Integrated distribution box state online monitoring method and system based on artificial intelligence

Abstract

The invention belongs to the technical field of power equipment state monitoring and fault diagnosis, and particularly relates to a comprehensive distribution box state online monitoring method and system based on artificial intelligence, and the method comprises the steps: obtaining the terminal temperature of a monitoring point, the load current of a loop, and the environment temperature outside a distribution box, and carrying out the time sequence alignment of collected data; calculating real-time heat dissipation power and load heat shock source intensity of the monitoring points based on the aligned data; based on the real-time heat dissipation power, the load heat shock source intensity and the terminal temperature sequence, the equivalent electric heating impedance modulus of the monitoring point is obtained; and constructing a sliding time window based on the equivalent electrothermal impedance modulus, calculating the contact health degree in the sliding time window, and determining the operation state of the distribution box according to the contact health degree. According to the invention, the interference of the load current and the environment temperature is decoupled by constructing the physical inversion model, the thermal hysteresis phenomenon is effectively overcome, and accurate monitoring of contact aging and loosening faults is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power equipment state monitoring and fault diagnosis. More specifically, the present application relates to an online monitoring method and system for the state of a comprehensive distribution box based on artificial intelligence. BACKGROUND

[0002] As a key node of the low-voltage distribution network, the comprehensive distribution box is exposed to the complex outdoor environment for a long time. The key parts inside the box, such as the circuit breaker contacts and busbar connection points, are prone to oxidation corrosion and bolt loosening caused by mechanical vibration due to long-term operation, which gradually increases the contact resistance. This contact deterioration can cause local abnormal temperature rise, and in severe cases, it can cause equipment burnout and even fire.

[0003] Currently, the state monitoring of the distribution box mainly relies on infrared temperature measurement or pre-embedded temperature sensors. The existing technology usually uses fixed temperature thresholds or temperature rise thresholds for judgment, i.e., when the temperature of the monitoring point exceeds the set value, an alarm is triggered.

[0004] However, this static monitoring method based on fixed thresholds has significant technical limitations. Specifically, the internal temperature of the distribution box is influenced by both the load current and the ambient temperature. In the summer high-load working condition, even if the device connection is normal, the conductor temperature may approach the alarm threshold, leading to frequent false alarms and increasing the workload of the maintenance personnel. On the contrary, in the winter or low-load working condition, even if the connection point has serious aging or loosening, causing the contact resistance to increase, due to the small total amount of Joule heat generated, its absolute temperature may still be lower than the alarm threshold, resulting in system false negatives until the device is completely burned out. In addition, due to the thermal capacity characteristics of metal conductors, there is a significant thermal hysteresis phenomenon in temperature changes relative to current changes, and direct comparison of real-time current and temperature cannot accurately reflect the current contact state.

[0005] Therefore, how to eliminate the interference of the environment and the load, and accurately extract the essential features reflecting the conductor contact state from the thermal hysteresis data, is a problem that needs to be solved in the current online monitoring field of distribution boxes. SUMMARY

[0006] To solve the technical problems of the above-mentioned existing technology being affected by load fluctuations and environmental temperature interference and thermal hysteresis phenomenon in complex working conditions, and being unable to accurately characterize the conductor contact state, resulting in false alarms and false negatives of the monitoring system, the present application provides solutions in the following aspects.

[0007] In a first aspect, the present application provides an online monitoring method for the state of an artificial intelligence-based comprehensive distribution box, comprising: acquiring the terminal temperature of a monitoring point, the load current of the circuit, and the ambient temperature outside the distribution box, performing time alignment processing on the terminal temperature, the load current, and the ambient temperature to obtain the aligned terminal temperature sequence, the load current sequence, and the ambient temperature sequence; calculating the real-time heat dissipation power of the monitoring point based on the aligned terminal temperature sequence and the ambient temperature sequence, and calculating the load heat source intensity of the monitoring point based on the aligned load current sequence; based on the real-time heat dissipation power, the load heat source intensity, and the terminal temperature sequence, obtaining the equivalent electrothermal impedance modulus of the monitoring point; constructing a sliding time window based on the equivalent electrothermal impedance modulus, calculating the contact health degree in the sliding time window, and determining the operating state of the distribution box according to the contact health degree.

