Intelligent monitoring and fault diagnosis system in manufacturing process of aerospace core forgings

Abstract

The invention discloses an intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings, and particularly relates to the technical field of intelligent monitoring and fault diagnosis. Performing correlation analysis on the preprocessed process parameter data to calculate a cross correlation coefficient; performing nonlinear fluctuation analysis on the preprocessed process parameter data to calculate a nonlinear fluctuation coefficient; according to the calculation result of the cross correlation coefficient and the nonlinear fluctuation coefficient, the interaction between different process parameters is accurately evaluated, a multivariable cross fault diagnosis index is output in time, the single fluctuation of the process parameters is considered, and the complex interaction effect between multiple process links is also considered. A more accurate basis is provided for the evaluation of the fault hidden danger, and after the fault hidden danger is diagnosed, according to the multivariable cross fault diagnosis index, the process parameters are dynamically adjusted, and the most suitable process parameter adjustment strategy is automatically selected.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent monitoring and fault diagnosis, and more specifically, to an intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings. Background Art

[0002] Aerospace core forgings are key load-bearing components in aerospace structures. They are usually subjected to extreme working environments and high-intensity mechanical loads, so their manufacturing precision and quality requirements are extremely stringent. The forging manufacturing process involves multiple complex process links, such as heating, pressing, annealing, cooling, etc. These processes are intertwined, and the manufacturing parameters of each link (such as temperature, pressure, heating rate, cooling rate, etc.) have strong nonlinear and multivariable characteristics. Even a small process deviation may have a significant impact on the quality, performance and safety of the final product, and thus affect the reliability of the entire aerospace system. During the forging manufacturing process, the intelligent monitoring system collects process data in real time, provides real-time feedback and analyzes the status of different production links, and can promptly issue alarms or automatically adjust process parameters when abnormalities or potential failures occur, to ensure the stability of the production process and the quality of parts.

[0003] However, due to the complexity of the forging manufacturing process, especially under extreme operating conditions such as high temperature, high pressure, and high speed, how to accurately diagnose faults and optimize process parameters in a multivariable, nonlinear environment is still a major challenge facing intelligent monitoring systems. Many existing traditional monitoring technologies are difficult to handle such a large amount of data and complex process parameters. Therefore, a new type of intelligent monitoring and fault diagnosis system is urgently needed to cope with the changing manufacturing environment and precise quality requirements, and ensure that the manufacturing process of core aerospace forgings is always in the best control state. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] Intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings, including data sensing module, correlation analysis module, nonlinear fluctuation module, fault diagnosis module, and adaptive optimization module;

[0007] The data sensing module is used to collect the process parameters in the forging manufacturing process through sensors and pre-process the process parameter data;

[0008] The correlation analysis module is used to perform correlation analysis on the pre-processed process parameter data, evaluate the degree of cross-correlation between different process parameter data, and calculate the cross-correlation coefficient;

[0009] The nonlinear fluctuation module is used to perform nonlinear fluctuation analysis on the preprocessed process parameter data, evaluate the nonlinear fluctuation degree of the process parameter data, and calculate the nonlinear fluctuation coefficient;

[0010] The fault diagnosis module is used to construct a multivariate cross fault diagnosis model based on the cross correlation coefficient and the nonlinear fluctuation coefficient, output a multivariate cross fault diagnosis index, and evaluate potential fault hazards in multivariate cross scenarios;

[0011] The adaptive optimization module is used to adaptively optimize the process parameters in the forging manufacturing process according to the potential failure hazards in multi-variable cross-scenario.

[0012] In a preferred embodiment, the process parameters include but are not limited to heating temperature, heating time, heating environment gas concentration, applied pressure, molding time, mold temperature, annealing temperature, annealing time, cooling time, cooling medium, cooling temperature, and forging times.

