Method for determining structure parameters of few-mode optical fiber and electronic device

Abstract

This application discloses a method and electronic device for determining the structural parameters of a few-mode fiber, relating to the field of optical communication technology. The method includes using multiple structural parameters of a centrally recessed trench-assisted few-mode fiber as joint optimization parameters, generating multiple sets of structural parameters in a multi-dimensional structural parameter space using an optimization algorithm, and calculating the number of guided modes, the minimum effective refractive index difference between adjacent mode groups, and the bending loss of higher-order modes for each corresponding fiber structure using simulation methods. Structural parameter groups whose guided mode number does not meet the condition are assigned invalid values. The performance quantification value of each structural parameter group is calculated using a performance optimization relationship, and the structural parameter group is updated accordingly. This process is repeated until the condition is met to obtain the optimal structural parameters. This invention can solve the problem that related technologies cannot simultaneously consider a large effective refractive index difference between adjacent mode groups and a low bending loss of higher-order modes, and can effectively improve the mode spacing and bending robustness of centrally recessed trench-assisted few-mode fibers.

Description

Technical Field

[0001] This application relates to the field of optical communication technology, and in particular to a method for determining the structural parameters of a few-mode optical fiber and an electronic device thereof. Background Technology

[0002] As the transmission capacity of single-mode fiber (SMF) communication systems gradually approaches the Shannon limit, few-mode fiber (FMF) based on space division multiplexing can transmit multiple spatial modes simultaneously, thereby improving transmission capacity and being widely used in next-generation high-capacity optical communication.

[0003] However, when determining the structural parameters of few-mode fibers, related technologies rely on preset layer structures and refractive index values, making it difficult to simultaneously consider a large effective refractive index difference between adjacent mode groups and a low bending loss of higher-order modes in a multi-dimensional structural parameter space.

[0004] Therefore, improving the mode spacing and bending robustness of centrally recessed groove-assisted few-mode fibers is a technical problem that needs to be solved by those skilled in the art.

[0005] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] This application provides a method and electronic device for determining the structural parameters of a few-mode fiber, which can effectively improve the mode spacing and bending robustness of a centrally recessed trench-assisted few-mode fiber.

[0007] To solve the above-mentioned technical problems, this application provides the following technical solution: This application provides a method for determining the structural parameters of a few-mode optical fiber, including: Multiple structural parameters of a few-mode fiber with an axisymmetric structure and a refractive index profile that includes a central recess region, a dominant core region, a trench structure region, and a cladding region in a radial direction from the fiber axis are used as joint optimization parameters. Multiple candidate fiber profile structural parameter sets are generated in a multi-dimensional structural parameter space using a target optimization algorithm. The refractive indices of the central recess region and the trench structure region are both lower than the refractive index of the dominant core region. The following process is repeated for each set of candidate fiber profile structure parameters determined in each round until a preset termination condition is met, and the set of candidate fiber profile structure parameters determined in the last round is taken as the optimal fiber profile structure parameters for the few-mode fiber: The target numerical simulation method is used to calculate the number of guided modes of the corresponding fiber structure, the minimum effective refractive index difference between adjacent mode groups, and the bending loss of the target higher-order mode at the preset bending radius. For candidate fiber profile structure parameter groups whose number of guided modes does not meet the target mode number range, invalidation values ​​are assigned to prevent them from being selected. The performance optimization relationship used to calculate the fiber performance quantification value of each candidate fiber profile structure parameter group is determined according to at least a first term that is negatively correlated with the minimum effective refractive index difference and a second term that is positively correlated with bending loss. Based on the fiber performance quantification values ​​of each candidate fiber profile structure parameter group, the candidate fiber profile structure parameter group is updated to obtain the next round of candidate fiber profile structure parameter groups.

[0008] This application also provides an electronic device, including a memory and a processor, wherein the processor is used to implement the steps of the above-described method for determining the structural parameters of a few-mode fiber when executing a computer program stored in the memory.

[0009] The advantages of the technical solution provided in this application are that it uses multiple structural parameters of the axisymmetric few-mode fiber, consisting of the central concave region, the dominant core region, the outer trench structure region, and the cladding region, as joint optimization variables. The selected optimization algorithm employs performance optimization relationships that are at least negatively correlated with the minimum effective refractive index difference and positively correlated with bending loss to evaluate candidate fiber profile structural parameter groups. Simultaneously, candidate groups whose guided mode number does not meet the target mode number range are directly assigned invalid values ​​to exclude them. This allows the entire search process to automatically maximize the minimum effective refractive index difference between adjacent mode groups and minimize the bending loss of higher-order modes within a multi-dimensional parameter space including the central concave depth, trench depth, concave radius, trench position, and trench width. This avoids the limitations of manual parameter scanning, which, due to complex variable coupling relationships, can only search local sections and is unlikely to obtain a globally optimal structure. By eliminating candidate structural parameters with insufficient mode numbers in advance, the number of subsequent fine calculations of bending loss and mode spacing for invalid fiber structures is reduced, thus lowering the overall computational load of the simulation process. Simultaneously, since the minimum effective refractive index difference term and bending loss term in the performance optimization relationship follow physical laws—that is, a larger effective refractive index difference results in weaker intermode crosstalk, and a smaller bending loss results in stronger bending resistance—the optimization algorithm automatically converges in the direction of increasing the effective refractive index difference between adjacent mode groups and reducing the bending loss of higher-order modes during iteration. The final optimal fiber profile structural parameters enable the fiber to achieve a better minimum effective refractive index difference and lower bending loss between adjacent mode groups than relying on manual experience or parameter scanning design, while supporting the target number of guided modes. This improves the fiber's ability to suppress mode coupling in weakly coupled mode-division multiplexing systems. Furthermore, this application also provides corresponding electronic equipment for the method of determining few-mode fiber structural parameters, further enhancing the practicality of the method. The electronic equipment offers corresponding advantages.

[0010] The technical features mentioned above, those to be mentioned below, and those shown individually in the accompanying drawings can be arbitrarily combined, as long as the combined technical features are not contradictory. All feasible combinations of features are the technical content explicitly described in this application. Any one of the multiple sub-features contained in the same statement can be applied independently, without necessarily being applied together with other sub-features.

[0011] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of this application or related technologies, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A flowchart illustrating a method for determining the structural parameters of a few-mode optical fiber provided in this application; Figure 2 A schematic diagram of the refractive index profile of the few-mode fiber provided in this application under an exemplary embodiment; Figure 3 A flowchart illustrating another method for determining the structural parameters of a few-mode fiber provided in this application; Figure 4 The optimal cost function value generated and the optimal cost function value at runtime are provided for this application. A schematic diagram of the curves showing the change with algebra; Figure 5 A schematic diagram showing the changes in minimum effective refractive index difference and bending loss as algebra for the optimal individual provided in this application during operation. Figure 6 This is a structural diagram of an exemplary embodiment of the electronic device provided in this application. Detailed Implementation

[0014] To enable those skilled in the art to better understand the technical solutions of this application, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. The terms "first," "second," "third," "fourth," etc., used in the specification and the aforementioned drawings are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. The term "exemplary" means "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as superior to or better than other embodiments.

[0015] As the capacity of single-mode optical fiber communication systems gradually approaches the Shannon limit, spatial division multiplexing (SDM) has become the next-generation high-capacity optical communication technology direction to further improve the transmission capacity of a single fiber. In SDM systems, mode-division multiplexing (MDM) utilizes orthogonal spatial modes within the same fiber as independent data channels to enhance transmission capacity, and has been extensively studied. At the engineering implementation level, few-mode fiber is the physical carrier of MDM signals. However, during actual transmission and deployment, optical fibers are inevitably affected by factors such as bending, twisting, and fabrication errors. These disturbances disrupt the orthogonality between different spatial modes, causing inter-mode coupling and crosstalk, ultimately leading to a reduction in signal transmission quality. To achieve high-performance MDM transmission, it is necessary to effectively suppress mode crosstalk.

[0016] To suppress mode crosstalk and reduce the complexity of multiple-input multiple-output (MIMO) digital signal processing at the receiver, one related technique is based on full-input multiple-output (MIMO) digital equalization. However, as the number of modes increases, the dimensionality of the equalization matrix and computational power consumption rise significantly. When the differential mode group delay is large, this further burdens the digital signal processing. Another related technique uses vector mode or orbital angular momentum mode transmission. This method can reduce or even avoid complex full-MIMO processing to some extent. It enhances mode orthogonality through special fiber structures (such as ring cores or elliptical cores), but it has high requirements for mode excitation, multiplexing / demultiplexing, and polarization stability, limiting mode scalability. Yet another related technique employs a weak coupling scheme. This method still uses linearly polarized mode groups as the transmission channel and suppresses crosstalk by increasing the effective refractive index difference between adjacent mode groups, typically requiring a difference exceeding 0.5 × 10⁻⁶. - ³, ideally reaching or exceeding 1.0 × 10 - ³, requiring only low-dimensional intra-group MIMO processing. In this structure, the maximum degeneracy within a single mode group typically does not exceed 4, thus low-dimensional, low-complexity intra-group MIMO processing is sufficient to recover the transmitted signal. Compared to traditional strongly coupled MDM systems, weakly coupled FMF systems offer significant advantages in reducing DSP complexity and improving system scalability. Therefore, maximizing the minimum effective refractive index difference between adjacent mode groups (i.e., ...) is crucial. This has become an unavoidable problem to be solved in weakly coupled FMF (Fiber Optic Multifiber). Currently, it is generally believed that weakly coupled transmission fibers... Should exceed A more ideal design goal is to reach or exceed However, in expanding Maintaining other key performance indicators at the same time remains a significant challenge.

[0017] In traditional step-index few-mode fiber (SI-FMF), due to the limited degrees of freedom in its refractive index profile, the effective refractive index difference between some adjacent mode groups is small, such as... and The effective refractive index difference between modes is typically only about 0.6 × 10⁻⁶. - ³, which is insufficient to meet the mode isolation requirements of high-performance weakly coupled transmission. To further increase the mode spacing, ring-core few-mode fiber (RC-FMF) introduces a depression with the same refractive index as the cladding to suppress the mode spacing. The effective refractive index of the mode, thereby increasing and The effective refractive index difference between modes. However, an excessively deep central depression will not only reduce... and The effective refractive index difference between them will also lead to As the mode field expands outward, bending loss deteriorates. Depressed-core-few-mode fiber (DC-FMF) retains a higher central refractive index than the cladding, providing more flexible mode tuning space; however, without a matching bending suppression structure, the bending loss of higher-order modes remains high. While multi-ring core or multi-layer refractive index tuning structures can achieve even greater degrees of freedom, their numerous layers and complex structures require strict control over the fabrication process, resulting in relatively low fabrication tolerances. To address this issue, a moderate central depression can be combined with a trench-assisted structure—a centrally depressed trench-assisted few-mode fiber—to expand the minimum effective refractive index difference between adjacent mode groups in a weakly coupled FMF and improve bending performance.

[0018] A related technique improves key performance parameters of optical fibers, such as mode coupling, mode dispersion, chromatic dispersion, mode correlation loss, bending tolerance, and effective mode area, by numerically optimizing the refractive index distribution of the fiber cross-section. This method first constructs an initial fiber structure and uses a waveguide mode solver to obtain mode physical parameters, such as transverse mode field distribution, propagation constant, effective refractive index, group delay, dispersion coefficient, and effective mode area. A cost function is constructed using the difference between the actual and target propagation constants, the root mean square of the group delay of each mode, or the dispersion coefficient. Based on the gradient of the cost function with respect to the refractive index distribution, the refractive index profile is updated using gradient descent. Constraints such as axisymmetry, refractive index variation amplitude, and smoothing are introduced to iteratively obtain the optimized refractive index profile. During the iteration process, this method can also introduce certain design constraints, such as maintaining the axisymmetry of the fiber structure, limiting the maximum allowable refractive index variation amplitude in each iteration, and smoothing the refractive index update results, to avoid obtaining complex refractive index structures that are difficult to fabricate or overly dependent on mesh discretization. After a sufficient number of iterations, the initial refractive index profile gradually converges to the optimized refractive index profile, thereby obtaining an improved fiber structure, such as structural parameters like mode coupling, group delay, and chromatic dispersion. However, this method uses gradient descent to update the continuous refractive index, relying on the explicit gradient of the cost function, and is not suitable for handling non-analytical, non-convex, multi-constraint optimization problems obtained from COMSOL finite element simulations.

