A method for periodic monitoring of the operating state of an electrical distribution device

Abstract

The present application relates to the technical field of industrial big data, and particularly relates to a periodic monitoring method for operation state of power distribution equipment, comprising: constructing a space physics and working condition mapping relationship based on a physical space structure of a power distribution cabinet and historical operation data; combining multi-dimensional time sequence state to extract an operation baseline and a dynamic transition relationship; identifying and extracting current load characteristics and cabinet door state characteristics to determine disturbance characteristics, and correcting to obtain an equivalent thermal response under a closed cabinet state; calculating a dynamic anomaly index according to an equivalent state deviation, and executing operation state determination, monitoring period dynamic adjustment and load scheduling instruction generation accordingly. The present application decouples multi-dimensional characteristics and reconstructs a real operation working condition through mapping, thereby guaranteeing the accuracy of operation state monitoring and the stability of continuous operation load regulation of a power distribution network.

Description

Technical Field

[0001] This invention belongs to the field of industrial big data technology and relates to a method for periodic monitoring of the operating status of power distribution equipment. Background Technology

[0002] In new power systems, compartmentalized distribution cabinets are located at the edge nodes of the distribution network, undertaking tasks such as power distribution and safety isolation. The stability of the cabinet's operating status directly affects power supply reliability and load distribution. As smart distribution networks gradually improve, distribution nodes continuously generate operational data such as thermal response, load changes, and cabinet door status. This data spans a long period of time and has multiple dimensions, serving as a basis for judging equipment status.

[0003] Existing monitoring methods do not fully utilize the aforementioned data. In practical applications, distribution cabinets are typically treated as single devices with fixed sampling cycles, supplemented by manual inspections, and little consideration is given to the physical isolation structure within the cabinet. Heat transfer from the fault location to the sensor is subject to attenuation and delay, therefore, surface temperature measurements may deviate from the actual temperature of the core connection points, and local anomalies can easily be masked.

[0004] On the other hand, the load on the distribution network constantly changes with the electricity consumption process, and the opening and closing of cabinet doors also alters the heat exchange conditions inside the cabinet. Fixed-interval sampling is insufficient to continuously describe these changes, and it is even more difficult to separate the temperature rise caused by load switching from the dynamic heat exchange disturbance caused by door opening. If the data used for dynamic load regulation, predictive maintenance, and network optimization is still relied upon directly, the data may become distorted. The monitoring of the operating status of power distribution equipment is affected by the dynamic changes in the operating environment and operating conditions, and the collected monitoring data is insufficient to characterize the true operating status of the parts of the equipment under test.

[0005] Based on the above problems, this invention proposes a method for periodic monitoring of the operating status of power distribution equipment. Summary of the Invention

[0006] This invention aims to reduce the impact of space heat transfer, load switching, and cabinet door opening and closing on monitoring results. The method establishes space thermal response and thermal inertia transition relationships, corrects and inverts operational data to obtain the equivalent thermal response value of the closed cabinet; then, based on this, it calculates the dynamic anomaly index, combines load phase and cabinet door status to determine the operating status, adjusts the monitoring cycle, and generates corresponding load scheduling instructions.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: A method for periodically monitoring the operating status of power distribution equipment includes the following steps: Acquire spatial location information, physical structure information, and historical operating data of the thermal response sensors inside the power distribution cabinet and the parts under test, and establish spatial thermal response relationship.

[0008] Based on historical operating data under closed cabinet conditions, different load cycle phases are extracted. The standard thermal response baseline of the test part is obtained by combining the spatial thermal response relationship. The thermal inertial transition relationship is established based on the thermal response changes during the phase transition between adjacent load cycles.

[0009] Acquire the operating data of the distribution cabinet, identify the current load cycle phase, phase transition state and cabinet door state, determine the load transition component by combining the thermal inertia transition relationship, and determine the cabinet door disturbance component based on the difference in thermal response before and after the cabinet door state change; perform closed cabinet equivalent correction inversion on the distribution cabinet operating data based on the cabinet door disturbance component and the spatial thermal response relationship, and filter out the load transition component by combining the current change increase and decrease direction to obtain the closed cabinet equivalent thermal response value of the part to be tested.

[0010] The dynamic anomaly index is obtained by measuring the deviation between the equivalent thermal response value of the closed cabinet and the corresponding standard thermal response baseline. The operating status is determined based on the dynamic anomaly index, and the next monitoring cycle of the thermal response sensor is adjusted. When the preset alarm conditions are met, an alarm is issued and a load dispatching command is generated.

[0011] Preferably, the process of establishing the spatial thermal response relationship includes: acquiring the linear spatial distance and relative radiation angle between the part to be measured and the thermal response sensor, as well as the environmental parameters inside the distribution cabinet; calculating the thermal radiation compensation coefficient based on the linear spatial distance and relative radiation angle, combined with the thermal radiation attenuation law; calculating the convective heat transfer compensation coefficient based on the physical structure information and environmental parameters inside the distribution cabinet, combined with the natural convection heat transfer model; and constructing a weighted calculation model based on the linear spatial distance, relative radiation angle, and environmental parameters, i.e., the spatial thermal response relationship, by combining the thermal radiation compensation coefficient and the convective heat transfer compensation coefficient.

[0012] Preferably, the process of extracting different load cycle phases and obtaining a standard thermal response baseline includes: collecting load current data and temperature data of the distribution cabinet in historical operating cycles, and performing outlier removal and low-frequency long-term drift data stripping to obtain historical operating data; using a clustering algorithm to divide the load current data in the historical operating data into multiple steady-state current intervals, and mapping each steady-state current interval to a load cycle phase; when the cabinet door is closed and the internal heat distribution reaches equilibrium, extracting the inversion temperature of the test part under each steady-state current interval as the steady-state temperature of the data reference point; performing curve fitting on each data reference point to establish a standard thermal response baseline with load current as input and steady-state temperature as output.

[0013] Preferably, the process of establishing the thermal inertial transition relationship includes: when a cross-interval jump in the load cycle phase is detected, extracting the time delay characteristics of the temperature tracking current change of the measured part and the initial temperature at the time of the jump; determining the temperature of the standard thermal response baseline corresponding to the target load cycle phase after the jump, as well as the thermal time constant; constructing a first-order thermal inertial transition relationship with the standard thermal response baseline temperature as the steady-state target, and with the signed difference between the initial temperature and the steady-state target decaying exponentially with time.

[0014] Preferably, the process of determining the load transition component by combining the thermal inertia transient relationship includes: obtaining the load current change in the current load cycle phase and calculating the corresponding steady-state thermal power increment; setting the thermal time constant in the thermal inertia transient relationship as the system adjustment parameter, taking the steady-state thermal power increment as the target amplitude input, and constructing a transient thermal response compensation function that approaches the steady state exponentially with time; during the phase transition, calculating the normal load transient temperature change of the measured part at the current moment through the transient thermal response compensation function, taking the direction of increase or decrease of current change as the physical polarity attribute, assigning corresponding positive and negative polarity values ​​to the normal load transient temperature change, and obtaining the load transition component.

[0015] Preferably, determining the cabinet door disturbance component based on the difference in thermal response before and after the cabinet door state change includes: monitoring the cabinet door state of the power distribution cabinet; when the cabinet door state changes from closed to open, collecting the ambient temperature parameters inside and outside the power distribution cabinet; calculating the signed temperature difference between the ambient temperature outside the power distribution cabinet and the internal air temperature, and calculating the environmental convective heat exchange value caused by the signed temperature difference based on the convective heat dissipation law; determining the physical transfer direction of heat loss and backflow based on the positive and negative polarity of the signed temperature difference, and equivalently mapping the environmental convective heat exchange value to the cabinet door disturbance component of the test part corresponding to the physical transfer direction.

[0016] Preferably, the specific process for calculating the closed-cabinet equivalent thermal response value of the part to be tested in the distribution cabinet includes: acquiring the operating data of the distribution cabinet, inputting the real-time temperature measurement value of the thermal response sensor into the spatial thermal response relationship, and inverting to obtain the basic inversion temperature; based on energy conservation and thermal offset logic, superimposing the basic inversion temperature with the cabinet door disturbance component with physical transmission direction attribute to obtain the environmental correction temperature after excluding boundary thermal disturbance; according to the transient thermal change trend characterized by the positive and negative polarity of the load conversion component, combined with the thermal time constant, performing a symbolic compensation algebra operation on the environmental correction temperature and the load conversion component based on tending towards thermal equilibrium to calculate the closed-cabinet equivalent thermal response value of the part to be tested.

[0017] Preferably, the specific process for obtaining the dynamic anomaly index includes: calculating the difference between the equivalent thermal response value of the closed cabinet at the current monitoring time and the corresponding standard thermal response baseline to obtain the absolute value of the temperature deviation; within a preset time window, summing the discrete sampling deviation data with a difference that is continuously greater than zero using the trapezoidal numerical integration method to obtain the integral area value representing the cumulative amount of continuous heat generation; and normalizing and weighting the absolute value of the temperature deviation and the integral area value to obtain the dynamic anomaly index.

[0018] Preferably, the process of judging the operating status and dynamically adjusting the monitoring cycle includes: setting a preset abnormal threshold, an initial monitoring cycle, and a minimum safe monitoring cycle; if the dynamic abnormal index does not reach the preset abnormal threshold, the operating status is determined to be normal, and the monitoring cycle is gradually extended to the initial monitoring cycle; if the dynamic abnormal index reaches the preset abnormal threshold, the operating status is determined to be suspicious, and the monitoring cycle is shortened; if the monitoring cycle is shortened to the minimum safe monitoring cycle, the operating status is determined to be abnormal, an alarm signal is issued, a load dispatching command is output, and the load is reduced; if the dynamic abnormal index still reaches the preset abnormal threshold, a trip linkage signal is output.

[0019] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention establishes a spatial thermal response model based on spatial distance, relative radiation angle, and physical property parameters, and determines the thermal inertia transition relationship by combining the load phase switching process, thereby restoring the load thermal state of the distribution node. This result can be used to analyze the impact of power flow fluctuations and load abrupt changes on the node's carrying capacity, and can be converted into a load margin index, reducing misjudgments caused by thermal decay and response lag, thus making load distribution and power dispatch more stable under complex operating conditions.

[0020] 2. This invention identifies load phase changes and cabinet door status changes separately, calculates load transition components and cabinet door disturbance components, and then corrects and inverts the operating data according to the energy balance relationship. This separates the heat dissipation caused by door opening during inspections from the heat accumulation caused by increased load, avoiding the accidental transmission of environmental disturbances into the load control loop and reducing unnecessary adjustment actions.

[0021] 3. This invention simultaneously considers the instantaneous deviation between the equivalent thermal response of the closed-circuit power supply and the standard baseline, as well as the cumulative amount of continuous heat generation over a certain period, thereby deriving a dynamic anomaly index. The control strategy employs load smoothing transfer, load reduction, and tripping protection sequentially according to the degree of anomaly, and the monitoring cycle is adjusted according to changes in the state. Compared to direct tripping upon reaching a threshold, this method can reduce the impact of sudden power flow changes and unplanned power outages on the distribution network. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating the overall process of the periodic monitoring method of the present invention. Figure 2This is a flowchart of the operation status judgment and dynamic adjustment of the monitoring cycle of the present invention; Figure 3 A flowchart is provided to establish the spatial thermal response relationship of this invention. Detailed Implementation

[0023] The technical solution of the present invention will be described below with reference to the accompanying drawings and embodiments. The listed embodiments are used to illustrate specific implementation methods and do not constitute a limitation on the scope of protection of the present invention.

[0024] Example 1: This embodiment illustrates the overall execution process of the method, covering the initial configuration of the compartmented distribution cabinet, long-term online thermal status monitoring, and alarm and adjustment after anomalies.