[0013] In a preferred embodiment, the logic for obtaining the cross-correlation coefficient is as follows:

[0014] For each pair of process parameters X i and X j , X i and X j The numerical range of X is divided into g1 and g2 intervals respectively. i and X j The data space is divided into a grid grouping scheme of g1×g2, where g1 represents X i The number of groups, g2 represents X j The number of groups is divided into two groups based on the grid grouping scheme. i and X j The sample points are assigned to the grid according to their values; the process parameters X are calculated i and X j The mutual information of Where I(X i ,X j |G) represents the process parameter X i and X j The mutual information under the grid grouping scheme G, p(x i ,x j |G) represents the process parameter X i and X j The joint probability distribution under the grid grouping scheme G, p(x i|G) represents the process parameter X i The marginal probability distribution under the grid grouping scheme G, p(x j |G) represents the process parameter X j Marginal probability distribution under the grid grouping scheme G, x i Indicates process parameter X i The sample point, x j Indicates process parameter X j Sample points; each pair of process parameters X i and X j The cross-correlation coefficient JCG is obtained by adding the mutual information of

[0015] In a preferred embodiment, the logic for obtaining the nonlinear fluctuation coefficient is as follows:

[0016] Obtain the time series of process parameters, divide the time series of process parameters into several time windows according to different window sizes W, and calculate the cumulative deviation value LJ for each window size W. The expression is as follows Where RW represents the difference between the maximum and minimum values ​​in the time window, and SW represents the standard deviation of the process parameter data in the time window; logarithmic fitting is performed on the cumulative deviation values ​​of multiple window sizes W, and a straight line is fitted by the least squares method: log(LJ)=Hz*log(W), where Hz represents the time series dependence index; the nonlinear fluctuation coefficient NFC is calculated, and the expression is as follows: NFC=|Hz-0.5|.

[0017] In a preferred embodiment, a multivariate cross fault diagnosis model is constructed based on the cross correlation coefficient and the nonlinear fluctuation coefficient, and a multivariate cross fault diagnosis index DBL is output. The model is based on the following formula: In the formula, NFC n represents the nonlinear fluctuation coefficient calculated by the nth process parameter, n∈{1,2,...,N}, N is a positive integer, JCG represents the cross-correlation coefficient, a1 and a2 represent the preset proportional coefficients of the nonlinear fluctuation coefficient and the cross-correlation coefficient respectively, and a1 and a2 are both greater than 0.

[0018] In a preferred embodiment, the multivariate cross fault diagnosis index is compared with a preset multivariate cross fault diagnosis index threshold to identify potential fault hazards in a multivariate cross scenario, as follows:

[0019] If the multivariate cross fault diagnosis index is greater than the multivariate cross fault diagnosis index threshold, it indicates that there is a potential fault hazard in the forging manufacturing process, and a fault hazard signal is generated;

[0020] If the multivariate cross fault diagnosis index is less than or equal to the multivariate cross fault diagnosis index threshold, it means that the current forging manufacturing process is relatively stable, there is no significant potential fault, and there is no need to generate a potential fault signal.

[0021] In a preferred embodiment, when a potential fault signal is generated, the process parameters in the forging manufacturing process are adaptively optimized as follows:

[0022] Define the state space ZT of process parameters in forging manufacturing process t For: ZT t = {GY m}, where GY m represents the value of the mth process parameter at time t;

[0023] Define the action space DZ for process parameter adjustment t For: DZ t ={ΔGY m}, where ΔGY m Represents the adjustment value of the mth process parameter at time t;

[0024] Constructing a reward function based on a multivariate cross-fault diagnosis index: JL t =-e DBL , among which JL t represents the reward value at time t, DBL represents the multivariate cross fault diagnosis index;

[0025] The optimal strategy is approximated by updating the process parameter optimization function, which represents the state ZT t Next, take action DZ t When , the cumulative reward value obtained in the future is updated as follows:

[0026]

[0027] Where Q(ZT t ,DZ t ) represents the state ZT t Next take action DZ t The expected cumulative reward value of JL t represents the reward value at time t, α represents the learning rate, which is used to control the amplitude of the update, and γ represents the discount factor, which is used to measure the importance of future rewards. Indicates the optimal action DZ selected at time t+1 * The corresponding Q(ZT t ,DZ t );

[0028] By continuously selecting different actions to update the process parameter optimization function, when Q(ZT t ,DZt ) is greater than the cumulative reward threshold, the process parameters in the forging manufacturing process are adjusted and optimized according to the selected action.

