled, and the verification set mode-locked prediction accuracy does not improve for 18 consecutive epochs.

[0102] The double pulse width collaborative prediction model provided by the embodiment of the application can control the priority and ensure that the lock is not lost, and the pulse parameter is adjusted more accurately. The attention layer automatically focuses on the parameters that have the greatest impact on the mode-locked, such as the grating interval deviation and the stretching amount; the GELU activation function can capture the subtle correlation between parameters, such as the influence of small-range stretching amount change on the wavelength; and the weighted loss function can control the mode-locked stability, and the penalty for the mode-locked state error is 5 times that of the adjustment error, thereby avoiding the model from losing the lock in order to match the pulse width.

[0103] Step S7: using a Bayesian optimization algorithm to optimize the double pulse width collaborative prediction model.

[0104] Step S7 can include steps S71 to S74.

[0105] Step S71: Selecting a gradient boosting tree as a surrogate model to construct an objective function; wherein the objective function includes a first objective function, a second objective function, a third objective function, and a fourth objective function, the first objective function is used to maximize the mode-locked stability, the second objective function is used to minimize the pulse width error of the first laser, the third objective function is used to minimize the pulse width error of the second laser, and the fourth objective function is used to minimize the first center wavelength error.

[0106] Using the Bayesian optimization algorithm to optimize the double pulse width collaborative prediction model can break through the limitation of single target optimization, and simultaneously achieve the collaborative goals of mode-locked stability, pulse width standard, wavelength accuracy, and hardware safety, and the target priority can be dynamically adjusted.

[0107] Selecting a gradient boosting tree (XGBoost) as a surrogate model, the objective function of the model includes: (11) The first objective function is used to maximize the mode-locked stability, the second objective function is used to minimize the pulse width error of the first laser, the third objective function is used to minimize the pulse width error of the second laser, and the fourth objective function is used to minimize the first center wavelength error.

[0108] Step S72: Construct a constraint function; in the constraint function, the weight values corresponding to different mode-locked states are different.

[0109] By introducing a priority weight system through the constraint function, a dynamic constraint system is constructed, and the constraint strength is dynamically adjusted according to the current mode-locked state includes: (12) wherein, is the mode-locked constraint weight, = 5.0 represents a loss of lock, = 1.0 represents a mode-locked stable, - is the other constraint weight, the value is 1.0, and the constraint function is defined the same as in the foregoing embodiments, including wavelength, pulse width, and hardware safety constraints.

[0110] Step S73: Construct a multi-objective expected improvement acquisition function that fuses target priority; in the multi-objective expected improvement acquisition function, the target priority coefficients corresponding to different mode-locked states are different.

[0111] Design a constraint multi-objective expected improvement (P-CMEI) acquisition function with priority, and fuse the target priority and the constraint satisfaction probability includes: (13) wherein, is a target priority coefficient, the first target priority coefficient =0.6, the second target priority coefficient to the fourth target priority coefficient are all 0.15; when the mode-locked state is stable, the first target priority coefficient =0.3, the second target priority coefficient to the third target priority coefficient are all 0.25, and the fourth target priority coefficient =0.2; is the expected improvement of the kth target; is the dynamic constraint satisfaction probability. By optimizing the output optimal adjustment amount, the multi-target collaborative optimization is realized.

[0112] Step S74: based on the target function, the constraint function and the multi-target expected improvement acquisition function, a double pulse width collaborative prediction model is collected.

[0113] The optimized double pulse width collaborative prediction model can dynamically adjust the priority, can guarantee the collaboration of the mode-locked stability, the pulse width accuracy and the wavelength accuracy, and realizes the collaborative target of the mode-locked stability and the pulse width adjustment. The dynamic constraint weight can intelligently adjust the weight priority. If the current is lost, the weight of recovering the mode-locked is improved, and the accuracy requirement of the pulse width or the wavelength is relaxed; if the current mode-locked is stable, the pulse parameters are mainly optimized; the improvement acquisition function will preferentially select the mode that can be mode-locked with a high probability and the pulse parameters are close to the target, control the collaboration of the pulse width adjustment and the mode-locked state, and finally realize the deviation of the target value and the actual output value <5%.

[0114] The control method of the intelligent control and regulation anti-locking collaborative output ultrafast laser system provided by the embodiments of the present application can further include step S8.

