A Gain Optimization Design Method for O-Band Raman Amplifiers Based on Tree Species Optimization Algorithm

Abstract

This invention discloses a method for optimizing the gain of an O-band Raman amplifier based on a tree species optimization algorithm, comprising the following steps: randomly selecting several sets of pump parameters as initial tree species; calculating the Raman amplifier gain value of the initial tree species, using it as the fitness value of the optimization algorithm, and adjusting the gain of the initial tree species to obtain a dominant tree species; randomly generating multiple sets of first and second tree species near the dominant tree species, and merging the first and second tree species as second-generation tree species; calculating the second-generation fitness value of the second-generation tree species, and performing several iterations of optimization on the second-generation fitness value according to the above steps to obtain the optimal fitness value; and outputting the corresponding optimal tree species based on the optimal fitness value. This invention solves the multi-objective optimization problem regarding pump parameters and effectively improves the transmission distance of O-band signals.

Description

Technical Field

[0001] This invention relates to the field of O-band Raman amplifier gain optimization technology, specifically to an O-band Raman amplifier gain optimization design method based on tree-type optimization algorithm. Background Technology

[0002] Ultra-long-distance fiber optic transmission systems have extremely broad application prospects in special communication channels, such as underwater fiber optic communication. Compared with other optical amplifiers, fiber Raman amplifiers have advantages such as full-band operation, ultra-long-distance amplification, and low noise figure. Currently, the common approach to designing gain-flat Raman amplifiers is to combine optimization algorithms with numerical solutions to coupled-wave equations. To match the commonly used low-loss fiber windows, most common algorithm-optimized Raman amplifiers operate in the low-loss C or C+L bands. However, performance optimization for Raman amplifiers with different pump structures in special bands, especially the O band, is lacking, as the transmission distance of O band signals is relatively short. Summary of the Invention

[0003] In view of this, the purpose of this invention is to provide a method for optimizing the gain of an O-band Raman amplifier based on a tree species optimization algorithm, so as to solve the technical problems mentioned in the prior art.

[0004] The O-band Raman amplifier gain optimization design method based on tree species optimization algorithm is characterized by the following steps:

[0005] S1. Randomly select several sets of pumping parameters as the initial tree species;

[0006] S2. Calculate the Raman amplifier gain value of the initial tree species, use it as the fitness value of the optimization algorithm, and adjust the gain of the initial tree species to obtain the dominant tree species;

[0007] S3. Randomly generate multiple groups of first tree species and second tree species near the dominant tree species, and merge the first tree species and the second tree species as second-generation tree species;

[0008] S4. Calculate the second-generation fitness value of the second-generation tree species, and perform several iterations to optimize the second-generation fitness value according to steps S2 and S3 to obtain the optimal fitness value.

[0009] S5. Output the corresponding optimal tree species based on the optimal fitness value.

[0010] Optionally, in S1, the optimal pump band search range and pump power search range corresponding to the amplified signal are selected according to the Raman amplifier with different structures.

[0011] Based on the selected pump band search range and pump power search range, a corresponding number of pump parameters are randomly generated.

[0012] Optionally, S2 further includes:

[0013] The initial tree species is used as the initial value of the stimulated Raman coupled-wave equation, and the stimulated Raman coupled-wave equation is solved by numerical analysis to obtain the signal output power of the initial tree species.

[0014] The gain value of the corresponding wavelength signal is calculated by comparing the signal output power of the initial tree species with the input signal power of an arbitrary wavelength signal.

[0015] Calculate the average gain and gain flatness of the initial tree species based on the gain values ​​of signals at different wavelengths;

[0016] The ratio between the average gain of the initial tree species and the gain flatness is used as the Raman amplifier gain value.

[0017] Optionally, in step S4, the proportion of each tree species in the second-generation tree species is adjusted by adjusting the probability threshold of the second-generation tree species, so as to balance the local search and global search capabilities of the algorithm and improve the optimization efficiency of the algorithm.

[0018] Optionally, the second-generation tree species includes tree species randomly generated from locally optimal tree species and tree species randomly generated in the global search.

[0019] Optionally, in step S5, the performance parameters of Raman amplifiers with different structures are compared and analyzed based on the optimal tree species.

