Firearm barrel wear prediction method and device based on escape optimization

Abstract

This application provides a method and apparatus for predicting gun barrel wear based on escape optimization. By transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector, a sample set corresponding to the amount of ablation wear is constructed. Through designing the search space of the escape optimization algorithm, combined with cross-validation and fitness ranking mechanisms, global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable early warning mechanism for ablation wear status. This method effectively addresses the shortcomings of traditional prediction techniques in parameter optimization and status early warning, providing a scientific basis for the use and maintenance of gun barrels.

Description

Technical Field

[0001] This application relates to the field of data processing, specifically to a method and apparatus for predicting gun barrel wear based on escape optimization. Background Technology

[0002] Existing techniques for predicting artillery barrel wear have significant shortcomings. Traditional prediction methods often employ a single model or simple optimization algorithm, failing to fully utilize the characteristic information of multi-source firing data, resulting in insufficient prediction accuracy and robustness. Furthermore, existing methods are prone to getting trapped in local optima when dealing with high-dimensional nonlinear prediction problems, making it difficult to obtain the globally optimal solution.

[0003] Furthermore, existing technologies also face bottlenecks in parameter optimization and prediction result correction. Most methods lack effective parameter search mechanisms and cannot dynamically adjust optimization strategies based on sample characteristics, thus affecting model performance. The methods for quantifying and correcting the uncertainty of prediction results are relatively simple, lacking systematic error analysis and correction mechanisms.

[0004] Existing systems have technical shortcomings in comprehensive load assessment and condition early warning. The lack of unified modeling for multi-dimensional characteristics such as temperature, pressure, and propellant charge makes it difficult to accurately depict wear evolution under actual operating conditions, affecting the accuracy and timeliness of early warnings. Solving these problems is of great significance for improving the safety and maintenance efficiency of artillery. Summary of the Invention

[0005] To address the problems in the existing technology, this application provides a method and device for predicting artillery barrel wear based on escape optimization. This method can establish a reliable early warning mechanism for ablation wear conditions, effectively solving the shortcomings of traditional prediction technologies in parameter optimization and condition early warning, and providing a scientific basis for the use and maintenance of artillery barrels.

[0006] To solve at least one of the above problems, this application provides the following technical solution:

[0007] Firstly, this application provides a method for predicting gun barrel wear based on escape optimization, including:

[0008] Historical ablation and wear data of multiple artillery pieces at different firing stages were collected. Temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters were extracted from the historical ablation and wear data. The temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters were transformed into a comprehensive load feature vector. A sample set of the correspondence between the comprehensive load feature vector and the ablation and wear amount was constructed. A stratified sampling method was used to divide the sample set of the correspondence into a training sample subset and a validation sample subset.

[0009] The search space of the escape optimization algorithm is established based on the training sample subset. The initial position distribution of the search particle swarm is set, and the position information of the search particle swarm is mapped to the prediction model parameters. The prediction error corresponding to each particle is calculated using the cross-validation method. The prediction error is used as the particle fitness function. The particle swarm is sorted by fitness according to the fitness function. The particle with the best fitness is selected as the global optimal solution. The velocity and position of the particles are updated based on the escape mechanism in each iteration. The search stops when the number of iterations reaches a preset value. The finally converged global optimal solution is used as the optimal prediction model parameters.

[0010] The system reads real-time firing data of the artillery to be predicted, extracts temperature, pressure, and propellant characteristic parameters from the real-time firing data, generates a comprehensive load feature vector by converting the temperature, pressure, and propellant characteristic parameters according to the same transformation rule, inputs the comprehensive load feature vector into a prediction model with the parameters of the optimal prediction model, calculates the predicted ablation wear values ​​at different firing times, performs Bayesian correction on the predicted ablation wear values ​​to obtain the corrected prediction results, constructs an ablation wear degradation trend curve of the artillery barrel based on the corrected prediction results, extracts wear change rate and wear acceleration characteristics from the degradation trend curve, and provides early warning of the ablation wear state of the artillery barrel based on the wear change rate and wear acceleration characteristics.

[0011] Furthermore, the process involves collecting historical ablation and wear data from multiple artillery pieces at different firing stages, extracting temperature, pressure, and propellant characteristic parameters from the historical ablation and wear data, converting these parameters into a comprehensive load feature vector, and constructing a sample set of correspondences between the comprehensive load feature vector and the ablation and wear amount, including:

[0012] During the firing of multiple artillery pieces, barrel temperature data, bore pressure data, and propellant loading data are collected. The barrel temperature data is normalized according to the overload factor to obtain temperature characteristic parameters. The bore pressure data is normalized according to the peak pressure to obtain pressure characteristic parameters. The propellant loading data is normalized according to the quality standard to obtain propellant loading characteristic parameters. Principal component analysis is used to perform dimensionality reduction on the temperature characteristic parameters, the pressure characteristic parameters, and the propellant loading characteristic parameters to obtain a feature vector reflecting the comprehensive load of the firing process.

[0013] The feature vectors are sampled in time sequence and the cumulative ablation wear corresponding to each firing moment is calculated. A mapping relationship between the feature vector sequence and the ablation wear is established. The mapping relationship is divided into a training sample set and a validation sample set using a stratified random sampling method. The data in the training sample set and the validation sample set are normalized and standardized to construct a standardized feature vector and ablation wear correspondence sample set.

[0014] Furthermore, the step of using stratified sampling to divide the corresponding sample set into a training sample subset and a validation sample subset includes:

[0015] Based on the distribution characteristics of ablation wear in the corresponding sample set, the sample data is divided into multiple wear level layers. The proportion of sample data in each wear level layer is calculated. The sample data in the wear level layer is randomly shuffled. According to the preset ratio of training set to validation set, a corresponding number of sample data are extracted from each wear level layer. The extracted sample data are used to construct training sample subsets and validation sample subsets respectively.

[0016] The training sample subset and the validation sample subset are subjected to data balance test. The distribution difference of each wear level sample in the training set and the validation set is calculated. When the distribution difference exceeds a preset threshold, stratified sampling is performed again until the data distribution of the training sample subset and the validation sample subset meets the balance requirement.

[0017] Further, the step of establishing a search space for the escape optimization algorithm based on the subset of training samples, setting the initial position distribution of the search particle swarm, mapping the position information of the search particle swarm to prediction model parameters, calculating the prediction error corresponding to each particle using cross-validation, using the prediction error as the particle fitness function, ranking the particle swarm according to the fitness function, and selecting the particle with the best fitness as the global optimal solution includes:

[0018] The boundary range of the search space is determined based on the feature dimensions of the training sample subset. An initial particle swarm is randomly generated in the search space. Each dimensional component of the initial particle swarm is mapped to the weight parameters and bias parameters of the prediction model. The parameter matrix of the prediction model is constructed based on the weight parameters and the bias parameters. The training sample subset is divided into multiple training sets and test sets using the K-fold cross-validation method.

[0019] The training set is input into the prediction model for training. The prediction error of the model is calculated using the test set. The cross-validation process is repeated for each particle to obtain the average prediction error. The average prediction error is used as the fitness value of the particle. The particle swarm is sorted in ascending order based on the fitness value. The position of the particle with the smallest fitness value is selected as the global optimal solution for the current iteration.

[0020] Furthermore, the step of updating the particle's velocity and position based on the escape mechanism in each iteration, stopping the search when the number of iterations reaches a preset value, and using the finally converged global optimal solution as the optimal prediction model parameters includes:

[0021] Calculate the Euclidean distance between each particle and the global optimal solution. When the Euclidean distance is less than a preset escape threshold, trigger the escape mechanism. Add an escape perturbation term to the particle's current velocity direction. Update the particle's velocity vector based on the escape perturbation term and inertia weight. Calculate the particle's new position based on the velocity vector. Map the position components that exceed the search space boundary and convert the mapped position information into new prediction model parameters.

[0022] The fitness value of the particle swarm is recalculated and the global optimal solution is updated after each iteration. The historical change trajectory of the global optimal solution is recorded. When the change of the global optimal solution is less than the convergence threshold or the number of iterations reaches the preset upper limit, the search is stopped. The global optimal solution at the final convergence is used as the parameters of the optimized prediction model.

[0023] Further, the process of generating a comprehensive load feature vector by applying the same transformation rule to the temperature feature parameters, pressure feature parameters, and charge feature parameters, inputting the comprehensive load feature vector into a prediction model with the optimal prediction model parameters, calculating the ablation wear prediction values ​​at different firing times, and performing Bayesian correction on the ablation wear prediction values ​​to obtain the corrected prediction results includes:

[0024] The temperature, pressure, and propellant characteristics collected in real time are substituted into the principal component analysis model. The temperature, pressure, and propellant characteristics are then subjected to dimensionality reduction transformation based on the feature transformation matrix obtained during the training phase to generate a feature vector that reflects the comprehensive load of the current firing process. The feature vector is then input into the prediction model, and the ablation wear prediction value at the current moment is calculated based on the optimal prediction model parameters.

[0025] A Bayesian prior distribution model is constructed, using historical prediction data as prior information. The posterior probability distribution of the predicted ablation wear value is calculated. The prediction results are corrected based on the mean and variance of the posterior probability distribution. The corrected prediction results at different firing times are calculated sequentially using the sliding time window method.

[0026] Further, the step of constructing an ablation and wear degradation trend curve for the gun barrel based on the corrected prediction results, extracting wear rate characteristics and wear acceleration characteristics from the degradation trend curve, and providing early warning of the ablation and wear state of the gun barrel based on the wear rate characteristics and wear acceleration characteristics includes:

[0027] The corrected prediction result sequence is subjected to time series fitting processing. A degradation trend curve reflecting the evolution process of ablation wear is constructed using the polynomial regression method. The first derivative of the degradation trend curve at different times is calculated to obtain the wear change rate feature. The second derivative of the degradation trend curve at different times is calculated to obtain the wear acceleration feature. A quantitative evaluation index for the ablation wear state of the gun barrel is established based on the wear change rate feature and the wear acceleration feature.

[0028] Based on historical data statistical analysis, the wear change rate threshold and wear acceleration threshold are determined. When the wear change rate characteristic exceeds the wear change rate threshold or the wear acceleration characteristic exceeds the wear acceleration threshold, an early warning mechanism is triggered. The early warning level is determined based on the preset early warning level classification standard to assess the ablation and wear state of the artillery barrel.

