A temperature compensation method and system based on fiber-optic gyroscope temperature field modeling

Abstract

This application discloses a temperature compensation method and system based on fiber optic gyroscope temperature field modeling, relating to the field of fiber optic gyroscope temperature compensation. The method includes: modeling and simulating the temperature of the fiber optic gyroscope to obtain simulated temperature data; calculating the deviation between the simulated temperature data and the simulated set temperature data; determining a fused input temperature and a temperature change rate based on the deviation and actual physical experimental temperature data; determining a temperature compensation value using a fiber optic gyroscope temperature compensation model based on the fused input temperature and the temperature change rate; and compensating for the variable-temperature zero-bias measurement value of the fiber optic gyroscope under variable-temperature conditions based on the temperature compensation value to obtain the temperature-compensated fiber optic gyroscope zero-bias. This application can comprehensively display the overall temperature distribution of the fiber optic gyroscope at any given time, thereby improving the accuracy of fiber optic gyroscope temperature compensation.

Description

Technical Field

[0001] This application relates to the field of fiber optic gyroscope temperature compensation, and in particular to a temperature compensation method and system based on fiber optic gyroscope temperature field modeling. Background Technology

[0002] The temperature field distribution of a fiber optic gyroscope directly affects its performance stability. Uneven temperature distribution leads to uneven thermal stress in the fiber optic ring, introducing Shupe error, which increases the zero-bias error. Temperature experiments can reveal the temperature change trend of a specific point on the fiber optic gyroscope over time. However, these experiments are time-consuming and only provide temperature data for a single point, failing to comprehensively reflect the overall temperature distribution of the fiber optic gyroscope and thus affecting temperature compensation. Summary of the Invention

[0003] The purpose of this application is to provide a temperature compensation method and system based on fiber optic gyroscope temperature field modeling, which can comprehensively display the overall temperature distribution of the fiber optic gyroscope at any time, thereby improving the accuracy of fiber optic gyroscope temperature compensation.

[0004] To achieve the above objectives, this application provides the following solution:

[0005] Firstly, this application provides a temperature compensation method based on fiber optic gyroscope temperature field modeling, including:

[0006] Modeling and temperature simulation of the fiber optic gyroscope were performed to obtain simulated temperature data;

[0007] Calculate the deviation between the simulated temperature data and the simulated set temperature data;

[0008] The fusion input temperature and temperature change rate are determined based on the aforementioned deviation and actual physical experiment temperature data.

[0009] Construct a temperature compensation model for fiber optic gyroscopes;

[0010] Based on the fused input temperature and the temperature change rate, the temperature compensation value is determined using a fiber optic gyroscope temperature compensation model.

[0011] The temperature compensation value is used to compensate the temperature-compensated zero-bias measurement value of the fiber optic gyroscope under variable temperature conditions, so as to obtain the temperature-compensated zero-bias of the fiber optic gyroscope.

[0012] Secondly, this application provides a temperature compensation system based on fiber optic gyroscope temperature field modeling, comprising:

[0013] The temperature modeling and simulation module is used to model and simulate the temperature of the fiber optic gyroscope to obtain simulated temperature data.

[0014] The deviation calculation module is used to calculate the deviation between the simulated temperature data and the simulated set temperature data;

[0015] A module for determining the fusion input temperature and temperature change rate is used to determine the fusion input temperature and temperature change rate based on the deviation and actual physical experiment temperature data.

[0016] The model building module is used to build a temperature compensation model for fiber optic gyroscopes.

[0017] The temperature compensation value determination module is used to determine the temperature compensation value based on the fused input temperature and the temperature change rate, using a fiber optic gyroscope temperature compensation model.

[0018] The temperature compensation module is used to compensate the temperature-dependent zero-bias measurement value of the fiber optic gyroscope under variable temperature conditions based on the temperature compensation value, so as to obtain the temperature-compensated zero-bias of the fiber optic gyroscope.

[0019] According to the specific embodiments provided in this application, this application has the following technical effects:

[0020] This application provides a temperature compensation method and system based on fiber optic gyroscope temperature field modeling. By modeling and simulating the temperature of the fiber optic gyroscope, the temperature information of the gyroscope at any point at any time can be obtained, intuitively reflecting the temperature field distribution of the entire fiber optic gyroscope and greatly shortening the experimental time. Based on the modeling and temperature simulation of the fiber optic gyroscope, the input temperature and temperature change rate are calculated and fused, and the temperature compensation value is predicted through the fiber optic gyroscope temperature compensation model. This achieves compensation for the variable temperature zero-bias measurement value of the fiber optic gyroscope under variable temperature conditions, improving the accuracy of fiber optic gyroscope temperature compensation. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A schematic flowchart of a temperature compensation method based on fiber optic gyroscope temperature field modeling provided in an embodiment of this application;

[0023] Figure 2 This is a schematic diagram of the finite element model of a fiber optic gyroscope.

[0024] Figure 3 The simulated temperature distribution cloud map of the entire fiber optic gyroscope at a certain moment of heating;

[0025] Figure 4The simulated temperature distribution cloud maps of the fiber optic ring at four stages are shown. (a) shows the temperature field distribution of the fiber optic ring at a certain moment of cooling, (b) shows the temperature field distribution of the fiber optic ring at a certain moment of holding at -40℃, (c) shows the temperature field distribution of the fiber optic ring at a certain moment of heating, and (d) shows the temperature field distribution of the fiber optic ring at a certain moment of holding at 70℃.

