Surge Protection Control Method and System for Electronic Devices Based on Overcurrent and Overvoltage Detection

Abstract

The present invention relates to the technical field of surge protection, and discloses a surge protection control method and system for electronic devices based on over-current and over-voltage detection. First, an initial current and voltage data set of the electronic device is obtained, an initial bidirectional long short-term memory neural network model is trained, and an error function is established; secondly, the error function is used as a fitness function, and an improved snow ablation optimization algorithm is used to optimize the parameters in the initial bidirectional long short-term memory neural network model to construct a final bidirectional long short-term memory neural network model; the initial current and voltage data set of the electronic device is input to obtain an over-current and over-voltage data set of the electronic device, and then the data is corrected; finally, a safety threshold is set, and the surge protection result of the electronic device is obtained by comparison, so as to realize the surge protection control of the electronic device. The present invention processes and analyzes the current and voltage data of the electronic device to achieve the purpose of surge protection control of the electronic device, and the method is objective and accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of surge protection, and specifically to an electronic device surge protection control method and system based on overcurrent and overvoltage detection. Background Art

[0002] In traditional electronic device surge protection control methods, due to the different sensitivities of different electronic devices, it is difficult to control surge protection in real time, which is likely to damage electronic devices; at the same time, technologies such as artificial intelligence are not used, and surge voltages need to be monitored manually, which not only wastes manpower and material resources, but also has a slow speed in processing surge-related data. When there are potential risks, protective measures cannot be taken in advance to ensure good surge protection effects. Summary of the Invention

[0003] In view of the problems in the related art, the present invention provides an electronic device surge protection control method and system based on overcurrent and overvoltage detection to overcome the above-mentioned technical problems existing in the existing related technologies.

[0004] To solve the above technical problems, the present invention is implemented through the following technical solutions:

[0005] The present invention is an electronic device surge protection control method based on overcurrent and overvoltage detection, including the following steps:

[0006] S1. Obtain the current and voltage of the electronic device to obtain an initial electronic device current and voltage data set, train an initial bidirectional long short-term memory neural network model and a current and voltage error set, and establish an error function according to the current and voltage error set;

[0007] S2. Use the error function as a fitness function, use an improved snow ablation optimization algorithm to optimize the parameters in the initial bidirectional long short-term memory neural network model to obtain optimized parameters, and use the optimized parameters to construct a final bidirectional long short-term memory neural network model;

[0008] S3. Input the initial electronic device current and voltage data set, output a current and voltage anomaly recognition result to obtain an electronic device overcurrent and overvoltage data set, and then perform data correction to obtain an electronic device surge data set;

[0009] S4. Dynamically set a safety threshold according to the electronic device surge data set, obtain an electronic device surge protection result by comparing the safety threshold, and then add an electronic device circuit protection device to achieve electronic device surge protection control.

[0010] The invention obtains a set of initial current and voltage data of an electronic device, trains a bidirectional long short-term memory neural network to obtain an initial bidirectional long short-term memory neural network model, calculates the current and voltage error, and then establishes an error function. Compared with other neural networks, it can better handle the temporal differences between normal data and abnormal data, where the abnormal data is overcurrent and overvoltage data, and it is suitable for processing temporal data with good detection effect. Secondly, taking the error function as the fitness function, uses an improved snow ablation optimization algorithm to optimize the parameters in the model to obtain optimized parameters, and constructs a final bidirectional long short-term memory neural network model. This method optimizes the weights in the model, speeds up the convergence speed of the neural network, reduces the training time. The improved snow ablation optimization algorithm simulates the sublimation and melting of snow, combines heat transfer and condensation strategies, and compared with traditional optimization algorithms, improves the deficiencies of the original population mechanism, accelerates the convergence speed of the algorithm, and greatly improves the optimization efficiency. Then, combined with the set of initial current and voltage data of the electronic device, a set of overcurrent and overvoltage data of the electronic device is obtained, and then data correction is performed to obtain a set of surge data of the electronic device. This method takes the abnormal data identified by the neural network model as overcurrent and overvoltage data, which is simple, and then performs two data corrections to eliminate the device acquisition error and perform voltage correction, increasing the reliability and accuracy of the data. Finally, a safety threshold is obtained according to the set of surge data of the electronic device, and by comparing the safety threshold and adding a circuit protection device to the electronic device circuit, surge protection control of the electronic device is realized.

