Beidou satellite autonomous communication method based on ship power station abnormal event triggering

Abstract

The invention discloses a Beidou satellite autonomous communication method based on ship power station abnormal event triggering, and the method is characterized in that a novel LightGBM processing equation is designed, a correction coefficient employs a three-segment function, the nonlinear features of a model are effectively increased, and the over-fitting phenomenon is avoided; a novel particle swarm optimization algorithm speed equation is designed, a complex nonlinear maximum-minimum coefficient item is found through introduction, the global optimization ability is improved, and then the optimal wavelet packet denoising layer number is determined to calculate the abnormal degree of ship power station data; a heat energy collaborative generative adversarial network method is provided, three-phase voltage and current of a ship are combined, the heat energy condition of the three-phase voltage and current of a ship power station is fully considered, the data collaborative relation of the voltage and current is effectively mined, and whether Beidou satellite short message communication between the ship and the shore is triggered or not is determined in a collaborative generative adversarial mode. According to the method, the problem that Beidou satellite short message communication is limited in frequency and non-autonomous is solved.

Description

Technical Field

[0001] The present invention relates to a Beidou satellite autonomous communication method, and in particular to a Beidou satellite autonomous communication method based on triggering of abnormal events in a ship power station. Background Art

[0002] The BeiDou satellite developed by China has been included in the International Search and Rescue Satellite Organization in 2022, which means that on the basis of maintaining high-precision positioning, the unique short message communication function of BeiDou satellite has been recognized by powerful countries such as the United States, Russia and the European Union. Nowadays, BeiDou satellites have been widely used in aerospace ships, vehicle traffic and geological monitoring, and it is urgent to establish an information perception system for both ship and shore based on BeiDou satellite short message communication. However, considering the particularity of information interaction between ocean-going ships and remote ground command centers, BeiDou satellite short message communication has the problem of limited communication frequency. When ocean-going ships need to communicate with remote ground command centers, the existing communication method is a time-triggered mechanism, that is, non-autonomous data transmission at a fixed communication frequency, which will further increase the bandwidth burden of BeiDou satellite short message communication. Therefore, in order to ensure the efficient application of BeiDou satellites in the information situation awareness system composed of both ship and shore, it is necessary to design a BeiDou satellite autonomous communication method based on the triggering of abnormal events in ship power stations to enhance the performance of my country's BeiDou satellites in the field of ships.

[0003] The information situational awareness system composed of both ship and shore is designed to transmit critical shipborne data to the remote ground command center. The critical data judgment involved in this process is reflected in the information situational awareness of abnormal events in the ship power station. During a single voyage, an ocean-going ship will experience multiple acceleration / deceleration operations of the generator set, as well as the start / shutdown of the auxiliary combined unit. The different working conditions of the power unit will cause fluctuations in the voltage and current temperature of the ship power station. The frequency and amplitude of the fluctuation can be effectively obtained by the local monitoring device. In addition, when a ship power unit or power system fails, it is also accompanied by random abnormal fluctuations in the voltage and current temperature curves of the ship power station, which can be effectively obtained in real time by shipborne sensors and situational awareness devices. The existing Beidou satellite short message communication method adopts a time trigger mechanism, which means that all shipborne critical data needs to be transmitted to the remote shore command center within each Beidou communication frequency. Considering that the amount and frequency of Beidou satellite single transmission data are limited, the remote shore command platform will have poor real-time situational awareness of the ship. The time trigger mechanism is no longer applicable to the information situational awareness system composed of both ship and shore, which is an unresolved problem at this stage. Summary of the invention

[0004] The present invention provides a Beidou satellite autonomous communication method based on triggering of abnormal events in ship power stations. The method can be applied to an information situation awareness system composed of both ship and shore parties to ensure the efficient operation of the situation awareness system and solve the problem of limited and non-autonomous frequency of Beidou satellite short message communication. Different from the Beidou satellite-based ship monitoring system, the method of the present invention abandons the original time-triggered communication mechanism and proposes a situation awareness algorithm based on the Thermal Energy Synergistic Generative Adversarial Network (TeSGAN). The degree of abnormality is determined by calculating the error between the generated data and the original data in different time periods, and based on this, Beidou satellite short message communication is triggered. It is an innovative autonomous communication method.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] A Beidou satellite autonomous communication system based on abnormal event triggering of a ship power station includes a ship power station three-phase voltage and current acquisition module, a ship power station three-phase voltage and current denoising preprocessing module, a ship power station three-phase voltage and current thermal energy input value module, a thermal energy collaborative generative adversarial network (TeSGAN) module, a thermal energy anomaly degree module for calculating multivariate data of a ship power station, a ship power station thermal energy anomaly degree reaching a judgment threshold module, and a Beidou satellite short message communication module for triggering ship-shore communication, wherein:

[0007] The ship power station three-phase voltage and current acquisition module is responsible for real-time monitoring of the ship power station three-phase voltage and current data waveform, and segmenting the data at a fixed period;

[0008] The ship power station three-phase voltage and current denoising preprocessing module is responsible for denoising the ship power station three-phase voltage and current collected by the ship power station three-phase voltage and current collection module, removing the noise in the voltage and current waveforms, and inputting the denoised data into the ship power station three-phase voltage and current thermal energy input value module;

[0009] The three-phase voltage and current thermal energy input value module of the ship power station is responsible for connecting with the local database of the ship power station, obtaining the three-phase voltage and current thermal energy input values ​​from the local database, and at the same time, pre-processing the lost data and using the piecewise linear interpolation method to fill in the lost data values;

[0010] The Thermal Synergistic Generative Adversarial Network (TeSGAN) module is responsible for mining the key data segments that trigger Beidou satellite communications;

[0011] The module for calculating the thermal energy anomaly degree of the multivariate data of the ship power station is responsible for calculating the thermal energy anomaly degree of the multivariate data of the ship power station, including the fusion anomaly degree of voltage and current;

[0012] The module for determining whether the abnormal degree of thermal energy of the ship power station reaches the judgment threshold is responsible for analyzing whether the abnormal degree of thermal energy of the ship power station reaches the set judgment threshold based on the abnormal degree of thermal energy of the multivariate data of the ship power station. If the set judgment threshold is not reached, a new section of three-phase voltage and current values ​​of the ship power station is reselected to perform a new round of calculation; if the set judgment threshold is reached, the Beidou satellite short message communication between the ship and the shore is triggered;

[0013] The module for triggering Beidou satellite short message communication between ship and shore is responsible for triggering Beidou satellite short message communication between ship and shore.

[0014] A method for realizing autonomous Beidou satellite communication based on triggering of abnormal events of ship power stations using the above system comprises the following steps:

[0015] Step 1: Use the ship power station three-phase voltage and current acquisition module to monitor the three-phase voltage and current data waveform of the ship power station in real time, and process the data in segments at a fixed period, where: LightGBM is used to classify the data, and the LightGBM processing equation is as follows:

[0016]

[0017] Among them, F (t) is the processing target at the current moment, λ i is the coefficient term, is the loss expression, l represents the loss function, y i is the output value of this module, is the estimated value of this module at time t-1, n is the data length of three-phase voltage and current, f t represents the tree, Ω(f t ) represents the regularization term of the tree, ε is a fixed correction parameter, and f t (x i ) is a new tree, The meaning is: the current moment prediction value is composed of the previous moment prediction value and the new tree;

[0018] Step 2: Input the data processed in step 1 into the three-phase voltage and current denoising preprocessing module of the ship power station, perform denoising preprocessing on the three-phase voltage and current of the ship power station, remove the noise in the voltage and current waveforms, and use the denoised data as the input value of the three-phase voltage and current thermal energy input value module of the ship power station, wherein: the wavelet packet denoising method is used to denoise the three-phase voltage and current of the ship power station, and the particle swarm optimization algorithm is designed to determine the number of wavelet packet layers. The specific equation is as follows:

[0019]

[0020] in, is the particle update speed at time φ+1, is the particle update speed at time φ, r 1 ∈(0,1),r 2 ∈(0,1) are two different random numbers, v id is the velocity value, φ is the discretized time representation, ε 1 ,ε 2 is the weight coefficient, and is an extreme value, is the actual position of the particle at time φ, ω max and ω min is the balance coefficient, K is the iterative maximum value, and Δ is the fixed preset value;

[0021] Step 3: Input the denoised data from step 2 into the three-phase voltage and current thermal energy input value module of the ship power station, pre-process the missing data, and use the piecewise linear interpolation method to fill in the missing data values;

[0022] Step 4: Input the data processed in step 3 into the Thermal Energy Synergistic Generative Adversarial Network (TeSGAN) module to generate voltage and current thermal energy data, where:

[0023] The thermal energy collaborative generation adversarial network (TeSGAN) module includes a voltage thermal energy input module, a current thermal energy input module, a voltage thermal energy generator module, a current thermal energy generator module, a voltage thermal energy output module, a current thermal energy output module, a voltage thermal energy discriminator module, and a current thermal energy discriminator module. The output values ​​of the voltage thermal energy input module and the current thermal energy input module are represented by x and y respectively; the output value of the voltage thermal energy generator module is represented by G; the output value of the current thermal energy generator module is represented by F; G1 and G2 are the output values ​​of the voltage thermal energy generator module when the voltage and current are used as inputs respectively; F1 is the input value fed back from the voltage thermal energy discriminator module to the current thermal energy generator module, and F2 is the input value fed back from the current thermal energy discriminator module to the voltage thermal energy generator module; D 1 ,D 2 They are the discrimination results of the voltage thermal energy discriminator module and the current thermal energy discriminator module, which specifically include the following processing:

