Fault determination method and device of power system, equipment, storage medium and program product

Abstract

The invention relates to a fault determination method and device of a power system, equipment, a storage medium and a program product. The method comprises the steps of obtaining energy consumption data, transmitted by a target network, of a power system, extracting target features in the energy consumption data, determining an initial load sudden drop monitoring result based on the target features, determining a target load sudden drop monitoring result according to the initial load sudden drop monitoring result, and determining an open-phase fault monitoring result according to the target load sudden drop monitoring result. The target terminal with the sampling frequency greater than the preset frequency is used for collecting the energy consumption data, and the target network is used for transmitting the energy consumption data, so that the high-density sampling data is obtained, and fault diagnosis is carried out according to the high-density sampling data, and therefore, the phase failure diagnosis efficiency is improved, and the timeliness of phase failure diagnosis is improved.

Description

Technical Field

[0001] This application relates to the field of robot control technology, and in particular to a method, apparatus, equipment, storage medium, and program product for determining faults in a power system. Background Technology

[0002] With the rapid development of society and the economy, the power system, as the core infrastructure for energy supply, is constantly expanding in scale and becoming increasingly complex in structure, leading to ever-increasing demands on power supply stability and reliability. Intelligent management has become a core trend in power system development, and monitoring and fault diagnosis technologies, as key supports for intelligent management, directly determine the power system's ability to cope with sudden faults and ensure continuous power supply.

[0003] However, current technologies in the field of power system monitoring and fault diagnosis suffer from low timeliness in monitoring phase loss faults. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, equipment, storage medium, and program product for determining faults in power systems that can improve the timeliness of phase loss fault monitoring, in response to the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for determining faults in a power system, including:

[0006] Acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by target terminals with a sampling frequency greater than a preset frequency. Target terminals include smart meters and sensor networks, and target networks include 5G networks and fiber optic networks.

[0007] Extract target features from energy consumption data and determine the initial load drop monitoring results based on the target features. The target features include time-domain features, frequency-domain features, and time-frequency features related to load drop. The time-domain features include the abrupt gradient of current and voltage over time and the permutation entropy. The frequency-domain features include the harmonic distortion rate and the spectral centroid shift. The time-frequency features include the energy entropy.

[0008] The target load drop monitoring results are determined based on the initial load drop monitoring results.

[0009] The phase loss fault monitoring results are determined based on the target load drop monitoring results.

[0010] In one exemplary embodiment, determining the initial load descent monitoring results based on target characteristics includes:

[0011] The target features are input into the target classifier function in the load drop identification model to obtain the initial load drop identification result;

[0012] When the initial load drop identification results are used to characterize the target features as abnormal features, the target features are input into the target probability model to obtain the initial load drop monitoring results; the initial load drop monitoring results include the category of load drop.

[0013] In an exemplary embodiment, determining the target load sag monitoring result based on the initial load sag monitoring result includes:

[0014] If the load drop category in the initial load drop monitoring results is motor stoppage, obtain the current current change gradient and the current spectrum centroid offset; if the current current change gradient is higher than the preset change gradient and the current spectrum centroid offset is lower than the preset spectrum centroid offset, determine that the target load drop monitoring results are classified as motor stoppage.

[0015] If the load drop category in the initial load drop monitoring results is short circuit fault, obtain the current harmonic distortion rate and the current energy entropy. If the current harmonic distortion rate is greater than the preset harmonic distortion rate and the current energy entropy is less than the preset energy entropy, determine that the target load drop monitoring results are classified as short circuit fault.

[0016] If the load drop category in the initial load drop monitoring results is "load cut-off according to plan", the three-phase voltage of the power grid is obtained. If the voltage difference between any two phases of the three-phase voltage is less than the preset voltage difference, the target load drop monitoring results are determined to be "load drop category is load cut-off according to plan".

[0017] In one exemplary embodiment, the method further includes:

[0018] If at least one phase voltage in the three-phase voltage is lower than the preset voltage, and the duration of the lower voltage phase is longer than the preset duration, the target load drop monitoring result is determined to be a load drop category of motor stoppage or short circuit fault.

[0019] In an exemplary embodiment, determining the phase loss fault monitoring result based on the target load drop monitoring result includes:

[0020] If the load drop category in the target load drop monitoring results is motor stoppage, the vibration frequency of the power equipment is obtained by installing vibration sensors on the power equipment including the stopped motor; if the vibration frequency is an abnormal vibration frequency, it is determined that a power outage event has occurred at the location of the power equipment, and the phase loss fault monitoring results are determined based on the location of the power equipment.

[0021] When the load drop category in the target load drop monitoring results is a short-circuit fault, the reference amplitude-frequency curve of the signal at the transmitting end of the line where the short-circuit fault occurred is compared with the actual amplitude-frequency curve at the receiving end of the line to obtain the amplitude attenuation. The reference phase difference frequency curve of the signal at the transmitting end is compared with the actual phase difference frequency curve at the receiving end to obtain the phase change. The phase loss fault monitoring results are determined based on the amplitude attenuation and / or phase change.

[0022] In one exemplary embodiment, the method further includes:

[0023] The level of phase loss fault is determined based on the monitoring results of phase loss fault and the monitoring results of sudden drop in target load;

[0024] The system provides a visual display of the target load drop monitoring results, phase loss fault monitoring results, and phase loss fault levels.

[0025] Secondly, this application also provides a fault determination device for a power system, the device comprising:

[0026] The acquisition module is used to acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by the target terminal with a sampling frequency greater than the preset frequency, the target terminal includes smart meters and sensor networks, and the target network includes 5G network and fiber optic network;

[0027] The first determining module is used to extract target features from energy consumption data and determine the initial load drop monitoring results based on the target features. The target features include time-domain features, frequency-domain features, and time-frequency features related to load drop. The time-domain features include the abrupt gradient of current and voltage over time and the permutation entropy. The frequency-domain features include the harmonic distortion rate and the spectral centroid shift. The time-frequency features include the energy entropy.