[0011] Preferably, S1 includes the following steps:

[0012] S11. Use a current sensor and a voltage sensor to collect the current and voltage at different positions of the electronic device, obtain the current data of the electronic device and the voltage data of the electronic device, and record the data collection time to obtain a collection time series , where represents the m th collection time point. Combining the collection time series, the current data of the electronic device and the voltage data of the electronic device, an initial set of current and voltage data of the electronic device is obtained , where represents the current data of the electronic device collected at the m th collection time point, represents the voltage data of the electronic device collected at the m th collection time point;

[0013] S12. Train a bidirectional long short-term memory neural network to obtain an initial bidirectional long short-term memory neural network model and a set of current and voltage errors. The specific steps are as follows:

[0014] S121. Set that the bidirectional long short-term memory neural network includes a forward long short-term memory neural network and a backward long short-term memory neural network. Divide the data set to be measured according to the time series and input it into the bidirectional long short-term memory neural network. By setting the forward weights and biases as well as the backward weights and biases, respectively output the output value of the forward long short-term memory neural network and the output value of the backward long short-term memory neural network, and merge the output values to obtain the output value of the bidirectional long short-term memory neural network;

[0015] S122. Re-obtain a new set of current and voltage data of the electronic device, regard the new set of current and voltage data of the electronic device as the sample data set, and set the acquisition time point for the sample data set , and the acquisition time point and the acquisition time point The corresponding current data and voltage data of the electronic device are input into the bidirectional long short-term memory neural network, and the output values of the current and voltage of the electronic device corresponding to the acquisition time point are output, and the current-voltage error at the acquisition time point is calculated. Continuously input the sample data set until the bidirectional long short-term memory neural network converges to obtain the initial bidirectional long short-term memory neural network model and the current-voltage error set;

[0016] S13. Set and respectively represent the i th current error and the i th voltage error in the current-voltage error set, and establish a current error function and a voltage error function. The formulas are as follows:

[0017] , ;

[0018] Among them, represents the current error function, represents the voltage error function, n represents the number of errors;

[0019] Take the mean value of the current error function and the voltage error function as the error function.

[0020] This invention obtains the initial bidirectional long short-term memory neural network model by training the bidirectional long short-term memory neural network. Compared with other neural networks, it can better handle the time series differences between normal data and abnormal data, where the abnormal data is overcurrent and overvoltage data. It is suitable for processing time series data and has good detection effects. Then calculate the current-voltage error to establish an error function, which is convenient for subsequent optimization.

[0021] Preferably, the S2 includes the following steps:

[0022] S21. During the iteration of the initial bidirectional long short-term memory neural network model, the error function is used as the fitness function, and the improved snow ablation optimization algorithm is used to optimize the parameters in the initial bidirectional long short-term memory neural network model. The parameters are weights, and the optimized parameters are obtained. The specific steps are as follows:

[0023] S211. Regarding the iteration process of the initial bidirectional long short-term memory neural network model as the search space, there is a snow population in the search space. The snow individuals in the snow population represent parameters, and the dimension of the snow population is j ; The snow population is transformed into water vapor at high temperature to obtain a water vapor population. The water vapor individuals in the water vapor population move in Brownian motion, and the heat transfer strategy is used to replace Brownian motion; Set the current iteration number as b , the heat attenuation factor as , and respectively represent random numbers between the intervals [0, 1]. When is less than the heat attenuation factor, the position of the b th iteration and the c th water vapor individual is denoted as ; Calculate the fitness function value corresponding to the water vapor individual in the water vapor population at the b th iteration, sort the fitness function values in descending order, and randomly select a water vapor individual from the top five water vapor individuals with the highest fitness function values, denoted as . At this time, the water vapor population moves towards the water vapor individual with a higher fitness function value, and the position of the water vapor individual is updated. The calculation formula is as follows:

[0024] ;

[0025] The water vapor individuals in the water vapor population transfer heat to each other. Denote the difference between the fitness function value corresponding to the b th iteration and the c th water vapor individual and the fitness function values corresponding to other water vapor individuals at the b th iteration as ; Set the position of the b th iteration and the c -1th water vapor individual as , and the position of the b th iteration and the c +1th water vapor individual as . At this time, the position of the water vapor individual ;

[0026] S212. Compare the fitness function values corresponding to the current positions of the water vapor individuals, find the current best fitness function value, and obtain the current optimal parameters at the current best fitness function value; the snow population turns into liquid water at low temperatures, and the condensation strategy is used to replace the snow melting process. Set the maximum number of iterations to B , the condensation parameter ; At the b th iteration, the average position of the water vapor individuals is denoted as , and at the b th iteration, the best position of the water vapor individuals is denoted as , represents a random number between the interval [-1, 1]. At this time, the position of the water vapor individuals is updated again, ;