[0024] L1=Σ x [logG2(x)]+Σ y [log(2-G1(F(y)))]+Σ y [logG1(y)]+Σ x [log(1-G2(G1(x)))]

[0025] Among them, L1 is the loss value of the generative adversarial network model without considering thermal energy synergy;

[0026] L2=Σ x [||xF(G2(x)|| 2 ]+∑ y [||yG(F1(y)|| 2 ]

[0027] Wherein, L2 is the heat energy synergistic part added by the present invention;

[0028] L2′=Σ x [||xF(G1(x)|| 2 ]+Σ y [||yG(F2(y)|| 2 ]

[0029] Wherein, L2′ is the heat energy synergistic part added by the present invention;

[0030] L(G,F,D 1 ,D 2 )=γL1+μL2+εL2′

[0031] Among them, L(G,F,D 1 ,D 2 ) is the loss function of the Thermal Synergistic Generative Adversarial Network (TeSGAN) module, which consists of three parts: L1, L2, and L2′. The three coefficients γ, μ, and ε are used as parameters to balance each sub-part;

[0032]

[0033] Among them, G * ,F * It is the final optimization objective function of the Thermal Energy Synergistic Generative Adversarial Network (TeSGAN) module, which means that the thermal energy synergistic generators reach an ideal Nash equilibrium state;

[0034] Step 5: input the data processed in step 4 into a module for calculating the thermal energy anomaly degree of multivariate data of a ship power station, and calculate the thermal energy anomaly degree of the multivariate data of the ship power station, including the fusion anomaly degree of voltage and current;

[0035] Step 6: Input the data processed in step 5 into the judgment threshold module of the thermal energy anomaly degree of the ship power station, and analyze whether the thermal energy anomaly degree of the ship power station reaches the judgment threshold based on the thermal energy anomaly degree. If the thermal energy anomaly degree is greater than the judgment threshold, the Beidou satellite short message communication module between the ship and the shore is triggered, and a single communication from the ship to the shore command platform is executed. However, if the thermal energy anomaly degree is less than the judgment threshold, the Beidou satellite short message communication module between the ship and the shore cannot be triggered, and it directly returns to the three-phase voltage and current module of the ship power station, and reselects a new section of the three-phase voltage and current value of the ship power station, and performs a new round of calculation until all the three-phase voltage and current thermal energy data to be processed are completely analyzed, then the whole process is terminated, and finally the Beidou satellite autonomous communication triggered by the abnormal event of the ship power station is realized.

[0036] Compared with the prior art, the present invention has the following advantages:

[0037] 1. The present invention can avoid the existing time trigger mechanism and adopt a trigger mechanism based on abnormal events of ship power stations to realize the autonomous communication of Beidou satellites.

[0038] 2. The present invention designs a new LightGBM processing equation, and the correction coefficient adopts a three-segment function, which effectively increases the nonlinear characteristics of the model and avoids the occurrence of overfitting.

[0039] 3. The present invention designs a novel particle swarm optimization algorithm velocity equation, which improves the global optimization capability by introducing a method to find complex nonlinear maximum-minimum coefficient terms, and further determines the optimal number of wavelet packet denoising layers.

[0040] 4. In order to calculate the abnormality of ship power station data, the present invention proposes a thermal-energy synergistic generative adversarial network (TeSGAN) method, which combines the three-phase voltage and current of the ship, fully considers the thermal energy situation of the three-phase voltage and current of the ship power station, and effectively mines the data synergistic relationship of voltage and current, and uses the synergistic generative adversarial mode to decide whether to trigger the Beidou satellite short message communication between the ship and the shore. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is the overall structure diagram of the Beidou satellite autonomous communication method based on the triggering of abnormal events in the ship power station;

[0042] Figure 2 The overall structure diagram of the Thermal Synergistic Generative Adversarial Network (TeSGAN) module;

[0043] Figure 3 This is the waveform diagram of the three-phase voltage data of the ship power station;

[0044] Figure 4 This is the waveform diagram of the three-phase current data of the ship power station;

[0045] Figure 5 This is the abnormal identification result of three-phase voltage data of ship power station based on TeSGAN;

[0046] Figure 6 This is the abnormal identification result of three-phase current data of ship power station based on TeSGAN;

[0047] Figure 7 This is a comparison chart of Beidou satellite autonomous communication performance triggered by abnormal events in ship power stations. DETAILED DESCRIPTION

[0048] The technical solution of the present invention is further described below in conjunction with the accompanying drawings, but is not limited thereto. Any modification or equivalent replacement of the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention should be included in the protection scope of the present invention.

