Electrical safety monitoring methods, systems, electronic devices and storage media

Abstract

This invention provides a method, system, electronic device, and storage medium for monitoring electrical safety. The method includes extracting and analyzing load characteristics to obtain a feature dataset; detecting the feature dataset based on a calculation window, a detection extraction window, and start / stop thresholds to obtain start / stop event characteristics; identifying the preliminary load type based on the start / stop event characteristics; performing hierarchical channel identification of the preliminary load type based on the start and end points; using the difference in the center frequency of each modal component as an arc feature; and forming a feature vector; identifying the arc feature based on a trained integrated deep RVFL neural network to obtain electrical safety information, and detecting the determined load type. This invention can effectively reduce the occurrence of inadequate monitoring of electrical equipment, effectively avoid misjudgments and omissions, and also avoid the problem of unmonitored series arcs due to their high randomness.

Description

Technical Field

[0001] This invention relates to the field of electricity monitoring technology, and in particular to an electricity safety monitoring method, system, electronic device, and storage medium. Background Technology

[0002] People are using electrical equipment in increasingly diverse and widespread ways, but electrical safety and regulations are often overlooked. Furthermore, improper use of electricity is a significant contributing factor to electrical fires. Monitoring electrical safety often relies on measuring threshold quantities such as leakage current and temperature to determine if an electrical fault has occurred. However, this method is primarily effective only after a fault has occurred or when its characteristics are relatively obvious, and it falls short in detecting potential electrical safety hazards and concealed electrical faults.

[0003] Existing technologies cannot identify the types of electrical appliances, and there are errors in the detection of low-power appliances and complex environments, which can easily lead to misjudgment and missed judgment. Due to the complexity of low-voltage AC electrical systems, the high randomness of series arcs, and the increase of various nonlinear loads in the circuit, the characteristics of fault arcs in the circuit may vary greatly under different load conditions and at different times. In addition, due to the limitation of hardware costs, complex algorithms are difficult to implement in practical engineering, resulting in poor power safety monitoring and the inability to provide timely warnings. Summary of the Invention

[0004] Therefore, the purpose of this invention is to provide an electrical safety monitoring method, system, electronic device, and storage medium to address the shortcomings of the prior art.

[0005] In a first aspect, the present invention provides a method for monitoring electrical safety, the method comprising:

[0006] Real-time collection of electricity consumption data, acquisition of load characteristics from the electricity consumption data, extraction and analysis of the load characteristics to obtain a feature dataset;

[0007] A sliding detection window is defined in the electricity consumption time series, and a start-stop threshold is set. The sliding detection window includes a calculation window and a detection extraction window. The feature dataset is detected based on the calculation window, the detection extraction window, and the start-stop threshold to obtain start-stop event features.

[0008] An SVM model is used in conjunction with a second-order oscillating particle swarm optimization algorithm, and the initial load type is identified based on the characteristics of the start-stop events using multi-feature loads.

[0009] A movement strategy for the detection and extraction window is set, and the detection and extraction window is moved based on the movement strategy to perform hierarchical channel identification on the preliminary load type in order to obtain a determined load type;

[0010] The zero-crossing points in the feature dataset are used as the current data of the series arc. The current data is normalized to obtain the processed current data. The processed current data is then decomposed using VDM to obtain the center frequency of each modal component. The difference in the center frequency of each modal component is used as the arc feature.

[0011] Extract the current increment coefficient and the fuzzy entropy value of each modal component from the arc features, form a feature vector based on the fuzzy entropy value and the center frequency, and input the feature vector into an integrated deep RVFL neural network so that the integrated deep RVFL neural network is trained based on the feature vector to obtain the trained integrated deep RVFL neural network.

[0012] The electric arc features are identified based on the trained integrated deep RVFL neural network to obtain electrical safety information, and the determined load type is detected by the continuous moving change method to obtain the current status type.

[0013] Compared with existing technologies, the beneficial effects of this invention are as follows: By detecting and extracting the feature dataset of the start-stop event characteristics, and combining the SVM model with the second-order oscillating particle swarm optimization algorithm, the initial load type of the electrical equipment can be obtained, thereby obtaining the load status of the electrical equipment and effectively reducing the situation of inadequate monitoring of the electrical equipment. By using the difference in center frequency as the arc feature, the situation where other features cannot reflect the electrical equipment when monitoring the power consumption of the electrical equipment is avoided. Furthermore, by performing hierarchical channel identification through the detection and extraction window, the load type is determined. And by identifying the arc feature through the trained integrated deep RVFL neural network, the situation of false judgment and false omission can be effectively avoided. It can also avoid the problem of high randomness of series arcs leading to the inability to monitor them, thereby avoiding the problem of difficulty in monitoring the fault arc features of the line under different load conditions and at different times.

[0014] Furthermore, the steps of real-time acquisition of electricity consumption data, obtaining load characteristics from the electricity consumption data, extracting and analyzing the load characteristics to obtain a feature dataset include:

[0015] The system collects electrical signals based on sensors and converts the electrical signals by a preset factor to collect current data.