[0020] Optionally, in step S5, the method for outputting the corresponding optimal tree species based on the optimal fitness value specifically includes:

[0021] The population position of the second-generation tree species is initialized based on the optimal fitness value;

[0022] Based on the correlation between the influence of variables on the results, determine the number of seeds generated by each tree in the second-generation tree species, and generate the corresponding seeds;

[0023] Calculate the fitness value of the seeds produced by each tree in the second-generation tree species, sort the seeds according to the fitness value, and determine the optimal tree species;

[0024] The optimal tree species is iteratively optimized according to the tree species optimization algorithm until the optimal tree species that meets the design requirements is output. The evolution of optical power under different pump parameters and the change of fitness value with the number of iterations are recorded.

[0025] The beneficial effects that this invention can produce include:

[0026] The present invention provides an O-band Raman amplifier gain optimization design method based on a tree-species optimization algorithm. By applying the tree-species optimization algorithm to Raman amplifier design, it solves the multi-objective optimization problem regarding pump parameters. By adjusting the probability threshold of the tree-species optimization algorithm, the local and global search capabilities of the optimization model are improved, thereby quickly obtaining the relationship between pump light parameters and Raman amplifier gain values. This effectively increases the transmission distance of O-band signals and provides a novel research approach for the design of Raman amplifiers in special bands. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the design of Raman amplifiers with different pumps and signal bandwidths in this invention;

[0028] Figure 2 This is a flowchart of the algorithm for numerical calculation and optimization of the Raman amplifier in this invention;

[0029] Figure 3 This is a schematic diagram illustrating the gain optimization results of a six-pump Raman amplifier with a signal bandwidth of 100nm, as described in this invention; where, Figure 3 (a) is a schematic diagram of the pump power of a six-pump Raman amplifier with a signal bandwidth of 100nm as a function of fiber length. Figure 3 (b) is a schematic diagram of the optical power of a 100nm signal bandwidth six-pump Raman amplifier at different wavelengths as a function of fiber length; Figure 3 (c) is a schematic diagram of the gain of a six-pumped Raman amplifier with a signal bandwidth of 100nm at different wavelengths; Figure 3 (d) is a schematic diagram of the gain flatness of a six-pumped Raman amplifier with a signal bandwidth of 100nm at different wavelengths;

[0030] Figure 4 This is a schematic diagram illustrating the gain optimization results of a six-pump Raman amplifier with a 50nm signal bandwidth used in this invention; wherein, Figure 4 (a) is a schematic diagram of the pump power of a six-pump Raman amplifier with a signal bandwidth of 50nm as a function of fiber length. Figure 4 (b) is a schematic diagram of the optical power of a 50nm signal bandwidth six-pump Raman amplifier at different wavelengths as a function of fiber length; Figure 4 (c) is a schematic diagram of the gain of signal light at different wavelengths in a six-pumped Raman amplifier with a signal bandwidth of 50nm. Figure 4 (d) is a schematic diagram of the gain flatness of a six-pumped Raman amplifier with a signal bandwidth of 50nm at different wavelengths;

[0031] Figure 5 This is a schematic diagram illustrating the gain optimization results of the four-pump Raman amplifier with a signal bandwidth of 100nm in this invention; wherein, Figure 5(a) is a schematic diagram of the pump power of a 100nm signal bandwidth four-pump Raman amplifier as a function of fiber length. Figure 5 (b) is a schematic diagram of the optical power of a 100nm signal bandwidth four-pump Raman amplifier at different wavelengths as a function of fiber length. Figure 5 (c) is a schematic diagram of the gain of a 100nm signal bandwidth four-pump Raman amplifier for signal light at different wavelengths; Figure 5 (d) is a schematic diagram of the gain flatness of a 100nm signal bandwidth four-pump Raman amplifier for different wavelengths of signal light;

[0032] Figure 6 This is a schematic diagram illustrating the gain optimization results of the four-pump Raman amplifier with a signal bandwidth of 50nm in this invention; wherein, Figure 6 (a) is a schematic diagram of the pump power of a 50nm signal bandwidth four-pump Raman amplifier as a function of fiber length. Figure 6 (b) is a schematic diagram of the optical power of a 50nm signal bandwidth four-pump Raman amplifier at different wavelengths as a function of fiber length. Figure 6 (c) is a schematic diagram of the gain of a 50nm signal bandwidth four-pump Raman amplifier for different wavelengths of signal light; Figure 6 (d) is a schematic diagram of the gain flatness of a 50nm signal bandwidth four-pump Raman amplifier for different wavelengths of signal light;