[0049] The present invention provides a Beidou satellite autonomous communication system based on triggering of abnormal events of ship power stations, such as Figure 1 As shown, the system mainly includes the following modules: a three-phase voltage and current acquisition module for a ship power station, a three-phase voltage and current denoising preprocessing module for a ship power station, a three-phase voltage and current thermal energy input value module for a ship power station, a thermal energy collaborative generative adversarial network (TeSGAN) module, a module for calculating the thermal energy anomaly degree of multivariate data of a ship power station, a module for determining the thermal energy anomaly degree of a ship power station reaches a judgment threshold, and a module for triggering Beidou satellite short message communication between the ship and the shore. The above modules are integrated and linked in a hierarchical and feedback manner, and each time a unit fixed-length data segment is used as the basis to realize the judgment of the abnormality degree of each data segment, thereby completely realizing the Beidou satellite autonomous communication triggered by abnormal events in the ship power station. Among them:

[0050] The ship power station three-phase voltage and current acquisition module is responsible for real-time monitoring of the ship power station three-phase voltage and current data waveform, and segmenting the data at a fixed period;

[0051] The ship power station three-phase voltage and current denoising preprocessing module is responsible for denoising the ship power station three-phase voltage and current collected by the ship power station three-phase voltage and current collection module, removing the noise in the voltage and current waveforms, and inputting the denoised data into the ship power station three-phase voltage and current thermal energy input value module;

[0052] The three-phase voltage and current thermal energy input value module of the ship power station is responsible for connecting with the local database of the ship power station and obtaining the three-phase voltage and current thermal energy input values ​​from the local database. At the same time, unlike the common input module, the three-phase voltage and current thermal energy input value module of the ship power station needs to pre-process the lost data and use the piecewise linear interpolation method to complete the lost data values;

[0053] The thermal collaborative generative adversarial network (TeSGAN) module is responsible for mining the key data segments that trigger Beidou satellite communications. This is because piecewise linear interpolation has barriers in dealing with nonlinear problems. The TeSGAN proposed in the present invention has very strong nonlinear characteristics and can effectively mine the key features in the ship power station data, providing accurate input values ​​for the next module.

[0054] The module for calculating the thermal energy anomaly degree of the multivariate data of the ship power station is responsible for calculating the thermal energy anomaly degree of the multivariate data of the ship power station, including the fusion anomaly degree of voltage and current;

[0055] The module for determining whether the abnormal degree of thermal energy of the ship power station reaches the judgment threshold is responsible for analyzing whether the abnormal degree of thermal energy of the ship power station reaches the set judgment threshold based on the abnormal degree of thermal energy of the multivariate data of the ship power station. If the set judgment threshold is not reached, a new section of three-phase voltage and current values ​​of the ship power station is reselected to perform a new round of calculation; if the set judgment threshold is reached, the Beidou satellite short message communication between the ship and the shore is triggered;

[0056] The module for triggering Beidou satellite short message communication between ship and shore is responsible for triggering Beidou satellite short message communication between ship and shore.

[0057] like Figure 2 As shown, the thermal energy collaborative generation adversarial network (TeSGAN) module mainly includes the following modules: voltage thermal energy input module, current thermal energy input module, voltage thermal energy generator module, current thermal energy generator module, voltage thermal energy output module, current thermal energy output module, voltage thermal energy discriminator module, current thermal energy discriminator module, wherein:

[0058] The voltage heat energy input module and the current heat energy input module constitute the coordinated heat energy input of the ship power station;

[0059] The voltage thermal energy input module is responsible for transmitting the three-phase voltage data of the ship power station to the voltage thermal energy generator module, and using the generated data of the voltage thermal energy generator module as the input of the voltage thermal energy output module;

[0060] The voltage thermal energy discriminator module is responsible for judging whether the error requirement is met according to the generated value and the original value output by the voltage thermal energy output module, and feeding back the result to the voltage thermal energy generation module and the current thermal energy generation module;

[0061] The current thermal energy input module is responsible for transmitting the current thermal energy data of the ship power station to the current thermal energy generator module, and using the generated data of the current thermal energy generator module as the input of the current thermal energy output module;

[0062] The current thermal energy discriminator module is responsible for judging whether the error requirement is met according to the generated value and the original value output by the current thermal energy output module, and feeding back the result to the current thermal energy generation module and the voltage thermal energy generation module;

[0063] The voltage thermal energy discriminator module and the current thermal energy discriminator module are combined to form a Nash equilibrium thermal energy cooperative discriminator, which is the key to evaluating whether the generation and confrontation results meet the error standard.