[0029] Technical effects and advantages of the present invention:

[0030] The present invention collects process parameters in the forging manufacturing process through sensors, pre-processes the process parameter data, performs correlation analysis on the pre-processed process parameter data, evaluates the cross-correlation degree between different process parameter data, and calculates the cross-correlation coefficient; performs nonlinear fluctuation analysis on the pre-processed process parameter data, evaluates the nonlinear fluctuation degree of the process parameter data, and calculates the nonlinear fluctuation coefficient; accurately evaluates the interaction between different process parameters according to the calculation results of the cross-correlation coefficient and the nonlinear fluctuation coefficient, and timely outputs a multivariate cross fault diagnosis index, which not only considers the single fluctuation of the process parameter, but also comprehensively considers the complex interaction effects between multiple process links, providing a more accurate basis for the evaluation of potential fault hazards; after the potential fault hazards are diagnosed, the process parameters are dynamically adjusted according to the multivariate cross fault diagnosis index, and the most appropriate process parameter adjustment strategy is automatically selected, so that the manufacturing process can be quickly restored to the optimal control state, thereby reducing the probability of failure and improving the reliability of the production process and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0032] Figure 1 Flowchart of a system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] Embodiment: The present invention provides as Figure 1 The intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings shown includes a data sensing module, a correlation analysis module, a nonlinear fluctuation module, a fault diagnosis module, and an adaptive optimization module;

[0035] The data sensing module is used to collect the process parameters in the forging manufacturing process through sensors and pre-process the process parameter data;

[0036] The correlation analysis module is used to perform correlation analysis on the pre-processed process parameter data, evaluate the degree of cross-correlation between different process parameter data, and calculate the cross-correlation coefficient;

[0037] The nonlinear fluctuation module is used to perform nonlinear fluctuation analysis on the preprocessed process parameter data, evaluate the nonlinear fluctuation degree of the process parameter data, and calculate the nonlinear fluctuation coefficient;

[0038] The fault diagnosis module is used to construct a multivariate cross fault diagnosis model based on the cross correlation coefficient and the nonlinear fluctuation coefficient, output a multivariate cross fault diagnosis index, and evaluate potential fault hazards in multivariate cross scenarios;

[0039] Adaptive optimization module, which is used to adaptively optimize the process parameters in the forging manufacturing process according to the potential failure hazards in multi-variable cross-scenario;

[0040] The data sensing module is used to collect the process parameters in the forging manufacturing process through sensors and pre-process the process parameter data;

[0041] The process parameters include but are not limited to heating temperature, heating time, heating environment gas concentration, applied pressure, molding time, mold temperature, annealing temperature, annealing time, cooling time, cooling medium, cooling temperature, and forging times;

[0042] Heating temperature: affects the plasticity and microstructure of forgings. Too low a temperature will increase the brittleness of the forgings, while too high a temperature may lead to excessive oxidation or grain coarsening.

[0043] Heating time: Too long or too short a heating time will affect the uniformity and mechanical properties of the material. Appropriate heating time ensures uniform heating of the forgings to avoid local overheating or overcooling;

[0044] Heating environment gas concentration: Atmospheres such as oxygen, nitrogen or hydrogen will affect the surface oxidation degree and surface quality of forgings;

[0045] Applying pressure: During the forging process, the applied pressure is crucial. Too high or too low pressure will cause deviations in the shape and size of the forgings and affect their mechanical properties.

[0046] Forming time: The time of forging forming affects the uniformity of material flow. Too fast forming speed may cause uneven material flow, thus affecting the quality of forgings.

[0047] Die temperature: Die temperature is crucial to controlling the surface quality of forgings and reducing wear. Too high or too low die temperature will affect the friction and heat conduction efficiency during the forming process.

[0048] Annealing temperature: Annealing temperature affects the microstructure change of forgings. Too high annealing temperature may cause grain growth, while too low temperature cannot effectively eliminate stress.

[0049] Annealing time: Annealing time affects the annealing effect. Too short annealing time cannot effectively eliminate residual stress, while too long annealing time may cause grain growth and affect the strength of the material.

[0050] Cooling time: Cooling time is crucial to the structural transformation of materials, especially in metal materials such as steel. Cooling time directly affects the hardness, toughness and residual stress of forgings.

[0051] Cooling medium: Different cooling media (such as water, oil, air, etc.) will lead to different cooling effects, thus affecting the microstructure and final performance of the forging;

[0052] Cooling temperature: Temperature control during the cooling process helps optimize the structure of the forging and avoid cracks and deformation caused by uneven cooling;

[0053] Forging times: Multiple forgings can improve the mechanical properties and organizational structure of forgings and avoid material defects in the first forging process;

[0054] It should be noted that the above process parameters can be collected by installing various high-precision sensors during the forging manufacturing process, such as temperature sensors, pressure sensors, displacement sensors, etc. These sensors can collect process parameters in the forging manufacturing process in real time;

[0055] Preprocessing of process parameter data includes data cleaning, data alignment, and data normalization;

[0056] Data cleaning includes the following:

[0057] Remove outliers: Identify and remove abnormal data due to sensor failure or external interference through statistical analysis or rule-based algorithms.