[0028] The second determining module is used to determine the target load drop monitoring result based on the initial load drop monitoring result;

[0029] The third determination module is used to determine the phase loss fault monitoring results based on the target load drop monitoring results.

[0030] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0031] Acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by target terminals with a sampling frequency greater than a preset frequency. Target terminals include smart meters and sensor networks, and target networks include 5G networks and fiber optic networks.

[0032] Extract target features from energy consumption data and determine the initial load drop monitoring results based on the target features. The target features include time-domain features, frequency-domain features, and time-frequency features related to load drop. The time-domain features include the abrupt gradient of current and voltage over time and the permutation entropy. The frequency-domain features include the harmonic distortion rate and the spectral centroid shift. The time-frequency features include the energy entropy.

[0033] The target load drop monitoring results are determined based on the initial load drop monitoring results.

[0034] The phase loss fault monitoring results are determined based on the target load drop monitoring results.

[0035] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0036] Acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by target terminals with a sampling frequency greater than a preset frequency. Target terminals include smart meters and sensor networks, and target networks include 5G networks and fiber optic networks.

[0037] Extract target features from energy consumption data and determine the initial load drop monitoring results based on the target features. The target features include time-domain features, frequency-domain features, and time-frequency features related to load drop. The time-domain features include the abrupt gradient of current and voltage over time and the permutation entropy. The frequency-domain features include the harmonic distortion rate and the spectral centroid shift. The time-frequency features include the energy entropy.

[0038] The target load drop monitoring results are determined based on the initial load drop monitoring results.

[0039] The phase loss fault monitoring results are determined based on the target load drop monitoring results.

[0040] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0041] Acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by target terminals with a sampling frequency greater than a preset frequency. Target terminals include smart meters and sensor networks, and target networks include 5G networks and fiber optic networks.

[0042] Extract target features from energy consumption data and determine the initial load drop monitoring results based on the target features. The target features include time-domain features, frequency-domain features, and time-frequency features related to load drop. The time-domain features include the abrupt gradient of current and voltage over time and the permutation entropy. The frequency-domain features include the harmonic distortion rate and the spectral centroid shift. The time-frequency features include the energy entropy.

[0043] The target load drop monitoring results are determined based on the initial load drop monitoring results.

[0044] The phase loss fault monitoring results are determined based on the target load drop monitoring results.

[0045] The aforementioned power system fault determination method, apparatus, computer equipment, storage medium, and program product acquire power system energy consumption data transmitted from a target network, extract target features from the energy consumption data, determine initial load drop monitoring results based on the target features, determine target load drop monitoring results based on the initial load drop monitoring results, and determine phase loss fault monitoring results based on the target load drop monitoring results. Because energy consumption data is collected through a target terminal with a sampling frequency greater than a preset frequency and energy consumption data is transmitted through a target network, high-density sampling data is obtained. Fault diagnosis is then performed based on the high-density sampling data, thereby improving the efficiency and timeliness of phase loss fault diagnosis. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart illustrating a fault determination method for a power system provided in an embodiment of this application;

[0048] Figure 2 This is a flowchart illustrating a method for determining initial load drop monitoring results provided in an embodiment of this application;

[0049] Figure 3 This is a flowchart illustrating a result display method provided in an embodiment of this application;

[0050] Figure 4 This is a schematic diagram of the overall process of a fault determination method for a power system provided in an embodiment of this application;

[0051] Figure 5 This is a schematic diagram of the structure of a fault determination device for a power system provided in an embodiment of this application;

[0052] Figure 6 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0054] In one exemplary embodiment, such as Figure 1 As shown, Figure 1 This is a flowchart illustrating a fault determination method for a power system provided in an embodiment of this application. The method can be applied to a server and includes the following steps:

[0055] S101, acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by the target terminal with a sampling frequency greater than the preset frequency, the target terminal includes smart meters and sensor networks, and the target network includes 5G network and fiber optic network.

[0056] Deploying smart meters and sensor networks, with a sampling frequency of, for example, once per minute, allows for the collection of energy consumption data. This data can include parameters such as current, voltage, power, temperature, and vibration, thus achieving high-density data collection. Even higher sampling frequencies can be used, such as collecting data every half minute, to obtain high-density energy consumption data. This collected data can then be uploaded to a server in milliseconds via 5G wireless communication networks and fiber optic networks.

[0057] The high-density energy consumption data collection network architecture is deployed in layers, including: terminal layer (smart meters and sensor networks) → edge gateway (RS-485 / Modbus protocol) → communication layer (5G wireless communication network and fiber optic network) → cloud platform (server).

[0058] The role of the terminal layer (smart meters and sensor networks) is to collect real-time data. The implementation methods include: electrical parameter acquisition: deploying high-frequency sampling smart meters in the distribution area / user side to synchronously collect parameters such as three-phase current, voltage, and power (active / reactive). The accuracy of electricity metering is ±0.5% through pulse counting circuit.

[0059] The energy consumption data acquired in this step can be either directly collected initial energy consumption data or energy consumption data obtained after preprocessing the initial energy consumption data through an edge gateway. The role of the edge gateway is to quickly remove noise from the collected initial energy consumption data, compress the data after noise removal to obtain energy consumption data with transmittable physical quantity characteristics, and then transmit the transmittable energy consumption data with physical quantity characteristics to the server.