[0029] Secondly, this application provides a gun barrel wear prediction device based on escape optimization, comprising:

[0030] The dataset construction module is used to collect historical ablation and wear data of multiple artillery pieces at different firing stages, extract temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters from the historical ablation and wear data, convert the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters into a comprehensive load feature vector, construct a correspondence sample set between the comprehensive load feature vector and the ablation and wear amount, and use a stratified sampling method to divide the correspondence sample set into a training sample subset and a validation sample subset;

[0031] The optimal parameter determination module is used to establish the search space of the escape optimization algorithm based on the training sample subset, set the initial position distribution of the search particle swarm, map the position information of the search particle swarm to the prediction model parameters, calculate the prediction error corresponding to each particle using the cross-validation method, use the prediction error as the particle fitness function, sort the particle swarm according to the fitness function, select the particle with the best fitness as the global optimal solution, update the velocity and position of the particles based on the escape mechanism in each iteration, stop the search when the number of iterations reaches a preset value, and use the finally converged global optimal solution as the optimal prediction model parameters.

[0032] The wear state prediction module is used to read real-time firing data of the artillery to be predicted, extract temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters from the real-time firing data, generate a comprehensive load feature vector by converting the temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters according to the same transformation rule, input the comprehensive load feature vector into a prediction model with the parameters of the optimal prediction model, calculate the predicted value of ablation wear at different firing times, perform Bayesian correction on the predicted value of ablation wear to obtain the corrected prediction result, construct an ablation wear degradation trend curve of the artillery barrel based on the corrected prediction result, extract wear change rate characteristics and wear acceleration characteristics from the degradation trend curve, and provide early warning of the ablation wear state of the artillery barrel based on the wear change rate characteristics and wear acceleration characteristics.

[0033] Thirdly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the escape optimization-based artillery barrel wear prediction method.

[0034] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the escape optimization-based gun barrel wear prediction method.

[0035] Fifthly, this application provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the escape optimization-based gun barrel wear prediction method.

[0036] As described above, this application provides a method and device for predicting gun barrel wear based on escape optimization. It constructs a sample set corresponding to the ablation wear amount by transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector. By designing the search space of the escape optimization algorithm and combining cross-validation and fitness ranking mechanisms, the global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable ablation wear state early warning mechanism. This method effectively solves the shortcomings of traditional prediction techniques in parameter optimization and state early warning, providing a scientific basis for the use and maintenance of gun barrels. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is one of the flowcharts illustrating the escape optimization-based artillery barrel wear prediction method in the embodiments of this application;

[0039] Figure 2 This is the second flowchart illustrating the escape optimization-based artillery barrel wear prediction method in the embodiments of this application.

[0040] Figure 3 This is the third flowchart illustrating the escape optimization-based artillery barrel wear prediction method in the embodiments of this application.

[0041] Figure 4 This is the fourth flowchart illustrating the escape optimization-based artillery barrel wear prediction method in the embodiments of this application.

[0042] Figure 5 This is the fifth flowchart illustrating the escape optimization-based artillery barrel wear prediction method in the embodiments of this application.

[0043] Figure 6 This is the sixth flowchart illustrating the escape optimization-based artillery barrel wear prediction method in the embodiments of this application.

[0044] Figure 7 This is the seventh flowchart illustrating the escape optimization-based artillery barrel wear prediction method in the embodiments of this application.

[0045] Figure 8 This is a structural diagram of the artillery barrel wear prediction device based on escape optimization in the embodiments of this application;

[0046] Figure 9 This is a schematic diagram of the structure of the electronic device in the embodiments of this application.

[0047] Figure label:

[0048] Electronic device 9600, central processing unit 9100, memory 9140, communication module 9110, input unit 9120, audio processor 9130, display 9160, power supply 9170, buffer memory 9141, application / function storage unit 9142, data storage unit 9143, driver storage unit 9144, antenna 9111, speaker 9131, microphone 9132. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, 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, 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.

[0050] The acquisition, storage, use, and processing of data in this application all comply with the relevant provisions of national laws and regulations.

[0051] To address the problems in existing technologies, this application provides a method and apparatus for predicting gun barrel wear based on escape optimization. By transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector, a sample set corresponding to the amount of ablation wear is constructed. The search space of the escape optimization algorithm is designed, and combined with cross-validation and fitness ranking mechanisms, global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable early warning mechanism for ablation wear status. This method effectively solves the shortcomings of traditional prediction techniques in parameter optimization and status early warning, providing a scientific basis for the use and maintenance of gun barrels.

[0052] To establish a reliable early warning mechanism for ablation wear conditions and effectively address the shortcomings of traditional prediction techniques in parameter optimization and condition early warning, thus providing a scientific basis for the use and maintenance of artillery barrels, this application provides an embodiment of an escape optimization-based method for predicting artillery barrel wear. See [link to relevant documentation]. Figure 1 The escape optimization-based method for predicting gun barrel wear specifically includes the following:

[0053] Step S101: Collect historical ablation and wear data of multiple artillery pieces at different firing stages, extract temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters from the historical ablation and wear data, convert the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters into a comprehensive load feature vector, construct a correspondence sample set between the comprehensive load feature vector and the ablation and wear amount, and use a stratified sampling method to divide the correspondence sample set into a training sample subset and a validation sample subset;

[0054] Optionally, this embodiment employs a distributed data acquisition system to monitor the artillery firing process from all angles. Barrel temperature acquisition utilizes a K-type thermocouple array, with 16 measuring points arranged axially and 4 layers of measuring points radially. The thermocouples have a measurement range of -50℃ to 1100℃, and the sampling frequency is set to 100Hz to ensure accurate capture of transient temperature changes. The installation position of each measuring point has been optimized through thermal field analysis to achieve precise monitoring of the temperature field in key areas.

[0055] This embodiment employs a piezoelectric sensor for bore pressure measurement, installed at two key locations: the bore bottom and the middle. The sensor's range covers 0-600 MPa, and the sampling frequency is set to 200 kHz, enabling high-precision acquisition of pressure fluctuations. The signal conditioning circuit utilizes a charge amplifier configuration, exhibiting excellent linearity and temperature stability.

[0056] This embodiment employs a high-precision weighing system to collect propellant loading data, achieving a weighing accuracy of 0.1g. The influence of ambient temperature and humidity is considered during the weighing process, and a compensation algorithm is used to improve measurement reliability. Propellant loading data acquisition also includes propellant temperature measurement to analyze the impact of temperature on propellant performance.

[0057] This embodiment presents an innovative method for extracting temperature features. Temperature normalization employs an overload factor F_t = (T - T_ref) / (T_max - T_ref), where T is the measured temperature, T_ref is the ambient reference temperature, and T_max is the maximum allowable temperature of the material. This normalization method takes into account material properties, making the features more physically meaningful.

[0058] This embodiment implements a systematic extraction of pressure features. Pressure features are processed using peak normalization: F_p = P / P_max, where P is the measured pressure and P_max is the design maximum pressure. The feature extraction process considers pressure waveform characteristics, including multiple dimensions such as rise time, duration, and pressure gradient.

[0059] This embodiment innovatively implements charge feature extraction. Charge feature normalization adopts the quality standard method: F_m = M / M_s, where M is the actual charge amount and M_s is the standard charge amount. Feature extraction considers the influence of charge density and filling coefficient, achieving a comprehensive characterization of the charge operation conditions.

[0060] This embodiment employs an improved principal component analysis method for feature fusion. Feature weights are adaptively determined using variance contribution rates, retaining principal components with a cumulative contribution rate of 95%. The feature transformation matrix W is obtained through eigenvector orthogonalization to ensure the independence of the transformed features.

[0061] This embodiment designs a time-series sampling strategy for the system. The sampling interval is determined based on the characteristic time scale of the ablation process to ensure the capture of key state changes. The cumulative ablation amount is calculated considering the material removal rate and is achieved through an integral method.

[0062] This embodiment implements an innovative stratified sampling scheme. First, the stratification boundaries are determined based on the statistical distribution of ablation amount, and the K-means clustering method is used to divide the samples into multiple wear level layers. Random sampling is used within each layer, and the sampling ratio is proportional to the sample size of that layer.

[0063] This embodiment employs a data balance evaluation method. It calculates the distribution difference of samples from each layer in the training and validation sets: D = Σ|P_train,i - P_valid,i|, where P_train,i and P_valid,i are the proportions of samples from the i-th layer in the training and validation sets, respectively. A resampling mechanism is triggered when D exceeds a preset threshold.

[0064] This embodiment establishes a reliable sample set construction framework through systematic data acquisition and processing. While ensuring data quality, this scheme provides comprehensive feature representation through multi-level analysis. The overall design fully considers the characteristics of artillery ablation, achieving end-to-end optimization from data acquisition to sample construction.

[0065] The sample set in this embodiment exhibits excellent representativeness and balance, achieving accurate characterization of the ablation process through multi-level analysis. This scheme demonstrates powerful data processing capabilities in practical applications, providing a reliable data foundation for subsequent model training. The balanced distribution of samples provides comprehensive training information for model learning.

[0066] The data processing framework in this embodiment has good adaptability and scalability, and can adapt to the data acquisition needs under different working conditions. Through feature adaptation and sample equalization, it achieves accurate characterization of complex ablation processes and provides reliable data support for model training.

[0067] Step S102: Establish the search space of the escape optimization algorithm based on the training sample subset, set the initial position distribution of the search particle swarm, map the position information of the search particle swarm to the prediction model parameters, calculate the prediction error corresponding to each particle using the cross-validation method, use the prediction error as the particle fitness function, sort the particle swarm by fitness according to the fitness function, select the particle with the best fitness as the global optimal solution, update the velocity and position of the particles based on the escape mechanism in each iteration, stop the search when the number of iterations reaches a preset value, and use the finally converged global optimal solution as the optimal prediction model parameters;

[0068] Optionally, this embodiment constructs a multi-dimensional search space based on the feature dimensions of the training samples. The value range of each dimension is determined through feature analysis: Xi_min = μi - 3σi, Xi_max = μi + 3σi, where μi and σi are the mean and standard deviation of the i-th feature, respectively. This boundary setting method ensures that 99.7% of the feature distribution is covered, while avoiding an excessively large search space.