[0058] Filling missing values: For missing data caused by temporary sensor failure or data transmission interruption, interpolation methods (such as linear interpolation, mean interpolation, etc.) can be used to fill in the missing values.

[0059] Smoothing: Use filtering algorithms (such as moving average filtering, Kalman filtering, etc.) to remove noise from the data and ensure the smoothness and continuity of the data.

[0060] Data alignment includes the following:

[0061] Time synchronization: Since the sampling frequencies of different sensors may be different, the data needs to be time-aligned to ensure that all parameters have a corresponding relationship at the same time point.

[0062] Data resampling: For sensor data with inconsistent sampling frequencies, they can be unified to the same sampling frequency through resampling (such as interpolation).

[0063] Data normalization includes the following:

[0064] Standardization: Standardize process parameters of different dimensions (such as temperature, pressure, time, etc.) so that they are within the same numerical range (such as between 0 and 1) to facilitate subsequent data analysis.

[0065] Normalization: According to specific needs, normalize the data to a specific range or distribution to reduce the impact of dimensional differences between different parameters on the analysis results.

[0066] The correlation analysis module is used to perform correlation analysis on the pre-processed process parameter data, evaluate the degree of cross-correlation between different process parameter data, and calculate the cross-correlation coefficient;

[0067] The cross-correlation coefficient in the present invention is used to measure the degree of cross-correlation between different process parameter data, especially the nonlinear coupling relationship between different process parameter data; in the manufacturing process of aerospace forgings, various process parameters such as temperature, pressure, heating rate and cooling rate often do not change independently, and there are complex interactions between them. Traditional single-variable fault diagnosis often ignores these complex interactions, while the cross-correlation coefficient can accurately capture the cross-coupling effect between these parameters. By calculating the cross-correlation coefficient between process parameter data, the potential mutual influence relationship between multiple process parameters can be revealed. The multivariate cross-fault diagnosis model constructed based on the cross-correlation coefficient can comprehensively consider the changes in each process parameter and the interactions between them, and accurately diagnose potential fault hazards by evaluating the cross-correlation coefficient. In the fault diagnosis process, the cross-correlation coefficient not only reflects the direct relationship between different process parameters, but also reveals the complex situation that some potential fault points may be caused by the combined action of multiple process factors. This model can more accurately capture multivariate cross-fault hazards that cannot be discovered by traditional methods.

[0068] Specifically, the multivariate cross-fault diagnosis model based on the cross-correlation coefficient has significant effects in the following aspects:

[0069] More accurate fault location: By analyzing the nonlinear coupling relationship between multiple process parameters, the model can locate the fault source more accurately and avoid misdiagnosis.

[0070] Integrated multi-dimensional data: Able to integrate data from different sensors and process parameters to reduce the errors that may be caused by single-dimensional data.

[0071] Early warning: Anomaly detection based on cross-correlation coefficients can capture potential faults at an early stage, respond in time, and reduce risks and losses in the production process.

[0072] In summary, the multivariable cross-fault diagnosis model constructed based on the cross-correlation coefficient can effectively improve the accuracy and real-time performance of the intelligent monitoring system, ensure that the manufacturing process of aerospace core forgings is always kept in the best control state, and help to ensure the quality of parts and the stability of the production process.

[0073] The logic for obtaining the cross-correlation coefficient is as follows:

[0074] For each pair of process parameters X i and X j , X i and X j The numerical range of X is divided into g1 and g2 intervals respectively. i and X j The data space is divided into a grid grouping scheme of g1×g2, where g1 represents X i The number of groups, g2 represents X j The number of groups is divided into two groups based on the grid grouping scheme. i and X j The sample points are assigned to the grid according to their values; the process parameters X are calculated i and X j The mutual information of Where I(X i ,X j |G) represents the process parameter X i and X j The mutual information under the grid grouping scheme G, p(x i ,x j |G) represents the process parameter X i and X j The joint probability distribution under the grid grouping scheme G, p(x i |G) represents the process parameter X i The marginal probability distribution under the grid grouping scheme G, p(x j |G) represents the process parameter X j Marginal probability distribution under the grid grouping scheme G, x i Indicates process parameter X i The sample point, x j Indicates process parameter X j Sample points; each pair of process parameters X i and X j The cross-correlation coefficient JCG is obtained by adding the mutual information of

[0075] It should be noted that the joint probability distribution refers to the process parameter X i and X j The probability of getting the same value is calculated by dividing the frequency of the same value by the total number of samples to get the joint probability distribution. The marginal probability distribution refers to the probability of a single variable. The frequency of a single variable is calculated by dividing the frequency of the single variable by the total number of samples to get the marginal probability distribution.