[0060] The communication layer achieves millisecond-level upload speeds through 5G / fiber dual-mode data transmission. Implementation methods include: 5G wireless transmission; URLLC (Ultra-Reliable Low-Latency Communication) slicing with an air interface latency of <1ms, supporting concurrent access of 2000 nodes / km²; data packets encapsulated using the Southern Network uplink communication protocol; and encrypted transmission via a dedicated communication network. The fiber optic backbone network uses OTN optical transmission (wavelength λ=1550nm) at core nodes, achieving a transmission rate of 100Gbps and an end-to-end latency of <2ms.

[0061] S102, extract target features from energy consumption data, and determine the initial load drop monitoring results based on the target features; the target features include time-domain features, frequency-domain features and time-frequency features related to load drop, the time-domain features include the abrupt change gradient of current over time, the abrupt change gradient of voltage over time and the permutation entropy, the frequency-domain features include the harmonic distortion rate and the spectral centroid shift, and the time-frequency features include the energy entropy.

[0062] Wherein, the abrupt change gradient: The abrupt change gradient is the instantaneous rate of change of current or voltage calculated by difference. Taking the abrupt change gradient of current as an example: The abrupt change gradient of current , For 1 minute, Let be the current at minute t. Let be the current at minute t-1.

[0063] Permutation entropy: Permutation entropy quantifies the local sequential pattern distribution of a time series. The specific calculation process includes: mapping a day's power sequence to a multi-dimensional phase space matrix and constructing a window vector; arranging the original values ​​within the window vector in ascending order to obtain the permutation result, recording the index position order of the original values ​​in the permutation result, generating a symbolic pattern sequence corresponding to the pattern based on the index position order; and counting the number of times the pattern occurs. Calculate the pattern probability Calculate the permutation entropy.

[0064] Where the power sequence is P={P1,P2,…,PN}, if the power is sampled once per minute for the transformer area during the day, then N is 1440. If each minute is a time point, then the formula for calculating the power per minute, which is also the formula for calculating the power at the t-th time point, is: That is, the power at the t-th time point is equal to the average of the product of voltage and current over one minute.

[0065] Mapping the one-dimensional power sequence P to a multi-dimensional phase space matrix, the reconstruction formula is:

[0066]

[0067] N is the sequence length, and m is the window length (embedding dimension). It's a time delay. Each line It is a vector of length m. For each window The values ​​within are sorted in ascending order to generate a symbolic pattern sequence: the sorting rule is to record the index position order of the original values ​​after sorting.

[0068] Pattern type (symbolic representation): If Xi=[1.2,0.9,0.5] → after sorting, it becomes [0.5,0.9,1.2] → original index order: position 3 → position 2 → position 1 → the pattern is represented as (3,2,1).

[0069] model The number of occurrences is the frequency of a specific order pattern in the index order of all windows.

[0070] This reflects the repetition rate of the pattern in the time series:

[0071] ;in, It is an indicator function, if equal ,but =1, otherwise 0.

[0072] Pattern probability: The sum of the probabilities of all patterns is 1.

[0073] Calculate permutation entropy Introduce correction terms Avoid division by zero:

[0074] ;in, It is the total number of patterns in m dimensions (the number of permutations of m elements).

[0075] The following describes the frequency domain characteristics associated with load drops: harmonic distortion rate (THD) and spectral centroid shift.

[0076] Harmonic distortion rate includes voltage harmonic distortion rate and current harmonic distortion rate. The voltage harmonic distortion rate (THD) is calculated as follows:

[0077] The harmonic distortion rate of the current is equal to . It is the effective value of the current of the h-th harmonic. It is the effective voltage value of the h-th harmonic. It is the effective value of the fundamental frequency (first harmonic) voltage. It is the effective value of the fundamental current. and Extracting the signal after decomposing its spectrum using Fast Fourier Transform.

[0078] Calculate the spectral centroid shift :

[0079] ; It is the frequency value (in Hertz) of the k-th frequency point, which is determined by the sampling rate and the number of FFT points. It is the complex amplitude value of the signal at the k-th frequency point after FFT, taking the absolute value. This indicates the energy at that frequency point.

[0080] The calculation process for energy entropy includes:

[0081] 1) Decompose the signal into sub-signals of different frequency bands:

[0082] The current or power signal is decomposed into sub-signals of multiple frequency bands, resulting in the coefficient sequence of each sub-band. ,in, Indicates the first A person with a belt Indicates the first The first sub-band coefficient Values.

[0083] 2) Calculate the first energy of sub-band : ,in, It is the first The length of the subband coefficient.

[0084] 3) Calculate the total energy, Etotal = the sum of the energies of each subband.

[0085] 4) Calculate the first subband energy probability The energy probability is equal to the first The quotient of the energy of the subband divided by the total energy.

[0086] 5) Calculate the energy entropy. The formula for energy entropy is H = , where L is the total number of subbands.

[0087] S103, determine the target load drop monitoring result based on the initial load drop monitoring result.

[0088] If the load drop category in the initial load drop monitoring results is motor stoppage, obtain the current current change gradient and the current spectrum centroid offset; if the current current change gradient is higher than the preset change gradient and the current spectrum centroid offset is lower than the preset spectrum centroid offset, determine that the target load drop monitoring results are classified as motor stoppage.

[0089] If the load drop category in the initial load drop monitoring results is short circuit fault, obtain the current harmonic distortion rate and the current energy entropy. If the current harmonic distortion rate is greater than the preset harmonic distortion rate and the current energy entropy is less than the preset energy entropy, determine that the target load drop monitoring results are classified as short circuit fault.