[0069] This embodiment employs an improved initialization strategy to generate the search particle swarm. Particle positions are determined using the Latin hypercube sampling method to ensure that initial particles are uniformly distributed in the search space. The velocity of each particle is initialized using a Gaussian distribution: Vi ~ N(0, 0.1×|Xi_max - Xi_min|), where Vi is the i-th velocity component.

[0070] This embodiment innovatively implements a parameter mapping mechanism. The mapping from position vectors to model parameters uses a piecewise function: W = sigmoid(X) for weight parameters and b = tanh(X) for bias parameters. This mapping method ensures the rationality of the parameter value range while maintaining the continuity of the parameter space.

[0071] This embodiment designs a complete cross-validation framework. A 5-fold cross-validation method is used, randomly dividing the training samples into 5 subsets. During validation, one subset is selected sequentially as the test set, and the rest are used as the training set. The prediction error is calculated using a weighted mean squared error: E = Σwi(yi - ŷi)² / N, where wi is the sample weight, reflecting the importance of different wear levels.

[0072] This embodiment implements an innovative fitness evaluation scheme. The fitness function combines prediction accuracy and model complexity: F = E + λ×R, where E is the prediction error, R is the parameter regularization term, and λ is the balance factor. This design avoids overfitting and improves the model's generalization ability.

[0073] This embodiment employs an improved escape mechanism update strategy. The distance between a particle and the global optimum is calculated using cosine similarity: cos_θ = X·Xg / (|X|·|Xg|), where X is the particle position and Xg is the global optimum. The escape mechanism is triggered when cos_θ exceeds a preset threshold.

[0074] This embodiment innovatively incorporates an escape perturbation term. The perturbation direction is orthogonal to the current velocity direction: D = V×n, where V is the velocity vector and n is a unit orthogonal vector. The perturbation magnitude is related to the particle's current fitness: |D| = α×exp(-β×F), where F is the fitness value, and α and β are control parameters.

[0075] This embodiment implements an adaptive velocity update strategy. The velocity update formula is: V = w×V + c1×r1×(PX)+ c2×r2×(GX) + D, where w is the inertia weight, which decays with the number of iterations; c1 and c2 are acceleration coefficients; r1 and r2 are random numbers; P is the individual optimal solution; G is the global optimal solution; and D is the escape perturbation term.

[0076] This embodiment employs a boundary handling mechanism to ensure search stability. When a particle's position exceeds the search space, a bounce strategy is used: Xi = Xi_min + r×(Xi_max - Xi_min), where r is a random number between (0,1). This approach maintains the diversity of the particle swarm.

[0077] This embodiment employs multiple convergence criteria. In addition to limiting the number of iterations, it also monitors the rate of change of the global optimum: ΔG = |Gt - Gt-1| / |Gt-1|, where Gt is the global optimum of the current iteration. Convergence is determined when ΔG is less than a threshold for multiple consecutive iterations.

[0078] This embodiment achieves an efficient parameter optimization process through an innovative optimization strategy. The scheme effectively avoids local optima through an escape mechanism, improving global search capabilities. The overall design fully considers the characteristics of the prediction model, enabling efficient exploration of the parameter space.

[0079] The optimization results of this embodiment demonstrate excellent convergence and stability, achieving accurate parameter selection for the prediction model through multi-level optimization. This scheme exhibits strong optimization capabilities in practical applications, providing a reliable guarantee for model performance.

[0080] The optimization framework in this embodiment exhibits good versatility and robustness, enabling it to adapt to parameter optimization needs in different feature spaces. Through escape mechanisms and adaptive strategies, it achieves efficient search of complex parameter spaces, providing reliable parameter support for prediction models.

[0081] Step S103: Read the real-time firing data of the artillery to be predicted, extract temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters from the real-time firing data, generate a comprehensive load feature vector by converting the temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters according to the same transformation rule, input the comprehensive load feature vector into the prediction model with the optimal prediction model parameters, calculate the ablation wear prediction value at different firing times, perform Bayesian correction on the ablation wear prediction value to obtain the corrected prediction result, construct the ablation wear degradation trend curve of the artillery barrel based on the corrected prediction result, extract the wear change rate feature and wear acceleration feature from the degradation trend curve, and provide early warning of the ablation wear state of the artillery barrel based on the wear change rate feature and the wear acceleration feature.

[0082] Optionally, this embodiment employs a real-time data acquisition system to dynamically monitor the artillery firing process. Temperature acquisition utilizes a high-response thermocouple array, with 16 measuring points evenly distributed along the barrel axis, and each measuring point having four layers of sensors arranged radially. The sampling frequency is set to 100Hz to ensure the capture of transient temperature changes. Signal conditioning employs cold junction compensation and digital filtering to improve measurement accuracy.

[0083] This embodiment achieves real-time acquisition of pressure data. A piezoelectric sensor is used to monitor pressure at key locations in the gun barrel, with a sampling frequency of 200kHz, enabling precise capture of the pressure waveform. Signal processing employs an adaptive threshold detection method to identify firing events, ensuring accurate data segmentation.

[0084] This embodiment designs a real-time monitoring scheme for charge parameters. A high-precision weighing system is used to measure the charge amount, while simultaneously monitoring the charge temperature and environmental parameters. Automatic compensation is implemented during data acquisition to eliminate the influence of environmental factors such as temperature and humidity.

[0085] This embodiment innovatively implements a feature extraction framework. Temperature features are obtained through thermal field reconstruction: T(r,z) = ΣΣaij·φij(r,z), where φij is the basis function and aij is the expansion coefficient. Pressure features are extracted through waveform analysis to extract key parameters, including peak value, rise time, and waveform integral. Charge features are combined with charge density and uniformity indices.

[0086] This embodiment employs a consistency transformation strategy to generate feature vectors. The feature transformation uses the principal component analysis matrix obtained during the training phase: F = W·X, where X represents the original features and W is the transformation matrix. This ensures that the features in the prediction phase remain consistent with those in the training phase.

[0087] This embodiment implements an innovative predictive computation framework. The predictive model adopts a recursive structure: y(t) = f(F(t),y(t-1); θ), where F(t) is the current feature vector, y(t-1) is the predicted value at the previous time step, and θ is the optimized model parameters. This design takes into account the cumulative effect of the ablation process.

[0088] This embodiment employs a Bayesian correction mechanism. The prior distribution is established based on historical data: p(θ) ~ N(μ0, Σ0), where μ0 and Σ0 are the mean and covariance of the parameters, respectively. The posterior distribution is calculated through Bayesian updates: p(θ|D) ∝ p(D|θ)·p(θ), where D represents the observed data.

[0089] This embodiment innovatively implements prediction result correction. The correction process employs a sliding window strategy, with the window length determined through autocorrelation analysis. Each predicted value is calculated using the posterior distribution expectation: ŷ = E[y|D,θ], simultaneously obtaining an estimate of the prediction uncertainty.

[0090] This embodiment uses spline interpolation to construct the degradation trend curve. The curve expression is: S(t) = Σci·Bi,k(t), where Bi,k is the k-th order B-spline basis function, and ci is the control point parameter. The interpolation nodes are adaptively arranged through curvature analysis to ensure that key features are preserved.

[0091] This embodiment implements a feature extraction strategy. The rate of change feature is calculated through numerical differentiation: v(t) = dS / dt, and the central difference scheme is used to improve accuracy. The acceleration feature is obtained through the second derivative: a(t) = d²S / dt², and smoothing is combined to reduce the influence of noise.

[0092] This embodiment designs a multi-level early warning mechanism. The early warning index comprehensively considers the rate of change and acceleration characteristics: W = w1·v(t) + w2·a(t), where w1 is the weighting coefficient. The early warning threshold is determined through statistical analysis of historical data, taking into account the characteristics of different operating conditions.

[0093] This embodiment achieves reliable condition assessment through innovative predictive analysis. The scheme provides accurate characterization of the ablation process through multi-level analysis. The overall design fully considers the characteristics of artillery use, achieving end-to-end optimization from data acquisition to condition early warning.

[0094] The prediction results of this embodiment demonstrate excellent accuracy and reliability, achieving precise prediction of the ablation process through multi-level analysis. This scheme exhibits strong predictive capabilities in practical applications, providing a reliable quantitative basis for equipment maintenance decisions.

[0095] The prediction framework in this embodiment has good adaptability and scalability, and can adapt to prediction needs under different operating conditions. Through feature adaptation and Bayesian correction, it achieves accurate prediction of complex ablation processes, providing reliable decision support for equipment management.

[0096] As described above, the escape optimization-based gun barrel wear prediction method provided in this application can construct a sample set corresponding to the ablation wear amount by transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector. By designing the search space of the escape optimization algorithm and combining cross-validation and fitness ranking mechanisms, global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable ablation wear state early warning mechanism. This method effectively addresses the shortcomings of traditional prediction techniques in parameter optimization and state early warning, providing a scientific basis for the use and maintenance of gun barrels.

[0097] In one embodiment of the escape optimization-based gun barrel wear prediction method of this application, see [link to relevant documentation]. Figure 2 It can also specifically include the following:

[0098] Step S201: During the firing of multiple artillery pieces, barrel temperature data, bore pressure data, and propellant loading data are collected. The barrel temperature data is normalized according to the overload factor to obtain temperature characteristic parameters. The bore pressure data is normalized according to the peak pressure to obtain pressure characteristic parameters. The propellant loading data is normalized according to the quality standard to obtain propellant loading characteristic parameters. Principal component analysis is used to perform dimensionality reduction on the temperature characteristic parameters, the pressure characteristic parameters, and the propellant loading characteristic parameters to obtain a feature vector reflecting the comprehensive load of the firing process.

[0099] Step S202: Perform time-series sampling on the feature vector and calculate the cumulative ablation wear corresponding to each firing moment. Establish the mapping relationship between the feature vector sequence and the ablation wear. Use a stratified random sampling method to divide the mapping relationship into a training sample set and a validation sample set. Normalize and standardize the data in the training sample set and the validation sample set to construct a standardized feature vector and ablation wear correspondence sample set.

[0100] Optionally, for monitoring the ablation and wear of the artillery barrel, the embodiment deploys a multi-dimensional sensing system. Temperature monitoring employs a high-precision thermocouple array with 16 measuring points arranged axially, and 4 sensors radially positioned at each measuring point, forming a 3D temperature field monitoring network. The sampling frequency is set to 100Hz to ensure accurate capture of transient temperature changes. Signal conditioning utilizes cold junction compensation and adaptive filtering to effectively suppress interference caused by firing vibrations.