[0064] The present invention also provides a Beidou satellite autonomous communication method based on the triggering of abnormal events in ship power stations. First, the three-phase voltage and current data waveforms of the ship power station are monitored in real time, and the data are processed in segments at a fixed period, and a new LightGBM is proposed to classify the voltage / current thermal energy data; then, the three-phase voltage and current thermal energy waveforms of the ship power station are denoised and preprocessed, and the present invention designs a new particle swarm optimization algorithm to determine the number of wavelet packet layers; then, the denoised data is input into the Beidou satellite autonomous communication module triggered by the abnormal event, and the thermal energy collaborative generative adversarial network (TeSGAN) module is the core sub-part of the module .... Then, the processed data is input into the judgment threshold module of the abnormal degree of thermal energy of the ship power station. The maximum judgment threshold of the abnormal degree is pre-defined artificially. If it is greater than the threshold, the Beidou satellite short message communication module between the ship and the shore is triggered, and a single communication from the ship to the shore command platform is executed. However, if it is less than the threshold, the Beidou satellite short message communication module between the ship and the shore cannot be triggered, and it is directly returned to the three-phase voltage and current module of the ship power station to execute the cycle process until all the three-phase voltage and current thermal energy data to be processed are completely analyzed, then the whole process is ended, and finally the Beidou satellite autonomous communication based on the triggering of abnormal events in the ship power station is realized. The specific steps are as follows:

[0065] Step 1: Use the ship power station three-phase voltage and current acquisition module to monitor the ship power station three-phase voltage and current data waveform in real time, and segment the data at a fixed period. In this process, in order to effectively classify the voltage and current data and segment the data at a fixed period, the new LightGBM of the present invention is used to classify the data. The LightGBM processing equation is as follows:

[0066]

[0067] Among them, F (t) is the processing target at the current moment, λ iis the coefficient term, is the loss expression, l represents the loss function, y i is the output value of this module, is the estimated value of this module at time t-1, n is the data length of three-phase voltage and current, f t represents the tree, Ω(f t ) represents the regularization term of the tree, ε is a fixed correction parameter, and f t (x i ) is a new tree, The meaning of is that the current prediction value is composed of the previous prediction value and the new tree. The improvement of the above formula is reflected in two aspects: λ is proposed i Correction coefficient; introduce ε into the loss expression. Specifically, λ i It is represented by a three-part function:

[0068]

[0069] Where c is a fixed value. is an approximation function, p, q are segmented decision thresholds, prob k (j) is the probability that k belongs to j, the symbol prob represents the probability, k represents the kth classification type, m is the total number of categories, and max is the maximum identifier. i It can effectively improve the nonlinear effect of the model, avoid overfitting, and intelligently classify the voltage and current data.

[0070] Step 2: Input the data processed in step 1 into the three-phase voltage and current denoising preprocessing module of the ship power station, perform denoising preprocessing on the three-phase voltage and current of the ship power station, remove the noise in the voltage and current waveforms, and use the denoised data as the input value of the three-phase voltage and current thermal energy input value module of the ship power station.

[0071] In this process, the wavelet packet denoising method is used to pre-process the three-phase voltage and current denoising of the ship power station. Considering the problem that the number of wavelet packet layers is difficult to determine, the present invention designs a new particle swarm optimization algorithm to determine the number of wavelet packet layers. The specific equation is as follows:

[0072]

[0073] The position equation is not an improvement of the present invention, so no additional explanation is given here. The above equation is a new velocity equation proposed by the present invention, in which: is the particle update speed at time φ+1, is the particle update speed at time φ, r 1 ∈(0,1),r 2 ∈(0,1) are two different random numbers, vid is the velocity value, φ is the discretized time representation, ε 1 ,ε 2 is the weight coefficient, and is an extreme value, is the actual position of the particle at time φ, ω max and ω min is the balance coefficient, K is the iterative maximum value, and Δ is a fixed preset value. The novel particle swarm optimization algorithm proposed in the present invention can enhance the global optimization capability.

[0074] Step 3: Input the denoised data from step 2 into the ship power station three-phase voltage and current thermal energy input value module. In this process, the segmentation and classification of voltage / current have been completed in step 1, and the preprocessing of voltage / current has been completed in step 2. The ship power station three-phase voltage and current thermal energy input value module needs to preprocess the missing data and use the piecewise linear interpolation method to fill in the missing data values.

[0075] Step 4: Input the data processed in step 3 into the Thermal Energy Synergistic Generative Adversarial Network (TeSGAN) module to generate voltage and current thermal energy data.

[0076] In this process, the thermal energy collaborative generation adversarial network (TeSGAN) module includes a voltage thermal energy input module, a current thermal energy input module, a voltage thermal energy generator module, a current thermal energy generator module, a voltage thermal energy output module, a current thermal energy output module, a voltage thermal energy discriminator module, and a current thermal energy discriminator module, wherein: the output values ​​of the voltage thermal energy input module and the current thermal energy input module are represented by x and y respectively; the output value of the voltage thermal energy generator module is represented by G; the output value of the current thermal energy generator module is represented by F; G1 and G2 are the output values ​​of the voltage thermal energy generator module when the voltage and current are used as inputs respectively; F1 is the input value fed back from the voltage thermal energy discriminator module to the current thermal energy generator module, and F2 is the input value fed back from the current thermal energy discriminator module to the voltage thermal energy generator module; D 1 ,D 2 They are the discrimination results of the voltage thermal energy discriminator module and the current thermal energy discriminator module, which specifically include the following processing:

[0077] L1=∑ x [logG2(x)]+∑ y [log(2-G1(F(y)))]+∑ y [logG1(y)]+∑ x [log(1-G2(G1(x)))]

[0078] Among them, L1 is the loss value of the generative adversarial network model without considering thermal energy synergy, which consists of four parts, including: the output part of the voltage thermal energy generator module, the current thermal energy output module, the output part of the current thermal energy generator module, and the voltage thermal energy output module.