[0025] Please see Figure 1 The present invention provides a method for periodic monitoring of the operating status of power distribution equipment. The technical solution is as follows: A method for periodic monitoring of the operating status of power distribution equipment includes the following steps: acquiring the spatial location information, physical structure information and historical operating data of the thermal response sensor inside the distribution cabinet and the part to be measured, and establishing a spatial thermal response relationship.

[0026] Based on historical operating data under closed cabinet conditions, different load cycle phases are extracted. The standard thermal response baseline of the test part is obtained by combining the spatial thermal response relationship. The thermal inertial transition relationship is established based on the thermal response changes during the phase transition between adjacent load cycles.

[0027] Acquire the operating data of the distribution cabinet, identify the current load cycle phase, phase transition state and cabinet door state, determine the load transition component by combining the thermal inertia transition relationship, and determine the cabinet door disturbance component based on the difference in thermal response before and after the cabinet door state change; perform closed cabinet equivalent correction inversion on the distribution cabinet operating data based on the cabinet door disturbance component and the spatial thermal response relationship, and filter out the load transition component by combining the current change increase and decrease direction to obtain the closed cabinet equivalent thermal response value of the part to be tested.

[0028] The dynamic anomaly index is obtained by measuring the deviation between the equivalent thermal response value of the closed cabinet and the corresponding standard thermal response baseline. The operating status is determined based on the dynamic anomaly index, and the next monitoring cycle of the thermal response sensor is adjusted. When the preset alarm conditions are met, an alarm is issued and a load dispatching command is generated.

[0029] Further, see Figure 3The process of establishing the spatial thermal response relationship includes: acquiring the linear spatial distance and relative radiation angle between the part to be measured and the thermal response sensor, as well as the environmental parameters inside the distribution cabinet; calculating the thermal radiation compensation coefficient based on the linear spatial distance and relative radiation angle, combined with the thermal radiation attenuation law; calculating the convective heat transfer compensation coefficient based on the physical structure information and environmental parameters inside the distribution cabinet, combined with the natural convection heat transfer model; and constructing a weighted calculation model based on the linear spatial distance, relative radiation angle, and environmental parameters, i.e., the spatial thermal response relationship, by combining the thermal radiation compensation coefficient and the convective heat transfer compensation coefficient.

[0030] Specifically, obtain the position coordinates of the part to be measured in the three-dimensional spatial coordinate system inside the power distribution cabinet. With thermal response sensor position coordinates The Euclidean distance formula is used to calculate the linear spatial distance. : Obtain the axial normal vector of the thermal response sensor Vector of the line connecting two points The angle between two vectors is calculated based on the vector dot product formula and used as the relative radiation angle. : Read environmental parameters stored in the system configuration database, including air density. air kinematic viscosity thermal conductivity of air Air expansion coefficient and specific heat capacity at constant pressure ; To address the radiation obstruction and multiple reflection effects caused by components such as metal partitions and wire harnesses within the distribution cabinet, the linear spatial distance... Spatial Boolean intersection detection is performed between the ray and the internal physical structure information of the distribution cabinet to extract the line-of-sight obstruction rate along the ray path. Based on linear spatial distance and relative radiation angle By incorporating the thermal radiation attenuation law with a non-line-of-sight occlusion penalty term, a modified thermal radiation compensation coefficient is constructed. Algebraic calculation formula: In the formula, This represents the surface emissivity of the material in the part being measured; in this embodiment, it is taken as [value missing]. ; Indicates the absorption coefficient of the air medium; This is a preset minimum anti-overflow constant used to prevent numerical collapse when the denominator of a division operation approaches zero. In this example, it is set to... ; The equivalent transmittance after multiple reflections within the cavity is determined based on the steady-state convergence extremum of the infrared radiation energy after multiple diffuse reflections on the inner surface of the fully enclosed metal cavity. The preferred value range is [value range missing]. to In this embodiment, we take .

[0031] To avoid large angle measurement conditions Numerical singularities that cause the energy decay coefficient to diverge; setting a limiting angle threshold. When the calculated relative radiation angle At that time, a mandatory order ; Values to In this embodiment, we take ; Medium absorption coefficient The method for determining the relative humidity is as follows: the current relative humidity is obtained based on a humidity sensor deployed inside the distribution cabinet. According to the formula It is obtained through dynamic calculation, where The dry air baseline absorbance is set to a value of , This is the humidity-sensitive gain coefficient, with a value of [value missing]. Determine qualitative dimensions based on the internal physical structure information of the distribution cabinet. , expressed as ,in Here are the vertical coordinates of the inner wall of the distribution cabinet. To accurately quantify the transient natural convection heat transfer intensity caused by the heating of the measured parts inside the distribution cabinet, this embodiment introduces a nonlinear iterative convergence architecture to coupledly solve for the convection heat transfer compensation coefficient. Specifically, initialize the number of iteration steps. And set the initial temperature rise estimate for the part to be measured. The initial temperature rise estimate is set based on the predicted temperature difference in the normal environment under the initial operating conditions of the equipment, and its preferred value range is: to In this embodiment, the specific value is... Perform the following fluid-temperature cycle iterative steps: Step 1: Dynamic temperature rise variable based on the current iteration step Calculate the Glashoff number And combined with Plant number Solving Rayleigh numbers : In the formula, To minimize the overflow prevention constant, this embodiment uses a fixed value. ; Step 2: Based on the coordinates of the part to be measured in the three-dimensional spatial topology model inside the distribution cabinet. Along a horizontal ray emitted in both directions parallel to the X-axis, retrieve the coordinates of the first intersection point with the nearest internal obstacle or inner wall surface. Coordinates of the second intersection point The feature's horizontal width is calculated. Then calculate the dimensionless aspect ratio. ; Step 3: Under narrow cavity, logarithmic terms It may diverge towards negative infinity. To limit the flow regime index to the actual range of laminar to turbulent flow, a two-way saturation cutoff function is added to the calculation formula, and the convection shape coefficient of the current iteration step is obtained. With flow index : In the formula, To avoid anomalies in the logarithmic function at zero, a preset minimum anti-overflow constant is used in this embodiment. ; and The lower and upper limits of the physical extrema of the convection shape coefficient are truncated based on the heat transfer reference constant of natural convection on a standard vertical flat plate in a confined space, thus limiting the calculation results to the physically feasible region. The preferred value ranges are as follows: to as well as to In this embodiment, , ; and The physical extreme values ​​of the flow regime index are defined by lower and upper limits, based on the flow regime evolution laws of thermodynamic natural convection (from laminar heat transfer). Power-order heat transfer to turbulent flow The absolute boundary of the power transition), and the preferred value ranges are respectively to as well as to In this embodiment, .

[0032] Step 4: Calculate the Rayleigh number corresponding to the current iteration step. A natural convection heat transfer model for a closed cabinet was constructed, and the Nusselt number was calculated. Based on the Nusselt number, air thermal conductivity, and characteristic vertical height, calculate the natural convection heat transfer coefficient of the enclosed cabinet corresponding to the current iteration step. It is used to characterize the intensity of natural convection heat transfer in the confined space surrounding the measured part when the cabinet door is closed. In the formula, To prevent the denominator of the characteristic length from approaching a very small positive number of zero, this example takes... Introducing a closed cabinet to reference the natural convection heat transfer coefficient. This is used to weight the convection term with the dimensionless thermal radiation compensation coefficient, and is not used as the baseline heat transfer coefficient for the door-opening disturbance model. The dimensionless convective heat transfer compensation coefficient is also calculated. : in, The natural convection heat transfer coefficient, pre-calibrated under conditions of cabinet door closure, reference ambient temperature, reference load, and predetermined sensor arrangement, is used for natural convection normalization of the enclosed cabinet. To prevent the normalized denominator from approaching a very small positive number of zero, this example sets it to... .

[0033] Step 5: Adjust the dimensionless thermal radiation compensation coefficient The dimensionless convective heat transfer compensation coefficient of the current iteration step By performing a weighted combination, the dimensionless spatial thermal response weighted value for the current iteration step is obtained. : The preferred range of values ​​for the weight is: to Based on the calculation results of the current iteration step in this embodiment, the radiative heat transfer weight is adjusted. Updated to and the convective heat transfer weight Updated to The numerical values ​​mentioned are only used to illustrate the calculation results under this working condition and do not replace the weight calculation relationship; Step 6: Calculate the dimensionless spatial thermal response weighted value obtained in the current iteration step. Calculate the estimated temperature of the part to be measured. Read the temperature measurement value of the thermal response sensor at the current sampling time. and the reference temperature of the cabinet environment The estimated temperature of the measured part is calculated based on the dimensionless spatial thermal response weighted value of the current iteration step: Update the dynamic temperature rise variable for the next cycle step based on the estimated temperature above: In the formula, This indicates the upper limit of safe temperature rise set according to the heat resistance limit of the insulation material of the distribution cabinet. In this embodiment, the value is fixed. ; Step 7: Calculate the absolute value of the residual of dynamic temperature rise between two adjacent iteration steps: when When the current iteration reaches the convergence condition, the iteration is terminated, and the corresponding dimensionless spatial thermal response weighted value is calculated. As the final scalar operator of the space thermal response relationship .when Furthermore, the current iteration count did not reach the maximum iteration count. season Return to step one and continue to the next iteration. And this iteration count has reached the maximum number of iterations. If the iteration fails to reach convergence, the current iteration is stopped, and it is determined that the iteration has not met the convergence condition. The output of the dimensionless spatial thermal response scalar operator from the most recent converged historical monitoring period is retrieved, and a low-confidence flag is added. The spatial thermal response with the low-confidence flag is only used for status display and subsequent data acquisition, and is not used to generate automatic load reduction, tripping, or load transfer commands. If no valid historical convergence value exists, the dimensionless spatial thermal response weighted value corresponding to the minimum absolute value of the temperature rise residual in the current iteration is selected as the output, and a low-confidence flag is added. The convergence threshold is specified in the provided text. The preferred value range is to In this embodiment, we take Maximum number of iterations The preferred value range is to In this embodiment, .

[0034] Through the above iterative convergence process, and by combining the spatial continuous flow state function to dynamically correct the natural convection state, the influence of fixed discrete parameters and large load fluctuations on the natural convection calculation results can be reduced, making the estimated steady-state temperature of the measured part closer to the actual thermal state, while improving the stability of the calculation process under different load conditions.

[0035] Furthermore, the process of extracting different load cycle phases and obtaining a standard thermal response baseline includes: collecting load current and temperature data of the distribution cabinet in historical operating cycles, and performing outlier removal and low-frequency long-term drift data stripping to obtain historical operating data; using a clustering algorithm to divide the load current data in the historical operating data into multiple steady-state current intervals, and mapping each steady-state current interval to a load cycle phase; when the cabinet door is closed and the internal heat distribution reaches equilibrium, extracting the inversion temperature of the test part under each steady-state current interval as the steady-state temperature of the data reference point; performing curve fitting on each data reference point to establish a standard thermal response baseline with load current as input and steady-state temperature as output.