[0101] In this embodiment, piezoelectric sensors are used for pressure data acquisition, deployed at key locations in the gun barrel. The sensor range covers 0-600 MPa, with a sampling frequency of 200 kHz. The signal conditioning circuit employs a charge amplifier configuration to ensure measurement stability under high temperature and high pressure conditions. A high-precision weighing system is used for propellant data acquisition, while simultaneously monitoring environmental parameters such as propellant temperature.

[0102] This embodiment innovatively implements feature normalization. Temperature feature normalization uses an overload factor: F_t = (T - T_ref) / (T_max - T_ref), where T is the measured temperature, T_ref is the ambient temperature, and T_max is the material's withstand temperature. This method correlates temperature features with material properties, enhancing the physical meaning of the features.

[0103] This embodiment designs a pressure characteristic normalization scheme. The pressure characteristic is normalized by peak value: F_p = P / P_max, where P is the measured pressure and P_max is the design maximum pressure. Simultaneously, pressure waveform characteristics, including rise time, duration, and pressure gradient, are considered to form a multi-dimensional pressure characteristic vector.

[0104] This embodiment implements the normalization of charge characteristics. The charge characteristics are standardized using the following quality standard: F_m = M / M_s, where M is the actual charge amount and M_s is the standard charge amount. Simultaneously, filling density and uniformity indices are considered to construct a complete charge characteristic system.

[0105] This embodiment employs an improved principal component analysis method for feature fusion. First, the feature matrix is ​​standardized to eliminate the influence of dimensions. The feature covariance matrix is ​​then decomposed using SVD to obtain eigenvalues ​​and eigenvectors. Principal components are selected based on cumulative contribution rate, retaining those with a contribution rate of 95%.

[0106] This embodiment innovatively designs a time-series sampling strategy. The sampling interval is determined based on the characteristic time scale of the ablation process, ensuring the capture of key state changes. The calculation of cumulative ablation amount takes into account the nonlinear characteristics of the material removal rate and is achieved through an improved integration method.

[0107] This embodiment implements an innovative stratified sampling scheme. First, based on the statistical distribution characteristics of ablation, K-means clustering is used to divide the samples into multiple wear level layers. Random sampling is used within each layer, with the sampling ratio proportional to the sample size of that layer, ensuring the representativeness of the samples.

[0108] This embodiment designs a data standardization framework. Standardization uses the z-score method: Z = (X-μ) / σ, where X is the original data, μ is the mean, and σ is the standard deviation. This process eliminates scale differences between different features and improves the stability of subsequent modeling.

[0109] This embodiment implements a sample set construction strategy. The corresponding sample set is organized in matrix form, with rows representing samples at different times and columns representing features and labels. The data structure design takes into account temporal correlations, facilitating subsequent model training.

[0110] This embodiment establishes a reliable sample construction framework through systematic data processing. While ensuring data quality, this scheme provides comprehensive feature representation through multi-level analysis. The overall design fully considers the characteristics of artillery ablation, achieving end-to-end optimization from data acquisition to sample construction.

[0111] The data processing results in this embodiment demonstrate excellent representativeness and balance, achieving accurate characterization of the ablation process through multi-level analysis. This scheme exhibits powerful data processing capabilities in practical applications, providing a reliable data foundation for subsequent model training.

[0112] The processing framework in this embodiment has good adaptability and scalability, and can adapt to the data acquisition needs under different working conditions. Through feature adaptation and sample equalization, it achieves accurate characterization of complex ablation processes and provides reliable data support for model training.

[0113] This embodiment innovatively solves key problems such as data imbalance and feature redundancy, significantly improving sample quality. Through stratified sampling and feature dimensionality reduction, high-quality training samples are constructed, laying the foundation for improving the performance of the prediction model.

[0114] In one embodiment of the escape optimization-based gun barrel wear prediction method of this application, see [link to relevant documentation]. Figure 3 It can also specifically include the following:

[0115] Step S301: Based on the distribution characteristics of ablation wear in the corresponding sample set, the sample data is divided into multiple wear level layers. The proportion of sample data in each wear level layer is calculated. The sample data in the wear level layer is randomly shuffled. According to the preset ratio of training set to validation set, a corresponding number of sample data are extracted from each wear level layer. The extracted sample data are used to construct training sample subsets and validation sample subsets respectively.

[0116] Step S302: Perform a data balance test on the training sample subset and the validation sample subset, calculate the distribution difference of each wear level sample in the training set and the validation set, and when the distribution difference exceeds a preset threshold, perform stratified sampling again until the data distribution of the training sample subset and the validation sample subset meets the balance requirements.

[0117] Optionally, this embodiment first analyzes the distribution characteristics of ablation wear and then uses a data clustering method to divide the samples into multiple wear level layers. Specifically, an improved K-means clustering algorithm is used, and the optimal number of level layers is determined by evaluating the silhouette coefficient and inter-cluster distance. In actual artillery applications, the wear state is usually divided into three main levels: slight wear, moderate wear, and severe wear. Each main level is further subdivided into several sub-layers based on subtle differences in the degree of wear.

[0118] This embodiment performs statistical analysis on the sample data of each wear level layer, calculating the proportion of each layer's samples in the total population. This analysis fully considers the frequency characteristics of different wear states during artillery use, reflecting the distribution pattern of wear states in actual application scenarios. For artillery with high usage frequency, slightly worn samples usually account for a larger proportion, while severely worn samples are relatively fewer.

[0119] This embodiment employs an improved Fisher-Yates algorithm in the sample shuffling process, ensuring randomness while maintaining the temporal correlation characteristics of the data. The shuffling process considers the correlation between adjacent shots, avoiding the dispersion of strongly correlated samples across different datasets. This approach ensures that the model can learn the continuous characteristics of wear evolution.

[0120] This embodiment employs a stratified proportional sampling strategy during sample extraction. First, the overall ratio of the training set to the validation set is determined, typically set to 7:3 or 8:2. Then, sampling is performed at the same ratio within each wear level stratum, ensuring that the proportion of samples from each level stratum remains consistent across different datasets. This sampling method avoids over-representation or under-representation of certain wear states in a specific dataset.

[0121] This embodiment innovatively implements a sample equalization processing mechanism. For wear level layers with a small number of samples, a data augmentation method based on a physical model is used to generate synthetic samples. The augmentation method takes into account the physical constraints of firing conditions, ensuring that the generated samples conform to the actual wear evolution law. This processing effectively improves the unevenness of sample distribution.

[0122] This embodiment designs a complete data balance verification scheme. The consistency of the distributions of the two datasets is evaluated by calculating the JS divergence of the sample distributions for each wear level in the training and validation sets. When the distribution difference exceeds a preset threshold, a resampling mechanism is triggered. This dynamic adjustment ensures the distribution similarity between different datasets.

[0123] This embodiment employs an adaptive adjustment strategy during the resampling process. Based on the distribution differences from the previous sampling, the sampling weights of each level are dynamically adjusted. Levels with significant distribution differences have their sampling weights increased, thereby improving their selection probability during the resampling process. This method accelerates the convergence speed of distribution equilibrium.

[0124] This embodiment implements an iteratively optimized sampling framework. After each round of resampling, the representativeness and diversity of the samples are evaluated simultaneously. Representativeness is assessed by intra-stratum sample variance, and diversity is assessed by inter-stratum sample distance. This multi-dimensional evaluation ensures the quality of the sampling results.

[0125] The stratified sampling scheme in this embodiment demonstrates excellent performance in practical applications. By considering the distribution characteristics and evolution patterns of wear states, it achieves the construction of high-quality training data. This scheme effectively solves the problems of sample imbalance and insufficient representativeness in traditional random sampling.

[0126] The data balance check mechanism in this embodiment ensures the reliability of model training. Through strict distribution consistency control, it avoids the negative impact of differences in the distribution of the training and validation sets on model performance. This mechanism provides an important guarantee for the model's generalization ability.

[0127] This embodiment significantly improves sample quality through innovative sampling strategies and balance control. While ensuring data representativeness, the scheme achieves a balanced distribution of samples across different wear states, laying a solid foundation for subsequent model training. This balanced sample distribution enables the model to fully learn the characteristics of different wear states, improving the accuracy and reliability of predictions.

[0128] In one embodiment of the escape optimization-based gun barrel wear prediction method of this application, see [link to relevant documentation]. Figure 4 It can also specifically include the following:

[0129] Step S401: Determine the boundary range of the search space based on the feature dimensions of the training sample subset, randomly generate an initial particle swarm in the search space, map each dimension component of the initial particle swarm to the weight parameters and bias parameters of the prediction model, construct the parameter matrix of the prediction model based on the weight parameters and the bias parameters, and divide the training sample subset into multiple training sets and test sets using the K-fold cross-validation method.

[0130] Step S402: Input the training set into the prediction model for training, calculate the prediction error of the model using the test set, repeat the cross-validation process for each particle to obtain the average prediction error, use the average prediction error as the fitness value of the particle, sort the particle swarm in ascending order based on the fitness value, and select the position of the particle with the smallest fitness value as the global optimal solution for the current iteration.

[0131] Optionally, this embodiment determines the search space boundary based on the feature distribution of the training samples. For each feature dimension, its range of variation is determined through statistical analysis, and an appropriate margin is left when setting the search boundary to avoid effective solutions being truncated by the boundary. This boundary setting method fully considers the physical characteristics of the artillery ablation process, ensuring that the search space covers all possible effective solutions.

[0132] This embodiment employs an improved Latin hypercube sampling method to generate the initial particle swarm. This method divides the search space into multiple subspaces, ensuring that the initial particles are uniformly distributed throughout the search space. The number of particles is dynamically determined based on the problem dimension, guaranteeing search space coverage while avoiding excessive computational burden.

[0133] This embodiment innovatively designs a parameter mapping mechanism. The mapping from particle positions to model parameters employs a non-linear transformation. A sigmoid function is used to map the weight parameters, ensuring their distribution remains within a reasonable range. A tanh function is used to map the bias parameters. This design considers the unique roles of different parameters in the prediction model.

[0134] This embodiment implements the dynamic construction of the parameter matrix. Based on the network structure of the prediction model, the mapped parameters are organized into a matrix according to hierarchical relationships. The matrix structure design takes into account the correlation between parameters, facilitating subsequent model calculations and gradient propagation.