[0027] At the end of the b th iteration, the entire process of this iteration of the water vapor population is completed, a new generation of water vapor population is generated, and the next iteration is entered. The parameters are continuously updated until the current iteration number reaches the maximum number of iterations, at which point the iteration stops, and the final water vapor population is obtained. The position of the water vapor individual corresponding to the best fitness function value is regarded as the optimized parameter;

[0028] S22. During the iteration of the initial bidirectional long short-term memory neural network model, according to the optimization process of the improved snow ablation optimization algorithm, the optimized parameters are used as the weights of the initial bidirectional long short-term memory neural network model to construct the final bidirectional long short-term memory neural network model.

[0029] In this invention, by using the error function as the fitness function and optimizing the parameters in the model using the improved snow ablation optimization algorithm, the optimized parameters are obtained, and the weights in the model are optimized, which speeds up the convergence speed of the neural network, reduces the training time. The improved snow ablation optimization algorithm simulates the sublimation and melting phenomena of snow, combines heat transfer and condensation strategies, and compared with traditional optimization algorithms, improves the deficiencies of the original population mechanism, accelerates the convergence speed of the algorithm, and greatly improves the optimization efficiency.

[0030] Preferably, the S3 includes the following steps:

[0031] S31. For the initial electronic device current-voltage data set, set the acquisition time point , and input the electronic device current data and electronic device voltage data corresponding to the acquisition time point and the acquisition time point into the final bidirectional long short-term memory neural network model to output the current-voltage anomaly recognition result. The current-voltage anomaly recognition result is the acquisition time point The corresponding current output value and voltage output value of the electronic device are recorded in sequence to obtain a set of current-voltage output values of the electronic device;

[0032] A current threshold and a voltage threshold are set. At the same acquisition time point, the set of current-voltage output values of the electronic device is compared with the initial set of current-voltage data of the electronic device. When the difference between the current output value of the electronic device in the set of current-voltage output values of the electronic device and the current data of the electronic device in the initial set of current-voltage data of the electronic device is greater than the current threshold, the corresponding current data of the electronic device is recorded as overcurrent data of the electronic device; when the difference between the voltage output value of the electronic device in the set of current-voltage output values of the electronic device and the voltage data of the electronic device in the initial set of current-voltage data of the electronic device is greater than the voltage threshold, the corresponding voltage data of the electronic device is recorded as overvoltage data of the electronic device, and an overcurrent-overvoltage data set of the electronic device is formed;

[0033] S32. Perform data correction on the overcurrent-overvoltage data set of the electronic device to obtain a surge data set of the electronic device. The specific steps are as follows:

[0034] S321. Perform the first data correction on the overcurrent-overvoltage data set of the electronic device. Count the acquisition time points of the overcurrent data of the electronic device in the overcurrent-overvoltage data set of the electronic device and the acquisition time points of the overvoltage data of the electronic device in the overcurrent-overvoltage data set of the electronic device. When the acquisition time points of the overcurrent data of the electronic device and the acquisition time points of the overvoltage data of the electronic device are different, delete the overcurrent data of the electronic device corresponding to the acquisition time point of the overcurrent data of the electronic device and delete the overvoltage data of the electronic device corresponding to the acquisition time point of the overvoltage data of the electronic device to obtain a preliminarily processed overcurrent-overvoltage data set of the electronic device;

[0035] S322. Perform the second data correction on the preliminarily processed overcurrent-overvoltage data set of the electronic device. Set sample points on the electronic device, and arbitrarily select sample points and record them as the first sample point and the second sample point respectively, with the current direction flowing from the first sample point to the second sample point; set the active voltage flowing from the first sample point to the second sample point as and the reactive voltage flowing from the first sample point to the second sample point as , the voltage at the first sample point is , the voltage at the first sample point; calculate the voltages at all sample points in sequence, set a sample point threshold, and when the difference between the voltage at the sample point and the overvoltage data of the electronic device in the preliminarily processed overcurrent-overvoltage data set of the electronic device is greater than the sample point threshold, delete the corresponding overvoltage data of the electronic device and the overcurrent data of the electronic device; complete the data correction, regard the overcurrent-overvoltage data of the electronic device as surge data, and obtain a surge data set of the electronic device.