[0026] Figure 5 A schematic diagram for setting up a physical experiment;

[0027] Figure 6 The diagram shows the maximum probe of the fiber optic ring at a certain moment during the cooling process. (a) is the maximum probe of the fiber optic ring at 1455s, (b) is the maximum probe of the fiber optic ring at 2233.3s, and (c) is the maximum probe of the fiber optic ring at 3045.3s.

[0028] Figure 7 The diagram shows the minimum probe of the fiber optic ring at a certain moment during the cooling process. (a) is the minimum probe of the fiber optic ring at 1455s, (b) is the minimum probe of the fiber optic ring at 2233.3s, and (c) is the minimum probe of the fiber optic ring at 3045.3s.

[0029] Figure 8 The diagram shows the maximum probe size of the fiber optic ring at a certain moment during the heating process. (a) is the maximum probe size of the fiber optic ring at 15561s, (b) is the maximum probe size of the fiber optic ring at 16648s, and (c) is the maximum probe size of the fiber optic ring at 18471s.

[0030] Figure 9 The diagram shows the minimum probe of the fiber optic ring at different times during the heating process, where (a) is the minimum probe of the fiber optic ring at 15561s, (b) is the minimum probe of the fiber optic ring at 16648s, and (c) is the minimum probe of the fiber optic ring at 18471s.

[0031] Figure 10 This is a schematic diagram of the local maximum probe size of the fiber optic ring at a certain moment during the heating process.

[0032] Figure 11 A schematic diagram for selecting a temperature reference point;

[0033] Figure 12 A comparison chart of simulated temperature data (T1, T2, T3) and actual physical experiment temperature data;

[0034] Figure 13 The diagram shows the deviation between the simulated temperature data and the simulated set temperature data. (a) represents the deviation between T1 and the simulated set temperature data, (b) represents the deviation between T2 and the simulated set temperature data, and (c) represents the deviation between T3 and the simulated set temperature data.

[0035] Figure 14 A schematic diagram of the improved sand cat swarm optimization algorithm;

[0036] Figure 15 This is a schematic diagram of the BP neural network structure. Detailed Implementation

[0037] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0039] In one exemplary embodiment, such as Figure 1 As shown, a temperature compensation method based on fiber optic gyroscope temperature field modeling is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is described using a server as an example, and includes the following steps S1 to S6. Wherein:

[0040] S1: Model and simulate the temperature of the fiber optic gyroscope to obtain simulated temperature data. The simulated temperature data includes key temperature feature points and temperature reference points; the key temperature feature points include the maximum temperature points during the cooling and heating processes, or the minimum temperature points during the cooling and heating processes.

[0041] S2: Calculate the deviation between the simulated temperature data and the simulated set temperature data.

[0042] S3: Determine the fusion input temperature and temperature change rate based on the aforementioned deviation and actual physical experiment temperature data.

[0043] S4: Construct a temperature compensation model for fiber optic gyroscopes.

[0044] S5: Based on the fused input temperature and the temperature change rate, the temperature compensation value is determined using a fiber optic gyroscope temperature compensation model.

[0045] S6: Based on the temperature compensation value, compensate the temperature-compensated zero-bias measurement value of the fiber optic gyroscope under variable temperature conditions to obtain the temperature-compensated zero-bias of the fiber optic gyroscope.

[0046] In a specific embodiment, step S1 specifically includes:

[0047] S11: Create the finite element model of the fiber optic gyroscope.

[0048] Fiber optic gyroscopes have a relatively complex internal structure, containing numerous holes, chamfers, and screw holes, which can affect the quality of mesh generation for finite element analysis using ANSYS software, leading to unnecessary calculation errors. Therefore, when creating a 3D model using Creo software, appropriate simplification is performed to create a finite element model, such as... Figure 2 As shown.

[0049] S12: Perform experimental simulation on the finite element model to obtain the simulated temperature distribution cloud map of the fiber optic gyroscope.

[0050] The simplified finite element model of the fiber optic gyroscope with the fiber optic loop was imported into a transient thermal analysis project for experimental simulation. The parameters determined included basic physical properties such as material density, thermal conductivity, and specific heat capacity, as well as estimations of the average heat generation of the circuit board and the convective heat transfer coefficient in the boundary conditions.

[0051] After completing the above steps, the finite element model is solved to obtain the temperature distribution cloud map of the fiber optic gyroscope and the temperature value at any point inside. Figure 3 The simulated temperature distribution cloud map of the fiber optic gyroscope at a certain moment during heating. Figure 4 The simulated temperature distribution cloud map shows the four stages of the fiber optic ring.

[0052] The simulation conditions should be consistent with the steps of the actual physical experiment. The actual physical experiment involves placing the fiber optic gyroscope in a high and low temperature test chamber to conduct a temperature test, ensuring that the simulation results match the actual test conditions and accurately reflect the performance of the fiber optic gyroscope at different temperatures. The fiber optic gyroscope is equipped with a temperature sensor, and the static zero-bias output data of the fiber optic gyroscope and the temperature data of the temperature sensor are measured through the host computer data acquisition software. The sampling frequency of the fiber optic gyroscope data output acquisition system is 1Hz.

