A method and apparatus for demodulating a fiber optic pressure sensor for sensing the force of a ship striking a bridge

Abstract

The application provides a kind of optical fiber pressure sensor demodulation method and device for sensing ship-bridge impact force, and relates to the field of optical fiber sensing technology.The method comprises: constructing an optical fiber pressure sensor demodulation system, an initial multi-modal model based on a double-branch radial basis function neural network, and using a hybrid optimization algorithm including an improved adaptive particle swarm global exploration and an enhanced sparrow search local optimization.In the global stage, a nonlinear adaptive PSO is used to obtain the global optimal solution, and in the local stage, the solution is fine-tuned by SSA and dynamic refocusing is achieved using multi-modal monitoring to obtain the optimal solution of the model.The characteristic vector is input into the model, the analytic signal is obtained by Hilbert transform, the envelope is extracted by extracting the modulus, and the real-time pressure prediction value is output by the weighted fusion layer.The application can significantly improve the convergence speed and accuracy of nonlinear spectral inversion, and meet the demand for real-time and high robustness monitoring of local high pressure in ship-bridge impact scenarios.

Description

Technical Field

[0001] This invention relates to the field of fiber optic sensing technology, and in particular to a demodulation method and apparatus for a fiber optic pressure sensor used to sense the impact force of a ship colliding with a bridge. Background Technology

[0002] Statistics show that ship-bridge collisions have become a major cause of bridge damage and even collapse. When a ship collides with a bridge structure (such as piers or anti-collision devices), transient, localized high pressure is generated in the contact area. The peak pressure can reach several megapascals or even higher, and it exhibits high spatial non-uniformity and temporal suddenness. Therefore, achieving high-precision, real-time dynamic monitoring of local pressure during the impact is crucial for assessing the severity of the impact, providing early warning of structural damage risks, and making subsequent maintenance decisions. Traditional ship collision monitoring methods mostly rely on accelerometers or strain gauges to indirectly estimate the impact force. However, these methods are difficult to accurately reflect the local stress distribution on the contact surface and are susceptible to interference from structural vibration modes, resulting in limited measurement resolution. In contrast, fiber optic pressure sensing technology, due to its advantages such as resistance to electromagnetic interference, corrosion resistance, small size, and distributed deployment, demonstrates unique potential in complex engineering environments. Especially in critical parts of bridges, such as inside the anti-collision fenders, embedded layers on the surface of piers, or composite material coatings, deploying miniaturized, high-sensitivity fiber optic pressure sensors can directly capture the pressure change process in the contact area at the moment of impact, thereby achieving accurate perception of the impact location, intensity, and spatiotemporal distribution characteristics. In recent years, "fiber hybrid structures" based on special optical fibers have become a research hotspot for realizing ultra-compact, high-performance micro-pressure sensing. A typical structure is the single-mode-hollow-core-coreless fiber cascade structure, which utilizes hollow-core fiber as a Fabry-Perot cavity or mode filter, while coreless fiber enhances mode field matching and interference effects, forming stable interference spectrum characteristics. When external pressure is applied to this structure, it causes micro-deformation or refractive index changes in the fiber, leading to wavelength drift or intensity modulation of the interference fringes in the transmission spectrum, thereby achieving optical encoding of pressure information. This type of structure has advantages such as small size (length can be less than 1 mm), high sensitivity, and ease of packaging and integration, making it particularly suitable for embedding within bridge anti-collision facilities for localized high-pressure monitoring. However, these sensors face numerous challenges in practical applications: First, their optical response is not only related to pressure but also significantly affected by environmental factors such as temperature and humidity, resulting in severe cross-sensitivity issues. Second, due to factors such as nonlinear material deformation and structural stress redistribution, the spectral signal and the measured pressure often exhibit a strong nonlinear relationship, making traditional linear fitting or physical modeling paradigms insufficient for high-precision demodulation requirements. Third, ship collision events are sudden and transient, demanding that sensor systems possess rapid response and efficient data processing capabilities to achieve real-time early warning. To address these issues, intelligent optimization algorithms have gradually been introduced into the field of fiber optic sensor signal demodulation. Particle Swarm Optimization (PSO) algorithms, with their advantages of fast convergence speed and few parameters, are suitable for initial global search, but they are prone to getting trapped in local optima when dealing with high-dimensional nonlinear problems, affecting demodulation accuracy. Sparrow Search Algorithm (SSA), by simulating the group behavior mechanism of discoverers-followers-watchers, performs excellently in local exploration and escaping traps, but its early exploration efficiency is low, affecting overall computational speed. A single algorithm cannot simultaneously achieve demodulation accuracy, speed, and robustness. Summary of the Invention

[0003] To address the technical challenge of simultaneously achieving accuracy, speed, and stability in the signal demodulation stage of existing fiber optic pressure sensing systems, this invention provides a fiber optic pressure sensor demodulation method and apparatus for sensing the impact force of a ship colliding with a bridge. The technical solution is as follows:

[0004] On the one hand, a demodulation method for a fiber optic pressure sensor used to sense the impact force of a ship colliding with a bridge is provided. This method is implemented by a fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, and includes: S1: Construct and initialize the fiber optic pressure sensor demodulation system. The system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is constructed based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration stage and a knowledge transfer and local optimization stage. The global exploration stage includes an improved adaptive particle swarm optimization algorithm. The knowledge transfer and local optimization stage includes an enhanced sparrow search algorithm, which includes three types of individuals: discoverers, followers, and watchers. The fiber optic pressure sensor is a single-mode fiber-hollow fiber-coreless fiber cascaded structure sensor and is installed at a preset position on the monitored structure. S2: Initialize the global exploration phase, knowledge transfer phase, and local optimization phase; S3: In the global exploration phase, the improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge, and the global optimal solution of the initial multimodal pressure demodulation model is obtained. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy. The global exploration phase is the steady-state operation of the bridge without impact. S4: In the knowledge transfer stage, the initial position of the discoverer in the enhanced sparrow search algorithm is directly set as the global optimal solution, and the followers and watchers are randomly distributed in the preset hypersphere. The knowledge transfer stage is when the bridge is hit, the parameters of the preset hypersphere are calculated jointly based on the hollow fiber material parameters and historical impact data. S5: In the local optimization stage, the enhanced sparrow search algorithm is used to fine-tune the model parameters, simulate the localized concentration and fine evolution of energy in the hollow fiber microstructure, introduce an adaptive learning rate mechanism for the follower, and realize dynamic refocusing based on the vigilant multimodal monitoring mechanism to obtain the optimal individual, which is used to calculate the local optimal solution of the initial multimodal pressure demodulation model. Based on the local optimal solution, the preset multimodal pressure demodulation model is obtained. S6: Based on the sensing front end of the interferometric fiber optic sensor, the transmission spectrum is collected from the single-mode-hollow-coreless fiber cascade structure sensor to generate a spectral feature vector. The spectral feature vector includes wavelength, intensity and envelope change rate. When the transmission spectrum is subjected to impact, it exhibits complex envelope drift and nonlinear shift of interference fringes. S7: Input the spectral feature vector into the preset multimodal pressure demodulation model, perform Hilbert transform on the spectral feature vector and analyze it to obtain the analytical signal, extract the modulus of the analytical signal to obtain the spectral envelope, and use peak detection and interpolation methods to extract the wavelength of the main interference peak and calculate its neighborhood slope. Through the weighted fusion layer, the real-time demodulated pressure prediction value is obtained.