[0033] Figure 7 This is a schematic diagram illustrating the gain optimization results of the dual-pump Raman amplifiers with signal bandwidths of 100nm and 50nm in this invention; wherein, Figure 7 (a) is a schematic diagram of the pump power of a dual-pump Raman amplifier with a signal bandwidth of 100nm as a function of fiber length. Figure 7 (b) is a schematic diagram of the pump power of a dual-pump Raman amplifier with a signal bandwidth of 50nm as a function of fiber length. Figure 7 (c) is a schematic diagram of the optical power of a dual-pumped Raman amplifier with a signal bandwidth of 100nm at different wavelengths as a function of fiber length. Figure 7 (d) is a schematic diagram of the variation of optical power of a dual-pumped Raman amplifier with different wavelengths as a function of fiber length for a 50nm signal bandwidth dual-pumped Raman amplifier. Figure 7 (f) is a schematic diagram of the gain of signal light at different wavelengths in a dual-pumped Raman amplifier with a signal bandwidth of 100nm. Figure 7 (e) is a schematic diagram of the signal light gain of a dual-pumped Raman amplifier with a signal bandwidth of 50nm at different wavelengths. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] The purpose of this invention is to provide a gain optimization design method for O-band Raman amplifiers based on a tree species optimization algorithm, achieving a flat pump configuration for Raman amplifier gain within the O-band operating bandwidth, comprising the following steps:

[0036] Step 1: Randomly select several sets of pump parameters (40 sets in this example) as the initial tree. Specifically, based on the Raman amplifier structure, select the optimal pump band search range and pump power search range corresponding to the amplified signal; based on the selected pump band search range and pump power search range, randomly generate a corresponding number of pump parameters.

[0037] Step 2: Calculate the Raman amplifier gain value of the initial tree species, use it as the fitness value of the optimization algorithm, and adjust the gain of the initial tree species to obtain the dominant tree species.

[0038] In the above, the initial tree species is used as the initial value for the stimulated Raman coupled-wave equation. Numerical analysis methods (such as the fourth-order Runge-Kutta method in the Runge-Kutta method, the specific steps of which are: dividing the part to be calculated into multiple intervals according to the fiber length; calculating the slope of the next target point in two adjacent intervals according to the given initial value; iterating multiple times until the end of the interval is reached, finally obtaining the numerical solution of the differential equation system) are employed. The algorithm controls the error of each step to within 10... -5 To achieve the ideal accuracy of fourth order, the stimulated Raman coupled wave equation is solved to obtain the signal output power of the initial tree; the signal output power of the initial tree is compared with the input signal power of any wavelength signal to calculate the gain value of the corresponding wavelength signal; based on the gain values ​​of different wavelength signals, the average gain and gain flatness of the initial tree are calculated; the ratio between the average gain and gain flatness of the initial tree is used as the gain value of the Raman amplifier.

[0039] Step 3: Randomly generate multiple sets of first and second tree species near the dominant tree species, and merge the first and second tree species as the second-generation tree species (i.e., pump parameters). The first tree species are defined as several sets of pump parameters randomly generated within a set search range; the second tree species are defined as several sets of pump parameters randomly generated near the pump parameters of the dominant tree species. By adjusting the probability threshold, a suitable ratio is selected to merge the first and second tree species to balance the local and global search capabilities of the optimization algorithm, ultimately obtaining the globally optimal solution. Specifically, taking 40 sets of dominant tree species as an example, firstly, 20 sets of first tree species are randomly generated from the 40 sets of dominant tree species. Then, 10 sets of second tree species are randomly generated near the remaining 20 sets of dominant tree species. Finally, the 20 sets of first tree species and the 10 sets of second tree species are merged in a ratio of 5:2-3 to obtain the second-generation tree species.