[0016] The steady-state characteristics of the current are obtained based on the current amplitude, current harmonic components, steady-state active power, and steady-state reactive power in the current data.

[0017] The transient characteristics of the current are obtained based on the transient current, transient active power, transient reactive power, voltage noise, and duration in the current data.

[0018] Furthermore, the calculation expression for the steady-state active power is as follows:

[0019]

[0020] In the formula, P represents steady-state active power, k represents the harmonic order, and V represents the harmonic power. k I represents the effective value of the k-th voltage harmonic. k θ represents the effective value of the k-th current harmonic. k This represents the phase difference of the k-th harmonic;

[0021] The formula for calculating the steady-state reactive power is:

[0022]

[0023] In the formula, Q represents steady-state reactive power.

[0024] Furthermore, the step of defining a sliding detection window in the electricity consumption time series and setting a start / stop threshold, wherein the sliding detection window includes a calculation window and a detection extraction window, and detecting the feature dataset based on the calculation window, the detection extraction window, and the start / stop threshold to obtain start / stop event features includes:

[0025] Define a calculation window and a detection extraction window within the electricity consumption time series, and set the window length of the calculation window and the detection extraction window;

[0026] The effective average current of the feature dataset is calculated based on the calculation window;

[0027] Set a start / stop threshold and compare the start / stop threshold with the effective average current. If the effective average current is greater than the start / stop threshold, it is determined that a transient event has occurred, and the starting point of the power consumption event is recorded.

[0028] The detection extraction window is continuously slid, and current data for several cycles in the feature dataset is detected based on the slid detection extraction window to detect voltage transient values. Start-stop event features in the feature dataset are obtained based on the starting point of the power consumption event and the voltage transient values.

[0029] Furthermore, the calculation expression for the start / stop threshold is as follows:

[0030]

[0031] In the formula, H min H maxL1 and L2 represent the maximum and minimum values ​​of the start / stop threshold, respectively; L1 represents the number of sampling points where the fluctuation increases; L2 represents the number of sampling points where the current of the electrical equipment increases; L represents the length of the calculation window; and I represents the maximum and minimum values ​​of the start / stop threshold, respectively. min λ represents the minimum amplitude of change in the identified electrical equipment, f represents the average value of the current effective value fluctuation and the safety margin, and λ represents the noise level.

[0032] Furthermore, the step of employing an SVM model combined with a second-order oscillating particle swarm optimization algorithm and identifying the preliminary load type based on multi-feature load characteristics of the start-stop event includes:

[0033] Each particle in the second-order oscillating particle swarm is assigned a random position and a random velocity, and the velocity magnitude and velocity direction of each particle are updated by comparing the fitness value, the particle's optimal value, and the global optimal value of the objective function of the second-order oscillating particle swarm algorithm.

[0034] The motion of each particle is corrected based on the updated velocity magnitude and velocity direction, so that each particle gradually approaches the global optimum. Based on the global optimum and the SVM model, the initial load type of the start-stop event characteristics is identified.

[0035] Furthermore, the step of setting a movement strategy for the detection extraction window, moving the detection extraction window based on the movement strategy, and performing hierarchical channel identification on the preliminary load type to determine the load type includes:

[0036] The movement strategy of the detection extraction window is set according to the current trajectory map of the feature dataset, so that the detection extraction window covers the current trajectory map based on the movement strategy;

[0037] Based on the detection extraction window and the effective value of the current, high-power loads in the start-stop event characteristics are screened.

[0038] The load type is determined based on the steady-state fundamental frequency, harmonic amplitude, transient active power, transient reactive power, and transient duration.

[0039] Secondly, the present invention also provides an electricity safety monitoring system, the system comprising:

[0040] The real-time acquisition module is used to acquire electricity consumption data in real time, obtain load characteristics from the electricity consumption data, extract and analyze the load characteristics to obtain a feature dataset;

[0041] A detection module is defined to define a sliding detection window in the electricity consumption time series and set a start-stop threshold. The sliding detection window includes a calculation window and a detection extraction window. The feature dataset is detected based on the calculation window, the detection extraction window and the start-stop threshold to obtain start-stop event features.

[0042] Combined with the identification module, it is used to identify the preliminary load type of the start-stop event characteristics based on the SVM model and the second-order oscillating particle swarm algorithm and multi-feature load.

[0043] A detection module is configured to set a movement strategy for the detection extraction window, move the detection extraction window based on the movement strategy, and perform hierarchical channel identification on the preliminary load type to obtain a determined load type.

[0044] The statistical processing module is used to statistically analyze the zero-crossing points in the feature dataset as current data of the series arc, normalize the current data to obtain processed current data, and perform VDM decomposition on the processed current data to obtain the center frequency of each modal component, and use the difference of the center frequency of each modal component as the arc feature.

[0045] An extraction training module is used to extract the current increment coefficient and the fuzzy entropy value of each modal component in the arc features, form a feature vector based on the fuzzy entropy value and the center frequency, and input the feature vector into an integrated deep RVFL neural network so that the integrated deep RVFL neural network is trained based on the feature vector to obtain a trained integrated deep RVFL neural network.