[0053] The initial conditions mainly include the initial ambient temperature and the initial temperature of the fiber optic gyroscope, both set to 20℃. During heating and cooling, the ambient temperature changes by 1℃ every 60 seconds. The holding time at -40℃ and 70℃ is set to 150 minutes. The physical test settings are as follows: Figure 5 As shown.

[0054] S13: Based on the simulated temperature distribution cloud map, determine the maximum and minimum temperature points during the cooling and heating processes.

[0055] S14: Extract the temperature node between the maximum and minimum temperature points as the temperature reference point.

[0056] Steps S13-S14 specifically include the following:

[0057] based on Figure 4This analysis examines the temperature field distribution of a fiber optic gyroscope at specific moments during four stages: cooling from 20°C to -40°C, holding at -40°C, heating from -40°C to 70°C, and holding at 70°C. During the heating and cooling processes, random moments are selected, and maximum and minimum temperature probes are added at these moments. Figures 6-9 As shown.

[0058] Analysis of the simulated temperature distribution cloud map and temperature probe of the fiber optic ring reveals that the node positions of the maximum and minimum temperatures interchange during both cooling and heating processes. During cooling, the maximum temperature point is fixed at node 231433, and the minimum temperature point is fixed at node 277993. Conversely, during heating, the maximum temperature point is located at node 277993, and the minimum temperature point is fixed at node 231433. In other words, the node positions of the temperature extreme points are fixed at 231433 and 277993, indicating that these two nodes are located at the most thermally sensitive positions in the fiber optic ring. Their response to temperature changes under different thermodynamic states is symmetrical and consistent. These positions may be located at key locations in the heat conduction path of the fiber optic ring or at key material interfaces, making them more likely to reach temperature extremes than other nodes during temperature changes.

[0059] In summary, nodes 231433 and 277993 are likely special heat-sensitive regions in the fiber optic ring. Influenced by the heat conduction path and the thermal properties of the material, they are the locations most prone to extreme values ​​during temperature changes. Furthermore, the temperature distribution of the fiber optic ring exhibits symmetry and consistency under different thermodynamic states. These patterns provide a theoretical basis for optimizing the temperature compensation model of the fiber optic gyroscope and help to understand the thermodynamic behavior of the fiber optic ring under different temperature changes.

[0060] Create a local maximum probe at a specific moment during the heating process, such as Figure 10 As shown. By Figure 10 As can be seen, the six local temperature maximum points are distributed in a roughly triangular pattern, with three vertices located on the surface of the fiber optic ring parallel to the upper and lower shells and close to the upper shell. At this moment, the maximum temperature difference at the local temperature maximum point is 0.1196℃, which is only 4.6% of the temperature range at the same time point (2.6052℃). This distribution pattern is consistent throughout the simulation process for both the local temperature maximum and minimum points (during cooling and the -40℃ holding process). Therefore, the local temperature maximum point can be selected as the temperature feature point of the fiber optic ring for temperature extraction. Similarly, the temperature minimum point can be selected as the feature point for temperature extraction. In this embodiment, the temperature maximum point is selected as the temperature feature point of the fiber optic ring for temperature extraction.

[0061] Key temperature feature points, nodes 231433 and 277993, were extracted from the fiber optic ring. Simultaneously, the node temperatures between the extreme temperatures were extracted as temperature reference points. As shown in the simulated temperature distribution cloud map of the fiber optic ring presented earlier, except for the nodes surrounding the maximum and minimum temperatures, the temperature on the surface of the fiber optic ring parallel to the upper and lower outer shells is relatively uniform. Therefore, [the following is a selection / selection]... Figure 11 The node at the indicated location has a node count of 181943.

[0062] Identify the nodes from which temperature information needs to be extracted, create a coordinate system at these nodes, and rename them: name node 231433 as T1, node 181943 as T2, and node 277993 as T3. Insert a temperature probe in the solution (taking T1 as an example, and the same applies to T2 and T3), select the T1 coordinate as the location, retrieve the results of this temperature probe, and copy and save the results.

[0063] Since the simulation step size is set to 10, meaning temperature data is collected every 10 seconds, the temperature data collected by the built-in temperature sensor on the fiber optic gyroscope is smoothed every 10 seconds, and the resulting actual physical experimental temperature data is denoted as T0. Comparing the simulated temperature data (T1, T2, T3) with the actual physical experimental temperature data (T0), as follows... Figure 12 As shown. Figure 12 In the simulation, there were some discrepancies between the simulated temperature data and the actual physical experiment temperature data during certain time periods, but the overall trend was consistent, reflecting that the finite element model accurately captured the changes in the temperature field of the fiber optic ring.

[0064] To demonstrate the reliability of simulated temperature data, the following methods can be used to quantitatively assess the consistency between simulated temperature data and actual physical experiment temperature data. These methods can provide intuitive numerical evaluations, thereby proving the reliability of simulated temperature data.

[0065] (1) Root Mean Square Error (RMSE).

[0066] RMSE is a standard method for measuring the difference between simulated temperature data and actual physical experimental temperature data. It quantifies the magnitude of the error by calculating the square root of the squared mean of the temperature difference. A smaller RMSE indicates that the simulated temperature (i.e., the simulated temperature data) is closer to the experimental temperature (i.e., the actual physical experimental temperature data). The calculation formula is:

[0067]

[0068] In the formula, and They are respectively the i-th t The simulated temperature and experimental temperature at each moment, n *This represents the total number of data points.

