A monitoring and prediction method and system for electromagnetic launch system

Abstract

The present invention relates to the field of measurement technology of electromagnetic field characteristics, and specifically to a monitoring prediction method and system for an electromagnetic launch system; comprising the following steps: S1, collecting monitoring data; S2, establishing a digital twin model; S3, building a topological mapping relationship; S4, training a prediction model; based on a machine learning model, training a prediction model and accessing the digital twin model; S5, model verification; S6, predicting the monitoring data based on the verified model; S7, if the predicted monitoring data exceeds the launch operation value, stop the launch and perform fault repair on the electromagnetic launch system. The present invention establishes a digital twin model and accesses a deep learning model, considers the coupling between monitoring parameters, and in addition to electromagnetic interference, also considers factors such as the integrity of the system structure, temperature, current and voltage, and sensitive equipment, and performs comprehensive and real-time monitoring and prediction of the data, so that the stability and accuracy of the electromagnetic launch system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of measuring electromagnetic field characteristics, and in particular to a monitoring and prediction method and system for an electromagnetic emission system. Background Art

[0002] The Electromagnetic Launch System (ELS) is an energy conversion system that uses electrical energy as energy to convert electromagnetic energy into kinetic energy. It is mainly composed of energy storage system, power change system, pulse launch system, control system and other parts. It is a large and complex system and is currently mainly used in the military and aerospace fields.

[0003] The electromagnetic launch system will be subject to various interferences during operation, including electromagnetic interference sources, sensitive equipment, environmental factors, system design and component integrity, etc. Minor faults will affect the accuracy of the electromagnetic launch system, so it is necessary to predict faults and manage its health. Monitoring the electromagnetic launch system is a key link to ensure its safe and efficient operation. The main monitoring quantities include: current and voltage, temperature, speed and acceleration, structural integrity, environmental factors (electromagnetic interference), etc. There are many monitoring quantities, and the correlation between the monitoring quantities will also affect the stability of the system operation.

[0004] Related technologies are disclosed in the prior art:

[0005] The Chinese patent with the publication number CN112069703A discloses a predictive optimization design method for an electromagnetic coil launch system; specifically, the following methods are disclosed: constructing an electromagnetic coil launcher sample, using the circuit model method or the finite element field-circuit coupling method to calculate the corresponding launch performance sample of the electromagnetic coil launcher sample, constructing an optimization target sample and a constraint target sample based on the launch performance sample, and constructing a prediction model training set through the electromagnetic coil launcher sample, the optimization target sample, and the constraint target sample; optimizing the support vector regression model training according to the prediction model training set, and constructing an electromagnetic coil parameter optimization model based on the electromagnetic coil performance prediction value output by the optimized support vector regression model, the optimization target value, and the predicted constraint target value, and further obtaining the optimized electromagnetic coil parameters through the intelligent optimization algorithm. The global optimization of the electromagnetic coil launch system parameters is achieved, which greatly improves the optimization efficiency of the electromagnetic coil launch system parameters.

[0006] The Chinese patent with the publication number CN116184088A discloses a method for fault detection of electromagnetic radiation emission system based on electromagnetic spectrum characteristics; specifically disclosed: including the following steps: S1. Determine the test frequency band and M common fault states of the electromagnetic radiation emission system, and test to obtain N groups of background environmental noise; S2. Under the mth fault state, test to obtain the radiation emission amplitude of N groups of electromagnetic radiation emission systems; S3. Extract the electromagnetic characteristics of N groups of emission spectrum under the mth fault state; S4. Obtain the emission electromagnetic spectrum of each group of fault states under M fault states; S5. Update the electromagnetic characteristics of N groups of emission spectrum under each fault state; S6. Construct and train a fault identification model; S7. Use a mature fault identification model to perform fault detection of the electromagnetic radiation emission system. The fault identification model is trained by using the relationship between the electromagnetic radiation emission characteristics of the equipment and the corresponding equipment, and is used for fault detection of the electromagnetic radiation emission system, which reduces the difficulty of fault detection and shortens the troubleshooting cycle.

