Seawater temperature, salinity and sound velocity inversion method based on Brillouin laser radar

Abstract

The invention discloses a seawater temperature, salinity and sound velocity inversion method based on a Brillouin laser radar, and belongs to the technical field of marine environment monitoring. According to the method, the temperature, salinity and sound velocity of seawater are inverted by measuring the Brillouin frequency shift and the Brillouin line width and combining the MLP neural network; the method specifically comprises the steps of collecting seawater parameter data of different sea areas, different seasons and different depths and corresponding frequency shifts and line widths; carrying out data preprocessing and standardization on the collected data; and establishing an inversion model by using an MLP neural network, and outputting temperature, salinity and sound velocity. The multi-layer perceptron MLP neural network processes frequency shift and line width data measured by the Brillouin laser radar, high-precision and real-time inversion of seawater temperature, salinity and sound velocity is achieved, the parameter range is remarkably expanded, the inversion precision is improved, and efficient and reliable marine environment monitoring is provided.

Description

Technical Field

[0001] This invention relates to the field of marine environmental monitoring technology, specifically to a method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar. Background Technology

[0002] Ocean exploration plays an increasingly important role in military and scientific research. As a result, countries around the world have elevated the development of marine economy, marine science and technology and marine equipment to the level of national strategy. my country has also raised its development goals in the marine field to an unprecedented level. Furthermore, the implementation of the "Belt and Road" initiative and the development of various aspects such as the deep blue sea all require the support and guarantee of marine exploration technology.

[0003] The detection of parameters such as seawater temperature, salinity, and especially sound velocity plays a crucial role in seabed mapping, disaster prediction, and military defense, and is essential for improving marine exploration and meeting the needs of numerous underwater applications. Acoustic communication is highly susceptible to the influence of seawater temperature, salinity, and pressure, and is considered one of the most difficult communication methods to implement. Currently, seawater temperature, salinity, and sound velocity parameters are mostly measured using traditional towed ship methods, which are affected by external factors such as the natural environment, resulting in low measurement efficiency and poor accuracy. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing methods for measuring seawater temperature, salinity, and sound velocity, which are characterized by low efficiency and poor accuracy. The invention proposes a seawater temperature, salinity, and sound velocity inversion method based on Brillouin lidar to solve the problems mentioned in the background section.

[0005] To address the aforementioned technical problems, this invention provides a method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar. By measuring the Brillouin frequency shift and Brillouin linewidth, and combining this with a multilayer perceptron (MLP) neural network, the temperature, salinity, and sound velocity of seawater can be inverted, accurately obtaining these parameters. Specifically, this includes:

[0006] S100: Collect seawater parameter data and their corresponding frequency shift and linewidth for different sea areas, seasons, and depths.

[0007] A comprehensive and diverse dataset was constructed, covering various marine environmental conditions, thereby enhancing the generalization ability and applicability of subsequent models, avoiding bias caused by limited data, and providing a reliable foundation for the inversion model.

[0008] S200. Perform data preprocessing and standardization on the collected data to remove outliers, duplicates, and incomplete data records.

[0009] Temperature, salinity, sound velocity, frequency shift, and linewidth are standardized to bring them to the same order of magnitude, facilitating neural network training. Data volume is expanded through interpolation and transformation to improve the model's generalization ability.

[0010] Improving data quality and consistency reduces the interference of noise and anomalies on model training, facilitating efficient processing by neural networks. Data augmentation enhances the model's generalization ability, reduces the risk of overfitting, and ensures stable learning across different scenarios, laying a solid foundation for subsequent neural network training.

[0011] S300. A multilayer perceptron (MLP) neural network suitable for handling regression problems is selected to establish an inversion model. Ocean parameters (temperature, salinity, and sound speed) are inverted based on frequency shift and linewidth. The MLP network structure is designed as follows:

[0012] The input layer features are Brillouin frequency shift and linewidth, where the Brillouin frequency shift v B The range is limited to 7.09 GHz ≤ v B ≤7.79GHz; Brillouin linewidth Γ B The range is limited to 0.33GHz≤Γ B ≤1.83GHz.

[0013] Brillouin shift v B The calculation formula is:

[0014]

[0015] In the formula, n is the refractive index of the medium, and λ is the incident light wavelength, usually taken as λ = 532 nm. s Let θ be the speed of sound in the medium, and θ be the scattering angle (θ = 180° for backscattering).

[0016] n and v s All are related to temperature, salinity, and pressure. n adopts the GomerT model, v s The Richards model is used.

