Method and device for predicting ablation abrasion loss of barrel based on combined model

Abstract

The embodiment of the invention provides a barrel ablation abrasion loss prediction method and device based on a combined model, and a double-layer prediction architecture is constructed by combining an adaptive neuron growth mechanism and a dynamic programming algorithm. Ablation abrasion loss identification is realized through a hierarchical neural network trained by temperature, pressure and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system adopts an escape search optimization operator to select an optimal wear evolution path, establishes a multi-factor early warning threshold model in combination with a wear rate and an acceleration index, and realizes accurate prediction and evaluation early warning of an ablation wear state. According to the method, the defects of the traditional technology in the aspects of model integration and state early warning are effectively overcome, and a reliable basis is provided for artillery maintenance decisions.

Description

Technical Field

[0001] The present application relates to the field of data processing, and in particular to a method and device for predicting barrel ablation wear based on a combined model. Background Art

[0002] The existing barrel ablation wear prediction technology has obvious shortcomings. Traditional prediction methods often use a single model structure and fail to effectively integrate the advantages of neural networks and dynamic programming, resulting in limited prediction accuracy and generalization ability. At the same time, the existing methods lack a reasonable constraint mechanism when dealing with the problem of time-series wear evolution, making it difficult to ensure the physical interpretability of the prediction results.

[0003] In addition, existing technologies also have bottlenecks in energy state modeling and path optimization. Most methods ignore the cumulative effect of energy loss and cannot accurately describe the evolution of ablation wear. The optimization algorithm lacks an escape search mechanism and is prone to falling into a local optimal solution, affecting the reliability of the prediction results.

[0004] The existing system has technical shortcomings in status warning and evaluation. The warning mechanism is too simple and lacks comprehensive consideration of wear rate and acceleration, making it difficult to achieve accurate multi-factor threshold warning. Solving these problems is of great significance to improving the safety of artillery use and the scientific nature of maintenance decisions. Summary of the invention

[0005] In response to the problems in the prior art, the present application provides a method and device for predicting barrel ablation and wear based on a combined model, which can effectively solve the shortcomings of traditional technology in model integration and status warning, and provide a reliable basis for artillery maintenance decisions.

[0006] In order to solve at least one of the above problems, the present application provides the following technical solutions: In a first aspect, the present application provides a method for predicting barrel ablation wear based on a combined model, comprising: Collect historical data of ablation wear of the gun barrel, extract temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establish a hierarchical neural network model using an adaptive neuron growth mechanism, train the hierarchical neural network model with the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to obtain an ablation wear identifier, construct a dynamic programming algorithm model of ablation wear changing with time, and input the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; Collecting sensor data of the gun barrel during continuous firing, inputting the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, constructing an energy state constraint function based on the distribution parameters, applying evolution constraints to the dynamic programming algorithm model, substituting the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generating a set of wear evolution paths that meet the energy constraints, calculating the cumulative energy loss of each path in the wear evolution path set, and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path; The state vector of the optimal wear evolution path at different shooting moments is extracted, an evolution state curve characterizing the ablation wear of the barrel is established according to the state vector, a wear rate index and a wear acceleration index are extracted from the evolution state curve, the wear rate index and the wear acceleration index are input into a preset multi-factor early warning threshold model, the ablation wear state of the barrel is determined based on the multi-factor early warning threshold model, and an early warning mechanism is triggered when the ablation wear state exceeds a preset threshold, and a state assessment report including a predicted value of the ablation wear amount and an early warning level is generated.

[0007] Furthermore, the historical data of ablation and wear of the gun barrel is collected, and temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters are extracted from the historical data. A hierarchical neural network model is established by using an adaptive neuron growth mechanism, and the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters are used as input variables to train the hierarchical neural network model to obtain an ablation wear identifier, and a dynamic programming algorithm model of ablation wear changing with time is constructed, including: Collecting historical data of ablation and wear of the gun barrel, performing denoising processing on the historical data based on the wavelet transform method, calculating the values ​​of the temperature sensor, the pressure sensor and the charge sensor at different times, converting the values ​​into temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters in combination with the sensor calibration curve, and performing normalization processing on the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters to generate a characteristic vector; A hierarchical neural network model is constructed using an adaptive neuron growth mechanism, the feature vector is input into the hierarchical neural network model, the hierarchical neural network model is trained based on a back-propagation algorithm to obtain an ablation wear identifier, an ablation wear state transfer matrix is ​​established using a Markov process, a dynamic programming algorithm model including the state transfer matrix is ​​constructed, and the output result of the ablation wear identifier is substituted into the dynamic programming algorithm model.

[0008] Furthermore, the step of inputting the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model includes: The output result of the ablation wear identifier is divided into multiple time series intervals according to the number of shootings, and the ablation wear data in each time series interval is piecewise linearly fitted to calculate the wear change rate and wear acceleration in different intervals, and a state transition probability matrix reflecting the evolution characteristics of ablation wear is constructed based on the wear change rate and the wear acceleration; The state transfer probability matrix is ​​substituted into the dynamic programming algorithm model to construct a staged wear evolution model that includes the temperature accumulation effect, pressure accumulation effect and charge accumulation effect during the shooting process. Physical boundary constraints are imposed on the staged wear evolution model, and the Bellman equation is used to iteratively calculate the state transfer process of each stage.

[0009] Further, the sensor data of the gun barrel during continuous firing is collected, the sensor data is input into the ablation wear identifier to obtain the distribution parameters of the ablation wear, an energy state constraint function is constructed based on the distribution parameters, an evolution constraint is imposed on the dynamic programming algorithm model, the dynamic programming algorithm model with the evolution constraint is substituted into the escape search operator for iterative optimization, and a set of wear evolution paths that meet the energy constraint is generated, including: Collecting data from temperature sensors, pressure sensors and charge sensors during continuous firing of the gun barrel, performing time domain sampling and feature extraction on the sensor data, inputting the extracted feature data into the ablation wear identifier, calculating the mean and variance distribution parameters of the ablation wear based on a Gaussian mixture model, and constructing an energy state constraint function including temperature energy dissipation, pressure energy loss and charge energy conversion using the distribution parameters; The energy state constraint function is substituted into the dynamic programming algorithm model to form a state transfer equation with evolution constraints, a global optimization operator based on the escape search mechanism is constructed, the state transfer equation is input into the global optimization operator for iterative optimization, and a Monte Carlo sampling method is used to generate multiple groups of initial solutions. Each group of initial solutions is subjected to escape search optimization to obtain a set of wear evolution paths that meet the energy constraints.

[0010] Furthermore, the calculating the cumulative energy loss of each path in the wear evolution path set and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path includes: Calculating the temperature cumulative loss, pressure cumulative loss and charge cumulative loss of each path in the wear evolution path set during the continuous shooting process, substituting the temperature cumulative loss, pressure cumulative loss and charge cumulative loss into an energy weight fusion model, calculating the comprehensive energy loss index of each path based on the energy weight fusion model, and normalizing the comprehensive energy loss index; A path evaluation criterion based on the energy minimization principle is constructed, and the normalized comprehensive energy loss index is sorted using the path evaluation criterion. The path with the smallest comprehensive energy loss index is selected from the wear evolution path set as the optimal wear evolution path, and the state transition sequence of the optimal wear evolution path at different shooting stages is calculated.

[0011] Furthermore, the extracting of the state vector of the optimal wear evolution path at different shooting moments, establishing an evolution state curve characterizing barrel ablation wear according to the state vector, and extracting a wear rate index and a wear acceleration index from the evolution state curve include: Extracting the state vector at each shooting moment from the optimal wear evolution path, constructing a state space model including a temperature state component, a pressure state component and a charge state component based on the state vector, performing continuous processing on the state space model using a cubic spline interpolation algorithm, and mapping the interpolated state space trajectory to a time-wear amount coordinate system to obtain an evolution state curve; The evolution state curve is numerically differentiated, the first-order derivative value of the curve at each sampling point is calculated to obtain the wear rate index, the rate of change of the first-order derivative value is calculated to obtain the wear acceleration index, and the wear rate index and the wear acceleration index are smoothed using a sliding average filtering algorithm.

[0012] Furthermore, the wear rate index and the wear acceleration index are input into a preset multi-factor early warning threshold model, and the barrel ablation wear state is determined based on the multi-factor early warning threshold model. When the ablation wear state exceeds the preset threshold, the early warning mechanism is triggered, and a state assessment report including the ablation wear amount prediction value and the early warning level is generated, including: The wear rate index and the wear acceleration index are normalized and input into a preset multi-factor early warning threshold model, the risk index of the current ablative wear state is calculated based on the multi-factor early warning threshold model, the risk index is mapped to a preset ablative wear grading standard by using a fuzzy comprehensive evaluation method, and the ablative wear state is determined by comparing the ablative wear grading standard with the preset threshold value; A warning trigger mechanism based on the remaining service life of the barrel is constructed. When the ablation wear state exceeds a preset threshold, the prediction interval and confidence level of the ablation wear amount of the barrel are calculated, the warning level is determined according to the prediction interval and the confidence level, and a status assessment report including the ablation wear amount prediction value and the warning level is generated.