[0076] The larger the cross-correlation coefficient, the more accurate the assessment of potential fault hazards in multivariable cross scenarios, and the more likely it is to capture multivariable cross fault hazards that cannot be discovered by traditional methods; the smaller the cross-correlation coefficient, the lower the probability of potential fault hazards in multivariable cross scenarios, and the more likely it is to effectively reduce the probability of misjudgment of fault hazards;

[0077] The nonlinear fluctuation module is used to perform nonlinear fluctuation analysis on the preprocessed process parameter data, evaluate the nonlinear fluctuation degree of the process parameter data, and calculate the nonlinear fluctuation coefficient;

[0078] The nonlinear fluctuation coefficient in the present invention is used to measure the nonlinear fluctuation degree of process parameter data over time. In the manufacturing process of core aerospace forgings, process parameters (such as temperature, pressure, etc.) are usually affected by multiple factors, and these influences may be nonlinear, especially in extreme working environments. The nonlinear fluctuation coefficient can help identify these nonlinear fluctuations, and on this basis, a multivariate cross fault diagnosis model is constructed to evaluate potential fault hazards; through the multivariate cross fault diagnosis model constructed based on the nonlinear fluctuation coefficient, the following beneficial effects can be achieved in the manufacturing process:

[0079] Identify potential failures in advance: The model can detect potential failures early by monitoring the nonlinear fluctuations of various process parameters in real time. For example, when certain process parameters fluctuate abnormally under specific cross-scenario conditions, the system can issue an alarm in time to help operators take measures to prevent failures.

[0080] Optimize the production process: Through in-depth analysis of nonlinear fluctuations, the process parameters in the production process can be fine-tuned to optimize the production process, reduce energy waste, and improve product consistency and quality.

[0081] Enhanced production stability: Through accurate diagnosis and timely adjustment, the failure rate can be effectively reduced, the stability and efficiency of the production line can be improved, and the high reliability of aerospace core forgings in the production process can be ensured.

[0082] When the nonlinear fluctuation coefficient of the process parameters is large, it means that these parameters have undergone more drastic or unstable changes during the production process, which may indicate that there are irregularities or abnormal fluctuations in the process, thereby increasing the probability of failure. When the nonlinear fluctuation coefficient is small, it means that the process parameters are relatively stable during the production process, indicating that the process is more stable and the probability of hidden faults is lower;

[0083] The logic for obtaining the nonlinear volatility coefficient is as follows:

[0084] Obtain the time series of process parameters, divide the time series of process parameters into several time windows according to different window sizes W, and calculate the cumulative deviation value LJ for each window size W. The expression is as follows Where RW represents the difference between the maximum and minimum values ​​in the time window, and SW represents the standard deviation of the process parameter data in the time window; logarithmic fitting is performed on the cumulative deviation values ​​of multiple window sizes W, and a straight line is fitted by the least square method: log(LJ)=Hz*log(W), where Hz represents the time series dependency index; the nonlinear fluctuation coefficient NFC is calculated, and the expression is as follows: NFC=|Hz-0.5|;

[0085] It should be noted that the time series dependence index obtained by least squares fitting is in the range of [0,1]:

[0086] Hz>0.5: It indicates that the time series is persistent, which is manifested by the fact that the growth trend of the series remains consistent, that is, it has self-similarity (long-term dependence), indicating that the current process is relatively stable, but if abnormal fluctuations occur, they may persist;

[0087] Hz<0.5: indicates that the time series has anti-persistence, that is, the series shows reverse dependence, which means that if the past changes were upward, then there may be a downward trend in the future;

[0088] Hz = 0.5: indicates that the time series is a random walk, which means there is no long-term dependence and the future changes of the series are difficult to predict;

[0089] The fault diagnosis module is used to construct a multivariate cross fault diagnosis model based on the cross correlation coefficient and the nonlinear fluctuation coefficient, output a multivariate cross fault diagnosis index, and evaluate potential fault hazards in multivariate cross scenarios;