[0036] Specifically, a one-dimensional sliding time window is constructed, and the discrete time span of the one-dimensional sliding time window is set to the number of sampling points. Discrete time span Balancing noise smoothing and real-time response, a preset time-domain sliding window duration is selected. Divide by the data sampling period And round down; duration of the time-domain sliding window The selectable range is minutes to minutes, this embodiment takes Minutes. Acquire data from historical operating cycles and calculate the discrete time span of each sampling point in the load current data sequence. Local standard deviation within; standard deviation greater than a preset rate of change threshold. Sampling points are removed, and a preset rate of change threshold is used. Used to identify current spikes caused by instantaneous grid switching or sudden pulses; selectable range is rated current. to In this embodiment, the rated current is taken as... A first-order high-pass filter is used to estimate the low-frequency variation components in historical temperature data, and the low-frequency complementary components corresponding to the high-pass filter are used as the low-frequency long-term drift data. The output of the high-pass filter is only used to identify drift characteristics in temperature changes and is not directly used as the absolute temperature of the measured part; the difference equation of the first-order high-pass filter is expressed as: In the formula, The current input temperature, The output temperature after filtering. These are the filter coefficients; filter coefficients The calculation rule is: based on the data sampling period With the preset equivalent cutoff frequency Constructing the discretization mapping formula: In the formula, the equivalent cutoff frequency is... According to the long cycle of interference that needs to be removed Setting, expressed as The preset target stripping cycle in this embodiment Set as Hours (i.e.) (seconds), used to filter out interference from day / night and seasonal environmental temperature regulation; the selectable range is calculated according to different sampling rates. to The sampling period of the corresponding data in this embodiment The specific value obtained by the above mapping formula is Before clustering, historical load currents are screened for thermal steady-state conditions to avoid the influence of random load changes and thermal inertia on the steady-state baseline: the first-order time-domain derivative of the load current is calculated using differential circuits or discrete difference equations. Set a sliding window for determining steady state. Sliding window for determining steady state The preferred discrete span minutes to minutes, this embodiment takes Minutes; thermal steady-state data are determined in the following manner: based on the current sampling time. As the endpoint, backtrack to the historical time domain by a preset thermal delay lag window duration. (This embodiment takes) minutes), if the entire length is The absolute value of the first time derivative of the load current within the backtracking time domain interval In each Within each sub-interval where the step size slides forward, the current fluctuation threshold is less than the current fluctuation threshold. Then determine the current time. The load current data is in thermal steady state. Among them, the current fluctuation threshold... Designed to tolerate normal, small-amplitude, transient power flow fluctuations. The selectable range is [range missing] of the distribution cabinet's rated current. to In this embodiment, the rated current is taken as... The preset thermal delay hysteresis window is used to match the time-domain hysteresis effect of the large cross-section metal busbar from current heating to temperature stabilization conduction. The selectable range is [insert range here]. minutes to minutes, this embodiment takes Minutes; the filtered results form a historical steady-state load current data set. ,in This represents the total number of normal sampling points after being filtered through steady-state physical constraints.

[0037] K-means clustering algorithm based on Euclidean distance was used to analyze the historical operating load current data set. Perform feature interval division and random initialization. Each cluster has a centroid. To ensure the reproducibility of load cycle phase division results corresponding to the same historical operating data, this embodiment uses a fixed cluster initialization seed and performs multiple independent initializations, selecting the clustering result with the smallest sum of squared errors. After clustering convergence, the load currents corresponding to each cluster centroid are sorted, and the load cycle phase number is determined according to the sorting results. If the number of effective steady-state samples is less than the preset minimum number of samples, no re-clustering is performed in this round, and the most recent effective steady-state current interval is used. The objective function is to minimize the sum of squared errors. Iteratively update the centroid position: In the formula, Indicates the first The set of load currents of each cluster. For the first The centroid mean of each cluster. Number of clusters. Used to refine the typical stage characteristics of daily electricity load in the distribution network, the selectable range is: to In this embodiment, After convergence, the boundary values ​​of each characteristic interval correspond to different steady-state current intervals. ,in It serves as a cluster index, mapping each feature interval to an independent load cycle phase. .

[0038] The current cabinet door monitoring status variables are obtained by using travel status sensors or proximity sensors deployed on the cabinet door. ,when (Indicating closure), the current moment is not in the door recovery period and the rate of temperature change over time conforms to... At this point, it is determined that the internal heat distribution has reached equilibrium. Equilibrium threshold. The setting is based on determining whether electromagnetic heating and convective heat dissipation inside the cabinet have reached a quasi-steady-state thermal equilibrium, and its preferred value range is... to In this embodiment, In this state, the scalar operator for the spatial thermal response relationship obtained after iterative convergence. The sensor temperature is inverted, and the average inversion temperature of the measured part within the corresponding load cycle phase is extracted. The specific temporal intersection alignment extraction rule is as follows: construct a steady-state mask set, traverse historical time-series data, and extract the cabinet door monitoring state variables that simultaneously meet the requirements. Rate of change of temperature over time Furthermore, sampling moments where the spatial thermal response iteration results have reached convergence are marked as quasi-steady-state sample points; spatial inversion results where the current spatial thermal response iteration has not reached convergence or have low-confidence markers are not included in the phase local quasi-steady-state point set, nor are they used for data benchmark steady-state temperature calculation and standard thermal response baseline update, nor are automatic load reduction, tripping, or load transfer commands generated based on these results. Then, temporal spatial feature alignment is performed for any load cycle phase. Among the quasi-steady-state sample points, the instantaneous value of the load current data falls within the steady-state current range corresponding to that phase. The intersection data pairs within the set are used to form a local quasi-steady-state point set for the phase; finally, the algebraic average of all inversion temperatures within the local quasi-steady-state point set is calculated as the steady-state temperature of the data reference point. Simultaneously calculate the algebraic average value of the load current within this point set as... For each quasi-steady-state sample point, the corresponding long-term drift estimate is first subtracted from the original absolute temperature to obtain a corrected temperature that retains the absolute temperature reference; or the steady-state temperature rise of the measured part relative to the ambient reference temperature inside the cabinet is calculated. Subsequent data reference points and the standard thermal response baseline are all established based on the corrected temperature or steady-state temperature rise, and the high-pass filter output is not directly used as the absolute temperature. The standard thermal response baseline function is established by performing a polynomial fitting on the set of multiple data reference points using the least squares method: In the formula, The input is the load current, and the output is the corresponding steady-state temperature. These are the fitted regression coefficients. The formula for solving the fitted regression matrix is: In the formula, the Vandermonde matrix Observation vector Before solving the fitted regression matrix, the number of valid baseline data points, the rank of the Vandermonde matrix, and the condition number are determined. If the number of valid baseline data points is insufficient to support the current fitting order, the matrix is ​​not of full rank, or the condition number exceeds the preset upper limit, the current solution is not adopted. Instead, the fitting order is reduced, regularized least squares is applied, or a backtracking process using the most recent valid standard thermal response baseline is performed sequentially. The current fitting result is then marked as a low-confidence state.

[0039] The preliminary physical screening and underlying fitting process in this step establishes an ideal thermal performance model of the distribution cabinet under the condition of eliminating transient interleaved data interference, weakens the causal inversion error caused by random fluctuations in power flow, and provides an absolute reference benchmark with physical confidence for subsequent multidimensional disturbance stripping.

[0040] Furthermore, the process of establishing the thermal inertial transition relationship includes: when a cross-interval jump in the load cycle phase is detected, extracting the time delay characteristics of the temperature tracking current change of the measured part and the initial temperature at the time of the jump; determining the temperature of the standard thermal response baseline corresponding to the target load cycle phase after the jump, as well as the thermal time constant; constructing a first-order thermal inertial transition relationship with the standard thermal response baseline temperature as the steady-state target, and with the signed difference between the initial temperature and the steady-state target decaying exponentially with time.

[0041] Specifically, when the real-time load current crosses the boundary point of adjacent steady-state current intervals, causing the load cycle phase to change... Abrupt change to target load cycle phase When the status monitoring is triggered, the moment of phase cross-interval transition is extracted from the current distribution cabinet operation data stream within the monitoring period. The basic inversion temperature of the part to be measured is used as the initial temperature. To accurately calculate the response hysteresis from discrete temperature sampling and eliminate the interference of high-frequency discrete noise in the extraction of the extreme value of the rate of change, time delay characteristics are used. The real-time dynamic acquisition process is as follows: a discrete first-order low-pass smoothing filter is constructed to smooth and denoise the base inversion temperature after the jump, resulting in a denoised continuous temperature sequence in the time domain; a first-order time-domain difference is performed on the denoised continuous temperature sequence to obtain the temperature change rate sequence, and a threshold for the abrupt response change rate is set. Its preferred value range is to In this embodiment, During the real-time data stream progression of status monitoring, the current transition time is taken as the starting point. Start timing counting from the beginning; when the value in the temperature change rate sequence first exceeds the threshold of the abrupt response change rate... And in the subsequent preset verification sliding window The internal temperature change rate continues to increase monotonically or remains at the threshold. When the threshold is exceeded, the moment when it is first exceeded is determined as the inflection point. Preset verification sliding window The discrete value is obtained by dividing the reference time span by the data sampling period, and its preferred range is [value range missing]. minutes to minutes, this embodiment takes Minutes; calculating the inflection point time using subtraction. With the time of the jump The difference, i.e. This serves as the time delay feature for tracking current changes at temperature. At the current moment No inflection point meeting the above conditions has been found yet. Previously, it was assumed that the system was in the response dead zone, and the first-order thermal inertia transition relationship was automatically forced to be a constant value. Until the turning point Triggered activation sets the maximum delay recognition time. If the accumulated waiting time since the load phase transition reaches the maximum delay recognition time and no inflection point meeting the conditions is detected, the current inflection point search terminates. At this point, the median effective time delay corresponding to the same historical load phase transition is preferentially used as the backoff value; if no effective historical value exists, the current load transition process is marked as a low-confidence state, the process is not used to update the thermal inertia transition relationship, and the alarm or manual review status is maintained. This ensures the causal realizability and timing logic closed loop of the algorithm under streaming media data. The target load current after the current transition is... Substituting the aforementioned standard thermal response baseline function In the process, the phase of the target load cycle is calculated. Corresponding standard thermal response baseline temperature Obtain the inherent thermal time constant of the internal equipment and metal busbar structure of the distribution cabinet. Inherent thermal time constant Based on the equivalent heat capacity of the pre-set busbar material in the distribution cabinet Dynamic equivalent thermal resistance under the current flow boundary The product is calculated and expressed as: Among them, equivalent heat capacity The structural constants, pre-calibrated based on the specific heat capacity and geometric mass of the busbar material, are directly stored in the system configuration database; dynamic equivalent thermal resistance The dynamic adaptive solution rule is: the natural convection heat transfer coefficient of the closed cabinet when the aforementioned natural convection iteration reaches the convergence condition. Combined with the characteristic heat transfer surface area of ​​the busbar Calculate the dynamic equivalent thermal resistance under the current flow boundary: In the formula, ,in This represents the iteration step corresponding to the convergence condition of the natural convection iteration. Dimensionless convective heat transfer compensation coefficient. It is used only for space thermal response weighting and is not involved in the calculation of dynamic equivalent thermal resistance and thermal time constant. It is constructed based on the standard thermal response baseline temperature. As a steady-state objective, and with initial temperature The first-order thermal inertial transient algebraic equation, in which the signed difference between the thermal inertial target and the steady-state target decays exponentially with time, serves as the first-order thermal inertial transient relation: In the formula, This indicates the current runtime after the transition occurred; when in the response dead zone, i.e. At that time, the first-order thermal inertia transient response is set to a constant value, i.e. This characterizes the state where the temperature of the measured part has not yet physically responded to the sudden change in current within the time delay characteristic segment. This is to reduce the first-order thermal inertia transition relationship at the segmented critical point. To address numerical step spikes caused by derivative discontinuities, this embodiment of the program executes a parameter with a preset length of [value missing] after the critical point. The transition boundary sequence is described. To reduce numerical spikes caused by derivative discontinuities at the piecewise critical points of the first-order thermal inertia transition relationship, cubic Hermite interpolation is used for smooth connection within a preset transition interval before and after the critical point. The interpolation boundary conditions include the function values ​​and first-order derivative values ​​at both ends of the critical interval, ensuring continuity of the function values ​​and first-order derivative before and after the critical point; a preset length is also described. The step size is obtained through adaptive calculation based on the baseline smoothing step size. In this embodiment, it is taken as... One sampling period.