[0008] The present application aligns the multi-source data in time sequence, constructs an inversion model containing real-time heat dissipation power and load heat source intensity based on the principle of thermodynamics, calculates the equivalent electrothermal impedance modulus, and then uses the contact health degree of the sliding time window for judgment; this method decouples the interference of load current fluctuation and ambient temperature change on the monitoring result in physical essence, reversely solves the contact state through energy conservation, effectively overcomes the thermal hysteresis effect of temperature change relative to current change; makes the system not only avoid false positives caused by high ambient temperature in summer high load conditions, but also sensitively capture weak abnormal signals caused by increased contact resistance in winter low load conditions, solves the technical problems of false positives in high load and false negatives in small current existing in traditional fixed threshold method, and realizes accurate monitoring of the operating state of the distribution box.

[0009] Preferably, the time alignment processing of the terminal temperature, the load current, and the ambient temperature comprises: setting a sampling period, and acquiring data sequences of a plurality of continuous sampling points; performing linear interpolation and timestamp calibration on the data sequences of the terminal temperature, the data sequences of the load current, and the data sequences of the ambient temperature to ensure that the terminal temperature, the load current, and the ambient temperature are synchronized in the time dimension.

[0010] Preferably, the real-time heat dissipation power of the monitoring point is equal to the product of the comprehensive heat dissipation coefficient of the monitoring point and the temperature difference, and the temperature difference is the terminal temperature at the last time minus the ambient temperature at the last time.

[0011] The present application quantifies the heat dissipation power of the conductor to the environment at the current time by constructing the calculation formula of the real-time heat dissipation power, introducing the comprehensive heat dissipation coefficient and the temperature difference between the terminal and the environment; this enables the model to dynamically distinguish between temperature changes caused by changes in heat dissipation conditions and temperature changes caused by increases in internal heat generation, providing an accurate correction for accurately removing environmental factor interference and restoring the real internal heat generation.

[0012] Preferably, the comprehensive heat dissipation coefficient of the monitoring point is obtained as follows: under the condition that the distribution box is in a healthy state and the ambient temperature is stable, the power is cut off after the equipment is run to a thermally stable state, and the natural cooling curve of the terminal temperature decreasing over time is recorded; the section of the natural cooling curve in which the temperature difference decays is selected, and the exponential decay equation is fitted using the least squares method, and the reciprocal of the extracted time constant is used as the comprehensive heat dissipation coefficient.

[0013] Preferably, the load thermal shock source intensity at the monitoring point is equal to the product of the square of the load current at the current moment and the sampling period.

[0014] This invention constructs the load thermal excitation source intensity based on Joule's law and calculates the product of the square of the current and the sampling period. This benchmark value represents the theoretical heat energy that the current should generate under standard resistance units. Using it as the denominator in subsequent calculations can transform nonlinear current fluctuations into a linear comparison benchmark, thereby effectively normalizing the influence of load current when calculating dynamic impedance.

[0015] Preferably, the equivalent electrothermal impedance modulus of the monitoring point is equal to the numerator value divided by the denominator value; the numerator value is the sum of the internal energy increment and the heat loss energy, wherein the internal energy increment is the product of the equivalent thermal capacity coefficient of the monitoring point and the difference between the terminal temperature at the current moment and the terminal temperature at the previous moment, and the heat loss energy is the product of the real-time heat dissipation power at the current moment and the sampling period; the denominator value is the sum of the load heat source intensity at the current moment and the small compensation constant.

[0016] This invention constructs a formula for calculating the equivalent electrothermal impedance modulus, comparing the sum of the actual internal energy increment absorbed by the conductor and the heat energy dissipated, i.e., the total heat generation, with the intensity of the load heat source; using the law of conservation of energy, the contact resistance, which is difficult to measure directly, is transformed into a calculable thermal characteristic quantity, realizing the online real-time quantification of the aging or loosening degree of the contact point without power outage to measure the resistance.

[0017] Preferably, the contact health level is equal to the sum of the mean value and the fluctuation value; the mean value is the product of the amplitude weighting coefficient and the average value of the equivalent electrothermal impedance modulus within the sliding time window; the fluctuation value is the product of the fluctuation weighting coefficient and the standard deviation of the equivalent electrothermal impedance modulus within the sliding time window.

[0018] The invention introduces a contact health calculation formula that includes a mean term and a standard deviation fluctuation term: the mean term can identify the continuous increase in impedance caused by oxidation and corrosion, and the fluctuation term can identify the violent oscillation of impedance caused by mechanical loosening under current impact; by the weighted combination of the two, the system can comprehensively cover the two common fault modes of distribution boxes, improve the ability to identify different types of contact anomalies, and use statistical characteristics to smooth out instantaneous noise interference.