[0135] This embodiment employs an improved K-fold cross-validation method. First, the training sample subset is stratified to ensure similar distributions of samples with different wear states within each fold. During validation, a rotation method is used, giving each sample a chance to serve as test data, thus improving the reliability of the validation results.

[0136] This embodiment innovatively implements a model training strategy. In each training iteration, a batch training method is employed, with the batch size dynamically adjusted based on the number of samples. An adaptive learning rate adjustment mechanism is used during training to ensure the stability of model convergence.

[0137] This embodiment designs a complete prediction error evaluation framework. The error calculation considers the importance weights of different wear states, giving higher weight to the prediction error of severe wear states. This weighting method is consistent with the degree of attention paid to different wear states in practical applications.

[0138] This embodiment implements an innovative error aggregation method. In the aggregation of multi-fold validation results, the stability of the prediction error at each fold is considered. Validation results with large fluctuations are filtered using analysis of variance, improving the representativeness of the average prediction error.

[0139] This embodiment employs an improved fitness evaluation scheme. Fitness calculation not only considers prediction error but also introduces a model complexity penalty term to avoid overfitting. This design improves the model's generalization ability while maintaining prediction accuracy.

[0140] This embodiment implements an efficient particle sorting mechanism. The sorting process employs an improved quicksort algorithm, which performs a detailed comparison of particles with similar fitness. The selection of the global optimal solution considers the stability of the solution and avoids interference from local noise.

[0141] This embodiment achieves efficient model parameter search through an innovative parameter optimization strategy. While ensuring search space coverage, the scheme provides reliable parameter evaluation through multi-level validation. The overall design fully considers the characteristics of the prediction model, enabling efficient exploration of the parameter space.

[0142] The optimization results of this embodiment demonstrate excellent performance and stability. Through multi-level verification and evaluation, the selected parameters are ensured to have good generalization ability. This scheme exhibits strong parameter optimization capabilities in practical applications, providing reliable parameter support for the prediction model.

[0143] The optimization framework in this embodiment exhibits good versatility and robustness, enabling it to adapt to parameter optimization needs in different feature spaces. Through innovative parameter mapping and evaluation mechanisms, it achieves efficient searching of complex parameter spaces, providing reliable performance assurance for the prediction model.

[0144] In one embodiment of the escape optimization-based gun barrel wear prediction method of this application, see [link to relevant documentation]. Figure 5 It can also specifically include the following:

[0145] Step S501: Calculate the Euclidean distance between each particle and the global optimal solution. When the Euclidean distance is less than a preset escape threshold, trigger the escape mechanism, add an escape perturbation term in the current velocity direction of the particle, update the particle's velocity vector according to the escape perturbation term and the inertial weight, calculate the particle's new position based on the velocity vector, perform mapping processing on the position components that exceed the search space boundary, and convert the mapped position information into new prediction model parameters.

[0146] Step S502: Recalculate the fitness value of the particle swarm after each iteration and update the global optimal solution. Record the historical change trajectory of the global optimal solution. Stop the search when the change of the global optimal solution in multiple consecutive iterations is less than the convergence threshold or the number of iterations reaches the preset upper limit. Use the global optimal solution at the final convergence as the optimized prediction model parameters.

[0147] Optionally, this embodiment uses Euclidean distance calculation to determine particle escape. For each particle, its distance to the current global optimal solution in the feature space is calculated. This distance calculation considers the weights of features in each dimension, reflecting the actual distribution of particles in the search space. The distance threshold is dynamically adjusted based on the scale of the search space to ensure the reasonable triggering of the escape mechanism.

[0148] This embodiment innovatively designs an escape perturbation mechanism. When a particle gets trapped in a local optimum, it escapes by superimposing a perturbation in the velocity direction. The perturbation amplitude is related to the particle's current fitness; the better the fitness, the smaller the perturbation. This design ensures a fine search for potential high-quality solutions.

[0149] This embodiment implements an adaptive speed update strategy. The inertia weight is dynamically adjusted during the iteration process, initially maintaining a large value to promote global search, and then decreasing it in the later stages to enhance local fine-grained search. The speed update considers both individual and group experience, enabling intelligent adjustment of the search direction.

[0150] This embodiment employs an innovative boundary handling method. When a particle's position exceeds the search space, instead of simple truncation or bounce, a periodic mapping approach is used. This method maintains the continuity of the search while avoiding the drastic interference of boundary handling on the search trajectory.

[0151] This embodiment incorporates a smoothing mechanism for parameter transformation. The transformation from location information to model parameters employs a continuously differentiable mapping function, ensuring the smoothness of the parameter space. This design facilitates the transfer of gradient information and improves the efficiency of parameter optimization.

[0152] This embodiment implements an innovative fitness evaluation framework. The fitness calculation after each iteration not only considers the current prediction error but also incorporates a smoothing factor based on historical performance, avoiding drastic fluctuations in the evaluation results. This stable evaluation mechanism contributes to the convergence of the search process.

[0153] This embodiment employs a multi-level optimal solution update strategy. The update of the global optimal solution considers the stability of the solution, avoiding being misled by temporary high-quality solutions through continuous multi-round verification. This strategy ensures the stability and reliability of the optimization process.

[0154] This embodiment innovatively designs a convergence criterion. By monitoring the changing trend of the global optimal solution, it combines the change amount and the rate of change as dual indicators to determine the convergence state. This design ensures the quality of the optimization results while avoiding the computational overhead caused by excessive iteration.

[0155] This embodiment implements dynamic monitoring of the optimization process. By recording the historical trajectory of the global optimal solution, the convergence characteristics of the optimization process are analyzed. This information can be used for dynamic adjustment of the optimization strategy, improving search efficiency.

[0156] This embodiment employs an improved termination mechanism. In addition to the basic iteration limit, a stability check is introduced. When the optimal solution changes very little over multiple iterations, a stable solution is considered to have been found, and the search is terminated promptly to avoid invalid computation.

[0157] This embodiment achieves an efficient parameter search process through an innovative optimization strategy. The scheme effectively avoids local optima through an escape mechanism, improving global search capabilities. The overall design fully considers the characteristics of the prediction model, enabling efficient exploration of the parameter space.

[0158] The optimization results of this embodiment demonstrate excellent convergence and stability. Accurate parameter selection for the prediction model was achieved through multi-level optimization. This scheme exhibits strong optimization capabilities in practical applications, providing a reliable guarantee for model performance.

[0159] The optimization framework in this embodiment exhibits good versatility and robustness, capable of adapting to parameter optimization needs in different feature spaces. Through escape mechanisms and adaptive strategies, it achieves efficient search of complex parameter spaces, providing reliable parameter support for the prediction model. This scheme significantly improves the model's performance in predicting artillery ablation and wear.

[0160] In one embodiment of the escape optimization-based gun barrel wear prediction method of this application, see [link to relevant documentation]. Figure 6 It can also specifically include the following:

[0161] Step S601: Substitute the real-time collected temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters into the principal component analysis model. Based on the feature transformation matrix obtained during the training phase, perform dimensionality reduction transformation on the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters to generate a feature vector reflecting the comprehensive load of the current firing process. Input the feature vector into the prediction model and calculate the ablation wear prediction value at the current moment based on the optimal prediction model parameters.

[0162] Step S602: Construct a Bayesian prior distribution model, using historical prediction data as prior information, calculate the posterior probability distribution of the ablation wear prediction value, correct the prediction result based on the mean and variance of the posterior probability distribution, and use the sliding time window method to calculate the corrected prediction result at different firing times in sequence.

[0163] Optionally, this embodiment employs an efficient feature extraction scheme in real-time data processing. Temperature feature parameters are acquired through a high-precision thermocouple array to capture the temperature distribution at each measuring point on the gun barrel in real time. Real-time filtering and compensation are performed during data acquisition to ensure the accuracy of the feature parameters. For transient temperature changes during artillery firing, an adaptive sampling strategy is adopted, increasing the sampling frequency at critical moments.

[0164] This embodiment employs an intelligent triggering mechanism for pressure feature extraction. A piezoelectric sensor array monitors barrel pressure fluctuations in real time, and a high-speed data acquisition system accurately captures pressure waveform characteristics. During data processing, a physical model is used to reconstruct the signal, improving the reliability of feature extraction.

[0165] This embodiment achieves real-time monitoring of charge characteristics. Charge quantity and filling status information are obtained through a high-precision weighing system and environmental parameter monitoring. The feature extraction process considers the influence of environmental factors such as temperature and humidity, and a compensation algorithm is used to improve the accuracy of feature parameters.

[0166] This embodiment innovatively implements a feature dimensionality reduction strategy. It utilizes the feature transformation matrix obtained during the training phase to perform a consistent transformation on real-time features. The transformation process maintains the correlation between features while significantly reducing data dimensionality, thereby improving the efficiency of subsequent predictions.

[0167] This embodiment designs a real-time computation framework for the prediction model. The model input layer employs a dynamic caching mechanism to ensure continuous input of feature vectors. A parallel processing strategy is used during computation to improve prediction speed. Prediction results are evaluated using confidence levels to ensure output reliability.

[0168] This embodiment innovatively constructs a Bayesian prior distribution model. The prior distribution is established based on the statistical characteristics of historical data, fully considering the wear evolution patterns under different operating conditions. The model parameters are determined through maximum likelihood estimation to ensure the accuracy of the prior information.

[0169] This embodiment implements dynamic updates to the posterior probability distribution. The posterior distribution is calculated using Bayesian inference, combining current predicted values ​​and historical data. The updates to the distribution parameters take into account the timeliness of the data, with recent data receiving higher weight.

[0170] This embodiment incorporates a correction mechanism for the prediction results. The correction process considers not only the mean of the posterior distribution but also uses variance information to assess the uncertainty of the prediction. The correction parameters are adjusted using an adaptive algorithm to ensure the stability of the correction effect.

[0171] This embodiment employs an innovative sliding window strategy. The window size is determined based on the wear evolution characteristics over a timescale, and the window sliding step size is dynamically adjusted according to the firing frequency. This approach ensures both the continuity of prediction and improves computational efficiency.

[0172] This embodiment implements time-series processing of the prediction results. Key features of wear evolution are identified through trend analysis of the data within the window. The time-series correlation of the prediction results is described using a state transition model, improving the accuracy of the prediction.

[0173] This embodiment achieves real-time assessment of artillery ablation wear through an innovative prediction framework. The scheme combines physical models and data-driven methods to provide reliable predictive capabilities. The overall design fully considers practical application needs, achieving end-to-end optimization from data acquisition to prediction output.