[0013] In a second aspect, the present application provides a barrel ablation wear prediction device based on a combined model, comprising: A wear evolution module is used to collect historical data of ablation wear of the gun barrel, extract temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establish a hierarchical neural network model using an adaptive neuron growth mechanism, train the hierarchical neural network model with the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to obtain an ablation wear identifier, construct a dynamic programming algorithm model of ablation wear changing with time, and input the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; A path determination module, used to collect sensor data when the gun barrel is continuously fired, input the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, construct an energy state constraint function based on the distribution parameters, impose evolution constraints on the dynamic programming algorithm model, substitute the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generate a set of wear evolution paths that meet the energy constraints, calculate the cumulative energy loss of each path in the wear evolution path set, and select the path with the smallest cumulative energy loss as the optimal wear evolution path; The state evaluation module is used to extract the state vector of the optimal wear evolution path at different shooting moments, establish an evolution state curve representing the ablation wear of the barrel according to the state vector, extract a wear rate index and a wear acceleration index from the evolution state curve, input the wear rate index and the wear acceleration index into a preset multi-factor early warning threshold model, determine the ablation wear state of the barrel based on the multi-factor early warning threshold model, trigger an early warning mechanism when the ablation wear state exceeds a preset threshold, and generate a state evaluation report including a predicted value of the ablation wear amount and an early warning level.

[0014] In a third aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the method for predicting barrel ablation and wear based on the combined model are implemented.

[0015] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for predicting barrel ablation and wear based on a combined model.

[0016] In a fifth aspect, the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the method for predicting the ablation and wear of a barrel based on a combined model.

[0017] It can be seen from the above technical solution that the present application provides a method and device for predicting the ablation wear of the barrel based on a combined model. By combining the adaptive neuron growth mechanism and the dynamic programming algorithm, a two-layer prediction architecture is constructed. The ablation wear is identified by a hierarchical neural network trained with temperature, pressure, and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system uses an escape search optimization operator to select the optimal wear evolution path, and combines the wear rate and acceleration indicators to establish a multi-factor warning threshold model, which realizes the accurate prediction and evaluation warning of the ablation wear state. This method effectively solves the shortcomings of traditional technology in model integration and status warning, and provides a reliable basis for artillery maintenance decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is one of the flow charts of the barrel ablation wear prediction method based on the combined model in the embodiment of the present application; Figure 2 The second flowchart of the barrel ablation wear prediction method based on the combined model in the embodiment of the present application; Figure 3 The third flowchart of the barrel ablation wear prediction method based on the combined model in the embodiment of the present application; Figure 4 FIG4 is a flow chart of a method for predicting barrel ablation wear based on a combined model in an embodiment of the present application; Figure 5 FIG5 is a flow chart of a method for predicting barrel ablation wear based on a combined model in an embodiment of the present application; Figure 6 FIG6 is a flow chart of a method for predicting barrel ablation wear based on a combined model in an embodiment of the present application; Figure 7 FIG7 is a flow chart of a method for predicting barrel ablation wear based on a combined model in an embodiment of the present application; Figure 8 It is a structural diagram of a barrel ablation wear prediction device based on a combined model in an embodiment of the present application; Fig. 9 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application.

[0020] Reference numerals: 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 program storage unit 9144, antenna 9111, speaker 9131, microphone 9132. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0023] Taking into account the problems existing in the prior art, the present application provides a method and device for predicting the ablation wear of the barrel based on a combined model. By combining an adaptive neuron growth mechanism and a dynamic programming algorithm, a two-layer prediction architecture is constructed. The ablation wear is identified by a hierarchical neural network trained with temperature, pressure, and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system uses an escape search optimization operator to select the optimal wear evolution path, and combines the wear rate and acceleration indicators to establish a multi-factor warning threshold model, which realizes accurate prediction and evaluation warning of the ablation wear state. This method effectively solves the shortcomings of traditional technology in model integration and status warning, and provides a reliable basis for artillery maintenance decisions.

[0024] In order to effectively solve the deficiencies of traditional technologies in model integration and status warning and provide a reliable basis for artillery maintenance decision-making, this application provides an embodiment of a method for predicting barrel ablation wear based on a combined model, see Figure 1 The barrel ablation wear prediction method based on the combined model specifically includes the following contents: Step S101: collecting historical data of ablation wear of the gun barrel, extracting temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establishing a hierarchical neural network model using an adaptive neuron growth mechanism, using the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to train the hierarchical neural network model to obtain an ablation wear identifier, constructing a dynamic programming algorithm model of ablation wear changing with time, and inputting the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; Optionally, this embodiment deploys a high-precision sensor array at key positions of the gun barrel to collect historical data. The temperature sensor adopts a high-temperature resistant K-type thermocouple, and a measuring point is arranged every 50mm along the axial direction of the barrel. A total of 20 temperature measuring points are set to achieve continuous monitoring of the temperature field of the tube wall. The sampling frequency of each temperature measuring point is 1kHz to ensure the capture of transient temperature changes. The pressure sensor uses a piezoelectric sensor with a sensitivity of 5pc / MPa. 8 measuring points are arranged at different cross-sections of the barrel, and the sampling frequency is 100kHz to obtain the chamber pressure fluctuation characteristics. The charging parameters obtain the charging amount through a weighing system with an accuracy of 0.1g, and temperature and humidity sensors are arranged to monitor the charging environment parameters.

[0025] This embodiment implements a multi-level data preprocessing solution. First, the original data is normalized, and different physical quantities are mapped to the [-1,1] interval to eliminate the dimension effect. Then, the db4 wavelet basis function is used to decompose the signal into 5 layers, and the soft threshold method is selected to remove high-frequency noise. When reconstructing the signal, an adaptive threshold strategy is adopted, and the threshold size is related to the local characteristics of the signal to ensure the retention of important feature information. For data points with missing samples, the cubic spline interpolation method is used to repair them.

[0026] This embodiment innovatively designs a feature extraction framework. Temperature features include parameters such as the maximum wall temperature, radial temperature gradient, axial temperature gradient, heating rate, cooling rate, etc., through which the thermal stress distribution is fully characterized. Pressure features include parameters such as maximum chamber pressure, pressure rise rate, pressure duration, pressure fluctuation frequency, etc., which reflect the dynamic load characteristics during the projectile movement. Charge features include parameters such as charge amount, charge density, filling coefficient, ambient temperature, ambient humidity, etc., which reflect the influence of launch conditions.

[0027] This embodiment constructs a network model based on an improved neuron growth mechanism. The initial network adopts a 4-6-1 structure, the input layer corresponds to three types of features: temperature, pressure, and charge, and the hidden layer adopts a ReLU activation function. During the training process, high error areas are identified by analyzing the distribution of samples in the feature space. When the local error exceeds the preset threshold, neurons are dynamically added to the area. The initial weights of the newly added neurons are obtained by interpolating the weights of the surrounding neurons to ensure a smooth transition of the network structure.

[0028] This embodiment implements an innovative network training strategy. Mini-batch training is adopted with a batch size of 32, and the weights are updated through the Adam optimizer. The initial value of the learning rate is set to 0.01, and the cosine annealing strategy is used for dynamic adjustment. To prevent overfitting, the L2 regularization term is introduced with a regularization coefficient of 0.001. During the training process, when the validation set error does not decrease for 5 consecutive rounds, the early stopping mechanism is triggered.

[0029] This embodiment designs a comprehensive verification mechanism. The data set is divided into a training set, a verification set, and a test set in a ratio of 7:2:1. The verification set contains data under different launch conditions, covering various combinations of temperature, pressure, and charge. The model performance is evaluated by calculating the mean square error, mean absolute error, and R square coefficient. Based on the verification results, the network structure and training parameters are optimized.

[0030] This embodiment innovatively constructs a dynamic programming model. The ablation wear process is divided into multiple stages according to the number of shots, and the state of each stage is characterized by the current wear amount and wear rate. A state transition probability matrix is ​​defined, and the element values ​​are obtained based on historical data statistics. A cost function that takes into account the heat accumulation effect is constructed, and the optimal state sequence is solved iteratively through the Bellman equation.

[0031] This embodiment achieves accurate modeling of ablation and wear of artillery barrels through the combination of adaptive neural networks and dynamic programming. The scheme fully considers the physical mechanism during the firing process and establishes a mapping relationship between characteristic parameters and wear. The prediction accuracy and generalization ability of the model are ensured through innovative network structure design and training strategies. The introduction of the dynamic programming model makes the prediction results have better time series consistency. The overall scheme shows excellent prediction performance in practical applications and provides a reliable basis for artillery maintenance decisions.

[0032] Step S102: collecting sensor data of the gun barrel during continuous firing, inputting the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, constructing an energy state constraint function based on the distribution parameters, applying evolution constraints to the dynamic programming algorithm model, substituting the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generating a set of wear evolution paths that meet the energy constraints, calculating the cumulative energy loss of each path in the wear evolution path set, and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path; Optionally, a real-time data acquisition system is designed in this embodiment. During the continuous firing of the artillery, the temperature sensor adopts a distributed acquisition architecture, each measuring point is configured with an independent signal conditioning circuit, and the sampling frequency is dynamically adjusted. When a firing signal is detected, the sampling frequency is increased to 5kHz to capture the transient temperature change characteristics; during the firing interval, it is reduced to 100Hz to monitor the cooling process. The pressure sensor adopts a synchronous triggering method and starts collecting after the firing signal arrives. The sampling frequency is set to 200kHz to record the complete pressure waveform. The charging parameters are recorded by a high-precision weighing system during each loading.

[0033] This embodiment implements a real-time data processing framework. The sliding window method is used to process the sensor data in segments, with a window length of 100ms and an overlap rate of 50%. Fast Fourier transform is performed in each window to extract frequency domain features. Interference signals are removed by adaptive threshold filtering to ensure the stability of feature extraction. The processed feature data is transmitted to the ablation wear identifier through a cache queue.

[0034] This embodiment innovatively constructs a distribution parameter estimation model. The output results of the ablation wear identifier are input into the Gaussian mixture model, and the model parameters are estimated using the expectation maximization algorithm. The mixture model contains three Gaussian components, corresponding to mild, moderate and severe wear states. By calculating the posterior probability, the probability distribution characteristics of the wear amount are obtained.