[0019] Another related technique first determines the fixed structural parameters of the few-mode fiber and the optimization range of the double-groove structure. Then, within this range, multiple sets of groove width and refractive index parameters are randomly generated to form the initial population of the genetic algorithm. Subsequently, for each set of parameters, the effective refractive index and transverse electric field distribution of the guided mode are calculated using the finite element method or finite difference method, and the inter-mode crosstalk value under preset bending conditions is calculated based on coupled-mode theory. Then, the crosstalk value is used as the fitness function to select excellent individuals with low crosstalk, and a new generation of population is generated through crossover and mutation. The above process is repeated until the convergence condition is met, and finally, the set of groove structure parameters with the lowest crosstalk is output as the optimal few-mode fiber design. However, this method takes the inter-mode crosstalk value under preset bending conditions as the only optimization objective, calculates the maximum crosstalk through coupled-mode theory, and uses the minimization of crosstalk as the optimization objective of the genetic algorithm. It cannot achieve both improved weakly coupled transmission capability and bending robustness.

[0020] Another related technique involves establishing a theoretical model of erbium-doped fiber and its amplifier, determining the range of amplifier parameters to be optimized, such as erbium doping concentration, core radius, erbium-doped radius, core-cladding refractive index difference, fiber length, pump wavelength, and signal power. An initial population is randomly generated, and numerical calculations are performed based on rate and propagation equations. Signal gain and bandwidth are used as cost functions, and a genetic algorithm is employed to select the optimal parameters, outputting the best parameters for different pump powers. This approach focuses on performance optimization of erbium-doped fiber amplifiers, such as improving center gain, expanding bandwidth, and reducing noise figure. However, this method, which targets EDFA (Electronic Diffusion Fiber Amplifier) ​​performance optimization, is not suitable for structural optimization of weakly coupled 10-mode center-recessed trench-assisted few-mode fibers.

[0021] Another related technique involves constructing training samples based on a general design model for ring-core fibers. The input variables are the core radius of each layer and the relative refractive index difference between each layer and the cladding. The output variable is the coupling integral coefficient between adjacent high-order mode groups. Samples are obtained using conventional electromagnetic field calculation methods, a backpropagation (BP) neural network is constructed and trained, and then a genetic algorithm is used to search for the ring-core fiber structure parameters with the minimum coupling integral coefficient under defined conditions, using the neural network as the prediction model. However, this method is geared towards ring-core fibers, with the optimization objective being the coupling integral coefficient between mode groups. This invention is geared towards weakly coupled 10-mode center-recessed trench-assisted few-mode fibers, and is therefore unsuitable for refractive index profile optimization of weakly coupled 10-mode center-recessed trench-assisted few-mode fibers. Furthermore, this approach uses the coupling integral coefficient between mode groups as the optimization objective, predicts using a BP neural network, and searches for the fiber structure with the minimum coupling integral coefficient in a genetic algorithm, failing to achieve the goal of improving weakly coupled transmission capability while simultaneously ensuring bending robustness.

[0022] Another related technology involves collecting raw process data such as torch gas flow rate during the optical fiber preform fabrication process. Using gas flow rate as input and preform quality as output, the data is preprocessed and quality rated to construct a BP neural network prediction model. A genetic algorithm is then used to optimize the initial weights and thresholds of the neural network, and finally, the optimal gas flow rate combination for predicting preform quality is obtained through further optimization using the genetic algorithm. This approach belongs to the optical fiber preform fabrication process optimization, focusing on the gas flow rate combination with the goal of improving preform quality. However, optimizing the refractive index profile parameters of many-mode optical fibers cannot achieve the goal of improving weakly coupled transmission performance.

[0023] As shown above, related technologies rely on manual experience and parameter scanning to determine the refractive index profile of center-recessed trench-assisted few-mode fibers. The parameter scanning method has limited search dimensions; the center-recessed parameters affect the effective refractive index distribution of different linear polarization modes, while the trench parameters affect the confinement capability of higher-order modes and bending loss. Multiple parameters are coupled together. One-dimensional or two-dimensional scanning can only reflect local sections in the multi-dimensional parameter space, making it difficult to obtain designs with optimal overall performance. With the increase in the number of design variables and parameter combinations, a large number of finite element mode simulations and bending loss calculations are required. Furthermore, manual scanning necessitates repeated adjustments to the range and step size, resulting in redundant calculations. The response of effective refractive index, mode spacing, and bending loss in few-mode fibers to structural parameters exhibits nonlinear and non-monotonic characteristics. Manual experience or local scanning easily yields locally optimal or empirical structures, failing to systematically search for globally optimal solutions. Furthermore, some current optimization methods employ local optimization algorithms such as gradient descent, requiring the cost function to be continuous and differentiable, and sensitive to the initial structure. However, the cost function based on finite element simulation consists of multiple non-analytical, non-convex, and multi-constraint stages, including mode solving, mode number determination, mode label identification, effective refractive index difference calculation, and bending loss calculation, making explicit gradient methods unsuitable. Additionally, current optimization methods often target only a single performance metric (such as mode spacing, crosstalk, or dispersion), while few-mode fibers must simultaneously satisfy multiple constraints, including the target number of modes, the minimum effective refractive index difference between adjacent mode groups, and bending loss of higher-order modes.

[0024] In view of this, in order to solve the problem that related technologies are difficult to simultaneously consider the number of target modes, the effective refractive index difference between adjacent mode groups, and the bending loss of higher-order modes in the multi-dimensional structural parameter space during the design of refractive index profiles of few-mode fibers, this invention, by jointly searching multiple structural parameters to be optimized, such as the depth of the central refractive index depression, the trench depth, the radius of the central depression, the position of the annular structure, and the trench width, can determine the optimal structural parameters for the central depression trench-assisted few-mode fiber within the parameter range that is feasible in the process, thereby reducing manual parameter scanning and manual trial and error, and improving the design efficiency of complex refractive index profiles. By unifying the constraints on the number of target modes, maximizing the minimum effective refractive index difference between adjacent mode groups, and the bending loss constraint of higher-order modes into the optimization closed loop of the same optimization algorithm, this optimization method can increase the minimum effective refractive index difference between adjacent mode groups, reduce the risk of inter-mode coupling, and suppress the bending loss of the highest-order mode under bending conditions, thereby improving the overall performance of weakly coupled mode-division multiplexing transmission fibers, while ensuring that the fiber supports multiple target spatial modes in the working band. After introducing the technical solution of this application, the various non-limiting embodiments of this application are described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] The design of the weakly coupled FMF aims to maximize (That is, the minimum effective refractive index difference) is used to suppress inter-mode crosstalk and reduce MIMO complexity. In SI-FMF, the effective refractive index distribution of different LP (linearly polarized) modes is not uniform, leading to crosstalk between some adjacent mode groups. The effective refractive index of a specific mode can be selectively controlled by adjusting the refractive index of a local region where the mode field energy is concentrated, based on perturbation theory. The transverse mode field before and after fiber perturbation satisfies the following scalar wave equation: (1) in, This represents the transverse mode field of the undisturbed optical fiber. This represents the transverse mode field of the optical fiber after the disturbance. This represents the propagation constant of the transverse mode field in an undisturbed optical fiber. This represents the propagation constant of the transverse mode field of the optical fiber after the disturbance. For free space wavenumber, , It is the wavelength in free space. It is a transverse Laplace operator. This represents the refractive index distribution before the disturbance. This represents the refractive index distribution after the perturbation. Assuming a weak refractive index perturbation, the mode field of the fiber before and after the perturbation is almost identical, that is... By multiplying the two equations in the above relation (1) by their respective conjugate transverse mode fields, subtraction is performed on the resulting expression, and the cross-sectional area is calculated. Performing integration on the above, due to the self-adjoint property of the operator, the Laplace terms cancel each other out, resulting in the following relation: (2) For weak perturbations, the changes in propagation constant and refractive index are very small, making , Under the weak-conductivity approximation, the propagation constant of the undisturbed mode is approximately equal to the product of the maximum refractive index of the fiber core and the wavenumber, i.e. ,in The maximum refractive index of the fiber core region, for By performing the expansion operation, we can obtain the following relation: (3) Ignoring second-order minor quantities, we can obtain the following relationship: (4) Similarly, the refractive index perturbation term can be approximately expressed as the following relationship: (5) Substituting the above relations (4) and (5) into the above relation (2), the above relation (2) can be simplified to: (6) If the refractive index perturbation exists only in a local region Within, and the disturbance amplitude is constant. ,Right now Then the propagation constant can be expressed as: (7) use The relationship between the effective refractive index of the mode and the first-order perturbation expression can be expressed as: (8) As can be seen from equation (8), when a mode has greater mode power overlap in the perturbation region, the effective refractive index of that mode is... It is more sensitive to local refractive index changes. Therefore, by introducing a central refractive index depression, groove structure, or high refractive index ring in an appropriate region, the effective refractive index of different modes can be selectively controlled, thereby redistributing the effective refractive index difference between modes. Furthermore, it improves weakly coupled transmission performance. Based on the above analysis, the method for optimizing fiber performance by controlling the shape and distribution of the refractive index profile is as follows.

[0026] Please see first. Figure 1 In some embodiments of the method, the method includes the following steps: S101: Multiple structural parameters of a few-mode fiber with an axisymmetric structure and whose refractive index profile includes a central recess region, a dominant core region, a trench structure region, and a cladding region in the radial direction from the fiber axis are used as joint optimization parameters. Multiple candidate fiber profile structural parameter sets are generated in the multidimensional structural parameter space using the objective optimization algorithm.

[0027] Few-mode fiber refers to an optical waveguide that supports a limited number (usually a few to a dozen) spatial modes within its operating wavelength band. It is used in mode-division multiplexing optical communication systems, where different spatial modes serve as independent channels for transmitting multiple optical signals. This requires a large effective refractive index difference between adjacent mode groups to suppress mode coupling, while also requiring higher-order modes to have low radiation loss under bending conditions. In this embodiment, the few-mode fiber is a centrally recessed trench-assisted fiber structure. The axisymmetric structure means that the refractive index distribution of the fiber remains unchanged after rotating at any angle around the fiber axis. This structure can be achieved through a rotary drawing process of the fiber preform, resulting in angular symmetry in the mode field distribution. The refractive index profile refers to the radial distribution curve of the refractive index on the cross-section of the fiber, which determines the eigenmode distribution and propagation characteristics of the electromagnetic field in the optical waveguide. For example... Figure 2As shown, a few-mode fiber includes at least a central recess region, a dominant core region, an outer low-refractive-index trench region, and a cladding region along the fiber radial direction. The refractive indices of the central recess region and the trench structure region are both low-refractive-index, lower than that of the dominant core region. The refractive index of the trench structure region is equal to or slightly less than that of the cladding refractive index. The fiber cross-sectional structural parameters include the radius of the central recess. Dominant core region radius trench starting radius Groove width d, center depression depth and trench depth The central recessed region, located near the fiber axis, can employ a step-type, gradually varying, parabolic, or smoothly modified profile. It is used to adjust the mode with a central field strength distribution, for example... and The effective refractive index of the modes is redistributed to increase the spacing between adjacent mode groups. The dominant core region can be step-index or gradually variable, serving to provide light guiding capability, that is, to concentrate light energy in the core region for propagation, enabling the fiber to support the target LP mode group. The outer low-refractive-index trench region is a low-refractive-index annular region located outside the dominant core region. The trench can be single-groove, double-groove, or multi-groove, and its function is to enhance the confinement capability of higher-order modes and reduce the radiation loss of higher-order modes under bending conditions, thereby suppressing radiation loss caused by bending. The cladding region refers to the outermost region of the fiber, used to form a complete light guiding boundary. Compared with black-box optimization methods that lack physical constraints, this invention combines automated optimization capability, physical interpretability, and process transferability. It is convenient to fine-tune parameters based on doping diffusion, profile smoothing, and dimensional deviations in actual fabrication, and it is also convenient to re-import the measured refractive index profile into the simulation model for feedback optimization.