[0174] The prediction results of this embodiment demonstrate excellent real-time performance and accuracy. Through multi-level analysis and correction, accurate prediction of the ablation wear process is achieved. This scheme exhibits strong predictive capabilities in practical applications, providing a reliable basis for equipment maintenance decisions.

[0175] The prediction framework in this embodiment has good adaptability and scalability, and can adapt to prediction needs under different operating conditions. Through feature adaptation and Bayesian correction, it achieves accurate prediction of complex ablation processes, providing reliable decision support for equipment management.

[0176] In one embodiment of the escape optimization-based gun barrel wear prediction method of this application, see [link to relevant documentation]. Figure 7 It can also specifically include the following:

[0177] Step S701: Perform time-series fitting on the corrected prediction result sequence, construct a degradation trend curve reflecting the ablation wear evolution process using the polynomial regression method, calculate the first derivative of the degradation trend curve at different times to obtain the wear change rate feature, calculate the second derivative of the degradation trend curve at different times to obtain the wear acceleration feature, and establish a quantitative evaluation index for the ablation wear state of the gun barrel based on the wear change rate feature and the wear acceleration feature.

[0178] Step S702: Determine the wear change rate threshold and wear acceleration threshold based on historical data statistical analysis. When the wear change rate characteristic exceeds the wear change rate threshold or the wear acceleration characteristic exceeds the wear acceleration threshold, trigger the early warning mechanism. Determine the early warning level of the ablation and wear state of the gun barrel based on the preset early warning level classification standard.

[0179] Optionally, this embodiment employs a polynomial regression method to fit the corrected predicted sequence. During the fitting process, the polynomial order is first determined based on the characteristics of the firing data, and the optimal fitting parameters are selected through cross-validation. For the wear characteristics of artillery at different stages of use, a piecewise fitting strategy is adopted to ensure that the degradation trend curve accurately reflects the wear evolution process.

[0180] This embodiment innovatively extracts wear change rate features. By calculating the first derivative of the degradation trend curve, the wear change rate at different times is obtained. The calculation process employs a numerical differentiation method and introduces a smoothing mechanism to reduce the impact of noise. The change rate feature directly reflects the evolution speed of the wear process.

[0181] This embodiment designs a wear acceleration feature extraction scheme. The second derivative of the degradation trend curve is calculated to obtain the acceleration characteristics of wear changes. An adaptive step-size strategy is used during the calculation to ensure the stability of feature extraction. The acceleration features reflect the development trend of the wear process.

[0182] This embodiment implements an innovative state assessment index system. The assessment indexes comprehensively consider the rate of change and acceleration characteristics to construct a multi-dimensional state representation. The index weights are determined through a combination of expert experience and data analysis to ensure the scientific validity of the assessment results.

[0183] This embodiment employs a data-driven threshold determination method. A distribution model of wear characteristics is established through statistical analysis of historical data. The threshold setting considers the wear characteristic distribution under different operating conditions, and a reasonable threshold range is determined using the quantile method.

[0184] This embodiment innovatively designs a dynamic threshold adjustment mechanism. Based on newly added data, the feature distribution model is continuously updated to achieve adaptive adjustment of the threshold. This dynamic mechanism ensures the matching of the warning standard with actual usage conditions.

[0185] This embodiment implements a multi-level early warning mechanism. The early warning level classification considers two dimensions: wear level and development trend, constructing a two-dimensional early warning matrix. Different early warning levels correspond to different management measures, providing a basis for equipment maintenance decisions.

[0186] This embodiment designs a warning level determination process. The determination process adopts a fuzzy decision-making method, comprehensively considering multiple feature indicators. By establishing a fuzzy rule base, intelligent determination of the warning level is achieved.

[0187] This embodiment innovatively achieves dynamic updates of early warning information. As new monitoring data is continuously input, the early warning status is updated in real time. The update process takes into account the lag characteristics of status transitions, avoiding frequent fluctuations in early warning levels.

[0188] This embodiment employs a tiered early warning processing strategy. Different early warning levels trigger different response mechanisms, escalating from suggestive warnings to mandatory warnings. Early warning information includes wear and tear assessments and trend predictions, providing comprehensive support for maintenance decisions.

[0189] This embodiment achieves accurate assessment of the ablation and wear status of artillery through an innovative evaluation and early warning framework. This scheme combines physical characteristics and statistical analysis to provide reliable early warning capabilities. The overall design fully considers practical application needs, achieving end-to-end optimization from status assessment to early warning output.

[0190] The assessment and early warning results of this embodiment demonstrate excellent accuracy and reliability. Through multi-dimensional analysis and dynamic adjustment, a precise assessment of the ablation and wear condition is achieved. This scheme exhibits strong early warning capabilities in practical applications, providing a scientific basis for equipment maintenance.

[0191] The assessment and early warning framework in this embodiment exhibits good adaptability and scalability, enabling it to meet assessment needs under different operating conditions. Through feature adaptation and dynamic threshold adjustment, it achieves accurate assessment of complex ablation processes, providing reliable decision support for equipment management. This scheme significantly improves the scientific rigor and effectiveness of artillery maintenance and support.

[0192] To establish a reliable early warning mechanism for ablation wear conditions and effectively address the shortcomings of traditional prediction techniques in parameter optimization and condition early warning, thereby providing a scientific basis for the use and maintenance of artillery barrels, this application provides an embodiment of an escape optimization-based artillery barrel wear prediction device for implementing all or part of the escape optimization-based artillery barrel wear prediction method. See [link to embodiment]. Figure 8 The escape-optimized artillery barrel wear prediction device specifically includes the following components:

[0193] The dataset construction module 10 is used to collect historical ablation and wear data of multiple artillery pieces at different firing stages, extract temperature feature parameters, pressure feature parameters, and charge feature parameters from the historical ablation and wear data, convert the temperature feature parameters, pressure feature parameters, and charge feature parameters into a comprehensive load feature vector, construct a correspondence sample set between the comprehensive load feature vector and the ablation and wear amount, and use a stratified sampling method to divide the correspondence sample set into a training sample subset and a validation sample subset;

[0194] The optimal parameter determination module 20 is used to establish the search space of the escape optimization algorithm based on the training sample subset, set the initial position distribution of the search particle swarm, map the position information of the search particle swarm to the prediction model parameters, calculate the prediction error corresponding to each particle using the cross-validation method, use the prediction error as the particle fitness function, sort the particle swarm according to the fitness function, select the particle with the best fitness as the global optimal solution, update the velocity and position of the particles based on the escape mechanism in each iteration, stop the search when the number of iterations reaches a preset value, and use the finally converged global optimal solution as the optimal prediction model parameters.

[0195] The wear state prediction module 30 is used to read the real-time firing data of the artillery to be predicted, extract temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters from the real-time firing data, generate a comprehensive load feature vector by converting the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters according to the same transformation rule, input the comprehensive load feature vector into a prediction model with the optimal prediction model parameters, calculate the ablation wear prediction value at different firing times, perform Bayesian correction on the ablation wear prediction value to obtain the corrected prediction result, construct an ablation wear degradation trend curve of the artillery barrel based on the corrected prediction result, extract wear change rate characteristics and wear acceleration characteristics from the degradation trend curve, and provide early warning of the ablation wear state of the artillery barrel based on the wear change rate characteristics and wear acceleration characteristics.

[0196] As described above, the escape optimization-based gun barrel wear prediction device provided in this application can construct a sample set corresponding to the ablation wear amount by transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector. By designing the search space of the escape optimization algorithm and combining cross-validation and fitness ranking mechanisms, global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable ablation wear state early warning mechanism. This method effectively solves the shortcomings of traditional prediction techniques in parameter optimization and state early warning, providing a scientific basis for the use and maintenance of gun barrels.

[0197] From a hardware perspective, in order to establish a reliable early warning mechanism for ablation and wear conditions, effectively address the shortcomings of traditional prediction technologies in parameter optimization and condition early warning, and provide a scientific basis for the use and maintenance of artillery barrels, this application provides an embodiment of an electronic device for implementing all or part of the escape optimization-based artillery barrel wear prediction method. The electronic device specifically includes the following components:

[0198] The system comprises a processor, memory, a communications interface, and a bus; wherein the processor, memory, and communications interface communicate with each other via the bus; the communications interface is used to transmit information between the escape-optimized artillery barrel wear prediction device and core business systems, user terminals, and related databases and other related equipment; the logic controller can be a desktop computer, tablet computer, or mobile terminal, etc., and this embodiment is not limited to these. In this embodiment, the logic controller can be implemented with reference to the embodiments of the escape-optimized artillery barrel wear prediction method and the escape-optimized artillery barrel wear prediction device in the embodiments, the contents of which are incorporated herein, and repeated details will not be described again.

[0199] It is understood that the user terminal may include smartphones, tablet computers, network set-top boxes, portable computers, desktop computers, personal digital assistants (PDAs), in-vehicle devices, smart wearable devices, etc. Among these, the smart wearable devices may include smart glasses, smartwatches, smart bracelets, etc.

[0200] In practical applications, parts of the escape optimization-based gun barrel wear prediction method can be executed on the electronic device side as described above, or all operations can be completed in the client device. The choice can be made based on the processing power of the client device and the limitations of the user's usage scenario. This application does not impose any limitations on this. If all operations are completed in the client device, the client device may further include a processor.

[0201] The aforementioned client device may have a communication module (i.e., a communication unit) that can communicate with a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side; in other implementation scenarios, it may also include a server on an intermediate platform, such as a server on a third-party server platform that has a communication link with the task scheduling center server. The server may include a single computer device, a server cluster consisting of multiple servers, or a distributed server structure.

[0202] Figure 9This is a schematic block diagram illustrating the system configuration of the electronic device 9600 according to an embodiment of this application. Figure 9 As shown, the electronic device 9600 may include a central processing unit 9100 and a memory 9140; the memory 9140 is coupled to the central processing unit 9100. It is worth noting that... Figure 9 This is an example; other types of structures can also be used to supplement or replace this structure to achieve telecommunications functions or other functions.