[0017] Brillouin line width Γ B The calculation formula is:

[0018]

[0019] In the formula, ρ is the density of seawater, which is generally around 1.023 × 10³; η S Let η be the shear viscosity coefficient of seawater. b is the volume viscosity coefficient of seawater.

[0020] κ is the thermal conductivity, γ is the adiabatic coefficient, and γ = C p / C v Cp For the specific heat capacity at constant pressure, C v This is the specific heat capacity at constant volume.

[0021] The output layer features include seawater temperature, salinity, and sound speed.

[0022] The hidden layers consist of multiple layers of neurons, with each layer containing a certain number of neurons. The optimal number of layers and neurons per layer are determined experimentally.

[0023] The activation functions chosen are ReLU, Sigmoid, and Tanh to enhance the model's nonlinear fitting ability.

[0024] Hyperparameters were tuned using cross-validation and grid search to achieve optimal performance. Hyperparameters included the learning rate, batch size, and number of iterations. The preprocessed dataset was divided into training, validation, and test sets in a 7:2:1 ratio.

[0025] The neural network model is trained using training set data, and the model parameters are adjusted using the backpropagation algorithm to minimize the error between the predicted and actual values. Specifically:

[0026] The maximum temperature error is 0.079℃, and the relative error is 0.212%.

[0027] The maximum salinity error is 0.122‰, and the relative error is 0.156%.

[0028] The maximum error in sound velocity is 0.124 m / s, and the relative error is 0.015%.

[0029] The inversion model is integrated with the pulsed laser, polarization optical path, and signal acquisition system to ensure real-time data transmission and processing.

[0030] The inversion accuracy and generalization ability of the model are analyzed on the validation set, and the model is adjusted synchronously based on the validation results.

[0031] Test the model's performance on the test set to ensure that the model can obtain accurate data in different marine environments.

[0032] The trained MLP model is deployed to the Brillouin lidar system for real-time online inversion. The inversion model is applicable within the following ranges: Temperature: 10℃≤T≤30℃; Salinity: 15‰≤S≤35‰.

[0033] It achieves high-precision inversion, has a wide range of applicability, supports real-time online deployment, improves the efficiency and reliability of marine parameter monitoring, and solves the problems of low efficiency and poor accuracy of traditional methods.

[0034] The present invention also provides a seawater temperature, salinity and sound velocity inversion system based on Brillouin lidar, including a pulsed laser, a polarization optical path, a marine environment simulation system and a signal acquisition system.

[0035] Pulsed lasers are used to output laser sources, including the DG535 digital delay pulse generator and the injected Nd:YAG pulsed laser.

[0036] The pulsed laser source is a seed-injected Nd:YAG pulsed laser. The Nd:YAG pulsed laser outputs vertically polarized light at 532 nm. Its output parameters are: repetition frequency 10 Hz, pulse width 8 ns, beam diameter approximately 19 mm, divergence angle approximately 0.45 mrad, maximum output energy approximately 4 J / Pulse, and linewidth approximately 90 MHz in a single longitudinal mode.

[0037] The polarization optical path is used for polarization conversion. It consists of a half-wave plate, a polarizing beam splitter (PBS), and a quarter-wave plate, and can realize the conversion between vertically polarized light, horizontally polarized light, and circularly polarized light.

[0038] The marine environment simulation system is used to simulate the seawater environment and control temperature and pressure, including a marine simulation experimental chamber and an experimental chamber control system.

[0039] The ocean simulation experimental chamber in the ocean simulation system is 1m long. The control system is used to control the temperature and pressure inside the chamber. It can achieve a seawater temperature change with an accuracy of 0.1℃ within the range of 0-40℃ and a pressure change with an accuracy of 0.01MPa within the range of 0-10MPa.

[0040] The signal acquisition system is used to acquire and analyze scattered light signals, including a Fabry-Perot etalon (FPE) and an enhanced charge-coupled device (ICCD). The FPE is used to split Rayleigh scattering and SBS signals, with a free spectral range (FSR) of 20.1 GHz. The ICCD is used to acquire the spectrum of the scattered light signals, with a resolution of 1024 × 256 and a pixel size of 26 μm × 26 μm.

[0041] The neural network model is integrated with other modules of the lidar system (such as pulsed lasers, polarization optical paths, and signal acquisition systems) to ensure smooth data transmission and processing. The deployed neural network model is used to process Brillouin shift and linewidth data in real time to retrieve seawater parameters including temperature, salinity, and sound velocity.