[0028] Among them, the structural parameter to be optimized can be the radius of the central depression. Dominant core region radius trench starting radius Groove width d, center depression depth and trench depth All or any combination of the parameters can be used. For example, five structural parameters to be optimized can be used as joint optimization parameters. For instance, the central depression depth, trench depth, central depression radius, trench starting position, and trench width can be used as joint optimization parameters. The joint optimization parameters can be expressed as follows: Dominant core region radius As a fixed reference parameter, it can be set. These are fixed structural parameters. Of course, other geometric dimensions or refractive index parameters, such as the dominant core radius, can also be extended as joint optimization parameters according to the actual optical fiber design requirements.

[0029] In this embodiment, the target optimization algorithm refers to an iterative optimization method used to search for the optimal solution in the parameter space. It can be any algorithm capable of calling performance optimization relationships to calculate the quantified fiber performance values ​​of each candidate fiber profile structural parameter group and updating the candidate fiber profile structural parameter group in the current round, such as genetic algorithms, differential evolution algorithms, particle swarm optimization algorithms, simulated annealing algorithms, Bayesian optimization algorithms, surrogate model-assisted optimization algorithms, reinforcement learning search algorithms, hybrid genetic algorithms, or multi-objective evolutionary algorithms. The multidimensional structural parameter space refers to an abstract space composed of multiple structural parameters to be optimized as coordinate axes. Each point in the space corresponds to a set of structural parameter combinations. For example, if there are five structural parameters to be optimized, the corresponding multidimensional structural parameter space is a five-dimensional structural parameter space. The candidate fiber profile structural parameter group refers to a set of fiber profile structural parameters to be evaluated generated during the optimization process; this set of fiber profile structural parameter groups represents the specific values ​​of a set of structural parameters to be optimized. When the first round of calculation begins, the candidate fiber profile structure parameter set can be a set of fiber profile structure parameters initialized by the target optimization algorithm, a set of fiber profile structure parameters randomly generated by it, or a set of fiber profile structure parameters input by the user. That is, the candidate fiber profile structure parameter set determined in the first round can be a set of fiber profile structure parameters generated in any of the above methods. The fiber profile structure parameter set selected in each subsequent round is based on the fiber profile structure parameter set of the current round. The target optimization algorithm generates the candidate fiber profile structure parameter set for the next round in the corresponding multi-dimensional structure parameter space by executing the following steps in S104.

[0030] S102: For each candidate fiber profile structure parameter group, use the target numerical simulation method to calculate the number of guided modes of the corresponding fiber structure, the minimum effective refractive index difference between adjacent mode groups, and the bending loss of the target higher-order mode at the preset bending radius.

[0031] The target numerical simulation method is capable of numerically solving the refractive index distribution of an optical fiber cross-section. It obtains the propagation constants and mode field distributions of each mode by discretizing the fiber cross-section and solving the eigenvalue problem, and outputs the effective refractive index (including the complex effective refractive index) and mode field distribution of the guided modes. The target numerical simulation method can employ the finite element method, finite difference method, beam propagation method, mode matching method, or time-domain finite difference method. Correspondingly, the mode solving platform is not limited to COMSOL Multiphysics; it can also use self-developed finite element mode solvers, finite difference mode solvers, vector finite element solvers, optical waveguide simulation software, or other solvers capable of outputting the effective refractive index and complex effective refractive index. The number of guided modes refers to the total number of modes that satisfy the waveguide condition (the real part of the effective refractive index is greater than the cladding refractive index), where each spatial mode contains two polarization degenerate states under the weakly guided approximation. The minimum effective refractive index difference between adjacent mode groups refers to the difference in effective refractive index calculated among adjacent mode groups arranged from high to low effective refractive index after degenerate modes with similar effective refractive indices are grouped into the same mode group, and the minimum value among these differences is taken. The larger this value, the more difficult it is to meet the phase matching condition between different mode groups, and the weaker the energy coupling between modes. The target high-order mode refers to the mode with the highest order and strongest bending sensitivity in the target mode set. Its bending loss is usually the largest, so it can be used as an evaluation object for bending robustness. The bending radius refers to the radius of curvature of the arc when the optical fiber bends during actual deployment or testing. The smaller the value, the more severe the bending. Bending loss refers to the energy lost due to radiation leakage when the optical signal propagates in a bent optical fiber, usually expressed in decibels per revolution or per unit length. The preset bending radius can be set according to the actual application scenario of the optical fiber. For example, it can be set to 10 mm or less in access networks or data center interconnection, and 30 mm in long-distance transmission.

[0032] S103: For candidate fiber profile structure parameter groups whose number of guided modes does not meet the target mode number range, assign invalidation values ​​to prevent them from being selected, and determine the performance optimization relationship for calculating the fiber performance quantification value of each candidate fiber profile structure parameter group according to at least a first term negatively correlated with the minimum effective refractive index difference and a second term positively correlated with bending loss.

[0033] The invalidation value refers to a specific value assigned to the fiber structure corresponding to an unqualified candidate fiber profile structure parameter group. This value prevents the candidate fiber profile structure parameter group from being selected as optimal or even selected in subsequent comparisons. The specific value can be a positive number much larger than the performance quantification value of a normal candidate fiber profile structure parameter group, or it can be an identifier directly marked as invalid within the algorithm. The target mode quantity range refers to the number of target modes that the fiber can support throughout the entire operating wavelength band. Target modes may include the fundamental mode. Various higher-order modules, such as , , , , There are six LP (linear polarization mode) mode groups, corresponding to 10 spatial modes and 20 polarization degenerate modes. The performance optimization relationship refers to the correspondence between multiple performance indicators of candidate fiber profile structure parameter groups, which are combined into a single scalar result according to defined operational rules. This scalar result is used to compare the overall performance of fiber structures corresponding to different candidate fiber profile structure parameter groups. The fiber performance quantification value refers to the scalar result calculated for the candidate fiber profile structure parameter groups according to the performance optimization relationship; its magnitude is directly used to determine the merits of the candidate structure.

[0034] After calculating the three performance indicators as described above, the first step is to determine whether the number of guided modes meets the target mode number range. For example, if the target supports 10 spatial modes, since each spatial mode contains two orthogonal polarization states, the total number of corresponding polarization degenerate modes is 20. Therefore, the target mode number range can be set to 20 to 24, allowing for a maximum of two additional spatial modes of guided modes. For candidate structures whose guided mode number does not meet this range, an invalidation value is directly assigned to prevent them from being selected. For example, a very large positive number is assigned as the performance quantization value in the genetic algorithm, or it is marked as invalid and excluded from subsequent comparisons in the sorting algorithm. For candidate structures whose guided mode number meets the target mode number range, the fiber performance quantization value is calculated according to the performance optimization relationship. This performance optimization relationship includes at least a first term negatively correlated with the minimum effective refractive index difference and a second term positively correlated with bending loss. The negative correlation of the first term indicates that the larger the minimum effective refractive index difference, the smaller this term's value, and the more beneficial its contribution to the performance quantization value. The positive correlation of the second term indicates that the larger the bending loss, the larger this term's value, and the more detrimental its contribution to the performance quantization value. The specific form of the performance optimization relationship can be a weighted sum of the first and second terms. For example, the first term takes the negative weighted value of the minimum effective refractive index difference, and the second term takes the positive weighted value of the bending loss. The sum of the two terms gives a comprehensive quantitative value.

[0035] S104: Update the candidate fiber profile structure parameter group according to the fiber performance quantification value of each candidate fiber profile structure parameter group to obtain a new candidate fiber profile structure parameter group, and then jump to execute S102.

[0036] The fiber performance quantification value reflects the quality of the fiber structure corresponding to the candidate fiber profile structure parameter group. The update process is the process of selecting the candidate fiber profile structure parameter group for the next round based on the current round's candidate fiber profile structure parameter group. For example, the candidate fiber profile structure parameter group with the best performance can be selected from the fiber performance quantification values ​​to directly participate in the next round of calculation. Some candidate fiber profile structure parameter groups with good performance can have some structural parameters interchanged with the best-performing candidate fiber profile structure parameter group or other good-performing candidate fiber profile structure parameter groups to generate new fiber profile structure parameter groups. Of course, random perturbations can also be added to some parameters of the newly generated fiber profile structure parameter groups to generate new fiber profile structure parameter groups. These newly generated fiber profile structure parameter groups are used as the candidate fiber profile structure parameter groups for the next round, and S102-S104 are executed again.

[0037] For example, when the target optimization algorithm is a genetic algorithm, the update process is as follows: Several candidate structures with the smallest performance quantization values ​​in the current generation are retained as elite individuals. Parent individuals are selected according to their performance quantization values ​​for crossover (exchanging some parameters between two selected parent parameter sets) and mutation (adding random perturbations to some parameters) to generate a new generation of candidate parameter sets. When the target optimization algorithm is a particle swarm optimization algorithm, the update process includes: updating the velocity and position of each particle based on its current position and velocity, its historical best position, and its global best position.

[0038] S105: If the preset termination condition is met, the candidate fiber profile structure parameter group determined in the last round will be used as the optimal fiber profile structure parameter for the few-mode fiber.

[0039] The preset termination condition is a pre-set termination condition for the iterative calculation. It can be reaching a pre-set maximum number of iterations, or the change in the optimal performance quantization value over multiple consecutive generations being less than a pre-set convergence threshold. For example, the termination condition could be reaching 100 iterations, or the change in the optimal performance quantization value over multiple consecutive generations (e.g., 10 generations) being less than 1%. After the termination condition is met, the candidate fiber profile structure parameter set with the best fiber performance in the last generation is determined as the optimal fiber profile structure parameter for the few-mode fiber, and this parameter set is output for subsequent fiber fabrication or system design. Of course, it can also output the minimum effective refractive index difference between adjacent mode sets and the bending loss of the target higher-order mode at a preset bending radius.

[0040] In the technical solution provided in this application embodiment, multiple structural parameters of the axisymmetric few-mode fiber, consisting of the central concave region, the dominant core region, the outer trench structure region, and the cladding region, are used as joint optimization variables. The selected optimization algorithm employs performance optimization relationships that are at least negatively correlated with the minimum effective refractive index difference and positively correlated with bending loss to evaluate candidate fiber profile structural parameter groups. Simultaneously, candidate groups whose guided mode number does not meet the target mode number range are directly assigned invalid values ​​to exclude them. This allows the entire search process to automatically maximize the minimum effective refractive index difference between adjacent mode groups and minimize the bending loss of higher-order modes within a multi-dimensional parameter space including the central concave depth, trench depth, concave radius, trench position, and trench width. This avoids the limitations of manual parameter scanning, which, due to complex variable coupling relationships, can only search local sections and struggles to obtain globally optimal structures. By eliminating candidate structural parameters with insufficient mode numbers in advance, the number of subsequent fine calculations of bending loss and mode spacing for invalid fiber structures is reduced, thus lowering the overall computational load of the simulation process. Simultaneously, because the minimum effective refractive index difference term and bending loss term in the performance optimization relationship follow physical laws—that is, a larger effective refractive index difference results in weaker intermode crosstalk, and a smaller bending loss results in stronger bending resistance—the optimization algorithm automatically converges in the direction of increasing the effective refractive index difference between adjacent mode groups and reducing the bending loss of higher-order modes during iteration. The final optimal fiber profile structural parameters enable the fiber to achieve a better minimum effective refractive index difference and lower bending loss between adjacent mode groups than designs relying on manual experience or parameter scanning, while supporting the target number of guided modes. This improves the fiber's ability to suppress mode coupling in weakly coupled mode division multiplexing systems.

[0041] In the above embodiments, no limitation is made on how to use the target numerical simulation method to calculate the number of guided modes of the corresponding optical fiber structure for the candidate optical fiber profile. Based on the above embodiments, the present invention also provides an exemplary method for calculating the number of guided modes, which may include the following: At the long-wavelength end of the operating band, numerical simulation calculations are performed on the ideal fiber structure corresponding to each candidate fiber profile structure parameter. The real part of the effective refractive index of each mode is extracted, and the mode with the real part of the effective refractive index greater than the cladding refractive index is determined as the guided mode. The total number of polarization degenerate modes of the guided mode is counted. For candidate fiber profile structure parameter groups whose total number of polarization degenerate modes of the guided mode is not within the range of the target number of modes, invalidation values ​​are assigned, and they are set not to participate in the fiber performance quantification calculation process. For candidate fiber profile structure parameter groups whose total number of polarization degenerate modes of the guided mode is within the range of the target number of modes, they are set to participate in the fiber performance quantification calculation process.