[0019] Preferably, determining the operating status of the distribution box based on the contact health level includes: determining that the distribution box is in normal condition when the contact health level is less than a preset health threshold; and determining that the distribution box has a contact abnormality when the contact health level is greater than or equal to the preset health threshold.

[0020] Preferably, the health threshold is obtained by: collecting contact health data of the distribution box during its healthy steady-state operation; calculating the mean and standard deviation of the contact health data; and adding three times the standard deviation to the mean as the health threshold.

[0021] In a second aspect, the present invention provides an artificial intelligence-based integrated distribution box status online monitoring system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned artificial intelligence-based integrated distribution box status online monitoring method is implemented.

[0022] By adopting the above technical solution, the above-mentioned AI-based integrated distribution box status online monitoring method is generated into a computer program and stored in a memory for loading and execution by a processor. This allows for the creation of terminal devices based on the memory and processor, making them convenient to use.

[0023] The beneficial effects of this invention are as follows:

[0024] This invention aligns multi-source data in a time sequence and constructs an inversion model based on thermodynamic principles, incorporating real-time heat dissipation power and load heat source intensity, to calculate the equivalent electrothermal impedance modulus. The contact health is then determined using a sliding time window. This method physically decouples the interference of load current fluctuations and ambient temperature changes on the monitoring results. By inversely solving for the contact state through energy conservation, it effectively overcomes the thermal hysteresis effect of temperature changes relative to current changes. This allows the system to avoid false alarms caused by high ambient temperatures under high load conditions in summer and to accurately detect weak abnormal signals caused by increased contact resistance under low load conditions in winter. It solves the technical problems of high-load false alarms and low-current missed alarms inherent in traditional fixed threshold methods, achieving accurate monitoring of the distribution box's operating status. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating the online monitoring method for the status of an integrated distribution box based on artificial intelligence in this invention;

[0026] Figure 2 This is a schematic diagram illustrating the effects of data acquisition and time-series preprocessing.

[0027] Figure 3 This is a schematic diagram illustrating the comparison of impedance inversion effects based on the thermal equilibrium differential equation;

[0028] Figure 4 This is a schematic diagram illustrating the comprehensive condition assessment and fault determination. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0031] This invention discloses an online monitoring method for the status of integrated distribution boxes based on artificial intelligence, referring to... Figure 1 This includes steps S1-S4:

[0032] S1. Obtain the terminal temperature of the monitoring point, the load current of the circuit, and the ambient temperature outside the distribution box, and perform time-series alignment processing on the collected data.

[0033] It should be noted that the electromagnetic environment in industrial settings is complex, and the raw data collected by sensors often contains high-frequency noise. Furthermore, due to differences in the response speed and data transmission path between temperature sensors and current transformers, there may be slight misalignments in the raw data along the timeline. To ensure the accuracy of subsequent derivations based on physical models, it is essential to ensure that the current at the same moment corresponds to its generated thermal effect in the data. Therefore, time alignment and filtering / noise reduction processing are required for the multi-source data.

[0034] Specifically, a multi-channel synchronous acquisition device is used to acquire the terminal temperature of the monitoring point, the load current of the circuit, and the ambient temperature outside the distribution box. The sampling period is set to [value missing]. , obtain continuous The data sequence of each sampling point is used. Linear interpolation and timestamp calibration are performed on the terminal temperature sequence, load current sequence, and ambient temperature sequence to ensure strict synchronization in the time dimension, resulting in an aligned terminal temperature sequence. Load current sequence and ambient temperature sequence .in, The sampling point number, This represents the number of sampling points.

[0035] For example, Figure 2This is a schematic diagram of the data acquisition and timing preprocessing effects; the diagram includes the preprocessed terminal temperature data sequence and the corresponding load current data sequence; in the high load interference area on the left and the contact deterioration fault area on the right, the terminal temperature shows an upward trend.

[0036] S2. Calculate the real-time heat dissipation power of the monitoring point based on the aligned terminal temperature sequence and ambient temperature sequence, and calculate the load heat source intensity of the monitoring point based on the aligned load current sequence.