[0042] The units for frequency shift and linewidth are GHz, the unit for temperature T is °C, the unit for salinity S is ‰, and the unit for sound velocity is m / s. This fitting model is applicable to the frequency shift measurement range of 7.09 GHz ≤ vB ≤ 7.79 GHz and the linewidth measurement range of 0.33 GHz ≤ ΓB ≤ 1.83 GHz.

[0043] In summary, the effects of temperature T, salinity S, and sound speed on frequency shift v were obtained. B and line width Γ BThe inversion model.

[0044] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0045] 1. For specific sea areas or application scenarios, this invention can perform customized calibration processes to improve measurement accuracy and applicability. Through data fusion technology, multi-parameter data are comprehensively analyzed to provide more comprehensive marine environmental monitoring information.

[0046] 2. Existing methods for sound velocity detection primarily rely on indirect CTD measurements, which are not only time-consuming and labor-intensive but also unable to provide rapid real-time monitoring. This invention trains a large amount of experimental data using a neural network algorithm to establish a rapid inversion model for temperature, salinity, and especially sound velocity. This improves inversion accuracy by further expanding the measurement range of temperature, salinity, and sound velocity, ensuring accurate data can be obtained in various marine environments.

[0047] 3. In view of the shortcomings of the existing Brillouin lidar inversion model in terms of inversion accuracy and parameter range, this invention uses frequency shift and linewidth as independent variables and adopts MLP neural network to invert temperature, salinity and sound speed, thereby realizing the synchronous monitoring of ocean temperature, salinity and sound speed.

[0048] Expanding the temperature inversion range from 15℃≤T≤30℃ to 10℃≤T≤30℃ improved the temperature inversion accuracy by 0.12℃. Expanding the salinity inversion range from 30‰≤S≤35‰ to 15‰≤S≤35‰ improved the accuracy by 0.31‰. The maximum errors of the temperature, salinity, and sound velocity inversion models were 0.079℃, 0.122‰, and 0.124 m / s, respectively, with relative errors of 0.212%, 0.156%, and 0.015%. This provides more reliable and accurate data for marine monitoring, playing a vital role in the development of marine economy and marine science and technology. Attached Figure Description

[0049] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0050] Figure 1 This is a schematic diagram illustrating the principle of the seawater temperature, salinity, and sound velocity inversion method based on Brillouin lidar of this invention.

[0051] Figure 2 This is a flowchart illustrating the steps of the seawater temperature, salinity, and sound velocity inversion method based on Brillouin lidar according to the present invention.

[0052] Figure 3 This is a schematic diagram illustrating the inversion of seawater temperature, salinity, and sound velocity based on Brillouin lidar according to the present invention.

[0053] (a) shows the temperature inversion data under different linewidths; (b) shows the temperature inversion data under different frequency shifts.

[0054] (c) shows the salinity inversion data at different linewidths; (d) shows the salinity inversion data at different frequency shifts.

[0055] (e) shows the inversion data of sound velocity at different linewidths; (f) shows the inversion data of sound velocity at different frequency shifts. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] This invention provides a method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar. By measuring the Brillouin frequency shift and Brillouin linewidth, and combining this with a multilayer perceptron (MLP) neural network, the temperature, salinity, and sound velocity of seawater can be inverted, accurately obtaining these parameters. Specifically, it includes:

[0058] S100: Collect seawater parameter data and their corresponding frequency shift and linewidth for different sea areas, seasons, and depths.

[0059] A comprehensive and diverse dataset was constructed, covering various marine environmental conditions, thereby enhancing the generalization ability and applicability of subsequent models, avoiding bias caused by limited data, and providing a reliable foundation for the inversion model.

[0060] S200. Perform data preprocessing and standardization on the collected data to remove outliers, duplicates, and incomplete data records.

[0061] Temperature, salinity, sound velocity, frequency shift, and linewidth are standardized to bring them to the same order of magnitude, facilitating neural network training. Data volume is expanded through interpolation and transformation to improve the model's generalization ability.

[0062] Improving data quality and consistency reduces the interference of noise and anomalies on model training, facilitating efficient processing by neural networks. Data augmentation enhances the model's generalization ability, reduces the risk of overfitting, and ensures stable learning across different scenarios, laying a solid foundation for subsequent neural network training.

[0063] S300. A multilayer perceptron (MLP) neural network suitable for handling regression problems is selected to establish an inversion model. Ocean parameters (temperature, salinity, and sound speed) are inverted based on frequency shift and linewidth. The MLP network structure is designed as follows:

[0064] The input layer features are Brillouin frequency shift and linewidth, where the Brillouin frequency shift v B The range is limited to 7.09 GHz ≤ v B ≤7.79GHz; Brillouin linewidth Γ B The range is limited to 0.33GHz≤Γ B ≤1.83GHz.