[0040] Step 4: Calculate the second-generation fitness value of the second-generation tree species, and perform several iterations to optimize the second-generation fitness value according to Step 2 and Step 3 to obtain the optimal fitness value (i.e., gain value).

[0041] Furthermore, by adjusting the probability threshold of the second-generation tree species, the proportion of each tree species in the second-generation tree species is adjusted to balance the local search and global search capabilities of the algorithm and improve the optimization efficiency. The second-generation tree species include tree species randomly generated from locally optimal tree species and tree species randomly generated during the global search.

[0042] Step 5: Output the optimal tree species (i.e., optimal pump parameters) based on the optimal fitness value. Then, compare and analyze the performance parameters of Raman amplifiers with different structures based on the optimal tree species. The method for outputting the optimal tree species based on the optimal fitness value specifically includes: first, initializing the population position of the second-generation tree species based on the optimal fitness value; then, determining the number of seeds generated by each tree in the second-generation tree species based on the correlation of the variables' influence on the results, and generating the corresponding seeds; next, calculating the fitness value of the seeds generated by each tree in the second-generation tree species, sorting the seeds according to the fitness value, and determining the optimal tree species; finally, iteratively optimizing the optimal tree species using a tree species optimization algorithm until the optimal tree species that meets the design requirements is output; and recording the evolution of different optical power under the optimal pump parameters and the change of fitness value with the number of iterations.

[0043] like Figure 1As shown, this invention optimizes the gain design of Raman amplifiers with different pump and signal bandwidths, specifically employing signal light with a signal bandwidth of 50nm (1280nm-1330nm) and 100nm (1260nm-1360nm); Raman amplifiers with dual-pump, quad-pump, and six-pump structures; wavelength division multiplexers, wavelength demultiplexers, and 30km single-mode fiber (gain medium). By using gain flatness and average gain as comparison targets, the influence of Raman amplifiers with different pump structures on the amplification results of signal light under 50nm and 100nm signal bandwidths is compared and analyzed. Based on the Raman gain spectrum of G652 silica fiber, the suitable pump wavelength search range for amplification in this band is determined, and the transmission loss of pump light and signal light is then determined. Specific parameters are shown in Table 1.

[0044] Table 1: Raman Amplifier Simulation Parameter Settings

[0045]

[0046]

[0047] Furthermore, the evolution of optical power under different pumping parameters and the change of fitness value with the number of iterations were recorded.

[0048] Table 2: Tree Species Optimization Algorithm Parameter Settings

[0049] Population size (N) 80 Maximum number of iterations (R) 40 Variable Dimension (D) 4 / 8 / 12 Pump power search range / W [0.01,1.5] Pump wavelength search range / nm [1000,1259]

[0050] As shown in Table 2, the tree species optimization algorithm proposed in this invention uses the same optimization parameters, population size and number of iterations, and optimizes the Raman amplifier design for different pump numbers (such as dual-pump structure, quad-pump structure, six-pump structure, and pump wavelength and pump power of each pump structure) by adjusting different variable dimensions (such as 4 dimensions, 8 dimensions, 12 dimensions).

[0051] Example 1

[0052] A Raman amplifier with a six-pump structure is optimized. Based on the Raman amplifier parameters and algorithm parameters, the gain performance of the six-pump Raman amplifier under 100 signal beams (i.e., signal beams with a signal bandwidth of 100nm) and 50 signal beams (i.e., signal beams with a signal bandwidth of 50nm) is as follows: Figure 3 and Figure 4 As shown in the figure. The variation in amplification of 100 signal beams when the six pump optical power propagates along a 30km optical fiber is as follows: Figure 3 As shown in (a), while the pump light amplifies 100 signal beams, the pump light with a shorter wavelength transfers some energy to the second-order pump light to amplify it; as... Figure 3As shown in (b), the power evolution of the 100-channel signal optical transmission along a 30km fiber exhibits a trend of initial amplification followed by attenuation, ultimately achieving signal gain and loss cancellation, and convergence at the Raman amplifier output. The gain of the 100-channel signal optical transmission over 30km with and without pump light is shown in Figure [Figure number missing]. Figure 3 As shown in (c), the gain of the 100 signal beams (1260-1270nm and 1350-1360nm) located at both edges of the operating band is poor. This is because the pump light loss in this band is relatively large, and it cannot provide sufficient gain for the 100 signal beams at both ends of the bandwidth during long-distance transmission. Calculations show that the average gain of the 100 signal beams at this point is 9dB. The gain flatness (i.e., the absolute value of the difference between the gain and the average gain) of the 100 signal beams at different wavelengths is as follows: Figure 3 As shown in (d), the gain flatness is 3.3dB, which shows that the gain flatness is greatly improved after the wavelength of 1340nm.

