Method for monitoring electricity consumption of distribution box

Abstract

The application relates to the field of power monitoring, and discloses a power monitoring method for a distribution box, which is used to solve the problem of zero point deviation when power monitoring is performed. The method comprises the following steps: installing a power monitoring device at a line access point in the distribution box, obtaining data deviation information of the power monitoring device, evaluating an offset index according to the data deviation information, and performing zero point deviation judgment. If it is judged that the power monitoring device has zero point drift, initial power data is collected through the power monitoring device, the initial power data is corrected to zero points according to the offset index, corrected power data is obtained, a power anomaly index is calculated according to the corrected power data, power anomaly judgment is performed according to the power anomaly index, and power anomaly warning is performed if it is judged that the power is abnormal. The accuracy of power data collection is effectively improved, the accuracy of operation and maintenance strategies is improved, and the safety risk probability is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power monitoring, more particularly to a power monitoring method for a distribution box. BACKGROUND

[0002] With the development of smart grid and industrial automation, power safety and energy consumption management have become a key link in the management of industrial enterprises, power operation units and large public buildings. As an important node for power distribution and load protection in the power system, the real-time monitoring of the running state and power consumption behavior of the distribution box is of great significance for ensuring equipment safety, optimizing energy use and improving operation and maintenance efficiency.

[0003] The traditional power monitoring method for the distribution box mainly records data through power monitoring equipment, and identifies abnormalities according to the recorded data to determine whether there are overload, leakage, three-phase imbalance, voltage drop and other abnormal conditions.

[0004] However, in actual application, during the long-term operation of the monitoring equipment, it will be affected by external environment and its own factors, resulting in zero point drift, causing errors in the data collected by the monitoring equipment, greatly affecting the accuracy of subsequent power analysis, load identification and abnormality judgment, and even misleading operation and maintenance strategies, increasing power consumption costs or causing safety risks.

[0005] In view of the above problems, the present application provides a solution. SUMMARY

[0006] In order to overcome the above-mentioned defects of the prior art, the present application provides a power monitoring method for a distribution box to solve the problems existing in the background art.

[0007] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0008] The power monitoring method for the distribution box comprises the following steps: step 1, installing a power monitoring equipment at the internal line connection of the distribution box; step 2, obtaining data offset information of the power monitoring equipment, the data offset information including three-phase current data, power factor data, acceleration data, temperature data and humidity data, evaluating an offset index according to the data offset information, and determining zero point drift of the power monitoring equipment according to the offset index; step 3, if it is determined that the power monitoring equipment has zero point drift, collecting initial power data through the power monitoring equipment, correcting the initial power data according to the offset index to obtain corrected power data; step 4, calculating a power abnormality index according to the corrected power data, and determining power abnormality according to the power abnormality index; step 5, if it is determined that there is power abnormality, performing power abnormality warning; and step 6, storing power consumption data, analysis calculation results and warning events in the form of time series into a cloud database.

[0009] Preferably, the offset index obtaining step is: setting a monitoring time window, obtaining three-phase current data and power factor data in the time window, and calculating a power data offset coefficient according to the three-phase current data and the power factor data; setting an acceleration sensor in the power monitoring device, obtaining acceleration data through the acceleration sensor, obtaining the cumulative running time of the power monitoring device, and calculating a hardware running influence coefficient according to the acceleration data and the cumulative running time; setting a temperature sensor and a humidity sensor in the power monitoring device, obtaining temperature data and humidity data through the temperature sensor and the humidity sensor, and calculating a physical environment influence coefficient according to the temperature data and the humidity data; normalizing the power data offset coefficient, the hardware running influence coefficient and the physical environment influence coefficient, and calculating the offset index according to the normalized power data offset coefficient, the normalized hardware running influence coefficient and the normalized physical environment influence coefficient. The specific obtaining step is: OT = a1 × PO + a2 × OI + a3 × EI; in the formula, OT represents the offset index, PO represents the normalized power data offset coefficient, OI represents the normalized hardware running influence coefficient, EI represents the normalized physical environment influence coefficient, a1, a2 and a3 represent the weight coefficient of the power data offset coefficient, the weight coefficient of the hardware running influence coefficient and the weight coefficient of the physical environment influence coefficient.

