Anti-interference leakage protection device and method

Abstract

The present application belongs to the field of leakage protection technology, and discloses an anti-interference leakage protection device and method, the device comprising: an induction module, an amplification module, a rectifier module, an analog-to-digital conversion module, a processing module, and a relay module; the amplification module is used to receive the leakage signal of the induction module and amplify it to obtain an amplified signal; the rectifier module is used to receive the amplified signal and rectify the amplified signal to obtain a DC signal; the analog-to-digital conversion module is used to receive the DC signal and convert the DC signal into a digital signal; the processing module is used to filter the received digital signal and calculate the leakage current value based on the digital signal; when the leakage current value is greater than a preset circuit breaker threshold, a tripping signal is sent to the relay module; the relay module is used to perform a circuit disconnection operation upon receiving the tripping signal. The present application can prevent the leakage protection device from receiving electromagnetic interference during detection, resulting in false detection and false triggering of tripping.

Description

Technical Field

[0001] The present application relates to the technical field of leakage protection, and in particular to an anti-interference leakage protection device and method. Background Art

[0002] Existing leakage protectors usually amplify and rectify the leakage current signal after receiving it, and then compare the value of the DC signal with the threshold to determine whether there is leakage. After determining leakage, they send a trip signal to trigger the tripping action to ensure power safety. In order to improve the stability of detection, some manufacturers will also add a pulse generator to the leakage protector to achieve the widening of the tripping signal to ensure the stable triggering of the tripping action.

[0003] However, electromagnetic induction between circuits or devices may affect the shape and parameters of the induced current generated by the induction coil, causing false detection of the leakage protector and false tripping of the switch, affecting the power user experience. Summary of the Invention

[0004] The present application provides an interference-resistant leakage protection device and method, which can prevent the leakage protection device from being subject to electromagnetic interference during detection, thereby causing false detection, false triggering and tripping.

[0005] In a first aspect, an embodiment of the present application provides an anti-interference leakage protection device, comprising: a sensing module, an amplifying module, a rectifying module, an analog-to-digital conversion module, a processing module, and a relay module connected in sequence;

[0006] The amplifying module is used to receive the leakage signal of the sensing module and amplify it to obtain an amplified signal;

[0007] The rectifier module is used to receive the amplified signal and rectify the amplified signal to obtain a DC signal;

[0008] The analog-to-digital conversion module is used to receive a DC signal and convert the DC signal into a digital signal;

[0009] The processing module is used to filter the received digital signal and calculate the leakage current value based on the digital signal; when the leakage current value is greater than the preset circuit breaker threshold, a trip signal is sent to the relay module;

[0010] The relay module is used to perform a circuit disconnection operation when receiving a trip signal.

[0011] Furthermore, the processing module is specifically used to obtain the range value of the sensing module, the gain value of the amplification module and the number of bits of the analog-to-digital conversion module; and calculate the leakage current value based on the range value, gain value, number of bits and digital signal.

[0012] Furthermore, the analog-to-digital conversion module includes a first analog-to-digital conversion unit, a second analog-to-digital conversion unit and a judgment unit;

[0013] The first analog-to-digital conversion unit is used to convert the DC signal into a first current signal;

[0014] The second analog-to-digital conversion unit is used to convert the DC signal into a second current signal;

[0015] The judgment unit is used to calculate the difference between the first current signal and the second current signal; when the difference is less than a preset calibration threshold, the average value of the first current signal and the second current signal is used as the digital signal.

[0016] Furthermore, the judgment unit is further configured to perform parameter calibration on the first analog-to-digital conversion unit and the second analog-to-digital conversion unit according to the difference when the difference is greater than a preset calibration threshold.

[0017] Furthermore, the processing module is specifically configured to filter the digital signal using a median filter algorithm or a Kalman filter.

[0018] Furthermore, the processing module is also used to extract the current current characteristics of the digital signal; calculate the similarity between the current current characteristics and the historical leakage current characteristics; when the similarity is greater than a preset similarity threshold, calculate the leakage current value based on the digital signal.

[0019] Furthermore, the processing module is further configured to:

[0020] Input the historical leakage current signal into the trained generative adversarial network to obtain a leakage current signal set;

[0021] The leakage current signal set and the normal current signal set are used to train the dilated convolutional neural network to obtain the current recognition model.

[0022] Input the digital signal into the current identification model to obtain the identification result;

[0023] When the identification result is true, the leakage current value is calculated based on the digital signal.

[0024] Furthermore, the device also includes a sensor module;

[0025] The sensor module is used to obtain the working environment parameters of the leakage protection device;

[0026] The processing module is further used to process the current characteristics of the digital signal based on the gain value of the amplification module to obtain the actual current characteristics; and send the working environment parameters, the actual current characteristics and the leakage current value to the mobile terminal via wireless communication.

[0027] Furthermore, the processing module is further configured to:

[0028] Obtain historical leakage current signals and initialize the network parameters of the generative adversarial network;

[0029] Extracting historical leakage current characteristics of historical leakage current signals and working environment parameters of leakage protection devices;

[0030] Instructing the generator to generate a false leakage current signal based on historical leakage current characteristics and working environment parameters;

[0031] Inputting the false leakage current signal and the corresponding historical leakage current signal into a discriminator to obtain a first discrimination probability corresponding to the false leakage current signal and a second discrimination probability corresponding to the historical leakage current signal;

[0032] Determine whether the first discriminant probability is equal to the second discriminant probability;

[0033] If not, the network parameters of the generator and the discriminator are updated according to the first discrimination probability and the second discrimination probability.

[0034] Furthermore, the processing module is further configured to:

[0035] After sending the trip signal, user feedback information is received; if the user feedback information is the preset reset information, the digital signal is placed in the leakage current signal set, otherwise it is placed in the normal current signal set.

[0036] Furthermore, the processing module is also used to record the leakage protection response time starting from the moment the digital signal is received and ending at the moment the trip signal is sent; and send the leakage protection response time to the mobile terminal via wireless communication.