[0036] The invention obtains the over-current and over-voltage data set of the electronic device by comparing thresholds. By using the abnormal data identified by the neural network model as the over-current and over-voltage data, the method is simple. Then, two data corrections are performed to exclude the device acquisition error and perform voltage correction from the aspects of space and time sequence respectively, increasing the reliability and accuracy of the data, and obtaining the surge data set of the electronic device.

[0037] Preferably, S4 includes the following steps:

[0038] S41. The surge data set of the electronic device includes the acquisition time point and the over-current and over-voltage data of the electronic device, denoted as , where represents the g th acquisition time point, represents the over-current data of the electronic device acquired at the g th acquisition time point, represents the over-voltage data of the electronic device acquired at the g th acquisition time point; calculate the average value of the over-current data of the electronic device and the average value of the over-voltage data of the electronic device, and use the average value of the over-current data of the electronic device and the average value of the over-voltage data of the electronic device as the safety threshold;

[0039] S42. The electronic device is powered on with current again, and the current and voltage are collected in real time. When the real-time collected current and voltage are greater than the safety threshold, the electronic device generates a surge, issues an alarm, obtains the surge protection result of the electronic device, automatically cuts off the power supply or switches to the safe mode, and adds a circuit protection device to the circuit to realize the surge protection control of the electronic device.

[0040] The invention also discloses a system for the surge protection control method of an electronic device based on over-current and over-voltage detection, which specifically includes: an error detection module, a model parameter optimization module, an abnormal recognition and data correction module, and a surge protection control module;

[0041] The error detection module is used to construct a neural network model to obtain the current and voltage error and establish an error function;

[0042] The model parameter optimization module is used to optimize the parameters in the neural network model by using an optimization algorithm;

[0043] The abnormal recognition and data correction module is used to identify the over-current and over-voltage data set in the current and voltage data set of the electronic device, and then perform data correction;

[0044] The surge protection control module is used to compare the safety threshold to obtain the surge protection result of the electronic device and realize the surge protection control of the electronic device.

[0045] The invention has the following beneficial effects:

[0046] 1. The invention trains a bidirectional long short-term memory neural network to obtain an initial bidirectional long short-term memory neural network model. Compared with other neural networks, it can better handle the temporal differences between normal data and abnormal data, where the abnormal data is overcurrent and overvoltage data, which is applicable to processing temporal data and has good detection effects. Then, the current-voltage error is calculated to establish an error function for subsequent optimization.

[0047] 2. The invention uses the error function as a fitness function and optimizes the parameters in the model using an improved snow ablation optimization algorithm to obtain optimized parameters, and optimizes the weights in the model, which speeds up the convergence rate of the neural network and reduces the training time. The improved snow ablation optimization algorithm simulates the sublimation and melting phenomena of snow, combines heat transfer and condensation strategies. Compared with traditional optimization algorithms, it improves the deficiencies of the original population mechanism, accelerates the convergence rate of the algorithm, and greatly improves the optimization efficiency.

[0048] 3. The invention obtains the overcurrent and overvoltage data set of the electronic device by comparing thresholds. By using the abnormal data identified by the neural network model as overcurrent and overvoltage data, the method is simple. Then, two data corrections are performed to exclude the device acquisition error and perform voltage correction in terms of space and time respectively, increasing the reliability and accuracy of the data, and obtaining the surge data set of the electronic device.

[0049] Of course, it is not necessary for any product implementing the present invention to simultaneously achieve all the above-mentioned advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0051] Figure 1 FIG. is a schematic flow chart of the surge protection control of an electronic device provided by the surge protection control system of the electronic device based on overcurrent and overvoltage detection of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0053] In the description of the present invention, it should be understood that terms such as "opening", "upper", "lower", "top", "middle", "inner", etc. indicating orientation or positional relationships are only for the convenience of describing the invention and simplifying the description, rather than indicating or implying that the components or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation on the invention.

[0054] Embodiment 1

[0055] Please refer to Figure 1 , the electronic device surge protection control method based on over-current and over-voltage detection of the present invention includes the following steps:

[0056] S1. Obtain the current and voltage of the electronic device to obtain an initial electronic device current-voltage data set, train an initial bidirectional long short-term memory neural network model and a current-voltage error set, and establish an error function according to the current-voltage error set;

[0057] The S1 includes the following steps:

[0058] S11. Use a current sensor and a voltage sensor to collect the current and voltage at different positions of the electronic device to obtain electronic device current data and electronic device voltage data, and record the data acquisition time to obtain an acquisition time series , where represents the m th acquisition time point. Combine the acquisition time series, electronic device current data, and electronic device voltage data to obtain an initial electronic device current-voltage data set , where represents the electronic device current data collected at the m th acquisition time point, represents the electronic device voltage data collected at the m th acquisition time point;