[0090] A multivariate cross fault diagnosis model is constructed based on the cross correlation coefficient and the nonlinear fluctuation coefficient, and the multivariate cross fault diagnosis index DBL is output. The formula based on the model is as follows: In the formula, NFC nrepresents the nonlinear fluctuation coefficient calculated by the nth process parameter, n∈{1,2,...,N}, N is a positive integer, JCG represents the cross-correlation coefficient, a1 and a2 represent the preset proportional coefficients of the nonlinear fluctuation coefficient and the cross-correlation coefficient respectively, and a1 and a2 are both greater than 0;

[0091] It should be noted that before constructing a multivariate cross fault diagnosis model, it is necessary to ensure that the cross correlation coefficient and the nonlinear fluctuation coefficient are normalized. Common normalization methods include Min-Max normalization and Z-Score standardization. a1 and a2 are set according to actual conditions. For example, the expert empowerment method is adopted, that is, experts in related fields are invited to determine the preset proportional coefficients of various indicators through professional opinion surveys and comprehensive evaluations.

[0092] It can be seen from the above calculation expression that the larger the cross-correlation coefficient and the larger the nonlinear fluctuation coefficient, the larger the multivariate cross fault diagnosis index, which means that the potential fault hidden danger in the forging manufacturing process is greater. On the contrary, the smaller the cross-correlation coefficient and the smaller the nonlinear fluctuation coefficient, the smaller the multivariate cross fault diagnosis index, which means that the process is stable and the fault hidden danger is smaller.

[0093] The present invention constructs a multivariate cross fault diagnosis model based on the cross correlation coefficient and the nonlinear fluctuation coefficient, which can effectively improve the stability and reliability of the aerospace core forging manufacturing process. Specifically, the beneficial effects of the model are reflected in the following aspects:

[0094] Early warning of potential faults: By comprehensively analyzing the cross-correlation coefficient and the nonlinear fluctuation coefficient, the complex coupling relationship between process parameters and the trend of nonlinear fluctuations can be identified in advance, which helps to timely discover potential faults in the production process. When the multivariate cross-fault diagnosis index is high, it indicates that the changes in process parameters show large nonlinear fluctuations, and the production system may be in an unstable state. At this time, preventive measures can be taken to reduce the probability of failure.

[0095] Process optimization and adaptive adjustment: Multivariate cross-fault diagnosis models can conduct in-depth analysis of the complex interactions between different process parameters and identify which parameters have a greater impact on failures in the production process.

[0096] Improve the accuracy and efficiency of the production process: By real-time monitoring of the cross-correlation coefficient and nonlinear fluctuation coefficient, the mutual influence between various parameters in the process can be accurately evaluated. By optimizing control, the parameter adjustment lag and misjudgment that may occur in traditional monitoring methods can be avoided, ensuring that the production process is always in the best control state, improving production efficiency and accuracy.

[0097] Enhance the fault tolerance of the system: By utilizing the nonlinear fluctuation coefficient and cross-correlation coefficient between the process parameters reflected in the model, potential process fluctuations and cross-faults can be identified, which enables the entire production system to better adapt to environmental changes and process adjustments, and enhance the system's fault tolerance and stability.

[0098] Improve quality consistency and product reliability: By comprehensively considering the nonlinear relationship and fluctuation characteristics between process parameters in multivariable cross-scenario, the process accuracy and stability of each link can be ensured, thereby improving the quality consistency and reliability of the final product, especially in high-precision fields such as aerospace, and reducing the quality failure rate caused by process fluctuations.

[0099] In summary, the combination of the cross-correlation coefficient and the nonlinear fluctuation coefficient provides a more accurate quantitative basis for fault diagnosis and process optimization, making the manufacturing process more intelligent and efficient, significantly reducing the risk of failure, and improving the controllability of production and product quality.

[0100] The multivariate cross fault diagnosis index is compared with the preset multivariate cross fault diagnosis index threshold to identify potential fault hazards in the multivariate cross scenario, as follows:

[0101] If the multivariate cross fault diagnosis index is greater than the multivariate cross fault diagnosis index threshold, it means that there are potential fault hazards in the forging manufacturing process, and corresponding measures need to be taken immediately to intervene and deal with them, generating a fault hazard signal;

[0102] If the multivariate cross fault diagnosis index is less than or equal to the multivariate cross fault diagnosis index threshold, it means that the current forging manufacturing process is relatively stable and there is no significant potential fault hazard. It is sufficient to continue with the normal operation without generating a potential fault signal.