[0005] Preferably, in step S1, a system for demodulating the fiber optic pressure sensor is constructed and initialized. This system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is constructed based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration phase and a knowledge transfer and local optimization phase. The global exploration phase includes an improved adaptive particle swarm optimization algorithm. The knowledge transfer and local optimization phase includes an enhanced sparrow search algorithm, which includes three types of individuals: discoverers, followers, and watchdogs. The fiber optic pressure sensor is a cascaded structure of single-mode fiber-hollow fiber-coreless fiber and is installed at a preset location on the monitored structure. S11: Construct an initial multimodal pressure demodulation model. The initial multimodal pressure demodulation model uses a dual-branch radial basis function neural network as the demodulation model architecture. The dual-branch radial basis function neural network includes an envelope feature extraction branch, a fringe feature extraction branch, and an RBF sub-network. The envelope feature extraction branch is used to perform a Hilbert transform on the original transmission spectrum to extract the modulus of its analytical signal and obtain the spectral envelope. The fringe feature extraction branch is used to extract the wavelength of the main interference peak using peak detection and interpolation methods and calculate its neighborhood slope to obtain the fringe feature. The spectral envelope reflects the overall light intensity suppression caused by the compression of the hollow fiber wall thickness. The fringe feature reflects the phase drift caused by the change in the refractive index of the air inside the hollow fiber. The RBF sub-network is used to merge the spectral envelope and fringe feature and output the pressure prediction value through a weighted fusion layer. S12: Configure the initial structural parameters of the multimodal pressure demodulation model, including the input layer dimension, the number of hidden layer nodes, and the initial parameter range; S13: Initialize the hybrid optimization algorithm module, which includes a global exploration phase, a knowledge transfer phase, and a local optimization phase.

[0006] Preferably, the initialization of the global exploration phase, knowledge transfer phase, and local optimization phase in S2 includes: S21: In the global exploration phase, the improved adaptive particle swarm optimization algorithm is tested and its parameters are set. The parameter settings include the population size of the subswarm optimization, the maximum number of iterations, and the form of the fitness function. S22: Set the timing for switching from particle swarm optimization algorithm to sparrow search algorithm, preset the number of iterations as the threshold to trigger the stage migration, and maintain the consistency of the fitness function; S23: Set the population size of the sparrow search algorithm and refine the population role composition, wherein the population size includes the ratio of discoverers, followers and watchers; S24: Set the maximum number of iterations for the knowledge transfer and local optimization phases.

[0007] Preferably, in the global exploration phase of S3, an improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge, obtaining the global optimal solution of the initial multimodal pressure demodulation model. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy. The global exploration phase is the steady-state operation of the bridge without impact, including: S31: Based on the improved adaptive particle swarm optimization algorithm, the search range of the initial particle swarm is obtained after setting a nonlinear adaptive strategy. The nonlinear adaptive strategy includes a function that decreases with the number of iterations. The function that decreases with the number of iterations is used to achieve large-scale exploration in the early stage and convergence and focusing in the later stage. S32: A dynamic adjustment mechanism of cognitive and social factors is adopted to adjust the intensity of individual memory guidance and the intensity of group cooperation. The dynamic adjustment mechanism includes increasing the cognitive factor when the population is dispersed and increasing the social factor when the population is clustered. S33: By numerically simulating the stress wave propagation process of the bridge's macrostructure, the wide-area diffusion characteristics of energy spreading from the impact point to the surrounding area are simulated. The numerical simulation is based on the finite element and equivalent stiffness models. S34: Represent the position of each particle in the parameter space as a set of possible combinations of demodulation model parameters to obtain the motion state of the particle, which is described by two variables: velocity and position. S35: An adaptive strategy is adopted to dynamically collect the changing trends of inertial weights, cognitive factors, and social factors, and key parameters are adjusted through adaptive update rules to perform iterative search and locate the approximate position of the global optimal region. The adaptive strategy includes a composite fitness function and constraint terms. The composite fitness function includes a mean square error term between the predicted pressure and the calibrated pressure and a second-order difference smoothing term of the pressure sequence, which is used to evaluate the current position of each particle and obtain the fitness. The constraint terms include physical rationality constraints and material mechanical limitations, specifically including the elastic modulus of quartz glass, the wall thickness of hollow fiber, the change in the radius of hollow fiber, and empirical coefficients. S36: Based on the approximate location of the global optimal region, compare and replace the historical best position and the global best position of the particle. The rule includes updating the corresponding position when the current fitness is better than the historical best. S37: By iterating to the maximum number of iterations or the convergence threshold, the global optimal solution is obtained, which is used as the initial parameter values ​​for the multimodal pressure demodulation model in subsequent stages.

[0008] Preferably, in the knowledge transfer phase of S4, the initial position of the discoverer in the enhanced sparrow search algorithm is directly set as the global optimal solution, and the followers and watchers are randomly distributed within a preset hypersphere. The knowledge transfer phase involves calculating the parameters of the preset hypersphere based on the hollow fiber material parameters and historical impact data after a bridge collision, including: S41: Based on the hollow fiber material parameters and historical impact data, the parameters of the preset hypersphere are jointly calculated to carry out a local fine search in the neighborhood of the global optimal solution. The parameters of the preset hypersphere include, but are not limited to, the center position, radius, shape factor and sampling density. S42: Set the initial position of the discoverer in the enhanced sparrow search algorithm directly as the global optimal solution; S43: Randomly distribute followers and vigilants within a predefined hypersphere to obtain their initial individual distribution.

[0009] Preferably, in the local optimization stage of S5, an enhanced sparrow search algorithm is used to fine-tune the model parameters, simulating the localized concentration and fine evolution of energy in the hollow fiber microstructure. An adaptive learning rate mechanism is introduced for the follower, and dynamic refocusing is achieved based on the vigilant multimodal monitoring mechanism to obtain the optimal individual. This optimal individual is used to calculate the local optimal solution of the initial multimodal pressure demodulation model. Based on the local optimal solution, a preset multimodal pressure demodulation model is obtained, including: S51: When entering the local optimization stage, a knowledge transfer strategy oriented towards the response characteristics of hollow fiber microstructures is adopted to switch the search mode from the improved adaptive particle swarm optimization algorithm to the enhanced sparrow search algorithm, so as to obtain a high-precision optimization mode that can be entered without retracing the entire solution space. The high-precision optimization mode is used to fine-tune the model parameters using the enhanced sparrow search algorithm. The knowledge transfer strategy includes a seamless mapping that takes the endpoint of stress wave propagation as the starting point of local fine search. S52: First, simulate the localization and fine evolution of energy in the hollow fiber microstructure, simulate the physical process of stress wave propagation to the preset position of the monitored structure, and then collect the energy localization and interference spectrum shift characteristics in the hollow fiber segment to obtain the response guidance from wide-area low energy to local high energy. S53: Based on the response guidance from wide-area low energy to local high energy, the discoverer is updated through a hybrid strategy that introduces Levy flight, resulting in an iterative position of the discoverer with improved jump-out capability. S54: An adaptive learning rate mechanism is adopted to adjust the step size and update the position of followers. The optimal individual is obtained through dynamic refocusing control. The dynamic refocusing control includes a multimodal monitoring mechanism of the watchers to collect the fitness improvement and real-time spectral residuals of consecutive generations. When the fitness improvement of five consecutive generations is less than a set threshold, a mutation operation is triggered. When the spectral residual suddenly increases beyond the threshold, the positions of some watchers are reset to the prior parameter range corresponding to the new observation features. The adaptive learning rate mechanism includes small-amplitude Gaussian noise, which is used to maintain population vitality. S55: Based on the optimal individual, a local optimal solution is obtained. After parameter solidification and model reconfiguration, a preset multimodal pressure demodulation model is obtained. The local optimal solution includes a high-precision parameter set.