[0115] Step S8: using the double pulse width collaborative prediction model to control the intelligent control and regulation anti-locking collaborative output ultrafast laser system.

[0116] In step S8, the intelligent control and regulation anti-locking collaborative output ultrafast laser system outputs the first laser and the second laser while maintaining a stable mode-locked state.

[0117] According to the mode-locked stability priority principle, the control method of the intelligent control and regulation anti-locking collaborative output ultrafast laser system provided by the embodiments of the present application can control the intelligent control and regulation anti-locking collaborative output ultrafast laser system to perform step-by-step adjustment and emergency mechanism to ensure the safety and controllability of the adjustment process. First, the mode-locked state is judged. If the current is lost, the first displacement table moves at a speed v Lg1=5μm / ms back to the last stable grating pair spacing history value, while the pump power is temporarily reduced by 10%, reducing the intracavity gain to assist the recovery of mode locking; secondly, adjust , according to and the corresponding relationship of λ c , the center wavelength is calibrated to ±1nm, avoiding the indirect influence of wavelength deviation on mode locking stability; then adjust ΔLg1, pause for 2ms every 5μm adjustment, and detect the signal-to-noise ratio SNR synchronously, if SNR < 50dB, immediately back 3μm to prevent loss of lock caused by dispersion imbalance; finally, adjust and ΔD3 in turn, pause for 2ms after each adjustment, and collect pulse width data in real time to ensure that the pulse width matches the target value. If the output power fluctuation is > 3% or SNR < 50dB, restart the optimization process to achieve dynamic stability.

[0118] The results of the specific embodiment are as follows in Table 1: Table 1 Center wavelength and pulse width control table ; The prior art lacks a perfect loss of lock prediction and emergency mechanism, and is prone to cause mode locking imbalance due to environmental interference and parameter adjustment. The control method of the intelligent regulation and control anti-loss of lock cooperative output ultrafast laser system provided in the embodiment can run through the whole process of anti-loss of lock: the weighted loss function used in the model training stage gives the mode locking state a training priority 5 times that of the pulse width adjustment; the dynamic constraint multi-objective Bayesian optimization is used in the optimization adjustment stage, and the mode locking constraint weight is automatically increased when the loss of lock occurs; the step-by-step strategy of loss of lock back, wavelength calibration, dispersion optimization, and pulse width adjustment is used in the execution stage, and real-time monitoring at 10ms intervals such as monitoring of power fluctuation, SNR, and pulse width deviation is used, and once the loss of lock sign occurs, the grating pair spacing back, pump power fine adjustment, and other emergency actions are triggered. This mechanism ensures that the system can still maintain stable mode locking pulse output under environmental interference such as temperature fluctuation and vibration, and solves the core problem of the traditional system that is prone to loss of lock during long-term operation.

[0119] In the control process, first, the grating pair spacing can be adjusted with high precision by controlling the first displacement table through the main control board, and the adjustment precision and stability of the system are higher than those of manual adjustment of the grating pair spacing; secondly, the base module can be intelligently controlled by combining the target pulse width input by the upper computer, and higher precision and more controllable center wavelength tuning can be realized by applying precise tension to the optical fiber, so that the required pulse width can be directly and accurately output, the operation is intelligentized, the predictability of the result is realized, and the filtering performance and wavelength positioning accuracy of the system are also improved.

[0120] The control method of the intelligent regulation and lock-prevention collaborative output ultrafast laser system does not need to rely on the time base of the electrical signal, and can directly realize the output of picosecond and femtosecond level ultra-short pulses, the pulse width adjustment range is wider and more in line with the core application requirements of ultrafast lasers. The grating is used to separate the incident light wavelength, and the wavelength selection needs to be realized by adjusting the grating angle, but the efficiency of the diffraction grating is strongly related to the angle, and the angle change will greatly affect the overall efficiency of the system; in the present application, the diffraction grating angle is fixed at the position of maximum efficiency, and only by adjusting the distance between the grating pairs to change the negative dispersion value provided by them, the pulse width adjustment is realized, which not only ensures the stability of the system output efficiency, but also improves the accuracy of the pulse width adjustment through the precise dispersion regulation.

[0121] Figure 5 is a flowchart of a second intelligent regulation and lock-prevention collaborative output ultrafast laser system control method provided by the embodiments of the present application.

[0122] Referring to Figure 5 , in one specific implementation, the first laser output by the intelligent regulation and lock-prevention collaborative output ultrafast laser system can be femtosecond laser, and the second laser can be picosecond laser.