[0069] (2) Mean Absolute Error (MAE).

[0070] MAE is another standard for measuring the error between simulated and experimental temperatures. It only considers the absolute value of the error. The smaller the MAE, the closer the simulated temperature is to the experimental temperature. Its calculation formula is:

[0071]

[0072] (3) Relative Error (RE).

[0073] Relative error is used to evaluate the proportion of the deviation between the simulated temperature and the experimental temperature, expressed as a percentage. It helps to quantify the magnitude of the deviation between the simulated temperature and the experimental temperature. The smaller the value, the smaller the relative error between the simulated temperature and the experimental temperature, and the higher the reliability of the simulation. Its calculation formula is:

[0074]

[0075] (4) Pearson Correlation Coefficient (PCCs).

[0076] The Pearson correlation coefficient is an indicator that measures the linear correlation between simulated and experimental temperatures. It is calculated by dividing the covariance of two sets of data by their respective standard deviations, yielding a value between -1 and 1. The closer the value is to 1, the stronger the linear relationship between the two. The formula is as follows:

[0077]

[0078] In the formula, and These are the average values ​​of the simulation data and the experimental data, respectively.

[0079] (5) Index of Agreement (IA).

[0080] IA is an index used to evaluate the consistency between simulated and experimental temperatures. It is defined as a value between 0 and 1, where 1 represents complete consistency and 0 represents complete inconsistency. The closer IA is to 1, the more consistent the simulated and experimental temperatures are. Its calculation formula is as follows:

[0081]

[0082] (6) Regression Analysis.

[0083] The goodness of fit between simulated and experimental temperatures was analyzed by performing linear regression on the two data points. The regression equation and coefficient of determination R0 were then used to analyze the results. 2 The coefficient of determination (R²) is used to measure the degree of agreement between simulated and experimental temperatures. The closer the R² is to 1, the better the simulated temperature can explain variations in the experimental temperature. 2 The calculation formula is:

[0084]

[0085] In the formula, These are the simulated values ​​obtained from linear regression fitting.

[0086] The consistency between the simulated temperature and the actual temperature was quantitatively evaluated using the above six methods. The accuracy evaluation results of the finite element model are shown in Table 1.

[0087] Table 1

[0088]

[0089]

[0090] As shown in the table above, the three temperature nodes T1, T2, and T3 all have high reliability and can provide stable input data for the fiber optic gyroscope temperature compensation model.

[0091] In one specific embodiment, step S2 specifically includes:

[0092] Compare T1, T2, and T3 with the simulation set temperature data (T) that has been smoothed for 10 seconds. sets The differences between the simulated temperature data and the simulated set temperature data are calculated using the following formulas:

[0093] dT1=T1-T sets ;

[0094] dT2=T2-T sets ;

[0095] dT3=T3-T sets

[0096] The deviation between the simulated temperature and the set temperature is as follows: Figure 13 As shown.

[0097] In a specific embodiment, step S3 specifically includes:

[0098] The calculated deviation was added to the actual physical experiment temperature data T0 to obtain the new temperature T1. * T2 * and T3 * The calculation formula is as follows:

[0099] T1 * =T0+dT1;

[0100] T2 * =T0 + dT2;

[0101] T3 * =T0+dT3

[0102] This processing method can effectively correct the deviations in the simulated temperature data, making the simulated temperature data closer to the real temperature, thereby improving the input accuracy of the fiber optic gyroscope temperature compensation model.

[0103] T1 * T2 * and T3 * The average of the three values ​​is used to obtain the fused input temperature T, which contains rich temperature information. * For T * The rate of temperature change dT is obtained by performing the time derivative. * It can reflect the instantaneous thermal effects inside the fiber optic gyroscope, helping the fiber optic gyroscope temperature compensation model to better adapt to changes in complex temperature fields.

[0104] In the above processing, based on the simulated temperature data (T1, T2, and T3) representing three different temperature characteristic regions within the fiber optic ring, a more comprehensive fused input temperature T is generated. * This temperature measurement provides a more comprehensive reflection of the thermodynamic state of the entire fiber optic ring. This multi-point fusion approach captures more complex temperature field changes than a single node temperature, especially under non-uniform temperature gradients. Averaging these three temperatures eliminates potential local errors and improves the representativeness of the temperature data. Simultaneously, utilizing the temperature residual effectively corrects deviations in the simulated temperature data, resulting in a more accurate final T0. * It more closely approximates the actual temperature, improving the input accuracy of the fiber optic gyroscope temperature compensation model.

[0105] In one specific embodiment, step S4 specifically includes:

[0106] S41: The BP neural network is optimized using an improved sand cat swarm optimization algorithm; the improved sand cat swarm optimization algorithm is to introduce chaotic perturbation and Gaussian mutation mechanism into the sand cat swarm optimization algorithm.

[0107] The Sand Cat Swarm Optimization (SCSO) algorithm simulates two main foraging behaviors of sand cats: searching for prey and attacking prey. This strategy of mimicking the natural behavior of sand cats makes the SCSO algorithm more efficient in the global search process, helping to avoid getting trapped in local optima. Moreover, compared with traditional BP neural networks, BP neural networks optimized by SCSO converge to a satisfactory solution faster during the iteration process, thus reducing training time and improving learning efficiency.