[0042] In this context, an ideal fiber structure refers to an idealized model where the fiber is straight and free from any bending or twisting. The operating band refers to the wavelength range within which the fiber is designed and used to transmit optical signals, such as the C-band (1530 nm to 1565 nm) or the C+L band (1530 nm to 1630 nm). The long-wavelength refers to the longest wavelength within the operating band, such as 1630 nm in the C+L band. At this wavelength, the equivalent normalized frequency of higher-order modes is the lowest, and the number of guided modes is the fewest. Therefore, if the target number of guided modes can be supported at the long-wavelength end, it will also be supported at other shorter wavelengths within the band. The real part of the effective refractive index is the real part of the ratio of the mode propagation constant to the vacuum wavenumber; its magnitude determines the phase propagation characteristics of the mode. The cladding refractive index refers to the refractive index of the outermost layer of the fiber. The guided mode criterion is that the real part of the effective refractive index of the mode is greater than the cladding refractive index. In this case, the light energy is confined in the fiber core and propagates axially. If the real part of the effective refractive index is less than or equal to the cladding refractive index, the mode is a radiation mode or a leakage mode and cannot be transmitted over long distances in the fiber. A polarization degenerate mode refers to a mode in which two orthogonal polarization states of the same spatial mode have the same effective refractive index due to the circular symmetry of the fiber. In mode calculation, these modes appear as two modes with equal effective refractive indices. The target mode number range refers to the acceptable range of the number of guided modes that the fiber is expected to support, including a lower limit and an upper limit.

[0043] In this embodiment, numerical simulation calculations are first performed on the ideal fiber structure corresponding to the cross-sectional structural parameters of each candidate fiber at the long-wavelength end of the working band. For example, a two-dimensional axisymmetric or full-section fiber model can be established in COMSOL Multiphysics, setting the material refractive index to the quartz refractive index corresponding to the working wavelength. A perfectly matched layer is set outside the computational domain to absorb boundary radiation. The eigenmodes are solved using the finite element method. The eigenvalue search range of the solver should cover the effective refractive indices of all possible guided modes. Typically, several search points are uniformly set between the cladding refractive index and the maximum core refractive index. The real part of the effective refractive index of all solved modes is extracted and compared with the cladding refractive index one by one. Modes with a real part of effective refractive index greater than the cladding refractive index are identified as guided modes, and the total number of polarization degenerate modes of these guided modes is counted. Since the numerical solution may solve a polarization degenerate mode as two modes with slightly different effective refractive indices (numerical error), modes with an effective refractive index difference less than a pre-set error threshold can be considered degenerate modes during the statistical analysis. For candidate fiber profile structure parameter groups whose total number of polarization degenerate modes is outside the target mode range, an invalidation value is assigned to prevent them from being selected. Taking a genetic algorithm as an example, the fiber performance quantization value of this candidate fiber profile structure parameter group is set to a positive number much larger than the fiber performance quantization values ​​corresponding to all normal candidate fiber profile structure parameter groups, for example, a maximum value, so that the probability of this candidate structure being selected in subsequent selection operations approaches zero. Alternatively, a validity flag can be set inside the algorithm, and this flag can be set to invalid, so that the candidate structure is skipped directly in subsequent comparison, selection, and crossover operations and does not participate in any calculations. For candidate fiber profile structure parameter groups whose total number of polarization degenerate modes is within the target mode range, the candidate structure is set to participate in the subsequent fiber performance quantization calculation process, that is, to calculate the minimum effective refractive index difference between its adjacent mode groups and the bending loss of the target higher-order modes.

[0044] Taking the target numerical simulation method as finite element simulation, the ideal fiber structure as a straight fiber model, and the operating wavelength as the C+L band as an example, the structural parameters of the i-th candidate fiber profile can be expressed as follows: , Indicates the total number of candidate fiber profile structural parameters, such as For each candidate fiber profile structural parameter, the candidate fiber structure is automatically... The data is incorporated into the finite element method (FEM) simulation model. The FEM model automatically updates the fiber cross-sectional geometry, material refractive index distribution, and mesh settings based on the candidate fiber's cross-sectional structural parameters. To ensure the fiber supports the target number of modes (20 polarization degenerate modes) across the entire C+L band, a mode count check is first performed at the long-wavelength end. Since higher-order modes are more likely to approach cutoff at the long-wavelength end, they are prioritized for... Solve the straight fiber model and count the total number of polarization degenerate modes for each guided mode. The guided mode can be represented as: ,in, The cladding refractive index, Let represent the real part of a complex number. The criterion can be written as: if the effective real part of the refractive index of a mode is greater than the cladding refractive index, then the mode is considered a guided mode, which can be expressed as: The range of target patterns can be expressed as ,in, This represents the number of polarization degenerate modes that satisfy the guided mode criterion. This target mode number range ensures that the candidate fiber structure should support at least twenty polarization degenerate modes corresponding to the ten target spatial modes, while avoiding supporting too many additional higher-order modes to prevent introducing unnecessary mode crosstalk and system complexity. If the candidate fiber structure does not meet the constraint of this target mode number range, it is deemed unqualified, assigned a large penalty fitness value for the candidate fiber profile structure parameters, and skipped from subsequent simulations. This method of determining the number of guided modes based on the effective refractive index of the modes is relatively accurate. Traditional step-index few-mode fiber (SI-FMF) determines the number of guided modes based on the normalized frequency constant. To evaluate the number of guide molds, among which The core radius is... The refractive index of the fiber core, This represents the cladding refractive index. However, for complex refractive index profiles, this method is difficult to accurately determine the number of guided modes and can usually only be used as a preliminary reference.

[0045] As can be seen from the above, considering that a large number of randomly generated candidate structures may not support the target number of guided modes during the optimization iteration process, it would be computationally expensive to perform complete mode spacing and bending loss calculations for each candidate structure. In this embodiment, guided mode number screening is first performed at the long-wavelength end, and a penalty is imposed on candidate structures that do not meet the target mode number range, and subsequent time-consuming simulations are skipped. Compared with estimating the number of guided modes based solely on the normalized frequency constant, this hierarchical screening mechanism can more accurately control the target mode number. Subsequently, the mode spacing is calculated at the center wavelength, and the higher-order mode loss is calculated in the bending model, which reduces the total number of finite element simulation calls, reduces the overall optimization time, and reduces the overall optimization computation cost. At the same time, it can also ensure that the final output fiber structure supports the target number of guided modes throughout the entire working band.

[0046] Based on the embodiments, the present invention further specifies the parameter preset step before the target numerical simulation method is executed, which may include the following: The total number of polarization degenerate modes corresponding to the target spatial mode is used as the lower limit of the target mode quantity range, and the sum of the total number of polarization degenerate modes and the preset margin is used as the upper limit of the target mode quantity range; the working band and the preset bending radius are obtained; the target mode quantity range, the working band, and the preset bending radius are input into the finite element simulation model.

[0047] The target spatial pattern refers to the set of spatial patterns intended to be used as independent transmission channels in a modular division multiplexing system, for example... , , , , , There are six mode groups, corresponding to 10 spatial modes. The total number of polarization degenerate modes refers to the number of polarization states obtained by summing all target spatial modes according to their respective degeneracy. For example, 10 spatial modes correspond to 20 polarization degenerate modes. The lower limit value indicates the minimum number of guided modes that the fiber needs to support. The preset margin refers to the number of polarization degenerate modes corresponding to the excess guided modes that the fiber is allowed to support. The margin can be freely set based on the principle of ensuring the number of target modes while avoiding unnecessary inter-mode crosstalk introduced by supporting too many additional higher-order modes. For example, a margin of 4 means that a maximum of 2 additional spatial modes are allowed, corresponding to 4 polarization degenerate modes. In this case, the upper limit of the target mode number range is the lower limit of 20 plus the margin of 4, which equals 24. If the system is sensitive to crosstalk from additional higher-order modes, the margin can be set to 2 or 0; if the system has good MIMO equalization capability, the margin can be set to 6 or greater. The operating band refers to the wavelength range used by an optical communication system, such as the C-band (1530 nm to 1565 nm), L-band (1565 nm to 1625 nm), or C+L band (1530 nm to 1630 nm). The preset bending radius refers to the required degree of fiber optic bending in practical engineering applications; for example, the radius of the fiber optic winding coil in an access network is typically 10 mm to 30 mm. The finite element simulation model is a numerical model established based on the finite element method for calculating optical waveguide modes. It includes the geometry of the fiber cross-section, the material refractive index distribution, boundary conditions, and solver settings.

[0048] In this embodiment, before calculating the number of guided modes using the target numerical simulation method, the total number of polarization degenerate modes supporting the target spatial mode can be used as the lower limit of the target mode number range. For example, if the target spatial mode is... , , , , , There are a total of six mode groups, among which and Each contains two polarization degeneracy modes. , , , Each mode contains 4 polarization degenerate modes, for a total of 20 polarization degenerate modes. 20 is used as the lower limit for the target mode count. The sum of this total number of polarization degenerate modes and a preset margin is used as the upper limit for the target mode count range. With a preset margin of 4, the upper limit is 20 + 4 = 24. The operating band can be obtained through user input or system presets, such as reading the C+L band (1530 nm to 1630 nm) from the optical communication system design requirements. The preset bending radius can be set according to the deployment scenario, for example, 10 mm. Taking the simulation settings in COMSOL Multiphysics as an example, the target mode count range is used as the criterion for guided mode statistics. The long-wavelength end and center wavelength of the operating band are used as the solution wavelengths for guided mode number selection and mode spacing calculation, respectively. The preset bending radius is used as the bending radius parameter for the bent fiber model. These parameters are set once before optimization begins and are automatically applied each time the finite element simulation is called, without the need for repeated input. In addition, weak coupling criteria, fiber fabrication process constraints, material refractive index range, and structural size constraints can be input as needed, and those skilled in the art can flexibly set them according to actual needs.

[0049] As can be seen from the above, considering that different candidate fiber structures need to adopt a unified performance evaluation benchmark, otherwise the comparison results will lose meaning, this embodiment pre-sets the target mode number range, working band and bending radius and fixes them into the simulation model, ensuring that each candidate fiber structure is evaluated under the same external constraints. The final output fiber structure parameters can meet the specified target mode number, working band and bending radius requirements, avoiding the optimization results from deviating from actual needs due to inconsistent evaluation benchmarks.

[0050] Considering that the central concave structure can cause changes in the effective refractive index order of some linear polarization mode groups, simply grouping them according to the magnitude of the effective refractive index can easily lead to mode misjudgment and distortion in the calculation of the minimum effective refractive index difference between adjacent mode groups. Therefore, this invention also provides the following calculation method for the minimum effective refractive index difference between adjacent mode groups, which may include: Numerical simulations were performed on an ideal fiber structure at the center wavelength of the operating band to extract the effective refractive index of polarization degenerate modes that meet the target mode number range. The real parts of the effective refractive indices were then sorted numerically to obtain an effective refractive index sequence. Based on the number of polarization degenerate modes in adjacent mode groups and the continuous arrangement of each polarization degenerate mode, the effective refractive indices corresponding to adjacent mode groups were selected from the effective refractive index sequence to form a refractive index cross subset. The adjacent mode groups include a first mode group and a second mode group. According to the order in which the first mode group precedes the second mode group, and based on the number of polarization degenerate modes contained in each of the first and second mode groups, the refractive index cross subsets were constructed as a first candidate group. According to the order in which the second mode group precedes the second mode group, and based on the second mode group and the first... The number of polarization degenerate modes contained in each mode group is used to construct a second candidate group by cross-refractive index subsets. Based on the effective refractive index interval at the boundary between the first and second mode groups, the effective refractive index intervals corresponding to the first and last polarization degenerate modes in the first mode group, the effective refractive index intervals corresponding to the first and last polarization degenerate modes in the second mode group, and a preset safety factor, the boundary comparison scores of the first and second candidate groups are calculated respectively. The candidate group with the larger boundary comparison score is selected as the polarization degenerate mode division result of the adjacent mode groups. The average effective refractive index of each polarization degenerate mode in the same mode group is calculated, and the average effective refractive index difference of each adjacent mode group is calculated. The minimum value of the average effective refractive index difference is taken as the minimum effective refractive index difference between adjacent mode groups.