[0010] Preferably, the power data offset coefficient obtaining step is: in the monitoring time window, obtaining real-time three-phase current data, and performing difference calculation on any two current data to obtain a current difference, performing mean value calculation on the current difference to obtain a three-phase difference average value; linearly fitting the three-phase difference average value according to time sequence to obtain a three-phase fitting slope, and the specific obtaining step is: in the formula, k1 represents the three-phase fitting slope, m is the number of three-phase current data samples, Δavg(t) is the three-phase difference average value at time point t, and t is the time point number; in the monitoring time window, obtaining real-time power factor data, calculating a power factor standard deviation in a sliding window, and performing root mean square calculation on the power factor standard deviations of all windows to obtain a power factor deviation degree factor; normalizing the three-phase fitting slope and the power factor deviation degree factor, and calculating the power data offset coefficient according to the normalized three-phase fitting slope and the normalized power factor deviation degree factor.

[0011] Preferably, the hardware operation influence coefficient obtaining step is: obtaining three-axis acceleration of the power monitoring device in continuous operation through an acceleration sensor, and calculating a resultant acceleration according to the three-axis acceleration; calculating the standard deviation of the resultant acceleration and the range of the resultant acceleration at each time point in a monitoring time window; calculating an operation vibration degree factor according to the standard deviation of the resultant acceleration and the range of the resultant acceleration, obtaining a total operation duration of the power monitoring device, and calculating an operation duration factor according to the total operation duration; and calculating the hardware operation influence coefficient according to the operation vibration degree factor and the operation duration factor.

[0012] Preferably, the physical environment influence coefficient obtaining step is: collecting temperature data and humidity data of the surface of the shell of the power monitoring device in the monitoring time window through a temperature sensor and a humidity sensor; calculating the standard deviation of the temperature in the monitoring time window, denoted as a temperature fluctuation degree; calculating a humidity exposure factor according to the humidity data using an exponential amplification function; calculating a synergistic fluctuation according to the temperature fluctuation degree and the humidity exposure factor; and calculating the physical environment influence coefficient according to the synergistic fluctuation.

[0013] Preferably, the zero-point offset determination step according to the offset index is: comparing the offset index with an offset threshold value, if the offset index is greater than or equal to the offset threshold value, determining that the power monitoring device has zero-point drift; and if the offset index is less than the offset threshold value, determining that the power monitoring device has no zero-point drift.

[0014] Preferably, the zero-point correction step according to the offset index is: obtaining the offset index and the offset threshold value, calculating the offset index and the offset threshold value to obtain an offset intensity ratio; constructing a correction factor according to the offset intensity ratio using a hyperbolic tangent function; obtaining a minimum value of the collected items in the monitoring time window, and calculating the corrected data according to the minimum value and the correction factor, the specific obtaining steps being: JZ(t) = Z(t) - CZ x Z min ; wherein, JZ(t) represents the corrected power data, Z(t) represents the initial power data, CZ represents the correction factor, and Z min represents the minimum value.

[0015] Preferably, the power anomaly index obtaining step is: constructing a power consumption data sample set according to the corrected power data in a set time window, performing standardization processing on all parameters in the set to obtain a standardized data set, using K-Means clustering method to divide all parameters in the standardized data set into a normal cluster and an abnormal cluster, obtaining a normal cluster center of the normal cluster, calculating the Euclidean distance between the power consumption data at each time point t and the normal cluster center as an abnormal deviation, obtaining a jump factor by calculating the multi-parameter change rate of the current time point and the previous time point, and adding the abnormal deviation of each time point and the jump factor to obtain the power anomaly index.

[0016] Preferably, the power anomaly determination step according to the power anomaly index is: comparing the power anomaly index with an anomaly threshold, if the power anomaly index is greater than or equal to the anomaly threshold, determining that the power is abnormal; and if the power anomaly index is less than the anomaly threshold, determining that the power is not abnormal.