[0042] The above processing expresses the response delay of heat conduction and the process of recovery to steady state as a continuous time-varying relationship. Compared with direct segmentation, the numerical abrupt changes at the critical positions are smaller, which can serve as a time reference for subsequent identification of transient load components and separation of abnormal heating.

[0043] Furthermore, the process of determining the load transition component by combining the thermal inertia transient relationship includes: obtaining the load current change in the current load cycle phase and calculating the corresponding steady-state thermal power increment; setting the thermal time constant in the thermal inertia transient relationship as the system adjustment parameter, taking the steady-state thermal power increment as the target amplitude input, and constructing a transient thermal response compensation function that approaches the steady state exponentially with time; during the phase transition, calculating the normal load transient temperature change of the measured part at the current moment through the transient thermal response compensation function, taking the direction of increase or decrease of current change as the physical polarity attribute, assigning corresponding positive and negative polarity values ​​to the normal load transient temperature change, and obtaining the load transition component.

[0044] In this embodiment, according to Joule's law, the heating power is proportional to the square of the current. Based on the load current from the initial load current... Switch to target load current Calculate the steady-state heat power increment based on the change in load current. Mapping expression: In the formula, The equivalent AC / DC resistance of the part to be measured. The specific acquisition rules are as follows: During initial configuration, the Joule thermal power mapping distortion caused by individual equipment differences is reduced by injecting a test current of known scale into the busbar under test and reading the voltage drop across the busbar under test during offline measurement. The selectable range is... to In this embodiment, In order to increase the steady-state thermal power increment Coupled to the temperature dynamics model, considering that the theoretical convective thermal resistance cannot fully cover the cross-radiation of multiple heat sources, transient compensation may not converge to the standard baseline. Therefore, this embodiment introduces historical steady-state consistency calibration. The previously established standard thermal response baseline function is retrieved. Calculate the target load current With initial load current The steady-state temperature rise difference at the baseline; using the dynamic equivalent thermal resistance calculated based on the Nusselt number above. Constructing an empirically consistent calibration factor Algebraic calculation formula: In the formula, To preset a minimum anti-overflow constant, this embodiment takes a fixed value of... Using empirical consistency calibration factors The temperature rise response given by the physical model is calibrated, and the target amplitude of the non-negative ideal steady-state temperature rise response that meets the convergence requirement is calculated. : This is used to ensure that the steady-state temperature rise response amplitude calculated by the theoretical Joule thermal model is consistent with the steady-state temperature difference characterized by the standard thermal response baseline. When the denominator does not trigger the minimum protection condition, the calibrated ideal steady-state temperature rise response amplitude is equal to the absolute value of the baseline steady-state temperature difference corresponding to the target load current and the initial load current; this amplitude is only used to determine the final value of the transient response under normal load and is not updated using the current real-time temperature measurement value or diagnostic deviation. The empirical consistency calibration in this embodiment is only used to determine the final value of the transient response corresponding to normal load changes and does not accept online write-back of subsequent long-cycle contact resistance estimation results. The aforementioned inherent thermal time constant is set. As a system regulation parameter, a transient thermal response compensation function is constructed that approaches steady state exponentially with time: In the formula, The effective duration from the end of the response dead zone to the current sampling time is represented as: ; and only if At that time, the transient thermal response compensation function is activated. ;when Forced transient thermal response compensation function The output value is This ensures the numerical and slope continuity of the thermal inertia decay model at the time domain boundary.

[0045] When the load phase switches, transient calculations are initiated under the following conditions: the first time derivative of the real-time load current reaches a threshold, i.e., the following conditions are met. At that time, Set as And begin calculating the load transition components under unsteady conditions.

[0046] After entering the transient state, as long as the steady-state screening conditions have not been met again, even It has dropped to the following, Still remain as The calculation of the separation of load dynamic components continues. The flag will be cleared once the steady-state baseline can be retrieved again. Threshold When used in conjunction with steady-state screening conditions, it can prevent the transient process from being prematurely terminated due to a short-term current drop. Its preferred value range is... to In this embodiment, we take .

[0047] The real-time temperature measurement value at the current sampling moment is obtained by a thermal response sensor deployed in the power distribution cabinet. Simultaneously acquire the ambient temperature reference temperature inside the distribution cabinet, as measured by an ambient temperature sensor deployed at the ambient temperature reference location inside the cabinet. Considering that Celsius temperature is an interval scale, to ensure the unique physical meaning of the compensation operation and avoid systematic deviations in the reference temperature caused by direct multiplication scaling, the actual temperature rise of the real-time temperature measurement relative to the reference temperature of the cabinet environment is first extracted. Then, this is compared with the final spatial thermal response scalar operator corresponding to the convergence condition of the aforementioned iteration. Perform product inversion, then algebraically compensate and add back the reference temperature to reconstruct the basic inversion temperature of the measured part. : During the phase transition, the aforementioned transient thermal response compensation function is invoked to calculate the transient temperature change of the measured part caused by the normal load change at the current moment. The load change direction coefficient is defined. When the target load current is greater than the initial load current, let When the target load current is less than the initial load current, let Based on the ideal steady-state temperature rise response amplitude and thermal time constant, the load transition component is calculated: In the formula, The physical polarity attribute used to characterize load current changes; the load transition component is positive during load increase and negative during load decrease. The load transition component is determined solely by load current changes, historical steady-state consistency calibration results, dynamic equivalent thermal resistance, and thermal time constant; the difference between the current real-time temperature measurement and the predicted temperature is not used as input. Therefore, additional temperature rises caused by contact defects or other abnormal factors in the current measurement results will not be included in the normal load transient model. The transient thermal response compensation function at the segmentation point... Abrupt changes in the derivative may occur. To reduce the resulting numerical glitches, the program sets the time domain depth to [value missing]. A first-in-first-out circular buffer.

[0048] When the calculation time of the transient thermal response compensation function is close to the segmentation point, the function output samples before and after the segmentation point are written into the buffer in chronological order. (Cumulative) After sampling for one period and filling the buffer, all samples are read. After accumulating a preset number of sampling points and filling the buffer, the function output samples before and after the critical point are read, and a cubic Hermite interpolation function is constructed based on the function values ​​at both ends of the critical interval and the first derivative value. The modified transient thermal response compensation function is used for load switching component calculation. The output of the modified transient thermal response compensation function is used for load switching component calculation. This embodiment will... Set as Each sampling period, making It achieves smooth transitions in the critical region and maintains first-order continuity throughout the entire time range.

[0049] This smoothing process is performed only when the non-steady-state flag is valid, and is mutually exclusive with the aforementioned steady-state screening conditions. Therefore, dynamic power components will not be repeatedly subtracted in the steady-state interval, which can also reduce the interference of transient load changes on defect heating identification.

[0050] Furthermore, determining the cabinet door disturbance component based on the difference in thermal response before and after the cabinet door state change includes: monitoring the cabinet door state of the distribution cabinet; when the cabinet door state changes from closed to open, collecting the ambient temperature parameters inside and outside the distribution cabinet; calculating the signed temperature difference between the ambient temperature outside the distribution cabinet and the internal air temperature, and calculating the environmental convective heat exchange value caused by the signed temperature difference based on the convective heat dissipation law; determining the physical transfer direction of heat loss and backflow based on the positive and negative polarity of the signed temperature difference, and equivalently mapping the environmental convective heat exchange value to the cabinet door disturbance component of the test part corresponding to the physical transfer direction.

[0051] Specifically, limit switches installed along the edge of the distribution cabinet frame are used to monitor the cabinet door status variables in real time. When identified When a transition occurs and satisfies the characteristics of changing from closed to open, this moment is recorded as the initial moment of door opening. Simultaneously, the first ambient temperature sensor deployed in the power distribution room environment is used to acquire external ambient temperature parameters. Read the internal air temperature measured by the second ambient temperature sensor installed above the inside of the distribution cabinet. The signed temperature difference is calculated using subtraction algebraic operations. : When the cabinet door is opened, the three-dimensional air mixing alters the temperature field inside the cabinet. To describe this process, a transient convective heat transfer attenuation model is used; the current time is calculated. Duration since the door opened Introducing a transient nonlinear attenuation factor for door opening, the time-varying convective heat transfer coefficient under the current door opening condition is calculated. : In the formula, This is the reference convective heat transfer coefficient when the door is open. It represents the reference convective heat transfer intensity under the air exchange state inside and outside the distribution cabinet after the cabinet door is opened. It is only used in the cabinet door disturbance model and is not used as a normalized reference value for the natural convection compensation coefficient of the closed cabinet. This is a dimensionless door opening disturbance enhancement coefficient, used to represent the degree of enhancement in the convection of hot and cold air at the instant the door is opened. This represents the duration of the current moment relative to the initial moment the door opened. To reconstruct the mixing time constant for space air, This refers to the standard temperature rise design value inside the cabinet; the standard temperature rise design value inside the cabinet. Defined as the nominal temperature difference between the internal air and the external environment of a distribution cabinet when it reaches steady-state thermal equilibrium under full-load rated operating conditions, this reference temperature is written as an initialization constant into the system configuration database. It is retrieved from a pre-stored cabinet type-temperature rise lookup table in the system configuration database based on the rated capacity and insulation class of the distribution cabinet. Different rated capacities or insulation classes correspond to different... The value used in this embodiment is the nominal cabinet type under full load rated operating conditions. As the denominator of the dimensionless temperature difference term; the convective heat transfer coefficient of the door opening reference. The selectable range is to In this embodiment, ; The selectable range is to In this embodiment, Space air reconfiguration mixing time constant The time required for the cabinet opening to reach convection equilibrium with the indoor air is used to indicate the specific adaptive correction rule: based on a preset baseline mixing time. Based on this, multiplied by a flow field disturbance correction factor determined by the instantaneous temperature difference at the current door opening, the air reconfiguration evolution process accelerates as the ambient temperature difference increases. The dynamic correction calculation formula is expressed as follows: In the formula, The reference mixing time constant is set to a value of , The temperature difference flow field sensitivity factor has a value of [value missing]. ; different temperature differences marked Under real-time evolution, the space air reconfiguration mixing time constant Based on the above formula, the selectable range is: to This allows the transient decay model to be applied to different temperature difference boundaries. Based on the integral form of Newton's law of cooling, the environmental convective heat exchange value induced by the signed temperature difference is calculated. Algebraic calculation formula: In the formula, This refers to the effective heat dissipation cross-sectional area of ​​the opening after the distribution cabinet door is fully opened; effective heat dissipation cross-sectional area of ​​the opening. As a physical structural constant bound to the current distribution cabinet model, it is pre-stored in the system configuration database for real-time calculation and retrieval. In this embodiment, the value is taken according to the standard single-door specification. To maintain consistent input parameters. Considering that the global environmental convective heat exchange caused by opening the cabinet door is shared by all components inside the distribution cabinet, if this global heat exchange value is... Mapping this energy to a single measurement site will result in significant overcompensation. Therefore, a local energy sharing coefficient is introduced. ; Obtain the effective exposed surface area of ​​the part to be tested. The total effective heat exchange surface area of ​​all components and inner walls inside the distribution cabinet The algebraic formula for calculating the local energy sharing coefficient based on area proportion is as follows: In the formula, The preset minimum overflow prevention constant (value) To avoid the risk of division by zero; and As physical structural constants bound to the internal three-dimensional topology of the distribution cabinet, they are pre-stored in the system configuration database.

[0052] The local energy sharing coefficient is related to the cabinet compartment structure, internal partitions, component layout, local air ducts, and the location of the part under test. The area ratio is only an initial estimate and needs to be calibrated through local temperature control tests under cabinet door opening and closing conditions. For parts under test that have not completed calibration for the corresponding cabinet type, the area ratio result is not used to directly generate automatic control commands.