[0103] Adaptive optimization module, which is used to adaptively optimize the process parameters in the forging manufacturing process according to the potential failure hazards in multi-variable cross-scenario;

[0104] When a potential fault signal is generated, the process parameters in the forging manufacturing process are adaptively optimized as follows:

[0105] Define the state space ZT of process parameters in forging manufacturing process t For: ZT t = {GY m}, where GY m represents the value of the mth process parameter at time t;

[0106] Define the action space DZ for process parameter adjustment t For: DZ t ={ΔGY m}, where ΔGY m Represents the adjustment value of the mth process parameter at time t;

[0107] Constructing a reward function based on a multivariate cross-fault diagnosis index: JL t =-e DBL , among which JL t represents the reward value at time t, DBL represents the multivariate cross fault diagnosis index;

[0108] The optimal strategy is approximated by updating the process parameter optimization function, which represents the state ZT t Next, take action DZ t When , the cumulative reward value obtained in the future is updated as follows:

[0109]

[0110] Where Q(ZT t ,DZ t ) represents the state ZT t Next take action DZ t The expected cumulative reward value of JL t represents the reward value at time t, α represents the learning rate, which is used to control the amplitude of the update, and γ represents the discount factor, which is used to measure the importance of future rewards. Indicates the optimal action DZ selected at time t+1 * The corresponding Q(ZT t ,DZ t );

[0111] By continuously selecting different actions to update the process parameter optimization function, when Q(ZT t ,DZ t ) is greater than the cumulative reward threshold, the process parameters in the forging manufacturing process are adjusted and optimized according to the selected action;

[0112] The present invention collects process parameters in the forging manufacturing process through sensors, pre-processes the process parameter data, performs correlation analysis on the pre-processed process parameter data, evaluates the cross-correlation degree between different process parameter data, and calculates the cross-correlation coefficient; performs nonlinear fluctuation analysis on the pre-processed process parameter data, evaluates the nonlinear fluctuation degree of the process parameter data, and calculates the nonlinear fluctuation coefficient; accurately evaluates the interaction between different process parameters according to the calculation results of the cross-correlation coefficient and the nonlinear fluctuation coefficient, and timely outputs a multivariate cross fault diagnosis index, which not only considers the single fluctuation of the process parameter, but also comprehensively considers the complex interaction effects between multiple process links, providing a more accurate basis for the evaluation of potential fault hazards; after the potential fault hazards are diagnosed, the process parameters are dynamically adjusted according to the multivariate cross fault diagnosis index, and the most appropriate process parameter adjustment strategy is automatically selected, so that the manufacturing process can be quickly restored to the optimal control state, thereby reducing the probability of failure and improving the reliability of the production process and product quality.

[0113] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0114] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0115] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0116] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0117] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. Intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings, characterized by: It includes data sensing module, correlation analysis module, nonlinear fluctuation module, fault diagnosis module and adaptive optimization module; The data sensing module is used to collect the process parameters in the forging manufacturing process through sensors and pre-process the process parameter data; The correlation analysis module is used to perform correlation analysis on the pre-processed process parameter data, evaluate the degree of cross-correlation between different process parameter data, and calculate the cross-correlation coefficient; The nonlinear fluctuation module is used to perform nonlinear fluctuation analysis on the preprocessed process parameter data, evaluate the nonlinear fluctuation degree of the process parameter data, and calculate the nonlinear fluctuation coefficient; The fault diagnosis module is used to construct a multivariate cross fault diagnosis model based on the cross correlation coefficient and the nonlinear fluctuation coefficient, output a multivariate cross fault diagnosis index, and evaluate potential fault hazards in multivariate cross scenarios; The adaptive optimization module is used to adaptively optimize the process parameters in the forging manufacturing process according to the potential failure hazards in multi-variable cross-scenario.

2. The intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings according to claim 1 is characterized in that: The process parameters include but are not limited to heating temperature, heating time, heating environment gas concentration, applied pressure, molding time, mold temperature, annealing temperature, annealing time, cooling time, cooling medium, cooling temperature, and forging times.