[0017] The technical effects and advantages of the present application are:

[0018] The power monitoring device is installed at the line access inside the distribution box, the data offset information of the power monitoring device is obtained, the offset index is obtained according to the data offset information, the zero point offset is determined, if it is determined that the power monitoring device has zero point drift, the initial power data is collected through the power monitoring device, the initial power data is corrected according to the offset index, the corrected power data is obtained, the power anomaly index is calculated according to the corrected power data, the power anomaly is determined according to the power anomaly index, if the power anomaly is determined, the power anomaly warning is performed, the accuracy of power consumption data collection is effectively improved, the accuracy of operation and maintenance strategy is improved, and the safety risk probability is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 The overall flowchart of the present application. DETAILED DESCRIPTION

[0020] The technical solutions in the present application will be described clearly and completely in combination with the drawings in the present application, and in addition, the forms of each structure described in the following embodiments are only examples, and the power consumption monitoring method for the distribution box is not limited to each structure described in the following embodiments, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0021] The present application provides a power consumption monitoring method for a distribution box, as shown in the accompanying drawings, comprising the following steps: Figure 1

[0022] ​Step 1: install a power monitoring device at the line access inside the distribution box, the power monitoring device including voltage, current, and power sensors;

[0023] Step 2: obtain data offset information of the power monitoring device, the data offset information including three-phase current data, power factor data, acceleration data, temperature data, and humidity data, evaluate an offset index according to the data offset information, and determine zero-point offset of the power monitoring device according to the offset index;

[0024] By obtaining the data offset information of the power monitoring device, the offset index is evaluated, and whether the device has zero-point offset is determined accordingly, which has significant real-time and precision advantages. This method can dynamically identify baseline offset problems caused by sensor performance degradation, environmental changes, or installation offset under the condition of continuous operation of the device, avoiding distortion of power data caused by long-term cumulative error. Compared with the traditional manual verification or shutdown detection method, this method can realize intelligent monitoring and early warning of measurement stability and data reliability without interrupting the running process, thereby improving the accuracy, safety, and operation efficiency of the system.

[0025] In this embodiment, it needs to be specifically pointed out that the offset index obtaining step is:

[0026] Set a monitoring time window, obtain three-phase current data and power factor data within the time window, calculate a power data offset coefficient according to the three-phase current data and the power factor data, the three-phase current data refers to the current values flowing through the A-phase, B-phase, and C-phase circuits in the three-phase power system, which reflects the current distribution of the three-phase load and is an important basic parameter for judging system load balance, power consumption characteristics, and fault identification. Under normal operating conditions, the three-phase currents should be approximately equal, and if there is a significant difference, it may indicate that one phase is overloaded, poorly connected, or the sensor is drifting, etc.

[0027] An acceleration sensor is arranged in the power monitoring device, acceleration data is obtained through the acceleration sensor, the cumulative running time of the power monitoring device is obtained, and a hardware running influence coefficient is calculated according to the acceleration data and the cumulative running time;

[0028] A temperature sensor and a humidity sensor are arranged in the power monitoring device, temperature data and humidity data are obtained through the temperature sensor and the humidity sensor, and a physical environment influence coefficient is calculated according to the temperature data and the humidity data;

[0029] The power data offset coefficient, the hardware running influence coefficient, and the physical environment influence coefficient are normalized, the offset index is calculated according to the normalized power data offset coefficient, the hardware running influence coefficient, and the physical environment influence coefficient, and the specific obtaining steps are:

[0030] OT = a1 x PO + a2 x OI + a3 x EI;