[0065] Brillouin shift v B The calculation formula is:

[0066]

[0067] In the formula, n is the refractive index of the medium, and λ is the incident light wavelength, usually taken as λ = 532 nm. s Let θ be the speed of sound in the medium, and θ be the scattering angle (θ = 180° for backscattering).

[0068] n and v s All are related to temperature, salinity, and pressure. n adopts the GomerT model, v s The Richards model is used.

[0069] Brillouin line width Γ B The calculation formula is:

[0070]

[0071] In the formula, ρ is the density of seawater, which is generally around 1.023 × 10³; η S Let η be the shear viscosity coefficient of seawater. b is the volume viscosity coefficient of seawater.

[0072] κ is the thermal conductivity, γ is the adiabatic coefficient, and γ = C p / C v C p For the specific heat capacity at constant pressure, C v This is the specific heat capacity at constant volume.

[0073] The output layer features include seawater temperature, salinity, and sound speed.

[0074] The hidden layers consist of multiple layers of neurons, with each layer containing a certain number of neurons. The optimal number of layers and neurons per layer are determined experimentally.

[0075] The activation functions chosen are ReLU, Sigmoid, and Tanh to enhance the model's nonlinear fitting ability.

[0076] Hyperparameters were tuned using cross-validation and grid search to achieve optimal performance. Hyperparameters included the learning rate, batch size, and number of iterations. The preprocessed dataset was divided into training, validation, and test sets in a 7:2:1 ratio.

[0077] The neural network model is trained using training set data, and the model parameters are adjusted using the backpropagation algorithm to minimize the error between the predicted and actual values. Specifically:

[0078] The maximum temperature error is 0.079℃, and the relative error is 0.212%.

[0079] The maximum salinity error is 0.122‰, and the relative error is 0.156%.

[0080] The maximum error in sound velocity is 0.124 m / s, and the relative error is 0.015%.

[0081] The inversion model is integrated with the pulsed laser, polarization optical path, and signal acquisition system to ensure real-time data transmission and processing.

[0082] The inversion accuracy and generalization ability of the model are analyzed on the validation set, and the model is adjusted synchronously based on the validation results.

[0083] Test the model's performance on the test set to ensure that the model can obtain accurate data in different marine environments.

[0084] The trained MLP model is deployed to the Brillouin lidar system for real-time online inversion. The inversion model is applicable within the following ranges: Temperature: 10℃≤T≤30℃; Salinity: 15‰≤S≤35‰.

[0085] It achieves high-precision inversion, has a wide range of applicability, supports real-time online deployment, improves the efficiency and reliability of marine parameter monitoring, and solves the problems of low efficiency and poor accuracy of traditional methods.

[0086] The present invention also provides a seawater temperature, salinity and sound velocity inversion system based on Brillouin lidar, including a pulsed laser, a polarization optical path, a marine environment simulation system and a signal acquisition system.

[0087] Pulsed lasers are used to output laser sources, including the DG535 digital delay pulse generator and the injected Nd:YAG pulsed laser.

[0088] The pulsed laser source is a seed-injected Nd:YAG pulsed laser. The Nd:YAG pulsed laser outputs vertically polarized light at 532 nm. Its output parameters are: repetition frequency 10 Hz, pulse width 8 ns, beam diameter approximately 19 mm, divergence angle approximately 0.45 mrad, maximum output energy approximately 4 J / Pulse, and linewidth approximately 90 MHz in a single longitudinal mode.

[0089] The polarization optical path is used for polarization conversion. It consists of a half-wave plate, a polarizing beam splitter (PBS), and a quarter-wave plate, and can realize the conversion between vertically polarized light, horizontally polarized light, and circularly polarized light.

[0090] The marine environment simulation system is used to simulate the seawater environment and control temperature and pressure, including a marine simulation experimental chamber and an experimental chamber control system.

[0091] The ocean simulation experimental chamber in the ocean simulation system is 1m long. The control system is used to control the temperature and pressure inside the chamber. It can achieve a seawater temperature change with an accuracy of 0.1℃ within the range of 0-40℃ and a pressure change with an accuracy of 0.01MPa within the range of 0-10MPa.

[0092] The signal acquisition system is used to acquire and analyze scattered light signals, including a Fabry-Perot etalon (FPE) and an enhanced charge-coupled device (ICCD). The FPE is used to split Rayleigh scattering and SBS signals, with a free spectral range (FSR) of 20.1 GHz. The ICCD is used to acquire the spectrum of the scattered light signals, with a resolution of 1024 × 256 and a pixel size of 26 μm × 26 μm.