[0010] Preferably, step S7 involves inputting the spectral feature vector into a preset multimodal pressure demodulation model, performing a Hilbert transform on the spectral feature vector and analyzing it to obtain an analytical signal, extracting the magnitude of the analytical signal to obtain the spectral envelope, simultaneously using peak detection and interpolation methods to extract the wavelength of the main interference peak, calculating its neighborhood slope, and obtaining the real-time demodulated pressure prediction value through a weighted fusion layer, including: S71: After obtaining the preset multimode pressure demodulation model, the spectral feature vector is obtained from the single-mode-hollow-coreless fiber cascade structure sensor and input into the dual-branch radial basis function neural network of the preset multimode pressure demodulation model. The spectral feature vector includes, but is not limited to, wavelength, intensity and envelope change rate. S72: Perform Hilbert transform on the spectral feature vector based on the envelope feature extraction branch and analyze to obtain the analytical signal. Extract the magnitude of the analytical signal to obtain the spectral envelope, which is used to reflect the overall light intensity modulation caused by the compression of the hollow fiber wall thickness. S73: Based on the fringe feature extraction branch, the peak detection and interpolation method is used simultaneously to extract the wavelength of the main interference peak and calculate its neighborhood slope to obtain the fringe feature. The fringe feature reflects the phase drift caused by the change in the air refractive index inside the hollow fiber. S74: Input the envelope features and stripe features into the independent RBF subnetworks of the dual-branch radial basis function neural network for feature modeling. The dual-branch radial basis function neural network is built on the basic architecture of the multimodal pressure demodulation model and is used to approximate complex nonlinear input-output relationships and accurately invert the pressure value corresponding to the externally applied impact force. S75: The outputs of the envelope feature extraction branch and the stripe feature extraction branch are fused by a weighted fusion layer to obtain the pressure prediction value for real-time demodulation.

[0011] On the other hand, a fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge is provided. This device is applied to a fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge. The device includes: Initial Model Module: Used to construct and initialize the fiber optic pressure sensor demodulation system. The system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is constructed based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration phase and a knowledge transfer and local optimization phase. The global exploration phase includes an improved adaptive particle swarm optimization algorithm. The knowledge transfer and local optimization phase includes an enhanced sparrow search algorithm, which includes three types of individuals: discoverers, followers, and watchdogs. The fiber optic pressure sensor is a cascaded structure of single-mode fiber-hollow fiber-coreless fiber and is installed at a preset position on the monitored structure. Optimization initialization module: Used to initialize the global exploration phase, knowledge transfer phase, and local optimization phase; Adaptive Particle Swarm Module: During the global exploration phase, an improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge and obtain the global optimal solution of the initial multimodal pressure demodulation model. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy, and the global exploration phase is the steady-state operation of the bridge without impact. Sparrow Search Module: Used in the knowledge transfer phase to directly set the initial position of the discoverer in the enhanced sparrow search algorithm as the global optimal solution, and the followers and watchers are randomly distributed in the preset hypersphere. The knowledge transfer phase is when the bridge is hit, the parameters of the preset hypersphere are calculated based on the hollow fiber material parameters and historical impact data. Local optimization module: During the local optimization stage, it enables the enhanced sparrow search algorithm to fine-tune the model parameters, simulates the localized concentration and fine evolution of energy in the hollow fiber microstructure, introduces an adaptive learning rate mechanism for followers, and achieves dynamic refocusing based on the vigilant multimodal monitoring mechanism to obtain the optimal individual, which is used to calculate the local optimal solution of the initial multimodal pressure demodulation model, and obtains the preset multimodal pressure demodulation model based on the local optimal solution; Feature vector module: used to acquire transmission spectra from a single-mode-hollow-coreless fiber cascade structure sensor based on the sensing front end of an interferometric fiber optic sensor, and generate spectral feature vectors. The spectral feature vectors include wavelength, intensity and envelope change rate. When the transmission spectrum is subjected to impact, it exhibits complex envelope drift and nonlinear shift of interference fringes. Demodulation module: It is used to input the spectral feature vector into the preset multimodal pressure demodulation model, perform Hilbert transform on the spectral feature vector and analyze it to obtain the analytical signal, extract the modulus of the analytical signal to obtain the spectral envelope, and use peak detection and interpolation methods to extract the wavelength of the main interference peak and calculate its neighborhood slope. Through the weighted fusion layer, the real-time demodulated pressure prediction value is obtained.

[0012] On the other hand, a fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge is provided. The fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge includes: a processor; a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the method described in any one of the above-described fiber optic pressure sensor demodulation methods for sensing the impact force of a ship colliding with a bridge is implemented.

[0013] On the other hand, a computer-readable storage medium is provided, characterized in that the computer-readable storage medium stores program code, which can be invoked by a processor to execute the method as described in any one of claims 1 to 7.

[0014] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: A hybrid intelligent optimization method for fiber optic sensing, designed for high-precision pressure demodulation, is proposed. This method establishes an intelligent switching and knowledge transfer mechanism between Positional Search (PSO) and Localized Optimal Search (SSA). The globally optimal solution obtained in the PSO phase is used as the initial position of the SSA discoverer. Furthermore, the physical response prior of the HCF microstructure determines the initial neighborhood of the follower and the watcher, achieving a smooth transition between the two algorithms. This method fully leverages the efficiency of PSO in global exploration and the precision of SSA in local development, overcoming the limitations of traditional serial splicing modes. It significantly improves the demodulation accuracy and convergence speed of ultra-compact fiber optic hybrid structure sensors under strongly nonlinear conditions, providing a pressure sensing solution with real-time performance, accuracy, and robustness for high-dynamic, high-reliability applications such as bridge collision avoidance monitoring. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge, provided by an embodiment of the present invention. Figure 2 This is a flowchart of the demodulation process of an optical fiber pressure sensor for sensing the impact force of a ship colliding with a bridge, provided by an embodiment of the present invention. Figure 3 This is a block diagram of a fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, provided in an embodiment of the present invention. Figure 4 This is a schematic diagram of the structure of a fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0019] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0020] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0021] List of abbreviations, their English equivalents, and key term definitions: 1. PSO, Particle Swarm Optimization Algorithm 2. SSA, Sparrow Search Algorithm 3. SMF-HCF-NCMF, Single-Mode Fiber - Hollow-Core Fiber - No-Core Fiber Cascaded Structure. 4. DBRBFNN, Dual-Branch Radial Basis Function Neural Network 5. ESSA, Enhanced Sparrow Search Algorithm 6. APSO, Adaptive Particle Swarm Optimization.