[0203] In one embodiment, the escape-optimized gun barrel wear prediction method can be integrated into the central processing unit 9100. The central processing unit 9100 can be configured to perform the following control:

[0204] Step S101: Collect historical ablation and wear data of multiple artillery pieces at different firing stages, extract temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters from the historical ablation and wear data, convert the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters into a comprehensive load feature vector, construct a correspondence sample set between the comprehensive load feature vector and the ablation and wear amount, and use a stratified sampling method to divide the correspondence sample set into a training sample subset and a validation sample subset;

[0205] Step S102: Establish the search space of the escape optimization algorithm based on the training sample subset, set the initial position distribution of the search particle swarm, map the position information of the search particle swarm to the prediction model parameters, calculate the prediction error corresponding to each particle using the cross-validation method, use the prediction error as the particle fitness function, sort the particle swarm by fitness according to the fitness function, select the particle with the best fitness as the global optimal solution, update the velocity and position of the particles based on the escape mechanism in each iteration, stop the search when the number of iterations reaches a preset value, and use the finally converged global optimal solution as the optimal prediction model parameters;

[0206] Step S103: Read the real-time firing data of the artillery to be predicted, extract temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters from the real-time firing data, generate a comprehensive load feature vector by converting the temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters according to the same transformation rule, input the comprehensive load feature vector into the prediction model with the optimal prediction model parameters, calculate the ablation wear prediction value at different firing times, perform Bayesian correction on the ablation wear prediction value to obtain the corrected prediction result, construct the ablation wear degradation trend curve of the artillery barrel based on the corrected prediction result, extract the wear change rate feature and wear acceleration feature from the degradation trend curve, and provide early warning of the ablation wear state of the artillery barrel based on the wear change rate feature and the wear acceleration feature.

[0207] As described above, the electronic device provided in this application constructs a sample set corresponding to the ablation wear amount by transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector. By designing the search space of the escape optimization algorithm and combining cross-validation and fitness ranking mechanisms, global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable ablation wear state early warning mechanism. This method effectively addresses the shortcomings of traditional prediction techniques in parameter optimization and state early warning, providing a scientific basis for the use and maintenance of artillery barrels.

[0208] In another embodiment, the escape-optimized gun barrel wear prediction device can be configured separately from the central processing unit 9100. For example, the escape-optimized gun barrel wear prediction device can be configured as a chip connected to the central processing unit 9100, and the escape-optimized gun barrel wear prediction method function can be realized through the control of the central processing unit.

[0209] like Figure 9 As shown, the electronic device 9600 may further include: a communication module 9110, an input unit 9120, an audio processor 9130, a display 9160, and a power supply 9170. It is worth noting that the electronic device 9600 does not necessarily need to include these components. Figure 9 All components shown; in addition, the electronic device 9600 may also include Figure 9 For components not shown, please refer to existing technology.

[0210] like Figure 9 As shown, the central processing unit 9100, sometimes also referred to as a controller or operating control, may include a microprocessor or other processor device and / or logic device, which receives inputs and controls the operation of various components of the electronic device 9600.

[0211] The memory 9140 may be, for example, one or more of a cache, flash memory, hard drive, removable media, volatile memory, non-volatile memory, or other suitable devices. It may store the aforementioned failure-related information, and also store a program for executing that information. The central processing unit 9100 may execute the program stored in the memory 9140 to perform information storage or processing, etc.

[0212] Input unit 9120 provides input to central processing unit 9100. Input unit 9120 may be, for example, a keypad or touch input device. Power supply 9170 provides power to electronic device 9600. Display 9160 displays images and text. Display may be, for example, an LCD display, but is not limited thereto.

[0213] The memory 9140 can be a solid-state memory, such as a read-only memory (ROM), random access memory (RAM), a SIM card, etc. It can also be a memory that retains information even when power is off, can be selectively erased, and contains more data; examples of this type of memory are sometimes referred to as EPROMs. The memory 9140 can also be some other type of device. The memory 9140 includes a buffer memory 9141 (sometimes referred to as a buffer). The memory 9140 may include an application / function storage unit 9142 for storing application programs and function programs or processes for executing the operation of the electronic device 9600 via the central processing unit 9100.

[0214] The memory 9140 may also include a data storage unit 9143 for storing data, such as contacts, digital data, pictures, sounds, and / or any other data used by the electronic device. The driver storage unit 9144 of the memory 9140 may include various drivers for the electronic device for communication functions and / or for performing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0215] The communication module 9110 is a transmitter / receiver that sends and receives signals via the antenna 9111. The communication module 9110 (transmitter / receiver) is coupled to the central processing unit 9100 to provide input signals and receive output signals, which is the same as in a conventional mobile communication terminal.

[0216] Based on different communication technologies, multiple communication modules 9110 can be configured in the same electronic device, such as cellular network modules, Bluetooth modules, and / or wireless LAN modules. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide audio output via the speaker 9131 and receive audio input from the microphone 9132, thereby realizing typical telecommunications functions. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. Additionally, the audio processor 9130 is coupled to a central processing unit 9100, enabling on-device recording via the microphone 9132 and on-device playback of stored audio via the speaker 9131.

[0217] Embodiments of this application also provide a computer-readable storage medium capable of implementing all steps of the escape-optimized artillery barrel wear prediction method with a server or client as the execution subject in the above embodiments. The computer-readable storage medium stores a computer program that, when executed by a processor, implements all steps of the escape-optimized artillery barrel wear prediction method with a server or client as the execution subject in the above embodiments. For example, when the processor executes the computer program, it implements the following steps:

[0218] Step S101: Collect historical ablation and wear data of multiple artillery pieces at different firing stages, extract temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters from the historical ablation and wear data, convert the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters into a comprehensive load feature vector, construct a correspondence sample set between the comprehensive load feature vector and the ablation and wear amount, and use a stratified sampling method to divide the correspondence sample set into a training sample subset and a validation sample subset;

[0219] Step S102: Establish the search space of the escape optimization algorithm based on the training sample subset, set the initial position distribution of the search particle swarm, map the position information of the search particle swarm to the prediction model parameters, calculate the prediction error corresponding to each particle using the cross-validation method, use the prediction error as the particle fitness function, sort the particle swarm by fitness according to the fitness function, select the particle with the best fitness as the global optimal solution, update the velocity and position of the particles based on the escape mechanism in each iteration, stop the search when the number of iterations reaches a preset value, and use the finally converged global optimal solution as the optimal prediction model parameters;

[0220] Step S103: Read the real-time firing data of the artillery to be predicted, extract temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters from the real-time firing data, generate a comprehensive load feature vector by converting the temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters according to the same transformation rule, input the comprehensive load feature vector into the prediction model with the optimal prediction model parameters, calculate the ablation wear prediction value at different firing times, perform Bayesian correction on the ablation wear prediction value to obtain the corrected prediction result, construct the ablation wear degradation trend curve of the artillery barrel based on the corrected prediction result, extract the wear change rate feature and wear acceleration feature from the degradation trend curve, and provide early warning of the ablation wear state of the artillery barrel based on the wear change rate feature and the wear acceleration feature.

[0221] As described above, the computer-readable storage medium provided in this application constructs a sample set corresponding to the amount of ablation wear by transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector. By designing the search space of the escape optimization algorithm and combining cross-validation and fitness ranking mechanisms, global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable ablation wear state early warning mechanism. This method effectively addresses the shortcomings of traditional prediction techniques in parameter optimization and state early warning, providing a scientific basis for the use and maintenance of artillery barrels.

[0222] Embodiments of this application also provide a computer program product capable of implementing all steps in the escape optimization-based artillery barrel wear prediction method described above, where the execution subject is a server or client. When executed by a processor, this computer program / instruction implements the steps of the escape optimization-based artillery barrel wear prediction method. For example, the computer program / instruction implements the following steps:

[0223] Step S101: Collect historical ablation and wear data of multiple artillery pieces at different firing stages, extract temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters from the historical ablation and wear data, convert the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters into a comprehensive load feature vector, construct a correspondence sample set between the comprehensive load feature vector and the ablation and wear amount, and use a stratified sampling method to divide the correspondence sample set into a training sample subset and a validation sample subset;

[0224] Step S102: Establish the search space of the escape optimization algorithm based on the training sample subset, set the initial position distribution of the search particle swarm, map the position information of the search particle swarm to the prediction model parameters, calculate the prediction error corresponding to each particle using the cross-validation method, use the prediction error as the particle fitness function, sort the particle swarm by fitness according to the fitness function, select the particle with the best fitness as the global optimal solution, update the velocity and position of the particles based on the escape mechanism in each iteration, stop the search when the number of iterations reaches a preset value, and use the finally converged global optimal solution as the optimal prediction model parameters;

[0225] Step S103: Read the real-time firing data of the artillery to be predicted, extract temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters from the real-time firing data, generate a comprehensive load feature vector by converting the temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters according to the same transformation rule, input the comprehensive load feature vector into the prediction model with the optimal prediction model parameters, calculate the ablation wear prediction value at different firing times, perform Bayesian correction on the ablation wear prediction value to obtain the corrected prediction result, construct the ablation wear degradation trend curve of the artillery barrel based on the corrected prediction result, extract the wear change rate feature and wear acceleration feature from the degradation trend curve, and provide early warning of the ablation wear state of the artillery barrel based on the wear change rate feature and the wear acceleration feature.

[0226] As described above, the computer program product provided in this application constructs a sample set corresponding to the ablation wear amount by transforming temperature, pressure, and propellant characteristics into a comprehensive load feature vector. By designing the search space of the escape optimization algorithm and combining cross-validation and fitness ranking mechanisms, global optimization of the prediction model parameters is achieved. The system employs Bayesian correction to improve prediction accuracy and extracts wear change rate and acceleration features based on the degradation trend curve, establishing a reliable ablation wear state early warning mechanism. This method effectively addresses the shortcomings of traditional prediction techniques in parameter optimization and state early warning, providing a scientific basis for the use and maintenance of artillery barrels.