[0123] The control method of the intelligent regulation and lock-prevention collaborative output ultrafast laser system can be realized by the following steps S01 to S06.

[0124] Step S01: Obtain the dispersion filter component basic parameters: the first grating pair distance set [Lg 11 , Lg 12 , …, Lg 1i ], the stretching amount set ΔX of the base module [ΔX1, ΔX2, …, ΔX i ], and the center wavelength set λ c corresponding to the stretching amount [λ1, λ2, …, λ i ], to obtain the quantitative correspondence between the stretching amount and the center wavelength.

[0125] The correspondence is: .

[0126] This step can refer to the aforementioned embodiment step S1.

[0127] Step S02: Establish a double pulse width collaborative prediction model, realize the linkage of the grating pair distance and the stretching amount through , constrain the stretching amount ≤3mm, the total adjustment amount of the grating pair distance ≤1cm, and maintain the constant optical path difference of the two arms to ensure the stability of the filter.

[0128] Generally, dispersion regulation and spectral filtering are performed independently, which easily leads to unbalanced net dispersion in the cavity, pulse broadening or filtering failure. The embodiment of the application establishes a synergistic adjustment model of the dispersion filtering assembly, binds dispersion compensation and spectral filtering depth, which not only ensures the accuracy of dispersion compensation, but also avoids the distortion of filtering characteristics due to independent adjustment, and significantly improves the stability of pulse output.

[0129] Dual-pulse synergistic output + anti-lost lock dual protection: for the first time, dual-pulse synergistic regulation is combined with displacement table over-movement anti-lost lock mechanism, through safety range calibration, adjustment amount limitation, real-time lost lock monitoring and rollback mechanism, to ensure that the displacement table of the dispersion management module does not move beyond the range allowed by mode locking, solving the problem of lost lock caused by over-movement of components in traditional lasers.

[0130] Step S03: Collecting historical operating parameters of the second grating pair in the pulse width first adjustment assembly and the adjustable fiber stretcher in the pulse width second adjustment assembly, clarifying dispersion constraints, and adapting to the adjustable demand of femtosecond and picosecond pulse width.

[0131] Step S04: Collecting system operating data at 1kHz synchronously, eliminating abnormal samples such as lost lock, constructing a 10-dimensional feature vector and doing standardization processing, and proportionally splitting into training set and test set.

[0132] The above steps S02-S04 can refer to the aforementioned embodiment steps S2-S5.

[0133] Step S05: Training a nonlinear prediction model with fusion attention mechanism, according to the input feature vector, ensuring mode locking priority, and outputting accurate matching parameters .

[0134] This step can refer to the aforementioned embodiment step S6.

[0135] Step S06: Using a multi-objective Bayesian optimization algorithm to dynamically adjust the target priority, outputting the optimal adjustment amount, realizing accurate pulse width matching of femtosecond fs and picosecond ps and synergistic stable adjustable output.

[0136] This step can refer to the aforementioned embodiment steps S7-S8.

[0137] The traditional mode-locked laser is limited by the fixed bandwidth of the filter device, and it is difficult to flexibly switch the pulse type and the adjustable range of the pulse width is narrow. Embodiments of the application adjust the filter 3dB bandwidth by stretching, and at the same time, a double pulse width adjustment module is matched, so that 127fs-326fs femtosecond pulse output and 6ps-150ps picosecond pulse output can be finally realized, and the center wavelength can cover the adjustable range of 1023.3nm-1039.5nm. Compared with the existing multi-pulse laser system, the design is more suitable for the pulse requirements of different scenes such as ultra-precision machining and micro-detection, and solves the problem of single pulse characteristics of the traditional system.

[0138] It should be noted that other embodiments of the application will readily occur to those skilled in the art upon consideration of the specification and practice of the application disclosed herein. The application is intended to include all such variations as fall within the general scope of the application and the general principles and features defined herein.

[0139] It should be understood that the application is not limited to the precise construction that has been described above and shown in the accompanying drawings, and that various modifications and changes can be made by those skilled in the art without departing from the scope of the application. The true scope of the application is set forth in the attached claims.