[0090] If the load drop category in the initial load drop monitoring results is "load cut-off according to plan", the three-phase voltage of the power grid is obtained. If the voltage difference between any two phases of the three-phase voltage is less than the preset voltage difference, the target load drop monitoring results are determined to be "load drop category is load cut-off according to plan".

[0091] In this step, determining the target load descent monitoring result based on the initial load descent monitoring result further verifies the accuracy of the initial load descent monitoring result. For example, if the load descent category in the initial load descent monitoring result is planned load shedding, the three-phase voltage of the power grid is obtained. If the voltage difference between any two phases is less than a preset voltage difference, the target load descent monitoring result is determined to be a planned load shedding. If at least one phase voltage is lower than a preset voltage, and the duration of the lower voltage phase is longer than a preset duration, the target load descent monitoring result is determined to be a motor stoppage or short-circuit fault. In other words, the load descent category in the initial load descent monitoring result is not planned load shedding, but a type of short-circuit fault during motor stoppage.

[0092] If at least one phase voltage in the three-phase voltage is lower than the preset voltage, and the duration of the phase voltage lower than the preset voltage is longer than the preset duration, the target load drop monitoring result can be further determined according to the method for judging motor stoppage or short circuit faults to determine which type of fault is motor stoppage or short circuit fault.

[0093] For example, if at least one phase voltage in the three-phase voltage is lower than the preset voltage, and the duration of the phase voltage lower than the preset voltage is longer than the preset duration, the current current change gradient and the current spectrum centroid offset are obtained; if the current current change gradient is higher than the preset change gradient, and the current spectrum centroid offset is lower than the preset spectrum centroid offset, the target load drop monitoring result is determined to be a load drop category of motor stoppage.

[0094] For example, if at least one phase voltage in the three-phase voltage is lower than the preset voltage, and the duration of the phase voltage lower than the preset voltage is longer than the preset duration, the current harmonic distortion rate and the current energy entropy are obtained. If the current harmonic distortion rate is greater than the preset harmonic distortion rate, and the current energy entropy is less than the preset energy entropy, the target load drop monitoring result is determined to be a short circuit fault.

[0095] S104, Determine the phase failure monitoring results, including the power outage area, based on the target load drop monitoring results.

[0096] If the load drop category in the target load drop monitoring results is motor stoppage, the vibration frequency of the power equipment is obtained by installing vibration sensors on the power equipment including the stopped motor; if the vibration frequency is an abnormal vibration frequency, it is determined that a power outage event has occurred at the location of the power equipment, and the phase failure monitoring results are determined based on the location of the power equipment; the phase failure monitoring results may include the power outage area.

[0097] When the load drop category in the target load drop monitoring results is a short-circuit fault, the amplitude attenuation is obtained by comparing the reference amplitude-frequency curve of the signal at the transmitting end of the line where the short-circuit fault occurred with the actual amplitude-frequency curve at the receiving end of the line. Similarly, the phase change is obtained by comparing the reference phase difference frequency curve of the signal at the transmitting end with the actual phase difference frequency curve at the receiving end. The phase loss fault monitoring results are then determined based on the amplitude attenuation and / or phase change. The phase loss fault monitoring results may include the power outage area.

[0098] The method provided in this embodiment acquires energy consumption data of the power system transmitted by the target network, extracts target features from the energy consumption data, determines the initial load drop monitoring result based on the target features, determines the target load drop monitoring result based on the initial load drop monitoring result, and determines the phase loss fault monitoring result based on the target load drop monitoring result. Since energy consumption data is collected by a target terminal with a sampling frequency greater than a preset frequency and the energy consumption data is transmitted by the target network, high-density sampling data is acquired. Fault diagnosis is then performed based on the high-density sampling data, thereby improving the efficiency and timeliness of phase loss fault diagnosis.

[0099] In one exemplary embodiment, such as Figure 2 As shown, Figure 2 This is a flowchart illustrating a method for determining initial load drop monitoring results provided in an embodiment of this application. The method includes the following steps:

[0100] S201, the target features are input into the target classifier function in the load drop identification model to obtain the initial load drop identification result.

[0101] The load descent identification model can be a Support Vector Machine (SVM). The initial load descent identification result is either normal operation or a load descent. If the initial load descent identification result is a load descent, then the initial load descent identification result is used to characterize the target feature as an abnormal feature. If the initial load descent identification result is normal operation, then the initial load descent identification result is used to characterize the target feature as a normal feature.

[0102] The specific formula for the classifier function is as follows:

[0103] .

[0104] If f(x) is greater than zero, it is determined to be a sudden load drop; otherwise, it is normal operation. Here, K(xi,x) is the RBF kernel function: K(xi,x)=exp(-γ||xi-x||²), where γ is the kernel width parameter, controlling the range of influence of the samples; the larger γ is, the more complex the decision boundary. It is the squared Euclidean distance, representing a measure of similarity between feature vectors.

[0105] These are support vector weights. The value is greater than 0 and less than C, where C is the penalty coefficient, which controls the tolerance for misclassification.

[0106] ∈{-1,+1}, These are category labels (-1 = normal, +1 = sharp drop).

[0107] The load descent identification model can be obtained as follows: The time-domain, frequency-domain, and time-frequency features related to load descent from energy consumption data samples collected by target terminals with a sampling frequency greater than a preset frequency are input into the classifier function of the initial load descent identification model to obtain initial load descent identification result samples. Based on the load descent labels corresponding to the initial load descent identification result samples and the energy consumption data samples, the parameters of the classifier function are optimized. The load descent labels corresponding to the energy consumption data samples are used to characterize whether a load descent has occurred.

[0108] Among them, the penalty coefficient and kernel width parameter γ are parameters to be optimized. By adjusting the classifier parameters, the classifier function achieves the best overall performance in terms of false positive rate, recall rate, and detection latency, that is, to achieve fewer false positives, more detections, and faster response.