[0031] In the formula, OT represents the offset index, PO represents the normalized power data offset coefficient, the higher the power data offset coefficient, the more likely the monitoring data does not reflect the true load state, but is disturbed by the sensor itself error, so the offset index rises, which helps the system to identify and respond to potential sensor zero drift problems in time, and ensures the accuracy and stability of data acquisition, OI represents the normalized hardware operation influence coefficient, when the hardware operation influence coefficient rises, for example, the sensor vibration intensity increases, the continuous operation time becomes longer, etc., it means that the physical stress or aging risk of the sensor is also increasing, which is more likely to cause internal reference voltage drift, measurement baseline change and other problems, resulting in the synchronous rise of the overall offset index, EI represents the normalized physical environment influence coefficient, temperature rise may cause thermal expansion, parameter drift or reference voltage instability of internal electronic devices; high humidity environment may cause device surface condensation, insulation performance degradation, and even leakage or weak short circuit phenomenon; and strong electromagnetic interference may cause interference coupling to the analog signal link, inducing false sampling, false pulse or reference signal offset. These adverse environmental conditions will have cumulative effects over time, accelerating the offset of the sensor measurement reference, so that the collected data no longer accurately reflects the true power load state, and then causes the continuous accumulation of systematic measurement error, reducing the stability and reliability of the entire power monitoring system, a1, a2, a3 represent the weight coefficients of the power data offset coefficient, the weight coefficients of the hardware operation influence coefficient and the weight coefficients of the physical environment influence coefficient, and a1+a2+a3=1, a1, a2, a3 are obtained by the analytic hierarchy process, which is a mathematical method for multi-index decision, mainly through constructing a multi-level decision model, and combining expert judgment to compare each factor in pairs, so as to determine the relative importance weight of each index. The basic steps include: constructing a hierarchical structure model of target-criteria-index, forming a pair-wise comparison matrix, calculating the eigenvector to obtain the weight, and finally performing consistency check to verify the rationality of the judgment. In the present application, the "power data offset coefficient", "hardware operation influence coefficient" and "physical environment influence coefficient" are compared by the analytic hierarchy process, and different weights are given according to the influence degree of zero drift in actual scene, so as to ensure that the weighted summation process has scientificity and logical consistency, and improve the credibility and accuracy of the offset index.

[0032] In the present embodiment, it needs to be specifically pointed out that the power data offset coefficient acquisition step is:

[0033] In the monitoring time window, real-time three-phase current data is acquired, and the difference between any two current data is calculated to obtain a three-phase current difference, and the mean value of the three-phase current difference is calculated to obtain a three-phase difference average value;

[0034] The three-phase difference average value is linearly fitted according to the time sequence to obtain a three-phase fitting slope, and the specific acquisition steps are as follows:

[0035]

[0036] In the formula, k1 represents the three-phase fitting slope, represents the inter-phase current imbalance trend, the slope is large, and the offset trend is stronger, m is the number of three-phase current data samples, and avg(t) is the three-phase difference average value at time point t, and t is the time point number;

[0037] In the monitoring time window, real-time power factor data is acquired, and the power factor standard deviation is calculated in each 30-second sliding window, and the power factor standard deviation of all windows is calculated to obtain a power factor deviation degree factor;

[0038] The three-phase fitting slope and the power factor deviation degree factor are normalized, and the power data offset coefficient is calculated according to the normalized three-phase fitting slope and the power factor deviation degree factor, and the specific acquisition steps are as follows:

[0039]

[0040] In the formula, PO represents the power data offset coefficient, k1 represents the three-phase fitting slope, and k2 represents the power factor deviation degree factor. The formula realizes nonlinear fusion of two core offset characteristics by squaring and summing the three-phase fitting slope and the power factor deviation degree factor and then taking the square root. The principle of the formula is to measure the overall offset strength under the joint action of the two dimensionless offset factors using the Euclidean distance form, thereby avoiding the subjective influence caused by artificially setting the weight in weighted summation. The effect of the formula is that when any index abnormally increases, the overall offset coefficient will significantly increase, effectively enhancing the response capability to local abnormal changes. The formula can be used to sensitively identify non-real deviations in power data that may be caused by sensor zero drift, and improve the intelligent perception and judgment accuracy of the system to data stability.

[0041] In this embodiment, it needs to be specifically pointed out that the hardware running influence coefficient acquisition step is:

[0042] The three-axis acceleration in the continuous running process of the power monitoring device is acquired by the acceleration sensor, and the resultant acceleration is calculated according to the three-axis acceleration, and the specific acquisition steps are as follows:

[0043]

[0044] In the formula, A(t) represents the synthetic acceleration of the power monitoring device at time point t, A x (t) represents the synthetic acceleration of the power monitoring device at time point t, A y (t) represents the synthetic acceleration of the power monitoring device at time point t, A z (t) represents the instantaneous acceleration value of the power monitoring device in the X, Y, and Z directions at time point t;

[0045] The standard deviation of the synthetic acceleration and the range of the synthetic acceleration at each time point in the monitoring time window are calculated to quantify the vibration fluctuation and intensity;