[0007] However, the above prior art still has the following problems:

[0008] 1. The existing electromagnetic system monitoring method has a single monitoring quantity, that is, it does not consider the coupling between the monitoring parameters. It is difficult to effectively monitor the electromagnetic launch system of the load by relying on a single monitoring data. In addition to electromagnetic interference, factors such as the integrity of the system structure, temperature, current and voltage, and sensitive equipment will affect the electromagnetic launch system. If the data is not fully monitored, the electromagnetic launch system will have problems such as poor stability and low accuracy.

[0009] 2. The existing electromagnetic system monitoring and prediction methods obtain the optimal parameter design through optimization algorithms, but do not consider the impact of the electromagnetic launch system on the system itself during operation, and do not adjust the parameters in a timely manner. As a result, the optimization is prone to deviations. When performing fault prediction and health management on the electromagnetic launch system, the service life and frequency of use of the system should be taken into consideration, and the system should be corrected in a timely manner to avoid the problem of error accumulation.

[0010] 3. One of the factors affecting the electromagnetic launch system is electromagnetic interference. There are many reasons for electromagnetic interference, including environmental factors. The existing technology does not monitor and predict the system in real time, and there are problems such as untimely data updates.

[0011] Therefore, how to effectively monitor and predict the electromagnetic launch system in real time is a technical problem that needs to be solved urgently. Summary of the invention

[0012] To achieve the purpose of the present invention, the present invention is implemented by the following technical solution: a monitoring and prediction method for an electromagnetic emission system, comprising the following steps:

[0013] S1, collect monitoring data; analyze the influencing factors, determine the monitoring quantity and monitoring points, arrange collection sensors at the monitoring points, and standardize the monitoring quantity to form a monitoring data set P i ;

[0014] S2. Establish a digital twin model: Build a digital simulation model based on the system design, access the acquisition sensor interface, emission parameters and result parameters, and form a digital twin model;

[0015] S3, build topological mapping relationship; collect historical transmission data to form transmission data pool {R k {[P]}}, R k is the result parameter, including coordinates and elevation, [P] is the R k Monitoring data of the results, including multiple sets of monitoring data sets;

[0016] S4. Training the prediction model: Using the launch data pool as training data, based on the machine learning model, the prediction model is trained and connected to the digital twin model;

[0017] S5. Model verification; setting target parameter W k , the monitoring prediction result P is obtained through the prediction model get , the actual monitoring data set P is obtained through the digital twin model i+1 , calculate the prediction error ε:

[0018]

[0019] Among them, ε1, ε2, ε3, ε4, and ε5 are respectively the voltage, current, the influence of sensitive equipment, the electromagnetic interference vector, and the error value of the system stress state; if the error value is greater than the error threshold, the monitoring data corresponding to the maximum error value is screened out, and a new topological mapping relationship is formed after adjusting the monitoring data. The monitoring prediction result P is obtained through the prediction model. get+1 , until the error value is less than the error threshold;

[0020] S6. Predict the monitoring data based on the verified model;

[0021] S7. If the predicted monitoring data exceeds the launch operation value, the launch is stopped and the electromagnetic launch system is repaired.

[0022] Furthermore, the monitoring data set P in step S1 i as follows:

[0023] P i = {U i ,I i ,[S i (n)},B i ,Fi}

[0024] Among them, P i represents the monitoring data set with serial number i, i∈N, N is the total number of monitoring, which is a natural number set; U i and I i Represents current and voltage respectively; {S i (n)} represents the influence of sensitive equipment in the electromagnetic transmission system, n is the equipment number; B i F is the impact of electromagnetic interference, including natural factors; i It is the stress state of the system and characterizes the integrity of the main components of the system.

[0025] Furthermore, the stress state of the system is a stress-strain curve diagram or an impact fluctuation curve diagram.

[0026] Furthermore, the electromagnetic interference influence degree is calculated as follows:

[0027] S21. Determine the source of electromagnetic interference X in the surrounding environment env (m), including multiple interference sources;

[0028] S22. Obtain the current natural factor X nat (g), including atmospheric noise and solar noise;

[0029] S23. Calculate the impact of electromagnetic interference as follows:

[0030] B i =K m ×f(X env (m),X nat (g))

[0031] Among them, f(X env (m),X nat (g)) is the influence function, which is obtained by fitting multiple sets of data. m is the influence coefficient.