[0053] Environmental convective heat exchange value Combined with local energy sharing coefficient The equivalent mapping is the cabinet door disturbance component of the measured part corresponding to the physical transmission direction. Directly multiplying the heat dissipation power by the thermal resistance may result in instantaneous overcompensation. Considering the thermal inertia of the metal busbar and the cabinet door disturbance component... The calculation is performed using a time-domain integral dynamic accumulation model based on the local equivalent heat capacity of the measured part and incorporating an energy dissipation attenuation term. The integral formula is as follows: In the formula, The integral time variable is the duration of the door opening. The local equivalent heat capacity of the part to be measured is given a fixed value in this embodiment. ; The introduced time constant for heat dissipation in the cabinet door boundary layer is used to characterize the nonlinear damping of the metal busbar's self-thermal equilibrium under convective conditions. The selectable range is... to In this embodiment, A time-domain integral dynamic accumulation model with a heat dissipation decay term is developed by introducing a heat dissipation decay term into the integrand. The time-physics decay operator, characterized by [specific feature], ensures that the integral term converges over time in the time domain, avoiding unbounded growth of the integral during prolonged door opening. The time-domain integral dynamic accumulation model directly preserves the signed temperature difference. The true sign of the heat loss and the direction of physical transfer of external heat backflow are reflected in the cabinet door disturbance component. In the numerical changes, there is no need to set multiple sets of mutually exclusive conditions, thus simplifying the calculation process and obtaining the scalar. This indicates the temperature value representing the disturbance component of the cabinet door.

[0054] The continuity of the aforementioned cabinet door state variables is confirmed only if the cabinet door state is continuous. Only when all sampling points remain consistent is the corresponding cabinet door state change confirmed as a valid state transition. In the formula, The number of continuous sampling points required to confirm the cabinet door status is used to reduce misjudgments of status caused by limit switch contact bounce or short-term false triggering.

[0055] When the main controller confirms that the cabinet door has changed from open to closed, it records the corresponding moment as the closing moment. After the cabinet door is closed, the accumulation of heat exchange caused by convection in the open environment is stopped. The cabinet door disturbance component that has not yet dissipated at the moment of closing is used as the initial value for recovery. The cabinet door disturbance component during the recovery period is calculated according to the following formula: In the formula, This refers to the cabinet door disturbance component that has not yet dissipated when the cabinet door is closed. Let be the time constant for heat dissipation at the cabinet door boundary layer. After the cabinet door is closed, the door disturbance component does not immediately become zero, but rather gradually decays over time from its residual value at the moment of closing. When the following conditions are met... Or the door closure recovery time reaches the preset maximum recovery time. The event is considered complete upon determining the end of the cabinet door disturbance event. The cabinet door disturbance component is set to zero, and the corresponding opening and closing times, as well as the integral cache, are cleared. Where, The threshold for terminating cabinet door disturbance recovery is set. The maximum recovery time is preset. If the cabinet door is confirmed to have changed from closed to open again during the door recovery period, the cabinet door disturbance component that has not yet completely decayed is used as the initial state of a new round of cabinet door events, the opening time is re-recorded, and the calculation of the cabinet door disturbance component is restarted. When the cabinet door status signal is missing or does not meet the preset confidence conditions, the current sampled data is marked as a low confidence state, the cabinet door status switch is not triggered based on the sampled data, and the use of the sampled data to update the standard thermal response baseline is suspended. The cabinet door disturbance parameters in this embodiment are only applicable to the calibrated cabinet type, compartment structure, and sensor arrangement conditions. When changing the power distribution cabinet model, internal partition structure, component arrangement, or sensor position, the local energy sharing coefficient, local equivalent heat capacity, and cabinet door boundary layer heat dissipation time constant should be re-determined.

[0056] Through the above calculations, the air mixing and heat dissipation process after the cabinet door is opened is converted into a cabinet door disturbance component that varies with time. After the cabinet door is closed, the remaining cabinet door disturbance component gradually decays according to an exponential law, and the event reset is completed after the recovery termination condition is met. The local equivalent heat capacity and dissipation terms keep the cabinet door disturbance component bounded during the cabinet door opening phase and the door closing recovery phase, which can be used for subsequent equivalent inversion of the closed cabinet.

[0057] Furthermore, the specific process for calculating the closed-cabinet equivalent thermal response value of the tested part of the distribution cabinet includes: acquiring the operating data of the distribution cabinet, inputting the real-time temperature measurement value of the thermal response sensor into the spatial thermal response relationship, and inverting to obtain the basic inversion temperature; based on energy conservation and thermal offset logic, superimposing the basic inversion temperature with the cabinet door disturbance component with physical transmission direction attribute to obtain the environmental correction temperature after excluding boundary thermal disturbance; according to the transient thermal change trend characterized by the positive and negative polarity of the load conversion component, combined with the thermal time constant, performing a symbolic compensation algebra operation on the environmental correction temperature and the load conversion component based on tending towards thermal equilibrium to calculate the closed-cabinet equivalent thermal response value of the tested part.

[0058] Specifically, the first step is to obtain the real-time temperature measurement value at the current sampling time. The scalar operator corresponding to the final spatial thermal response relationship when the aforementioned iterations reach the convergence condition. Perform product inversion, then algebraically compensate and add back the reference temperature to reconstruct the basic inversion temperature of the measured part. : The second step, based on the law of conservation of energy and boundary layer thermal hedging logic, introduces a method that varies with the duration of the door opening. The cabinet door disturbance component in the current moment of dynamic evolution The calculation is performed according to the aforementioned time-domain integral dynamic accumulation model during the cabinet door opening phase, and according to the aforementioned exponential decay model during the recovery phase after the cabinet door is closed. The basic inversion temperature is then calculated. The cabinet door disturbance component at the current moment Subtraction algebraic superposition is performed to offset the dynamic heat loss or heat backflow caused by the unsteady turbulent interaction of external cold and hot air, generating the environmental correction temperature after eliminating boundary thermal disturbances. : The third step is to obtain the load switching components. Extract load conversion components The corresponding physical polarity sign yields the transient thermal change trend, i.e., the trend with the effective duration of the phase transition. A sign-reverse hedging mechanism tending towards a new thermal equilibrium state; among which, the effective duration Defined as the time from the end of the response dead zone mentioned above. up to the current sampling time The effective duration, expressed as Furthermore, the sign-reverse hedging adjustment mechanism performs piecewise control in the time domain: when it is in the response dead zone, i.e. At that time, due to Forced to The final closed-cabinet equivalent thermal response value is automatically equal to the ambient corrected temperature. This ensures the dual continuity of physical values ​​and slopes throughout the entire computational chain at the time-domain boundary where the dead zone and transient decay intersect; and satisfies the condition when activating the transient hedging solution. At that time, combined with the aforementioned inherent thermal time constant With transient thermodynamic equilibrium correction factor By subtracting the time Corrected load switching components The equivalent thermal response value of the closed cabinet at the test location is obtained by calculation. : In the formula, This is a transient thermodynamic equilibrium correction factor used to characterize the degree of thermal capacitive coupling among multiple components. It is pre-calibrated and stored in the system configuration database based on the number of components and their spatial stacking density within the distribution cabinet. It is used to correct the amplitude nonlinear distortion of the transient component derived from a single inherent thermal time constant in a complex multiphase heat transfer system. The preferred value range is [value range missing]. to In this embodiment, the value is taken as... .

[0059] The aforementioned equivalent thermal response value of the closed cabinet By adjusting the ambient temperature With the transient thermodynamic equilibrium correction factor Corrected load switching components Algebraic subtraction is performed to mitigate the impact of normal transient temperature rises caused by sudden power flow changes on monitoring results and avoid repeated compensation of the load transition component, which could lead to distortion of the steady-state thermal response. In this calculation, the absolute value of the load transition component is not taken, nor is its polarity changed; instead, its physical polarity is preserved, and corresponding algebraic calculations are performed based on the temperature rise or fall trend corresponding to the load change. When the load experiences a normal sudden increase or decrease, the load transition component can offset the transient temperature fluctuations caused by power flow changes. When there are contact defects at the busbar connection points and abnormal heating occurs, the continuous positive temperature rise caused by the defect is not eliminated as a normal load disturbance and is therefore retained in the calculation results. The resulting equivalent thermal response value for the closed-circuit unit is... A time-varying temperature scalar that characterizes the temperature state of the part being measured.

[0060] Through the above processing, on the basis of correcting the disturbance caused by cabinet door opening, the transient thermal response caused by load change can be further separated, so that the disturbance caused by cabinet opening and the disturbance caused by load conversion can be compensated respectively, thereby obtaining the thermal response result of the tested part under the equivalent state of cabinet closure, reducing misjudgment caused by environmental changes and dynamic load fluctuations.

[0061] Furthermore, the specific process for obtaining the dynamic anomaly index includes: calculating the difference between the equivalent thermal response value of the closed cabinet at the current monitoring time and the corresponding standard thermal response baseline to obtain the absolute value of the temperature deviation; within a preset time window, summing the discrete sampling deviation data with a difference that is continuously greater than zero using the trapezoidal numerical integration method to obtain the integral area value representing the cumulative amount of continuous heat generation; and normalizing and weighting the absolute value of the temperature deviation and the integral area value to obtain the dynamic anomaly index.

[0062] Specifically, the current monitoring time Calculated equivalent thermal response value of the closed cabinet The aforementioned standard thermal response baseline function is obtained by inputting the current real-time operating load current. The obtained baseline temperature is then subtracted. To preserve the direction of deviation of the closed-cabin equivalent thermal response value from the standard thermal response baseline, a signed diagnostic bias is first defined for the current monitoring time. : Based on this, the absolute value of the signed diagnostic deviation is taken to obtain the absolute value of the temperature deviation: Set the discrete time span of the preset time window as This is used to construct the current monitoring time. The following backtracking time window Time span The setting is based on the principle of smoothing transient temperature disturbances and fully capturing the continuous heat dissipation cycle in the early stage of defects, and its preferred value range is: to In this embodiment, ; Keep window The complete and valid sampling sequence within the window and the original timestamps of each sampling point; for each sampling time within the window Calculate the positive diagnostic bias: The trapezoidal numerical integration method is used to sum all discrete sampling deviation data that satisfy the above constraints along the time axis within the window to obtain the integral area value representing the cumulative amount of continuous heat generation. : In the formula, and windows respectively The first in chronological order The effective discrete sampling time and the first There are two valid discrete sampling times, which are adjacent sampling times in the complete valid sampling sequence; For discrete time step index, and ; For window Total number of valid discrete sampling points. For signed diagnostic bias... At the sampling time, its positive diagnostic deviation Pick However, it is not removed from the time series. If there are intervals of missing measurements, abnormal timestamps, or unqualified sensing quality between adjacent sampling times, the corresponding trapezoidal integral term is not included in the integration area value. When performing trapezoidal numerical integration, the actual time-domain interval between two adjacent valid sampling times is used. When performing area accumulation, the time interval between discontinuous sampling moments is no longer forcibly replaced with the standard sampling period. .like and If there are sampling intervals with missing measurements, abnormal timestamps, or substandard sensor quality, the corresponding trapezoidal integral term will not be included in the integral area value. The accumulation is performed, and the current window is marked as a low-confidence state, thereby avoiding connecting integrals across invalid time intervals, making... Reflects a continuous positive heating trend over the entire timeline. Sets the maximum limit for temperature deviation. With the maximum limit of the integral area As a normalization benchmark; maximum limit of temperature deviation The range is determined based on the maximum allowable temperature rise corresponding to the equipment's insulation heat resistance class, and the selectable range is as follows: to In this embodiment, Maximum limit value of integral area The maximum allowable temperature rise is determined based on the integral upper limit corresponding to a preset time window, and calculated according to the following formula: The maximum normalization operator was used to evaluate the absolute value of the temperature deviation. and integral area value Normalization and amplitude limiting are performed to obtain the dynamic anomaly index. : In the formula, The instantaneous deviation weighting coefficient is used. The cumulative energy weighting coefficients satisfy the constraints. In order to enable the final diagnostic indicators to identify minor degradation defects, the weighting coefficients... and The adaptive configuration rule is set as follows: based on the aforementioned signed diagnostic bias Adjust the time-domain rate of change; When the rate of change of the signed diagnostic bias is detected to increase significantly and exceed the threshold of the derivative's infinitesimal change, the weight of the instantaneous term is increased, making... This is to improve the sensitivity of the dynamic anomaly index to the risk of instantaneous overheating. The optimal range for both weights is [missing value]. to In this embodiment, the initial value is determined based on the historical average operating conditions. The calculation results represent a dimensionless index of the degree of anomaly. In the calculation of the dynamic anomaly index, trapezoidal numerical integration is used to accumulate signed diagnostic deviations that are continuously greater than zero over a certain time range. Therefore, anomaly judgment does not only use the temperature rise at a single moment, but also considers the instantaneous temperature rise and the heat accumulation over a period of time.