[0022] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0023] This invention provides a method for demodulating an optical fiber pressure sensor to sense the impact force of a ship colliding with a bridge. This method can be implemented using an optical fiber pressure sensor demodulation device, which can be a terminal or a server. Figure 1 The flowchart shown is for a fiber optic pressure sensor demodulation method used to sense the impact force of a ship colliding with a bridge. The processing flow of this method may include the following steps:

[0024] S1: Construct and initialize the fiber optic pressure sensor demodulation system. The system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is constructed based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration stage and a knowledge transfer and local optimization stage. The global exploration stage includes an improved adaptive particle swarm optimization algorithm. The knowledge transfer and local optimization stage includes an enhanced sparrow search algorithm, which includes three types of individuals: discoverers, followers, and watchers. The fiber optic pressure sensor is a single-mode fiber-hollow fiber-coreless fiber cascaded structure sensor and is installed at a preset position on the monitored structure. Preferably, S1 includes: S11: Construct an initial multimodal pressure demodulation model. The initial multimodal pressure demodulation model uses a dual-branch radial basis function neural network as the demodulation model architecture. The dual-branch radial basis function neural network includes an envelope feature extraction branch, a fringe feature extraction branch, and an RBF sub-network. The envelope feature extraction branch is used to perform a Hilbert transform on the original transmission spectrum to extract the modulus of its analytical signal and obtain the spectral envelope. The fringe feature extraction branch is used to extract the wavelength of the main interference peak using peak detection and interpolation methods and calculate its neighborhood slope to obtain the fringe feature. The spectral envelope reflects the overall light intensity suppression caused by the compression of the hollow fiber wall thickness. The fringe feature reflects the phase drift caused by the change in the refractive index of the air inside the hollow fiber. The RBF sub-network is used to merge the spectral envelope and fringe feature and output the pressure prediction value through a weighted fusion layer. S12: Configure the initial structural parameters of the multimodal pressure demodulation model, including the input layer dimension, the number of hidden layer nodes, and the initial parameter range; S13: Initialize the hybrid optimization algorithm module, which includes a global exploration phase, a knowledge transfer phase, and a local optimization phase.

[0025] In some embodiments, when a single-mode fiber-hollow fiber-coreless fiber cascade structure (SMF-HCF-NCMF) is subjected to an impact, its transmission spectrum exhibits complex envelope drift and nonlinear shift of interference fringes.

[0026] It should be noted that the spectral envelope reflects the overall intensity modulation caused by HCF wall thickness compression.

[0027] It should be further explained that the multimodal pressure demodulation model is based on the radial basis function neural network (DBRBFNN) architecture, which can effectively approximate complex nonlinear input-output relationships.

[0028] S2: Initialize the global exploration phase, knowledge transfer phase, and local optimization phase; Preferably, S2 includes: S21: In the global exploration phase, the improved adaptive particle swarm optimization algorithm is tested and its parameters are set. The parameter settings include the population size of the subswarm optimization, the maximum number of iterations, and the form of the fitness function. S22: Set the timing for switching from particle swarm optimization algorithm to sparrow search algorithm, preset the number of iterations as the threshold to trigger the stage migration, and maintain the consistency of the fitness function; S23: Set the population size of the sparrow search algorithm and refine the population role composition, wherein the population size includes the ratio of discoverers, followers and watchers; S24: Set the maximum number of iterations for the knowledge transfer and local optimization phases.

[0029] In some embodiments, the hybrid optimization algorithm is initialized as follows: the population size is 50 in the Particle Swarm Optimization (PSO) stage; the population size is 30 in the Sparrow Search Algorithm (SSA) stage, with 5 discoverers, 15 followers, and 10 watchdogs; the maximum number of iterations is 200, and PSO is switched to SSA after 100 iterations.

[0030] S3: In the global exploration phase, the improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge, and the global optimal solution of the initial multimodal pressure demodulation model is obtained. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy. The global exploration phase is the steady-state operation of the bridge without impact. Preferably, S3 includes: S31: Based on the improved adaptive particle swarm optimization algorithm, the search range of the initial particle swarm is obtained after setting a nonlinear adaptive strategy. The nonlinear adaptive strategy includes a function that decreases with the number of iterations. The function that decreases with the number of iterations is used to achieve large-scale exploration in the early stage and convergence and focusing in the later stage. S32: A dynamic adjustment mechanism of cognitive and social factors is adopted to adjust the intensity of individual memory guidance and the intensity of group cooperation. The dynamic adjustment mechanism includes increasing the cognitive factor when the population is dispersed and increasing the social factor when the population is clustered. S33: By numerically simulating the stress wave propagation process of the bridge's macrostructure, the wide-area diffusion characteristics of energy spreading from the impact point to the surrounding area are simulated. The numerical simulation is based on the finite element and equivalent stiffness models. S34: Represent the position of each particle in the parameter space as a set of possible combinations of demodulation model parameters to obtain the motion state of the particle, which is described by two variables: velocity and position. S35: An adaptive strategy is adopted to dynamically collect the changing trends of inertial weights, cognitive factors, and social factors, and key parameters are adjusted through adaptive update rules to perform iterative search and locate the approximate position of the global optimal region. The adaptive strategy includes a composite fitness function and constraint terms. The composite fitness function includes a mean square error term between the predicted pressure and the calibrated pressure and a second-order difference smoothing term of the pressure sequence, which is used to evaluate the current position of each particle and obtain the fitness. The constraint terms include physical rationality constraints and material mechanical limitations, specifically including the elastic modulus of quartz glass, the wall thickness of hollow fiber, the change in the radius of hollow fiber, and empirical coefficients. S36: Based on the approximate location of the global optimal region, compare and replace the historical best position and the global best position of the particle. The rule includes updating the corresponding position when the current fitness is better than the historical best. S37: By iterating to the maximum number of iterations or the convergence threshold, the global optimal solution is obtained, which is used as the initial parameter values ​​for the multimodal pressure demodulation model in subsequent stages.

[0031] In some embodiments, the inertia weights employ a nonlinear adaptive strategy to balance early exploration with later development: ,in , This ensures a broad initial search followed by focused convergence, where t is the number of iterations. This represents the maximum number of iterations.

[0032] Cognitive factor c1(t) and social factor c2(t) are determined based on population diversity. Dynamic adjustment: For particle swarm scale, Let be the position vector of the particle. The location of the population center This is the minimum value of the cognitive factor. This represents the maximum value of the cognitive factor. The minimum value of the social factor. The social factor reaches its maximum value, indicating a dispersed population. When the concentration is large, it enhances individual memory guidance (c1 increases), and when aggregation ( Smaller timeframes enhance group collaboration (increase c2) to avoid premature convergence.

[0033] It should be noted that the APSO algorithm searches in parallel in the parameter space using multiple "particles," each representing a set of possible combinations of model parameters. Its motion state is described by two variables: velocity and position, and the update rule is as follows:

[0034] in: For inertial weights, and Let represent the velocity and position of the i-th particle at the t-th iteration, respectively; The ability to control particles to maintain their original tendency of motion; and Adjust the degree of influence between individual experience and group optimality; r1, r2 are random numbers in the interval [0,1]; This is the particle's own historical optimal position; This is the current globally optimal position.

[0035] To improve search efficiency, the algorithm introduces an adaptive strategy to dynamically adjust key parameters. The fitness function is defined as: in To predict the mean square error between the pressure and the calibration value; These are the smoothness constraint weighting coefficients; It is the weighting coefficient for physical rationality constraints; This is the second-order difference term of the pressure sequence, used to smooth the output; Physical rationality constraints: The elastic modulus of quartz glass is 72 GPa. HCF wall thickness (typical value 12 μm). This represents the change in the HCF radius. This is an empirical coefficient (taken as 1.5). This constraint prevents the inversion results from violating the laws of materials mechanics.

[0036] It should be further explained that this phase runs until... The next iteration outputs the global optimal solution. This serves as the starting point for the next stage of optimization.