[0059] S12. Train a bidirectional long short-term memory neural network to obtain an initial bidirectional long short-term memory neural network model and a current-voltage error set. The specific steps are as follows:

[0060] S121. Set that the bidirectional long short-term memory neural network includes a forward long short-term memory neural network and a backward long short-term memory neural network. Divide the data set to be measured according to the time series and input it into the bidirectional long short-term memory neural network. By setting the forward weight and bias and the backward weight and bias, respectively output the output value of the forward long short-term memory neural network and the output value of the backward long short-term memory neural network, and merge the output values to obtain the output value of the bidirectional long short-term memory neural network;

[0061] S122. Re-obtain a new set of current and voltage data of the electronic device, regard the new set of current and voltage data of the electronic device as a sample data set, and set the acquisition time point for the sample data set , the acquisition time point and the acquisition time point corresponding electronic device current data and electronic device voltage data are input into a bidirectional long short-term memory neural network, and the electronic device current output value and electronic device voltage output value corresponding to the acquisition time point are output, and the current-voltage error at the acquisition time point is calculated; continuously input the sample data set until the bidirectional long short-term memory neural network converges, and obtain an initial bidirectional long short-term memory neural network model and a current-voltage error set;

[0062] S13. Set and respectively represent the i -th current error and the i -th voltage error in the current-voltage error set, establish a current error function and a voltage error function, and the formulas are as follows:

[0063] , ;

[0064] Among them, represents the current error function, represents the voltage error function, n represents the number of errors;

[0065] Take the mean value of the current error function and the voltage error function as the error function;

[0066] S2. Use the error function as the fitness function, and use an improved snow ablation optimization algorithm to optimize the parameters in the initial bidirectional long short-term memory neural network model to obtain optimized parameters, and use the optimized parameters to construct a final bidirectional long short-term memory neural network model;

[0067] The S2 includes the following steps:

[0068] S21. During the iteration of the initial bidirectional long short-term memory neural network model, use the error function as the fitness function, and use an improved snow ablation optimization algorithm to optimize the parameters in the initial bidirectional long short-term memory neural network model. The parameters are weights, and the optimized parameters are obtained. The specific steps are as follows:

[0069] S211. Regard the iteration process of the initial bidirectional long short-term memory neural network model as a search space. There is a snow population in the search space. The snow individuals in the snow population represent parameters, and the dimension of the snow population isj ; The snow population is transformed into water vapor at high temperature to obtain a water vapor population. The water vapor individuals in the water vapor population move by Brownian motion, and the heat transfer strategy is used to replace Brownian motion. Set the current iteration number to b , the heat attenuation factor is , and respectively represent random numbers between the intervals [0, 1]. When is less than the heat attenuation factor, the position of the b -th iteration and the c -th water vapor individual is recorded as ; Calculate the fitness function value corresponding to the water vapor individual in the water vapor population at the b -th iteration, sort the fitness function values in descending order, and randomly select a water vapor individual from the top five water vapor individuals with the highest fitness function values, recorded as . At this time, the water vapor population moves towards the water vapor individual with a higher fitness function value, and the position of the water vapor individual is updated. The calculation formula is as follows:

[0070] ;

[0071] The water vapor individuals in the water vapor population transfer heat to each other. Denote the difference between the fitness function value corresponding to the b -th iteration and the c -th water vapor individual and the fitness function values corresponding to other water vapor individuals at the b -th iteration as ; Set the position of the b -th iteration and the c -1 water vapor individual as , and the position of the b -th iteration and the c +1 water vapor individual as . At this time, the position of the water vapor individual ;

[0072] S212. Compare the fitness function value corresponding to the current water vapor individual position to find the current best fitness function value, and obtain the current optimal parameters at the current best fitness function value. The snow population is transformed into liquid water at low temperature, and the condensation strategy is used to replace the snow melting process. Set the maximum iteration number to B , the condensation parameter ; Denote the average position of the water vapor individuals at the b -th iteration as , and the best position of the water vapor individuals at the b -th iteration as , Represents a random number between the interval [-1, 1]. At this time, the individual position of water vapor is updated again. For updating. ;

[0073] The b th iteration ends, completing the entire process of this iteration of the water vapor population, generating the next generation of water vapor population, entering the next iteration, and continuing to update the parameters until the current iteration number reaches the maximum iteration number, at which point the iteration stops, obtaining the final water vapor population, and regarding the water vapor individual position corresponding to the best fitness function value as the optimized parameter;