[0063] This can reduce the impact of instantaneous sensor noise on the judgment results, reduce the possibility of false alarms caused by single data mutations, and improve the ability to identify early contact defects and minor deterioration defects.

[0064] Further, see Figure 2 The process of judging the operating status and dynamically adjusting the monitoring cycle includes: setting a preset abnormal threshold, an initial monitoring cycle, and a minimum safe monitoring cycle; if the dynamic abnormal index does not reach the preset abnormal threshold, the operating status is determined to be normal, and the monitoring cycle is gradually extended to the initial monitoring cycle; if the dynamic abnormal index reaches the preset abnormal threshold, the operating status is determined to be suspicious, and the monitoring cycle is shortened; if the monitoring cycle is shortened to the minimum safe monitoring cycle, the operating status is determined to be abnormal, an alarm signal is issued, a load dispatching command is output, and the load is reduced; if the dynamic abnormal index still reaches the preset abnormal threshold, a trip linkage signal is output.

[0065] Specifically, a preset exception threshold is set in the system memory. Initial monitoring period With minimum safety monitoring cycle Initial monitoring period The setting is based on meeting the daily inspection frequency of the equipment while taking into account the processor's computing power load, and its preferred value range is: to In this embodiment, Minimum safety monitoring cycle The setting is based on the shortest thermal equilibrium transition time during the matching flow degradation stage, and its preferred value range is: to In this embodiment, ; Obtain the actual monitoring cycle being executed at the current moment. The dynamic anomaly index obtained from the above calculation is applied using a conditional branch state machine. Perform numerical logic judgment: if the dynamic anomaly index Preset abnormal threshold If the condition is met, the power distribution cabinet is determined to be operating normally; a preset abnormal threshold is set. The setting is based on filtering out minor environmental noise fluctuations and anchoring the quantitative critical point of early equipment degradation; its preferred value range is... to In this embodiment, If the condition is determined to be normal, the reset mechanism for the current monitoring cycle is triggered, and the current monitoring cycle is determined by subtraction. Is it less than the initial monitoring period? ;like If the current monitoring period is extended, the monitoring period will be extended step by step using a step-by-step addition formula. The next monitoring period will then be adjusted accordingly. Represented as: In the formula, Preset delay step size; Preset delay step size The setting is based on achieving smooth recovery of the cycle and preventing missed detections due to state rebound; its preferred value range is... to In this embodiment, If the current monitoring period Not less than the initial monitoring period Then the cycle extension operation will not be executed and will be... Then proceed directly to the next monitoring cycle; if the dynamic abnormal index Preset abnormal threshold If the operating status of the distribution cabinet is deemed suspicious, the dynamic cycle shortening mechanism will be automatically triggered, and a dynamic anomaly index will be constructed. Add an adaptive adjustment function that decreases exponentially: In the formula, The periodic decay adjustment constant; the periodic decay adjustment constant The setting is based on the response sensitivity of the monitoring frequency as the fault worsens, and its preferred value range is: to In this embodiment, ; after calculating the adjusted next monitoring period Then, the next monitoring cycle is further determined through conditional branching. Should the minimum safety monitoring cycle be shortened? If the next monitoring cycle Minimum safety monitoring cycle If the current operating state has not completely deteriorated, the state machine will directly control the entry into the next monitoring cycle.

[0066] When determining the next monitoring period within a continuous dynamic sampling period Decreasing closed-loop convergence to minimum safety monitoring period Furthermore, the results measured at the current moment still satisfy the dynamic anomaly index. Preset abnormal threshold If the condition is abnormal, the distribution cabinet's operating status is determined to be abnormal; at this time, a pulse alarm signal is sent to the distribution network master station dispatch center, and the pre-stored node heat-load sensitivity matrix is ​​retrieved. ; Nodal heat-load sensitivity matrix To characterize the distribution of the target total load reduction current among the smart branch circuit breakers, the node thermal-load sensitivity matrix is ​​non-negatively normalized before generating the target load reduction vector, and its dimension is strictly set to [value missing]. In mathematical terms, it degenerates and is equivalent to a column vector, where Let be the number of intelligent branch circuit breakers under this distribution network node, where each element is non-negative and the sum of the elements is . The node heat-load sensitivity matrix corresponds to the current distribution node topology. It is obtained through offline power flow sensitivity calculation and pre-stored in the system configuration database. The main controller calls the node heat-load sensitivity matrix to solve the dynamic load dispatching command for the node where the distribution cabinet is located. The dynamic load dispatching command includes the target load reduction vector for each intelligent branch circuit breaker. ; Obtain the total load current of the current distribution network node Calculate the target total load descent current based on the current dynamic anomaly index: Then, based on the node heat-load sensitivity matrix, the target total load reduction current is distributed to each intelligent branch circuit breaker to obtain the target load reduction vector: In the formula, It is the load control gain coefficient, used to quantify the linear mapping strength of thermal overload risk to the reduction in electricity load. Its preferred value range is... to In this embodiment, 0 is taken. . This is a preset maximum allowable power reduction and load balancing constraint value, and its preferred range is: to In this embodiment, .

[0067] After all security checks pass, the main controller determines the target load reduction vector. The system controls each intelligent branch circuit breaker connected to the distribution cabinet to reduce the load limit of the corresponding outgoing lines, thereby completing the load reduction operation. After load reduction, the integral area value... The historical sliding back window used The system may still retain deviation data from before the load reduction. If the original window is used for integration, the accumulated anomalies before the load reduction will still affect subsequent judgments, potentially causing false trips. Therefore, this embodiment is configured such that when the main controller sends a pulse alarm signal to the distribution network master station and initiates load reduction operation, the backtracking window will be used. The start time is updated to the current unloading time, while historical deviation data stored before that time is cleared. Thus, within the preset observation duration... Inside, the integral area value Trapezoidal integration is performed only based on newly sampled data after load reduction, ensuring the dynamic anomaly index reflects the actual cooling status of the equipment after load reduction. The load reduction operation duration reaches the preset observation period. Then, the time rate of change of the current dynamic anomaly index is calculated using the discrete difference method. Then, the load reduction effect is evaluated based on the set logical judgment conditions. When the conditions are met... And the current dynamic anomaly index Preset abnormal threshold When the dynamic anomaly index does not show a downward trend and remains above the anomaly threshold, the equipment is determined to be in an abnormal state. After confirming the protection interlock status and communication effectiveness, a trip linkage signal is sent to the existing protection execution chain of the distribution network to disconnect the corresponding load. This method does not change the original relay protection settings and hardware interlock relationships. When the following conditions are met... When the current dynamic anomaly index is determined to be declining, indicating the equipment is in the heat convection cooling phase after load reduction, the state machine releases the trip lockout, and the control continues to maintain the current load reduction operation state, entering the next monitoring cycle. For cases where the judgment result is normal, in a suspicious state but has not yet reached the minimum cycle threshold, or in the normal cooling phase after load reduction, the dynamic anomaly index is recalculated after entering the next monitoring cycle, and the new calculation result is input into the conditional branch state machine again to execute the next round of dynamic anomaly threshold comparison and state judgment. Preset observation duration. Used to wait for the thermal convection cooling response of metal components after unloading, the selectable range is: to In this embodiment, we take .

[0068] After the above processing, status awareness, monitoring cycle adjustment, and active control are linked together. The distribution cabinet no longer only undergoes inspections at fixed intervals, but adjusts the monitoring cycle length and takes corresponding controls based on the actual status to reduce the possibility of node failures continuing to develop.

[0069] This implementation method corrects for thermal disturbances caused by cabinet door opening and closing and load changes, while simultaneously examining the current temperature rise and heat accumulation over a set time period. This yields the equivalent thermal response when the cabinet is closed and determines the equipment status. Adjusting the next monitoring cycle based on the assessment results reduces false alarms caused by environmental or load fluctuations and improves the reliability of contact defect identification and anomaly monitoring.

[0070] Example 2: Example 2 follows the spatial thermal response relationship, transient load component separation method, and closed-cabinet equivalent inversion process from Example 1. Building upon this, it further explains the slow-varying contact resistance tracking and parameter updating, cross-radiation separation of adjacent heat sources, and the remaining time prediction and smooth transition control based on dynamic thermal margin.

[0071] When determining the load switching components, the equivalent AC / DC resistance of the measured part is also tracked over a long period. Specifically, a long-cycle evolution time window is established, the temperature rise drift residual of the equivalent thermal response value of the closed cabinet relative to the standard thermal response baseline is extracted, and the root mean square value of the effective load current is obtained; combined with Joule's law of heating and dynamic equivalent thermal resistance, the dynamic resistance increment is calculated in reverse, and the equivalent AC / DC resistance is corrected according to the dynamic resistance increment to obtain the current actual contact resistance used in the next long-cycle stage.

[0072] Specifically, busbar contacts and connections are subjected to long-term current and thermal stress, and surface oxidation and mechanical loosening can cause the contact resistance to increase slowly. If the transient load component always uses the initial resistance, the calculation reference will gradually deviate from the actual state over time, so this parameter needs to be updated on a long-term basis.

[0073] To mitigate the aforementioned impacts, the main controller is configured with a long-cycle evolution time window. The time span is selectable. Heavenly In this embodiment, the day is taken Day. Within this window, extract each sampling time. Corresponding equivalent thermal response value of closed cabinet Using the aforementioned formula The calculations were performed, and corrections have been made for spatial path loss, cabinet door opening disturbances, and thermal response caused by transient power flow changes. This was based on the load current at each sampling time. Determine the corresponding target value from the standard thermal response baseline. .calculate Equivalent thermal response value of enclosed cabinet With target value The average difference between them yields the temperature rise drift residual. : In the formula, To exclude transient fluctuations, the total number of valid steady-state sampling points is determined only if the sampling points simultaneously satisfy the first-order time-domain derivative of the load current. Furthermore, data is included when the duration exceeds the preset thermal delay hysteresis window to reduce the impact of sudden power flow changes on slow aging assessment. Simultaneously, the root mean square value of the load current within the long-period evolution time window is obtained. And the historical average value of the aforementioned dynamic equivalent thermal resistance within this window. Based on the Joule heating model and Ohm's law of heat transfer, the dynamic resistance increment is calculated inversely. Based on this, a slowly varying heat power relationship is established, and the formula is: Considering the possibility of occasional anomalies and outliers within a single long-period window, an exponential moving average method is used to update the contact resistance. Let... This is the equivalent AC / DC resistance that solidified at the end of the previous long cycle. If the equivalent AC / DC resistance reference value is determined during the initial configuration, then it is based on the dynamic resistance increment within the current window. The calculated absolute resistance observation value is Current actual contact resistance for: In the formula, An aging-updated forgetting factor is used to reduce single-assessment errors and parameter mutations caused by outlier data. The selectable range is... to In this embodiment, we take After the update is complete, the main controller will... Write to the system configuration database and replace the original fixed parameters with it. Subsequent steady-state thermal power increment The load transfer components are calculated using the updated contact resistance.