Claims

1. A smart control anti-lockout cooperative output ultrafast laser system, characterized in that, include: The pump module is configured to transmit pump light; A first beam splitter is disposed in the optical path of the pump light, and the first input terminal of the first beam splitter is connected to the pump module; A dispersion filter component is disposed on the side of the first beam splitter away from the pump module. The input end of the dispersion filter component is connected to the first output end of the first beam splitter. The dispersion filter component is configured to perform dispersion compensation to output laser light. The pulse width adjustment component includes a first pulse width adjustment component and a second pulse width adjustment component, wherein the first pulse width adjustment component and the second pulse width adjustment component are respectively connected to the dispersive filter component; The first pulse width modulation component is configured to perform dispersion compensation on the laser using a first pulse width modulation mechanism to output a first laser; the second pulse width modulation component is configured to perform dispersion compensation on the laser using a second pulse width modulation mechanism to output a second laser. The first pulse width adjustment mechanism differs from the second pulse width adjustment mechanism, and the pulse widths of the first laser and the second laser are different.

2. The intelligent control anti-lockout cooperative output ultrafast laser system according to claim 1, characterized in that, The dispersion filtering component includes a second beam splitter and a first grating pair module; the first input terminal of the second beam splitter is connected to the first output terminal of the first beam splitter, and the first output terminal of the second beam splitter is connected to the first grating pair module; the first grating pair module is configured to receive the laser and perform dispersion compensation on the laser. The dispersion filtering component further includes a third beam splitter and a base module; the first input end of the third beam splitter is connected to the first grating pair module; the second output end of the second beam splitter is connected to the second input end of the third beam splitter, and the connecting optical fiber between the second output end of the second beam splitter and the second input end of the third beam splitter is fixed on the base module. The optical path from the first output end of the second beam splitter to the first input end of the third beam splitter is the first optical path, and the optical path from the second output end of the second beam splitter to the second input end of the third beam splitter via the connecting optical fiber between the base modules is the second optical path. The first optical path and the second optical path are different to filter the laser.

3. The intelligent control anti-lockout cooperative output ultrafast laser system according to claim 2, characterized in that, The first pulse width adjustment component is connected to the dispersive filtering component, and the first pulse width adjustment component includes a third fiber collimator and a second grating pair module; The input end of the third fiber collimator is connected to the output end of the third beam splitter. The third fiber collimator is configured to collimate the laser and then direct it into the second grating pair module. The second grating pair module is configured to use the first pulse width adjustment mechanism to compensate for the dispersion value of the first fiber connection segment. The first fiber connection segment includes the connecting fiber between the third beam splitter and the third fiber collimator.

4. The intelligent control anti-lockout cooperative output ultrafast laser system according to claim 1, characterized in that, The dispersion filtering component is also connected to the second input terminal of the first beam splitter, and the second pulse width adjustment component is connected to the second output terminal of the first beam splitter. The second pulse width adjustment component includes a circulator and an adjustable fiber stretcher. The first port of the circulator is connected to the second output terminal of the first beam splitter, and the second port of the circulator is connected to the adjustable fiber stretcher. The intelligent control anti-lockout cooperative output ultrafast laser system also includes a fourth fiber collimator. The third port of the circulator is connected to the fourth fiber collimator. The adjustable fiber stretcher is configured to compensate for the dispersion value of the second fiber connection segment using the second pulse width adjustment mechanism. The second fiber connection segment includes the connecting fiber between the first beam splitter and the fourth fiber collimator.

5. The intelligent control anti-lockout cooperative output ultrafast laser system according to claim 2, characterized in that, The first grating pair module includes a first grating pair and a first displacement stage; the first grating in the first grating pair is disposed on the first displacement stage, and the first displacement stage is configured to drive the first grating to move along the normal direction of the first grating to adjust the first grating pair spacing of the first grating pair; the first grating pair module is also configured to perform dispersion compensation based on the first grating pair spacing.