[0035] This embodiment designs an energy state constraint function. The constraint function considers three energy forms: thermal energy loss, mechanical energy loss, and chemical energy conversion. Thermal energy loss is calculated through temperature field distribution, taking into account heat conduction and convection heat transfer. Mechanical energy loss is evaluated based on pressure fluctuation characteristics, including friction work and deformation work. Chemical energy conversion is estimated through charge parameters, taking into account the impact of combustion efficiency.

[0036] This embodiment implements the constraint optimization of the dynamic programming model. The energy constraint is converted into the state transition cost, and the energy penalty term is added to the Bellman equation. When the state transition causes the energy loss to exceed the threshold, the corresponding transition probability is reduced. This mechanism ensures that the predicted wear evolution path complies with the principle of energy conservation.

[0037] This embodiment innovatively designs an escape search algorithm. The algorithm initializes multiple search paths, each path contains N state points, where N is the prediction step length. During the search process, candidate solutions are generated through local perturbations. The perturbation amplitude is related to the degree of satisfaction of the energy constraint. The more severe the violation of the constraint, the greater the perturbation. The taboo table mechanism is introduced to avoid repeated searches, and the taboo length is dynamically adjusted.

[0038] This embodiment implements a path optimization strategy. During the iteration process, the elite solution set is maintained, and several paths with the lowest energy loss are saved. The elite solution is updated after each iteration, and the paths with large energy loss are eliminated. The iteration termination conditions include the maximum number of iterations and the convergence criterion to ensure the convergence of the algorithm.

[0039] This embodiment designs a path evaluation system. For the generated wear evolution path set, the cumulative energy loss of each path is calculated. The energy loss includes the sum of thermal energy, mechanical energy and chemical energy loss at each time step. In the calculation process, the time series characteristics of energy loss are considered and the time weight factor is introduced.

[0040] This embodiment innovatively realizes the optimal path selection. Based on the Pareto optimization principle, the two goals of minimizing energy loss and path smoothness are comprehensively considered. By calculating the normalized value of energy loss and path curvature, a comprehensive evaluation index is constructed. The path with the best comprehensive evaluation is selected as the final prediction result.

[0041] This embodiment significantly improves the physical rationality of wear prediction by introducing energy constraints and escape search mechanisms. The scheme fully considers the energy conversion law during the launch process to ensure that the prediction results conform to the laws of physics. Through the multi-objective optimization strategy, while ensuring the accuracy of the prediction, the interpretability of the prediction results is improved. The overall scheme shows excellent prediction performance in practical applications and provides reliable support for artillery maintenance decisions.

[0042] Step S103: extracting the state vector of the optimal wear evolution path at different shooting moments, establishing an evolution state curve characterizing the ablation wear of the barrel according to the state vector, extracting a wear rate index and a wear acceleration index from the evolution state curve, inputting the wear rate index and the wear acceleration index into a preset multi-factor early warning threshold model, determining the ablation wear state of the barrel based on the multi-factor early warning threshold model, triggering an early warning mechanism when the ablation wear state exceeds a preset threshold, and generating a state assessment report including a predicted value of the ablation wear amount and an early warning level.

[0043] Optionally, this embodiment implements a state vector extraction module for the optimal wear evolution path. At each firing moment, a multidimensional state vector including temperature state quantity, pressure state quantity and charge state quantity is extracted. The temperature state quantity includes the tube wall temperature distribution, thermal stress distribution and temperature gradient, which are acquired in real time through the thermocouple array. The pressure state quantity includes the maximum chamber pressure, pressure waveform characteristics and pressure fluctuation frequency, which are collected by the pressure sensor network. The charge state quantity includes charge parameters and environmental parameters to achieve a comprehensive characterization of the firing conditions.

[0044] This embodiment designs an innovative state space modeling method. A three-dimensional state space model is used, and the coordinate axes correspond to temperature, pressure, and charge characteristics respectively. The extracted state vector is mapped into the state space to form a discrete state point sequence. The state points are connected by a cubic spline interpolation algorithm to ensure the continuity and smoothness of the state trajectory. The tension coefficient adaptive adjustment strategy is adopted during the interpolation process to avoid unreasonable oscillations in the trajectory.

[0045] This embodiment realizes the construction of the evolution state curve. The state space trajectory is projected onto the time-wear volume plane to generate a state curve that characterizes the ablation wear evolution process. The ordinate of the curve represents the cumulative wear volume, and the abscissa represents the number of shots. The shape of the curve is optimized by piecewise polynomial fitting, and the fitting order is adaptively selected according to the data complexity to ensure that the curve can accurately reflect the wear trend without overfitting.

[0046] This embodiment innovatively designs a feature extraction algorithm. The evolution state curve is numerically differentiated, and the first-order derivative is calculated using the central difference format to obtain the wear rate index. To improve the calculation accuracy, the differential step size is dynamically adjusted according to the local characteristics of the curve. The second-order derivative is further calculated to obtain the wear acceleration index, which reflects the change trend of the wear rate. The noise effect in the differentiation process is eliminated by sliding average filtering.

[0047] This embodiment implements a multi-factor early warning model. The model input includes wear rate index, wear acceleration index and state space characteristics. The fuzzy logic reasoning mechanism is adopted to design a fuzzy rule base to characterize the relationship between each index and the wear state. The fuzzy rules are formulated by combining expert knowledge and data analysis to ensure the rationality and completeness of the rules.

[0048] This embodiment designs an adaptive threshold determination mechanism. Based on the statistical analysis of historical data, a distribution model of wear indicators is established. The quantile method is used to determine the warning threshold, and the threshold is dynamically updated with the sample distribution. Multi-level warning thresholds are set to correspond to different levels of danger to achieve graded warning.

[0049] This embodiment innovatively implements an early warning trigger mechanism. When the monitoring index exceeds the corresponding threshold, the early warning program is started. The early warning level is determined based on the degree of over-limit and duration. In order to avoid frequent alarms, a hysteresis judgment mechanism is introduced to ensure the stability of the early warning. At the same time, the prediction results are corrected online through the Kalman filter algorithm to improve the reliability of the early warning.

[0050] This embodiment designs a status assessment report generation module. The report contains information such as current wear, predicted wear, warning level and confidence interval. The predicted wear is obtained by extrapolating the status curve, and the confidence interval is determined based on the prediction error analysis. The warning level is intuitively displayed in a color-coded manner for quick identification.

[0051] This embodiment achieves accurate assessment of the ablation and wear state of the gun barrel through multi-dimensional state analysis and multi-factor early warning mechanism. The scheme combines physical models with data-driven methods to establish a mapping relationship between wear evolution characteristics and early warning indicators. Through innovative feature extraction algorithms and early warning judgment mechanisms, the accuracy and reliability of state assessment are improved. The overall scheme shows excellent early warning effects in practical applications and provides a scientific basis for gun maintenance and guarantee.

[0052] From the above description, it can be seen that the barrel ablation wear prediction method based on the combined model provided in the embodiment of the present application can construct a two-layer prediction architecture by combining the adaptive neuron growth mechanism and the dynamic programming algorithm. The ablation wear is identified by a hierarchical neural network trained with temperature, pressure, and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system uses the escape search optimization operator to select the optimal wear evolution path, and combines the wear rate and acceleration indicators to establish a multi-factor warning threshold model, which realizes the accurate prediction and evaluation warning of the ablation wear state. This method effectively solves the shortcomings of traditional technology in model integration and status warning, and provides a reliable basis for artillery maintenance decisions.

[0053] In one embodiment of the barrel ablation wear prediction method based on the combined model of the present application, see Figure 2 , and can also include the following: Step S201: collecting historical data of ablation and wear of the gun barrel, performing denoising processing on the historical data based on the wavelet transform method, calculating the values ​​of the temperature sensor, the pressure sensor and the charge sensor at different times, converting the values ​​into temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters in combination with the sensor calibration curve, and performing normalization processing on the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters to generate a characteristic vector; Step S202: construct a hierarchical neural network model using an adaptive neuron growth mechanism, input the feature vector into the hierarchical neural network model, train the hierarchical neural network model based on a back propagation algorithm to obtain an ablation wear identifier, establish an ablation wear state transfer matrix using a Markov process, construct a dynamic programming algorithm model including the state transfer matrix, and substitute the output result of the ablation wear identifier into the dynamic programming algorithm model.

[0054] Optionally, this embodiment uses a distributed data acquisition architecture to record the historical data of the gun barrel. A high-precision sensor array is arranged at the key position of the gun barrel, including 20 K-type thermocouple temperature measurement points evenly distributed along the axial direction, and each measurement point is equipped with a signal conditioning circuit and a 16-bit analog-to-digital converter. Pressure monitoring uses 8 piezoelectric sensors with a sensitivity of 5pc / MPa and a sampling frequency of 200kHz. The charging parameters are obtained through a high-precision weighing system and environmental monitoring equipment to achieve comprehensive monitoring of the launch conditions.

[0055] This embodiment innovatively implements a signal denoising scheme based on wavelet transform. The db4 wavelet basis function is selected to perform a 5-layer decomposition of the original signal to obtain wavelet coefficients of different scales. By analyzing the statistical characteristics of the wavelet coefficients, an adaptive threshold method is used to distinguish between signal and noise components. The threshold size is related to the local characteristics of the signal, and noise suppression is achieved while maintaining useful information. The signal reconstruction adopts a soft threshold processing method to ensure the smoothness of the reconstructed signal.

[0056] This embodiment designs a high-precision sensor calibration system. The temperature sensor uses a black body furnace as a standard source, performs multi-point calibration in the range of 25-1000°C, and establishes a mapping relationship between thermoelectric potential and temperature. The pressure sensor uses a pressure standard generator for dynamic calibration, covering a range of 0-600MPa. The calibration curve uses a piecewise polynomial fit to ensure conversion accuracy in different working ranges.