[0051] The effective refractive index sequence refers to an ordered list formed by arranging the real parts of the effective refractive indices of all guided modes that satisfy the target mode range in descending order of value. Each element in the sequence corresponds to a polarization degenerate mode. Adjacent mode groups refer to two LP mode groups that are adjacent in terms of effective refractive index magnitude, for example... Groups and Grouping. A continuous arrangement of polarization degenerate modes refers to all polarization degenerate modes within the same LP mode group occupying consecutive positions in the effective refractive index sequence. This is because the polarization degenerate modes within the same mode group have the same or very similar effective refractive indices. A refractive index cross subset refers to a local subsequence selected from the complete effective refractive index sequence that contains all polarization degenerate modes from two adjacent mode groups. This subset is used for grouping when mode order crossover is possible. The first mode group and the second mode group refer to two specific adjacent LP mode groups that may experience order crossover. The first candidate grouping refers to the result of assigning modes in the refractive index cross subset to the mode group assuming the effective refractive index of the first mode group is higher than that of the second mode group. The second candidate grouping refers to the result of assigning modes in the refractive index cross subset to the mode group assuming the effective refractive index of the second mode group is higher than that of the first mode group. Boundary contrast score is a numerical value used to measure the clarity of the inter-group interval relative to the intra-group split under a certain candidate grouping method. A higher score indicates a more reasonable physical mode ordering corresponding to that grouping method. The effective refractive index interval at the boundary refers to the difference in effective refractive index between two adjacent modes located at the end of the first mode group and the beginning of the second mode group within a candidate group. A larger difference indicates a stronger phase mismatch between the two mode groups. The effective refractive index interval between the first and last polarization degenerate modes within a mode group refers to the difference between the polarization degenerate mode with the highest effective refractive index and the polarization degenerate mode with the lowest effective refractive index within the same mode group. This difference reflects the degree of splitting within the mode group due to numerical errors or weak degeneracy splitting. The preset safety factor is a very small positive number used to prevent the denominator from being zero when calculating the boundary contrast fraction. The average effective refractive index is the arithmetic mean of the effective refractive indices of all polarization degenerate modes within the same LP mode group, used to characterize the overall phase level of that mode group.

[0052] The target numerical simulation method is finite element simulation, the ideal fiber structure is a straight fiber model, the operating wavelength is the C+L band, and the number of target modes is within the range of [missing information]. Adjacent pattern groups are and Taking pattern groups as an example, for , , , , , These six LP (linear polarization mode) groups, in the common order of decreasing effective refractive index (when no overlap occurs), are: , , , , , .

[0053] For candidate fiber profile parameters that satisfy the constraint of the target mode number range, further analysis is performed in COMSOL using FEM at the center operating wavelength. Solve for the modes of a straight fiber. Extract the effective refractive index of the first twenty polarization degenerate modes and arrange them in descending order of their real parts, obtaining the effective refractive index sequence, which can be expressed as: This step is used to subsequently identify each LP mode group and calculate the effective refractive index difference between adjacent mode groups. During optimization, the central concave structure significantly affects modes with a central field intensity distribution, especially... Pattern, therefore and There may be overlap in the effective refractive index order between pattern groups. If grouping is directly fixed according to the magnitude of the effective refractive index, it may lead to incorrect pattern labeling, resulting in inaccurate calculation of the minimum effective refractive index difference and misleading the search direction of the target optimization algorithm. To solve this problem, this embodiment proposes a pattern label-guided implementation method based on local effective refractive index clustering features and boundary contrast scoring: Select a local subset from the twenty sorted effective refractive indices that may intersect. , The mode group contains two polarization degenerate modes. The mode group contains four polarization degenerate modes, according to and Different orderings of pattern groups are used to construct a first candidate group A and a second candidate group B. The first candidate group A is represented as follows: Correspondingly, include , include The second candidate group B is represented as... Correspondingly, include , include To select the group that better conforms to the physical pattern order from the first candidate group A and the second candidate group B, the boundary comparison score of the first candidate group and the second candidate group can be calculated by applying the following boundary comparison scoring formula. The boundary comparison scoring formula can be expressed as: .

[0054] in, It is a small positive number, used to avoid the denominator being zero. Indicates in and The clarity of the boundaries when setting boundary between mode groups. Indicates in and The degree of clarity of the boundary when setting mode group boundaries. A higher score indicates a more pronounced inter-group interval compared to intra-group splitting, and the more likely this grouping method corresponds to the true LP mode group order. The corresponding effective refractive index difference set is defined as: .

[0055] when If the condition is met, the first candidate group A is selected; otherwise, the second candidate group B is selected. This pattern label-guided method can automatically handle the process during the genetic algorithm iteration. and The system addresses the mode group crossover issue, ensuring that the minimum effective refractive index difference in subsequent calculations has a clear physical meaning. After completing mode label identification, the system averages the effective refractive indices of polarization degenerate modes within the same LP mode group to obtain the average effective refractive index of each LP mode group. The average effective refractive index of each LP mode group can be expressed as: Subsequently, the effective refractive index difference between adjacent LP mode groups is calculated, and the minimum value is taken as the weak coupling evaluation index of the candidate structure. .in, A larger value indicates a stronger phase mismatch between adjacent mode groups and a weaker intermode coupling caused by random perturbations. Therefore, this value should be increased as much as possible during the parameter optimization process of weakly coupled few-mode fibers.

[0056] Of course, in the above-mentioned pattern cross-discrimination process, clustering can be replaced by methods such as pattern-field overlap integral tracking, field distribution feature identification, symmetry discrimination, machine learning pattern classification, or multi-criteria fusion, as long as physical pattern groups can be identified and erroneous calculations can be avoided during optimization iterations. This invention does not impose any limitations on this.

[0057] As can be seen from the above, this embodiment, in the process of optimizing structural parameters, introduces pattern label guidance based on clustering and boundary contrast discrimination. This enables automatic identification of possible pattern sequence overlaps under large-scale structural parameter perturbations, avoiding pattern label errors caused by sorting solely based on the effective refractive index. Therefore, The evaluation object always corresponds to the adjacent LP mode group in the real physical sense, thereby improving the stability of the optimization direction and the reliability of the optimization results.

[0058] Based on the above embodiments, the present invention further defines the calculation process for bending loss, which may include the following: A finite element model of a bent fiber is constructed by replacing the material refractive index with the bending equivalent refractive index. The bending equivalent refractive index is determined based on the material refractive index, the transverse coordinate along the bending direction, and the bending radius. A target matching layer is set outside the computational region of the bent fiber finite element model to absorb the leakage field radiated outward. The refractive index distribution information of the bent fiber finite element model is calculated, and the complex effective refractive index of the target higher-order mode is extracted from it. Based on the imaginary part of the complex effective refractive index, the working wavelength, and the preset bending radius, the single-turn bending loss of the target higher-order mode at the preset bending radius is determined.

[0059] The bending equivalent refractive index refers to the transformation of the geometric curvature effect of a bent fiber into the lateral distribution of refractive index in a straight fiber. This equivalent transformation, based on the conformal mapping principle, converts the electromagnetic field problem in a bent waveguide into an equivalent refractive index problem in a straight waveguide, allowing the calculation of radiation loss caused by bending under a straight waveguide model. The bent fiber finite element model is a finite element model of the fiber cross-section established based on the bending equivalent refractive index distribution. Its geometry is that of a straight fiber, but the refractive index distribution varies exponentially or linearly along the lateral direction to simulate the bending effect. The target matching layer is an absorbing boundary condition set outside the computational domain. Its material parameters are complex numbers, enabling the electromagnetic waves leaking from the core to the boundary to attenuate without reflection, thus simulating an infinitely large open boundary. In the finite element method, the target matching layer is a perfectly matched layer. Its number of layers, thickness, and scaling factor are set according to the operating wavelength and bending radius; for example, 8 layers with a thickness twice the wavelength. The complex effective refractive index is the complex form of the propagation constant in the bending mode. Its real part characterizes the phase propagation characteristics, and its imaginary part characterizes the leakage loss characteristics. Single-turn bending loss refers to the proportion of energy lost when an optical signal propagates through a complete circle in an optical fiber, measured in decibels (dB).

[0060] For example, to evaluate the bending robustness of candidate fiber structures, a finite element model of the bent fiber is further established. In this model, the bending effect of the fiber can be described using an equivalent refractive index transformation. The equivalent refractive index after bending can be expressed as: .

[0061] in, For the bending equivalent refractive index, For the refractive index of the material, Let be the lateral coordinate along the direction of curvature. A preset bending radius, such as 10mm, is used. To absorb the leakage field radiating outwards under bending conditions, a PML (perfectly matched layer) is set outside the calculation region. The PML simulates an open boundary, absorbing the leakage light field and preventing boundary reflection from affecting the bending loss calculation. (The bend radius is...) Under the given conditions, considering the target pattern set, Belonging to higher-order modes, these typically exhibit greater bending sensitivity, and their losses can reflect the candidate structure's ability to confine higher-order modes. The target higher-order mode can be... The bending loss can be calculated from the imaginary part of the complex effective refractive index. Correspondingly, the target higher-order mode at the preset bending radius can be calculated according to the following relationship. Single-turn bending loss: .

[0062] in, This indicates the bending loss per revolution. This indicates taking the imaginary part of a complex number. Indicates the operating wavelength.

[0063] As can be seen from the above, this embodiment maps the bent fiber to a calculable straight waveguide model through equivalent refractive index transformation and sets a perfectly matched layer to absorb leakage radiation, thereby obtaining the complex effective refractive index of the mode in the bent state. It can calculate the determined high-order mode bending loss value for each candidate fiber structure during the optimization iteration process, so that the bending loss term in the performance optimization relationship can be quantified and calculated, thereby guiding the target optimization algorithm to search in the direction of reducing bending loss.

[0064] Based on the above embodiments, the present invention also provides an exemplary method for determining performance optimization relationships, which may include the following: Obtain the mode spacing term weight and mode magnitude balance factor corresponding to the minimum effective refractive index difference; obtain the bending loss term weight and loss magnitude balance factor corresponding to bending loss; wherein, the mode spacing term weight is greater than the bending loss term weight; take the negative value of the ratio of the product of the minimum effective refractive index difference and the mode spacing term weight to the mode magnitude balance factor as the first term negatively correlated with the minimum effective refractive index difference; take the positive value of the ratio of the product of bending loss and the bending loss term weight to the loss magnitude balance factor as the second term positively correlated with bending loss; take the sum of the first and second terms as the performance optimization relation; for each target candidate fiber profile structure parameter group whose guided mode number meets the target mode number range, calculate the fiber performance quantification value of each target candidate fiber profile structure parameter group based on the performance optimization relation.

[0065] The mode spacing term weight refers to the weighting coefficient assigned to the minimum effective refractive index difference term in the performance optimization relationship, and its magnitude reflects the importance of this performance index in the overall optimization. The mode magnitude balancing factor is a scaling factor used to normalize the minimum effective refractive index difference to around 1, making its numerical range similar to other terms. The bending loss term weight refers to the weighting coefficient assigned to the bending loss term in the performance optimization relationship. The loss magnitude balancing factor is a scaling factor used to normalize the bending loss to around 1. The mode spacing term weight is greater than the bending loss term weight, reflecting the design emphasis of prioritizing mode spacing over bending loss in weakly coupled transmission. For example, the mode spacing term weight can be 0.9, and the bending loss term weight can be 0.1. In actual implementation, the values ​​of the mode spacing term weight, mode magnitude balancing factor, bending loss term weight, and loss magnitude balancing factor can be adjusted according to specific design requirements. If the system has stricter requirements for bending loss, the weight of the bending loss term can be increased from 0.1 to 0.2 or 0.3; if the fabrication process can achieve a wider refractive index control range, the mode order of magnitude balance factor can be adjusted accordingly to keep all values ​​on the same order of magnitude. The first term is negatively correlated, its value decreases as the minimum effective refractive index difference increases, and the second term is positively correlated, its value increases as bending loss increases. The target candidate fiber profile structure parameter set is a set of candidate fiber profile structure parameters that guide the number of modes to meet the target mode number range. Based on the above performance optimization relationship, the fiber performance quantification value of each target candidate structure is calculated. This performance quantification value is a scalar; the smaller the value, the larger the min Δneff and the smaller L_bend, that is, the better the overall fiber performance. Maximization term and higher-order mode bending loss term Dimensional normalization is performed, and weights are allocated according to the requirements of weakly coupled transmission, so that mode crosstalk suppression is dominant. At the same time, soft constraints are used to prevent bending loss divergence, ultimately obtaining an optical fiber structure that achieves a reasonable trade-off between weakly coupled transmission capability and bending robustness.