[0079] L2=∑ x [||xF(G2(x)|| 2 ]+∑ y [||yG(F1(y)|| 2 ]

[0080] Among them, L2 is the thermal energy coordination part added by the present invention, and its structure consists of two parts, namely: the numerical result of the voltage thermal energy discriminator module input to the current thermal energy generator module, and the numerical result of the voltage thermal energy discriminator module fed back to the voltage thermal energy generator module.

[0081] L2′=∑ x [||xF(G1(x)|| 2 ]+∑ y [||yG(F2(y)|| 2 ]

[0082] Among them, L2′ is the thermal energy coordination part added by the present invention, and its structure consists of two parts, namely: the numerical result of the current thermal energy discriminator module input to the voltage thermal energy generator module, and the numerical result of the current thermal energy discriminator module fed back to the current thermal energy generator module.

[0083] L(G,F,D 1 ,D 2 )=γL1+μL2+εL2′

[0084] Among them, L(G,F,D 1 ,D 2 ) is the loss function of the thermal collaborative generative adversarial network (TeSGAN) module of the present invention, and is composed of three parts: L1, L2, and L2′. The three coefficients γ, μ, and ε are used as parameters to balance each sub-part.

[0085]

[0086] Among them, G * ,F * It is the final optimization objective function of the thermal energy collaborative generation adversarial network (TeSGAN) module of the present invention, which means that the ideal Nash equilibrium state is achieved between the thermal energy collaborative generators.

[0087] Step 5: Input the data processed in step 4 into the module for calculating the thermal energy anomaly degree of multivariate data of ship power station, and calculate the thermal energy anomaly degree of multivariate data of ship power station, including the fusion anomaly degree of voltage and current. In this process, the voltage and current thermal energy data generated by the Thermal Energy Synergistic Generative Adversarial Network (TeSGAN) module are taken as the object, and the real ship three-phase voltage and current thermal energy data are taken as the benchmark, and the thermal energy anomaly degree of each segmented data is calculated in the form of absolute average value.

[0088] Step 6: Input the data processed in step 5 into the judgment threshold module of the abnormal degree of thermal energy of the ship power station, and analyze whether the abnormal degree of thermal energy of the ship power station reaches the judgment threshold based on the abnormal degree of thermal energy. In this process, the maximum judgment threshold that the abnormal degree of thermal energy can reach is pre-defined by humans. If the abnormal degree of thermal energy is greater than the judgment threshold, the Beidou satellite short message communication module between the ship and the shore is triggered, and a single communication from the ship to the command platform on the shore is executed. However, if the abnormal degree of thermal energy is less than the judgment threshold, the Beidou satellite short message communication module between the ship and the shore cannot be triggered, and it directly returns to the three-phase voltage and current module of the ship power station, reselects a new section of the three-phase voltage and current value of the ship power station, and performs a new round of calculations until all the three-phase voltage and current thermal energy data to be processed are completely analyzed, then the whole process is terminated, and finally the Beidou satellite autonomous communication based on the triggering of abnormal events of the ship power station is realized.

[0089] In order to verify the performance of the above steps 1 to 6, the present invention uses hardware to implement the Beidou satellite autonomous communication method designed by the present invention based on the triggering of abnormal events of ship power stations. Figure 3 The waveform diagrams of the three-phase voltage data and the three-phase current data of the ship power station to be processed are shown in FIG. Figure 4 The results of abnormal identification of three-phase voltage data of ship power station based on TeSGAN and the results of abnormal identification of three-phase current data of ship power station based on TeSGAN are shown in the figure. Figure 5 The comparison results of the Beidou satellite autonomous communication performance based on the triggering of abnormal events in ship power stations are shown in the figure. Through the experimental curves and the data obtained, it can be analyzed that: under the action of the Beidou satellite autonomous communication method based on the triggering of abnormal events in ship power stations proposed by the present invention, the thermal collaborative generation adversarial network (TeSGAN) module can effectively detect abnormal data segments, and at the moment when each abnormal data segment is detected, the Beidou satellite short message communication module between the ship and the shore is triggered in real time. Compared with the Beidou short message time triggering mechanism in the prior art, the Beidou satellite autonomous communication method based on the triggering of abnormal events in ship power stations designed by the present invention effectively reduces the frequency of Beidou short message communication, thereby improving the communication efficiency of Beidou satellite short messages.