[0053] Specifically, the amplification of 50 signal beams when the six pump lights propagate along a 30km optical fiber is as follows: Figure 4 As shown in (a), pump 6 has the highest input power (1.16W), while pump 4 has the lowest (0.01W). The optical power of the second-order pump light, pump 2, first increases and then decreases, ultimately transferring all energy except for losses to the signal light. (Comparison) Figure 3 (a) It is evident that, under the same optimization conditions, the amplification performance of the 50nm signal bandwidth Raman amplifier is significantly improved. The variation in optical power of 50 signals transmitted along a 30km optical fiber is shown below. Figure 4 As shown in (b), similar to the Raman amplifier with a signal bandwidth of 100nm, the power of the 50-channel signal light also generally exhibits a trend of first amplification and then attenuation, but the amplification effect is more obvious. Furthermore, after eliminating signal bands that are difficult to amplify, the 50-channel signal light exhibits better convergence at the fiber optic end. The gain of the 50-channel signal light after transmission for 30km with and without pump light is shown in Figure [Figure number missing]. Figure 4 As shown in (c), the average gain is 22dB. The gain flatness of the 50 signal beams varies with wavelength as follows: Figure 4 As shown in (d), the gain flatness is 0.52dB.

[0054] Example 2

[0055] In this invention, according to Figure 4 (a) shows the pump parameters. Even after optimization, there is still a pump with almost zero power. In order to reduce costs, we consider reducing the number of pumps while keeping the original operating parameters unchanged. That is, we adopt a four-pump structure to optimize the design of the Raman amplifier.

[0056] In a 100-channel signal optical system with four pump Raman amplifiers, the variation of pump optical power during transmission along a 30km optical fiber is as follows: Figure 5 As shown in (a), based on the initial pump power, it can be seen that the optimized four-pump structure avoids the resource waste that occurs in the six-pump structure, and there is almost no remaining pump light energy at the fiber end; Figure 5 (b) shows the variation of the power of 100 signal beams with transmission distance at different wavelengths in a 100-channel signal beam quad-pump Raman amplifier. Compared to a 100nm signal bandwidth six-pump Raman amplifier, the Raman amplifier with a 100nm signal bandwidth four-pump structure has a larger average signal gain, but this also results in a larger gain flatness. The gain of the 100 signal beams with wavelength in the four-pump Raman amplifier structure is shown below. Figure 5 As shown in (c), the average gain is 11.9 dB. In the four-pump Raman amplifier structure, the gain flatness of the 100 signal optical channels varies with wavelength as follows: Figure 5 As shown in (c), the gain flatness is 5.5dB.

[0057] In this embodiment, as Figure 6 (a) shows the variation of pump power in a four-pump Raman amplifier with 50 signal beams propagating along a 30km optical fiber; as shown in Figure 1. Figure 6 (b) shows the variation of signal power with fiber length in a four-pumped Raman amplifier with 50 signal beams; Figure 6 (c) shows the signal gain variation with different wavelengths in a 50-channel optical four-pump Raman amplifier structure, with an average gain of 17.5 dB; Figure 6 (d) shows the change in signal light gain flatness with wavelength under a 50-channel signal light four-pump Raman amplifier structure, with a gain flatness of 4dB.

[0058] Example 3

[0059] Based on the performance of Raman amplifiers with different structures, a Raman amplifier with a dual-pump structure was designed under the premise that the original operating parameters remain unchanged, and the performance of 100nm signal bandwidth and 50nm signal bandwidth was compared and analyzed.