[0093] The neural network model is integrated with other modules of the lidar system (such as pulsed lasers, polarization optical paths, and signal acquisition systems) to ensure smooth data transmission and processing. The deployed neural network model is used to process Brillouin shift and linewidth data in real time to retrieve seawater parameters including temperature, salinity, and sound velocity.

[0094] The units for frequency shift and linewidth are GHz, the unit for temperature T is °C, the unit for salinity S is ‰, and the unit for sound velocity is m / s. This fitting model is applicable to the frequency shift measurement range of 7.09 GHz ≤ vB ≤ 7.79 GHz and the linewidth measurement range of 0.33 GHz ≤ ΓB ≤ 1.83 GHz.

[0095] In summary, the effects of temperature T, salinity S, and sound speed on frequency shift v were obtained. B and line width Γ B The inversion model.

[0096] Example 1:

[0097] Please see Figure 1The schematic diagram of the present invention shows that the seed-injected Nd:YAG pulsed laser emits laser light, which becomes horizontally polarized light after passing through a half-wave plate, is incident at Brewster angle and passes through a PBS, then becomes circularly polarized light after passing through a quarter-wave plate, and finally enters the ocean simulation system after being focused.

[0098] When the energy density of the incident light at the focal point reaches the threshold of stimulated Brillouin scattering, backscattering is excited. The scattered light is backscattered at 180°, reflected by the PBS, collimated, filtered, and then enters the FPE. The scattered signal is then acquired and analyzed by the ICCD.

[0099] The incident light intensity is monitored by an optical power meter (PM), and the timing of the entire system is controlled by a high-precision pulse delay generator (DG535) and a reverse-flow control.

[0100] Example 2:

[0101] Please see Figure 2 The flowchart of the present invention shows that when a laser is incident on water, it induces elastic scattering and inelastic scattering.

[0102] In elastic scattering, Rayleigh scattering is dominant, characterized by the frequency of the scattered light being the same as the frequency of the incident light.

[0103] Inelastic scattering is mainly manifested as Brillouin scattering, which originates from the interaction between the light wave field and the sound wave field of the medium, thus triggering the Doppler effect.

[0104] The pump light generates Stokes and anti-Stokes components under the action of the acoustic field.

[0105] When the pump light wavelength is fixed, Brillouin scattering causes a frequency shift in the scattered light, the amount of which depends on the scattering angle and the internal properties of the medium.

[0106] The frequency difference between the incident light and the scattered light is defined as the sound wave frequency, called the Brillouin frequency shift, and its expression is as follows:

[0107]

[0108] In the formula, n is the refractive index of the medium, and λ is the incident light wavelength, with a value of 532 nm. s Let θ be the speed of sound in the medium, and θ be the scattering angle (θ = 180° for backscattering).

[0109] n and v s All are related to temperature, salinity, and pressure. n adopts the GomerT model, v s Using the Richards model:

[0110] n(T,S,λ,P)=n0+n1λ -2 -n2λ-4 -n3T 2 +(n4-n5T)(n6S)+(n7P)(n8-n9S)(n 10 -n 11 T) (2)

[0111]

[0112] In the formula, S is the seawater salinity, T is the seawater temperature, and P is the seawater pressure. The refractive index n is related to the speed of sound v. s The coefficients are shown in Table 1 and Table 2.

[0113] Table 1. Coefficients of refractive index n(T,S,λ,P)

[0114] <![CDATA[n0=1.3247]]> <![CDATA[n6=4×10 -5 ]]> <![CDATA[n1=3.3×10 3 ]]> <![CDATA[n7=1.45×10 -5 ]]> <![CDATA[n2=3.2×10 7 ]]> <![CDATA[n8=1.021]]> <![CDATA[n3=-2.5×10 -6 ]]> <![CDATA[n9=6×10 -4 ]]> <![CDATA[n4=5]]> <![CDATA[n 10 =1]]> <![CDATA[n5=2×10 -2 ]]> <![CDATA[n 11 =4.5 × 10 -3 ]]>

[0115] Table 2 Speed ​​of Sound υ s The coefficients of (T,S,P)