Claims

1. A Beidou satellite autonomous communication system based on abnormal events triggered by ship power stations, characterized in that The system includes a ship power station three-phase voltage and current acquisition module, a ship power station three-phase voltage and current denoising preprocessing module, a ship power station three-phase voltage and current thermal energy input value module, a thermal energy collaborative generation adversarial network module, a thermal energy anomaly degree module for calculating multivariate data of the ship power station, a ship power station thermal energy anomaly degree reaching judgment threshold module, and a Beidou satellite short message communication module for triggering ship-shore communication, wherein: The ship power station three-phase voltage and current acquisition module is responsible for real-time monitoring of the ship power station three-phase voltage and current data waveform, and segmenting the data at a fixed period; The ship power station three-phase voltage and current denoising preprocessing module is responsible for denoising the ship power station three-phase voltage and current collected by the ship power station three-phase voltage and current collection module, removing the noise in the voltage and current waveforms, and inputting the denoised data into the ship power station three-phase voltage and current thermal energy input value module; The three-phase voltage and current thermal energy input value module of the ship power station is responsible for connecting with the local database of the ship power station, obtaining the three-phase voltage and current thermal energy input values ​​from the local database, and at the same time, pre-processing the lost data and using the piecewise linear interpolation method to fill in the lost data values; The thermal energy collaborative generation adversarial network module is responsible for mining the key data segments that trigger Beidou satellite communications; The module for calculating the thermal energy anomaly degree of the multivariate data of the ship power station is responsible for calculating the thermal energy anomaly degree of the multivariate data of the ship power station, including the fusion anomaly degree of voltage and current; The module for determining whether the abnormal degree of thermal energy of the ship power station reaches the judgment threshold is responsible for analyzing whether the abnormal degree of thermal energy of the ship power station reaches the set judgment threshold based on the abnormal degree of thermal energy of the multivariate data of the ship power station. If the set judgment threshold is not reached, a new section of three-phase voltage and current values ​​of the ship power station is reselected to perform a new round of calculation; if the set judgment threshold is reached, the Beidou satellite short message communication between the ship and the shore is triggered; The module for triggering Beidou satellite short message communication between ship and shore is responsible for triggering Beidou satellite short message communication between ship and shore.

2. The Beidou satellite autonomous communication system based on the triggering of abnormal events of ship power stations according to claim 1 is characterized in that The thermal energy collaborative generation adversarial network module includes a voltage thermal energy input module, a current thermal energy input module, a voltage thermal energy generator module, a current thermal energy generator module, a voltage thermal energy output module, a current thermal energy output module, a voltage thermal energy discriminator module, and a current thermal energy discriminator module, wherein: The voltage heat energy input module and the current heat energy input module constitute the coordinated heat energy input of the ship power station; The voltage thermal energy input module is responsible for transmitting the three-phase voltage data of the ship power station to the voltage thermal energy generator module, and using the generated data of the voltage thermal energy generator module as the input of the voltage thermal energy output module; The voltage thermal energy discriminator module is responsible for judging whether the error requirement is met according to the generated value and the original value output by the voltage thermal energy output module, and feeding back the result to the voltage thermal energy generation module and the current thermal energy generation module; The current thermal energy input module is responsible for transmitting the current thermal energy data of the ship power station to the current thermal energy generator module, and using the generated data of the current thermal energy generator module as the input of the current thermal energy output module; The current thermal energy discriminator module is responsible for judging whether the error requirement is met according to the generated value and the original value output by the current thermal energy output module, and feeding back the result to the current thermal energy generation module and the voltage thermal energy generation module; The voltage thermal energy discriminator module and the current thermal energy discriminator module are combined to form a Nash equilibrium thermal energy cooperative discriminator to evaluate whether the generated and antagonistic results meet the error standard.