[0032] Furthermore, the influence of the sensitive device is calculated as follows:

[0033]

[0034] Among them, C n (E) represents the measurement result of the device numbered n when the magnetic field strength is E, C n (0) is the measurement result without magnetic field, and γ is the environmental adjustment factor.

[0035] Furthermore, the machine learning model in step S4 is LightGBM, and the learning steps are as follows:

[0036] S31, data preparation, using the emission data pool as training data;

[0037] S32, model construction, setting parameters such as learning rate, number of trees, maximum number of leaf nodes, etc.;

[0038] S33, model training, expanding the training data, and using the expanded training data to train the model;

[0039] S34, model evaluation and optimization, build a validation set, and adjust setting parameters to optimize the model.

[0040] Furthermore, in step S5, the monitoring data corresponding to the maximum error value is screened out, and a new topological mapping relationship is formed after the monitoring data is adjusted as follows:

[0041] S41, maximum screening error, ε max =MAX(ε1,ε2,ε3,ε4,ε5);

[0042] S42, replace monitoring data, if ε max =ε1, then P io = {U i +ΔU,I i ,{S i (n)},B i ,F i};

[0043] S43, construct the emission data pool data: {W k {P io}};

[0044] S44, integrating the constructed transmission data pool data into the transmission data pool to form a new topological mapping relationship.

[0045] Furthermore, the troubleshooting of the electromagnetic launch system in step S7 includes:

[0046] S51, directional inspection of the equipment corresponding to the monitoring data exceeding the transmission operation value;

[0047] S52. Synchronize maintenance data to the digital twin model.

[0048] Furthermore, the multiple sets of monitoring data sets in step S3 are data sets that have undergone similar data processing.

[0049] A monitoring and prediction system for an electromagnetic launch system, comprising:

[0050] Data acquisition module, used to obtain data obtained by the acquisition sensor and environmental factors;

[0051] Digital twin module, used to build digital twin models and display them visually;

[0052] Data analysis module, used to analyze the correlation of monitoring data and build topological mapping relationships;

[0053] Training module, used to train the prediction model;

[0054] The verification module is used to verify and optimize the prediction model.

[0055] The beneficial effects of the present invention are:

[0056] 1. The present invention establishes a digital twin model and connects it to a deep learning model. It considers the coupling between monitoring parameters. In addition to electromagnetic interference, it also considers factors such as the integrity of the system structure, temperature, current and voltage, and sensitive equipment. It monitors and predicts the data in a comprehensive and real-time manner, thereby improving the stability and accuracy of the electromagnetic launch system.

[0057] 2. The present invention takes into account the impact of the electromagnetic launch system on the system itself during operation, and based on the advantages of the digital twin model such as fast data update, it performs real-time predictive analysis on the monitoring results. When there is an error, it adjusts the data in the data pool to form a new topological mapping relationship, thereby continuously verifying and optimizing the model to avoid the problem of error accumulation.

[0058] 3. The present invention takes into account environmental interference sources and natural noise interference, supplements the monitoring data, and also takes into account the influence of sensitive equipment, thereby improving the accuracy of the prediction results. By predicting the monitoring data, it is possible to effectively discover possible faults in the system before launch, stop the launch in time, and avoid launch accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION

[0060] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with examples. The examples are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0061] Embodiment 1

[0062] according to Figure 1 As shown, this embodiment provides a monitoring and prediction method for an electromagnetic launch system, comprising the following steps:

[0063] S1, collect monitoring data; analyze the influencing factors, determine the monitoring quantity and monitoring points, arrange collection sensors at the monitoring points, and standardize the monitoring quantity to form a monitoring data set P i ;

[0064] S2. Establish a digital twin model: Build a digital simulation model based on the system design, access the acquisition sensor interface, emission parameters and result parameters, and form a digital twin model;

[0065] S3, build topological mapping relationship; collect historical transmission data to form transmission data pool {R k {[P]}}, R k is the result parameter, including coordinates and elevation, [P] is the result parameter k Monitoring data of the results, including multiple sets of monitoring data sets;

[0066] S4. Training the prediction model: Using the launch data pool as training data, based on the machine learning model, the prediction model is trained and connected to the digital twin model;