[0037] In a second aspect, an embodiment of the present application provides an anti-interference leakage protection method, comprising:

[0038] The amplifying module amplifies the received leakage signal to obtain an amplified signal;

[0039] The rectifier module rectifies the amplified signal to obtain a DC signal;

[0040] The analog-to-digital conversion module converts the DC signal into a digital signal;

[0041] The processing module filters the received digital signal; calculates the leakage current value based on the digital signal; determines whether the leakage current value is greater than a preset circuit breaker threshold, and if so, sends a trip signal to the relay module;

[0042] When the relay module receives the trip signal, it performs a circuit disconnection operation.

[0043] Furthermore, the above-mentioned calculation of the leakage current value based on the digital signal includes:

[0044] Obtain the range value of the sensing module, the gain value of the amplification module, and the number of bits of the analog-to-digital conversion module;

[0045] Calculates the exponential with base 2 raised to the power of the number of bits;

[0046] Subtract 1 from the exponent and multiply it by the gain value to get the first product;

[0047] Multiply the digital signal by the range value to obtain a second product;

[0048] Divide the second product by the first product to obtain the leakage current value.

[0049] Furthermore, the method further comprises:

[0050] The processing module extracts the current characteristics of the digital signal;

[0051] Calculate the similarity between the current current characteristics and the historical leakage current characteristics;

[0052] When the similarity is greater than a preset similarity threshold, the leakage current value is calculated based on the digital signal.

[0053] Furthermore, the method further comprises:

[0054] The processing module inputs the historical leakage current signal into the trained generative adversarial network to obtain a set of leakage current signals;

[0055] The leakage current signal set and the normal current signal set are used to train the dilated convolutional neural network to obtain the current recognition model.

[0056] Input the digital signal into the current identification model to obtain the identification result;

[0057] When the identification result is true, the leakage current value is calculated based on the digital signal.

[0058] Furthermore, the method further comprises:

[0059] The processing module obtains the historical leakage current signal and initializes the network parameters of the generative adversarial network;

[0060] Extracting historical leakage current characteristics of historical leakage current signals and working environment parameters of leakage protection devices;

[0061] Instructing the generator to generate a false leakage current signal based on historical leakage current characteristics and working environment parameters;

[0062] Inputting the false leakage current signal and the corresponding historical leakage current signal into a discriminator to obtain a first discrimination probability corresponding to the false leakage current signal and a second discrimination probability corresponding to the historical leakage current signal;

[0063] Determine whether the first discriminant probability is equal to the second discriminant probability;

[0064] If not, the network parameters of the generator and the discriminator are updated according to the first discrimination probability and the second discrimination probability.

[0065] Furthermore, the generator generates a false leakage current signal based on historical leakage current characteristics and working environment parameters, including:

[0066] The generator calculates the environmental adjustment factor based on the preset mutation threshold and working environment parameters;

[0067] Multiply the environmental adjustment factor and random noise to obtain the noise vector;

[0068] Multiply the historical leakage current characteristics and the preset extraction weight to obtain a feature vector;

[0069] Add the noise vector and the eigenvector to obtain the false current data;

[0070] Perform inverse Fourier transform on the false current data to obtain a false leakage current signal.

[0071] Furthermore, the updating of the network parameters of the generator and the discriminator according to the first discrimination probability and the second discrimination probability includes:

[0072] Calculate the binary cross entropy loss function of the first discriminant probability and the second discriminant probability respectively;

[0073] Add the two binary cross entropy loss functions to get the discriminant loss function;

[0074] Calculate the updated network parameters of the discriminator based on the discriminant loss function and the discriminant update gradient;

[0075] The updated network parameters of the generator are calculated based on the binary cross entropy loss function of the first discriminant probability and the generated update gradient.

[0076] Furthermore, the method further comprises:

[0077] After sending the trip signal, the processing module receives user feedback information;

[0078] If the user feedback information is the preset reset information, the digital signal is placed in the leakage current signal set; otherwise, the digital signal is placed in the normal current signal set;

[0079] The current recognition model is trained again based on the leakage current signal set and the normal current signal set.

[0080] Furthermore, the method further comprises:

[0081] The processing module records the leakage protection response time starting from the moment of receiving the digital signal and ending at the moment of sending the trip signal; and sends the leakage protection response time to the mobile terminal via wireless communication.

[0082] In summary, compared with the prior art, the technical solutions provided by the embodiments of the present application have at least the following beneficial effects:

[0083] An embodiment of the present application provides an anti-interference leakage protection device. Unlike conventional leakage protectors that directly judge the leakage current signal of an analog signal, the present application adds an analog-to-digital conversion module behind the rectifier module to convert the DC signal of the analog signal into a digital signal, and then uses the processing module to filter the digital signal and calculate the leakage current value. It fully utilizes the advantages of digital signals having high stability during transmission and being less susceptible to electromagnetic interference, thereby avoiding the situation where the leakage protection device is subject to electromagnetic interference, resulting in false detection and false triggering and tripping. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 A structural diagram of an anti-interference leakage protection device provided as an exemplary embodiment of the present application.

[0085] Figure 2 This is a diagram of the internal structure of an analog-to-digital conversion module provided as an exemplary embodiment of the present application.

[0086] Figure 3 A structural diagram of an anti-interference leakage protection device provided as another exemplary embodiment of the present application.

[0087] Figure 4 A flowchart of an anti-interference leakage protection method provided by an exemplary embodiment of the present application.

[0088] Figure 5 A flowchart of the steps for generative adversarial network training is provided for an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0089] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments.

[0090] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work shall fall within the scope of protection of this application.

[0091] See Figure 1 An embodiment of the present application provides an anti-interference leakage protection device, comprising: an induction module, an amplification module, a rectification module, an analog-to-digital conversion module, a processing module and a relay module connected in sequence.

[0092] The sensing module typically uses a zero-sequence current transformer (CCT), which surrounds the neutral and live wires in the circuit and is used to detect leakage current. When the circuit is operating normally, the vector sum of the currents in the neutral and live wires is zero, and the CCT generates no induced electromotive force. However, when leakage occurs, the vector sum of the currents is non-zero, and the secondary coil of the CCT generates an induced electromotive force, sending a leakage signal to the amplification module.