[0109] Multi-objective optimization:

[0110] minFPR (false positive rate); maxRecall (recall rate); minttdelay (detection delay).

[0111] FPR (False Alarm Rate) = FP / (FP + TN) * 100%, where FP (False Positive) is the number of times normal load was mistakenly identified as a sudden drop, and TN (True Negative) is the number of times normal load was correctly identified.

[0112] Recall = TP / (TP + FN) * 100%, where TP (TruePositive): the number of true dropouts that were correctly detected. FN (FalseNegative): the number of true dropouts that were not detected (false negatives).

[0113] tdelay (detection delay) = talert - to, where to represents the actual start time of the fault and talert represents the system alarm time.

[0114] The optimization process for the classifier function is as follows:

[0115] Initialize the population: Generate 100 sets of parameter combinations (e.g., γ∈[0.01,1], C∈[1,100]).

[0116] Non-dominated ordination: Calculate the three objective values ​​(FPR, Recall, tdelay) for each individual. Stratify by Pareto front (front 1 is the optimal solution set).

[0117] Parent selection: Prioritize the first individual from the frontier.

[0118] Crossover operator: generates a combination of child parameters (e.g., γchild=0.5(γp1+γp2)).

[0119] Elite Preservation: Merge parent and offspring generations, preserving one leading individual for the next generation.

[0120] Constraints: tdelay < 1s (industrial real-time requirements), FPR < 5% (to reduce misoperation), and Recall > 95%.

[0121] S202, when the initial load drop identification result is used to characterize the target feature as an abnormal feature, the target feature is input into the target probability model to obtain the initial load drop monitoring result; the initial load drop monitoring result includes the category of load drop.

[0122] The target probability model can be a Gaussian Mixture Model (GMM). GMM models can assist clustering in probability calibration. The corresponding target probability model is:

[0123] .

[0124] Mixed weights: are the prior probabilities of cluster k, ∑ =l.

[0125] Ɲ(·) represents a Gaussian distribution: ∑k is the mean vector, and ∑k is the covariance matrix.

[0126] K represents the number of clusters. In this embodiment, the number of clusters K=3, corresponding to three categories of sudden load drops: motor stoppage, short circuit fault, and planned load shedding. That is, the categories of sudden load drops include motor stoppage, short circuit fault, and planned load shedding.

[0127] The output f(x) of the Support Vector Machine (SVM) is used for initial binary classification, i.e., the initial load drop is identified as normal / abnormal. Then, based on the abnormal sample x (e.g., the target feature is an abnormal feature, i.e., the target feature is an abnormal sample), the probability of it belonging to each GMM cluster is calculated using the target probability model:

[0128] For example, p1>0.8 indicates motor stoppage (characteristics: high ΔI, low fc); p2>0.7 indicates short circuit fault (characteristics: THD>15%, extremely high HE). The droop type can be determined based on the maximum pk, i.e., the droop category is determined to be motor stoppage.

[0129] In this embodiment, by inputting the target features into the target classifier function of the load descent identification model, an initial load descent identification result is obtained. If the initial load descent identification result is used to characterize the target features as anomalous, the target features are then input into the target probability model to obtain an initial load descent monitoring result that includes the category of load descent. By utilizing both the target classifier function and the target probability model, the accuracy of the obtained initial load descent monitoring result is improved.

[0130] In an exemplary embodiment, determining the target load descent monitoring result based on the initial load descent monitoring result can be achieved in the following way:

[0131] If the initial load descent monitoring results indicate that the load descent category is motor stoppage, the current current surge gradient and the current spectral centroid offset are obtained. If the current current surge gradient is higher than the preset surge gradient, and the current spectral centroid offset is lower than the preset spectral centroid offset, the target load descent monitoring result is determined to be classified as motor stoppage. By further obtaining the current current surge gradient and the current spectral centroid offset when the initial load descent monitoring results indicate motor stoppage, the accuracy of the obtained load descent category is further verified.

[0132] If the initial load sag monitoring results indicate a short-circuit fault, the current harmonic distortion rate and energy entropy are obtained. If the current harmonic distortion rate is greater than a preset harmonic distortion rate and the current energy entropy is less than a preset energy entropy, the target load sag monitoring result is confirmed to indicate a short-circuit fault. By further obtaining the current harmonic distortion rate and energy entropy when the initial load sag monitoring results indicate a short-circuit fault, the accuracy of the obtained load sag classification is improved.

[0133] If the initial load sag monitoring results indicate that the load sag is classified as planned load shedding, the three-phase voltages of the power grid are acquired. If the voltage difference between any two phases is less than a preset voltage difference, the target load sag monitoring result is determined to be classified as planned load shedding. By further acquiring the three-phase voltages of the power grid when the initial load sag monitoring results indicate planned load shedding, the accuracy of the obtained load sag classification is further verified.

[0134] It should be noted that when load is cut off according to plan, such as when some industrial loads need to be cut off according to plan, this is considered a normal load drop, and the phase failure monitoring results can be determined to be that there is no phase failure.

[0135] In this embodiment, by using a method that matches the category of load drop in the initial load drop monitoring results, the accuracy of the load drop category in the initial load drop monitoring results is further verified, thereby improving the accuracy of the final target load drop monitoring results.

[0136] In one exemplary embodiment, the method further includes:

[0137] If at least one phase voltage in the three-phase voltage is lower than the preset voltage, and the duration of the lower voltage phase is longer than the preset duration, the target load drop monitoring result is determined to be a load drop category of motor stoppage or short circuit fault.

