Reverse Design And Performance Optimization Of Elliptical Ring Core Few-Mode Fiber Based On Neural Network

With the rapid development of emerging technologies such as cloud computing,big data and artificial intelligence,human information and data traffic is growing explosively,making the demand for traffic carried by optical network infi-astructure continue to increase.fiber optic technology.The space division multiplexing technology based on the spatial dimension further improves the transmission capacity and has great potential.The transmission modes of the fewmode fiber are orthogonal to each other,and can be used as the space channel required for space division multiplexing,which greatly improves the transmission capacity.Ring-core few-mode fibers with annular refractive index profiles have great potential to increase the capacity of low-complexity space-division multiplexing systems,and are characterized by low nonlinearity,low transmission loss,microbending loss,and splicing loss.Because of its powerful prediction and optimization capabilities,machine learning algorithms are used by researchers to combine them with reverse engineering to design complex fiber structures.Therefore,this paper mainly studies the air hole-assisted elliptical ring-core fewmode fiber and groove-assisted elliptical-ring-core few-mode fiber that support multiple transmission modes and have a large effective refractive index difference,combined with a neural network-based reverse design The scheme carries out structural design and performance optimization for it.The main work of this paper is as follows:(1)An air hole-assisted elliptical ring core few-mode fiber composed of three elliptical ring cores and two air holes on both sides of the short semi-axis of the elliptical ring is proposed,which can support 16 polarizations.transmission mode,and adjacent modes can achieve effective separation.Quantitative analysis and preliminary optimization of several structural parameters such as the radius of the elliptical ring,the ellipticity,the ratio of the major and semi-axes of the three elliptical rings,the radius of the air hole and the distance between the air hole and the elliptical ring are carried out.The optimized minimum effective refractive index difference is 1.498×10-4,and its broadband characteristics,effective mode area and dispersion are analyzed.Combined with the reverse design scheme based on neural network,the structural parameters of the air holeassisted elliptical ring core few-mode fiber are optimized.When optimizing the neural network,the influence of parameters such as activation function,optimizer,number of hidden layer neurons,and learning rate on the neural network was analyzed,and a well-trained neural network model was obtained.The average correlation coefficient was 0.995.Finally,we obtained an optimized fiber structure parameters,the minimum effective refractive index difference of the structure is 1.779×10-4,which is improved by about 19%,and has smaller effective mode area and dispersion in the broadband.(2)The basic structure of a trench-assisted elliptical ring core few-mode fiber with 6 elliptical rings is proposed,and the influence of the structure and number of elliptical rings and the refractive index difference on the performance is analyzed,and a preliminary optimization is carried out at the same time.A trenchassisted elliptical ring core few-mode fiber with 5 elliptical rings is obtained,which can support 20 polarization modes.The reverse design and performance optimization of the optical fiber is carried out by using the reverse design scheme based on neural network,including the semi-major axis of the elliptical ring,the ratio of the semi-major axis of the inner ring,the ratio of the semi-major axis of the outer ring,the refractive index difference of the elliptical ring,etc.parameter.The average correlation coefficient between the predicted data of the neural network model and the actual simulation data is 0.997.Finally,the structural parameters of an optimized trench-assisted elliptical ring core few-mode fiber are obtained.The minimum effective refractive index difference is 1.658×10-4,which is about 30%higher than that of the preliminary optimized fiber structure.The broadband characteristics,effective mode area and dispersion in broadband are simulated and analyzed.The results show that the optical fiber with this structure has good transmission characteristics in broadband.In this paper,the air hole-assisted elliptical ring-core few-mode fiber structure and the groove-assisted elliptical ring-core few-mode fiber structure are designed based on the simulation.The network and reverse design optimized its structural parameters,and finally obtained two elliptical ring-core few-mode fiber structures with good transmission performance.This paper combines the traditional optical fiber design with the reverse design based on neural network,and provides a fast solution for the structural design and performance optimization of complex optical fibers in the future.