[0093] The amplifying module is used to receive the leakage signal from the sensing module and amplify it to obtain an amplified signal.

[0094] Among them, the amplification module consists of a switched capacitor amplifier and an IA amplifier connected to each other. The switched capacitor amplifier uses a combination of switches and capacitors to achieve signal sampling, holding and amplification, while the instrumentation (IA) amplifier is an amplifier with differential input and single-ended output, and has the characteristics of high input impedance, low noise and high common-mode rejection ratio (CMRR).

[0095] In a switched-capacitor amplifier (IA), the leakage signal is first sampled and held in a switched-capacitor network, and then amplified by the periodic switching of the switches. The amplifier's gain is determined by the ratio of the capacitors and the switching frequency. A key advantage of switched-capacitor amplifiers is that they can achieve high-gain and wide-bandwidth amplification on integrated circuits, while also maintaining low power consumption and a high level of integration. IA amplifiers typically consist of three operational amplifiers: two differential input stages providing high input impedance, and the third operational amplifier serving as a gain stage. The IA amplifier's gain can be adjusted using external resistors or internal settings.

[0096] This application combines a switched capacitor amplifier with an IA amplifier, leveraging the high gain and wide bandwidth of the switched capacitor amplifier for initial amplification. The IA amplifier then provides further precision amplification and improved common-mode rejection. This combined amplification module provides high-precision, low-noise, and highly stable signal amplification.

[0097] The rectifier module is used to receive the amplified signal and rectify the amplified signal to obtain a DC signal.

[0098] Rectifier modules are usually composed of semiconductor devices such as rectifier diodes or thyristors, which are responsible for converting alternating current (AC) into direct current (DC). Rectifier modules mainly include the following types:

[0099] 1) Half-wave rectification: A simple rectifier circuit consisting of a diode, which only uses the positive half-cycle or negative half-cycle of the AC power for rectification, outputs pulsating DC power, and has a low current utilization rate. It is suitable for high voltage and low current occasions.

[0100] 2) Full-wave rectification: It is composed of two diodes and a transformer secondary coil with a center tap. It can use the positive and negative half cycles of the alternating current for rectification, outputting a smoother direct current and improving the current utilization rate.

[0101] 3) Bridge rectifier: Using four diodes to form a bridge structure can also achieve the effect of full-wave rectification.

[0102] 4) Synchronous rectification: Synchronous rectification technology is used to control the switching timing of the rectifier devices to achieve synchronization of voltage and current waveforms during the rectification process, reduce switching losses, and improve rectification efficiency.

[0103] The analog-to-digital conversion module is used to receive a DC signal and convert the DC signal into a digital signal.

[0104] Specifically, the analog-to-digital conversion module can be implemented using an ADC (Analog-to-Digital Converter). The working principle of the ADC includes the following steps:

[0105] 1) Obtain discrete samples of an analog signal (in this application, a DC signal) at a certain time interval.

[0106] 2) Between two samples, the sample value remains unchanged for quantization and encoding.

[0107] 3) Convert the sampled analog signal into discrete digital values.

[0108] 4) Represent the quantized sample value as binary data to obtain a digital signal.

[0109] The processing module is used to filter the received digital signal and calculate the leakage current value based on the digital signal; when the leakage current value is greater than the preset circuit breaker threshold, a trip signal is sent to the relay module.

[0110] The relay module is used to perform a circuit disconnection operation when receiving a trip signal.

[0111] Among them, the relay module can be implemented by a relay. When the relay receives a trip signal, the coil of the relay generates magnetic force according to the trip signal. This magnetic force drives the switch to trip, thereby cutting off the circuit.

[0112] The above embodiment provides an anti-interference leakage protection device. Different from the conventional leakage protector that directly judges the leakage current signal of the analog signal, the present application adds an analog-to-digital conversion module behind the rectifier module to convert the DC signal of the analog signal into a digital signal, and then uses the processing module to filter the digital signal and calculate the leakage current value. It fully utilizes the advantages of digital signals having high stability during transmission and not being easily affected by electromagnetic interference, thereby avoiding the situation where the leakage protection device is subject to electromagnetic interference, resulting in false detection and false triggering and tripping.

[0113] In some embodiments, the processing module is specifically used to obtain the range value of the sensing module, the gain value of the amplification module and the number of bits of the analog-to-digital conversion module; and calculate the leakage current value based on the range value, gain value, number of bits and digital signal.

[0114] Specifically, the leakage current value of the digital signal can be calculated based on the following formula:

[0115]

[0116] Where I is the calculated leakage current value, is a digital signal, G is the gain value of the amplifier module, B is the number of bits of the ADC, and A is the current range that the current transformer can monitor. If the full scale is 50A, then A is 50.

[0117] For example, consider a 12-bit ADC with an input voltage range of 0 to 5 volts (5V corresponds to the maximum value of the 12-bit ADC), a current transformer with a full-scale range of 5 amps, and a gain of 1 (1V output corresponds to 1A of actual current). For each digital value D read by the ADC, the corresponding leakage current value I can be calculated using the following formula:

[0118]

[0119] The above embodiment calculates the leakage current value based on the parameters of the analog-to-digital conversion module, the amplification module, and the sensing module, which can restore the true situation of the leakage signal before amplification and ensure the authenticity and accuracy of the leakage judgment.

[0120] See Figure 2 In some embodiments, the analog-to-digital conversion module includes a first analog-to-digital conversion unit, a second analog-to-digital conversion unit, and a judgment unit; the first analog-to-digital conversion unit is used to convert the DC signal into a first current signal.

[0121] The second analog-to-digital conversion unit is used to convert the DC signal into a second current signal.

[0122] The judgment unit is used to calculate the difference between the first current signal and the second current signal; when the difference is less than a preset calibration threshold, the average value of the first current signal and the second current signal is used as the digital signal.