[0037] It should be noted that the temperature change of a conductor is the result of a dynamic interplay between internal heat generation and external heat dissipation. Traditional monitoring methods ignore the dynamic changes in heat dissipation conditions and only focus on the temperature result. This invention returns to the essence of thermodynamics, analyzing the rate of temperature change as the difference between heat generation and heat dissipation. By calculating the current heat dissipation flux, it is possible to quantify how much the conductor's temperature should have decreased if there were no current input, thus providing a benchmark for subsequently estimating the actual heat generation.

[0038] Specifically, the real-time heat dissipation power at the monitoring point satisfies the expression:

[0039]

[0040] In the formula, Indicates the monitoring point at the 1st Instantaneous heat dissipation due to temperature difference at any given moment, unit: watt; The overall heat dissipation coefficient of the monitoring point is expressed in watts per degree Celsius. Indicates the first Terminal temperature at any given time; Indicates the first The ambient temperature at any given time. The temperature difference between the monitoring point and the environment. The larger the conductor, the greater the heat power dissipated to the environment, leading to... Increase; when the overall heat dissipation coefficient The larger the size, the stronger the device's heat dissipation capacity. Increase.

[0041] Among them, the overall heat dissipation coefficient The method for obtaining the coefficient of performance is as follows: At night, when the distribution box is in a healthy state and the ambient temperature is stable, the equipment is run to a thermally stable state and then the power is cut off. The natural cooling curve of the terminal temperature decreasing over time is recorded. The section with the most significant temperature difference decay in the cooling curve is selected, and the exponential decay equation is fitted using the least squares method. The reciprocal of the extracted time constant is the comprehensive heat dissipation coefficient. In this embodiment, the overall heat dissipation coefficient is... Set to 2.0, unit: watts / degree Celsius.

[0042] Furthermore, the intensity of the load heat source at the monitoring point satisfies the expression:

[0043]

[0044] In the formula, Indicates the first The basic ability of a conductor to generate Joule heat at any given moment, measured in ampere square seconds (A²·s). Indicates the first Load current at any given moment; Indicates the sampling period. When the load current... When the current increases, according to Joule's law, the resulting thermal effect is proportional to the square of the current, leading to an increase in the intensity of the load heat source. Significantly increased.

[0045] S3. Based on the real-time heat dissipation power, load heat source intensity, and terminal temperature sequence, obtain the equivalent electrothermal impedance modulus of the monitoring point.

[0046] It should be noted that the actual heat generated by the terminals depends on the current flowing through them and the contact resistance. Since contact resistance cannot be directly measured in actual operation, and there is a thermal hysteresis between current fluctuations and temperature changes, directly observed data cannot reflect changes in contact resistance. This invention uses the inverse solution of the law of conservation of energy to compare the actually observed net heat with the intensity of the load heat source, constructing an equivalent electrothermal impedance modulus. This index can decouple the influence of load fluctuations and directly characterize the aging degree of the contact point (i.e., the magnitude of the contact resistance).

[0047] Specifically, the equivalent electrothermal impedance modulus of the monitoring point satisfies the following expression:

[0048]

[0049] In the formula, Indicates the first Equivalent thermoelectric modulus at time t, in ohms; The equivalent heat capacity coefficient of the monitoring point is expressed in joules per degree Celsius. Indicates the first Terminal temperature at any given time; Indicates the first Terminal temperature at any given time; Indicates the first Real-time heat dissipation power at any given moment; Indicates the sampling period; Indicates the first The intensity of the load heat source at any given time, in ampere squared second (A²·s); This represents a small compensation constant.

[0050] Molecular part Physically, this represents the conductor in the first... The total actual heat energy generated within each sampling period, in joules, of which The internal energy increment used to increase the temperature of the conductor, This represents the heat energy lost to the environment during this period. The denominator is... This represents the integral of the square of the current over time. According to Joule's law... It can be seen that the ratio of the actual total heat generated to the integral of the square of the current is the resistance value. Therefore, It numerically reflects the current contact resistance. When the contact point is in good condition, the contact resistance is low. Maintain a low level; when contact points age or loosen, causing contact resistance to increase, the actual total heat increases, leading to... Increase.

[0051] Among them, the infinitesimal compensation constant The setting method is as follows: Select an extremely small positive number, in ampere-squared seconds (A²·s), ensuring that its value is much smaller than the thermal efficiency reference generated by the normal operating current, but can serve as the dominant term in the denominator under no-load conditions. In this embodiment, Set as This is to avoid calculation divergence caused by a denominator of zero.