[0046] The identification and detection module is used to identify the arc features based on the trained integrated deep RVFL neural network to obtain electrical safety information, and to detect the determined load type through the continuous moving change method to obtain the current status type.

[0047] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described power safety monitoring method.

[0048] Fourthly, the present invention provides a storage medium on which a computer program is stored, which, when executed by a processor, implements the above-described method for monitoring electrical safety. Attached Figure Description

[0049] Figure 1 This is a flowchart of the electricity safety monitoring method in the first embodiment of the present invention;

[0050] Figure 2This is a structural block diagram of the electricity safety monitoring system in the second embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of the hardware structure of the electronic device in the third embodiment of the present invention.

[0052] Explanation of key component symbols:

[0053] 10. Real-time data acquisition module;

[0054] 20. Define the detection module;

[0055] 30. Combined with the recognition module;

[0056] 40. Configure the detection module;

[0057] 50. Statistical Processing Module;

[0058] 60. Extract the training module;

[0059] 70. Identification and detection module;

[0060] 80. Bus; 81. Processor; 82. Memory; 83. Communication interface.

[0061] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0062] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate several embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0063] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0065] Example 1

[0066] Please see Figure 1 The figure shows an electricity safety monitoring method according to the first embodiment of the present invention, the method comprising steps S1 to S7:

[0067] S1, collect electricity consumption data in real time, obtain load characteristics from the electricity consumption data, extract and analyze the load characteristics to obtain a feature dataset;

[0068] Specifically, step S1 includes steps S11 to S13:

[0069] S11, Based on the acquisition sensor, the power consumption signal is acquired, and the power consumption signal is converted by a preset multiple to acquire current data;

[0070] Understandably, data acquisition devices for electrical equipment collect measurement values, including voltage and current. Since the actual operating voltage for residential electrical loads is generally 220V, it is necessary to convert the actual voltage and current data to the appropriate multiples to ensure accurate load data acquisition.

[0071] It is worth noting that, in this embodiment, the conversion expression is:

[0072] V0=G×(V IN +-V IN- );

[0073]

[0074] In the formula, V0 represents the output voltage of the operational amplifier, V IN+ V represents the voltage at the non-inverting input terminal. IN- Represents the voltage at the inverting input terminal, G represents the amplification factor, and R represents the voltage at the inverting input terminal. G This indicates the resistance at the non-inverting and inverting input terminals.

[0075] S12, obtain the steady-state characteristics of the current based on the current amplitude, current harmonic components, steady-state active power and steady-state reactive power in the current data;

[0076] It is understood that steady-state characteristics mainly refer to the continuous characteristics exhibited by electrical equipment during steady-state operation, such as the effective value of current, current amplitude, current harmonic components, steady-state active power, steady-state reactive power, and the VI curve. In this embodiment, the calculation expression for the steady-state active power is:

[0077]

[0078] In the formula, P represents steady-state active power, k represents the harmonic order, and V represents the harmonic power. k I represents the effective value of the k-th voltage harmonic.k θ represents the effective value of the k-th current harmonic. k This represents the phase difference of the k-th harmonic;

[0079] The formula for calculating the steady-state reactive power is:

[0080]

[0081] In the formula, Q represents steady-state reactive power.

[0082] S13, Obtain the transient characteristics of the current based on the transient current, transient active power, transient reactive power, voltage noise and duration in the current data;

[0083] It is understandable that transient characteristics are mainly the changes in electrical quantities such as current, voltage, and power during the state transition of electrical equipment, including transient current, transient active power, transient reactive power, voltage noise, and duration.

[0084] S2, Define a sliding detection window in the electricity consumption time series and set a start-stop threshold. The sliding detection window includes a calculation window and a detection extraction window. Detect the feature dataset based on the calculation window, the detection extraction window and the start-stop threshold to obtain start-stop event features.

[0085] It needs to be explained that the feature dataset can be modeled as a statistical graph of changing states. Therefore, by defining two sliding windows within the time series of electricity consumption, namely the calculation window and the detection and extraction window, the feature dataset modeled as a statistical graph of changing states can be detected through the calculation window and the detection and extraction window. In conjunction with the start and stop threshold, the start and stop event characteristics of electrical equipment can be detected, that is, the current and voltage change characteristics when the electrical equipment starts and stops.

[0086] Specifically, step S2 includes steps S21 to S24:

[0087] S21, Define a calculation window and a detection extraction window within the electricity consumption time series, and set the window length of the calculation window and the detection extraction window;

[0088] It is understandable that a calculation window and a detection extraction window are defined during the time period when the device is consuming electricity, and the length of the calculation window and the detection extraction window are set according to the size of the feature dataset modeled as a statistical graph of changing states.

[0089] S22, Calculate the effective average current of the feature dataset based on the calculation window;

[0090] It is understandable that by traversing the feature data in the change state statistics chart through the calculation window, the effective average current of the electrical equipment during the power consumption period can be calculated. The specific calculation process is to statistically analyze the power consumption current at each time point in the change state statistics chart, and then calculate the effective average current.