[0037] S4: In the knowledge transfer stage, the initial position of the discoverer in the enhanced sparrow search algorithm is directly set as the global optimal solution, and the followers and watchers are randomly distributed in the preset hypersphere. The knowledge transfer stage is when the bridge is hit, the parameters of the preset hypersphere are calculated jointly based on the hollow fiber material parameters and historical impact data. Preferably, S4 includes: S41: Based on the hollow fiber material parameters and historical impact data, the parameters of the preset hypersphere are jointly calculated to carry out a local fine search in the neighborhood of the global optimal solution. The parameters of the preset hypersphere include, but are not limited to, the center position, radius, shape factor and sampling density. S42: Set the initial position of the discoverer in the enhanced sparrow search algorithm directly as the global optimal solution; S43: Randomly distribute followers and vigilants within a predefined hypersphere to obtain their initial individual distribution.

[0038] In some embodiments, the specific implementation is as follows: Discoverer initialization: The initial positions of the five discoverers in SSA are directly set to the global optimal solution obtained in the PSO phase. This operation is equivalent to using the "end point of stress wave propagation" as the "starting point of local fine-grained search," achieving seamless integration. Follower and watchdog initialization: The remaining 25 individuals are initialized with... Centered on, with radius Random distribution within the hypersphere:

[0039] It is the scaling factor. It is a random vector. It is the solution space dimension, radius It is not based on experience. The initial radius of the hypersphere is determined by both the HCF material parameters and historical impact data. in, This represents the change in the optimal parameters of the model before and after the m-th historical impact event; HCF wall thickness; The elastic modulus of the material; , is an empirical coefficient; It represents the total number of historical impact events. The thinner the HCF wall, the softer the material ( The larger the value, the more sensitive it is to external pressure, and the greater the parameter fluctuation. Therefore, the initial search neighborhood should be wider to reflect the engineering mapping of "structural vulnerability → search robustness". If the current event is warned of a strong impact by the accelerometer, then... It can be further enlarged to To cope with stronger nonlinear responses.

[0040] This mechanism enables a natural transition from "wide-area probing" to "local fine-tuning," making full use of the global exploration results of PSO, allowing SSA to enter a high-precision optimization mode without retracing the entire solution space, significantly reducing demodulation delay.

[0041] S5: In the local optimization stage, the enhanced sparrow search algorithm is used to fine-tune the model parameters, simulate the localized concentration and fine evolution of energy in the hollow fiber microstructure, introduce an adaptive learning rate mechanism for the follower, and realize dynamic refocusing based on the vigilant multimodal monitoring mechanism to obtain the optimal individual, which is used to calculate the local optimal solution of the initial multimodal pressure demodulation model. Based on the local optimal solution, the preset multimodal pressure demodulation model is obtained. Preferably, S5 includes: S51: When entering the local optimization stage, a knowledge transfer strategy oriented towards the response characteristics of hollow fiber microstructures is adopted to switch the search mode from the improved adaptive particle swarm optimization algorithm to the enhanced sparrow search algorithm, so as to obtain a high-precision optimization mode that can be entered without retracing the entire solution space. The high-precision optimization mode is used to fine-tune the model parameters using the enhanced sparrow search algorithm. The knowledge transfer strategy includes a seamless mapping that takes the endpoint of stress wave propagation as the starting point of local fine search. S52: First, simulate the localization and fine evolution of energy in the hollow fiber microstructure, simulate the physical process of stress wave propagation to the preset position of the monitored structure, and then collect the energy localization and interference spectrum shift characteristics in the hollow fiber segment to obtain the response guidance from wide-area low energy to local high energy. S53: Based on the response guidance from wide-area low energy to local high energy, the discoverer is updated through a hybrid strategy that introduces Levy flight, resulting in an iterative position of the discoverer with improved jump-out capability. S54: An adaptive learning rate mechanism is adopted to adjust the step size and update the position of followers. The optimal individual is obtained through dynamic refocusing control. The dynamic refocusing control includes a multimodal monitoring mechanism of the watchers to collect the fitness improvement and real-time spectral residuals of consecutive generations. When the fitness improvement of five consecutive generations is less than a set threshold, a mutation operation is triggered. When the spectral residual suddenly increases beyond the threshold, the positions of some watchers are reset to the prior parameter range corresponding to the new observation features. The adaptive learning rate mechanism includes small-amplitude Gaussian noise, which is used to maintain population vitality. S55: Based on the optimal individual, a local optimal solution is obtained. After parameter solidification and model reconfiguration, a preset multimodal pressure demodulation model is obtained. The local optimal solution includes a high-precision parameter set.

[0042] In some embodiments, the discoverer update combines Levy flight with enhanced jump-out capability, employing a hybrid strategy.

[0043] The follower introduces an adaptive learning rate mechanism, which gradually reduces the step size as it approaches the optimal solution, thereby improving search accuracy. Adaptive learning rate following mechanism:

[0044] , During iteration, the follower position is updated as follows: , in, Let be the position vector of the follower in the i-th iteration. This is the globally optimal position vector. The adaptive learning rate for the i-th iteration. The initial learning rate, This is the learning rate decay coefficient. It is an exponential decay factor. To minimize the learning rate, The maximum number of iterations, To maintain population vitality, small-amplitude Gaussian noise is used.

[0045] The vigilant multimodal monitoring mechanism (convergence monitoring and anomaly detection) will detect when the fitness improvement over 5 consecutive generations is less than 1%. If the system determines that it may be trapped in a local optimum, it will trigger a mutation operation. If the real-time spectral residual suddenly increases beyond the threshold, the positions of some vigilant individuals will be reset to the prior parameter range corresponding to the new observation features, thus achieving dynamic refocusing.

[0046] It should be noted that the optimization strategy is dynamically adjusted based on the system's operating status. Population size can be scaled on demand, reducing the computational load on embedded devices. The final parameters are determined by the optimal individual output by ESSA.

[0047] The weights of different sub-objectives (accuracy, stability, response speed) are objectively determined using the entropy weight method: in, It's weight. It is the information entropy of the sub-target. It is the sum of the "entropy weight correction terms" of the sub-goals. It is the negative scaling factor of information entropy. Let i be the normalized contribution of the i-th dimension parameter under the j-th objective. It is the natural logarithm.

[0048] S6: Based on the sensing front end of the interferometric fiber optic sensor, the transmission spectrum is collected from the single-mode-hollow-coreless fiber cascade structure sensor to generate a spectral feature vector. The spectral feature vector includes wavelength, intensity and envelope change rate. When the transmission spectrum is subjected to impact, it exhibits complex envelope drift and nonlinear shift of interference fringes. S7: Input the spectral feature vector into the preset multimodal pressure demodulation model, perform Hilbert transform on the spectral feature vector and analyze it to obtain the analytical signal, extract the modulus of the analytical signal to obtain the spectral envelope, and use peak detection and interpolation methods to extract the wavelength of the main interference peak and calculate its neighborhood slope. Through the weighted fusion layer, the real-time demodulated pressure prediction value is obtained.