[0052] For example, Figure 3 This is a schematic diagram comparing the impedance inversion effect based on the thermal balance differential equation. The solid line in the figure represents the change curve of the equivalent electrothermal impedance modulus obtained by inversion calculation based on the method of this invention combined with the current gating mechanism, and the dashed line represents the change curve of the actual contact resistance as the true value for comparison. The figure shows that in the high load interference zone in the middle period, although the load current fluctuates drastically, the equivalent electrothermal impedance modulus obtained by inversion remains in a low stable state without any false increase. In the fault zone on the right, it can closely follow the linear increase of the actual contact resistance, which verifies the effective decoupling of the load current and the environmental thermal effect of this invention.

[0053] S4. Construct a sliding time window based on the equivalent electrothermal impedance modulus, calculate the contact health within the sliding time window, and determine the operating status of the distribution box based on the contact health.

[0054] It should be noted that the equivalent electrothermal impedance modulus at a single point may be affected by sensor noise or electromagnetic interference, resulting in spikes, and different types of faults exhibit different behaviors in the impedance domain. Oxidation corrosion typically manifests as a steady increase in impedance value, while mechanical loosening often manifests as a sharp fluctuation in contact resistance under current surges. To improve the robustness of monitoring, alarms should not be based solely on single-point values, but rather on the statistical distribution characteristics of the index over a period of time, using a comprehensive evaluation of the mean and volatility to identify anomalies.

[0055] Specifically, establish a length of For a sliding time window, the state evaluation function satisfies the expression:

[0056]

[0057] In the formula, Indicates the first Constant contact with health status; This represents the magnitude weighting coefficient; Indicates the volatility weighting coefficient; Indicates the length of the sliding time window; This represents the average value of the equivalent electrothermal impedance modulus within the sliding time window.

[0058] The first term in the expression is the mean, reflecting the average level of contact resistance. This term increases significantly when oxidation and corrosion at the contact point cause a continuous increase in resistance. The second term is the standard deviation, reflecting the stability of the contact resistance. This term increases significantly when mechanical loosening at the contact point causes drastic fluctuations in impedance due to current surges. Through weighted summation, It can comprehensively reflect both aging and loosening failure modes.

[0059] Wherein, the length of the sliding time window The setting method is as follows: considering the data update frequency and the sensitivity of the fault response, a time length that can cover the typical current fluctuation cycle is selected. In this embodiment, if the sampling period... The length of the sliding time window is 1 second. Set to 60, which corresponds to 1 minute of data, because this length can smooth out instantaneous noise while retaining minute-level fault characteristics.

[0060] Among them, the magnitude weighting coefficient With volatility weighting coefficient The value of is in the range [0,1], and In the state evaluation function, the magnitude weighting coefficient is used. Adjusting the weighting of the mean term in response to a sustained increase in impedance allows for accurate identification of long-term contact deterioration caused by oxidation and corrosion, while also using a fluctuation weighting coefficient. The sensitivity of the fluctuation term to the impedance standard deviation is adjusted to effectively capture the severe oscillations caused by mechanical loosening under current surges. This weighted combination of both methods achieves comprehensive monitoring of aging and loosening fault modes in the distribution box. Therefore, in specific allocation... and The settings need to balance the sensitivity to two types of faults: if the distribution box is in an environment with high vibration, such as near railways or heavy machinery, the fluctuation weighting coefficient can be appropriately increased. The value of the fluctuation weight coefficient at this time. The value range is (0.5, 0.7]; if in a high-humidity or corrosive environment, the amplitude weighting coefficient should be emphasized. At this point, the amplitude weighting coefficient The value range is (0.5, 0.7]; in this embodiment, the amplitude weighting coefficient is... Set the fluctuation weighting coefficient to 0.33. The value is set to 0.67; in other embodiments, the implementer can set the amplitude weighting coefficient according to the environment in which the distribution box is located. With volatility weighting coefficient .

[0061] Furthermore, based on the health level of contact Fault determination based on size: If Less than the preset health threshold The distribution box is determined to be in normal condition; if Greater than or equal to the preset health threshold The system determined that there was an abnormal contact in the distribution box.

[0062] Among them, health threshold The setting method is as follows: Collect contact health data of the distribution box during its healthy steady-state operation (e.g., within one week after new equipment is put into operation and passes thermal stabilization), calculate the mean and standard deviation of this data set, and take the mean plus three times the standard deviation as the benchmark value. In this embodiment, the health threshold... It can be dynamically set according to the statistical characteristics of the actual monitoring data.