[0091] S23, set a start / stop threshold and compare the start / stop threshold with the effective average current. If the effective average current is greater than the start / stop threshold, it is determined that a transient event has occurred, and the starting point of the power consumption event is recorded.

[0092] It is understood that, in this embodiment, the calculation expression for the start / stop threshold is:

[0093]

[0094] In the formula, H min H max L1 and L2 represent the maximum and minimum values ​​of the start / stop threshold, respectively; L1 represents the number of sampling points where the fluctuation increases; L2 represents the number of sampling points where the current of the electrical equipment increases; L represents the length of the calculation window; and I represents the maximum and minimum values ​​of the start / stop threshold, respectively. min λ represents the minimum amplitude of change in the identified electrical equipment, f represents the average value of the current effective value fluctuation and the safety margin, and λ represents the noise level.

[0095] It should be noted that by comparing the start-stop threshold with the effective average current, that is, by comparing the effective average current of the electrical equipment with the start and stop current thresholds, when the start-stop threshold is less than the effective average current, it is determined that a transient event has occurred, and the time point of the transient event is calculated and recorded as the starting point of the electricity consumption time.

[0096] S24, continuously slide the detection extraction window, and detect current data of several cycles in the feature dataset based on the sliding detection extraction window to detect voltage transient values, and obtain start-stop event features in the feature dataset based on the starting point of the electricity consumption event and the voltage transient values.

[0097] Understandably, by repeatedly looping through the feature data in the statistical chart of changing states within the detection extraction window, current data for several power consumption cycles of the electrical equipment can be extracted. Based on the current data from these cycles, the transient voltage values ​​of the electrical equipment can be detected. Current data has more electrical characteristics than power data. Subtracting the current before and after an event requires them to be in phase. Since the phase of the mains voltage is relatively stable during operation, the phase is unified using the transient voltage value sampled synchronously with the current. The data before and after the event are both taken with the voltage cycle value crossing zero as the starting phase of the total current. Five cycles of current data are used to extract steady-state load characteristics, while transient characteristics are extracted from the current data between the load start and end points.

[0098] S3, adopting an SVM model and combining it with a second-order oscillating particle swarm optimization algorithm, and identifying the preliminary load type of the start-stop event characteristics based on multi-feature load;

[0099] Specifically, step S3 includes steps S31 to S32:

[0100] S31, assign a random position and a random velocity to each particle in the second-order oscillating particle swarm, and update the velocity magnitude and velocity direction of each particle by comparing the fitness value, the particle's optimal value and the global optimal value of the objective function of the second-order oscillating particle swarm algorithm.

[0101] Understandably, each particle in the particle swarm optimization algorithm has its own position and velocity information. At the beginning of the algorithm, each particle is assigned a random position and velocity. The particle's velocity magnitude and direction are updated by comparing its own objective function fitness value with the particle's optimal value and the global optimal value, thus correcting the particle's motion process and making the particle gradually approach the global optimal value.

[0102] Furthermore, the second-order oscillating particle swarm optimization algorithm introduces a second-order oscillation element into the particle velocity update formula. The algorithm exhibits oscillatory convergence in the early stages, demonstrating good global search capability, and asymptotic convergence in the later stages, maintaining good global search capability. Therefore, it can find the global optimum faster and more accurately. The particle velocity update formula in this case is:

[0103] V i (t+1)=ω×V i (t)+c1r1(p i -x i (t))+c2r2(p g -x i (t));

[0104] In the formula, V i (t+1) represents the particle update rate, ω represents the inertia weight, c1 and c2 represent the first and second learning factors respectively, r1 and r2 represent two different random numbers uniformly distributed between 0 and 1, and V i (t) represents the velocity of the i-th particle, p i p represents the individual optimal value of a particle. g x represents the global optimum. i This represents the position of the i-th particle.

[0105] S32, the motion process of each particle is corrected according to the updated velocity magnitude and velocity direction of each particle, so that each particle gradually approaches the global optimum, and the initial load type of the start-stop event characteristics is identified based on the global optimum and the SVM model.

[0106] Understandably, since the SVM model is a binary classification model, compared to manually setting the parameters of the SVM model, the second-order oscillating particle swarm optimization algorithm can find more suitable parameters to achieve higher recognition accuracy and avoid getting trapped in local optima.

[0107] S4, set the movement strategy of the detection extraction window, move the detection extraction window based on the movement strategy and perform hierarchical channel identification on the preliminary load type to obtain the determined load type;

[0108] Specifically, step S4 includes steps S41 to S43:

[0109] S41, set the movement strategy of the detection extraction window according to the current trajectory map of the feature dataset, so that the detection extraction window covers the current trajectory map based on the movement strategy;

[0110] It is understandable that the feature dataset can be modeled as a power consumption data graph of various electrical devices. The power consumption data graph includes data such as current and voltage. The current data in the feature dataset is extracted and then modeled as a current trajectory graph. The movement strategy of the detection extraction window is set through the current trajectory graph of the feature dataset. In this embodiment, the movement strategy of the detection extraction window is that the movement path direction of the detection extraction window needs to be able to cover the current trajectory graph.