[0123] Specifically, during actual use, the analog-to-digital converter may produce some nonlinearity and errors due to the increase in usage time, resulting in a deterioration in the conversion accuracy of the digital signal, affecting the calculation and detection accuracy of the digital signal by the processing module. Therefore, the present application adopts two independent analog-to-digital conversion units, that is, two independent analog-to-digital converters to quantize the same DC signal, and the judgment unit calculates the difference between the two output current signals to determine whether the current analog-to-digital conversion module has errors and nonlinearity. If it is less than the preset calibration threshold, it means that the two analog-to-digital converters are very stable. At this time, the average value of the two output current signals is taken as the digital signal to improve the accuracy of the digital signal.

[0124] Furthermore, the judgment unit is further configured to perform parameter calibration on the first analog-to-digital conversion unit and the second analog-to-digital conversion unit according to the difference when the difference is greater than a preset calibration threshold.

[0125] Specifically, since the analog-to-digital conversion module usually has errors and nonlinearities during use, in order to ensure that it does not affect the operation of the leakage protection device, the analog-to-digital conversion module can use a background calibration algorithm, such as the LMS (least mean square) algorithm, for calibration when errors and nonlinearities occur. This calibration method will not affect the analog-to-digital converter's conversion of DC signals, but is performed silently in the background, by iteratively adjusting the weight coefficient until the outputs of the two ADCs match (that is, the difference in the output signals is less than the preset calibration threshold), thereby achieving calibration.

[0126] The background calibration algorithm in the above embodiment can perform parameter calibration in real time without interrupting the normal operation of the ADC, thereby improving the accuracy and robustness of analog-to-digital conversion and avoiding false leakage detection caused by ADC errors.

[0127] In one embodiment, the processing module is specifically configured to filter the digital signal using a median filter algorithm or a Kalman filter.

[0128] Among them, median filtering is a nonlinear digital filtering technology, which is usually used to remove noise from signals, especially salt-and-pepper noise, while keeping the edge information from being greatly affected.

[0129] The Kalman filter is a highly efficient recursive filter that can estimate the operating state of a leakage protection device from noisy measurements. Its core concept is to use a dynamic model combined with measurement data to perform state estimation in a recursive manner. The Kalman filter is particularly well-suited to environments with noise and uncertainty.

[0130] Specifically, in the Kalman filter, the state estimate and the error covariance matrix Initialize.

[0131] Then perform state prediction and error covariance prediction:

[0132]

[0133]

[0134] Finally calculate the Kalman gain , update state estimates and error covariance:

[0135]

[0136]

[0137]

[0138] Where A is the state transition matrix, B is the control matrix, u is the control vector, Q is the process noise covariance matrix, H is the measurement matrix, R is the measurement noise covariance matrix, z is the measurement value, K is the Kalman gain, and I is the identity matrix.

[0139] In some embodiments, the processing module is further used to extract the current current characteristics of the digital signal; calculate the similarity between the current current characteristics and the historical leakage current characteristics; and calculate the leakage current value based on the digital signal when the similarity is greater than a preset similarity threshold.

[0140] Among them, the current current characteristics and the historical leakage current characteristics can be obtained by Fourier algorithm analysis.

[0141] Specifically, power supply voltage fluctuations and other conditions may cause an instantaneous current difference between the current of the live wire and the current of the neutral wire, thereby generating an induced current; although the filtering processing in the analog-to-digital conversion module and the processing module of the present application can prevent the leakage protection device from being affected by electromagnetic interference and causing false detection, the accuracy of leakage signal recognition caused by power supply fluctuations and circulating current is not high. This is because electromagnetic interference is usually caused by the start-up of high-power electrical appliances or the presence of industrial control electronic equipment such as inverters and soft starters near the leakage protection device, which may generate an interfering magnetic field, thereby affecting the waveform and parameters of the leakage signal.

[0142] However, the effects of power supply fluctuations and circulating currents are not directly added to the actual leakage signal. Instead, they cause the sensing module to directly generate a "leakage signal" from scratch. This "leakage signal" does not contain any real leakage signal components, so it cannot be accurately identified through operations such as analog-to-digital conversion and filtering alone, resulting in malfunction of the relay module.

[0143] Therefore, after filtering, the processing module of this application determines whether the denoised signal is a real leakage signal rather than a signal caused by other reasons by calculating the similarity with the real historical leakage current characteristics. If the similarity is high enough, it means that it is a real leakage signal, and then the leakage current value is calculated, which further reduces the probability of false detection.

[0144] In some embodiments, the processing module is further configured to:

[0145] The historical leakage current signal is input into the trained generative adversarial network to obtain a leakage current signal set; the leakage current signal set and the normal current signal set are used to train the void convolutional neural network to obtain a current recognition model; the digital signal is input into the current recognition model to obtain a recognition result; when the recognition result is true, the leakage current value is calculated based on the digital signal.

[0146] Among them, the signal types of the historical leakage current signal, the leakage current signal set and the normal current signal set are all digital types, and a recognition result of true indicates that the model recognizes that the input digital signal is a leakage current.

[0147] Specifically, although the feature similarity calculation method can filter out some non-leakage signals, in the actual operation of the circuit, the leakage situation is not static. Judging only based on a certain real historical leakage current feature may be too one-sided. While reducing the probability of false detection, it may increase the probability of missed detection, that is, the recognition accuracy is weak.

[0148] Therefore, the present application first uses a trained generative adversarial network to generate a leakage current signal set from real historical leakage current signals, and the leakage current signal set includes a variety of different real leakage current signals.

[0149] Then, a data-rich set of leakage current signals and a set of normal current signals (normal current signals are extremely easy to obtain, so there is no need to use a generative adversarial network) are used to train a void convolutional neural network to obtain a current recognition model.

[0150] Finally, the current identification model is used to determine whether the currently received digital signal is a real leakage signal.

[0151] It is understandable that the current recognition model can learn the characteristics of the real leakage signal during training, and can still correctly identify the leakage signal even if it is slightly affected by changes in the working environment.

[0152] The above embodiment first utilizes a generative adversarial network to generate a variety of leakage current signals of different types as training samples, thereby improving the generalization ability and recognition accuracy of the current recognition model.

[0153] See Figure 3 In some embodiments, the device further includes a sensing module.

[0154] The sensor module is used to obtain the working environment parameters of the leakage protection device.