[0066] Based on the above embodiments, the performance optimization relationship can be further extended through the following relationship, which may include the following: calculate any one or any combination of the effective mode field area, dispersion coefficient, and differential mode group delay between mode groups for each mode group, as the third term; calculate the target bending loss of at least one target reference high-order mode under a preset bending radius, and adjust the second term according to the positive value of the ratio of the product of each target bending loss and the weight of the bending loss term, and the ratio of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product of the product and ...

[0067] The effective mode area refers to the area of ​​energy distribution of the mode electric field across the cross-section of the optical fiber. It can be calculated, for example, as the integral of the square of the mode electric field intensity over the entire cross-section divided by the square of the peak electric field intensity. A larger effective mode area indicates a weaker nonlinear effect. The dispersion coefficient refers to the rate at which the mode group velocity changes with wavelength. It is calculated as the second derivative of the propagation constant with respect to wavelength. The dispersion coefficient affects the broadening of the optical pulse during transmission. The differential mode group delay refers to the difference in propagation time per unit length between different mode groups. It is calculated as the difference in the reciprocals of the velocities of each mode group at the same wavelength. A larger differential mode group delay requires more taps for MIMO equalization at the receiver. The target reference higher-order mode refers to... It is also necessary to examine other higher-order modes of bending loss, such as , Etc. The combined method for bending loss of multiple modes can be to take the maximum value (representing the worst case), take the average value (representing the overall level), or use a weighted sum (assigning different weights according to the importance of the modes).

[0068] In the specific implementation process, any one or any combination of the effective mode area, dispersion coefficient, and differential mode group delay between mode groups for each mode group is calculated and added as the third term to the performance optimization relation. For example, if it is desired to increase the effective mode area of ​​each mode group to reduce nonlinear effects, the average value A_eff_avg of the effective mode area of ​​all mode groups can be calculated, multiplied by a negative weight, and added to the performance optimization relation so that the performance quantization value decreases when the effective mode area increases. If it is desired to control the differential mode group delay, the sum or maximum value of the absolute values ​​of the differential mode group delays between adjacent mode groups can be calculated, multiplied by a positive weight, and added to the performance optimization relation so that the performance quantization value increases when the differential mode group delay increases. The weight and balance factor of the third term are set according to the same principles as the first and second terms to ensure that the values ​​of each term are on the same order of magnitude. Meanwhile, in the bent fiber finite element model, in addition to Outside of the mode, additional extraction , , The imaginary part of the complex effective refractive index of each mode is used to calculate the single-turn bending loss of each mode in the manner described in the above embodiments. Then, the second term is adjusted by the positive value of the ratio of the target bending loss to the weight of the bending loss term. The maximum value of the bending loss of multiple modes is substituted into the second term to constrain the bending loss of all higher-order modes from exceeding the target value; or the average value is taken to control the overall bending loss level.

[0069] The performance of weakly coupled few-mode fibers in practical transmission systems depends not only on mode spacing and bending loss, but also on factors such as nonlinear effects, dispersion, and mode delay differences. This embodiment incorporates additional performance indicators such as effective mode field area, dispersion coefficient, and differential mode group delay into the performance optimization relationship, and comprehensively considers multi-mode bending loss, thereby expanding the scope of optimization objectives. The objective optimization algorithm can simultaneously take into account multiple transmission performance indicators, and the output fiber structure has a more comprehensive performance in practical system applications, avoiding the unfavorable situation where the mode spacing meets the requirements but nonlinear impairment is severe or dispersion is too large.

[0070] In practical applications, the performance optimization relationship can be represented by an objective cost function, which can be expressed as: ; in, This is the cost function value, which is also the quantified value of the fiber performance in the above embodiments. For the weight of the pattern interval, For example, the weights for the bending loss term are... Normalization factor 0.1 and 0.1 represent the numerical magnitudes of the balancing mode interval term and the bending loss term, respectively. The first term in the objective cost function is negative, indicating that the genetic algorithm minimizes the cost function. At that time, there will be a tendency to increase The second term is positive, indicating that the greater the bending loss, the larger the cost function value, and the less optimal the candidate structure. Therefore, this cost function can simultaneously achieve the dual optimization objectives of increasing mode spacing and suppressing bending loss. In weakly coupled mode division multiplexing systems, suppressing mode crosstalk is the primary objective, and mode crosstalk is mainly determined by the effective refractive index difference between adjacent mode groups. Therefore, setting a larger cost function is crucial. Prioritize increasing the pattern interval; at the same time, As a soft constraint, it prevents excessive bending losses in higher-order modes. Furthermore, the objective cost function can be extended to a multi-objective form, for example, by incorporating effective area, dispersion, differential mode group delay (DMGD), fabrication error sensitivity, profile smoothness, upper limit of doping concentration, or measured profile deviation terms; bending losses can also be determined from... The single-mode average value is extended to the maximum value, average value, or weighted sum of multiple higher-order modes, which does not affect the implementation of the present invention.

[0071] Based on the above embodiments, the present invention further specifies the target optimization algorithm as a genetic algorithm. The smaller the value of the fiber performance quantification, the better the fiber performance represented by the corresponding candidate fiber profile structure parameter group. In the implementation of the genetic algorithm, the process of updating the candidate fiber profile structure parameter group is as follows: the candidate fiber profile structure parameter group with the smallest fiber performance quantification value in the current generation is taken as the first type of candidate fiber profile structure parameter and directly retained to the next generation; multiple parent parameter groups are selected from the current generation according to the fiber performance quantification value, and some parameters are exchanged between any two parent parameter groups to generate a child parameter group; random perturbation values ​​are added to some structural parameters in the child parameter group within the value boundary range to obtain the second type of candidate fiber profile structure parameter; the first type of candidate fiber profile structure parameter group, parent parameter group, child parameter group and second type of candidate fiber profile structure parameter group are taken as the new candidate fiber profile structure parameter group.

[0072] Genetic algorithms are iterative search methods that simulate the mechanisms of natural selection and genetic mutation in biological evolution. Since fiber optic profile design problems involve non-convex objective cost functions that are not closed differentiable expressions, genetic algorithms are well-suited for this optimization task. They can handle complex design constraints, require no gradient information, and provide global search capabilities for multi-dimensional parameter spaces. In the current generation of optimization, the candidate fiber profile structure parameters with the smallest fiber performance quantification values ​​are selected as the first type of candidate fiber profile structure parameters and treated as elite individuals. These are not subjected to crossover or mutation operations to prevent the loss of superior genes during evolution and are directly retained to the next generation. Specifically, the current population is sorted by performance quantification values ​​from smallest to largest. The top two optimal individuals (the number of elite individuals is 8% of the population size) are selected and copied to the next generation population, without participating in subsequent selection, crossover, or mutation operations. This operation ensures that the optimal solution is not destroyed by random operations. In the process of selecting multiple parent parameter sets from the current generation based on fiber performance quantification values, the selection probability of each individual in the current population can be calculated. The selection probability is proportional to the reciprocal of the performance quantification value; the smaller the performance quantification value, the larger the reciprocal, and the higher the selection probability. Then, 20 parent individuals are randomly selected according to these probabilities. From the selected parent individuals, they are randomly paired, and a simulated binary crossover operator is used on each pair of parent individuals to exchange the encoded values ​​on one or more dimensions of the parameter vector. For each structural parameter of each offspring individual, a mutation probability of 0.1 is used to determine whether mutation occurs. If mutation occurs, a random perturbation following a normal or uniform distribution is added within the search range of the parameter, with the perturbation amplitude being 5% to 10% of the parameter search range. The first set of candidate fiber profile structure parameters (elite individuals), the parent parameter set (selected parent individuals, or whether to retain them depending on the algorithm configuration), the offspring parameter set (generated through crossover), and the second set of candidate fiber profile structure parameters (generated through mutation) are merged to form a new set of candidate fiber profile structure parameters for simulation and evaluation of the next generation. The size of the new population remains the same as the previous generation, such as 25 individuals. Two elite individuals are directly retained, and the remaining 23 individuals are selected from the offspring and mutated individuals based on their performance quantification values, or all are generated through crossover and mutation.

[0073] As can be seen from the above, this embodiment ensures that the optimal parameter combination is not lost through elite retention, the crossover operation combines excellent parameter fragments into the same candidate solution, and the mutation operation maintains the exploration breadth of the parameter space. The combination of the three enables the algorithm to effectively converge to the globally better combination of fiber profile parameters in the multidimensional structural parameter space, avoiding getting trapped in local extrema.

[0074] Based on the above embodiments, this embodiment further defines the method for generating the initial candidate parameter group, which may include the following: The radius of the dominant optical core region is used as a fixed reference structural parameter. The depth of the central depression, the depth of the trench, the radius of the central depression, the starting position of the trench, and the width of the trench are used as structural parameters to be optimized. According to the constraints of the optical fiber fabrication process, the corresponding parameter search boundary range is set for each structural parameter to be optimized. Multiple sets of structural parameters to be optimized are randomly generated within the search boundary range of each parameter to serve as the initial candidate optical fiber profile structural parameter sets for the genetic algorithm.

[0075] The dominant core radius refers to the radial distance from the fiber axis to the outer boundary of the dominant core region. It determines the waveguide capability of the fiber; a larger core radius supports more guided modes, while a smaller core radius helps reduce bending loss. Fixed reference structural parameters are structural parameters that remain constant during optimization and are not used as optimization variables. Their values ​​are predetermined based on design experience or process convenience. The central recess depth refers to the refractive index difference (expressed as a relative value) between the central recess region and the dominant core region. Its magnitude determines whether the central recess pair has a central field strength distribution pattern (e.g., ...). , The trench depth refers to the refractive index difference (expressed as a relative value) between the trench structure region and the cladding, and its magnitude determines the binding strength of the trench on higher-order modes. The central depression radius refers to the outer boundary radius of the central depression region, and its magnitude determines the radial range of the central depression effect. The trench starting position refers to the radial distance from the inner boundary of the trench structure region to the fiber axis, and the trench width refers to the radial thickness of the trench structure region. The parameter search boundary range refers to the lower and upper limits of the allowed values ​​of each structural parameter to be optimized during the optimization process. This range is determined by the actual capabilities of the fiber preform fabrication process (such as the achievable range of doping concentration, interface broadening caused by diffusion effects, and dimensional control accuracy during the fiber drawing process). For example, the central depression depth is constrained by the upper limit of doping concentration, and the trench width and position are constrained by the dimensional control accuracy of diffusion effects during the fiber drawing process.

[0076] In this embodiment, considering manufacturing feasibility and production tolerance, corresponding parameter search boundary ranges are set for the structural parameters to be optimized. These boundaries are used to prevent the genetic algorithm from generating structural parameters that are impossible to fabricate or lack light-guiding significance. For example, the parameter search boundary ranges for the central recess depth, trench depth, central recess radius, trench position, and trench width are: .

[0077] in, The units are all in μm.

[0078] Once the parameter search boundary range is determined, values ​​are randomly selected within the given parameter search boundary range to form an initial population. The population size or candidate set size can be set according to the computing resources and optimization accuracy requirements, for example, 25 or 50 groups.

[0079] As can be seen from the above, the candidate structures in the initial population of this embodiment are distributed throughout the entire feasible parameter space, which avoids inefficient simulation caused by the initial population deviating from the effective region. At the same time, fixing the baseline parameters reduces the dimensionality of the optimization variables, which helps to accelerate the convergence speed of the algorithm.

[0080] It should be noted that there is no strict order of execution for the steps in this application. As long as they conform to a logical order, these steps can be executed simultaneously or in a certain preset order. Figure 1 This is just an illustrative example and does not mean that this is the only possible execution order.

[0081] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention also provides several exemplary application scenarios. The method for determining the structural parameters of a few-mode fiber provided in this application can be implemented as a computer program product, installed and run in an optical fiber design and simulation analysis workstation, for the purpose of implementing automated optimization processing of the refractive index profile of a few-mode fiber.