[0116] j＝0 j＝1 j＝2 j＝3 j＝4 j＝5 <![CDATA[A 0j ]]> 1.389 <![CDATA[-1.262 × 10 -2 ]]> <![CDATA[7.164 × 10 -5 ]]> <![CDATA[2.006×10 -6 ]]> <![CDATA[-3.21×10 -8 ]]> 0 <![CDATA[A 1j ]]> <![CDATA[9.4742×10 -6 ]]> <![CDATA[-1.258×10 -5 ]]> <![CDATA[-6.4885×10 -8 ]]> <![CDATA[1.0507×10 -8 ]]> <![CDATA[-2.0122×10 -10 ]]> 0 <![CDATA[A 2j ]]> <![CDATA[-3.9064×10 -7 ]]> <![CDATA[9.1041×10 -9 ]]> <![CDATA[-16002×10 -10 ]]> <![CDATA[7.988×10 -12 ]]> 0 0 <![CDATA[A 3j ]]> <![CDATA[1.1×10 -7 ]]> <![CDATA[6.649×10 -12 ]]> <![CDATA[-3.389×10 -13 ]]> 0 0 0 <![CDATA[B 0j ]]> <![CDATA[-1.922×10 -2 ]]> <![CDATA[-4.42×10 -5 ]]> 0 0 0 0 <![CDATA[B 1j ]]> <![CDATA[7.3637×10 -5 ]]> <![CDATA[1.7945×10 -7 ]]> 0 0 0 0 <![CDATA[C 0j ]]> 1402.388 5.03711 <![CDATA[-5.8085 × 10 -2 ]]> <![CDATA[3.3420×10 -4 ]]> <![CDATA[-1478×10 -6 ]]> <![CDATA[3.1464×10 -9 ]]> <![CDATA[C 1j ]]> 0.1536 <![CDATA[6.8982×10 -4 ]]> <![CDATA[-8.1788×10 -6 ]]> <![CDATA[1.3621×10 -7 ]]> <![CDATA[-6.1185×10 -10 ]]> 0 <![CDATA[C 2j ]]> <![CDATA[3.126×10 -5 ]]> <![CDATA[-1.7107×10 -6 ]]> <![CDATA[2.5974×10 -8 ]]> <![CDATA[-2.5335×10 -10 ]]> <![CDATA[1.0405×10 -12 ]]> 0 <![CDATA[C 3j ]]> <![CDATA[-9.7729×10 -9 ]]> <![CDATA[3.8504×10 -10 ]]> <![CDATA[-2.3643×10 -12 ]]> 0 0 0 <![CDATA[D0]]> <![CDATA[1.727×10 -3 ]]> … … … … … <![CDATA[D1]]> <![CDATA[-7.9836×10 -6 ]]> … … … … …

[0117] From formulas (2) and (3), we know that n is a function of T, S, λ, and P, and v s It is a function of T, S, and P. Substituting formulas (2) and (3) into formula (1) yields v. B Expressions for T, S, λ, and P:

[0118]

[0119] In seawater parameters, the viscosity coefficient, characterizing the liquid's resistance to deformation, is a key physical quantity describing the internal frictional properties of seawater. The full width at half maximum (FWHM) of Brillouin scattering, defined as the linewidth, is determined by the viscosity coefficient and other thermodynamic parameters. The Brillouin linewidth varies with the viscosity coefficient and other thermodynamic parameters according to the Lorentz curve.

[0120]

[0121] In the formula: ρ is the density of seawater, as shown in formula (6), its value is approximately 1.023 × 10⁻⁶. 3 kg / m 3 .

[0122]

[0123] η S The shear viscosity coefficient of seawater depends on seawater temperature and salinity and is calculated using the empirical formula proposed by Richards.

[0124] η b κ is the volume viscosity of seawater; γ is the thermal conductivity; γ = Cp / C v C p For the specific heat capacity at constant pressure, C v This is the specific heat capacity at constant volume.

[0125] η S and η b As shown in formulas (7) and (8), it can be seen that the viscosity coefficient is a function of temperature and salinity. This formula is applicable to 0℃≤T≤40℃ and 0‰≤S≤40‰.

[0126]

[0127] η b (S,T)=η S (S,T)×(4.8840-0.2488T+7.3×10 -3 T 2 -0.6125×10 -5 T 3 (8)

[0128] Density ρ and shear viscosity η s The coefficients are shown in Tables 3 and 4.