[0057] This embodiment implements a feature parameter extraction framework. Temperature features include parameters such as maximum temperature, temperature gradient, heating rate, and cooling rate. Pressure features include parameters such as maximum chamber pressure, pressure rise rate, pressure duration, and pressure fluctuation frequency. Charge features include charge quantity, charge density, filling coefficient, and environmental parameters. Each feature is mapped to a unified interval through normalization processing to eliminate the dimension effect.

[0058] This embodiment innovatively designs an adaptive neuron growth mechanism. The initial network structure adopts the minimum topology, the input layer corresponds to the feature vector dimension, and the output layer is the ablation wear amount. The number of hidden layer nodes is dynamically adjusted during the training process, and the location of the newly added nodes is determined based on local error analysis. The node growth process takes into account the network complexity constraints to avoid excessive growth leading to a decrease in generalization ability.

[0059] This embodiment implements an improved back propagation training algorithm. The batch stochastic gradient descent method is used to update the weights, and the batch size is adjusted dynamically. The learning rate adopts an adaptive adjustment strategy, maintaining a large value in the early stage to accelerate convergence, and gradually reducing it in the later stage to ensure stability. The momentum term and weight decay term are introduced to optimize the training process.

[0060] This embodiment designs an innovative verification mechanism. K-fold cross-validation is used to evaluate model performance, and the verification set contains data from different working conditions. Verification indicators include prediction accuracy, generalization ability, and stability. Based on the verification results, the network structure and training parameters are dynamically adjusted to achieve model optimization.

[0061] This embodiment constructs a state transition model based on the Markov process. The ablation and wear process is discretized into multiple states, and the transition probability between states is obtained through historical data statistics. Considering the impact of the shooting conditions on the transition probability, a conditional transfer matrix is ​​established. The elements of the transfer matrix are dynamically updated as samples accumulate to maintain the adaptability of the model.

[0062] This embodiment innovatively implements a dynamic programming algorithm model. The ablation wear prediction problem is transformed into an optimal path search problem, and the state space is defined by the wear amount and the operating condition parameters. A cost function that considers physical constraints is constructed, and the optimal state sequence is iteratively solved through the Bellman equation. A heuristic strategy is introduced into the optimization process to improve the solution efficiency.

[0063] This embodiment achieves accurate modeling of ablation and wear of artillery barrels through the combination of adaptive neural networks and dynamic programming. The scheme fully considers the physical mechanism during the firing process and establishes a mapping relationship between characteristic parameters and wear. The prediction accuracy and generalization ability of the model are ensured through innovative network structure design and training strategies. The introduction of the dynamic programming model makes the prediction results have better time series consistency. The overall scheme shows excellent prediction performance in practical applications and provides a reliable basis for artillery maintenance decisions.

[0064] In one embodiment of the barrel ablation wear prediction method based on the combined model of the present application, see Figure 3 , and can also include the following: Step S301: dividing the output result of the ablation wear identifier into multiple time series intervals according to the number of shootings, performing piecewise linear fitting on the ablation wear data in each time series interval, calculating the wear change rate and wear acceleration in different intervals, and constructing a state transition probability matrix reflecting the ablation wear evolution characteristics based on the wear change rate and the wear acceleration; Step S302: Substitute the state transition probability matrix into the dynamic programming algorithm model, construct a staged wear evolution model that includes the temperature accumulation effect, pressure accumulation effect and charge accumulation effect during the shooting process, impose physical boundary constraints on the staged wear evolution model, and use the Bellman equation to iteratively calculate the state transition process of each stage.

[0065] Optionally, this embodiment implements a time series partitioning module. Based on the output results of the ablation wear identifier, an adaptive partitioning strategy is used to divide the time series. The partitioning process takes into account factors such as shooting intensity, continuous firing time, and cooling interval to ensure that the data in each interval has similar physical characteristics. For the high-intensity continuous firing stage, a smaller time window is used; for the intermittent firing stage, the time window is appropriately increased to achieve accurate modeling of different firing modes.

[0066] This embodiment innovatively designs a piecewise linear fitting algorithm. In each time series interval, a dynamic programming method is used to determine the optimal segmentation point. The selection of segmentation points is based on the error minimization criterion, while considering the constraint of the number of segments. The weighted least squares method is used to perform linear fitting on each segment of data, and the weight is related to the reliability of the data point. The wear change rate is obtained by analyzing the slope of the fitted line, and the wear acceleration is calculated by the change in the slope of adjacent segments.

[0067] This embodiment realizes the construction of the state transition probability matrix. Based on the wear change rate and acceleration, the ablation wear state is discretized into multiple levels. By analyzing the transition relationship between states at adjacent moments, the state transition frequency is statistically obtained. The Bayesian estimation method is used to calculate the transition probability, and the prior knowledge is introduced to constrain the probability distribution to ensure the physical rationality of the transfer matrix.

[0068] This embodiment designs an innovative cumulative effect model. The temperature cumulative effect takes into account thermal stress cycles and material fatigue, and determines the influence of temperature on material properties through thermal-mechanical coupling analysis. The pressure cumulative effect reflects the accumulation of material damage caused by chamber pressure impact, and is modeled based on damage mechanics theory. The charge cumulative effect describes the coupling effect of chemical corrosion and mechanical wear, and considers the influence of charge characteristics on the wear process.

[0069] This embodiment implements a staged wear evolution model. The continuous wear process is decomposed into multiple evolution stages, each of which corresponds to a specific combination of working conditions. Markov chains are used to describe state transitions within the stages, and the transfer of cumulative effects is considered between stages. Model parameters are obtained through historical data training to ensure the accuracy of prediction results.

[0070] This embodiment innovatively designs physical boundary constraints. Based on the principles of material science and mechanics, constraints including strength limit, strain limit and temperature limit are established. Constraints are dynamically updated with cumulative effects to reflect the evolution of material properties during use. Feasible domain analysis is used to ensure that the prediction results meet physical laws.

[0071] This embodiment implements an improved Bellman equation solution algorithm. A forward dynamic programming method is used to perform state transfer calculations, and the optimal value function is maintained at each stage. A heuristic search strategy is introduced to reduce the computational space and improve the solution efficiency. The state transfer sequence is optimized through error back propagation to ensure the acquisition of the global optimal solution.

[0072] This embodiment designs an iterative optimization mechanism. In the process of solving the Bellman equation, the state transition cost is dynamically adjusted to keep the prediction result consistent with the physical constraints. An adaptive step size strategy is used to control the iteration process to ensure the convergence of the algorithm. The prediction performance is evaluated through cross-validation and the model parameters are optimized.

[0073] This embodiment significantly improves the reliability of wear prediction by introducing cumulative effects and physical constraints. The scheme combines statistical modeling with physical mechanisms to achieve an accurate description of the complex ablation wear process. Innovative state transition modeling and iterative optimization strategies ensure the physical rationality of the prediction results. The overall scheme shows excellent prediction performance in practical applications and provides a scientific basis for artillery maintenance decisions.

[0074] This embodiment realizes the dynamic prediction of the evolution process of ablation and wear of the gun barrel through reasonable time series division and state transition modeling. The scheme fully considers the physical mechanism during the shooting process and establishes the transition relationship between wear states. By introducing cumulative effects and physical constraints, the reliability of the prediction results is improved. The overall scheme provides important support for gun life assessment and maintenance decision-making.

[0075] In one embodiment of the barrel ablation wear prediction method based on the combined model of the present application, see Figure 4 , and can also include the following: Step S401: collecting temperature sensor, pressure sensor and charge sensor data of the gun barrel during continuous firing, performing time domain sampling and feature extraction on the sensor data, inputting the extracted feature data into the ablation wear identifier, calculating the mean and variance distribution parameters of the ablation wear based on the Gaussian mixture model, and constructing an energy state constraint function including temperature energy dissipation, pressure energy loss and charge energy conversion using the distribution parameters; Step S402: Substitute the energy state constraint function into the dynamic programming algorithm model to form a state transfer equation with evolution constraints, construct a global optimization operator based on the escape search mechanism, input the state transfer equation into the global optimization operator for iterative optimization, use the Monte Carlo sampling method to generate multiple groups of initial solutions, and perform escape search optimization on each group of initial solutions to obtain a set of wear evolution paths that meet the energy constraints.

[0076] Optionally, a high-precision data acquisition system is designed in this embodiment. The temperature sensor adopts a K-type thermocouple array, and a measuring point is arranged every 50 mm along the axial direction of the barrel to achieve continuous monitoring of the temperature field. The pressure sensor adopts a piezoelectric sensor, and 8 measuring points are arranged at key positions of the barrel, with a sampling frequency of 200kHz. The charging parameters are obtained through a high-precision weighing system and environmental monitoring equipment to ensure the accuracy and integrity of the data.

[0077] This embodiment implements an innovative time domain sampling strategy. An adaptive sampling method is used to dynamically adjust the sampling frequency during the launch process. When a firing signal is detected, the pressure sensor sampling frequency is increased to the maximum value to capture the details of the pressure waveform; the temperature sensor sampling frequency is increased accordingly to record transient temperature changes. The sampling frequency is reduced during the launch interval to achieve a balance between the amount of data and the amount of information.

[0078] This embodiment designs a multi-level feature extraction framework. Temperature features include maximum temperature, temperature gradient, heating rate and cooling rate. Pressure features include maximum chamber pressure, pressure rise rate, pressure duration and pressure fluctuation frequency. Charge features include charge quantity, charge density, filling coefficient and environmental parameters. Multi-scale analysis of the original signal is performed through wavelet transform to extract feature information of each scale.