[0046] The running vibration intensity factor is calculated according to the standard deviation of the synthetic acceleration and the range of the synthetic acceleration, and the specific acquisition steps are as follows:

[0047]

[0048] In the formula, VF represents the running vibration intensity factor, σ A represents the standard deviation of the synthetic acceleration, and R A represents the range of the synthetic acceleration. The vibration intensity factor is constructed in the form of a product function, which combines stability and intensity characteristics. This expression enhances the dual sensitivity to strong vibration and high fluctuation, avoiding linear superposition distortion;

[0049] The total running time of the power monitoring device is obtained, and the running time factor is calculated according to the total running time, and the specific acquisition steps are as follows:

[0050] TF = log(1 + T 总 );

[0051] In the formula, TF represents the running time factor, T 总 represents the total running time. The logarithmic function can suppress the linear amplification effect of long-time devices on the result, enhancing the recognition of mid-term device state changes;

[0052] The hardware running influence coefficient is calculated according to the running vibration intensity factor and the running time factor, and the specific acquisition steps are as follows:

[0053]

[0054] In the formula, OI represents the hardware running influence coefficient, VF represents the running vibration intensity factor, and TF represents the running time factor. The Euclidean norm form is used for calculation, which can effectively reflect the enhancement of the overall influence index when any factor significantly rises, and has direction sensitivity and amplification characteristics.

[0055] In this embodiment, it needs to be specifically pointed out that the physical environment influence coefficient acquisition step is as follows:

[0056] The temperature data and humidity data of the surface of the power monitoring device shell in the monitoring time window are collected by the temperature sensor and the humidity sensor;

[0057] The standard deviation of the temperature in the monitoring time window is calculated, denoted as the temperature fluctuation degree, which is used to reflect the thermal stability of the shell;

[0058] The humidity exposure factor is calculated according to the humidity data using an exponential amplification function, and the specific acquisition steps are as follows:

[0059]

[0060] In the formula, HE represents the humidity exposure factor, n is the number of time points in the monitoring time window, H(i) represents the humidity value at time point i, and the denominator is the normalization constant e 5 -1≈147.41, which is used to ensure that HE∈[0, 1], and an exponential function e 0.05·H(i) The influence of high humidity environment can be amplified;

[0061] The synergistic fluctuation is calculated according to the temperature fluctuation degree and the humidity exposure factor, and the specific acquisition steps are as follows:

[0062]

[0063] In the formula, XB represents the synergistic fluctuation, σ T represents the temperature fluctuation degree, and HE represents the humidity exposure factor. In the calculation of the synergistic fluctuation, since the temperature standard deviation and the humidity exposure factor belong to different dimensions, the numerical scale difference is large, and if they are directly combined, the expression ability of the humidity factor in the synergistic effect will be reduced. Therefore, the coefficient 10 is introduced to amplify the normalized humidity factor, so that its numerical interval matches the temperature standard deviation, so as to improve the comprehensive sensitivity of the model to environmental unstable factors under high humidity environment, so as to more accurately reflect the potential influence of physical environment on sensor performance stability;

[0064] The physical environment impact coefficient is calculated according to the synergistic fluctuation, and the specific acquisition steps are as follows:

[0065] EI=ln(1+XB);

[0066] In the formula, EI represents the physical environment impact coefficient, and XB represents the synergistic fluctuation. The natural logarithm function is used to compress extreme values and enhance the resolution of the middle segment.

[0067] In this embodiment, it needs to be specifically pointed out that the zero point offset determination step of the power monitoring device according to the offset index is as follows:

[0068] The offset index is compared with an offset threshold value, if the offset index is greater than or equal to the offset threshold value, it is determined that the power monitoring device has zero point drift, if the offset index is less than the offset threshold value, it is determined that the power monitoring device has no zero point drift, the offset threshold value is obtained by an adaptive threshold method, the adaptive threshold method is a strategy of dynamically adjusting the determination threshold according to the system running state, and the basic principle is to automatically calculate and update the optimal threshold based on the statistical characteristics of historical data and the current real-time data fluctuation range, so as to adapt to the changes of environment, load or device state. In the present application, the adaptive threshold method is used to set the offset threshold value for zero point drift judgment, and the mean and standard deviation of the recent offset index are analyzed by a sliding window, the threshold level required for current judgment is dynamically adjusted by combining the judgment accuracy performance in historical drift events, so as to avoid misjudgment or omission caused by insufficient adaptability of the fixed threshold in different application scenarios, and improve the robustness and accuracy of zero point drift judgment.