[0049] Preferably, such as Figure 2 As shown, S7 includes: S71: After obtaining the preset multimode pressure demodulation model, the spectral feature vector is obtained from the single-mode-hollow-coreless fiber cascade structure sensor and input into the dual-branch radial basis function neural network of the preset multimode pressure demodulation model. The spectral feature vector includes, but is not limited to, wavelength, intensity and envelope change rate. S72: Perform Hilbert transform on the spectral feature vector based on the envelope feature extraction branch and analyze to obtain the analytical signal. Extract the magnitude of the analytical signal to obtain the spectral envelope, which is used to reflect the overall light intensity modulation caused by the compression of the hollow fiber wall thickness. S73: Based on the fringe feature extraction branch, the peak detection and interpolation method is used simultaneously to extract the wavelength of the main interference peak and calculate its neighborhood slope to obtain the fringe feature. The fringe feature reflects the phase drift caused by the change in the air refractive index inside the hollow fiber. S74: Input the envelope features and stripe features into the independent RBF subnetworks of the dual-branch radial basis function neural network for feature modeling. The dual-branch radial basis function neural network is built on the basic architecture of the multimodal pressure demodulation model and is used to approximate complex nonlinear input-output relationships and accurately invert the pressure value corresponding to the externally applied impact force. S75: The outputs of the envelope feature extraction branch and the stripe feature extraction branch are fused by a weighted fusion layer to obtain the pressure prediction value for real-time demodulation.

[0050] The above is an introduction to the method embodiments. The following describes the solution described in this application through device embodiments.

[0051] Figure 3 This is a block diagram illustrating a fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, according to an exemplary embodiment. The device is used in a fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge. (Refer to...) Figure 3 The device includes an initial model module, an optimization initialization module, an adaptive particle swarm module, a sparrow search module, a local optimization module, a feature vector module, and a demodulation module.

[0052] Initial Model Module: Used to construct and initialize the fiber optic pressure sensor demodulation system. The system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is constructed based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration phase and a knowledge transfer and local optimization phase. The global exploration phase includes an improved adaptive particle swarm optimization algorithm. The knowledge transfer and local optimization phase includes an enhanced sparrow search algorithm, which includes three types of individuals: discoverers, followers, and watchdogs. The fiber optic pressure sensor is a cascaded structure of single-mode fiber-hollow fiber-coreless fiber and is installed at a preset position on the monitored structure. Optimization initialization module: Used to initialize the global exploration phase, knowledge transfer phase, and local optimization phase; Adaptive Particle Swarm Module: During the global exploration phase, an improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge and obtain the global optimal solution of the initial multimodal pressure demodulation model. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy, and the global exploration phase is the steady-state operation of the bridge without impact. Sparrow Search Module: Used in the knowledge transfer phase to directly set the initial position of the discoverer in the enhanced sparrow search algorithm as the global optimal solution, and the followers and watchers are randomly distributed in the preset hypersphere. The knowledge transfer phase is when the bridge is hit, the parameters of the preset hypersphere are calculated based on the hollow fiber material parameters and historical impact data. Local optimization module: During the local optimization stage, it enables the enhanced sparrow search algorithm to fine-tune the model parameters, simulates the localized concentration and fine evolution of energy in the hollow fiber microstructure, introduces an adaptive learning rate mechanism for followers, and achieves dynamic refocusing based on the vigilant multimodal monitoring mechanism to obtain the optimal individual, which is used to calculate the local optimal solution of the initial multimodal pressure demodulation model, and obtains the preset multimodal pressure demodulation model based on the local optimal solution; Feature vector module: used to acquire transmission spectra from a single-mode-hollow-coreless fiber cascade structure sensor based on the sensing front end of an interferometric fiber optic sensor, and generate spectral feature vectors. The spectral feature vectors include wavelength, intensity and envelope change rate. When the transmission spectrum is subjected to impact, it exhibits complex envelope drift and nonlinear shift of interference fringes. Demodulation module: It is used to input the spectral feature vector into the preset multimodal pressure demodulation model, perform Hilbert transform on the spectral feature vector and analyze it to obtain the analytical signal, extract the modulus of the analytical signal to obtain the spectral envelope, and use peak detection and interpolation methods to extract the wavelength of the main interference peak and calculate its neighborhood slope. Through the weighted fusion layer, the real-time demodulated pressure prediction value is obtained.

[0053] A fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, the fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge includes: a processor; a memory, the memory storing computer-readable instructions, which, when executed by the processor, implement the method described in any one of the above-described fiber optic pressure sensor demodulation methods for sensing the impact force of a ship colliding with a bridge.

[0054] A computer-readable storage medium, characterized in that the computer-readable storage medium stores program code, the program code being invoked by a processor to execute the method as described in any one of claims 1 to 7.

[0055] Figure 4 This is a schematic diagram of a fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, provided in an embodiment of the present invention. Figure 4 As shown, the fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge may include the above-mentioned... Figure 3 The illustrated fiber optic pressure sensor demodulation device is used to sense the impact force of a ship colliding with a bridge. Optionally, the fiber optic pressure sensor demodulation device 410 for sensing the impact force of a ship colliding with a bridge may include a first processor 2001.

[0056] Optionally, the fiber optic pressure sensor demodulation device 410 for sensing the impact force of a ship colliding with a bridge may also include a memory 2002 and a transceiver 2003.

[0057] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0058] The following is combined Figure 4 The following is a detailed description of the various components of the fiber optic pressure sensor demodulation device 410 used to sense the impact force of a ship colliding with a bridge: The first processor 2001 is the control center of the fiber optic pressure sensor demodulation device 410 used to sense the impact force of the ship colliding with the bridge. It can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0059] Optionally, the first processor 2001 can perform various functions of the fiber optic pressure sensor demodulation device 410 for sensing the impact force of a ship colliding with a bridge by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0060] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 4 CPU0 and CPU1 are shown in the diagram.

[0061] In a specific implementation, as one example, the fiber optic pressure sensor demodulation device 410 for sensing the impact force of a ship colliding with a bridge may also include multiple processors, such as... Figure 4 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0062] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0063] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently, and may be connected via the interface circuit of the fiber optic pressure sensor demodulation device 410 for sensing the impact force of the ship colliding with the bridge. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0064] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0065] Alternatively, transceiver 2003 may include a receiver and a transmitter. Figure 4 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0066] Optionally, the transceiver 2003 can be integrated with the first processor 2001 or exist independently, and can be connected to the interface circuit of the fiber optic pressure sensor demodulation device 410 for sensing the impact force of the ship colliding with the bridge. Figure 4 (Not shown in the image) is coupled to the first processor 2001, and this embodiment of the invention does not specifically limit this.

[0067] It should be noted that, Figure 4 The structure of the fiber optic pressure sensor demodulation device 410 shown in the figure for sensing the impact force of a ship hitting a bridge does not constitute a limitation on the router. Actual knowledge structure identification devices may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0068] Furthermore, the technical effect of the fiber optic pressure sensor demodulation device 410 used to sense the impact force of a ship colliding with a bridge can be referred to the technical effect of the fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge described in the above method embodiments, and will not be repeated here.

[0069] It should be understood that the first processor 2001 in the embodiments of the present invention may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0070] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0071] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0072] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0073] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0074] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0075] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. 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 implementations should not be considered beyond the scope of this invention.