[0138] If the load drop category in the initial load drop monitoring results is "load shedding according to plan", the three-phase voltage of the power grid is obtained; if at least one phase voltage is lower than the preset voltage and the duration of the phase voltage lower than the preset voltage is longer than the preset duration, the current current change gradient and the current spectrum centroid offset are obtained; if the current current change gradient is higher than the preset change gradient and the current spectrum centroid offset is lower than the preset spectrum centroid offset, the target load drop monitoring result is determined to be a load drop category of motor stoppage.

[0139] For example, if the load drop category in the initial load drop monitoring results is load shedding according to plan, the three-phase voltage of the power grid is obtained; if at least one phase voltage is lower than the preset voltage and the duration of the phase voltage lower than the preset voltage is longer than the preset duration, the current harmonic distortion rate and the current energy entropy are obtained; if the current harmonic distortion rate is greater than the preset harmonic distortion rate and the current energy entropy is less than the preset energy entropy, the target load drop monitoring result is determined to be a short-circuit fault.

[0140] In this embodiment, the three-phase voltage of the power grid is obtained when the load drop category in the initial load drop monitoring results is "load shedding according to plan." If at least one phase voltage is lower than a preset voltage, and the duration of the lower voltage phase is longer than a preset duration, the target load drop monitoring result is determined to be a load drop category of motor stoppage or short-circuit fault. This allows for correction of the load drop category in the initial load drop monitoring results, improving the accuracy of the obtained target load drop monitoring results.

[0141] In an exemplary embodiment, the above-described S104, which determines the phase-loss fault monitoring results including the power outage area based on the target load drop monitoring results, can be achieved in the following way:

[0142] When the load drop category in the target load drop monitoring results is motor stoppage, the vibration frequency of the power equipment, including the stopped motor, is obtained by installing vibration sensors on the power equipment. If the vibration frequency is abnormal, it is determined that a power outage event has occurred at the location of the power equipment, and the phase loss fault monitoring results are determined based on the location of the power equipment. For example, if the vibration frequency of a vibration sensor installed on a power equipment changes abnormally, the physical location of the equipment can be matched in a Geographic Information System (GIS) map based on the equipment identifier, and the phase loss fault monitoring results, including the power outage area, can be determined based on the physical location.

[0143] When the load drop category in the target load drop monitoring results is a short-circuit fault, the amplitude attenuation is obtained by comparing the reference amplitude-frequency curve of the signal at the transmitting end of the line where the short-circuit fault occurred with the actual amplitude-frequency curve at the receiving end of the line. Similarly, the phase change is obtained by comparing the reference phase difference frequency curve of the signal at the transmitting end with the actual phase difference frequency curve at the receiving end. The phase loss fault monitoring result is determined based on the amplitude attenuation and / or phase change. For example, if a short-circuit fault occurs on a line, the amplitude attenuation and / or phase change of that line can be obtained. Based on the amplitude attenuation and / or phase change, it is determined that a phase loss fault exists on that line. For instance, if the signal attenuation of the line exceeds 20 dB or is completely lost, a phase loss fault is determined. Then, based on the line identifier, the physical location of the line identifier is matched in the GIS map. Based on this physical location, the phase loss fault monitoring result, including the power outage area, is determined.

[0144] In one exemplary embodiment, such as Figure 3 As shown, Figure 3 This is a flowchart illustrating a result display method provided in an embodiment of this application. The method includes the following steps:

[0145] S301. Determine the phase loss fault level based on the phase loss fault monitoring results and the target load drop monitoring results.

[0146] Specifically, classifying power outage events according to the category of load drop in the target load drop monitoring results and the severity represented by the category is the core mechanism of power system emergency response.

[0147] For example, a localized fault-spreading load drop (such as a main line phase failure triggering a power grid cascading effect) may involve cross-regional power outages and can be classified as a major phase failure-level power outage. In contrast, a single-point load drop at the equipment level (such as a single transformer breakdown) may involve a power outage affecting a single region and can be classified as a minor phase failure-level power outage.

[0148] S302 provides a visual display of the target load drop monitoring results, phase loss fault monitoring results, and phase loss fault levels.

[0149] Specifically, the monitoring results of sudden drop in target load, the monitoring results of phase loss fault, and the level of phase loss fault can be uploaded to a 3D visualization platform for visualization.

[0150] Digital twin mapping: Locate the power outage area using a GIS map (accuracy of 0.01 latitude and longitude) and render the equipment temperature field and current distribution in real time (UE engine).

[0151] Fault alarm classification:

[0152] Level 1 alarm (e.g., short circuit fault): audible and visual alarm and SMS push notification (response time <10s).

[0153] Level 2 alarm (e.g., overload): Platform pop-up notification.

[0154] Smart work order:

[0155] Automatically generate emergency repair routes (based on Amap API).

[0156] Push faulty equipment maintenance records (such as the number of circuit breaker operations and insulation aging history).

[0157] Implementation example: A sudden power outage in an industrial park.

[0158] Scenario: A lightning strike caused a short circuit in the 10kV distribution cabinet, and the three-phase voltage dropped to 30% of the rated value (phase A was lost).

[0159] Processing flow:

[0160] t=0s: The voltage sensor detects a sudden drop of 65% in the voltage of phase A, ΔI=180A / s, and the phase failure flag is activated.

[0161] t=200ms: The edge terminal confirms the anomaly, and the compressed data is transmitted to the cloud platform via 5G.

[0162] t=420ms: The AI ​​model determined that it was a short circuit power outage (confidence level 98%), and located the fault in distribution cabinet No. 3.

[0163] t=10s: An alarm pops up on the monitoring center's large screen, and a work order is dispatched to the nearest emergency repair team; the system automatically switches to the backup line to restore power.