[0063] For example, Figure 4 This diagram illustrates the comprehensive condition assessment and fault determination. The solid line in the diagram represents the change curve of the contact health quantification value calculated based on the equivalent electrothermal impedance modulus through sliding window statistical analysis. The filled area above the health threshold represents the alarm triggering area. The diagram shows that in the high-load interference area, the contact health quantification value always remains below the health threshold, avoiding false alarms. However, in the contact deterioration fault area, the value rapidly rises and exceeds the health threshold, entering the alarm triggering area, thus achieving accurate identification of abnormal conditions.

[0064] This invention also discloses an artificial intelligence-based integrated distribution box status online monitoring system, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the artificial intelligence-based integrated distribution box status online monitoring method according to this invention is implemented.

[0065] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. An online monitoring method for the status of integrated distribution boxes based on artificial intelligence, characterized in that, include: The terminal temperature of the monitoring point, the load current of the circuit, and the ambient temperature outside the distribution box are obtained. The terminal temperature, load current, and ambient temperature are time-aligned to obtain the aligned terminal temperature sequence, load current sequence, and ambient temperature sequence. The real-time heat dissipation power of the monitoring point is calculated based on the aligned terminal temperature sequence and ambient temperature sequence, and the load heat source intensity of the monitoring point is calculated based on the aligned load current sequence. Based on the real-time heat dissipation power, the intensity of the load heat source, and the terminal temperature sequence, the equivalent electrothermal impedance modulus of the monitoring point is obtained. A sliding time window is constructed based on the equivalent electrothermal impedance modulus, the contact health within the sliding time window is calculated, and the operating status of the distribution box is determined based on the contact health.

2. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 1, characterized in that, The timing alignment process for terminal temperature, load current, and ambient temperature includes: Set a sampling period to obtain a data sequence of multiple consecutive sampling points; Linear interpolation and timestamp calibration are performed on the data sequences of terminal temperature, load current, and ambient temperature to ensure that the terminal temperature, load current, and ambient temperature are synchronized in the time dimension.

3. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 1, characterized in that, The real-time heat dissipation power of the monitoring point is equal to the product of the comprehensive heat dissipation coefficient of the monitoring point and the temperature difference, where the temperature difference is the terminal temperature at the previous moment minus the ambient temperature at the previous moment.

4. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 3, characterized in that, The method for obtaining the overall heat dissipation coefficient of the monitoring point is as follows: With the distribution box in good condition and the ambient temperature stable, the equipment was operated to a thermally stable state and then the power was cut off. The natural cooling curve of the terminal temperature decreasing over time was recorded. The section of the natural cooling curve in which the temperature difference decays is selected, and the exponential decay equation is fitted using the least squares method. The reciprocal of the extracted time constant is used as the comprehensive heat dissipation coefficient.

5. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 1, characterized in that, The load heat source intensity at the monitoring point is equal to the product of the square of the load current at the current moment and the sampling period.

6. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 1, characterized in that, The equivalent electrothermal impedance modulus of the monitoring point is equal to the numerator value divided by the denominator value; the numerator value is the sum of the internal energy increment and the heat loss energy, wherein the internal energy increment is the product of the equivalent thermal capacity coefficient of the monitoring point and the difference between the current terminal temperature and the previous terminal temperature, and the heat loss energy is the product of the current real-time heat dissipation power and the sampling period; the denominator value is the sum of the current load heat source intensity and the small compensation constant.

7. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 1, characterized in that, The contact health level is equal to the sum of the mean value and the fluctuation value; the mean value is the product of the amplitude weighting coefficient and the average value of the equivalent electrothermal impedance modulus within the sliding time window; the fluctuation value is the product of the fluctuation weighting coefficient and the standard deviation of the equivalent electrothermal impedance modulus within the sliding time window.

8. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 1, characterized in that, The step of determining the operating status of the distribution box based on the contact health status includes: When the contact health level is less than the preset health threshold, the distribution box is determined to be in normal condition. When the contact health level is greater than or equal to a preset health threshold, it is determined that there is a contact abnormality in the distribution box.

9. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 8, characterized in that, The health threshold is obtained as follows: Collect contact health data of the distribution box during its healthy steady-state operation; Calculate the mean and standard deviation of the contact health data; The mean plus three times the standard deviation is used as the health threshold.

10. An integrated distribution box status online monitoring system based on artificial intelligence, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement the online monitoring method for the status of an integrated distribution box based on artificial intelligence according to any one of claims 1-9.

Images