[0079] This embodiment innovatively constructs a Gaussian mixture model. The model contains three Gaussian components, corresponding to light, moderate and severe wear states. The weight, mean and variance of each component are estimated by the expectation maximization algorithm. The model parameters are dynamically updated as data accumulates to maintain an accurate description of the wear state. By calculating the posterior probability, the probability distribution characteristics of the wear amount are obtained.

[0080] This embodiment implements the energy state constraint function. Temperature energy dissipation considers three mechanisms: heat conduction, convection heat transfer, and radiation heat transfer, and determines energy loss through thermodynamic analysis. Pressure energy loss includes friction work and deformation work, and energy conversion is calculated based on mechanical principles. Charge energy conversion considers chemical reaction efficiency and mechanical work conversion efficiency to achieve a quantitative description of energy balance.

[0081] This embodiment designs an innovative state transfer equation. The energy constraint is converted into a state transfer cost, and an energy penalty term is added to the Bellman equation. When the state transfer causes the energy loss to exceed the threshold, the corresponding transfer probability is reduced. It is ensured through iterative optimization that the state transfer sequence satisfies the energy conservation principle.

[0082] This embodiment implements an escape search optimization algorithm. The algorithm maintains a set of candidate solutions, each of which corresponds to a possible wear evolution path. New solutions are generated through local perturbations, and the perturbation amplitude is related to the degree of satisfaction of the energy constraint. The taboo table mechanism is introduced to avoid repeated searches, and the taboo length is dynamically adjusted. The elite retention strategy is used to save the optimal solution to ensure the convergence of the algorithm.

[0083] This embodiment innovatively designs a Monte Carlo sampling method. Based on the statistical characteristics of historical data, a probability distribution model of the initial solution is constructed. Multiple groups of initial solutions are generated through importance sampling, and each group of solutions contains N state points. The sampling process takes physical constraints into account to ensure the rationality of the initial solution. Each group of initial solutions is locally optimized to select the path with the least energy loss.

[0084] This embodiment implements an iterative optimization strategy. During the escape search process, the search parameters are dynamically adjusted, including the perturbation amplitude, taboo length, and termination condition. The optimization effect is evaluated through cross-validation to avoid falling into the local optimum. The generated path set is energy evaluated to screen feasible solutions that meet the constraints.

[0085] This embodiment significantly improves the physical rationality of wear prediction by introducing energy constraints and escape search mechanisms. The scheme fully considers the energy conversion law during the launch process to ensure that the prediction results conform to the laws of physics. Through the multi-objective optimization strategy, while ensuring the accuracy of the prediction, the interpretability of the prediction results is improved. The overall scheme shows excellent prediction performance in practical applications and provides reliable support for artillery maintenance decisions.

[0086] In one embodiment of the barrel ablation wear prediction method based on the combined model of the present application, see Figure 5 , and can also include the following: Step S501: Calculate the temperature cumulative loss, pressure cumulative loss and charge cumulative loss of each path in the wear evolution path set during the continuous shooting process, substitute the temperature cumulative loss, pressure cumulative loss and charge cumulative loss into the energy weight fusion model, calculate the comprehensive energy loss index of each path based on the energy weight fusion model, and normalize the comprehensive energy loss index; Step S502: Construct a path evaluation criterion based on the energy minimization principle, use the path evaluation criterion to sort the normalized comprehensive energy loss index, select the path with the smallest comprehensive energy loss index from the wear evolution path set as the optimal wear evolution path, and calculate the state transition sequence of the optimal wear evolution path in different shooting stages.

[0087] Optionally, this embodiment implements a multi-dimensional energy loss calculation module. The temperature accumulation loss is obtained through thermodynamic analysis, considering three mechanisms: heat conduction, convection heat transfer and radiation heat transfer. In each firing cycle, the temperature field distribution of the tube wall is calculated, and the heat loss is evaluated based on Fourier's heat conduction law. The convection heat transfer loss is calculated by the heat exchange between the tube wall and the projectile and the propellant gas. The radiation heat transfer loss is estimated based on the Stefan-Boltzmann law.

[0088] This embodiment innovatively designs a pressure cumulative loss evaluation method. By analyzing the action process of the pressure wave on the tube wall, the loss of pressure work converted into mechanical energy is calculated. Friction work loss is considered, including the sliding friction between the projectile and the tube wall and the viscous loss of the propellant gas. The plastic deformation work is evaluated to reflect the energy loss of the tube wall material under high pressure. The pressure loss of the complete shooting process is obtained by cumulative calculation.

[0089] This embodiment implements the analysis of charge cumulative loss. Based on the chemical reaction kinetics, the energy conversion efficiency during charge combustion is calculated. The conversion loss of chemical energy to mechanical energy is considered, including incomplete combustion loss and heat loss. The influence of charge parameters on energy conversion is evaluated, such as charge density, loading coefficient, etc. By establishing an energy balance equation, a quantitative description of charge cumulative loss is achieved.

[0090] This embodiment designs an innovative energy weight fusion model. The weight coefficients of various energy losses are determined by the hierarchical analysis method, and the weight values ​​reflect the degree of influence of different energy forms on the wear process. The weight coefficients are determined by combining expert evaluation and data analysis. The model considers the coupling effect between energy losses and reflects the interaction through cross terms.

[0091] This embodiment implements a comprehensive energy loss evaluation system. The weighted energy loss is normalized to eliminate the dimension effect. By constructing a dimensionless index, the comparability between different paths is achieved. The normalization process uses the maximum and minimum method to maintain the relative relationship of energy loss. The stability analysis of the normalized results is performed to ensure the reliability of the evaluation.

[0092] This embodiment innovatively constructs a path evaluation criterion. Based on the energy minimization principle, the wear evolution process is equivalent to a state transition sequence with minimal energy loss. The evaluation criterion includes energy efficiency terms and path smoothness terms, and comprehensively considers the rationality of energy loss and state transition. The multi-objective problem is processed by the Pareto optimization method to obtain a balanced evaluation result.

[0093] This embodiment designs a path sorting mechanism. A fast non-dominated sorting algorithm is used to stratify the path set, and each layer contains paths with similar energy loss. Congestion calculation is introduced to maintain the diversity of solutions. The optimal path is selected through the elite strategy to ensure the stability of the selection result. A sensitivity analysis is performed on the selected path to verify its robustness.

[0094] This embodiment realizes the state transition sequence extraction. The optimal wear evolution path is segmented, and each segment corresponds to a shooting stage. The transition characteristics between adjacent states are analyzed, including the transition direction and the transition rate. A state transition diagram is established to intuitively display the wear evolution process. The rationality of the transfer sequence is verified by Markov chain analysis.

[0095] This embodiment innovatively designs a verification mechanism. The accuracy of path selection is evaluated through cross-validation, and the verification set contains data from different working conditions. The prediction error and confidence interval are calculated to quantify the reliability of the prediction results. The rationality of state transfer is verified based on physical constraints to ensure that the prediction results are consistent with reality.

[0096] This embodiment achieves an accurate description of the evolution process of ablation and wear of the gun barrel through energy loss analysis and path optimization. The scheme combines thermodynamics, mechanics and chemistry principles to establish a complete energy evaluation system. The physical rationality of the prediction results is ensured through innovative path evaluation criteria and optimization strategies. The overall scheme shows excellent prediction performance in practical applications and provides a scientific basis for gun maintenance decisions.

[0097] In one embodiment of the barrel ablation wear prediction method based on the combined model of the present application, see Figure 6 , and can also include the following: Step S601: extracting the state vector at each shooting moment from the optimal wear evolution path, constructing a state space model including a temperature state component, a pressure state component and a charge state component based on the state vector, performing continuous processing on the state space model using a cubic spline interpolation algorithm, and mapping the interpolated state space trajectory to a time-wear amount coordinate system to obtain an evolution state curve; Step S602: perform numerical differentiation operation on the evolution state curve, calculate the first-order derivative value of the curve at each sampling point to obtain the wear rate index, calculate the rate of change of the first-order derivative value to obtain the wear acceleration index, and use a sliding average filtering algorithm to smooth the wear rate index and the wear acceleration index.

[0098] Optionally, this embodiment implements a state vector extraction module. From the optimal wear evolution path, multi-dimensional state information is extracted according to the firing sequence. The temperature state component includes the tube wall temperature distribution, thermal stress distribution and temperature gradient, which are acquired in real time through the thermocouple array. The pressure state component includes the maximum chamber pressure, pressure waveform characteristics and pressure fluctuation frequency, which are collected by the pressure sensor network. The charge state component includes charge parameters and environmental parameters to achieve a comprehensive characterization of the firing conditions.

[0099] This embodiment innovatively designs a state space modeling method. A three-dimensional state space model is used, and the coordinate axes correspond to temperature, pressure, and charge characteristics respectively. The extracted state vector is mapped to the state space to form a discrete state point sequence. The construction of the state space takes into account the coupling relationship between physical quantities to ensure that the model has physical meaning. The state space dimension is reduced by principal component analysis to retain key feature information.

[0100] This embodiment implements a cubic spline interpolation algorithm. For discrete state point sequences, a piecewise cubic polynomial is used for interpolation fitting. The interpolation function satisfies the continuity conditions of the first-order derivative and the second-order derivative at the nodes to ensure the smoothness of the state trajectory. The curve shape is adjusted by introducing a tension coefficient to avoid excessive oscillation. The endpoint derivative constraint is used to ensure physical rationality at the boundary.

[0101] This embodiment designs a trajectory mapping mechanism. The three-dimensional state space trajectory is projected onto the time-wear volume plane to generate a state curve that characterizes the ablation wear evolution process. The projection process maintains the topological relationship between the state points to ensure that the evolution characteristics are not distorted. The curve representation at different scales is achieved through coordinate transformation, which facilitates the analysis of wear characteristics at different time scales.