[0111] S42, based on the detection extraction window and the effective value of the current, filter the high-power loads in the start-stop event features;

[0112] Understandably, by using the detection extraction window and the effective value of the current to filter high-power loads in the start-stop event features, in specific implementation, the effective value of the current is used as a threshold, and then the current features in the start-stop event features are compared with the effective value of the current as the threshold through the detection extraction window, thereby filtering out high-power loads in the start-stop event features and completing the first step of identification.

[0113] S43, based on the steady-state fundamental frequency, harmonic amplitude, transient active power, transient reactive power and transient duration as load type characteristics, and determine the load type based on the load type characteristics;

[0114] It is understandable that steady-state fundamental frequency, harmonic amplitude, transient active power, transient reactive power, and transient duration are used as feature vectors and input into the SVM model for feature learning, thereby forming a load identification model. The load identification model identifies load type characteristics, thereby confirming the load type and completing the second step of identification.

[0115] S5, count the zero-crossing points in the feature dataset as the current data of the series arc, normalize the current data to obtain the processed current data, and perform VDM decomposition on the processed current data to obtain the center frequency of each modal component, and use the difference of the center frequency of each modal component as the arc feature.

[0116] It is understandable that due to the combustion characteristics of electric arcs, a temporary "zero-crossing" phenomenon occurs when the current crosses zero. Based on this characteristic of electric arcs, the number of zero-crossing points in the periodic current data can be used as a steady-state feature. The current data is first normalized, limiting the current amplitude to between -1 and 1 to facilitate subsequent fault feature extraction and analysis. VMD decomposition is performed on the line current data under normal load operation and with series fault arcs to obtain the corresponding IMF components, and their fuzzy entropy is calculated. The greater the disorder of the waveform of each IMF component, the greater its fuzzy entropy value. In this embodiment, the obtained data is decomposed into 5 different IMF components using the VMD algorithm. The fuzzy entropy value and center frequency of each load's IMF component are calculated separately, and combined with the time-domain characteristic current increment coefficient to form a feature vector, forming the feature vector library required for training the integrated deep RVFL neural network. It is worth noting that the center frequencies of different loads vary significantly, and this can be used as arc features to form feature vectors for arc identification, thereby obtaining arc features.

[0117] S6, extract the current increment coefficient and the fuzzy entropy value of each modal component from the arc features, form a feature vector based on the fuzzy entropy value and the center frequency, and input the feature vector into the integrated deep RVFL neural network so that the integrated deep RVFL neural network is trained based on the feature vector to obtain the trained integrated deep RVFL neural network;

[0118] Understandably, extracting the current increment coefficient and the fuzzy entropy values ​​and center frequencies of each modal component to form feature vectors, and inputting them into an integrated deep RVFL neural network for fault arc feature learning, can achieve better recognition results.

[0119] It is worth noting that when the current periodicity of electrical equipment is poor, the harmonic content of the current is high. Without an electric arc, its modal component complexity is high, and the fuzzy entropy value is large. With an electric arc, the current distortion is severe, the IMF1 waveform struggles to maintain normal fluctuations, the fuzzy entropy value is small, and the waveform uncertainty of other modal components increases, high-frequency harmonic components rise, waveform complexity continues to increase, and the fuzzy entropy value is even larger. The current waveform of an electric kettle is a sine wave with low harmonic content, but some noise interference still exists. Without an electric arc, the IMF fuzzy entropy value is large, while other IMF fuzziness is small. With an electric arc, the harmonic content increases, waveform complexity increases, and the fuzzy entropy value increases. This further demonstrates that the change in fuzzy entropy conforms to the change in waveform and can effectively reflect the waveform complexity index.

[0120] S7. Based on the trained integrated deep RVFL neural network, the arc features are identified to obtain electrical safety information, and the determined load type is detected by the continuous moving change method to obtain the current status type.

[0121] Understandably, the trained ensemble deep RVFL neural network can better identify fault conditions with arc characteristics. Furthermore, by detecting and determining the sudden leakage current at the moment the load type is determined using the continuous moving change method, it is necessary to first record the residual current of each power supply cycle and the residual current at a fixed time interval in the previous period, then compare and calculate the results. The detected residual current values ​​for each power supply cycle are recorded as I0, I1, I2, ..., I... m We can obtain the change in current at any time t, that is, the new fault leakage current generated in each cycle, expressed as:

[0122] I Δt =I t -I m ;

[0123] In the formula, I Δt I represents the new leakage current generated at time t. t I represents the residual current at time t. m This represents the residual current at time m.