[0155] Among them, the sensing module includes infrared sensor, temperature sensor, humidity sensor, etc.

[0156] Working environment parameters include infrared radiation value, temperature value, humidity value, etc.

[0157] The processing module is further used to process the current characteristics of the digital signal based on the gain value of the amplification module to obtain the actual current characteristics; and send the working environment parameters, the actual current characteristics and the leakage current value to the mobile terminal via wireless communication.

[0158] Specifically, the digital signal is obtained based on the amplified leakage signal, and therefore it is necessary to process the current characteristics according to the gain value of the amplified signal to obtain the actual current characteristics before the leakage signal is amplified.

[0159] This application adds a wireless communication function to the processing module, realizes remote communication with the user's mobile terminal, ensures that the user can obtain the current value and current characteristics of each actual leakage, and facilitates the user to repair and process the circuit.

[0160] In some embodiments, the processing module is further configured to:

[0161] Obtain a historical leakage current signal and initialize the network parameters of a generative adversarial network; extract the historical leakage current characteristics of the historical leakage current signal and the working environment parameters of the leakage protection device; enable the generator to generate a false leakage current signal based on the historical leakage current characteristics and the working environment parameters; input the false leakage current signal and the corresponding historical leakage current signal into a discriminator to obtain a first discrimination probability corresponding to the false leakage current signal and a second discrimination probability corresponding to the historical leakage current signal.

[0162] Determine whether the first discrimination probability is equal to the second discrimination probability; if not, update the network parameters of the generator and the discriminator according to the first discrimination probability and the second discrimination probability; if so, end the training and obtain a trained generative adversarial network.

[0163] Specifically, changes in the working environment of the leakage protection device are a major factor affecting the induction of leakage signals. Therefore, this application uses the working environment parameters of the leakage protection device as variables, and lets the generator generate false leakage current signals to train the discriminator in the generative adversarial network. When the discriminator cannot distinguish between false leakage current signals and historical leakage current signals, that is, when it judges that the probability of these two current signals being true is the same, it means that the generator has been able to generate data close to the real leakage signal.

[0164] After the training is completed, the generator is continued to generate a false leakage current signal, and the false leakage current signal generated at this time is put into the leakage current signal set together with the historical leakage current signal.

[0165] Furthermore, the processing module is further configured to:

[0166] After sending the trip signal, user feedback information is received; if the user feedback information is the preset reset information, the digital signal is placed in the leakage current signal set, otherwise it is placed in the normal current signal set.

[0167] Specifically, the present application also sets up a user feedback mechanism in the leakage protection device. The user can send user feedback information using the button on the leakage protection device or through remote communication of the mobile terminal. When the processing module receives the user feedback information, it first controls the relay module to restore the circuit path, and then determines whether the user feedback information is the preset reset information. The preset reset information indicates that the leakage signal is a real leakage signal, that is, the recognition is correct this time. The digital signal is placed in the leakage current signal set, and the trained current recognition model is updated to ensure the dynamic update of the current recognition model.

[0168] In some embodiments, the processing module is further used to record the leakage protection response time starting from the moment the digital signal is received and ending at the moment the trip signal is sent; and send the leakage protection response time to the mobile terminal via wireless communication.

[0169] Specifically, the leakage protection response time can also be sent to the user as part of the leakage measurement for storage, so that the user can promptly find out whether the response time of the leakage protection device is too long and needs to be replaced.

[0170] See Figure 4 Another embodiment of the present application provides an anti-interference leakage protection method, including:

[0171] In step S11 , the amplifying module amplifies the received leakage signal to obtain an amplified signal.

[0172] In step S12, the rectifier module rectifies the amplified signal to obtain a DC signal.

[0173] Step S13: the analog-to-digital conversion module converts the DC signal into a digital signal.

[0174] Specifically, discrete samples of an analog signal (a DC signal in this application) are acquired within a certain time interval; between two samples, the sampled value is kept unchanged for quantization and encoding; the sampled analog signal is converted into discrete digital values; the quantized sample values are represented as binary data to obtain a digital signal.

[0175] In step S14, the processing module filters the received digital signal; calculates the leakage current value based on the digital signal; determines whether the leakage current value is greater than a preset circuit breaker threshold, and if so, sends a trip signal to the relay module.

[0176] Step S15: When the relay module receives the trip signal, it performs a circuit disconnection operation.

[0177] In some embodiments, the above-mentioned calculation of the leakage current value based on the digital signal includes:

[0178] Step S141, obtaining the range value of the sensing module, the gain value of the amplifying module, and the number of bits of the analog-to-digital conversion module.

[0179] Step S142, calculating an exponent with 2 as the base and the number of bits as the power.

[0180] In step S143 , the exponent is decremented by 1 and then multiplied by the gain value to obtain a first product.

[0181] In step S144 , the digital signal is multiplied by the range value to obtain a second product.

[0182] Step S145 , dividing the second product by the first product to obtain a leakage current value.

[0183] Specifically, the leakage current value of the digital signal can be calculated based on the following formula:

[0184]

[0185] Where I is the calculated leakage current value, is a digital signal, G is the gain value of the amplifier module, B is the number of bits of the ADC, and A is the current range that the current transformer can monitor. If the full scale is 50A, then A is 50.

[0186] For example, consider a 12-bit ADC with an input voltage range of 0 to 5 volts (5V corresponds to the maximum value of the 12-bit ADC), a current transformer with a full-scale range of 5 amps, and a gain of 1 (1V output corresponds to 1A of actual current). For each digital value D read by the ADC, the corresponding leakage current value I can be calculated using the following formula:

[0187]

[0188] The above embodiment calculates the leakage current value based on the parameters of the analog-to-digital conversion module, the amplification module, and the sensing module, which can restore the true situation of the leakage signal before amplification and ensure the authenticity and accuracy of the leakage judgment.

[0189] In some embodiments, the method further comprises:

[0190] In step S21 , the processing module extracts the current characteristics of the digital signal.

[0191] Step S22: Calculate the similarity between the current current characteristics and the historical leakage current characteristics.