[0082] In some embodiments of this method, the fiber optic design and simulation analysis workstation includes at least one processor and a memory connected to the processor. The memory stores executable instructions. When the instructions are executed by the processor, the workstation performs the following operations: reads the number of target modes, working bands, preset bending radii, and search boundaries of each structural parameter to be optimized, set by the user through an input interface, into the memory; calls the target optimization algorithm to generate multiple candidate fiber profile structural parameter sets within the search boundaries as the first round of candidate fiber profile structural parameter sets; for each candidate parameter set, calls the finite element simulation solver to establish a straight fiber model and a bent fiber model, respectively... The algorithm calculates the number of guided modes, the minimum effective refractive index difference between adjacent mode groups, and the bending loss of the target higher-order mode at a preset bending radius. Based on the performance optimization relationship, which includes a first term negatively correlated with the minimum effective refractive index difference and a second term positively correlated with the bending loss, it calculates the fiber performance quantization value corresponding to each candidate parameter group. Based on the fiber performance quantization value of each candidate parameter group, it updates the candidate parameter groups through selection, crossover, and mutation operations, and repeats the simulation and quantization calculation until the preset termination condition is met. The final optimal fiber profile structure parameters are presented to the user through the output interface or directly transmitted to the parameter configuration terminal of the fiber preform fabrication equipment.

[0083] In a practical application scenario, this computer program product can be deployed on a dedicated workstation used in the fiber preform production process. The process designer inputs the refractive index adjustment range achievable by the current fabrication process (e.g., maximum center recess depth of 0.018, maximum trench depth of 0.006, manufacturing tolerances for various geometric dimensions, etc.) into the workstation, and sets the target to support 10 spatial modes (corresponding to 20 polarization degenerate modes) and a bending radius of 10 mm in the C+L band. The workstation automatically runs the above method, outputting a set of optimal structural parameter combinations within hours to days. The process designer converts this parameter combination into preform doping concentration and drawing speed control parameters, which are then fed into the fiber drawing tower's control system, thereby fabricating a center-recessed trench-assisted few-mode fiber that meets the requirements for weakly coupled transmission.

[0084] In another practical application scenario, this computer program product can be integrated into the simulation module of an optical communication system link design platform. When simulating a mode-division multiplexing transmission system, system designers need to select or design appropriate few-mode fiber parameters. The platform calls this program product and automatically calculates the optimal fiber profile parameters based on the system's requirements for the number of modes, operating bands, and bending resistance. These parameters are then provided to the link simulator to calculate system capacity, crosstalk level, and digital signal processing complexity, assisting system designers in evaluating the impact of different fiber parameters on overall transmission performance.

[0085] In another practical application scenario, this computer program product can run on a virtual machine on a cloud computing platform, providing few-mode fiber structure parameter optimization services to multiple users as a service. Users submit optimization requirements (target number of modes, operating band, bending radius, fabrication constraints, etc.) through a web page or application programming interface. The cloud platform automatically allocates computing resources to execute the above methods and returns the optimization results to the users. This deployment method allows design units without high-performance computing hardware to obtain optimization results remotely, lowering the technical application threshold.

[0086] The specific hardware configurations of the aforementioned workstations, link design platforms, or cloud computing platforms include, but are not limited to: a central processing unit based on x86 or ARM architecture, a graphics processor for accelerating finite element matrix solutions, a solid-state drive for storing simulation data and intermediate results, and a human-computer interaction interface or network interface for receiving user input and outputting optimization results. Computer program products exist in the form of compiled executable files, interpreted scripting language programs, or dynamically linked libraries on the non-volatile storage media of the workstation or platform.

[0087] Finally, this invention also presents a method for determining the optimal fiber profile structure parameters for few-mode fibers using a genetic algorithm (GA) and finite element simulation, employing MATLAB and COMSOL Multiphysics. MATLAB is used to control the iteration, parameter selection, crossover, mutation, fitness recording, and termination determination of the GA, while COMSOL Multiphysics is used to calculate the modal characteristics and bending loss of each candidate fiber profile structure parameter set. Figure 2 The initial structure of the few-mode fiber, with a central concave radius Dominant core region radius trench starting radius Groove width d, center depression depth and trench depth As a joint optimization parameter, it will dominate the core region radius. As fixed reference parameters, the effective refractive index of different LP modes is adjusted by the central concave structure to increase the effective refractive index difference between adjacent mode groups, thereby reducing inter-mode coupling. Secondly, the binding ability of higher-order modes is enhanced by the low-refractive-index grooves on the outside, reducing the radiation loss of the fiber under bending conditions. Thirdly, the five structural parameters are automatically optimized using a genetic algorithm to ensure that the fiber simultaneously meets the requirements of the target number of modes, a larger mode spacing, and a lower bending loss. Finally, a closed-loop optimization process is formed through finite element simulation and performance verification, such as... Figure 3 As shown, it may include the following: The optimization target is determined to be DCTA-FMF. The input design requirements include at least: target number of modes, operating wavelength, bending radius, weak coupling criterion, fiber fabrication process constraints, material refractive index range, and structural size constraints. A parameterized refractive index profile model of the centrally recessed trench-assisted few-mode fiber is established. This model includes at least a central recessed region, a dominant core region, an outer low-refractive-index trench region, and a cladding region along the fiber radially. Multiple structural parameters to be optimized are defined as follows: , The parameter search boundary range for each structural parameter to be optimized is: An initial population is randomly generated within the given parameter boundaries, with an initial population size of [value missing]. Then the i-th candidate fiber profile parameter set can be expressed as Genetic algorithms perform iterative calculations according to generations, with a maximum number of generations. The iterative process can be represented as For each candidate individual in the g-th generation, the system sequentially performs finite element simulation, mode number check, mode label identification, mode interval calculation, bending loss calculation, and objective cost function evaluation.

[0088] For each candidate fiber profile parameter group, automatically The candidate fiber profile parameters are written into the finite element simulation model, which automatically updates the fiber cross-sectional geometry, material refractive index distribution, and mesh settings based on these candidate parameters. To ensure the fiber supports the set target mode count range throughout the C+L band, a mode count check is performed at the long-wavelength end. If the candidate fiber profile parameter set does not meet the constraint of the target mode count range, it is deemed unqualified, assigned a large penalty fitness value in the cost function, and skipped from subsequent fine simulations. For candidate fiber profile parameter sets that meet the constraints, the FEM is used in COMSOL at the center operating wavelength. The mode selection for a straight fiber is determined. The effective refractive indices of the first twenty polarization degenerate modes are extracted and arranged in descending order of their real parts to obtain an effective refractive index sequence. Then, the minimum effective refractive index difference is calculated using the method described in the previous embodiment. Furthermore, bending loss is calculated based on the method described in the previous embodiment. Once the minimum adjacent effective refractive index difference and bending loss are calculated, the target cost function is constructed, and the cost function values ​​for candidate fiber profile parameter sets that satisfy the target mode number range are calculated. After completing the mode number check, mode tag identification, minimum effective refractive index difference calculation, bending loss calculation, and cost function calculation for the current candidate individual, the evaluation results for that candidate individual are saved. The saved content may include the candidate fiber profile parameter set for the current round. Pattern sorting results , and cost function value , can be represented as Then determine whether all individuals in the current population have been evaluated. If not, then... The process returns to the step of writing candidate parameters into the finite element model and continues to evaluate the next candidate individual. After all candidate individuals in the current generation have been evaluated, genetic operations are performed based on the cost function value, including elite retention, selection, crossover, and mutation. Elite retention is used to directly retain several individuals with the best performance in the current generation, avoiding the loss of excellent parameter combinations in subsequent random operations. The selection operation selects parent individuals based on fitness values, so that candidate structures with better objective cost functions have a higher reproduction probability. The crossover operation is used to exchange some parameters between two parent individuals to generate new candidate structures. The mutation operation is used to randomly perturb some parameters to increase population diversity and prevent the algorithm from getting trapped in local optima. After generating the next generation of candidate parameters, the generation number is updated to... Then, the finite element evaluation process is restarted. It is then determined whether the maximum number of iterations has been reached. Alternatively, the optimal objective cost function may vary less than a preset threshold for several consecutive generations. When the termination condition is met, the optimal fiber profile structure parameters are output. It also outputs the corresponding performance indicators, including the minimum effective refractive index difference between adjacent modes, the bending loss of the target higher-order modes, and the cost function value.

[0089] like Figure 4 As shown, the optimal cost function value is obtained in the first 20 generations of the genetic algorithm iteration. The rapid and monotonic decline indicates the effectiveness of the global search in the multidimensional parameter space. The curve then flattens out and converges stably after approximately 60 generations. The final convergent objective cost function value is... The corresponding optimal fiber optic profile structure parameters are: In this optimal state, the trench depth is... and trench width All reached their upper limits, indicating that, within manufacturing feasibility, stronger trench confinement is beneficial for suppressing higher-order modes. Bending loss. Due to manufacturing constraints, these parameters were not further increased. Figure 5 This shows the minimum effective refractive index difference between adjacent mode groups for the optimal individual during operation. With bending loss The curve shows the change with the number of iterations. Ultimately, the minimum adjacent effective refractive index difference reaches... Compared to SI-FMF, the effective refractive index difference is significantly improved, and it also exceeds the generally accepted... Weak coupling threshold. The bending loss of the higher-order mode is The above results demonstrate that GA optimization can maximize mode spacing while suppressing higher-order mode bending loss, provided that weak coupling is maintained. The optimized fiber parameters can be directly used in subsequent fiber fabrication.

[0090] As can be seen from the above, this embodiment unifies the structural variables of the fiber refractive index profile, the target mode number constraint, mode cross-discrimination, maximization of the effective refractive index difference between adjacent mode groups, and the bending loss constraint of higher-order modes into the same automatic optimization closed loop. Therefore, it overcomes the problem that traditional manual parameter scanning can only find local optimizations in one-dimensional or two-dimensional parameter spaces. Specifically, this embodiment uses the central depression depth, trench depth, central depression radius, trench position, and trench width as joint optimization variables. A genetic algorithm is used to automatically search for candidate profiles in the five-dimensional structural parameter space, and finite element simulation is used to evaluate the number of guided modes, the effective refractive index difference, and the bending loss of each candidate structure. Thus, this invention can achieve coordinated optimization of mode spacing and bending robustness under the condition of multi-variable coupling, rather than relying on manual experience to adjust parameters one by one. According to experimental results, the minimum effective refractive index difference between adjacent mode groups of the optimized structure can reach approximately... ,at the same time The bending loss of the mode at a bending radius of 10mm is approximately This demonstrates that the technical solution provided by the present invention can simultaneously achieve a large weakly coupled mode spacing and a low higher-order mode bending loss.

[0091] This application also provides a corresponding apparatus for determining the structural parameters of few-mode optical fibers, further enhancing the practicality of the method. The apparatus can be described from both a functional module perspective and a hardware perspective. The apparatus for determining the structural parameters of few-mode optical fibers provided in this application is described below. This apparatus is used to implement the method for determining the structural parameters of few-mode optical fibers provided in this application. In this embodiment, the apparatus may include or be divided into one or more program modules. These program modules are stored in a storage medium and executed by one or more processors to complete the method for determining the structural parameters of few-mode optical fibers disclosed in this embodiment. The program module referred to in this embodiment is a series of computer program instruction segments capable of performing specific functions, which is more suitable than the program itself for describing the execution process of the apparatus for determining the structural parameters of few-mode optical fibers in the storage medium. The following description will specifically introduce the functions of each program module in this embodiment. The apparatus for determining the structural parameters of few-mode optical fibers described below can be referred to in correspondence with the method for determining the structural parameters of few-mode optical fibers described above.

[0092] From the perspective of functional modules, the few-mode fiber structure parameter determination device provided in this embodiment may include: The structural parameter optimization module is used to jointly optimize multiple structural parameters of a few-mode fiber with an axisymmetric structure whose refractive index profile, from the fiber axis radially, includes a central recess region, a dominant core region, a trench structure region, and a cladding region. The refractive indices of the central recess region and the trench structure region are both lower than those of the dominant core region. Multiple candidate fiber profile structural parameter sets are generated in the multi-dimensional structural parameter space using a target optimization algorithm. For each candidate fiber profile structural parameter set, those with a guided mode number that does not meet the target mode number range are assigned invalidation values ​​to prevent selection. A performance optimization relationship is determined based on at least a first term negatively correlated with the minimum effective refractive index difference and a second term positively correlated with bending loss, used to calculate the fiber performance quantification values ​​of each candidate fiber profile structural parameter set. The candidate fiber profile structural parameter sets are updated based on the fiber performance quantification values ​​of each candidate fiber profile structural parameter set to obtain the next round of candidate fiber profile structural parameter sets. Repeat the above process until the preset termination condition is met, and use the candidate fiber profile structure parameter set determined in the last round as the optimal fiber profile structure parameter for the few-mode fiber: The simulation module is used to calculate the number of guided modes of the corresponding fiber structure, the minimum effective refractive index difference between adjacent mode groups, and the bending loss of the target higher-order mode at a preset bending radius using the target numerical simulation method.