[0067] S5. Model verification; setting target parameter W k , the monitoring prediction result P is obtained through the prediction model get , the actual monitoring data set P is obtained through the digital twin model i+1 , calculate the prediction error ε:

[0068]

[0069] Among them, ε1, ε2, ε3, ε4, and ε5 are respectively the voltage, current, the influence of sensitive equipment, the electromagnetic interference vector, and the error value of the system stress state; if the error value is greater than the error threshold, the monitoring data corresponding to the maximum error value is screened out, and a new topological mapping relationship is formed after adjusting the monitoring data. The monitoring prediction result P is obtained through the prediction model. get+1 , until the error value is less than the error threshold;

[0070] S6. Predict the monitoring data based on the verified model;

[0071] S7. If the predicted monitoring data exceeds the launch operation value, the launch is stopped and the electromagnetic launch system is repaired.

[0072] More specifically, the monitoring data set P in step S1 i as follows:

[0073] P i = {U i ,I i ,{S i (n)},B i ,F i}

[0074] Among them, P i represents the monitoring data set with serial number i, i∈N, N is the total number of monitoring, which is a natural number set; U i and Ii Represents current and voltage respectively; {S i (n)} represents the influence of sensitive equipment in the electromagnetic transmission system, n is the equipment number; B i F is the impact of electromagnetic interference, including natural factors; i It is the stress state of the system and characterizes the integrity of the main components of the system.

[0075] More specifically, the stress state of the system is a stress-strain curve diagram or an impact fluctuation curve diagram.

[0076] More specifically, the electromagnetic interference impact degree is calculated as follows:

[0077] S21. Determine the source of electromagnetic interference X in the surrounding environment env (m), including multiple interference sources;

[0078] S22. Obtain the current natural factor X nat (g), including atmospheric noise and solar noise;

[0079] S23. Calculate the impact of electromagnetic interference as follows:

[0080] B i =K m ×f(X env (m),X nat (g))

[0081] Among them, f(X env (m),X nat (g)) is the influence function, which is obtained by fitting multiple sets of data. m is the influence coefficient.

[0082] More specifically, the influence of the sensitive device is calculated as follows:

[0083]

[0084] Among them, C n (E) represents the measurement result of the device numbered n when the magnetic field strength is E, C n (0) is the measurement result without magnetic field, and γ is the environmental adjustment factor.

[0085] More specifically, the machine learning model in step S4 is LightGBM, and the learning steps are as follows:

[0086] S31, data preparation, using the emission data pool as training data;

[0087] S32, model construction, setting parameters such as learning rate, number of trees, maximum number of leaf nodes, etc.;

[0088] S33, model training, expanding the training data, and using the expanded training data to train the model;

[0089] S34, model evaluation and optimization, build a validation set, and adjust setting parameters to optimize the model.

[0090] More specifically, the monitoring data corresponding to the maximum error value is screened out in step S5, and a new topological mapping relationship is formed after the monitoring data is adjusted as follows:

[0091] S41, maximum screening error, ε max =MAX(ε1,ε2,ε3,ε4,ε5);

[0092] S42, replace monitoring data, if ε max =ε1, then P io = {U i +ΔU,I i ,{S i (n)},B i ,F i};

[0093] S43, construct the emission data pool data: {W k {P io}};

[0094] S44, integrating the constructed transmission data pool data into the transmission data pool to form a new topological mapping relationship.

[0095] More specifically, the troubleshooting of the electromagnetic launch system in step S7 includes:

[0096] S51, directional inspection of the equipment corresponding to the monitoring data exceeding the transmission operation value;

[0097] S52. Synchronize maintenance data to the digital twin model.

[0098] More specifically, the multiple groups of monitoring data sets in step S3 are data sets that have undergone similar data processing.

[0099] Embodiment 2

[0100] The present invention also provides a monitoring and prediction system for an electromagnetic launch system, comprising: a data acquisition module for acquiring data obtained by acquisition sensors and environmental factors; a digital twin module for building a digital twin model and performing a visual display; a data analysis module for analyzing the correlation of monitoring data and constructing a topological mapping relationship; a training module for training a prediction model; and a verification module for verifying and optimizing the prediction model.