[0074] S22. During the iteration process of the initial bidirectional long short-term memory neural network model, according to the optimization process of the improved snow ablation optimization algorithm, using the optimized parameter as the weight of the initial bidirectional long short-term memory neural network model, the final bidirectional long short-term memory neural network model is constructed;

[0075] S3. Input the initial electronic device current-voltage data set, output the current-voltage anomaly recognition result, obtain the electronic device overcurrent-overvoltage data set, and then perform data correction to obtain the electronic device surge data set;

[0076] The S3 includes the following steps:

[0077] S31. For the initial electronic device current-voltage data set, set the acquisition time point , and input the electronic device current data and electronic device voltage data corresponding to the acquisition time point and the acquisition time point into the final bidirectional long short-term memory neural network model, output the current-voltage anomaly recognition result, where the current-voltage anomaly recognition result is the electronic device current output value and electronic device voltage output value corresponding to the acquisition time point , and record them in sequence to obtain the electronic device current-voltage output value set;

[0078] Set the current threshold and voltage threshold. At the same acquisition time point, compare the electronic device current-voltage output value set and the initial electronic device current-voltage data set. When the difference between the electronic device current output value in the electronic device current-voltage output value set and the electronic device current data in the initial electronic device current-voltage data set is greater than the current threshold, record the corresponding electronic device current data as the electronic device overcurrent data; when the difference between the electronic device voltage output value in the electronic device current-voltage output value set and the electronic device voltage data in the initial electronic device current-voltage data set is greater than the voltage threshold, record the corresponding electronic device voltage data as the electronic device overvoltage data, and form the electronic device overcurrent-overvoltage data set;

[0079] S32. Perform data correction on the over-current and over-voltage data set of the electronic device to obtain a surge data set of the electronic device. The specific steps are as follows:

[0080] S321. Perform the first data correction on the over-current and over-voltage data set of the electronic device. Statistically analyze the acquisition time points of the over-current data of the electronic device in the over-current and over-voltage data set of the electronic device, and the acquisition time points of the over-voltage data of the electronic device in the over-current and over-voltage data set of the electronic device. When the acquisition time points of the over-current data of the electronic device and the acquisition time points of the over-voltage data of the electronic device are different, delete the over-current data of the electronic device corresponding to the acquisition time point of the over-current data of the electronic device, and delete the over-voltage data of the electronic device corresponding to the acquisition time point of the over-voltage data of the electronic device, to obtain a preliminarily processed over-current and over-voltage data set of the electronic device;

[0081] S322. Perform the second data correction on the preliminarily processed over-current and over-voltage data set of the electronic device. Set sample points on the electronic device, and select any sample points and denote them as the first sample point and the second sample point respectively, with the current direction flowing from the first sample point to the second sample point; set the active voltage flowing from the first sample point to the second sample point as , and the reactive voltage flowing from the first sample point to the second sample point as , the voltage at the first sample point as , the voltage at the first sample point; calculate the voltages at all sample points in sequence, set a sample point threshold, and when the difference between the voltage at the sample point and the over-voltage data of the electronic device in the preliminarily processed over-current and over-voltage data set of the electronic device is greater than the sample point threshold, delete the corresponding over-voltage data of the electronic device and the over-current data of the electronic device; complete the data correction, regard the over-current and over-voltage data of the electronic device as surge data, and obtain a surge data set of the electronic device;

[0082] S4. Dynamically set a safety threshold according to the surge data set of the electronic device, obtain a surge protection result of the electronic device by comparing the safety threshold, and then add an electronic device circuit protection device to achieve surge protection control of the electronic device;

[0083] The S4 includes the following steps:

[0084] S41. The surge data set of the electronic device contains an acquisition time point and over-current and over-voltage data of the electronic device, denoted as , where represents the g th acquisition time point, represents the over-current data of the electronic device acquired at the g th acquisition time point, represents the gOvervoltage data of the electronic device collected at each acquisition time point; calculate the average value of the overcurrent data of the electronic device and the average value of the overvoltage data of the electronic device, and use the average value of the overcurrent data of the electronic device and the average value of the overvoltage data of the electronic device as the safety threshold;

[0085] S42. The electronic device is energized again, and the current voltage is collected in real time. When the real-time collected current voltage is greater than the safety threshold, the electronic device generates a surge, issues an alarm, obtains the surge protection result of the electronic device, automatically cuts off the power supply or switches to the safe mode, and adds a circuit protection device to the circuit to realize the surge protection control of the electronic device.