[0164] In one exemplary embodiment, temperature gradient monitoring can also be performed: an infrared thermal imager or embedded thermocouple is used to collect the temperature of the equipment casing or internal coils in real time. A phase failure causes a surge in local current in the coil, forming a hot spot area (such as a sudden temperature rise of 15-20°C), and the faulty phase is located by comparing the temperature difference of the three-phase windings.

[0165] After a phase failure occurs, the remaining two phase windings will bear the load current of the original three phases, causing a surge in current. This surge in current will significantly increase the copper loss of the windings (which is proportional to the square of the current), leading to a rapid rise in the temperature of the failed phase. The current of the normal phase does not increase abnormally, and the temperature remains stable or only rises slightly. Ultimately, this results in a temperature difference characteristic where the temperature of the faulty phase is significantly higher than that of the other two phases. Therefore, the faulty phase can be located by comparing the temperature difference of the three-phase windings.

[0166] In this embodiment, the phase loss fault level is determined based on the phase loss fault monitoring results and the target load drop monitoring results. The target load drop monitoring results, phase loss fault monitoring results, and phase loss fault level are visualized and displayed. This allows for quick identification of the fault level and power outage area based on the displayed target load drop monitoring results, phase loss fault monitoring results, and phase loss fault level, facilitating rapid fault elimination and reducing the escalation of the fault.

[0167] like Figure 4 As shown, Figure 4 This is a schematic diagram of the overall flow of a fault determination method for a power system provided in an embodiment of this application. The method includes the following steps:

[0168] S401, acquire energy consumption data of the power system transmitted by the target network.

[0169] S402, the target features of the energy consumption data are input into the target classifier function in the load drop identification model to obtain the initial load drop identification result.

[0170] S403, when the initial load drop identification result is used to characterize the target feature as an abnormal feature, the target feature is input into the target probability model to obtain the initial load drop monitoring result.

[0171] S404, Determine the target load drop monitoring results based on the initial load drop monitoring results.

[0172] S405, determine the phase loss fault monitoring results based on the target load drop monitoring results.

[0173] S406. Determine the phase loss fault level based on the phase loss fault monitoring results and the target load drop monitoring results.

[0174] S407 provides a visual display of the target load drop monitoring results, phase loss fault monitoring results, and phase loss fault levels.

[0175] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0176] Based on the same inventive concept, this application also provides a power system fault determination device for implementing the power system fault determination method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more power system fault determination device embodiments provided below can be found in the limitations of the power system fault determination method described above, and will not be repeated here.

[0177] In one exemplary embodiment, such as Figure 5 As shown, Figure 5 This is a schematic diagram of a fault determination device for a power system provided in an embodiment of this application. The fault determination device 500 includes:

[0178] The acquisition module 501 is used to acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by the target terminal with a sampling frequency greater than a preset frequency, the target terminal includes smart meters and sensor networks, and the target network includes 5G networks and fiber optic networks;

[0179] The first determining module 502 is used to extract target features from energy consumption data and determine the initial load drop monitoring results based on the target features. The target features include time-domain features, frequency-domain features, and time-frequency features related to the load drop. The time-domain features include the abrupt gradient of current and voltage over time and the permutation entropy. The frequency-domain features include the harmonic distortion rate and the spectral centroid shift. The time-frequency features include the energy entropy.

[0180] The second determining module 503 is used to determine the target load drop monitoring result based on the initial load drop monitoring result;

[0181] The third determining module 504 is used to determine the phase loss fault monitoring results based on the target load drop monitoring results.

[0182] In an exemplary embodiment, the first determining module 502 is specifically used to input the target features into the target classifier function in the load drop identification model to obtain the initial load drop identification result; when the initial load drop identification result is used to characterize the target features as abnormal features, the target features are input into the target probability model to obtain the initial load drop monitoring result; the initial load drop monitoring result includes the category of load drop.

[0183] In an exemplary embodiment, the second determining module 503 is specifically configured to: 1) When the load drop category in the initial load drop monitoring result is motor stoppage, obtain the current current mutation gradient and the current spectral centroid offset; 2) When the current current mutation gradient is higher than a preset mutation gradient and the current spectral centroid offset is lower than a preset spectral centroid offset, determine that the target load drop monitoring result is a load drop category of motor stoppage; 3) When the load drop category in the initial load drop monitoring result is short-circuit fault, obtain the current harmonic distortion rate and the current energy entropy; 4) When the current harmonic distortion rate is greater than a preset harmonic distortion rate and the current energy entropy is less than a preset energy entropy, determine that the target load drop monitoring result is a load drop category of short-circuit fault; 5) When the load drop category in the initial load drop monitoring result is planned load shedding, obtain the three-phase voltage of the power grid; 6) When the voltage difference between any two phases of the three-phase voltage is less than a preset voltage difference, determine that the target load drop monitoring result is a load drop category of planned load shedding.

[0184] In an exemplary embodiment, the second determining module 503 is further configured to determine that the target load drop monitoring result is a load drop of motor stoppage or short circuit fault when at least one phase voltage in the three-phase voltage is lower than a preset voltage and the duration of the phase voltage lower than the preset voltage is longer than a preset duration.

[0185] In an exemplary embodiment, the third determining module 504 is specifically configured to: when the load drop category in the target load drop monitoring results is motor stoppage, acquire the vibration frequency of the power equipment by using a vibration sensor installed on the power equipment including the stopped motor; when the vibration frequency is an abnormal vibration frequency, determine that a power outage event has occurred at the location of the power equipment, and determine the phase loss fault monitoring result based on the location of the power equipment; when the load drop category in the target load drop monitoring results is short circuit fault, compare the reference amplitude frequency curve of the signal at the transmitting end of the line where the short circuit fault occurred with the actual amplitude frequency curve at the receiving end of the line to obtain the amplitude attenuation, and compare the reference phase difference frequency curve of the signal at the transmitting end with the actual phase difference frequency curve at the receiving end to obtain the phase change, and determine the phase loss fault monitoring result based on the amplitude attenuation and / or phase change.