[0129] Table 3. Coefficients of density ρ(T,S)

[0130] j＝0 j＝1 j＝2 j＝3 <![CDATA[E 0j ]]> 999.842595 0.824493 <![CDATA[-5.72466×10 -3 ]]> <![CDATA[4.8314×10 -4 ]]> <![CDATA[E 1j ]]> <![CDATA[6.793952×10 -2 ]]> <![CDATA[-4.0899×10 -3 ]]> <![CDATA[1.0277×10 -4 ]]> 0 <![CDATA[E 2j ]]> <![CDATA[-9.095290×10 -3 ]]> <![CDATA[7.6438×10 -5 ]]> <![CDATA[-1.6546×10 -6 ]]> 0 <![CDATA[E 3j ]]> <![CDATA[1.001685×10 -4 ]]> <![CDATA[-8.2467×10 -7 ]]> 0 0 <![CDATA[E 4j ]]> <![CDATA[-1.120083×10 -6 ]]> <![CDATA[5.3875×10 -9 ]]> 0 0 <![CDATA[E 5j ]]> <![CDATA[6.536332×10 -9 ]]> 0 0 0

[0131] Table 4 Shear viscosity coefficient η s The coefficients of (T,S)

[0132] j＝0 j＝1 j＝2 j＝3 <![CDATA[Q j ]]> <![CDATA[1.79×10 -2 ]]> <![CDATA[-6.1299×10 -4 ]]> <![CDATA[1.4467×10 -5 ]]> <![CDATA[-1.6826×10 -7 ]]> <![CDATA[R j ]]> <![CDATA[2.4727×10 -5 ]]> <![CDATA[4.8429×10 -7 ]]> <![CDATA[-4.7172×10 -8 ]]> <![CDATA[7.5986×10 -10 ]]>

[0133] The frequency shift and linewidth of the pump light in the medium were obtained using Brillouin lidar.

[0134] Based on Brillouin frequency shift v B (T, S) and Brillouin linewidth Γ B (T, S) An inversion model for temperature, salinity, and sound velocity is established. However, the inversion speed and accuracy of this model need to be improved.

[0135] Temperature inversion model T(ν) B ,Γ B Salinity inversion model S(v) B ,Γ B and the speed of sound v s (v B ,Γ B The calculation formulas for the inversion model are shown in formulas (9), (10), and (11), respectively:

[0136]

[0137] Temperature inversion model T(v) B ,Γ B Salinity inversion model S(v) B ,Γ B The coefficient of ) and the speed of sound v s (v B ,Γ B The inversion models are shown in Tables 5, 6 and 7.

[0138] Table 5 Temperature Retrieval Model T(v) B ,Γ B coefficient

[0139] <![CDATA[t1=-467.08]]> <![CDATA[t6=-4.401]]> <![CDATA[t2=-192.8]]> <![CDATA[t7=-1.137]]> <![CDATA[t3=3.48]]> <![CDATA[t8=0.7598]]> <![CDATA[t4=5.87]]> <![CDATA[t9=0.1761]]> <![CDATA[t5=-12.27]]> <![CDATA[t 10 =-2.09]]>

[0140] Table 6. Salinity Inversion Model S(v) B ,Γ B coefficient

[0141] <![CDATA[s1=779.28]]> <![CDATA[s5=24.55 <!-- 8 -->]]> <![CDATA[s2=-481.8]]> <![CDATA[s6=9.266]]> <![CDATA[s3=71.26]]> <![CDATA[s7=-6.211]]> <![CDATA[s4=-2.688]]> <![CDATA[s8=0.1634]]>

[0142] Table 7. Speed ​​of sound v s (v B ,Γ B Inversion model

[0143] <![CDATA[v1=-3047.5]]> <![CDATA[v5=-3.29]]> <![CDATA[v2=1489]]> <![CDATA[v6=3.451]]> <![CDATA[v3=-178.5]]> <![CDATA[v7=-0.7809]]> <![CDATA[v4=8.059]]> <![CDATA[v8=0.3882]]>

[0144] Example 3:

[0145] Please see Figure 3 Inversion values ​​at different frequency shifts and linewidths.

[0146] Depend on Figure 3 (a) shows that as the Brillouin frequency shift increases, the temperature T increases slowly with a power function trend.

[0147] Depend on Figure 3 (b) It can be seen that as the Brillouin linewidth increases, the temperature T decreases exponentially, changing rapidly in the range of 0.33 GHz to 0.6 GHz and slowly in the range of 0.6 GHz to 1.83 GHz.

[0148] Depend on Figure 3 (c) It can be seen that as the Brillouin shift increases, the salinity S increases in a power function trend and the growth rate is relatively fast.

[0149] Depend on Figure 3 (d) shows that as the Brillouin linewidth increases, the salinity S increases in a logarithmic manner; the larger the linewidth, the slower the salinity changes.