[0192] Step S23 : when the similarity is greater than a preset similarity threshold, calculating the leakage current value based on the digital signal.

[0193] Among them, the current current characteristics and the historical leakage current characteristics can be obtained by Fourier algorithm analysis.

[0194] Specifically, although the filtering processing of the present application can prevent the leakage protection device from being affected by electromagnetic interference and causing false detection, the accuracy of identifying leakage signals caused by power supply fluctuations and circulating currents is not high. This is because electromagnetic interference is usually caused by the start-up of high-power electrical appliances or the presence of industrial control electronic equipment such as inverters and soft starters near the leakage protection device, which may generate an interfering magnetic field, thereby affecting the waveform and parameters of the leakage signal.

[0195] However, the impact of power supply fluctuations and circulating currents is not directly added to the actual leakage signal. Instead, the sensing module directly generates a "leakage signal" from scratch. This "leakage signal" does not contain any real leakage signal components, so it cannot be identified through operations such as analog-to-digital conversion and filtering alone, resulting in malfunction of the relay module.

[0196] Therefore, after filtering, this application allows the processing module to determine whether the denoised signal is a real leakage signal rather than a signal caused by other reasons by calculating the similarity with the real historical leakage current characteristics. If the similarity is high enough, it means that it is a real leakage signal, and then the leakage current value is calculated, which further reduces the probability of false detection.

[0197] In some embodiments, the method further comprises:

[0198] In step S31, the processing module inputs the historical leakage current signals into the trained generative adversarial network to obtain a set of leakage current signals. Specifically, while the feature similarity calculation method can filter out some non-leakage signals, in actual circuit operation, leakage conditions are not static. Judging based solely on a single, actual historical leakage current signature can be overly biased, reducing the probability of false detection while potentially increasing the probability of missed detection, resulting in lower recognition accuracy.

[0199] Therefore, the present application first uses a trained generative adversarial network to generate a leakage current signal set from real historical leakage current signals, and the leakage current signal set includes a variety of different real leakage current signals.

[0200] In step S32, a dilated convolutional neural network is trained using the leakage current signal set and the normal current signal set to obtain a current recognition model. The dilated convolutional neural network used here is one-dimensional because one-dimensional dilated convolutional neural networks (1D-DCNN) perform well in processing sequence data.

[0201] Step S33: input the digital signal into the current identification model to obtain an identification result.

[0202] Step S34: when the identification result is true, calculating the leakage current value based on the digital signal.

[0203] The above embodiment first utilizes a generative adversarial network to generate a variety of leakage current signals of different types as training samples, thereby improving the generalization ability and recognition accuracy of the current recognition model.

[0204] See Figure 5 In some embodiments, the method further comprises:

[0205] In step S41 , the processing module obtains a historical leakage current signal and initializes network parameters of a generative adversarial network.

[0206] Specifically, the generative adversarial network includes a generator and a discriminator, and the network parameters are initialized to express the weight parameters and bias parameters of the generator and discriminator as a normal distribution of 0-0.01.

[0207] Step S42 , extracting historical leakage current characteristics of the historical leakage current signal and working environment parameters of the leakage protection device.

[0208] Specifically, the historical leakage current characteristics are extracted as follows:

[0209]

[0210] in, is the historical leakage current characteristic, FFT is the fast Fourier transform operation, is the historical leakage current signal, is the Fourier transform weight, which can be calculated using the least squares method:

[0211]

[0212] in, is the signal value at time t, and N is the number of time points.

[0213] Step S43: Instruct the generator to generate a false leakage current signal based on the historical leakage current characteristics and the working environment parameters.

[0214] Step S44 , inputting the false leakage current signal and the corresponding historical leakage current signal into a discriminator to obtain a first discrimination probability corresponding to the false leakage current signal and a second discrimination probability corresponding to the historical leakage current signal.

[0215] Step S45: determine whether the first discrimination probability is equal to the second discrimination probability.

[0216] Step S46: If not, update the network parameters of the generator and the discriminator according to the first discrimination probability and the second discrimination probability.

[0217] Specifically, changes in the working environment of the leakage protection device are a major factor affecting the induction of leakage signals. Therefore, this application uses the working environment parameters of the leakage protection device as variables, and lets the generator generate false leakage current signals to train the discriminator in the generative adversarial network. When the discriminator cannot distinguish between false leakage current signals and historical leakage current signals, that is, when it judges that the probability of these two current signals being true is the same, it means that the generator has been able to generate data close to the real leakage signal.

[0218] After the training is completed, the generator is continued to generate a false leakage current signal, and the false leakage current signal generated at this time is put into the leakage current signal set together with the historical leakage current signal.

[0219] In some embodiments, the generator generates a false leakage current signal based on historical leakage current characteristics and working environment parameters, which may specifically include the following steps:

[0220] In step S431, the generator calculates the environmental adjustment factor according to the preset variation threshold and the working environment parameters.

[0221] Among them, environmental regulatory factors The calculation method is:

[0222]

[0223] in, is the preset variation threshold, D is the number of types of working environment parameters, is the value of the environmental parameter of type i, is the historical average value of the environmental parameter of type i.

[0224] Step S432: Multiply the environment adjustment factor and the random noise to obtain a noise vector.

[0225] Step S433 : multiplying the historical leakage current feature by the preset extraction weight to obtain a feature vector.

[0226] Step S434: Add the noise vector and the characteristic vector to obtain false current data.

[0227] Specifically, the false current data The calculation formula is:

[0228]

[0229] in, is the preset extraction weight, and n is random noise.

[0230] Step S435 , performing inverse Fourier transform on the false current data to obtain a false leakage current signal.

[0231] The above embodiment generates a false leakage current signal using the working environment parameters as variables, which enables the generator to simulate the situation in which the signal induction of the leakage protection device changes due to changes in the working environment during a specific application process, thereby improving the generalization ability of the current identification model and being able to identify leakage signals affected by the environment.

[0232] In some embodiments, updating the network parameters of the generator and the discriminator according to the first discrimination probability and the second discrimination probability may specifically include the following steps:

[0233] Step S461 , respectively calculating the binary cross entropy loss function of the first discriminant probability and the second discriminant probability.