[0086] Embodiment 2

[0087] The present invention also discloses a system for the surge protection control method of an electronic device based on overcurrent and overvoltage detection, which specifically includes: an error detection module, a model parameter optimization module, an anomaly recognition and data correction module, and a surge protection control module;

[0088] The error detection module is used to construct a neural network model to obtain the current voltage error and establish an error function;

[0089] The model parameter optimization module is used to optimize the parameters in the neural network model by using an optimization algorithm;

[0090] The anomaly recognition and data correction module is used to identify the overcurrent and overvoltage data sets in the current voltage data set of the electronic device, and then perform data correction;

[0091] The surge protection control module is used to compare the safety threshold to obtain the surge protection result of the electronic device, and realize the surge protection control of the electronic device.

[0092] In the description of this specification, the descriptions referring to terms such as "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0093] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and changes can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principle and practical application of the invention, so that those skilled in the art in the technical field can well understand and utilize the invention.

Claims

1. An electronic device surge protection control method based on overcurrent and overvoltage detection, characterized in that It includes the following steps: S1. Obtain the current and voltage of the electronic device to get the initial electronic device current and voltage data set, train to obtain the initial bidirectional long short-term memory neural network model and the current and voltage error set, and establish an error function according to the current and voltage error set; S2. Use the error function as the fitness function to optimize the parameters in the initial bidirectional long short-term memory neural network model to obtain the optimized parameters, and use the optimized parameters to construct the final bidirectional long short-term memory neural network model; S3. Input the initial electronic device current and voltage data set, output the current and voltage anomaly recognition result to get the electronic device overcurrent and overvoltage data set, and then perform data correction to get the electronic device surge data set; S4. Dynamically set the safety threshold according to the electronic device surge data set, obtain the electronic device surge protection result by comparing the safety threshold, and then add the electronic device circuit protection device to realize the electronic device surge protection control; The S1 includes the following steps: S11. Collect the current and voltage of the electronic device to obtain the electronic device current data and the electronic device voltage data, and record the data collection time to obtain the collection time series, and form the initial electronic device current and voltage data set; S12. Train the bidirectional long short-term memory neural network to obtain the initial bidirectional long short-term memory neural network model and the current and voltage error set; S13. Obtain the current error function and the voltage error function according to the current and voltage error set, and use the mean value of the current error function and the voltage error function as the error function; The S12 includes the following steps: S121. Set that the bidirectional long short-term memory neural network includes a forward long short-term memory neural network and a backward long short-term memory neural network, divide the data set to be measured according to the time series, input it into the bidirectional long short-term memory neural network, and respectively output the forward long short-term memory neural network output value and the backward long short-term memory neural network output value by setting the forward weight and bias and the backward weight and bias, and merge the output values to obtain the output value of the bidirectional long short-term memory neural network; S122. Re-obtain a new set of electronic device current-voltage data, regard the new set of electronic device current-voltage data as a sample data set, and set the acquisition time point for the sample data set , and input the electronic device current data and the electronic device voltage data corresponding to the acquisition time point and the acquisition time point into the bidirectional long short-term memory neural network, and output the electronic device current output value and the electronic device voltage output value corresponding to the acquisition time point , and calculate the current-voltage error at the acquisition time point . Continuously input the sample data set until the bidirectional long short-term memory neural network converges, and obtain the initial bidirectional long short-term memory neural network model and the current-voltage error set; The S2 includes the following steps: S21. During the iteration of the initial bidirectional long short-term memory neural network model, use the error function as the fitness function and use the improved snow ablation optimization algorithm to optimize the parameters in the initial bidirectional long short-term memory neural network model. The parameters are the weights to obtain the optimized parameters; S22. During the iteration of the initial bidirectional long short-term memory neural network model, according to the optimization process of the improved snow ablation optimization algorithm, use the optimized parameters as the weights of the initial bidirectional long short-term memory neural network model to construct the final bidirectional long short-term memory neural network model.