[0101] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A monitoring and prediction method for an electromagnetic launch system, characterized in that: The following steps are involved: S1. Collect monitoring data; analyze the influencing factors, determine the monitoring quantity and monitoring points, arrange collection sensors at the monitoring points, and standardize the monitoring quantity to form a monitoring data set , the monitoring data set as follows: ； in, Indicates the serial number is Monitoring data sets, , is the total number of monitoring, which is a set of natural numbers; and Represent current and voltage respectively; It indicates the influence of sensitive equipment in the electromagnetic transmission system, and n is the equipment number; The impact of electromagnetic interference, including natural factors; It is the stress state of the system and characterizes the integrity of the main components of the system; The stress state of the system is a stress-strain curve diagram or an impact fluctuation curve diagram; The electromagnetic interference influence degree is calculated as follows: S21. Determine the source of electromagnetic interference in the surrounding environment , including multiple interference sources; S22. Obtain the current natural factors , including atmospheric noise and solar noise; S23. Calculate the impact of electromagnetic interference as follows: ； in, is the influence function, which is obtained by fitting multiple sets of data. is the influence coefficient; The impact of the sensitive equipment is calculated as follows: ； in, Indicates the number The device is in a magnetic field strength of The measurement results when is the measurement result when there is no magnetic field, is the environmental adjustment factor; S2. Establish a digital twin model: Build a digital simulation model based on the system design, access the acquisition sensor interface, emission parameters and result parameters, and form a digital twin model; S3. Build topological mapping relationships; collect historical launch data to form a launch data pool , are the result parameters, including coordinates and elevations, To obtain Monitoring data of the results, including multiple sets of monitoring data sets; S4. Training the prediction model: Using the emission data pool as training data, based on the machine learning model, the prediction model is trained and connected to the digital twin model; the machine learning model is LightGBM, and the learning steps are as follows: S31, data preparation, using the emission data pool as training data; S32, model construction, setting parameters of learning rate, number of trees, and maximum number of leaf nodes; S33, model training, expanding the training data, and using the expanded training data to train the model; S34, model evaluation and optimization, build a validation set, and adjust setting parameters to optimize the model; S5. Model verification; setting target parameters , the monitoring prediction results are obtained through the prediction model , the actual monitoring data set is obtained through the digital twin model , calculate the prediction error : ； in, They are voltage, current, influence of sensitive equipment, electromagnetic interference vector and system stress state error value respectively; if the error value is greater than the error threshold, the monitoring data corresponding to the maximum error value is screened out, and a new topological mapping relationship is formed after adjusting the monitoring data, and the monitoring prediction result is obtained through the prediction model. , until the error value is less than the error threshold; The monitoring data corresponding to the maximum error value is screened out, and a new topological mapping relationship is formed after adjusting the monitoring data as follows: S41, screening maximum error, ; S42, replace monitoring data, if ,but ; S43, constructing the emission data pool data: ; S44, integrating the constructed transmission data pool data into the transmission data pool to form a new topological mapping relationship; S6. Predict the monitoring data based on the verified model; S7. If the predicted monitoring data exceeds the launch operation value, the launch is stopped and the electromagnetic launch system is repaired.

2. A monitoring and prediction method for an electromagnetic launch system according to claim 1, characterized in that: The troubleshooting of the electromagnetic launch system in step S7 includes: S51, directional inspection of the equipment corresponding to the monitoring data exceeding the transmission operation value; S52. Synchronize maintenance data to the digital twin model.

3. A monitoring and prediction method for an electromagnetic launch system according to claim 1, characterized in that: The multiple monitoring data sets in step S3 are data sets that have undergone similar data processing.

4. A monitoring and prediction system for an electromagnetic launch system, used to implement a monitoring and prediction method for an electromagnetic launch system as claimed in any one of claims 1 to 3, characterized in that: include: Data acquisition module, used to obtain data obtained by the acquisition sensor and environmental factors; Digital twin module, used to build digital twin models and display them visually; Data analysis module, used to analyze the correlation of monitoring data and build topological mapping relationships; Training module, used to train the prediction model; The verification module is used to verify and optimize the prediction model.

Images