[0234] In step S462 , the two binary cross entropy loss functions are added together to obtain a discriminant loss function.

[0235] Specifically, the discriminant loss function , where x is the historical leakage current signal, is the second discriminant probability that the discriminator judges to be true; z is the data input to the generator, is the virtual leakage current signal output by the generator, is the first discriminant probability that the discriminator is true, and L is the binary cross entropy loss function.

[0236] Step S463: Calculate the updated network parameters of the discriminator based on the discriminant loss function and the discriminant update gradient.

[0237] Specifically, the parameter update formula of the discriminator can be expressed as:

[0238]

[0239] in, are the parameters of the discriminator, Update the gradient for discrimination.

[0240] Step S464, based on the binary cross entropy loss function of the first discriminant probability and the generated updated gradient, the updated network parameters of the generator are calculated. Specifically, the loss function of the generator is expressed as:

[0241]

[0242] The parameter update formula of the generator can be expressed as:

[0243]

[0244] in, are the parameters of the generator, To generate updated gradients.

[0245] In practical applications, the loss function of the generator Also often used , because when When it is close to 1, the loss of the generator will decrease, that is, the generator hopes that the discriminator will have a higher probability of identifying the samples it generates as real.

[0246] The discriminator and generator reach Nash equilibrium by alternately optimizing their own parameters. At this equilibrium point, the discriminator cannot distinguish between real data and generated data, and the data generated by the generator is as close to the real data distribution as possible.

[0247] In some embodiments, the method further comprises:

[0248] Step S51: After sending the trip signal, the processing module receives user feedback information.

[0249] Step S52: if the user feedback information is preset reset information, the digital signal is placed in the leakage current signal set; otherwise, the digital signal is placed in the normal current signal set.

[0250] Specifically, the present application also sets up a user feedback mechanism in the leakage protection device. The user can use the button on the leakage protection device or send user feedback information through remote communication of the mobile terminal. When the processing module receives the user feedback information, it first controls the relay module to restore the circuit path, and then determines whether the user feedback information is the preset reset information. The preset reset information indicates that the leakage signal is a real leakage signal, that is, the identification is correct this time, and the digital signal is placed in the leakage current signal set.

[0251] Step S53 : training the current recognition model again based on the leakage current signal set and the normal current signal set.

[0252] In some embodiments, the method further comprises:

[0253] The processing module records the leakage protection response time starting from the moment of receiving the digital signal and ending at the moment of sending the trip signal; and sends the leakage protection response time to the mobile terminal via wireless communication.

[0254] Specifically, the leakage protection response time can also be sent to the user as part of the leakage measurement for storage, so that the user can promptly find out whether the response time of the leakage protection device is too long and needs to be replaced.

[0255] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0256] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. An anti-interference leakage protection device, characterized in that: include: The sensing module, amplifying module, rectifying module, analog-to-digital conversion module, processing module and relay module are connected in sequence; The amplifying module is used to receive the leakage signal of the sensing module and amplify it to obtain an amplified signal; The rectifier module is used to receive the amplified signal and rectify the amplified signal to obtain a DC signal; The analog-to-digital conversion module is used to receive a DC signal and convert the DC signal into a digital signal; specifically, the analog-to-digital conversion module includes a first analog-to-digital conversion unit, a second analog-to-digital conversion unit and a judgment unit; the first analog-to-digital conversion unit is used to convert the DC signal into a first current signal; the second analog-to-digital conversion unit is used to convert the DC signal into a second current signal; the judgment unit is used to calculate the difference between the first current signal and the second current signal; when the difference is less than a preset calibration threshold, the average value of the first current signal and the second current signal is used as the digital signal; when the difference is greater than the preset calibration threshold, the first analog-to-digital conversion unit and the second analog-to-digital conversion unit are parameter calibrated according to the difference; The processing module is used to filter the received digital signal and calculate the leakage current value based on the digital signal; when the leakage current value is greater than a preset circuit breaker threshold, send a trip signal to the relay module; The processing module is further configured to obtain a historical leakage current signal and initialize network parameters of a generative adversarial network; extract historical leakage current characteristics of the historical leakage current signal and operating environment parameters of the leakage protection device; The historical leakage current characteristics are extracted as follows: in, is the historical leakage current characteristic, FFT is the fast Fourier transform operation, is the historical leakage current signal, is the Fourier transform weight, which is calculated using the least squares method: in, is the signal value at time t, and N is the number of time points; the generator generates a false leakage current signal based on the historical leakage current characteristics and the working environment parameters; the generator calculates the environmental adjustment factor according to the preset variation threshold and the working environment parameters; wherein the environmental adjustment factor The calculation method is: in, is the preset variation threshold, D is the number of types of working environment parameters, is the value of the environmental parameter of type i, is the historical average value of the environmental parameter of type i; The generator multiplies the environmental adjustment factor and random noise to obtain a noise vector; multiplies the historical leakage current feature and a preset extraction weight to obtain a feature vector; adds the noise vector and the feature vector to obtain false current data; and performs an inverse Fourier transform on the false current data to obtain the false leakage current signal; The processing module is further configured to input the false leakage current signal and the corresponding historical leakage current signal into a discriminator to obtain a first discrimination probability corresponding to the false leakage current signal and a second discrimination probability corresponding to the historical leakage current signal; determine whether the first discrimination probability is equal to the second discrimination probability; and if not, update network parameters of the generator and the discriminator according to the first discrimination probability and the second discrimination probability; The processing module is further configured to input historical leakage current signals into a trained generative adversarial network to obtain a leakage current signal set; train a dilated convolutional neural network using the leakage current signal set and the normal current signal set to obtain a current recognition model; input the digital signal into the current recognition model to obtain a recognition result; when the recognition result is true, obtain the range value of the sensing module, the gain value of the amplification module, and the number of bits of the analog-to-digital conversion module; calculate an exponent with a base of 2 and a power of the number of bits; subtract 1 from the exponent and multiply it by the gain value to obtain a first product; multiply the digital signal by the range value to obtain a second product; and divide the second product by the first product to obtain a leakage current value; The relay module is used to perform a circuit disconnection operation when receiving a trip signal.