[0108] For example, in a D-dimensional optimization problem, the sand cat is a 1×D-dimensional array representing the solution to the problem, defined as:

[0109] X i = [x1,x2,…,x D ];

[0110] In the formula, X i For the i-th sand cat, each variable has values ​​x1, x2, ... x. D It is a floating-point number, and each x must be between the upper and lower boundaries.

[0111] The main steps of the sand cat swarm optimization algorithm are as follows:

[0112] (1) Randomly initialize the location of the sand cat group.

[0113] According to the size of the problem (N) pop ×N D ), (pop=1,…,s), create an initialization matrix using the sand cat population:

[0114]

[0115] In the formula, X is the population matrix of the sand cat group. i (i = 1, 2, ..., s) represents the i-th sand cat. ij Let be the j-th dimension coordinate of the i-th sand cat, where j = 1, 2, ..., d. The fitness cost of each sand cat is obtained by evaluating a defined fitness function, F:

[0116]

[0117] In the formula, F i (i = 1, 2, ..., s) represents the fitness function value calculated for the i-th sand cat, f(X) represents the fitness of X (the population matrix of the sand cat population) obtained using the fitness function f, and f(x) represents the fitness of X (the population matrix of the sand cat population). i1 ,x i2 ,…,x id The meaning is similar to f(X), and the part in parentheses represents the position coordinates of the i-th sand cat.

[0118] (2) Calculate the location of the sand cat group and update the global optimal location information.

[0119] (3) Calculate the conversion factor R.

[0120] After initialization, the sand cat pack begins its search for or attack of prey, i.e., the exploration or development phase. The main parameter controlling the transition between the exploration and development phases is the conversion coefficient R, which is calculated using the following formula:

[0121] R = 2r G rand(0,1)-r G ;

[0122] In the formula, rand(0,1) represents a random number between 0 and 1. G The sensitivity coefficient for the sand cat group is used to guide the transition coefficient R to achieve inter-stage transfer control. Its calculation formula is:

[0123]

[0124] In the formula, z is the current iteration number, Z max S represents the maximum number of iterations. M For the auditory coefficient of the sand cat group, S M The value is 2.

[0125] (4) Search for prey.

[0126] When the transformation coefficient |R|>1, the sand cat enters the exploration phase. At this time, the i-th sand cat in the sand cat group is determined based on its current position. Best candidate position And the cat's sensitivity range r i The location is updated according to the following formula to find other possible best prey locations.

[0127]

[0128] r i =r G rand(0,1);

[0129] In the formula, P OS (z+1) represents the updated position.

[0130] (5) Attack the prey.

[0131] When the transformation coefficient |R|≤1, the sand cat enters the development phase. At the current iteration number z, the i-th sand cat in the sand cat group is determined based on the best candidate position. With current location Generate a random location The calculation formula is as follows:

[0132]

[0133] Assuming the sand cat's sensitivity range is a circle, a roulette wheel selection algorithm is used to randomly select an angle θ for each sand cat. Each sand cat in the group can then move along different circumferential directions in the search space. The sand cat updates its position according to the following formula to attack its prey.

[0134]

[0135] In the formula, P OS (z+1) represents the position when the iteration number is z+1.

[0136] The Sand Cat Swarm Optimization algorithm improves the global search capability of neural networks by balancing the exploration and development phases, conducting multi-dimensional searches in the space, and dynamically adjusting the search strategy. By optimizing weights and biases, adjusting network parameters such as activation function parameters and learning rate, it reduces training errors and improves the network's ability to fit complex problems, making network training more stable and efficient.

[0137] (6) Determine whether to output the optimal threshold.

[0138] Determine if the number of iterations has reached the maximum number of iterations. If the maximum number of iterations has been reached, output the optimal threshold. If the maximum number of iterations has not been reached, calculate the location of the sand cat group and update the global optimal location information.

[0139] In SCSO, even after the sand cat population completes its position update through the exploration or development phase, some inferior solutions may still exist. Since the algorithm lacks a population mutation mechanism, these inferior solutions can cause the algorithm to get trapped in local optima and struggle to escape. Therefore, to improve the algorithm's global search capability and avoid premature convergence, it is necessary to introduce a mutation mechanism in SCSO to maintain the diversity of the sand cat population and ensure the exploration and optimization of better solutions within the complex search space.

[0140] The initial distribution of the population directly affects the convergence speed and solution accuracy of SCSO. Using a chaotic sequence instead of the traditional random initialization method can generate a population distribution with higher diversity and ergodicity in the initial stage of SCSO, thereby expanding the coverage of SCSO in the search space. By applying chaotic perturbation to the optimal individual, SCSO can effectively escape local optima when it stagnates in the local search and continue the global search. This strategy not only improves the quality of the solution but also significantly enhances the efficiency and robustness of SCSO in solving complex problems. Based on this, in this embodiment, this application introduces chaotic perturbation and Gaussian mutation mechanisms into the Sand Cat swarm optimization algorithm to obtain an improved Sand Cat swarm optimization algorithm.

[0141] The improvement process for the sand cat swarm optimization algorithm includes:

[0142] (1) Replace random initialization with chaotic sequence.