[0076] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0077] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0078] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0079] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0080] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a 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 includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0081] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A demodulation method for an optical fiber pressure sensor used to sense the impact force of a ship colliding with a bridge, characterized in that, The method includes: S1: Construct and initialize the fiber optic pressure sensor demodulation system. The system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is constructed based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration stage and a knowledge transfer and local optimization stage. The global exploration stage includes an improved adaptive particle swarm optimization algorithm. The knowledge transfer and local optimization stage includes an enhanced sparrow search algorithm, which includes three types of individuals: discoverers, followers, and watchers. The fiber optic pressure sensor is a single-mode fiber-hollow fiber-coreless fiber cascaded structure sensor and is installed at a preset position on the monitored structure. S2: Initialize the global exploration phase, knowledge transfer phase, and local optimization phase; S3: In the global exploration phase, the improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge, and the global optimal solution of the initial multimodal pressure demodulation model is obtained. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy. The global exploration phase is the steady-state operation of the bridge without impact. S4: In the knowledge transfer stage, the initial position of the discoverer in the enhanced sparrow search algorithm is directly set as the global optimal solution, and the followers and watchers are randomly distributed in the preset hypersphere. The knowledge transfer stage is when the bridge is hit, the parameters of the preset hypersphere are calculated jointly based on the hollow fiber material parameters and historical impact data. S5: In the local optimization stage, the enhanced sparrow search algorithm is used to fine-tune the model parameters, simulate the localized concentration and fine evolution of energy in the hollow fiber microstructure, introduce an adaptive learning rate mechanism for the follower, and realize dynamic refocusing based on the vigilant multimodal monitoring mechanism to obtain the optimal individual, which is used to calculate the local optimal solution of the initial multimodal pressure demodulation model. Based on the local optimal solution, the preset multimodal pressure demodulation model is obtained. S6: Based on the sensing front end of the interferometric fiber optic sensor, the transmission spectrum is collected from the single-mode-hollow-coreless fiber cascade structure sensor to generate a spectral feature vector. The spectral feature vector includes wavelength, intensity and envelope change rate. When the transmission spectrum is subjected to impact, it exhibits complex envelope drift and nonlinear shift of interference fringes. S7: Input the spectral feature vector into the preset multimodal pressure demodulation model, perform Hilbert transform on the spectral feature vector and analyze it to obtain the analytical signal, extract the modulus of the analytical signal to obtain the spectral envelope, and use peak detection and interpolation methods to extract the wavelength of the main interference peak and calculate its neighborhood slope. Through the weighted fusion layer, the real-time demodulated pressure prediction value is obtained.

2. The fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge according to claim 1, characterized in that, The S1 describes the construction and initialization of a fiber optic pressure sensor demodulation system. This system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration phase and a knowledge transfer and local optimization phase. The global exploration phase includes an improved adaptive particle swarm optimization algorithm, and the knowledge transfer and local optimization phase includes an enhanced sparrow search algorithm. The enhanced sparrow search algorithm includes three types of individuals: discoverers, followers, and watchdogs. The fiber optic pressure sensor is a cascaded structure of single-mode fiber-hollow fiber-coreless fiber and is installed at a preset location on the monitored structure. S11: Construct an initial multimodal pressure demodulation model. The initial multimodal pressure demodulation model uses a dual-branch radial basis function neural network as the demodulation model architecture. The dual-branch radial basis function neural network includes an envelope feature extraction branch, a fringe feature extraction branch, and an RBF sub-network. The envelope feature extraction branch is used to perform a Hilbert transform on the original transmission spectrum to extract the modulus of its analytical signal and obtain the spectral envelope. The fringe feature extraction branch is used to extract the wavelength of the main interference peak using peak detection and interpolation methods and calculate its neighborhood slope to obtain the fringe feature. The spectral envelope reflects the overall light intensity suppression caused by the compression of the hollow fiber wall thickness. The fringe feature reflects the phase drift caused by the change in the refractive index of the air inside the hollow fiber. The RBF sub-network is used to merge the spectral envelope and fringe feature and output the pressure prediction value through a weighted fusion layer. S12: Configure the initial structural parameters of the multimodal pressure demodulation model, including the input layer dimension, the number of hidden layer nodes, and the initial parameter range; S13: Initialize the hybrid optimization algorithm module, which includes a global exploration phase, a knowledge transfer phase, and a local optimization phase.

3. The fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge according to claim 1, characterized in that, The initialization of the global exploration phase, knowledge transfer phase, and local optimization phase in S2 includes: S21: In the global exploration phase, the improved adaptive particle swarm optimization algorithm is tested and its parameters are set. The parameter settings include the population size of the subswarm optimization, the maximum number of iterations, and the form of the fitness function. S22: Set the timing for switching from particle swarm optimization to sparrow search algorithm, preset the number of iterations as the threshold to trigger the stage migration, and maintain the consistency of the fitness function; S23: Set the population size of the sparrow search algorithm and refine the population role composition, wherein the population size includes the ratio of discoverers, followers and watchers; S24: Set the maximum number of iterations for the knowledge transfer and local optimization phases.

4. The fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge according to claim 1, characterized in that, In the global exploration phase of S3, an improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge, obtaining the global optimal solution of the initial multimodal pressure demodulation model. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy. The global exploration phase is the steady-state operation of the bridge without impact, including: S31: Based on the improved adaptive particle swarm optimization algorithm, the search range of the initial particle swarm is obtained after setting a nonlinear adaptive strategy. The nonlinear adaptive strategy includes a function that decreases with the number of iterations. The function that decreases with the number of iterations is used to achieve large-scale exploration in the early stage and convergence and focusing in the later stage. S32: A dynamic adjustment mechanism of cognitive and social factors is adopted to adjust the intensity of individual memory guidance and the intensity of group cooperation. The dynamic adjustment mechanism includes increasing the cognitive factor when the population is dispersed and increasing the social factor when the population is clustered. S33: By numerically simulating the stress wave propagation process of the bridge's macrostructure, the wide-area diffusion characteristics of energy spreading from the impact point to the surrounding area are simulated. The numerical simulation is based on the finite element and equivalent stiffness models. S34: Represent the position of each particle in the parameter space as a set of possible combinations of demodulation model parameters to obtain the motion state of the particle, which is described by two variables: velocity and position. S35: An adaptive strategy is adopted to dynamically collect the changing trends of inertial weights, cognitive factors, and social factors, and key parameters are adjusted through adaptive update rules to perform iterative search and locate the approximate position of the global optimal region. The adaptive strategy includes a composite fitness function and constraint terms. The composite fitness function includes a mean square error term between the predicted pressure and the calibrated pressure and a second-order difference smoothing term of the pressure sequence, which is used to evaluate the current position of each particle and obtain the fitness. The constraint terms include physical rationality constraints and material mechanical limitations, specifically including the elastic modulus of quartz glass, the wall thickness of hollow fiber, the change in the radius of hollow fiber, and empirical coefficients. S36: Based on the approximate location of the global optimal region, compare and replace the historical best position and the global best position of the particle. The rule includes updating the corresponding position when the current fitness is better than the historical best. S37: By iterating to the maximum number of iterations or the convergence threshold, the global optimal solution is obtained, which is used as the initial parameter values ​​for the multimodal pressure demodulation model in subsequent stages.

5. The fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge according to claim 1, characterized in that, In the knowledge transfer phase of S4, the initial position of the discoverer in the enhanced sparrow search algorithm is directly set as the global optimal solution. Followers and watchers are randomly distributed within a preset hypersphere. This knowledge transfer phase involves calculating the parameters of the preset hypersphere based on hollow fiber material parameters and historical impact data after a bridge collision, including: S41: Based on the hollow fiber material parameters and historical impact data, the parameters of the preset hypersphere are jointly calculated to carry out a local fine search in the neighborhood of the global optimal solution. The parameters of the preset hypersphere include, but are not limited to, the center position, radius, shape factor and sampling density. S42: Set the initial position of the discoverer in the enhanced sparrow search algorithm directly as the global optimal solution; S43: Randomly distribute followers and vigilants within a predefined hypersphere to obtain their initial individual distribution.