[0150] Depend on Figure 3(e) shows that as the Brillouin frequency shift increases, the sound speed υs increases with a power function trend.

[0151] Depend on Figure 3 (f) shows that as the Brillouin linewidth increases, the sound speed υs increases in a logarithmic manner, and the larger the linewidth, the slower the sound speed changes.

[0152] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0153] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar, characterized in that: This method retrieves seawater temperature, salinity, and sound speed by measuring Brillouin frequency shift and Brillouin linewidth, combined with a multilayer perceptron (MLP) neural network; specifically, it includes: S100: Collect seawater parameter data and their corresponding frequency shift and linewidth for different sea areas, seasons, and depths; S200. Perform data preprocessing and standardization on the collected data; The S300 uses an MLP neural network to build an inversion model, outputting temperature, salinity, and sound speed.

2. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 1, characterized in that: The data preprocessing and standardization include: S201 Removes outliers, duplicates, and incomplete data records; S202 standardizes temperature, salinity, sound velocity, frequency shift, and linewidth; S203 expands the data volume through data interpolation and transformation.

3. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 1, characterized in that: The MLP neural network includes an input layer, a hidden layer, and an output layer; The input layer features are Brillouin frequency shift and linewidth; the output layer features are seawater temperature, salinity, and sound speed; the hidden layer consists of Q-layer neurons, and the activation functions used are ReLU, Sigmoid, and Tanh. The hyperparameters, including learning rate, batch size, and number of iterations, were tuned using cross-validation and grid search. The preprocessed dataset was divided into training, validation, and test sets in a ratio of 7:2:

1. The neural network model is trained using training set data, and the model parameters are adjusted using the backpropagation algorithm to minimize the error between the predicted and actual values. Analyze the model's inversion accuracy and generalization ability on the validation set, and adjust the model synchronously based on the validation results; Test the model's performance on the test set to ensure that the model can obtain accurate data in different marine environments.

4. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 1, characterized in that: In the inversion model, the Brillouin frequency shift v B The range is limited to 7.09 GHz ≤ v B ≤7.79GHz; Brillouin linewidth Γ B The range is limited to 0.33GHz≤Γ B ≤1.83GHz.

5. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 1, characterized in that: The accuracy of the inversion model is: The maximum temperature error is 0.079℃, and the relative error is 0.212%. The maximum salinity error was 0.122‰, and the relative error was 0.156%. The maximum error in sound velocity is 0.124 m / s, and the relative error is 0.015%.

6. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 1, characterized in that: The Brillouin frequency shift v B The calculation formula is: In the formula, n is the refractive index of the medium, λ is the incident light wavelength, and v s Let θ be the speed of sound in the medium, and θ be the scattering angle. n and v s All are related to temperature, salinity, and pressure. n adopts the GomerT model, v s The Richards model is used.

7. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 6, characterized in that: The Brillouin line width Γ B The calculation formula is: In the formula, ρ is the density of seawater, and η is the density of seawater. S Let η be the shear viscosity coefficient of seawater. b The viscosity coefficient of seawater; κ is the thermal conductivity, γ is the adiabatic coefficient, and γ = C p / C v C p For the specific heat capacity at constant pressure, C v This is the specific heat capacity at constant volume.

8. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 1, characterized in that: In S400, the trained MLP model is deployed to the Brillouin lidar system for real-time online inversion; the applicable range of the inversion model is: temperature: 10℃≤T≤30℃; salinity: 15‰≤S≤35‰. The inversion model is integrated with the pulsed laser, polarization optical path, and signal acquisition system to ensure real-time data transmission and processing.

9. A seawater temperature, salinity, and sound velocity inversion system based on Brillouin lidar, characterized in that: The system includes a pulsed laser, a polarization optical path, a marine environment simulation system, and a signal acquisition system; The pulsed laser is used to output a laser source; The polarization optical path is used for polarization conversion; The marine environment simulation system is used to simulate the seawater environment and control temperature and pressure. The signal acquisition system is used to acquire and analyze scattered light signals.

10. The method for inverting seawater temperature, salinity, and sound velocity based on Brillouin lidar according to claim 9, characterized in that: The pulsed laser includes a digital delay pulse generator and an injection pulsed laser; The polarization beam path consists of a half-wave plate, a polarizing beam splitter, and a quarter-wave plate, and performs conversion between vertically polarized light, horizontally polarized light, and circularly polarized light. The marine environment simulation system includes a marine simulation experimental chamber and an experimental chamber control system; The signal acquisition system includes a Fabry-Perot etalon and an enhanced charge-coupled device.

Images