[0124] In summary, the power safety monitoring method in the above embodiments of the present invention obtains start-stop event features by detecting and extracting feature datasets through a detection window, and obtains the preliminary load type of the power equipment by combining an SVM model with a second-order oscillating particle swarm optimization algorithm. This enables the determination of the load status of the power equipment, effectively reducing the possibility of inadequate monitoring of the power equipment. By using the difference in center frequency as an arc feature, it avoids situations where other features cannot reflect the power equipment when monitoring its power consumption. Furthermore, by performing hierarchical channel identification through the detection and extraction window, the load type is determined. The arc feature is identified by a trained integrated deep RVFL neural network, which effectively avoids misjudgment and missed judgment. It also avoids the problem of high randomness in series arcs leading to unmonitored conditions, thus avoiding the difficulty in monitoring the fault arc features of the line under different load conditions and at different times.

[0125] Example 2

[0126] Please see Figure 2 The image shows an electricity safety monitoring system according to a second embodiment of the present invention. The system includes:

[0127] The real-time acquisition module 10 is used to acquire electricity consumption data in real time, obtain load characteristics from the electricity consumption data, extract and analyze the load characteristics to obtain a feature dataset;

[0128] A detection module 20 is defined to define a sliding detection window in the power consumption time series and set a start-stop threshold. The sliding detection window includes a calculation window and a detection extraction window. The feature dataset is detected based on the calculation window, the detection extraction window and the start-stop threshold to obtain start-stop event features.

[0129] Combined with the identification module 30, it is used to identify the preliminary load type of the start-stop event characteristics based on the SVM model and the second-order oscillating particle swarm algorithm and multi-feature load.

[0130] The detection module 40 is configured to set the movement strategy of the detection extraction window, move the detection extraction window based on the movement strategy, and perform hierarchical channel identification on the preliminary load type to obtain the determined load type.

[0131] The statistical processing module 50 is used to statistically analyze the zero-crossing points in the feature dataset as current data of the series arc, normalize the current data to obtain processed current data, and perform VDM decomposition on the processed current data to obtain the center frequency of each modal component, and use the difference of the center frequency of each modal component as the arc feature.

[0132] The extraction training module 60 is used to extract the current increment coefficient and the fuzzy entropy value of each modal component in the arc feature, form a feature vector based on the fuzzy entropy value and the center frequency, and input the feature vector into the integrated deep RVFL neural network so that the integrated deep RVFL neural network is trained based on the feature vector to obtain the trained integrated deep RVFL neural network.

[0133] The identification and detection module 70 is used to identify the arc features based on the trained integrated deep RVFL neural network to obtain electrical safety information, and to detect the determined load type through the continuous moving change method to obtain the current status type.

[0134] In some alternative embodiments, the real-time acquisition module 10 includes:

[0135] The acquisition and conversion unit is used to acquire electrical signals based on the acquisition sensor and convert the electrical signals by a preset multiple to acquire current data;

[0136] The first acquisition unit is used to acquire the steady-state characteristics of the current based on the current amplitude, current harmonic components, steady-state active power and steady-state reactive power in the current data.

[0137] The formula for calculating the steady-state active power is:

[0138]

[0139] In the formula, P represents steady-state active power, k represents harmonic order, and V k I represents the effective value of the k-th voltage harmonic. k θ represents the effective value of the k-th current harmonic. k This represents the phase difference of the k-th harmonic;

[0140] The formula for calculating the steady-state reactive power is:

[0141]

[0142] In the formula, Q represents steady-state reactive power;

[0143] The second acquisition unit acquires the transient characteristics of the current based on the transient current, transient active power, transient reactive power, voltage noise, and duration in the current data.

[0144] In some alternative embodiments, the definition detection module 20 includes:

[0145] A definition setting unit is used to define a calculation window and a detection extraction window within an electricity consumption time series, and to set the window length of the calculation window and the detection extraction window;

[0146] A calculation unit is used to calculate the effective average current of the feature dataset based on the calculation window;

[0147] A judgment unit is set to set a start-stop threshold and compare the start-stop threshold with the effective average current. If the effective average current is greater than the start-stop threshold, it is determined that a transient event has occurred and the starting point of the power consumption event is recorded.

[0148] The calculation expression for the start / stop threshold is:

[0149]

[0150] In the formula, H min H max L1 and L2 represent the maximum and minimum values ​​of the start / stop threshold, respectively; L1 represents the number of sampling points where the fluctuation increases; L2 represents the number of sampling points where the current of the electrical equipment increases; L represents the length of the calculation window; and I represents the maximum and minimum values ​​of the start / stop threshold, respectively. min λ represents the minimum amplitude of change in the identified electrical equipment, f represents the average value of the current effective value fluctuation and the safety margin, and λ represents the noise level.

[0151] The detection and acquisition unit is used to continuously slide the detection and extraction window and detect current data of several cycles in the feature dataset based on the sliding detection and extraction window to detect voltage transient values, and to acquire start-stop event features in the feature dataset based on the starting point of the power consumption event and the voltage transient values.

[0152] In some alternative embodiments, the combined identification module 30 includes:

[0153] The allocation and update unit is used to assign a random position and a random velocity to each particle in the second-order oscillating particle swarm, and update the velocity magnitude and velocity direction of each particle by comparing the fitness value, the particle's optimal value and the global optimal value of the objective function of the second-order oscillating particle swarm algorithm.