[0102] This embodiment innovatively implements a numerical differential algorithm. The first-order derivative of the state curve is calculated using the central difference format to obtain the wear rate index. The differential step size is dynamically adjusted according to the local characteristics of the curve to suppress the influence of noise while ensuring the calculation accuracy. The Richardson extrapolation is used to improve the derivative calculation accuracy and reduce the truncation error.

[0103] This embodiment designs a wear acceleration calculation method. The wear rate index is numerically differentiated to obtain the wear acceleration index. The calculation process uses a second-order precision differential format to ensure the accuracy of the acceleration estimation. An adaptive step control mechanism is introduced to avoid calculation instability at singular points. The rationality of the acceleration index is verified by physical constraints.

[0104] This embodiment implements a sliding average filtering algorithm. A variable-length sliding window is used to smooth the velocity and acceleration indicators. The window length is related to the local characteristics of the signal. A shorter window is used in fast-changing areas to retain details, and a longer window is used in stable areas to suppress noise. The filter weight adopts a Gaussian distribution to ensure the continuity of the smoothing result.

[0105] This embodiment innovatively designs an indicator verification mechanism. The rationality of the rate and acceleration indicators is verified through physical model analysis. The corresponding relationship between the indicators and the material damage mechanism is established to ensure that the indicators have clear physical meanings. The stability and reliability of the indicators are evaluated through cross-validation. Outliers are identified and processed to improve the robustness of the indicators.

[0106] This embodiment achieves accurate characterization of the ablation and wear process of the gun barrel through state space modeling and index extraction. The scheme fuses and analyzes multi-source sensor data to establish a complete description of state evolution. The accuracy of feature extraction is ensured through innovative interpolation algorithms and differential calculation methods. The overall scheme provides a reliable quantitative basis for wear state assessment and is of great value in guiding gun maintenance.

[0107] This embodiment shows excellent performance in practical applications. Through reasonable state space construction and trajectory processing, an accurate description of the complex wear process is achieved. Innovative numerical calculation methods and filtering strategies ensure the stability of index extraction. The overall solution provides a scientific basis for artillery life assessment and maintenance decision-making, and has important engineering application value.

[0108] In one embodiment of the barrel ablation wear prediction method based on the combined model of the present application, see Figure 7 , and can also include the following: Step S701: normalizing the wear rate index and the wear acceleration index and inputting them into a preset multi-factor warning threshold model, calculating the risk index of the current ablative wear state based on the multi-factor warning threshold model, mapping the risk index to a preset ablative wear grading standard using a fuzzy comprehensive evaluation method, and determining the ablative wear state by comparing the ablative wear grading standard with the preset threshold; Step S702: constructing a warning trigger mechanism based on the remaining service life of the barrel; when the ablation wear state exceeds a preset threshold, calculating the prediction interval and confidence level of the ablation wear amount of the barrel; determining the warning level according to the prediction interval and the confidence level; and generating a status assessment report including the ablation wear amount prediction value and the warning level.

[0109] Optionally, this embodiment implements a normalization processing module. The wear rate index is normalized to the maximum and minimum values ​​to eliminate the dimension effect. The wear acceleration index adopts the Z-score normalization method to achieve data centering and variance normalization. The normalization process considers the physical meaning of the index to ensure that the processing results have practical interpretability. The normalization effect is verified through mathematical and statistical analysis to ensure the rationality of the data distribution characteristics.

[0110] This embodiment innovatively designs a multi-factor warning threshold model. The model includes three dimensions: temperature accumulation factor, pressure damage factor, and charge effect factor. The weight coefficient of each factor is determined by physical analysis to reflect its influence on the wear state. The model structure adopts a hierarchical design to achieve fusion analysis of multi-scale features. The model parameters are optimized by combining expert knowledge base and data-driven.

[0111] This embodiment implements a method for calculating the hazard index. The comprehensive hazard is calculated based on the weighted sum criterion, and the weight coefficient is dynamically adjusted with the wear state. A nonlinear mapping function is introduced to reflect the complex relationship between the index and the hazard. The stability of the hazard calculation is evaluated through Monte Carlo simulation to ensure the reliability of the results.

[0112] This embodiment designs an innovative fuzzy comprehensive evaluation mechanism. A fuzzy set including slight wear, moderate wear and severe wear is established, and the fuzzy characteristics of the state are described by the membership function. The evaluation matrix is ​​constructed based on historical data and expert experience to reflect the contribution of various indicators to the wear level. The maximum membership principle is used to determine the final evaluation result.

[0113] This embodiment implements the ablative wear classification standard. The standard contains multiple grade intervals, each of which corresponds to specific wear characteristics. The interval division takes into account the material performance limit and the safety requirements for use to ensure the practicality of the classification results. The interval boundaries are determined by statistical analysis to achieve accurate division of the wear state.

[0114] This embodiment innovatively designs a warning trigger mechanism. The warning condition is established based on the remaining life of the barrel, and the warning is triggered when the wear state exceeds the safety threshold. The warning threshold is determined by reliability analysis, taking into account the material fatigue limit and safety margin. An adaptive adjustment mechanism is introduced to dynamically update the threshold according to the operating conditions.

[0115] This embodiment implements a confidence interval estimation method. The prediction interval is constructed using the Bootstrap resampling technique to reflect the uncertainty of the prediction result. The probability distribution of the predicted value is obtained through kernel density estimation, and the interval range under different confidence levels is calculated. A bias correction mechanism is introduced to improve the accuracy of interval estimation.

[0116] This embodiment designs a criterion for determining the warning level. The prediction interval and confidence level are mapped to a multi-level warning system, and different levels correspond to different maintenance measures. The warning level takes into account factors such as wear rate, accumulation, and development trend to achieve a comprehensive status assessment. The warning rules are optimized through the decision tree algorithm to improve the accuracy of the judgment.

[0117] This embodiment innovatively constructs a status assessment report generation module. The report content includes key information such as wear prediction value, prediction interval, confidence level and warning level. The prediction results are displayed through data visualization technology to intuitively reflect the evolution trend of wear status. The assessment report adopts a hierarchical structure to facilitate understanding and use by users at different levels.

[0118] This embodiment achieves accurate evaluation of the ablation and wear state of the gun barrel through multi-factor analysis and fuzzy judgment. The solution combines physical models with data analysis to establish a complete early warning decision framework. The reliability of the state evaluation is improved through innovative early warning mechanisms and evaluation methods. The overall solution provides a scientific basis for gun maintenance management and has important engineering application value.

[0119] This embodiment has shown excellent performance in engineering practice. Through a reasonable indicator system and evaluation mechanism, an accurate description of the complex wear process is achieved. The innovative early warning strategy and report generation method provide strong support for maintenance decisions. The overall solution significantly improves the scientificity and accuracy of artillery maintenance, which is of great significance to improving equipment reliability.

[0120] In order to effectively solve the deficiencies of traditional technologies in model integration and status warning and provide a reliable basis for artillery maintenance decision-making, the present application provides an embodiment of a barrel ablation wear prediction device based on a combined model for realizing all or part of the contents of the barrel ablation wear prediction method based on a combined model, see Figure 8 The barrel ablation wear prediction device based on the combined model specifically includes the following contents: The wear evolution module 10 is used to collect historical data of ablation wear of the gun barrel, extract temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establish a hierarchical neural network model using an adaptive neuron growth mechanism, train the hierarchical neural network model with the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to obtain an ablation wear identifier, construct a dynamic programming algorithm model of ablation wear changing with time, and input the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; A path determination module 20 is used to collect sensor data when the gun barrel is continuously fired, input the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, construct an energy state constraint function based on the distribution parameters, impose evolution constraints on the dynamic programming algorithm model, substitute the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generate a set of wear evolution paths that meet the energy constraints, calculate the cumulative energy loss of each path in the wear evolution path set, and select the path with the smallest cumulative energy loss as the optimal wear evolution path; The state evaluation module 30 is used to extract the state vector of the optimal wear evolution path at different shooting moments, establish an evolution state curve representing the ablation wear of the barrel according to the state vector, extract a wear rate index and a wear acceleration index from the evolution state curve, input the wear rate index and the wear acceleration index into a preset multi-factor early warning threshold model, determine the ablation wear state of the barrel based on the multi-factor early warning threshold model, trigger an early warning mechanism when the ablation wear state exceeds a preset threshold, and generate a state evaluation report including a predicted value of the ablation wear amount and an early warning level.

[0121] From the above description, it can be seen that the barrel ablation wear prediction device based on the combined model provided in the embodiment of the present application can construct a two-layer prediction architecture by combining the adaptive neuron growth mechanism and the dynamic programming algorithm. The ablation wear amount is identified by a hierarchical neural network trained with temperature, pressure, and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system uses the escape search optimization operator to select the optimal wear evolution path, and combines the wear rate and acceleration indicators to establish a multi-factor warning threshold model, which realizes the accurate prediction and evaluation warning of the ablation wear state. This method effectively solves the shortcomings of traditional technology in model integration and status warning, and provides a reliable basis for artillery maintenance decisions.

[0122] From the hardware level, in order to effectively solve the deficiencies of traditional technologies in model integration and status warning and provide a reliable basis for artillery maintenance decisions, the present application provides an embodiment of an electronic device for implementing all or part of the contents of the barrel ablation wear prediction method based on the combined model, and the electronic device specifically includes the following contents: Processor, memory, communication interface and bus; wherein the processor, memory and communication interface communicate with each other through the bus; the communication interface is used to realize information transmission between the barrel ablation wear prediction device based on the combined model and related devices such as the core business system, user terminal and related database; the logic controller can be a desktop computer, a tablet computer and a mobile terminal, etc., but the present embodiment is not limited thereto. In the present embodiment, the logic controller can be implemented with reference to the embodiment of the barrel ablation wear prediction method based on the combined model and the embodiment of the barrel ablation wear prediction device based on the combined model in the embodiment, and the contents thereof are incorporated herein, and the repeated parts are not repeated.