6. The fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge according to claim 1, characterized in that, In the local optimization phase of S5, an enhanced sparrow search algorithm is used to fine-tune the model parameters, simulating the localized concentration and refined evolution of energy in the hollow fiber microstructure. An adaptive learning rate mechanism is introduced for the follower, and dynamic refocusing is achieved based on the vigilant multimodal monitoring mechanism to obtain the optimal individual. This optimal individual is used to calculate the local optimal solution of the initial multimodal pressure demodulation model. Based on the local optimal solution, a preset multimodal pressure demodulation model is obtained, including: S51: When entering the local optimization stage, a knowledge transfer strategy oriented towards the response characteristics of hollow fiber microstructures is adopted to switch the search mode from the improved adaptive particle swarm optimization algorithm to the enhanced sparrow search algorithm, so as to obtain a high-precision optimization mode that can be entered without retracing the entire solution space. The high-precision optimization mode is used to fine-tune the model parameters using the enhanced sparrow search algorithm. The knowledge transfer strategy includes a seamless mapping that takes the endpoint of stress wave propagation as the starting point of local fine search. S52: First, simulate the localization and fine evolution of energy in the hollow fiber microstructure, simulate the physical process of stress wave propagation to the preset position of the monitored structure, and then collect the energy localization and interference spectrum shift characteristics in the hollow fiber segment to obtain the response guidance from wide-area low energy to local high energy. S53: Based on the response guidance from wide-area low energy to local high energy, the discoverer is updated through a hybrid strategy that introduces Levy flight, resulting in an iterative position of the discoverer with improved jump-out capability. S54: An adaptive learning rate mechanism is adopted to adjust the step size and update the position of followers. The optimal individual is obtained through dynamic refocusing control. The dynamic refocusing control includes a multimodal monitoring mechanism of the watchers to collect the fitness improvement and real-time spectral residuals of consecutive generations. When the fitness improvement of five consecutive generations is less than a set threshold, a mutation operation is triggered. When the spectral residual suddenly increases beyond the threshold, the positions of some watchers are reset to the prior parameter range corresponding to the new observation features. The adaptive learning rate mechanism includes small-amplitude Gaussian noise, which is used to maintain population vitality. S55: Based on the optimal individual, a local optimal solution is obtained. After parameter solidification and model reconfiguration, a preset multimodal pressure demodulation model is obtained. The local optimal solution includes a high-precision parameter set.

7. The fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge according to claim 1, characterized in that, The S7 process involves inputting the spectral feature vector into a preset multimodal pressure demodulation model, performing a Hilbert transform on the spectral feature vector and analyzing it to obtain an analytical signal, extracting the magnitude of the analytical signal to obtain the spectral envelope, simultaneously using peak detection and interpolation methods to extract the wavelength of the main interference peak, calculating its neighborhood slope, and obtaining the real-time demodulated pressure prediction value through a weighted fusion layer, including: S71: After obtaining the preset multimode pressure demodulation model, the spectral feature vector is obtained from the single-mode-hollow-coreless fiber cascade structure sensor and input into the dual-branch radial basis function neural network of the preset multimode pressure demodulation model. The spectral feature vector includes, but is not limited to, wavelength, intensity and envelope change rate. S72: Perform Hilbert transform on the spectral feature vector based on the envelope feature extraction branch and analyze to obtain the analytical signal. Extract the magnitude of the analytical signal to obtain the spectral envelope, which is used to reflect the overall light intensity modulation caused by the compression of the hollow fiber wall thickness. S73: Based on the fringe feature extraction branch, the peak detection and interpolation method is used simultaneously to extract the wavelength of the main interference peak and calculate its neighborhood slope to obtain the fringe feature. The fringe feature reflects the phase drift caused by the change in the air refractive index inside the hollow fiber. S74: Input the envelope features and stripe features into the independent RBF subnetworks of the dual-branch radial basis function neural network for feature modeling. The dual-branch radial basis function neural network is built on the basic architecture of the multimodal pressure demodulation model and is used to approximate complex nonlinear input-output relationships and accurately invert the pressure value corresponding to the externally applied impact force. S75: The outputs of the envelope feature extraction branch and the stripe feature extraction branch are fused by a weighted fusion layer to obtain the pressure prediction value in real time demodulation.

8. A fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, wherein the fiber optic pressure sensor demodulation device is used to implement the fiber optic pressure sensor demodulation method for sensing the impact force of a ship colliding with a bridge as described in any one of claims 1-7, characterized in that... The device includes: Initial Model Module: Used to construct and initialize the fiber optic pressure sensor demodulation system. The system includes an initial multimodal pressure demodulation model and a hybrid optimization algorithm module. The initial multimodal pressure demodulation model is constructed based on a dual-branch radial basis function neural network. The hybrid optimization algorithm module includes a global exploration phase and a knowledge transfer and local optimization phase. The global exploration phase includes an improved adaptive particle swarm optimization algorithm. The knowledge transfer and local optimization phase includes an enhanced sparrow search algorithm, which includes three types of individuals: discoverers, followers, and watchdogs. The fiber optic pressure sensor is a cascaded structure of single-mode fiber-hollow fiber-coreless fiber and is installed at a preset position on the monitored structure. The optimization initialization module is used to initialize the global exploration phase, knowledge transfer phase, and local optimization phase. Adaptive Particle Swarm Module: During the global exploration phase, an improved adaptive particle swarm optimization algorithm is used to simulate the wide-area diffusion process of stress waves in the macroscopic structure of the bridge and obtain the global optimal solution of the initial multimodal pressure demodulation model. The improved adaptive particle swarm optimization algorithm adopts a nonlinear adaptive strategy, and the global exploration phase is the steady-state operation of the bridge without impact. Sparrow Search Module: Used in the knowledge transfer phase to directly set the initial position of the discoverer in the enhanced sparrow search algorithm as the global optimal solution, and the followers and watchers are randomly distributed in the preset hypersphere. The knowledge transfer phase is when the bridge is hit, the parameters of the preset hypersphere are calculated based on the hollow fiber material parameters and historical impact data. Local optimization module: During the local optimization stage, it enables the enhanced sparrow search algorithm to fine-tune the model parameters, simulates the localized concentration and fine evolution of energy in the hollow fiber microstructure, introduces an adaptive learning rate mechanism for followers, and achieves dynamic refocusing based on the vigilant multimodal monitoring mechanism to obtain the optimal individual, which is used to calculate the local optimal solution of the initial multimodal pressure demodulation model, and obtains the preset multimodal pressure demodulation model based on the local optimal solution; Feature vector module: used to acquire transmission spectra from a single-mode-hollow-coreless fiber cascade structure sensor based on the sensing front end of an interferometric fiber optic sensor, and generate spectral feature vectors. The spectral feature vectors include wavelength, intensity and envelope change rate. When the transmission spectrum is subjected to impact, it exhibits complex envelope drift and nonlinear shift of interference fringes. Demodulation module: It is used to input the spectral feature vector into the preset multimodal pressure demodulation model, perform Hilbert transform on the spectral feature vector and analyze it to obtain the analytical signal, extract the modulus of the analytical signal to obtain the spectral envelope, and use peak detection and interpolation methods to extract the wavelength of the main interference peak and calculate its neighborhood slope. Through the weighted fusion layer, the real-time demodulated pressure prediction value is obtained.

9. A fiber optic pressure sensor demodulation device for sensing the impact force of a ship colliding with a bridge, characterized in that, The fiber optic pressure sensor demodulation processor for sensing the impact force of a ship colliding with a bridge; a memory storing computer-readable instructions, which, when executed by the processor, implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be invoked by a processor to execute the method as described in any one of claims 1 to 7.

Images