2. The anti-interference leakage protection device according to claim 1, characterized in that: The processing module is specifically configured to filter the digital signal using a median filter algorithm or a Kalman filter.

3. The anti-interference leakage protection device according to claim 1, characterized in that: The processing module is further configured to extract current characteristics of the digital signal; Calculating the similarity between the current current feature and the historical leakage current feature; and calculating the leakage current value based on the digital signal when the similarity is greater than a preset similarity threshold.

4. The anti-interference leakage protection device according to claim 1, characterized in that: Also includes a sensor module; The sensor module is used to obtain the working environment parameters of the leakage protection device; The processing module is further configured to process the current characteristics of the digital signal based on the gain value of the amplification module to obtain the actual current characteristics; The working environment parameters, actual current characteristics and leakage current values are sent to the mobile terminal via wireless communication.

5. The anti-interference leakage protection device according to claim 1, characterized in that: The processing module is further configured to: After sending the trip signal, receiving user feedback information; if the user feedback information is preset reset information, putting the digital signal into the leakage current signal set, otherwise putting it into the normal current signal set.

6. The anti-interference leakage protection device according to claim 5, characterized in that: The processing module is further configured to record a leakage protection response time starting from the moment the digital signal is received and ending at the moment the trip signal is sent; and to send the leakage protection response time to a mobile terminal via wireless communication.

7. An anti-interference leakage protection method, characterized in that: include: The amplifying module amplifies the received leakage signal to obtain an amplified signal; The rectifier module rectifies the amplified signal to obtain a DC signal; The analog-to-digital conversion module converts the DC signal into a digital signal; specifically, the analog-to-digital conversion module includes a first analog-to-digital conversion unit, a second analog-to-digital conversion unit, and a judgment unit; the first analog-to-digital conversion unit converts the DC signal into a first current signal; the second analog-to-digital conversion unit converts the DC signal into a second current signal; the judgment unit calculates a difference between the first current signal and the second current signal; when the difference is less than a preset calibration threshold, an average value of the first current signal and the second current signal is used as the digital signal; when the difference is greater than the preset calibration threshold, parameter calibration is performed on the first analog-to-digital conversion unit and the second analog-to-digital conversion unit according to the difference; The processing module filters the received digital signal; calculates the leakage current value based on the digital signal; determines whether the leakage current value is greater than a preset circuit breaker threshold, and if so, sends a trip signal to the relay module; The processing module obtains a historical leakage current signal and initializes network parameters of a generative adversarial network; extracts historical leakage current characteristics of the historical leakage current signal and working environment parameters of the leakage protection device; The historical leakage current characteristics are extracted as follows: in, is the historical leakage current characteristic, FFT is the fast Fourier transform operation, is the historical leakage current signal, is the Fourier transform weight, which is calculated using the least squares method: in, is the signal value at time t, and N is the number of time points; the generator generates a false leakage current signal based on the historical leakage current characteristics and the working environment parameters; the generator calculates the environmental adjustment factor according to the preset variation threshold and the working environment parameters; wherein the environmental adjustment factor The calculation method is: in, is the preset variation threshold, D is the number of types of working environment parameters, is the value of the environmental parameter of type i, is the historical average value of the environmental parameter of type i; The generator multiplies the environmental adjustment factor and random noise to obtain a noise vector; multiplies the historical leakage current feature and a preset extraction weight to obtain a feature vector; adds the noise vector and the feature vector to obtain false current data; and performs an inverse Fourier transform on the false current data to obtain the false leakage current signal; The processing module inputs the false leakage current signal and the corresponding historical leakage current signal into the discriminator to obtain a first discrimination probability corresponding to the false leakage current signal and a second discrimination probability corresponding to the historical leakage current signal; determines whether the first discrimination probability is equal to the second discrimination probability; if not, updates the network parameters of the generator and the discriminator according to the first discrimination probability and the second discrimination probability; Specifically, the processing module inputs the historical leakage current signal into a trained generative adversarial network to obtain a leakage current signal set; uses the leakage current signal set and the normal current signal set to train a void convolutional neural network to obtain a current recognition model; inputs the digital signal into the current recognition model to obtain a recognition result; when the recognition result is true, obtains the range value of the sensing module, the gain value of the amplification module, and the number of bits of the analog-to-digital conversion module; calculates an exponent with 2 as the base and the number of bits as the power; subtracts 1 from the exponent and multiplies it by the gain value to obtain a first product; multiplies the digital signal by the range value to obtain a second product; and divides the second product by the first product to obtain the leakage current value; When the relay module receives the trip signal, it performs a circuit disconnection operation.

8. The anti-interference leakage protection method according to claim 7, characterized in that: Also includes: The processing module extracts the current characteristics of the digital signal; Calculating the similarity between the current current characteristic and the historical leakage current characteristic; When the similarity is greater than a preset similarity threshold, a leakage current value is calculated based on the digital signal.

9. The anti-interference leakage protection method according to claim 7, characterized in that: The updating of network parameters of the generator and the discriminator according to the first discrimination probability and the second discrimination probability includes: Calculate the binary cross entropy loss function of the first discriminant probability and the second discriminant probability respectively; Add the two binary cross entropy loss functions to get the discriminant loss function; Calculate updated network parameters of the discriminator based on the discriminant loss function and the discriminant update gradient; The updated network parameters of the generator are calculated based on the binary cross entropy loss function of the first discriminant probability and the generated update gradient.

10. The anti-interference leakage protection method according to claim 7, characterized in that: Also includes: The processing module receives user feedback information after sending the trip signal; If the user feedback information is preset reset information, the digital signal is placed in the leakage current signal set; otherwise, the digital signal is placed in the normal current signal set; The current identification model is trained again based on the leakage current signal set and the normal current signal set.

11. The anti-interference leakage protection method according to claim 7, characterized in that: Also includes: The processing module records the leakage protection response time starting from the moment of receiving the digital signal and ending at the moment of sending the trip signal; The leakage protection response time is sent to the mobile terminal via wireless communication.

Images