[0069] Step 3: if it is determined that the power monitoring device has zero point drift, initial power data is collected by the power monitoring device, the initial power data is zero point corrected according to the offset index, and corrected power data is obtained, the power data includes voltage, current, active power, reactive power and power factor, etc.

[0070] Zero point correction of the initial power data can accurately eliminate systematic deviation caused by sensor zero point drift, prevent false high phenomenon of current, voltage or power value, and avoid false alarm, misjudgment fault and energy consumption statistical distortion caused thereby. Through this processing mode, not only the reliability and stability of the monitoring data are improved, but also the self-adaptive ability of the system to equipment aging and environmental interference is enhanced, the dynamic guarantee of monitoring accuracy is realized, the equipment use cycle is effectively prolonged, and the manual maintenance frequency is reduced.

[0071] In the present embodiment, it is necessary to specifically explain that the initial power data is zero point corrected according to the offset index to obtain the corrected power data, and the steps are as follows:

[0072] The offset index and the offset threshold value are obtained, the offset index and the offset threshold value are calculated by ratio, the offset intensity ratio is obtained, and the proportion of the current offset degree relative to the system tolerance range is reflected;

[0073] In order to avoid mutation and misadjustment caused by linear reduction, a hyperbolic tangent function is used to construct a correction factor according to the offset intensity ratio, and the specific acquisition steps are as follows:

[0074] CZ=tanh(R-1);

[0075] In the formula, CZ represents a correction factor, R represents a deviation intensity ratio, the correction factor is designed by using a nonlinear function, the correction range can be adaptively adjusted according to the severity of the drift, so that the non-real baseline deviation part in the original signal can be accurately deducted;

[0076] The minimum value of the acquisition item in the monitoring time window is obtained, and the minimum value is taken as a sensor drift "zero reference" for calculating the correction range, and the corrected data is calculated according to the minimum value and the correction factor, and the specific acquisition steps are as follows:

[0077] JZ(t) = Z(t) - CZ x Z min ;

[0078] In the formula, JZ(t) represents corrected power data, Z(t) represents initial power data, CZ represents a correction factor, Z min represents the minimum value.

[0079] Step 4: Calculate the power anomaly index according to the corrected power data, and determine the power anomaly according to the power anomaly index;

[0080] In this embodiment, it needs to be specifically pointed out that the power anomaly index acquisition step is as follows:

[0081] According to the corrected power data in the set time window, a power consumption data sample set is constructed n is the sample number, U t represents the voltage at time point t, I t represents the current at time point t, P t represents the active power at time point t, Q t represents the reactive power at time point t, PF t represents the power factor at time point t, and the normalized data set is obtained by performing standardization processing on all parameters in the set Eliminate the dimension effect and improve the sensitivity to abnormal fluctuations;

[0082] All parameters in the normalized data set are divided into normal cluster and abnormal cluster by using K-Means clustering method, which respectively represent normal operation sample and abnormal operation sample. K-Means clustering method is a kind of unsupervised learning clustering algorithm, and its basic idea is: through iteration, the given data set is divided into a predetermined number of clustering clusters, so that the similarity between the data points in the same cluster is maximum, and the difference between the different clusters is maximum. This method takes "minimizing the sum of squares of Euclidean distances from sample points to the center of the cluster to which they belong" as the objective function, first randomly selects K initial center points, and then repeatedly executes "assigning samples to the nearest center point" and "recalculating cluster center". Two steps until the cluster center is stable or the iteration number is reached;

[0083] It needs to be specifically pointed out that the use of K-Means clustering method for parameter classification is prior art, and the specific steps are not described in detail in this embodiment.