[0074] To avoid the repeated accumulation of trend values ​​across multiple long-period evolution time windows, the main controller only saves the contact state change trend formed in the previous window and uses this as a historical reference for comparing the current window trend. This contact state change trend is not written into the standard thermal response baseline function. It also does not directly replace the initial equivalent resistance parameters in the normal load transient model.

[0075] When the contact state change trend formed by adjacent long-cycle evolution time windows changes significantly, the main controller only outputs a trend change marker and keeps the current standard thermal response baseline and normal load transient model parameters unchanged to avoid jumps in the normal model output caused by contact resistance estimation results that have not been independently verified. Using the above window-by-window update method, the continuous change of resistance parameters between adjacent long-cycle evolution time windows can be maintained, avoiding numerical deviations caused by repeated calculations of historical increments. It can also be used to indicate that the contact state may undergo long-term, slow changes, providing a reference for subsequent manual verification, power outage detection, or independent voltage drop measurement.

[0076] Furthermore, in the specific process of calculating the equivalent thermal response value of the closed cabinet of the distribution cabinet under test, the method also includes a decoupling and stripping step for cross-radiative coupling interference from adjacent multiple heat sources: obtaining the heat source entities of the adjacent functional areas of the under test in the three-dimensional spatial topology as adjacent interference heat sources; calculating the transient thermal driving potential of the surface based on the real-time load current of the adjacent interference heat sources and the corresponding standard thermal response baseline; constructing a multi-dimensional cross-radiative decoupling matrix by combining the three-dimensional relative distance between the adjacent interference heat sources and the under test and the effective radiation field of view; calculating the cross-thermal radiation crosstalk components of each adjacent interference heat source of the under test, and performing algebraic offset stripping from the basic inversion temperature to obtain the pure inversion temperature that excludes internal topological thermal crosstalk, and using the pure inversion temperature to replace the original basic inversion temperature in subsequent environmental correction and equivalent inversion calculations.

[0077] The internal space of the compartmentalized distribution cabinet is relatively small, and the A, B, and C phase busbars and the contacts of the upstream and downstream switches are close together. When a phase is under high load, heat will be transferred to adjacent functional areas through near-field radiation and local micro-convection, causing the temperature measured by adjacent sensors to be higher.

[0078] To eliminate the aforementioned spatial thermal crosstalk, the main controller enables multi-heat source cross-coupling calculation. Based on the three-dimensional topology model of the distribution cabinet, it searches for the location to be measured. The spatial straight-line distance is less than the preset crosstalk radius of One nearby interfering heat source, among which Index for nearby interfering heat sources; preset crosstalk radius Based on the near-field thermal radiation attenuation boundary and phase-to-phase physical insulation spacing within the enclosed space of the distribution cabinet, this setting is used to cover the effective thermal impact range and reduce invalid calculations for distant, weakly coupled heat sources. The selectable range is... to In this embodiment, we take ; Obtain the neighboring interfering heat sources Real-time load current at the current moment Substituting this into the pre-established standard thermal response baseline function for each heat source location, we obtain its theoretical baseline temperature. Extract the air temperature inside the distribution cabinet at this time. Calculate the transient thermal driving potential of each neighboring interfering heat source at the current moment. : use The function truncates negative values ​​so that effective cross-radiation effects only occur when the temperature of a nearby interfering heat source is higher than the ambient air temperature; a multidimensional cross-radiation decoupling matrix is ​​constructed. Its matrix elements Defined as a nearby interfering heat source The part to be tested Spatial thermal coupling weighting coefficients; combined with three-dimensional relative distance and effective radiation field of view, The formula for calculation is: In the formula, For the minimum overflow prevention constant (in this embodiment, it is taken as...) To avoid the risk of division by zero when coordinates overlap, heat source With the part to be tested The straight-line span of the space between them; heat source The effective radiation projection area facing the part to be measured; Connect the center of the two and the heat source The relative field of view between surface normals; Let be the internal radiation transfer gain constant. The absorption coefficient of air medium; the internal radiative transfer gain constant. and This parameter represents the combined effect of infrared radiation attenuation and wall reflection within a sealed cavity and is pre-stored as a fixed physical parameter in the system configuration database. Offline calibration is performed based on standard air medium characteristics and the reflectivity of the metal cavity. Set as , Values The three-dimensional spatial angle coefficients are determined using the constraint matrix described above.

[0079] Non-negativity and upper bound constraints are applied to the spatial thermal coupling weight coefficients. If the original weight calculated based on the effective radiation field of view is less than zero, it is truncated to zero; if the original weight exceeds the preset single heat source coupling upper limit, it is truncated to the upper limit. For multiple adjacent interfering heat sources corresponding to the same test location, each weight is normalized or the total amount is limited so that the sum of the coupling weights of all adjacent interfering heat sources does not exceed the preset total coupling upper limit. And satisfy: ,in, Determined by real multi-heat source experiments.

[0080] Furthermore, considering the time lag of cross-phase heat transfer, a spatial heat transfer lag filter is introduced to calculate the cross-thermal radiation crosstalk components of each neighboring interfering heat source at the current moment for the part under test. : In the formula, The equivalent time constant for interphase heat transfer is obtained by pre-establishing a spatial linear span. A two-dimensional thermal inertia mapping table is generated, with the composite thermophysical properties of the air medium inside the cavity as the first dimension input and the specific heat capacity and equivalent thermal conductivity as the second dimension input. This table is stored in the non-volatile memory of the main controller. During real-time calculations, the main controller utilizes the currently retrieved spatial linear span... The equivalent time constant is obtained by retrieving the specific heat capacity of the configured medium from a two-dimensional thermal inertia mapping table using a bilinear interpolation algorithm. , The duration during which the heat source is in a high-current state; duration The current over-limit comparator in the control program accumulates the current over time, and its specific judgment rule is: the main controller monitors the load current of each nearby interfering heat source in real time. If and only if satisfying Automatically activate for this heat source. Set the high current flag and start a timer to accumulate the current continuous monitoring duration on the time axis. If the load current drops below the threshold value, the control program will put the high current flag into a cooling buffer state, at which point the timer will... Stop accumulating upwards and do not immediately clear to zero, but instead increment according to the current real-time sampling time. The advancement is based on the equivalent time constant of transphase heat transfer. The matched discrete negative exponential decay rate decreases with increasing step size until... After gradually decreasing to zero, the state machine is then reset to zero. This reflects the dynamic lag and gradual change characteristics of the heat conduction and dissipation process of metal materials across phases on the time axis, which can reduce the sudden changes in crosstalk components caused by current fluctuations and the resulting false alarms.

[0081] Among them, the high current carrying threshold Used to identify load transitions that produce significant near-field thermal radiation, the selectable range is the corresponding node rated current. Doubled In this embodiment, the fixed configuration is the rated current of the corresponding node. This ensures that the heat transfer time-delay filter maintains a clear causal relationship in real-time calculations; the spatial thermal response scalar operator obtained after iterative convergence is described above. Perform a product inversion on the real-time measured temperature of the measured part to obtain the aforementioned real-time measurement base inversion temperature. After that, the main controller will As the minuend, subtract the currently calculated cross-thermal radiation crosstalk component from the data stream. Thus, the de-crosstalk inversion temperature, which excludes internal topological thermal crosstalk, can be calculated. , can be represented as: Subsequently, the crosstalk inversion temperature will be retrieved. Write back to the main controller's operating data stream to replace the original basic inversion temperature. It also participates in the aforementioned calculation of the first-order thermal inertia transition difference and the compensation calculation of the time-domain integral of the cabinet door disturbance component.

[0082] This step combines the thermal state analysis of the distribution cabinet with the topological relationships between distribution network nodes. By pre-calibrating the influence coefficients of adjacent heat sources and combining the calibration results with the judgment results of the current over-limit state machine, it is possible to distinguish between the temperature rise caused by the self-heating of the tested part and the radiative heat transfer from adjacent high-load parts. This can reduce false alarms caused by mutual interference between multiple heat sources under operating conditions such as three-phase load imbalance, and can be used to reduce the impact of temperature rise crosstalk caused by adjacent high-load parts on the judgment of the state of the tested part.

[0083] Furthermore, when judging the operating status of the distribution cabinet and dynamically adjusting the monitoring cycle, the process also includes the steps of calculating the remaining safe time window based on dynamic thermal margin and smooth transfer control: obtaining the temporal evolution characteristics of the current dynamic anomaly index and calculating the anomaly deterioration rate and deterioration acceleration; constructing a temporal sliding extrapolation prediction model based on the anomaly deterioration rate and anomaly deterioration acceleration to calculate the remaining safe time window for the distribution cabinet to reach the preset thermal risk boundary; activating the smooth transfer mechanism when the remaining safe time window is less than the preset scheduling preparation delay; dynamically calculating the load transfer slope based on the remaining safe time window and transforming the original target load reduction vector into a smooth transfer load flow that changes continuously with time, coordinating with adjacent nodes in the distribution network with healthy thermal margins to perform non-sensory load takeover, and completing the flexible reconfiguration of the distribution network topology before reaching the thermal limit.

[0084] Specifically, in modern flexible distribution networks, traditional hard disconnection strategies are prone to causing secondary impacts on power flow; in order to achieve seamless self-healing of potential hazards in distribution cabinets, the main controller calculates the aforementioned dynamic anomaly index. Then, a predictive remaining safety time window calculation model is introduced; firstly, a one-dimensional discrete-time observation window is constructed to eliminate the problems caused by the monitoring cycle in the aforementioned operation status determination. The dynamic reduction leads to non-uniformly spaced discrete-time axes, causing order-of-magnitude distortion in differential calculations, which poses an engineering hazard. To address this, the main controller sends historical dynamic anomaly index data streams to a resampling hardware accelerator at a fixed standard time-domain resolution. Homogeneous linear interpolation is performed to construct a historical observation queue with equal step size; the current time is obtained using a second-order backward difference algorithm. The first time-domain derivative of the dynamic anomaly index in the homogeneous historical observation queue (Characterizing the rate of abnormal deterioration) and the second-order time-domain derivative (Characterizing the acceleration of abnormal deterioration): In the formula, This represents the fixed time interval between adjacent resampling times. Linear interpolation is only used to unify the time base for difference calculations and eliminate numerical distortion caused by non-equidistant sampling due to dynamic reduction of the monitoring period, rather than to improve the actual observation resolution of the anomaly index; when the time interval between two adjacent real monitoring periods is greater than the preset information window, the current time... The estimation results are marked as low confidence, at which point the remaining safe time window... The computation preferentially adopts only-dependent The first-order linear extrapolation model avoids the extrapolation bias introduced by the second-order extrapolation in the low sampling density range.

[0085] The system configuration database pre-stores the preset thermal risk boundary index of the power distribution cabinet. The index This is used to indicate the dynamic anomaly index limit when a distribution cabinet reaches an insulation failure or risk state. In this embodiment, Values And higher than the aforementioned preset abnormal threshold .