[0154] The correction and identification unit is used to correct the motion process of each particle according to the updated velocity magnitude and velocity direction, so that each particle gradually approaches the global optimum, and identifies the preliminary load type of the start-stop event characteristics based on the global optimum and the SVM model.

[0155] In some alternative embodiments, the setting detection module 40 includes:

[0156] An extraction unit is configured to set a movement strategy for the detection extraction window based on the current trajectory map of the feature dataset, so that the detection extraction window covers the current trajectory map based on the movement strategy.

[0157] The filtering unit is used to filter high-power loads in the start-stop event characteristics based on the detection extraction window and the effective value of the current.

[0158] The determination unit is used to determine the load type based on the steady-state fundamental frequency, harmonic amplitude, transient active power, transient reactive power and transient duration as load type characteristics.

[0159] The functions or operation steps implemented by the above modules and units are largely the same as those in the above method embodiments, and will not be repeated here.

[0160] The power safety monitoring system provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0161] Example 3

[0162] Please see Figure 3 The figure shown is a schematic diagram of the hardware structure of the electronic device in the third embodiment of the present invention.

[0163] The electronic device may include a processor 81 and a memory 82 storing computer program instructions.

[0164] Specifically, the processor 81 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement this application.

[0165] The memory 82 may include a mass storage device for data or instructions. For example, and not limitingly, the memory 82 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 82 may include removable or non-removable (or fixed) media. Where appropriate, the memory 82 may be internal or external to a data processing device. In a particular embodiment, the memory 82 is non-volatile memory. In a particular embodiment, the memory 82 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), Extended Data Out Dynamic Random-Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0166] The memory 82 can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program instructions executed by the processor 81.

[0167] The processor 81 reads and executes the computer program instructions stored in the memory 82 to implement the power safety monitoring method of the above embodiment 1.

[0168] In some embodiments, the electronic device may further include a communication interface 83 and a bus 80. For example, Figure 3 As shown, the processor 81, memory 82, and communication interface 83 are connected through bus 80 and complete communication with each other.

[0169] The communication interface 83 is used to enable communication between the various modules, devices, units, and / or equipment in this application. The communication interface 83 can also enable data communication with other components such as external devices, image / data acquisition devices, databases, external storage, and image / data processing workstations.

[0170] Bus 80 includes hardware, software, or both, that couples components of a device together. Bus 80 includes, but is not limited to, at least one of the following: data bus, address bus, control bus, expansion bus, and local bus. For example, and not as a limitation, bus 80 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 80 may include one or more buses. Although this application describes and illustrates a specific bus, this application considers any suitable bus or interconnection.

[0171] The electronic device can access the power safety monitoring system and execute the power safety monitoring method of this embodiment.

[0172] In addition, in conjunction with the electricity safety monitoring method in Embodiment 1 above, this application can provide a storage medium for implementation. This storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement the electricity safety monitoring method of Embodiment 1 above.

[0173] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0174] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for monitoring electrical safety, characterized in that, The method includes: Real-time collection of electricity consumption data, acquisition of load characteristics from the electricity consumption data, extraction and analysis of the load characteristics to obtain a feature dataset; A sliding detection window is defined in the electricity consumption time series, and a start-stop threshold is set. The sliding detection window includes a calculation window and a detection extraction window. The feature dataset is detected based on the calculation window, the detection extraction window, and the start-stop threshold to obtain start-stop event features. An SVM model is used in conjunction with a second-order oscillating particle swarm optimization algorithm, and the initial load type is identified based on the characteristics of the start-stop events using multi-feature loads. A movement strategy for the detection and extraction window is set, and the detection and extraction window is moved based on the movement strategy to perform hierarchical channel identification on the preliminary load type in order to obtain a determined load type; The zero-crossing points in the feature dataset are used as the current data of the series arc. The current data is normalized to obtain the processed current data. The processed current data is then decomposed using VDM to obtain the center frequency of each modal component. The difference in the center frequency of each modal component is used as the arc feature. Extract the current increment coefficient and the fuzzy entropy value of each modal component from the arc features, form a feature vector based on the fuzzy entropy value and the center frequency, and input the feature vector into an integrated deep RVFL neural network so that the integrated deep RVFL neural network is trained based on the feature vector to obtain the trained integrated deep RVFL neural network. The electric arc features are identified based on the trained integrated deep RVFL neural network to obtain electrical safety information, and the determined load type is detected by the continuous moving change method to obtain the current status type.

2. The method for monitoring electrical safety according to claim 1, characterized in that, The steps of real-time acquisition of electricity consumption data, obtaining load characteristics from the electricity consumption data, extracting and analyzing the load characteristics to obtain a feature dataset include: The system collects electrical signals based on sensors and converts the electrical signals by a preset factor to collect current data. The steady-state characteristics of the current are obtained based on the current amplitude, current harmonic components, steady-state active power, and steady-state reactive power in the current data. The transient characteristics of the current are obtained based on the transient current, transient active power, transient reactive power, voltage noise, and duration in the current data.