[0084] The normal clustering center of the normal clustering cluster is obtained, and for each time point t, the Euclidean distance between the power consumption data of the point and the normal clustering center is calculated as the abnormal deviation degree. The Euclidean distance is a commonly used multidimensional space distance measurement method, which is used to measure the straight line distance between two points. In mathematics, the Euclidean distance is expressed as the square root of the square sum of the difference of two vectors in each corresponding dimension, which reflects the geometric distance of the similarity degree of two data points in multidimensional space;

[0085] The jump factor is obtained by calculating the multi-parameter change rate of the current time point and the previous time point, and the specific acquisition steps are as follows:

[0086]

[0087] In the formula, WJ(t) represents the jump factor of the tth time point, 5 represents the number of power consumption data parameters, and Z x,t represents the standardized value of parameter x at time t, Z x,t-1 represents the standardized value of parameter x at time t-1.

[0088] The abnormal deviation degree and the jump factor of each time point are added to obtain the power abnormality index.

[0089] In this embodiment, it needs to be specifically pointed out that the power abnormality determination step according to the power abnormality index is as follows:

[0090] The power abnormality index is compared with the abnormal threshold value, if the power abnormality index is greater than or equal to the abnormal threshold value, it is determined that the power is abnormal; if the power abnormality index is less than the abnormal threshold value, it is determined that the power is not abnormal, and the abnormal threshold value is obtained by the adaptive threshold method.

[0091] Step 5: if the power is determined to be abnormal, the power abnormality warning is performed;

[0092] Step 6: the power consumption data, analysis and calculation results and warning events are stored in the cloud database in the form of time sequence, which is used for historical data backtracking and statistical analysis.

[0093] The power consumption data, analysis and calculation results, and early warning events are stored in a cloud database in a time series form, so that data can be continuously accumulated and structured managed, and reliable basis is provided for subsequent historical data backtracking, trend analysis, and intelligent model optimization. The advantage of this method is that not only can the running state and deviation behavior of the equipment be tracked and compared across the time dimension, but also data sharing and multi-terminal access can be realized through cloud storage, and the scalability and analysis depth of the system are improved. At the same time, the time series-based data structure facilitates the rapid execution of periodic fluctuation detection, abnormal pattern recognition, and other algorithms, and helps to build a more efficient power operation and maintenance management system and a predictive maintenance mechanism.

[0094] Finally, the above is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0095] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which shall be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for monitoring the power consumption of a distribution box, characterized in that, Includes the following steps: Step 1: Install power monitoring equipment at the connection point of the wiring inside the distribution box; Step 2: Obtain data offset information from the power monitoring equipment. The data offset information includes three-phase current data, power factor data, acceleration data, temperature data, and humidity data. The offset index is evaluated based on the data offset information, and the zero-point offset of the power monitoring equipment is determined based on the offset index. Step 3: If it is determined that the power monitoring equipment has zero-point drift, the initial power data is collected through the power monitoring equipment, and the initial power data is zero-point corrected according to the offset index to obtain the corrected power data; Step 4: Calculate the power anomaly index based on the corrected power data, and determine the power anomaly based on the power anomaly index; Step 5: If a power anomaly is detected, issue a power anomaly warning; Step 6: Store electricity consumption data, analysis results, and early warning events in a time-series format in the cloud database; The steps for obtaining the offset index are as follows: Set a monitoring time window, acquire three-phase current data and power factor data within the time window, and calculate the power data offset coefficient based on the three-phase current data and power factor data; An acceleration sensor is installed in the power monitoring equipment to obtain acceleration data and the cumulative running time of the power monitoring equipment. The hardware operation impact coefficient is calculated based on the acceleration data and the cumulative running time. Temperature and humidity sensors are installed in power monitoring equipment to acquire temperature and humidity data, and the physical environment influence coefficient is calculated based on the temperature and humidity data. The power data offset coefficient, hardware operation impact coefficient, and physical environment impact coefficient are normalized. The offset index is then calculated based on these normalized coefficients. The specific steps for obtaining the offset index are as follows: ; In the formula, This is expressed as the offset index. This represents the offset coefficient of the normalized power data. This is expressed as the normalized hardware performance impact coefficient. This is expressed as the normalized physical environment impact coefficient. , , These are represented as the weighting coefficients for the power data offset coefficient, the hardware operation impact coefficient, and the physical environment impact coefficient.