3. A method for realizing Beidou satellite autonomous communication based on triggering of abnormal events of ship power station by using the system described in any one of claims 1-2, characterized in that The method comprises the following steps: Step 1: Use the ship power station three-phase voltage and current acquisition module to monitor the three-phase voltage and current data waveform of the ship power station in real time, and process the data in segments at a fixed period; Step 2: Input the data processed in step 1 to the three-phase voltage and current denoising preprocessing module of the ship power station, perform denoising preprocessing on the three-phase voltage and current of the ship power station, remove the noise in the voltage and current waveforms, and use the denoised data as the input value of the three-phase voltage and current thermal energy input value module of the ship power station; Step 3: Input the denoised data from step 2 into the three-phase voltage and current thermal energy input value module of the ship power station, pre-process the missing data, and use the piecewise linear interpolation method to fill in the missing data values; Step 4: Input the data processed in step 3 into the thermal energy collaborative generation adversarial network module to generate voltage and current thermal energy data; Step 5: input the data processed in step 4 into a module for calculating the thermal energy anomaly degree of multivariate data of a ship power station, and calculate the thermal energy anomaly degree of the multivariate data of the ship power station, including the fusion anomaly degree of voltage and current; Step 6: Input the data processed in step 5 into the judgment threshold module of the thermal energy anomaly degree of the ship power station, and analyze whether the thermal energy anomaly degree of the ship power station reaches the judgment threshold based on the thermal energy anomaly degree. If the thermal energy anomaly degree is greater than the judgment threshold, the Beidou satellite short message communication module between the ship and the shore is triggered, and a single communication from the ship to the shore command platform is executed. However, if the thermal energy anomaly degree is less than the judgment threshold, the Beidou satellite short message communication module between the ship and the shore cannot be triggered, and it directly returns to the three-phase voltage and current module of the ship power station, and reselects a new section of the three-phase voltage and current value of the ship power station, and performs a new round of calculation until all the three-phase voltage and current thermal energy data to be processed are completely analyzed, then the whole process is terminated, and finally the Beidou satellite autonomous communication triggered by the abnormal event of the ship power station is realized.

4. The Beidou satellite autonomous communication method based on triggering of abnormal events of ship power stations according to claim 3 is characterized in that In the step 1, LightGBM is used to classify the data, and the LightGBM processing equation is as follows: Among them, F (t) is the processing target at the current moment, λ i is the coefficient term, is the loss expression, l represents the loss function, y i is the output value of this module, is the estimated value of this module at time t-1, n is the data length of three-phase voltage and current, f t represents the tree, Ω(f t ) represents the regularization term of the tree, ε is a fixed correction parameter, and f t (x i ) is a new tree, The meaning is: the current moment prediction value is composed of the previous moment prediction value and the new tree.

5. The Beidou satellite autonomous communication method based on triggering of abnormal events of ship power stations according to claim 4 is characterized in that The lambda i It is represented by a three-part function: Where c is a fixed value. is an approximation function, p, q are segmented decision thresholds, prob k (j) is the probability that k belongs to j, the symbol prob represents probability, k represents the kth classification type, m is the total number of categories, and max is the maximum identifier.

6. The Beidou satellite autonomous communication method based on triggering of abnormal events of ship power stations according to claim 3 is characterized in that In the step 2, the wavelet packet denoising method is used to pre-process the three-phase voltage and current of the ship power station, and a particle swarm optimization algorithm is designed to determine the number of wavelet packet layers. The specific equation is as follows: in, is the particle update speed at time φ+1, is the particle update speed at time φ, r1∈(0,1), r2∈(0,1) are two different random numbers, v id is the velocity value, φ is the discretized time representation, ε1, ε2 are weight coefficients, and is an extreme value, is the actual position of the particle at time φ, ω max and ω min is the balance coefficient, K is the iterative maximum value, and Δ is a fixed preset value.

7. The Beidou satellite autonomous communication method based on triggering of abnormal events of ship power stations according to claim 3 is characterized in that The specific steps of step 4 are as follows: The output values ​​of the voltage thermal energy input module and the current thermal energy input module are represented by x and y respectively; the output value of the voltage thermal energy generator module is represented by G; the output value of the current thermal energy generator module is represented by F; G1 and G2 are the output values ​​of the voltage thermal energy generator module when the voltage and current are used as input respectively; F1 is the input value fed back from the voltage thermal energy discriminator module to the current thermal energy generator module, and F2 is the input value fed back from the current thermal energy discriminator module to the voltage thermal energy generator module; D1 and D2 are the discrimination results of the voltage thermal energy discriminator module and the current thermal energy discriminator module, respectively, which specifically include the following processing: L1=Σ x [logG2(x)]+Σ y [log(2-G1(F(y)))]+Σ y [logG1(y)]+Σ x [log(1-G2(G1(x))] Among them, L1 is the loss value of the generative adversarial network model without considering thermal energy synergy; L2=Σ x [||x-F(G2(x)||2]+∑ y [||y-G(F1(y)||2] Wherein, L2 is the heat energy synergistic part added by the present invention; L2′=Σ x [||x-F(G1(x)||2]+Σ y [||y-G(F2(y)||2] Wherein, L2′ is the heat energy synergistic part added by the present invention; L(G,F,D1,D2)=γL1+μL2+εL2′ Among them, L(G,F,D1,D2) is the loss function of the thermal energy collaborative generative adversarial network module, which consists of three parts: L1, L2, and L2′. The three coefficients γ, μ, and ε are used as parameters to balance each sub-part; Among them, G * ,F * It is the final optimization objective function of the thermal energy collaborative generation adversarial network module, which means that the ideal Nash equilibrium state is achieved between the thermal energy collaborative generators.

Images