[0123] It is understandable that the user terminal may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, a smart watch, a smart bracelet, etc.

[0124] In practical applications, part of the barrel ablation wear prediction method based on the combined model can be executed on the electronic device side as described above, or all operations can be completed in the client device. The specific selection can be based on the processing capability of the client device and the limitations of the user's usage scenario. This application does not limit this. If all operations are completed in the client device, the client device may also include a processor.

[0125] The 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, and other implementation scenarios 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, or a server cluster consisting of multiple servers, or a server structure of a distributed device.

[0126] Fig. 9 FIG. 9 is a schematic block diagram of the system structure of the electronic device 9600 according to an embodiment of the present application. Fig. 9As shown, the electronic device 9600 may include a central processor 9100 and a memory 9140; the memory 9140 is coupled to the central processor 9100. It is worth noting that Fig. 9 is exemplary; other types of structures may also be used to supplement or replace this structure to implement telecommunication functions or other functions.

[0127] In one embodiment, the barrel ablation wear prediction method function based on the combined model can be integrated into the central processor 9100. The central processor 9100 can be configured to perform the following control: Step S101: collecting historical data of ablation wear of the gun barrel, extracting temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establishing a hierarchical neural network model using an adaptive neuron growth mechanism, using the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to train the hierarchical neural network model to obtain an ablation wear identifier, constructing a dynamic programming algorithm model of ablation wear changing with time, and inputting the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; Step S102: collecting sensor data of the gun barrel during continuous firing, inputting the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, constructing an energy state constraint function based on the distribution parameters, applying evolution constraints to the dynamic programming algorithm model, substituting the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generating a set of wear evolution paths that meet the energy constraints, calculating the cumulative energy loss of each path in the wear evolution path set, and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path; Step S103: extracting the state vector of the optimal wear evolution path at different shooting moments, establishing an evolution state curve characterizing the ablation wear of the barrel according to the state vector, extracting a wear rate index and a wear acceleration index from the evolution state curve, inputting the wear rate index and the wear acceleration index into a preset multi-factor early warning threshold model, determining the ablation wear state of the barrel based on the multi-factor early warning threshold model, triggering an early warning mechanism when the ablation wear state exceeds a preset threshold, and generating a state assessment report including a predicted value of the ablation wear amount and an early warning level.

[0128] From the above description, it can be seen that the electronic device provided in the embodiment of the present application constructs a two-layer prediction architecture by combining an adaptive neuron growth mechanism and a dynamic programming algorithm. The ablation wear amount is identified by a hierarchical neural network trained with temperature, pressure, and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system uses an escape search optimization operator to select the optimal wear evolution path, and combines the wear rate and acceleration indicators to establish a multi-factor warning threshold model, which realizes accurate prediction and evaluation warning of the ablation wear state. This method effectively solves the shortcomings of traditional technology in model integration and status warning, and provides a reliable basis for artillery maintenance decisions.

[0129] In another embodiment, the barrel ablation and wear prediction device based on the combined model can be configured separately from the central processing unit 9100. For example, the barrel ablation and wear prediction device based on the combined model can be configured as a chip connected to the central processing unit 9100, and the function of the barrel ablation and wear prediction method based on the combined model can be realized through the control of the central processing unit.

[0130] like Fig. 9 As shown, the electronic device 9600 may also 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 have to include Fig. 9 In addition, the electronic device 9600 may also include Fig. 9 For components not shown, reference may be made to the prior art.

[0131] like Fig. 9 As shown, the central processing unit 9100 is sometimes also referred to as a controller or an operation control, and may include a microprocessor or other processor device and / or logic device. The central processing unit 9100 receives input and controls the operation of various components of the electronic device 9600.

[0132] The memory 9140 may be, for example, one or more of a cache, a flash memory, a hard drive, a removable medium, a volatile memory, a non-volatile memory or other suitable devices. The above-mentioned information related to the failure may be stored, and a program for executing the relevant information may also be stored. The CPU 9100 may execute the program stored in the memory 9140 to implement information storage or processing, etc.

[0133] The input unit 9120 provides input to the central processing unit 9100. The input unit 9120 is, for example, a key or a touch input device. The power supply 9170 is used to provide power to the electronic device 9600. The display 9160 is used to display display objects such as images and texts. The display may be, for example, an LCD display, but is not limited thereto.

[0134] The memory 9140 may be a solid-state memory, such as a read-only memory (ROM), a random access memory (RAM), a SIM card, etc. It may also be a memory that saves information even when the power is off, can be selectively erased, and is provided with more data, examples of which are sometimes referred to as EPROMs, etc. The memory 9140 may 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, which is used to store application programs and function programs or processes for executing the operation of the electronic device 9600 through the central processor 9100.

[0135] 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 communication functions of the electronic device and / or for executing other functions of the electronic device (such as messaging applications, address book applications, etc.).

[0136] 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 processor 9100 to provide input signals and receive output signals, which may be the same as the case of a conventional mobile communication terminal.

[0137] Based on different communication technologies, multiple communication modules 9110 may be provided in the same electronic device, such as a cellular network module, a Bluetooth module and / or a wireless LAN module. The communication module 9110 (transmitter / receiver) is also coupled to a speaker 9131 and a microphone 9132 via an audio processor 9130 to provide an audio output via the speaker 9131 and receive an audio input from the microphone 9132, thereby realizing a common telecommunication function. The audio processor 9130 may include any suitable buffer, decoder, amplifier, etc. In addition, the audio processor 9130 is also coupled to the central processor 9100, so that recording can be performed on the local machine through the microphone 9132, and the sound stored on the local machine can be played through the speaker 9131.

[0138] The embodiment of the present application also provides a computer-readable storage medium capable of implementing all the steps in the method for predicting barrel ablation wear based on a combined model in the above embodiment, where the execution subject is a server or a client. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, all the steps in the method for predicting barrel ablation wear based on a combined model in the above embodiment are implemented. For example, when the processor executes the computer program, the following steps are implemented: Step S101: collecting historical data of ablation wear of the gun barrel, extracting temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establishing a hierarchical neural network model using an adaptive neuron growth mechanism, using the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to train the hierarchical neural network model to obtain an ablation wear identifier, constructing a dynamic programming algorithm model of ablation wear changing with time, and inputting the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; Step S102: collecting sensor data of the gun barrel during continuous firing, inputting the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, constructing an energy state constraint function based on the distribution parameters, applying evolution constraints to the dynamic programming algorithm model, substituting the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generating a set of wear evolution paths that meet the energy constraints, calculating the cumulative energy loss of each path in the wear evolution path set, and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path; Step S103: extracting the state vector of the optimal wear evolution path at different shooting moments, establishing an evolution state curve characterizing the ablation wear of the barrel according to the state vector, extracting a wear rate index and a wear acceleration index from the evolution state curve, inputting the wear rate index and the wear acceleration index into a preset multi-factor early warning threshold model, determining the ablation wear state of the barrel based on the multi-factor early warning threshold model, triggering an early warning mechanism when the ablation wear state exceeds a preset threshold, and generating a state assessment report including a predicted value of the ablation wear amount and an early warning level.

[0139] From the above description, it can be seen that the computer-readable storage medium provided in the embodiment of the present application constructs a two-layer prediction architecture by combining an adaptive neuron growth mechanism and a dynamic programming algorithm. The ablation wear amount is identified by a hierarchical neural network trained with temperature, pressure, and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system uses an escape search optimization operator to select the optimal wear evolution path, and combines the wear rate and acceleration indicators to establish a multi-factor warning threshold model, which realizes accurate prediction and evaluation warning of the ablation wear state. This method effectively solves the shortcomings of traditional technology in model integration and status warning, and provides a reliable basis for artillery maintenance decisions.

[0140] The embodiment of the present application also provides a computer program product capable of implementing all the steps of the barrel ablation wear prediction method based on the combined model in the above embodiment, where the execution subject is a server or a client. When the computer program / instruction is executed by a processor, the steps of the barrel ablation wear prediction method based on the combined model are implemented. For example, the computer program / instruction implements the following steps: Step S101: collecting historical data of ablation wear of the gun barrel, extracting temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establishing a hierarchical neural network model using an adaptive neuron growth mechanism, using the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to train the hierarchical neural network model to obtain an ablation wear identifier, constructing a dynamic programming algorithm model of ablation wear changing with time, and inputting the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; Step S102: collecting sensor data of the gun barrel during continuous firing, inputting the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, constructing an energy state constraint function based on the distribution parameters, applying evolution constraints to the dynamic programming algorithm model, substituting the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generating a set of wear evolution paths that meet the energy constraints, calculating the cumulative energy loss of each path in the wear evolution path set, and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path; Step S103: extracting the state vector of the optimal wear evolution path at different shooting moments, establishing an evolution state curve characterizing the ablation wear of the barrel according to the state vector, extracting a wear rate index and a wear acceleration index from the evolution state curve, inputting the wear rate index and the wear acceleration index into a preset multi-factor early warning threshold model, determining the ablation wear state of the barrel based on the multi-factor early warning threshold model, triggering an early warning mechanism when the ablation wear state exceeds a preset threshold, and generating a state assessment report including a predicted value of the ablation wear amount and an early warning level.