2. The power consumption monitoring method for distribution boxes according to claim 1, characterized in that: The steps for obtaining the power data offset coefficient are as follows: Within the monitoring time window, real-time three-phase current data is acquired, and the difference between any two phase current data is calculated to obtain the phase current difference. The average value of the phase current difference is then calculated to obtain the average value of the three-phase difference. The average three-phase difference is linearly fitted to the time series to obtain the three-phase fitting slope. The specific steps are as follows: ; In the formula, This represents the slope of the three-phase fitting, where m is the number of samples in the three-phase current data. This represents the average of the three-phase differences at time point t, where t is the time point number. Within the monitoring time window, real-time power factor data is acquired, the standard deviation of the power factor is calculated in the sliding window, and the root mean square of the standard deviation of the power factor in all windows is used to calculate the power factor deviation factor. The three-phase fitting slope and the power factor deviation factor are normalized, and the power data offset coefficient is calculated based on the normalized three-phase fitting slope and the power factor deviation factor.

3. The power consumption monitoring method for distribution boxes according to claim 1, characterized in that: The steps for obtaining the hardware operation impact coefficient are as follows: The triaxial acceleration of the power monitoring equipment during continuous operation is obtained by using an accelerometer, and the composite acceleration is calculated based on the triaxial acceleration. Calculate the standard deviation and range of the composite acceleration at each time point within the monitoring time window; The vibration level factor is calculated based on the standard deviation and range of the combined acceleration. The total operating time of the power monitoring equipment is obtained, and the operating time factor is calculated based on the total operating time. The hardware operation impact coefficient is calculated based on the operating vibration level factor and the operating time factor.

4. The power consumption monitoring method for distribution boxes according to claim 1, characterized in that: The steps for obtaining the physical environment influence coefficient are as follows: Temperature and humidity data of the surface of the power monitoring equipment casing are collected within the monitoring time window using temperature and humidity sensors. Calculate the standard deviation of temperature within the monitoring time window, and record it as the degree of temperature fluctuation; The humidity exposure factor is calculated from the humidity data using the exponential amplification function. Co-variance was calculated based on the degree of temperature fluctuation and humidity exposure factor; The physical environment impact coefficient is calculated based on the cooperative volatility.

5. The power consumption monitoring method for a distribution box according to claim 1, characterized in that: The steps for determining the zero-point offset of power monitoring equipment based on the offset index are as follows: The offset index is compared with the offset threshold. If the offset index is greater than or equal to the offset threshold, it is determined that the power monitoring equipment has zero-point drift; if the offset index is less than the offset threshold, it is determined that the power monitoring equipment has not zero-point drift.

6. The power consumption monitoring method for a distribution box according to claim 1, characterized in that: The step of performing zero-point correction on the initial power data based on the offset index to obtain the corrected power data is as follows: Obtain the offset index and offset threshold, and calculate the ratio of the offset index to the offset threshold to obtain the offset intensity ratio; Using the hyperbolic tangent function, a correction factor is constructed based on the offset intensity ratio; Obtain the minimum value of the collected items within the monitoring time window, and calculate the corrected data based on the minimum value and the correction factor. The specific steps are as follows: ; In the formula, This is represented as the corrected power data. Represented as initial power data, Represented as a correction factor, This is represented as the minimum value.

7. The power consumption monitoring method for a distribution box according to claim 1, characterized in that: The steps for obtaining the power anomaly index are as follows: Based on the corrected power data within a set time window, a power consumption data sample set is constructed, and all parameters in the set are standardized to obtain a standardized data set. The K-Means clustering method was used to divide all parameters in the standardized dataset into normal clusters and abnormal clusters; Obtain the normal cluster centers of normal clusters. For each time point t, calculate the Euclidean distance between the electricity consumption data at that point and the normal cluster centers, and use it as the anomaly deviation. The jump factor is obtained by calculating the rate of change of multiple parameters between the current time point and the previous time point; The power anomaly index is obtained by adding the abnormal deviation at each time point to the jump factor.

8. The power consumption monitoring method for a distribution box according to claim 1, characterized in that: The steps for determining power anomalies based on the power anomaly index are as follows: The power anomaly index is compared with the anomaly threshold. If the power anomaly index is greater than or equal to the anomaly threshold, the power is determined to be abnormal; if the power anomaly index is less than the anomaly threshold, the power is determined not to be abnormal.

Images