2. The surge protection control method for an electronic device based on overcurrent and overvoltage detection according to claim 1, characterized in that The S21 includes the following steps: S211. Consider the initial bidirectional long short-term memory neural network model as the search space during the iteration process. There is a snow population in the search space. The snow individuals in the snow population represent parameters, and the dimension of the snow population is j ; The snow population turns into water vapor at high temperature to obtain a water vapor population. The water vapor individuals in the water vapor population move in Brownian motion, and the heat transfer strategy is used to replace Brownian motion. Set the current iteration number as b , the heat decay factor is , and respectively represent random numbers between the interval [0, 1]. When is less than the heat decay factor, the position of the b -th iteration and the c -th water vapor individual is recorded as ; Calculate the fitness function value corresponding to the water vapor individuals in the water vapor population at the b -th iteration, sort the fitness function values in descending order, and randomly select a water vapor individual from the top five water vapor individuals with the highest fitness function values, recorded as . At this time, the water vapor population moves towards the water vapor individual with a higher fitness function value, and updates the position of the water vapor individual. The calculation formula is as follows: ； Water vapor individuals in the water vapor population transfer heat to each other. Denote the difference between the fitness function value corresponding to the b -th water vapor individual at the c -th iteration and the fitness function values corresponding to other water vapor individuals at the b -th iteration as ; Set the position of the b -1 water vapor individual at the c -th iteration as , and the position of the b +1 water vapor individual at the c -th iteration as . At this time, the position of the water vapor individual is ; S212. Compare the fitness function values corresponding to the current positions of the water vapor individuals to find the current best fitness function value, and obtain the current optimal parameters at the current best fitness function value; the snow population is converted into liquid water at low temperature, the condensation strategy is used to replace the snowmelt process, and the maximum number of iterations is set to B , the condensation parameter ; At the b th iteration, the average position of each water vapor individual is denoted as , and at the b th iteration, the best position of each water vapor individual is denoted as . represents a random number between the interval [-1, 1]. At this time, the position of each water vapor individual is updated again, ; The b iteration ends, completing the entire process of this iteration of the water vapor population, generating the next generation of the water vapor population, entering the next iteration, continuing to update the parameters, and stopping the iteration until the current iteration number reaches the maximum iteration number, obtaining the final water vapor population, and regarding the position of the water vapor individual corresponding to the best fitness function value as the optimized parameter.

3. The surge protection control method for an electronic device based on over-current and over-voltage detection according to claim 2, characterized in that, The S3 includes the following steps: S31. Input the initial electronic device current and voltage data set into the final bidirectional long short-term memory neural network model, output the current and voltage anomaly recognition result to get the electronic device current and voltage output value set; Set the current threshold and voltage threshold, compare the set of current and voltage output values of the electronic device with the current threshold and voltage threshold respectively, and find the over-current data and over-voltage data of the electronic device. Form a set of over-current and over-voltage data of the electronic device; S32. Perform data correction on the set of over-current and over-voltage data of the electronic device to obtain a set of surge data of the electronic device.

4. The surge protection control method for an electronic device based on overcurrent and overvoltage detection according to claim 3, wherein The S32 includes the following steps: S321. Perform the first data correction on the set of over-current and over-voltage data of the electronic device. By comparing the acquisition time points of the set of over-current and over-voltage data of the electronic device, filter out the preliminarily processed set of over-current and over-voltage data of the electronic device; S322. Perform the second data correction on the preliminarily processed set of over-current and over-voltage data of the electronic device. Set sample points on the electronic device, correct the voltage at the sample points, regard the over-current and over-voltage data of the electronic device as surge data, and obtain a set of surge data of the electronic device.

5. The surge protection control method for an electronic device based on over-current and over-voltage detection according to claim 4, wherein, The S4 includes the following steps: S41. For the set of surge data of the electronic device, use the average value of the over-current data and the average value of the over-voltage data of the electronic device as the safety threshold; S42. The electronic device is powered on again, and the current and voltage are collected in real time. When the real-time collected current and voltage are greater than the safety threshold, the electronic device generates a surge, issues an alarm, obtains the surge protection result of the electronic device, automatically cuts off the power supply or switches to the safe mode, and adds a circuit protection device to the circuit to achieve the surge protection control of the electronic device.

6. A system for implementing the surge protection control method of an electronic device based on overcurrent and overvoltage detection as described in any one of claims 1-5, characterized in that, Specifically include: An error detection module, a model parameter optimization module, an anomaly recognition and data correction module, and a surge protection control module; The error detection module is used to construct a neural network model to obtain the current and voltage error and establish an error function; The model parameter optimization module is used to optimize the parameters in the neural network model using an optimization algorithm; The anomaly recognition and data correction module is used to identify the over-current and over-voltage data set in the set of current and voltage data of the electronic device and then perform data correction; The surge protection control module is used to compare the safety threshold to obtain the surge protection result of the electronic device and achieve the surge protection control of the electronic device.

Images