[0143] SCSO initializes the population using a random function, which leads to insufficient diversity and uneven distribution of individuals in the sand cat population within the search space, thus affecting global search capabilities. Statistical Parametric Mapping (SPM), on the other hand, possesses a highly uniform distribution function and excellent pseudo-random properties. Using SPM to generate chaotic variables allows for a more balanced distribution of population resources within the search space, effectively avoiding redundant initialization, maintaining population diversity, and preventing SCSO from getting trapped in local optima. This expands the search space and improves global search capabilities. The iterative method of chaotic mapping is expressed as follows:

[0144]

[0145] In the formula, η∈(0,0.5), μ∈(0,0.5), and rand(0,1) is a random number between 0 and 1.

[0146] The sequences generated by SPM exhibit better ergodicity and randomness, with a highly uniform distribution and excellent performance. Therefore, this application chooses SPM to initialize SCSO. Due to their nonlinear and ergodic characteristics, chaotic sequences can effectively prevent the algorithm from getting trapped in local optima and expand the search space.

[0147] (2) SPM chaotic perturbation and Gaussian mutation mixed perturbation strategy.

[0148] To avoid generating inferior solutions and ensure the diversity of the sand cat population, SCSO introduces chaotic perturbation and Gaussian mutation mechanisms. The currently found optimal solution is perturbed, meaning the optimal individual in the sand cat population is searched again globally to ensure a better solution is found. If the new position of the sand cat generated through chaotic perturbation or Gaussian mutation is better than the original position, the current optimal position is replaced. Otherwise, the original position is retained.

[0149] For the sand cat in the current optimal position (i.e., the optimal individual P) OSb Introducing chaotic perturbation, the position of the optimal individual is perturbed through chaotic mapping, and then restored to the original domain to obtain a new position P. new (i.e., new individuals) have:

[0150]

[0151] P new =lb+f SPM (P s )·(ub-lb).

[0152] The current best individual P OSb The individual P is obtained by mapping its original domain [lb, ub] to the standardized interval [0, 1]. S This approach helps to standardize the scale of the variables, thus enabling a more efficient exploration of the entire search space when applying chaotic sequences for perturbation. The SPM chaotic mapping function f is then called. SPM (*) A chaotic sequence is generated to perturb the current optimal individual. The result after perturbation by the chaotic sequence is restored to the original domain, ensuring that the new position is still within the specified range. Due to its nonlinear and ergodic characteristics, the chaotic sequence can effectively prevent the algorithm from getting trapped in local optima and expand the search space.

[0153] Gaussian mutation of a sand cat population involves generating random numbers using a Gaussian distribution. The location of the sand cat colony is perturbed. The Gaussian distribution and the post-perturbed location update process are represented as follows:

[0154]

[0155] In the formula, σ is a random number between 0 and 1, and σ is the standard deviation of the Gaussian distribution.

[0156] Chaotic perturbation breaks out of local optima by subtly perturbing the optimal individual. Gaussian mutation introduces more randomness, increasing its exploration range in the search space. Combining the advantages of both chaotic perturbation and Gaussian mutation, a hybrid perturbation strategy of SPM and Gaussian mutation is formulated to mutate the SCSO to find the current optimal position perturbation.

[0157] Calculate the average fitness F of all individuals in the population. avg This value is used to distinguish which individuals need chaotic perturbation and which individuals need Gaussian mutation. Chaotic perturbation is suitable for better individuals, aiming to escape local optima through perturbation. Gaussian mutation uses random numbers that follow a normal distribution. To change the individual's position. If the fitness F of the current best sand cat individual is... i Greater than or equal to the average fitness F of the population avg If the fitness F is high, then chaotic perturbation is used to optimize the individual. Otherwise, if the individual's fitness F is low, then chaotic perturbation is used to optimize the individual. i Below average F avg Then, Gaussian mutation is performed on the sample to introduce randomness and increase the diversity of individuals, ensuring the comprehensiveness of the search process.

[0158] Whether the result is from chaotic perturbation or Gaussian mutation, it must be ensured that it remains within the domain. Calculate the new fitness F. new If the fitness is better than the fitness F of the current individual iIf the condition is met, then update the individual's position and fitness. Among all the updated individuals, find the optimal individual and update the global optimal solution.

[0159] The implementation process of the improved Sand Cat Swarm Optimization (ISCSO) algorithm, which optimizes the algorithm by replacing random initialization with chaotic sequences and using a mixed perturbation strategy of SPM and Gaussian mutation, is as follows: Figure 14 As shown.

[0160] By combining two different strategies—chaotic perturbation and Gaussian mutation—the algorithm maintains its global search capability while enhancing its local exploitation ability. Chaotic perturbation is suitable for well-performing individuals, subtly perturbing the optimal individual to escape local optima traps. Gaussian mutation introduces more randomness, especially for poorly performing individuals, increasing their exploration range in the search space through random mutation and improving population diversity. This combined strategy allows the algorithm to maintain its global search capability while avoiding premature convergence, increasing the probability of finding the global optimum.

[0161] Based on the above description, the implementation process of step S41 in this application includes:

[0162] Step 1: Initialize the structure of the BP neural network and set the relevant parameters to obtain the parameter combination. The parameter combination includes: the weight matrix from the input layer to the hidden layer, the bias vector of the hidden layer, the weight matrix from the hidden layer to the output layer, and the bias vector of the output layer. The structure of the BP neural network is as follows: Figure 15 As shown.