[0086] After obtaining the abnormal deterioration rate and the abnormal deterioration acceleration, the abnormal deterioration rate is used to determine whether the abnormality will continue to develop. When the abnormal deterioration rate satisfies... At that time, a second-order Taylor expansion time-series sliding extrapolation prediction model is used to extrapolate the changing trend of the dynamic anomaly index over time: Let the extrapolated dynamic anomaly index be equal to the preset thermal risk boundary index, and solve for the relationship between the two indices. The equation is given. The positive real roots obtained represent the time required for the distribution cabinet to evolve from its current state to the preset thermal risk boundary, assuming the current trend of change remains unchanged. When the abnormal deterioration acceleration satisfies... When the equation is in time, a positive real root is selected from the real roots as the remaining safe time window. When abnormal deterioration accelerates When this occurs, it indicates that the anomaly is still developing, but the rate of deterioration is decreasing; at this point, the extrapolation curve opens downwards. Only when the discriminant of the equation is not less than zero and a positive real root exists, the smaller positive real root is selected as the... This is used to determine if there is a short-term risk of the device exceeding its limits; if the current condition is met... Or the discriminant of the equation is less than zero, or in If no positive real root exists, then it is considered that an effective thermal collapse moment cannot be predicted based on the current trend, and let... This approach avoids false triggering of the smoothing transition mechanism due to multiple solutions to equations, lack of effective solutions, or short-term fluctuations in abnormal exponents, while also reducing repeated switching of the control state machine between different states. The system-level scheduling preparation delay required for the distribution network master station to complete topology reconfiguration and switch switching is set. System-level scheduling preparation delay This represents a time constant indicating the delay between master station communication and circuit breaker mechanical action, and is pre-stored in the system configuration database. In this embodiment, it is taken as... When the remaining safe time window calculated in real time meets the requirements... Time (of which) The preset safety redundancy margin represents the scheduling safety margin, and the selectable range is: to In this embodiment, The system determines that the distribution cabinet has a short-term risk of loss of control; at this time, it automatically stops executing the aforementioned step load reduction command and instead activates the smooth transfer mechanism; the main controller determines the risk based on the remaining safe time window. The degree of urgency is used to calculate the load transfer slope vector. Before calculating the load transfer slope, determine the remaining safe time window. and safety margin The size relationship. Only when When, call the following formula for calculating the load transfer slope; when When the above calculation formula is not invoked, the smooth transition process is not initiated, and the alarm state is maintained, the existing protection process or manual handling process of the distribution network is entered.

[0087] In the formula, The target load reduction vector obtained from the aforementioned calculation, The load transfer slope vector is defined for each loaded branch. The main controller generates candidate load transfer schemes based on the remaining safe time window and the target load reduction vector. These candidate load transfer schemes are not directly executed as switching or circuit breaker commands. Before issuing the transfer command, the distribution network master station acquires the real-time network topology, available capacity of each candidate receiving node, node voltage, branch current, protection blocking status, switch position status, and communication status, and performs independent safety checks on the candidate load transfer schemes. Only when all safety check conditions are met is the candidate scheme converted into a smooth transfer command for each control channel.

[0088] The main controller uses the pre-stored topology matrix to transfer the load transfer slope vector. Each component in the data is mapped to the corresponding adjacent healthy distribution node takeover channel, generating each channel over time. Continuously increasing smooth transfer load flow component command Its function mapping is expressed as: In the formula, This is the topology channel index for the assigned adjacent healthy nodes. The corresponding load transfer slope component is represented by a scalar, used to ensure that the scalar and vector are consistent in the channel dimension. The trigger time for the smooth transfer mechanism is determined; during the execution of the smooth transfer command, the cumulative transfer current of each control channel is compared. The target load reduction current of the corresponding branch When the accumulated transfer current reaches the corresponding target load reduction current, the transfer command for that channel is stopped, and the subsequent output of that channel is kept at the command value corresponding to the target load reduction current. The accumulated transfer current must not exceed the corresponding target load reduction current. If insufficient receiving node capacity, node voltage exceeding limit, protection lockout, abnormal switch status, or communication failure is detected during the transfer process, the current transfer process is stopped, and the existing protection process or manual handling process of the distribution network is entered. Through the distribution network master station, the transferred load, which varies according to the slope, is smoothly distributed to adjacent nodes, and its own dynamic anomaly index is much lower than that of adjacent nodes. Take over the healthy power distribution nodes.

[0089] This process connects the assessment of the thermal status within the distribution cabinet with the power flow scheduling of the distribution network. The dynamic anomaly index is first resampled at uniform time intervals, and then its rate of change is used to determine the anomaly development trend, thereby determining the load transfer speed. This allows for measures to be taken before the thermal risk continues to escalate, and also reduces the power flow impact caused by directly cutting off the load.

[0090] This embodiment tracks the long-term, slow changes in contact resistance, updates calculation parameters, and reduces benchmark deviations caused by contact aging. Radiative heat generated by adjacent high-current-carrying components is calculated and subtracted using a cross-radiation decoupling matrix. The remaining time before reaching the thermal failure boundary is estimated based on the dynamic thermal margin, and the load is gradually transferred. This improves the reliability of identifying abnormal heating and contact defects under long-term operating conditions, while reducing the risk of thermal damage to the distribution cabinet and mitigating the impact of sudden load changes on continuous power supply and distribution network stability.

Claims

1. A method for periodically monitoring the operating status of power distribution equipment, characterized in that, include: Acquire spatial location information, physical structure information and historical operating data of the thermal response sensors inside the power distribution cabinet and the parts under test, and establish spatial thermal response relationship; Based on historical operating data under closed cabinet conditions, phases of different load cycles are extracted. The standard thermal response baseline of the test part is obtained by combining the spatial thermal response relationship. The thermal inertial transition relationship is established based on the thermal response changes during the phase transition between adjacent load cycles. Acquire the operating data of the distribution cabinet, identify the current load cycle phase, phase transition state and cabinet door state, determine the load transition component by combining the thermal inertia transition relationship, and determine the cabinet door disturbance component based on the difference in thermal response before and after the cabinet door state change. Based on the relationship between cabinet door disturbance component and spatial thermal response, the operating data of the distribution cabinet is used to perform closed-cabinet equivalent correction and inversion, and the load conversion component is filtered out by combining the direction of current change to obtain the closed-cabinet equivalent thermal response value of the part to be tested. The dynamic anomaly index is obtained by measuring the deviation between the equivalent thermal response value of the closed cabinet and the corresponding standard thermal response baseline. The operating status is determined based on the dynamic anomaly index, and the next monitoring cycle of the thermal response sensor is adjusted. When the preset alarm conditions are met, an alarm is issued and a load dispatching command is generated.

2. The method for periodically monitoring the operating status of power distribution equipment according to claim 1, characterized in that, The process of establishing the spatial thermal response relationship includes: acquiring the linear spatial distance and relative radiation angle between the measured part and the thermal response sensor, as well as the environmental parameters inside the distribution cabinet; calculating the thermal radiation compensation coefficient based on the linear spatial distance and relative radiation angle, combined with the thermal radiation attenuation law; calculating the convective heat transfer compensation coefficient based on the physical structure information and environmental parameters inside the distribution cabinet, combined with the natural convection heat transfer model; and constructing a weighted calculation model based on the linear spatial distance, relative radiation angle, and environmental parameters, i.e., the spatial thermal response relationship, by combining the thermal radiation compensation coefficient and the convective heat transfer compensation coefficient.

3. The method for periodic monitoring of the operating status of power distribution equipment according to claim 1, characterized in that, The process of extracting different load cycle phases and obtaining a standard thermal response baseline includes: collecting load current and temperature data of the distribution cabinet in historical operating cycles, and performing outlier removal and low-frequency long-term drift data stripping to obtain historical operating data; using a clustering algorithm to divide the load current data in the historical operating data into multiple steady-state current intervals, and mapping each steady-state current interval to a load cycle phase; when the cabinet door is closed and the internal heat distribution reaches equilibrium, extracting the inversion temperature of the test part under each steady-state current interval as the steady-state temperature of the data reference point; performing curve fitting on each data reference point to establish a standard thermal response baseline with load current as input and steady-state temperature as output.

4. The method for periodically monitoring the operating status of power distribution equipment according to claim 3, characterized in that, The process of establishing the thermal inertia transition relationship includes: when a cross-interval jump in the load cycle phase is detected, extracting the time delay characteristics of the temperature tracking current change of the measured part and the initial temperature at the time of the jump; determining the temperature of the standard thermal response baseline corresponding to the target load cycle phase after the jump, as well as the thermal time constant; constructing a first-order thermal inertia transition relationship with the standard thermal response baseline temperature as the steady-state target, and with the signed difference between the initial temperature and the steady-state target decaying exponentially with time.

5. The method for periodic monitoring of the operating status of power distribution equipment according to claim 4, characterized in that, The process of determining the load transition component by combining the thermal inertia transient relationship includes: obtaining the load current change in the current load cycle phase and calculating the corresponding steady-state thermal power increment; setting the thermal time constant in the thermal inertia transient relationship as the system adjustment parameter, taking the steady-state thermal power increment as the target amplitude input, and constructing a transient thermal response compensation function that approaches the steady state exponentially with time; during the phase transition, calculating the normal load transient temperature change of the measured part at the current moment through the transient thermal response compensation function, taking the direction of increase or decrease of current change as the physical polarity attribute, assigning corresponding positive and negative polarity values ​​to the normal load transient temperature change, and obtaining the load transition component.

6. The method for periodic monitoring of the operating status of power distribution equipment according to claim 1, characterized in that, The method of determining the cabinet door disturbance component based on the difference in thermal response before and after the cabinet door status change includes: monitoring the cabinet door status of the power distribution cabinet; when the cabinet door status changes from closed to open, collecting the ambient temperature parameters inside and outside the power distribution cabinet; calculating the signed temperature difference between the ambient temperature outside the power distribution cabinet and the internal air temperature, and calculating the environmental convective heat exchange value caused by the signed temperature difference based on the convective heat dissipation law; determining the physical transfer direction of heat loss and backflow based on the positive and negative polarity of the signed temperature difference, and equivalently mapping the environmental convective heat exchange value to the cabinet door disturbance component of the test part corresponding to the physical transfer direction.

7. The method for periodically monitoring the operating status of power distribution equipment according to claim 6, characterized in that, The specific process for calculating the closed-circuit equivalent thermal response value of the tested part of the distribution cabinet includes: acquiring the operating data of the distribution cabinet, inputting the real-time temperature measurement values ​​of the thermal response sensors into the spatial thermal response relationship, and inverting to obtain the basic inversion temperature; based on energy conservation and thermal offset logic, superimposing the basic inversion temperature with the cabinet door disturbance component that has physical transmission direction attributes to obtain the environmental correction temperature after excluding boundary thermal disturbances; according to the transient thermal change trend characterized by the positive and negative polarities of the load conversion component, combined with the thermal time constant, performing a symbolic compensation algebra operation on the environmental correction temperature and the load conversion component based on tending towards thermal equilibrium to calculate the closed-circuit equivalent thermal response value of the tested part.

8. The method for periodically monitoring the operating status of power distribution equipment according to claim 1, characterized in that, The specific process for obtaining the dynamic anomaly index includes: calculating the difference between the equivalent thermal response value of the closed cabinet at the current monitoring time and the corresponding standard thermal response baseline to obtain the absolute value of the temperature deviation; within a preset time window, summing the discrete sampling deviation data with a difference that is continuously greater than zero using the trapezoidal numerical integration method to obtain the integral area value representing the cumulative amount of continuous heat generation; and normalizing and weighting the absolute value of the temperature deviation and the integral area value to obtain the dynamic anomaly index.

9. The method for periodic monitoring of the operating status of power distribution equipment according to claim 8, characterized in that, The process of judging the operating status and dynamically adjusting the monitoring cycle includes: setting a preset abnormal threshold, an initial monitoring cycle, and a minimum safe monitoring cycle; if the dynamic abnormal index does not reach the preset abnormal threshold, the operating status is determined to be normal, and the monitoring cycle is gradually extended to the initial monitoring cycle; if the dynamic abnormal index reaches the preset abnormal threshold, the operating status is determined to be suspicious, and the monitoring cycle is shortened; if the monitoring cycle is shortened to the minimum safe monitoring cycle, the operating status is determined to be abnormal, an alarm signal is issued, a load dispatching command is output, and the load is reduced; if the dynamic abnormal index still reaches the preset abnormal threshold, a trip linkage signal is output.

Images