[0186] In one exemplary embodiment, the fault determination device 500 further includes:

[0187] The fourth determination module is used to determine the level of phase loss fault based on the monitoring results of phase loss fault and the monitoring results of target load drop.

[0188] The display module is used to visually display the target load drop monitoring results, phase failure monitoring results, and phase failure level.

[0189] Each module in the aforementioned power system fault determination device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0190] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 6 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a fault determination method for a power system. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0191] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0192] In one exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method embodiments. The implementation principle and technical effects are similar to those of the above-described method embodiments, and will not be repeated here.

[0193] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the above method embodiment. Its implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0194] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the above method embodiments. Its implementation principle and technical effects are similar to those of the above method embodiments, and will not be repeated here.

[0195] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0196] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0197] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0198] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for determining faults in a power system, characterized in that, The method includes: Acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by the target terminal with a sampling frequency greater than a preset frequency, the target terminal includes smart meters and sensor networks, and the target network includes 5G networks and fiber optic networks; Extract target features from the energy consumption data, and determine the initial load drop monitoring results based on the target features; the target features include time-domain features, frequency-domain features, and time-frequency features related to load drop, the time-domain features include the abrupt gradient of current and voltage over time and the permutation entropy, the frequency-domain features include the harmonic distortion rate and the spectral centroid shift, and the time-frequency features include the energy entropy; The target load drop monitoring results are determined based on the initial load drop monitoring results. The phase loss fault monitoring results are determined based on the target load drop monitoring results.

2. The method according to claim 1, characterized in that, The determination of the initial load drop monitoring results based on the target characteristics includes: The target features are input into the target classifier function in the load drop identification model to obtain the initial load drop identification result; When the initial load drop identification result is used to characterize the target feature as an anomalous feature, the target feature is input into the target probability model to obtain the initial load drop monitoring result; the initial load drop monitoring result includes the category of load drop.

3. The method according to claim 2, characterized in that, The step of determining the target load drop monitoring result based on the initial load drop monitoring result includes: If the load drop category in the initial load drop monitoring result is motor stoppage, obtain the current current change gradient and the current spectrum centroid offset; if the current current change gradient is higher than the preset change gradient and the current spectrum centroid offset is lower than the preset spectrum centroid offset, determine that the target load drop monitoring result is a load drop category of motor stoppage. If the load drop category in the initial load drop monitoring result is a short circuit fault, the current harmonic distortion rate and the current energy entropy are obtained. If the current harmonic distortion rate is greater than the preset harmonic distortion rate and the current energy entropy is less than the preset energy entropy, the target load drop monitoring result is determined to be a short circuit fault. If the load drop category in the initial load drop monitoring result is load shedding according to plan, the three-phase voltage of the power grid is obtained. If the voltage difference between any two phases of the three-phase voltage is less than the preset voltage difference, the target load drop monitoring result is determined to be load drop category as load shedding according to plan.

4. The method according to claim 3, characterized in that, The method further includes: If at least one phase voltage in the three-phase voltage is lower than a preset voltage, and the duration of the phase voltage lower than the preset voltage is longer than a preset duration, the target load drop monitoring result is determined to be a load drop category of motor stoppage or short circuit fault.

5. The method according to any one of claims 1-4, characterized in that, The determination of the phase loss fault monitoring result based on the target load drop monitoring result includes: If the load drop category in the target load drop monitoring results is motor stoppage, the vibration frequency of the power equipment is obtained by a vibration sensor installed on the power equipment including the stopped motor; if the vibration frequency is an abnormal vibration frequency, it is determined that a power outage event has occurred at the location of the power equipment, and the phase loss fault monitoring result is determined based on the location of the power equipment. If the load drop category in the target load drop monitoring results is a short-circuit fault, the reference amplitude-frequency curve of the signal at the transmitting end of the line where the short-circuit fault occurred is compared with the actual amplitude-frequency curve at the receiving end of the line to obtain the amplitude attenuation. The reference phase difference frequency curve of the signal at the transmitting end is compared with the actual phase difference frequency curve at the receiving end to obtain the phase change. The phase loss fault monitoring result is determined based on the amplitude attenuation and the phase change.

6. The method according to any one of claims 1-4, characterized in that, The method further includes: The phase loss fault level is determined based on the phase loss fault monitoring results and the target load drop monitoring results; The monitoring results of the target load drop, the phase failure monitoring results, and the phase failure level are displayed visually.

7. A fault determination device for a power system, characterized in that, The device includes: The acquisition module is used to acquire energy consumption data of the power system transmitted by the target network; the energy consumption data is data collected by the target terminal with a sampling frequency greater than a preset frequency, the target terminal includes smart meters and sensor networks, and the target network includes 5G networks and fiber optic networks; The first determining module is used to extract target features from the energy consumption data and determine the initial load drop monitoring results based on the target features. The target features include time-domain features, frequency-domain features, and time-frequency features related to the load drop. The time-domain features include the abrupt gradients of current and voltage over time and the permutation entropy. The frequency-domain features include the harmonic distortion rate and the spectral centroid shift. The time-frequency features include the energy entropy. The second determining module is used to determine the target load drop monitoring result based on the initial load drop monitoring result; The third determining module is used to determine the phase loss fault monitoring result based on the target load drop monitoring result.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images