[0163] In this step, the number of neurons in the input, hidden, and output layers of the BP neural network is set, and the parameters of the sand cat swarm optimization algorithm are defined, including the sand cat auditory coefficient S. M =2, population size s=30, maximum number of iterations Z max =500, the upper and lower bounds of the search space lb=-1 and ub=1, the parameter dimension dim=33, and the solution space X wb This solution space includes all the values ​​and thresholds in the BP neural network, representing the solution space of the optimization problem, as follows:

[0164] X wb =[w ν ,b ν ,w o ,b o ];

[0165] In the formula, w ν Let b be the weight matrix from the input layer to the hidden layer. ν w is the bias vector of the hidden layer.o Let b be the weight matrix from the hidden layer to the output layer. o This is the bias vector for the output layer.

[0166] The formulas for calculating the hidden layer output matrix ν and the output layer output matrix α are as follows:

[0167] ν=f1(w ν x+b ν );

[0168] α=f2(w o η+b o );

[0169] In the formula, f1 is the logsig function and f2 is the pureline function.

[0170] The total number of these parameters constitutes the dimension of the search space for the sand cat swarm optimization algorithm. For example, if the network's input layer has n... in There are n nodes, and the hidden layer has n nodes. hidden There are n nodes, and the output layer has n nodes. out If there are n nodes, then the total number of parameters N is:

[0171] N=(n in ×n hidden )+n hidden +(n hidden ×n out )+n out .

[0172] Step 2: Using a chaotic mapping method, the original domain is obtained by initializing the sand cat population positions based on parameter combinations, and the fitness values ​​of individuals and the average fitness value of all individuals in the original domain are determined. This is called SPM chaotic mapping initialization of the sand cat population positions. Specifically, the position of each sand cat in the population is randomly generated, and each position corresponds to a weight and threshold combination in a BP neural network.

[0173] Step 3: Perform chaotic perturbation on individuals with fitness values ​​greater than or equal to the average fitness value, perform Gaussian mutation on individuals with fitness values ​​less than the average fitness value, and perform iterative updates on individuals until the maximum number of iterations or the convergence condition (such as the prediction error reaching the set target) is reached, to obtain the optimal individual, and use the optimal individual to update the position of individuals in the sand cat group to obtain the global optimal solution.

[0174] Step 4: Use the global optimal solution as the optimal parameter combination for the BP neural network to obtain the optimized BP neural network.

[0175] According to the ISCSO search mechanism, the sand cat swarm continuously updates its position to gradually approach the global optimum in the search space. During this process, the control parameter (i.e., the sand cat swarm sensitivity coefficient r)...G The sensitivity range is used to guide the sand cat population, ensuring that the algorithm performs a broad global search in the early stages and a precise local search in the later stages. ISCSO deconstructs each sand cat location (i.e., the parameter combination of the BP neural network) into specific weights and thresholds, and then assigns these parameters to the BP neural network to obtain the optimized BP neural network.

[0176] S42: Train and test the optimized BP neural network to obtain the fiber optic gyroscope temperature compensation model.

[0177] Step 1: Obtain the training sample set and the test sample set.

[0178] Step 2: Train and test the optimized BP neural network using training and testing sample sets until the prediction error of the optimized BP neural network reaches the set target, thus obtaining the fiber optic gyroscope temperature compensation model.

[0179] Specifically, the input temperature T of the sample fusion after 10 seconds of smoothing is taken as an example. * 'With the rate of change of sample temperature dT * As input, the temperature-variable zero-bias prediction value of the fiber optic gyroscope under varying temperature conditions after 100s smoothing is used as output, and the difference between the predicted value and the measured value is used as the temperature compensation value to compensate for the measured value.

[0180] The fiber optic gyroscope temperature compensation model can significantly reduce the impact of temperature changes on the output of the fiber optic gyroscope, improve zero-bias stability, and significantly improve compensation accuracy and stability. Especially in complex temperature scenarios, the fiber optic gyroscope temperature compensation model shows stronger adaptability, improves the accuracy and reliability of fiber optic gyroscope measurements, and provides a better solution for temperature compensation of fiber optic gyroscopes in practical applications.

[0181] Based on the same inventive concept, this application also provides a system for implementing the temperature compensation method based on fiber optic gyroscope temperature field modeling as described above. The solution provided by this system is similar to the implementation described in the above method. Therefore, the specific limitations in one or more embodiments of the temperature compensation system based on fiber optic gyroscope temperature field modeling provided below can be found in the limitations of the temperature compensation method based on fiber optic gyroscope temperature field modeling described above, and will not be repeated here.

[0182] In one exemplary embodiment, a temperature compensation system based on fiber optic gyroscope temperature field modeling is provided, comprising:

[0183] The temperature modeling and simulation module is used to model and simulate the temperature of fiber optic gyroscopes to obtain simulated temperature data.

[0184] The deviation calculation module is used to calculate the deviation between the simulated temperature data and the simulated set temperature data.

[0185] The module for determining the fusion input temperature and temperature change rate is used to determine the fusion input temperature and temperature change rate based on the deviation and actual physical experiment temperature data.

[0186] The model building module is used to build temperature compensation models for fiber optic gyroscopes.

[0187] The temperature compensation value determination module is used to determine the temperature compensation value based on the fused input temperature and the temperature change rate, using a fiber optic gyroscope temperature compensation model.

[0188] The temperature compensation module is used to compensate the temperature-dependent zero-bias measurement value of the fiber optic gyroscope under variable temperature conditions based on the temperature compensation value, so as to obtain the temperature-compensated zero-bias of the fiber optic gyroscope.