3. The intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings according to claim 1, characterized in that: The logic for obtaining the cross-correlation coefficient is as follows: For each pair of process parameters X i and X j , X i and X j The numerical range of X is divided into g1 and g2 intervals respectively. i and X j The data space is divided into a grid grouping scheme of g1×g2, where g1 represents X i The number of groups, g2 represents X j The number of groups is divided into two groups based on the grid grouping scheme. i and X j The sample points of are distributed into the grid according to their values; Calculate process parameters X i and X j The mutual information of Where I(X i ,X j |G) represents the process parameter X i and X j The mutual information under the grid grouping scheme G, p(x i ,x j |G) represents the process parameter X i and X j The joint probability distribution under the grid grouping scheme G, p(x i |G) represents the process parameter X i The marginal probability distribution under the grid grouping scheme G, p(x j |G) represents the process parameter X j Marginal probability distribution under the grid grouping scheme G, x i Indicates process parameter X i The sample point, x j Indicates process parameter X j Sample points; each pair of process parameters X i and X j The cross-correlation coefficient JCG is obtained by adding the mutual information of 4. The intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings according to claim 1, characterized in that: The logic for obtaining the nonlinear volatility coefficient is as follows: Obtain the time series of process parameters, divide the time series of process parameters into several time windows according to different window sizes W, and calculate the cumulative deviation value LJ for each window size W. The expression is as follows Where RW represents the difference between the maximum and minimum values ​​in the time window, and SW represents the standard deviation of the process parameter data in the time window; logarithmic fitting is performed on the cumulative deviation values ​​of multiple window sizes W, and a straight line is fitted by the least squares method: log(LJ)=Hz*log(W), where Hz represents the time series dependence index; the nonlinear fluctuation coefficient NFC is calculated, and the expression is as follows: NFC=|Hz-0.5|.

5. The intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings according to claim 1, characterized in that: A multivariate cross fault diagnosis model is constructed based on the cross correlation coefficient and the nonlinear fluctuation coefficient, and the multivariate cross fault diagnosis index DBL is output. The formula based on the model is as follows: In the formula, NFC n represents the nonlinear fluctuation coefficient calculated by the nth process parameter, n∈{1,2,...,N}, N is a positive integer, JCG represents the cross-correlation coefficient, a1 and a2 represent the preset proportional coefficients of the nonlinear fluctuation coefficient and the cross-correlation coefficient respectively, and a1 and a2 are both greater than 0.

6. The intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings according to claim 5, characterized in that: The multivariate cross fault diagnosis index is compared with the preset multivariate cross fault diagnosis index threshold to identify potential fault hazards in the multivariate cross scenario, as follows: If the multivariate cross fault diagnosis index is greater than the multivariate cross fault diagnosis index threshold, it indicates that there is a potential fault hazard in the forging manufacturing process, and a fault hazard signal is generated; If the multivariate cross fault diagnosis index is less than or equal to the multivariate cross fault diagnosis index threshold, it means that the current forging manufacturing process is relatively stable, there is no significant potential fault, and there is no need to generate a potential fault signal.

7. The intelligent monitoring and fault diagnosis system in the manufacturing process of aerospace core forgings according to claim 6, characterized in that: When a potential fault signal is generated, the process parameters in the forging manufacturing process are adaptively optimized as follows: Define the state space ZT of process parameters in forging manufacturing process t For: ZT t = {GY m }, where GY m represents the value of the mth process parameter at time t; Define the action space DZ for process parameter adjustment t For: DZ t ={ΔGY m }, where ΔGY m Represents the adjustment value of the mth process parameter at time t; Constructing a reward function based on a multivariate cross-fault diagnosis index: JL t =-e DBL , among which JL t represents the reward value at time t, DBL represents the multivariate cross fault diagnosis index; The optimal strategy is approximated by updating the process parameter optimization function, which represents the state ZT t Next, take action DZ t When , the cumulative reward value obtained in the future is updated as follows: Where Q(ZT t ,DZ t ) represents the state ZT t Next take action DZ t The expected cumulative reward value of JL t represents the reward value at time t, α represents the learning rate, which is used to control the amplitude of the update, and γ represents the discount factor, which is used to measure the importance of future rewards. Indicates the optimal action DZ selected at time t+1 * The corresponding Q(ZT t ,DZ t ); By continuously selecting different actions to update the process parameter optimization function, when Q(ZT t ,DZ t ) is greater than the cumulative reward threshold, the process parameters in the forging manufacturing process are adjusted and optimized according to the selected action.

Images