[0227] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0228] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (devices), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0229] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0230] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0231] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for predicting gun barrel wear based on escape optimization, characterized in that, The method includes: Historical ablation and wear data of multiple artillery pieces at different firing stages were collected. Temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters were extracted from the historical ablation and wear data. The temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters were transformed into a comprehensive load feature vector. A sample set of the correspondence between the comprehensive load feature vector and the ablation and wear amount was constructed. A stratified sampling method was used to divide the sample set of the correspondence into a training sample subset and a validation sample subset. The search space boundary is determined based on the feature distribution of the training sample subset. Each feature dimension is determined through statistical analysis. An initial particle swarm is generated through hypercube sampling, and the search space is divided into multiple subspaces. The initial particles are evenly distributed in the search space. The particle positions of the initial particles are mapped to model parameters to obtain weight parameters and bias parameters. According to the network structure of the prediction model, the mapped model parameters are organized into a matrix according to hierarchical relationships. The training sample subset is divided into multiple training sets through K-fold cross-validation. The importance weights of different wear states are determined by the prediction error evaluation framework. The validation results with large fluctuations in the K-fold cross-validation results are filtered by analysis of variance. The fitness of the particle swarm is sorted by a fast sorting method to obtain the global optimal solution. The velocity and position of the particles are updated based on the escape mechanism in each iteration. The search stops when the number of iterations reaches a preset value. The finally converged global optimal solution is used as the optimal prediction model parameters. The system reads real-time firing data of the artillery to be predicted, extracts temperature, pressure, and propellant characteristic parameters from the real-time firing data, generates a comprehensive load feature vector by converting the temperature, pressure, and propellant characteristic parameters according to the same transformation rule, inputs the comprehensive load feature vector into a prediction model with the parameters of the optimal prediction model, calculates the predicted ablation wear values ​​at different firing times, performs Bayesian correction on the predicted ablation wear values ​​to obtain the corrected prediction results, constructs an ablation wear degradation trend curve of the artillery barrel based on the corrected prediction results, extracts wear change rate and wear acceleration characteristics from the degradation trend curve, and provides early warning of the ablation wear state of the artillery barrel based on the wear change rate and wear acceleration characteristics.

2. The method for predicting gun barrel wear based on escape optimization according to claim 1, characterized in that, The process involves collecting historical ablation and wear data from multiple artillery pieces at different firing stages, extracting temperature, pressure, and propellant characteristic parameters from the historical ablation and wear data, converting these parameters into a comprehensive load feature vector, and constructing a sample set of correspondences between the comprehensive load feature vector and the ablation and wear amount, including: During the firing of multiple artillery pieces, barrel temperature data, bore pressure data, and propellant loading data are collected. The barrel temperature data is normalized according to the overload factor to obtain temperature characteristic parameters. The bore pressure data is normalized according to the peak pressure to obtain pressure characteristic parameters. The propellant loading data is normalized according to the quality standard to obtain propellant loading characteristic parameters. Principal component analysis is used to perform dimensionality reduction on the temperature characteristic parameters, the pressure characteristic parameters, and the propellant loading characteristic parameters to obtain a feature vector reflecting the comprehensive load of the firing process. The feature vector is sampled in time sequence and the cumulative ablation wear corresponding to each firing moment is calculated. A mapping relationship between the sequence of the feature vector and the ablation wear is established. The mapping relationship is divided into a training sample set and a validation sample set using a stratified random sampling method. The data in the training sample set and the validation sample set are normalized and standardized to construct a standardized feature vector and ablation wear correspondence sample set.

3. The method for predicting gun barrel wear based on escape optimization according to claim 1, characterized in that, The step of using stratified sampling to divide the corresponding sample set into training sample subsets and validation sample subsets includes: Based on the distribution characteristics of ablation wear in the corresponding sample set, the sample data is divided into multiple wear level layers. The proportion of sample data in each wear level layer is calculated. The sample data in the wear level layer is randomly shuffled. According to the preset ratio of training set to validation set, a corresponding number of sample data are extracted from each wear level layer. The extracted sample data are used to construct training sample subsets and validation sample subsets respectively. The training sample subset and the validation sample subset are subjected to data balance test. The distribution difference of each wear level sample in the training set and the validation set is calculated. When the distribution difference exceeds a preset threshold, stratified sampling is performed again until the data distribution of the training sample subset and the validation sample subset meets the balance requirement.

4. The method for predicting gun barrel wear based on escape optimization according to claim 1, characterized in that, The process involves establishing a search space for the escape optimization algorithm based on the subset of training samples, setting the initial position distribution of the search particle swarm, mapping the position information of the search particle swarm to prediction model parameters, calculating the prediction error for each particle using cross-validation, using the prediction error as the particle fitness function, ranking the particle swarm by fitness function, and selecting the particle with the best fitness as the global optimum. The boundary range of the search space is determined based on the feature dimensions of the training sample subset. An initial particle swarm is randomly generated in the search space. Each dimensional component of the initial particle swarm is mapped to the weight parameters and bias parameters of the prediction model. The parameter matrix of the prediction model is constructed based on the weight parameters and the bias parameters. The training sample subset is divided into multiple training sets and test sets using the K-fold cross-validation method. The training set is input into the prediction model for training. The prediction error of the model is calculated using the test set. The cross-validation process is repeated for each particle to obtain the average prediction error. The average prediction error is used as the fitness value of the particle. The particle swarm is sorted in ascending order based on the fitness value. The position of the particle with the smallest fitness value is selected as the global optimal solution for the current iteration.

5. The method for predicting gun barrel wear based on escape optimization according to claim 1, characterized in that, The process of updating the particle's velocity and position based on the escape mechanism in each iteration, stopping the search when the number of iterations reaches a preset value, and using the finally converged global optimal solution as the optimal prediction model parameters includes: Calculate the Euclidean distance between each particle and the global optimal solution. When the Euclidean distance is less than a preset escape threshold, trigger the escape mechanism. Add an escape perturbation term to the particle's current velocity direction. Update the particle's velocity vector based on the escape perturbation term and inertia weight. Calculate the particle's new position based on the velocity vector. Map the position components that exceed the search space boundary and convert the mapped position information into new prediction model parameters. The fitness value of the particle swarm is recalculated and the global optimal solution is updated after each iteration. The historical change trajectory of the global optimal solution is recorded. When the change of the global optimal solution is less than the convergence threshold or the number of iterations reaches the preset upper limit, the search is stopped. The global optimal solution at the final convergence is used as the parameters of the optimized prediction model.

6. The method for predicting gun barrel wear based on escape optimization according to claim 1, characterized in that, The process involves generating a comprehensive load feature vector from the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters according to the same transformation rule; inputting this comprehensive load feature vector into a prediction model with the parameters of the optimal prediction model; calculating the predicted ablation wear values ​​at different firing times; and performing Bayesian correction on the predicted ablation wear values ​​to obtain the corrected prediction results, including: The temperature, pressure, and propellant characteristics collected in real time are substituted into the principal component analysis model. The temperature, pressure, and propellant characteristics are then subjected to dimensionality reduction transformation based on the feature transformation matrix obtained during the training phase to generate a feature vector that reflects the comprehensive load of the current firing process. The feature vector is then input into the prediction model, and the ablation wear prediction value at the current moment is calculated based on the optimal prediction model parameters. A Bayesian prior distribution model is constructed, using historical prediction data as prior information. The posterior probability distribution of the predicted ablation wear value is calculated. The prediction results are corrected based on the mean and variance of the posterior probability distribution. The corrected prediction results at different firing times are calculated sequentially using the sliding time window method.

7. The method for predicting gun barrel wear based on escape optimization according to claim 1, characterized in that, The process involves constructing an ablation and wear degradation trend curve for the gun barrel based on the corrected prediction results, extracting wear rate characteristics and wear acceleration characteristics from the degradation trend curve, and providing early warning of the ablation and wear state of the gun barrel based on the wear rate characteristics and wear acceleration characteristics, including: The corrected prediction result sequence is subjected to time series fitting processing. A degradation trend curve reflecting the evolution process of ablation wear is constructed using the polynomial regression method. The first derivative of the degradation trend curve at different times is calculated to obtain the wear change rate feature. The second derivative of the degradation trend curve at different times is calculated to obtain the wear acceleration feature. A quantitative evaluation index for the ablation wear state of the gun barrel is established based on the wear change rate feature and the wear acceleration feature. Based on historical data statistical analysis, the wear change rate threshold and wear acceleration threshold are determined. When the wear change rate characteristic exceeds the wear change rate threshold or the wear acceleration characteristic exceeds the wear acceleration threshold, an early warning mechanism is triggered. The early warning level is determined based on the preset early warning level classification standard to assess the ablation and wear state of the artillery barrel.

8. A gun barrel wear prediction device based on escape optimization, characterized in that, The device includes: The dataset construction module is used to collect historical ablation and wear data of multiple artillery pieces at different firing stages, extract temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters from the historical ablation and wear data, convert the temperature characteristic parameters, pressure characteristic parameters, and charge characteristic parameters into a comprehensive load feature vector, construct a correspondence sample set between the comprehensive load feature vector and the ablation and wear amount, and use a stratified sampling method to divide the correspondence sample set into a training sample subset and a validation sample subset; The optimal parameter determination module is used to determine the search space boundary based on the feature distribution of the training sample subset, determine each feature dimension through statistical analysis, generate an initial particle swarm through hypercube sampling, divide the search space into multiple subspaces, and make the initial particles evenly distributed in the search space. The particle positions of the initial particles are mapped to model parameters to obtain weight parameters and bias parameters. According to the network structure of the prediction model, the mapped model parameters are organized into a matrix according to hierarchical relationships. The training sample subset is divided into multiple training sets through K-fold cross-validation. The importance weights of different wear states are determined by the prediction error evaluation framework. The validation results with large fluctuations in the K-fold cross-validation results are filtered by analysis of variance. The fitness of the particle swarm is sorted by a fast sorting method to obtain the global optimal solution. The velocity and position of the particles are updated based on the escape mechanism in each iteration. The search stops when the number of iterations reaches a preset value. The finally converged global optimal solution is used as the optimal prediction model parameters. The wear state prediction module is used to read real-time firing data of the artillery to be predicted, extract temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters from the real-time firing data, generate a comprehensive load feature vector by converting the temperature characteristic parameters, pressure characteristic parameters, and propellant characteristic parameters according to the same transformation rule, input the comprehensive load feature vector into a prediction model with the parameters of the optimal prediction model, calculate the predicted value of ablation wear at different firing times, perform Bayesian correction on the predicted value of ablation wear to obtain the corrected prediction result, construct an ablation wear degradation trend curve of the artillery barrel based on the corrected prediction result, extract wear change rate characteristics and wear acceleration characteristics from the degradation trend curve, and provide early warning of the ablation wear state of the artillery barrel based on the wear change rate characteristics and wear acceleration characteristics.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the escape optimization-based gun barrel wear prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the escape optimization-based gun barrel wear prediction method as described in any one of claims 1 to 7.

Images