[0189] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0190] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0191] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0192] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A temperature compensation method based on fiber optic gyroscope temperature field modeling, characterized in that, include: Modeling and temperature simulation of the fiber optic gyroscope were performed to obtain simulated temperature data; Calculate the deviation between the simulated temperature data and the simulated set temperature data; The fusion input temperature and temperature change rate are determined based on the aforementioned deviation and actual physical experiment temperature data. Construct a temperature compensation model for fiber optic gyroscopes; Based on the fused input temperature and the temperature change rate, the temperature compensation value is determined using a fiber optic gyroscope temperature compensation model. The temperature compensation value is used to compensate the temperature-compensated zero-bias measurement value of the fiber optic gyroscope under variable temperature conditions, so as to obtain the temperature-compensated zero-bias of the fiber optic gyroscope.

2. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 1, characterized in that, The simulated temperature data includes key temperature feature points and temperature reference points; the key temperature feature points include the maximum temperature points during the cooling and heating processes, or the minimum temperature points during the cooling and heating processes. Modeling and temperature simulation of the fiber optic gyroscope were performed to obtain simulated temperature data, specifically including: Create a finite element model of the fiber optic gyroscope; Experimental simulation was performed on the finite element model to obtain the simulated temperature distribution cloud map of the fiber optic gyroscope; Based on the simulated temperature distribution cloud map, the maximum and minimum temperature points during the cooling and heating processes are determined. Extract the temperature node between the maximum and minimum temperature points as the temperature reference point.

3. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 1, characterized in that, The fusion input temperature and temperature change rate are determined based on the aforementioned deviation and actual physical experimental temperature data, specifically including: The new temperature data is obtained by adding the deviation to the actual physical experiment temperature data; The average value of the new temperature data is calculated to obtain the fused input temperature; The temperature change rate is obtained by differentiating the fused input temperature over time.

4. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 1, characterized in that, The construction of a temperature compensation model for fiber optic gyroscopes includes: An improved sand cat swarm optimization algorithm is used to optimize the BP neural network; the improved sand cat swarm optimization algorithm is to introduce chaotic perturbation and Gaussian mutation mechanism into the sand cat swarm optimization algorithm; The optimized BP neural network was trained and tested to obtain the fiber optic gyroscope temperature compensation model.

5. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 4, characterized in that, An improved sand cat swarm optimization algorithm is used to optimize the BP neural network, specifically including: The structure of the BP neural network is initialized to obtain the parameter combination; the parameter combination includes: the weight matrix from the input layer to the hidden layer, the bias vector of the hidden layer, the weight matrix from the hidden layer to the output layer, and the bias vector of the output layer; The original domain is obtained by initializing the location of the sand cat group based on the parameter combination, and the fitness values ​​of individuals and the average fitness value of all individuals in the original domain are determined. Individuals with fitness values ​​greater than or equal to the average fitness value are subjected to chaotic perturbation, and individuals with fitness values ​​less than the average fitness value are subjected to Gaussian mutation. Individuals are iteratively updated until the maximum number of iterations or the convergence condition is reached, at which point the optimal individual is obtained. The optimal individual is then used to update the position of individuals in the sand cat group to obtain the global optimal solution. The global optimal solution is used as the optimal parameter combination of the BP neural network to obtain the optimized BP neural network.

6. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 5, characterized in that, The new individuals obtained after applying chaotic perturbation to individuals whose fitness values ​​are greater than or equal to the average fitness value are represented as follows: P new =lb+f SPM (P s )·(ub-lb); Where [lb, ub] represents the original domain, and lb and ub are the upper and lower bounds of the original domain, f SPM (*) denotes the chaotic mapping function, P new P represents a new individual. s It refers to an individual.

7. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 5, characterized in that, The new individuals obtained by performing Gaussian mutation on individuals whose fitness values ​​are less than the average fitness value are represented as follows: Among them, P new Represents a new individual. This represents a random number generated using a Gaussian distribution. This represents the optimal individual.

8. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 1, characterized in that, After modeling and simulating the temperature of the fiber optic gyroscope to obtain the simulated temperature data, the following steps are also included: Quantitatively assess the consistency between simulated temperature data and actual physical experimental temperature data.

9. The temperature compensation method based on fiber optic gyroscope temperature field modeling according to claim 8, characterized in that, The consistency between simulated temperature data and actual physical experimental temperature data is quantitatively assessed using root mean square error, mean absolute error, Pearson correlation coefficient, consistency index, and regression analysis.

10. A temperature compensation system based on fiber optic gyroscope temperature field modeling, characterized in that, include: The temperature modeling and simulation module is used to model and simulate the temperature of the fiber optic gyroscope to obtain simulated temperature data. The deviation calculation module is used to calculate the deviation between the simulated temperature data and the simulated set temperature data; A module for determining the fusion input temperature and temperature change rate is used to determine the fusion input temperature and temperature change rate based on the deviation and actual physical experiment temperature data. The model building module is used to build a temperature compensation model for fiber optic gyroscopes. The temperature compensation value determination module is used to determine the temperature compensation value based on the fused input temperature and the temperature change rate, using a fiber optic gyroscope temperature compensation model. The temperature compensation module is used to compensate the temperature-dependent zero-bias measurement value of the fiber optic gyroscope under variable temperature conditions based on the temperature compensation value, so as to obtain the temperature-compensated zero-bias of the fiber optic gyroscope.

Images