[0060] Depend on Figure 7 (a) and Figure 7 (b) It can be seen that the variation of pump power with fiber length is affected by Raman effect and loss. It is difficult for a dual-pump structure to achieve gain flatness across a 100nm signal bandwidth, while the gain effect is better after reducing the bandwidth. For example... Figure 7 (c) and Figure 7 As shown in (d), in a dual-pump Raman amplifier, the signal power of the 100-channel signal beam and the 50-channel signal beam varies differently with fiber length. Figure 7 (e) and Figure 7 As shown in (f), in a dual-pump Raman amplifier, the gain changes of different wavelengths of the 100-channel signal light and the 50-channel signal light are different.

[0061] In summary, based on the optimization results of the Raman amplifiers with the above three pump structures, a longitudinal comparison was made, as shown in Table 3.

[0062] Table 3: Performance Comparison Results of Raman Amplifiers with Different Pump Quantities and Signal Bandwidths

[0063]

[0064] The comparison results show that, under the same pump beam conditions, a narrower signal bandwidth results in better optimization, with both maximum gain and gain flatness being better. For the same signal bandwidth, a higher number of pump beams leads to better optimization, with both maximum gain and gain flatness being better. The six-pump Raman amplifier with a 50nm signal bandwidth (including 1290nm and 1310nm) exhibits the best overall performance, achieving an average gain of 22dB and a gain flatness of 0.52dB.

Claims

1. A method for optimizing the gain of an O-band Raman amplifier based on a tree species optimization algorithm, characterized in that, Includes the following steps: S1. Randomly select several sets of pump parameters as initial tree species; according to the O-band Raman amplifier with different structures, select the optimal pump band search range and pump power search range corresponding to the amplified signal; according to the selected pump band search range and pump power search range, randomly generate a corresponding number of pump parameters. S2. Calculate the Raman amplifier gain value of the initial tree species, use it as the fitness value of the optimization algorithm, and adjust the gain of the initial tree species to obtain the dominant tree species; wherein, The initial tree species is used as the initial value of the stimulated Raman coupled-wave equation, and the stimulated Raman coupled-wave equation is solved by numerical analysis to obtain the signal output power of the initial tree species. The gain value of the corresponding wavelength signal is calculated by comparing the signal output power of the initial tree species with the input signal power of an arbitrary wavelength signal. Calculate the average gain and gain flatness of the initial tree species based on the gain values ​​of signals at different wavelengths; The ratio between the average gain of the initial tree species and the gain flatness is used as the Raman amplifier gain value. S3. Randomly generate multiple groups of first tree species and second tree species near the dominant tree species, and merge the first tree species and the second tree species as second-generation tree species; S4. Calculate the second-generation fitness value of the second-generation tree species, and perform several iterations to optimize the second-generation fitness value according to steps S2 and S3 to obtain the optimal fitness value. S5. Output the corresponding optimal tree species based on the optimal fitness value.

2. The O-band Raman amplifier gain optimization design method based on tree species optimization algorithm according to claim 1, characterized in that, In step S4, the proportion of each tree species in the second-generation tree species is adjusted by adjusting the probability threshold of the second-generation tree species, so as to balance the local search and global search capabilities of the algorithm and improve the optimization efficiency of the algorithm.

3. The O-band Raman amplifier gain optimization design method based on tree species optimization algorithm according to claim 2, characterized in that, The second-generation tree species include tree species randomly generated from locally optimal tree species and tree species randomly generated in the global search.

4. The O-band Raman amplifier gain optimization design method based on tree species optimization algorithm according to claim 1, characterized in that, In step S5, the performance parameters of Raman amplifiers with different structures are compared and analyzed based on the optimal tree species.

5. The O-band Raman amplifier gain optimization design method based on tree species optimization algorithm according to claim 1, characterized in that, In step S5, the method for outputting the corresponding optimal tree species based on the optimal fitness value specifically includes: The population position of the second-generation tree species is initialized based on the optimal fitness value; Based on the correlation between the influence of variables on the results, determine the number of seeds generated by each tree in the second-generation tree species, and generate the corresponding seeds; Calculate the fitness value of the seeds produced by each tree in the second-generation tree species, sort the seeds according to the fitness value, and determine the optimal tree species; The optimal tree species is iteratively optimized according to the tree species optimization algorithm until the optimal tree species that meets the design requirements is output. The evolution of optical power under different pump parameters and the change of fitness value with the number of iterations are recorded.

Images