3. The method for monitoring electrical safety according to claim 2, characterized in that, The formula for calculating the steady-state active power is: In the formula, P represents steady-state active power, k represents the harmonic order, and V represents the harmonic power. k I represents the effective value of the k-th voltage harmonic. k θ represents the effective value of the k-th current harmonic. k This represents the phase difference of the k-th harmonic; The formula for calculating the steady-state reactive power is: In the formula, Q represents steady-state reactive power.

4. The method for monitoring electrical safety according to claim 1, characterized in that, The step of defining a sliding detection window in the electricity consumption time series and setting a start / stop threshold, wherein the sliding detection window includes a calculation window and a detection extraction window, and detecting the feature dataset based on the calculation window, the detection extraction window, and the start / stop threshold to obtain start / stop event features includes: Define a calculation window and a detection extraction window within the electricity consumption time series, and set the window length of the calculation window and the detection extraction window; The effective average current of the feature dataset is calculated based on the calculation window; Set a start / stop threshold and compare the start / stop threshold with the effective average current. If the effective average current is greater than the start / stop threshold, it is determined that a transient event has occurred, and the starting point of the power consumption event is recorded. The detection extraction window is continuously slid, and current data for several cycles in the feature dataset is detected based on the slid detection extraction window to detect voltage transient values. Start-stop event features in the feature dataset are obtained based on the starting point of the power consumption event and the voltage transient values.

5. The method for monitoring electrical safety according to claim 4, characterized in that, The calculation expression for the start / stop threshold is: In the formula, H min H max L1 and L2 represent the maximum and minimum values ​​of the start / stop threshold, respectively; L1 represents the number of sampling points where the fluctuation increases; L2 represents the number of sampling points where the current of the electrical equipment increases; L represents the length of the calculation window; and I represents the maximum and minimum values ​​of the start / stop threshold, respectively. min λ represents the minimum amplitude of change in the identified electrical equipment, f represents the average value of the current effective value fluctuation and the safety margin, and λ represents the noise level.

6. The method for monitoring electrical safety according to claim 1, characterized in that, The steps of using an SVM model combined with a second-order oscillating particle swarm optimization algorithm and identifying the initial load type based on multi-feature load characteristics include: Each particle in the second-order oscillating particle swarm is assigned a random position and a random velocity, and the velocity magnitude and velocity direction of each particle are updated by comparing the fitness value, the particle's optimal value, and the global optimal value of the objective function of the second-order oscillating particle swarm algorithm. The motion of each particle is corrected based on the updated velocity magnitude and velocity direction, so that each particle gradually approaches the global optimum. Based on the global optimum and the SVM model, the initial load type of the start-stop event characteristics is identified.

7. The method for monitoring electrical safety according to claim 1, characterized in that, The steps of setting a movement strategy for the detection extraction window, moving the detection extraction window based on the movement strategy, and performing hierarchical channel identification on the preliminary load type to determine the load type include: The movement strategy of the detection extraction window is set according to the current trajectory map of the feature dataset, so that the detection extraction window covers the current trajectory map based on the movement strategy; Based on the detection extraction window and the effective value of the current, high-power loads in the start-stop event characteristics are screened. The load type is determined based on the steady-state fundamental frequency, harmonic amplitude, transient active power, transient reactive power, and transient duration.

8. An electrical safety monitoring system, characterized in that, The system includes: The real-time acquisition module is used to acquire electricity consumption data in real time, obtain load characteristics from the electricity consumption data, extract and analyze the load characteristics to obtain a feature dataset; A detection module is defined to define a sliding detection window in the electricity consumption time series and set a start-stop threshold. The sliding detection window includes a calculation window and a detection extraction window. The feature dataset is detected based on the calculation window, the detection extraction window and the start-stop threshold to obtain start-stop event features. Combined with the identification module, it is used to identify the preliminary load type of the start-stop event characteristics based on the SVM model and the second-order oscillating particle swarm algorithm and multi-feature load. A detection module is configured to set a movement strategy for the detection extraction window, move the detection extraction window based on the movement strategy, and perform hierarchical channel identification on the preliminary load type to obtain a determined load type. The statistical processing module is used to statistically analyze the zero-crossing points in the feature dataset as current data of the series arc, normalize the current data to obtain processed current data, and perform VDM decomposition on the processed current data to obtain the center frequency of each modal component, and use the difference of the center frequency of each modal component as the arc feature. An extraction training module is used to extract the current increment coefficient and the fuzzy entropy value of each modal component in the arc features, form a feature vector based on the fuzzy entropy value and the center frequency, and input the feature vector into an integrated deep RVFL neural network so that the integrated deep RVFL neural network is trained based on the feature vector to obtain a trained integrated deep RVFL neural network. The identification and detection module is used to identify the arc features based on the trained integrated deep RVFL neural network to obtain electrical safety information, and to detect the determined load type through the continuous moving change method to obtain the current status type.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the power safety monitoring method as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the power safety monitoring method as described in any one of claims 1 to 7.

Images