6. A control method for an intelligent control anti-lockout cooperative output ultrafast laser system, characterized in that, In the intelligent control anti-lockout cooperative output ultrafast laser system applied as described in any one of claims 1-5, the control method of the intelligent control anti-lockout cooperative output ultrafast laser system includes: Based on the first grating pair spacing set in the dispersion filter component, the first stretching amount set of the base module in the dispersion filter component, and the first center wavelength set in the dispersion filter component, the correspondence between the stretching amount and the center wavelength of the base module is determined; the first center wavelength in the first center wavelength set corresponds one-to-one with the first stretching amount in the first stretching amount set. The stretching amount is used to compensate for the optical path difference in the dispersive filter component, thereby controlling the optical path difference to be constant. Based on the second grating pair spacing set and the corresponding first pulse width set in the first pulse width adjustment component, the dispersion value of the first optical fiber connection segment is compensated. Based on the second pulse width set of the second pulse width adjustment component and the dispersion value set of the adjustable fiber stretcher in the second pulse width adjustment component, the dispersion value of the second fiber connection segment is compensated. Establish a dataset; the dataset includes a multidimensional feature vector, which includes the first target center wavelength, the first target pulse width, the second target pulse width, the first measured center wavelength, the first measured pulse width, the second measured pulse width, the first grating spacing measured adjustment amount, the second grating spacing measured adjustment amount, the first measured stretching amount of the base module, and the measured dispersion value of the second pulse width adjustment component; The dataset is used to train a dual-pulse-width collaborative prediction model; the input of the trained dual-pulse-width collaborative prediction model is the multi-dimensional feature vector, and the output is the first grating pair spacing adjustment amount, the second grating pair spacing adjustment amount, the first stretching amount of the base module, the pump light power adjustment amount, and the dispersion value adjustment amount of the second pulse width adjustment component. The dual-pulse-width collaborative prediction model was optimized using a Bayesian optimization algorithm.

7. The control method for the intelligent control anti-lockout cooperative output ultrafast laser system according to claim 6, characterized in that, Based on the first grating pair spacing set in the dispersive filter component, the first stretching amount set of the base module in the dispersive filter component, and the first center wavelength set, the correspondence between the stretching amount and the center wavelength is determined, including: Within the stable mode-locked pump power range of the pump light, the set of the first grating pair spacing, the set of the first stretching amount, and the set of the first center wavelengths corresponding to the first stretching amount are collected in the first grating pair. The adjustment amount of the first grating pair spacing is determined based on the first grating pair spacing set; The correspondence between stretching amount and center wavelength is determined based on the first grating pair spacing set, the first center wavelength set, and the first stretching amount set; wherein, the absolute value of the maximum value of the first grating pair spacing adjustment amount is within a first preset range, the first stretching amount is within a second preset range, and the first preset range is different from the second preset range.

8. The control method for the intelligent control anti-lockout cooperative output ultrafast laser system according to claim 7, characterized in that, Based on the second grating pair spacing set and the corresponding first pulse width set in the first pulse width adjustment component, the dispersion value of the first fiber connection segment is compensated, including: Based on the second grating pair spacing set and the corresponding first pulse width set in the first pulse width adjustment component, the negative dispersion value of the first pulse width adjustment component is determined; The dispersion value of the first optical fiber connection segment is compensated based on the negative dispersion value; the absolute value of the negative dispersion value is greater than or equal to the absolute value of the dispersion value of the first optical fiber connection segment.

9. The control method for the intelligent control anti-lockout cooperative output ultrafast laser system according to claim 8, characterized in that, The dual-pulse-width collaborative prediction model is trained using the dataset, including: A dual-pulse-width collaborative prediction model is constructed. The dual-pulse-width collaborative prediction model is a nonlinear prediction model that integrates an attention mechanism. The dual-pulse-width collaborative prediction model includes an input layer, an attention layer, a hidden layer, and an output layer connected in sequence. The dual-pulse width collaborative prediction model is trained by combining the dataset and the anti-lockout priority.

10. The control method for the intelligent control anti-lockout cooperative output ultrafast laser system according to claim 9, characterized in that, The dual-pulse-width collaborative prediction model is optimized using a Bayesian optimization algorithm, including: Gradient boosting tree is selected as the surrogate model to construct an objective function; wherein, the objective function includes a first objective function, a second objective function, a third objective function and a fourth objective function, the first objective function is used to maximize mode-locking stability, the second objective function is used to minimize the pulse width error of the first laser, the third objective function is used to minimize the pulse width error of the second laser, and the fourth objective function is used to minimize the first center wavelength error; Construct a constraint function; in the constraint function, the weight values ​​corresponding to different mode-locking states are different; A multi-objective expectation improved acquisition function is constructed that integrates target priorities; in the multi-objective expectation improved acquisition function, the target priority coefficients corresponding to different mode-locking states are different; The dual-pulse width collaborative prediction model is optimized based on the objective function, the constraint function, and the multi-objective expectation improved acquisition function.

Images