[0093] The device for determining the structural parameters of a few-mode fiber mentioned above is described from the perspective of a functional module. Furthermore, this application also provides an electronic device, which is described from the perspective of hardware. Figure 6 This is a schematic diagram of the structure of an electronic device provided in one embodiment of this application. The electronic device includes a memory 601 and a processor 602. The memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any of the above embodiments of the method for determining the structural parameters of a few-mode fiber.

[0094] It is understood that if the few-mode fiber structure parameter determination method in the above embodiments is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and executes all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes, but is not limited to, various media capable of storing program code, such as: USB flash drive, mobile hard drive, read-only memory (ROM), random access memory (RAM), electrically erasable programmable ROM, register, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, removable disk, CD-ROM, magnetic disk, or optical disk. Based on this, this application also provides a non-volatile storage medium storing a computer program, which, when executed by a processor, performs the steps of the few-mode fiber structure parameter determination method as described in any of the above embodiments.

[0095] It is understood that if the few-mode fiber structure parameter determination method in the above embodiments is implemented as a software functional unit and sold or used as an independent product, the computer software product may not need to be stored in a physical storage medium. For example, it can be directly transmitted to a computer or other device with information processing capabilities via a wired or wireless network to execute all or part of the steps of the methods in the various embodiments of this application. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, or all or part of the technical solution, can be embodied in the form of a software product. Based on this, this application also provides a computer program product storing a computer program, which, when executed by a processor, performs the steps of the few-mode fiber structure parameter determination method as described in any of the above embodiments.

[0096] Any of the components, modules, units, parts, methods, and operations described herein can be implemented using software, firmware, hardware (e.g., fixed logic circuitry), manual processing, or any combination thereof. Alternatively or additionally, any functionality described herein can be performed at least in part by one or more hardware logic components, such as, but not limited to, a central processing unit (CPU), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), an application-specific standard product (ASSP), a system-on-a-chip (SoC), a complex programmable logic device (CPLD), a microprocessor (MCU), etc. The systems, computing devices, or apparatuses described herein encompass a wide range of means, devices, and machines for processing data, including, for example, one or more programmable processors, computers, SoCs, or combinations thereof. The apparatus may also include code that creates an execution environment for the computer program in question, such as code constituting processor firmware, a protocol stack, a database management system, an operating system, a cross-platform runtime environment, a virtual machine, or one or more combinations thereof. The aforementioned computer program (also known as a program, software, software application, app, script, or code) can be written in any form of programming language, including compiled or interpreted languages, declarative or procedural languages, and can be deployed in any form, including as a standalone program or as a module, component, subroutine, object, or other unit suitable for a computing environment.

[0097] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0098] The foregoing has provided a detailed description of the method, apparatus, electronic device, computer-readable storage medium, and computer program product for determining the structural parameters of a few-mode optical fiber. The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Several improvements and modifications can be made to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for determining the structural parameters of a few-mode optical fiber, characterized in that, include: Multiple structural parameters of a few-mode fiber with an axisymmetric structure and a refractive index profile that includes a central recess region, a dominant core region, a trench structure region, and a cladding region in a radial direction from the fiber axis are used as joint optimization parameters. Multiple candidate fiber profile structural parameter sets are generated in a multi-dimensional structural parameter space using a target optimization algorithm. The refractive indices of the central recess region and the trench structure region are both lower than the refractive index of the dominant core region. The following process is repeated for each set of candidate fiber profile structure parameters determined in each round until a preset termination condition is met, and the set of candidate fiber profile structure parameters determined in the last round is taken as the optimal fiber profile structure parameters for the few-mode fiber: The target numerical simulation method is used to calculate the number of guided modes of the corresponding fiber structure, the minimum effective refractive index difference between adjacent mode groups, and the bending loss of the target higher-order mode at the preset bending radius. For candidate fiber profile structure parameter groups whose number of guided modes does not meet the target mode number range, invalidation values ​​are assigned to prevent them from being selected. The performance optimization relationship used to calculate the fiber performance quantification value of each candidate fiber profile structure parameter group is determined according to at least a first term that is negatively correlated with the minimum effective refractive index difference and a second term that is positively correlated with bending loss. Based on the fiber performance quantification values ​​of each candidate fiber profile structure parameter group, the candidate fiber profile structure parameter group is updated to obtain the next round of candidate fiber profile structure parameter groups.

2. The method for determining the structural parameters of a few-mode fiber according to claim 1, characterized in that, The number of guided modes of the corresponding fiber structure for the corresponding candidate fiber profile is calculated using the target numerical simulation method, including: Numerical simulation calculations are performed on the ideal fiber structure corresponding to the cross-sectional structural parameters of each candidate fiber at the long wavelength end of the working band. Extract the real part of the effective refractive index of each mode, identify the mode whose real part of the effective refractive index is greater than the cladding refractive index as the guided mode, and count the total number of polarization degenerate modes of the guided mode. For candidate fiber profile structure parameter groups whose total number of polarization degenerate modes of the guided mode is not within the range of the target number of modes, invalidation values ​​are assigned, and they are set not to participate in the fiber performance quantification calculation process; for candidate fiber profile structure parameter groups whose total number of polarization degenerate modes of the guided mode is within the range of the target number of modes, they are set to participate in the fiber performance quantification calculation process.

3. The method for determining the structural parameters of a few-mode fiber according to claim 2, characterized in that, Before calculating the number of guided modes of the corresponding fiber structure for the corresponding candidate fiber profile using the target numerical simulation method, the following steps are also included: The total number of polarization degenerate modes corresponding to the target spatial mode is taken as the lower limit of the target mode number range, and the sum of the total number of polarization degenerate modes and the preset margin is taken as the upper limit of the target mode number range. Obtain the working band and preset bending radius; The target mode quantity range, the working waveband, and the preset bending radius are input into the finite element simulation model.

4. The method for determining the structural parameters of a few-mode fiber according to claim 1, characterized in that, The minimum effective refractive index difference between adjacent mode groups of the corresponding fiber structure on the candidate fiber profile is calculated using the target numerical simulation method, including: Numerical simulation calculations were performed on the ideal fiber structure at the center wavelength of the working band to extract the effective refractive index of the polarization degenerate modes that meet the target mode number range. The real part of the effective refractive index was sorted by value to obtain the effective refractive index sequence. Based on the number of polarization degenerate modes in adjacent mode groups and the continuous arrangement of each polarization degenerate mode, the effective refractive index corresponding to the adjacent mode group is selected from the effective refractive index sequence to form a refractive index cross subset; the adjacent mode group includes a first mode group and a second mode group. Following the order in which the first mode group precedes the second mode group, and based on the number of polarization degenerate modes contained in each of the first and second mode groups, the refractive index cross subset is constructed as a first candidate group; following the order in which the second mode group precedes the second mode group, and based on the number of polarization degenerate modes contained in each of the second and first mode groups, the refractive index cross subset is constructed as a second candidate group. Based on the effective refractive index interval at the boundary between the first mode group and the second mode group, the effective refractive index interval corresponding to the first and last polarization degenerate modes of the first mode group, the effective refractive index interval corresponding to the first and last polarization degenerate modes of the second mode group, and the preset safety factor, the boundary comparison scores of the first candidate group and the second candidate group are calculated respectively. The candidate group with the larger boundary contrast score is selected as the polarization degenerate mode partitioning result of the adjacent mode group; Calculate the average effective refractive index of each polarization degenerate mode within the same mode group, and calculate the average effective refractive index difference between adjacent mode groups. The minimum of the average effective refractive index differences is taken as the minimum effective refractive index difference between adjacent mode groups.

5. The method for determining the structural parameters of a few-mode fiber according to claim 1, characterized in that, The bending loss of the corresponding fiber structure in the target higher-order mode at a preset bending radius is calculated using the target numerical simulation method, including: A finite element model of a bent optical fiber is constructed by replacing the material refractive index with the bending equivalent refractive index; the bending equivalent refractive index is determined based on the material refractive index, the transverse coordinate along the bending direction, and the bending radius. A target matching layer is set outside the computational region of the bent optical fiber finite element model. The target matching layer is used to absorb the leakage field radiated outward. Calculate the refractive index distribution information of the bent fiber finite element model, and extract the complex effective refractive index of the target higher-order mode from it; Based on the imaginary part of the complex effective refractive index, the operating wavelength, and the preset bending radius, the single-turn bending loss of the target higher-order mode at the preset bending radius is determined.

6. The method for determining the structural parameters of a few-mode optical fiber according to any one of claims 1 to 5, characterized in that, Determine the performance optimization relationships used to calculate the quantized values ​​of fiber performance for each candidate fiber profile structural parameter group, including: Obtain the mode spacing term weight and mode magnitude balance factor corresponding to the minimum effective refractive index difference; Obtain the weight of the bending loss term and the loss magnitude balance factor corresponding to the bending loss; wherein, the weight of the mode interval term is greater than the weight of the bending loss term; The negative value of the ratio of the product of the minimum effective refractive index difference and the weight of the mode spacing term to the mode magnitude balance factor is taken as the first term negatively correlated with the minimum effective refractive index difference. The positive value of the ratio of the product of bending loss and the weight of the bending loss term to the loss magnitude balance factor is taken as the second term that is positively correlated with bending loss. The sum between the first term and the second term is used as the performance optimization formula; For each target candidate fiber profile structure parameter group whose number of guided modes meets the target mode number range, the fiber performance quantification value of each target candidate fiber profile structure parameter group is calculated based on the performance optimization relationship.

7. The method for determining the structural parameters of a few-mode fiber according to claim 6, characterized in that, The sum of the first and second terms is determined as the performance optimization relation, including: Calculate any one or any combination of the effective mode field area, dispersion coefficient, and differential mode group delay between mode groups for each mode group, and use it as the third term; Calculate the target bending loss of at least one target reference high-order mode at a preset bending radius, and adjust the second term according to the positive value of the ratio of the product of each target bending loss and the weight of the bending loss term. The sum of the first term, the adjusted second term, and the third term is taken as the performance optimization relation.

8. The method for determining the structural parameters of a few-mode optical fiber according to any one of claims 1 to 5, characterized in that, The target optimization algorithm is a genetic algorithm. The smaller the value of the fiber performance quantification, the better the fiber performance represented by the corresponding candidate fiber profile structure parameter set. Updating the candidate fiber profile structure parameter set includes: The candidate fiber profile structure parameter group with the smallest fiber performance quantification value in the current generation is taken as the first type of candidate fiber profile structure parameter and directly retained to the next generation. Based on the fiber performance quantification value, select multiple parent parameter groups from the current generation, and exchange some parameters between any two parent parameter groups to generate child parameter groups; For some structural parameters in the sub-parameter group, random perturbation values ​​are added within the value boundary range to obtain the second type of candidate fiber profile structural parameters. The first type of candidate fiber profile structure parameter group, the parent parameter group, the child parameter group, and the second type of candidate fiber profile structure parameters are used as the new candidate fiber profile structure parameter group.

9. The method for determining the structural parameters of a few-mode optical fiber according to claim 8, characterized in that, Before generating multiple candidate fiber profile structural parameter sets in the multidimensional structural parameter space using the objective optimization algorithm, the following steps are also included: The radius of the dominant optical core region is used as a fixed reference structural parameter, and the depth of the central depression, the depth of the trench, the radius of the central depression, the starting position of the trench, and the width of the trench are used as structural parameters to be optimized. The corresponding parameter search boundary range is set for each structural parameter to be optimized according to the constraints of the optical fiber fabrication process. Multiple sets of structural parameters to be optimized are randomly generated within the search boundary range of each parameter to serve as the initial candidate fiber profile structural parameter sets for the genetic algorithm.

10. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the method for determining the structural parameters of a few-mode fiber as described in any one of claims 1 to 9.

Images