[0141] From the above description, it can be seen that the computer program product provided in the embodiment of the present application constructs a two-layer prediction architecture by combining an adaptive neuron growth mechanism and a dynamic programming algorithm. The ablation wear amount is identified by a hierarchical neural network trained with temperature, pressure, and charge characteristic parameters, and an energy state constraint function is constructed based on sensor data. The system uses an escape search optimization operator to select the optimal wear evolution path, and combines the wear rate and acceleration indicators to establish a multi-factor warning threshold model, which realizes accurate prediction and evaluation warning of the ablation wear state. This method effectively solves the shortcomings of traditional technology in model integration and status warning, and provides a reliable basis for artillery maintenance decisions.

[0142] It should be understood by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented 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.

[0143] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0144] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0145] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0146] The present invention uses specific embodiments to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for predicting barrel ablation wear based on a combined model, characterized in that: The method comprises: Collect historical data of ablation wear of the gun barrel, extract temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establish a hierarchical neural network model using an adaptive neuron growth mechanism, train the hierarchical neural network model with the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to obtain an ablation wear identifier, construct a dynamic programming algorithm model of ablation wear changing with time, and input the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; Collecting sensor data of the gun barrel during continuous firing, inputting the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, constructing an energy state constraint function based on the distribution parameters, applying evolution constraints to the dynamic programming algorithm model, substituting the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generating a set of wear evolution paths that meet the energy constraints, calculating the cumulative energy loss of each path in the wear evolution path set, and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path; The state vector of the optimal wear evolution path at different shooting moments is extracted, an evolution state curve characterizing the ablation wear of the barrel is established according to the state vector, a wear rate index and a wear acceleration index are extracted from the evolution state curve, the wear rate index and the wear acceleration index are input into a preset multi-factor early warning threshold model, the ablation wear state of the barrel is determined based on the multi-factor early warning threshold model, and an early warning mechanism is triggered when the ablation wear state exceeds a preset threshold, and a state assessment report including a predicted value of the ablation wear amount and an early warning level is generated.

2. The barrel ablation wear prediction method based on the combined model according to claim 1 is characterized in that: The method collects historical data of ablation and wear of the gun barrel, extracts temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establishes a hierarchical neural network model by using an adaptive neuron growth mechanism, trains the hierarchical neural network model with the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to obtain an ablation wear amount identifier, and constructs a dynamic programming algorithm model of ablation wear changing with time, including: Collecting historical data of ablation and wear of the gun barrel, performing denoising processing on the historical data based on the wavelet transform method, calculating the values ​​of the temperature sensor, the pressure sensor and the charge sensor at different times, converting the values ​​into temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters in combination with the sensor calibration curve, and performing normalization processing on the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters to generate a characteristic vector; A hierarchical neural network model is constructed using an adaptive neuron growth mechanism, the feature vector is input into the hierarchical neural network model, the hierarchical neural network model is trained based on a back-propagation algorithm to obtain an ablation wear identifier, an ablation wear state transfer matrix is ​​established using a Markov process, a dynamic programming algorithm model including the state transfer matrix is ​​constructed, and the output result of the ablation wear identifier is substituted into the dynamic programming algorithm model.

3. The barrel ablation wear prediction method based on the combined model according to claim 1 is characterized in that: The step of inputting the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model includes: The output result of the ablation wear identifier is divided into multiple time series intervals according to the number of shootings, and the ablation wear data in each time series interval is piecewise linearly fitted to calculate the wear change rate and wear acceleration in different intervals, and a state transition probability matrix reflecting the evolution characteristics of ablation wear is constructed based on the wear change rate and the wear acceleration; The state transfer probability matrix is ​​substituted into the dynamic programming algorithm model to construct a staged wear evolution model that includes the temperature accumulation effect, pressure accumulation effect and charge accumulation effect during the shooting process. Physical boundary constraints are imposed on the staged wear evolution model, and the Bellman equation is used to iteratively calculate the state transfer process of each stage.

4. The barrel ablation wear prediction method based on the combined model according to claim 1, characterized in that: The sensor data of the gun barrel during continuous firing is collected, the sensor data is input into the ablation wear identifier to obtain the distribution parameters of the ablation wear, an energy state constraint function is constructed based on the distribution parameters, an evolution constraint is imposed on the dynamic programming algorithm model, the dynamic programming algorithm model with the evolution constraint is substituted into the escape search operator for iterative optimization, and a set of wear evolution paths that meet the energy constraint is generated, including: Collecting data from temperature sensors, pressure sensors and charge sensors during continuous firing of the gun barrel, performing time domain sampling and feature extraction on the sensor data, inputting the extracted feature data into the ablation wear identifier, calculating the mean and variance distribution parameters of the ablation wear based on a Gaussian mixture model, and constructing an energy state constraint function including temperature energy dissipation, pressure energy loss and charge energy conversion using the distribution parameters; The energy state constraint function is substituted into the dynamic programming algorithm model to form a state transfer equation with evolution constraints, a global optimization operator based on the escape search mechanism is constructed, the state transfer equation is input into the global optimization operator for iterative optimization, and a Monte Carlo sampling method is used to generate multiple groups of initial solutions. Each group of initial solutions is subjected to escape search optimization to obtain a set of wear evolution paths that meet the energy constraints.

5. The barrel ablation wear prediction method based on the combined model according to claim 1, characterized in that: The calculating the cumulative energy loss of each path in the wear evolution path set and selecting the path with the smallest cumulative energy loss as the optimal wear evolution path includes: Calculating the temperature cumulative loss, pressure cumulative loss and charge cumulative loss of each path in the wear evolution path set during the continuous shooting process, substituting the temperature cumulative loss, pressure cumulative loss and charge cumulative loss into an energy weight fusion model, calculating the comprehensive energy loss index of each path based on the energy weight fusion model, and normalizing the comprehensive energy loss index; A path evaluation criterion based on the energy minimization principle is constructed, and the normalized comprehensive energy loss index is sorted using the path evaluation criterion. The path with the smallest comprehensive energy loss index is selected from the wear evolution path set as the optimal wear evolution path, and the state transition sequence of the optimal wear evolution path at different shooting stages is calculated.

6. The barrel ablation wear prediction method based on the combined model according to claim 1, characterized in that: The extracting of the state vector of the optimal wear evolution path at different shooting moments, establishing an evolution state curve representing barrel ablation wear according to the state vector, and extracting a wear rate index and a wear acceleration index from the evolution state curve include: Extracting the state vector at each shooting moment from the optimal wear evolution path, constructing a state space model including a temperature state component, a pressure state component and a charge state component based on the state vector, performing continuous processing on the state space model using a cubic spline interpolation algorithm, and mapping the interpolated state space trajectory to a time-wear amount coordinate system to obtain an evolution state curve; The evolution state curve is numerically differentiated, the first-order derivative value of the curve at each sampling point is calculated to obtain the wear rate index, the rate of change of the first-order derivative value is calculated to obtain the wear acceleration index, and the wear rate index and the wear acceleration index are smoothed using a sliding average filtering algorithm.

7. The barrel ablation wear prediction method based on the combined model according to claim 1, characterized in that: The wear rate index and the wear acceleration index are input into a preset multi-factor early warning threshold model, and the barrel ablation wear state is determined based on the multi-factor early warning threshold model. When the ablation wear state exceeds the preset threshold, an early warning mechanism is triggered, and a state assessment report including a predicted value of ablation wear and an early warning level is generated, including: The wear rate index and the wear acceleration index are normalized and input into a preset multi-factor early warning threshold model, the risk index of the current ablative wear state is calculated based on the multi-factor early warning threshold model, the risk index is mapped to a preset ablative wear grading standard by using a fuzzy comprehensive evaluation method, and the ablative wear state is determined by comparing the ablative wear grading standard with the preset threshold value; A warning trigger mechanism based on the remaining service life of the barrel is constructed. When the ablation and wear state exceeds a preset threshold, the prediction interval and confidence level of the ablation and wear amount of the barrel are calculated, the warning level is determined according to the prediction interval and the confidence level, and a status assessment report including the ablation and wear prediction value and the warning level is generated.

8. A device for predicting barrel ablation wear based on a combined model, characterized in that: The device comprises: A wear evolution module is used to collect historical data of ablation wear of the gun barrel, extract temperature characteristic parameters, pressure characteristic parameters and charge characteristic parameters from the historical data, establish a hierarchical neural network model using an adaptive neuron growth mechanism, train the hierarchical neural network model with the temperature characteristic parameters, the pressure characteristic parameters and the charge characteristic parameters as input variables to obtain an ablation wear identifier, construct a dynamic programming algorithm model of ablation wear changing with time, and input the output result of the ablation wear identifier into the dynamic programming algorithm model to establish a staged wear evolution model; A path determination module, used to collect sensor data when the gun barrel is continuously fired, input the sensor data into the ablation wear identifier to obtain the distribution parameters of the ablation wear, construct an energy state constraint function based on the distribution parameters, impose evolution constraints on the dynamic programming algorithm model, substitute the dynamic programming algorithm model with the evolution constraints into the escape search operator for iterative optimization, generate a set of wear evolution paths that meet the energy constraints, calculate the cumulative energy loss of each path in the wear evolution path set, and select the path with the smallest cumulative energy loss as the optimal wear evolution path; The state evaluation module is used to extract the state vector of the optimal wear evolution path at different shooting moments, establish an evolution state curve representing the ablation wear of the barrel according to the state vector, extract a wear rate index and a wear acceleration index from the evolution state curve, input the wear rate index and the wear acceleration index into a preset multi-factor early warning threshold model, determine the ablation wear state of the barrel based on the multi-factor early warning threshold model, trigger an early warning mechanism when the ablation wear state exceeds a preset threshold, and generate a state evaluation report including a predicted value of the ablation wear amount and an early warning level.

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, the steps of the method for predicting the ablation and wear amount of the barrel based on the combined model as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for predicting the ablation wear of a barrel based on a combined model as described in any one of claims 1 to 7 are implemented.

Images