Artificial intelligence-based comprehensive distribution box state online monitoring method and system

Abstract

The present application 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, which comprises the following steps: acquiring the terminal temperature of a monitoring point, the load current of the loop and the ambient temperature outside the distribution box, and performing time sequence alignment processing on the collected data; calculating the real-time heat dissipation power and the load thermal excitation source intensity of the monitoring point based on the aligned data; acquiring the equivalent electro-thermal impedance modulus of the monitoring point based on the real-time heat dissipation power, the load thermal excitation source intensity and the terminal temperature sequence; constructing a sliding time window based on the equivalent electro-thermal impedance modulus, calculating the contact health degree in the sliding time window, and determining the running state of the distribution box according to the contact health degree. The present application decouples the interference of the load current and the ambient temperature by constructing a physical inversion model, effectively overcomes the thermal hysteresis phenomenon, and realizes the precise monitoring of the contact aging and loosening faults.

Description

Technical Field

[0001] This invention relates to the field of power equipment condition monitoring and fault diagnosis technology. More specifically, this invention relates to an online monitoring method and system for integrated distribution boxes based on artificial intelligence. Background Technology

[0002] As a critical node in low-voltage power distribution networks, integrated distribution boxes are exposed to complex outdoor environments for extended periods. Key components such as circuit breaker contacts and busbar connection points are highly susceptible to increased contact resistance due to oxidation and corrosion from long-term operation, as well as loosening of bolts caused by mechanical vibration. This contact degradation can lead to abnormal localized temperature rises, potentially causing equipment burnout or even fires.

[0003] Currently, status monitoring of distribution boxes mainly relies on infrared thermography or pre-embedded temperature sensors. Existing technologies typically use fixed temperature thresholds or temperature rise thresholds for judgment, that is, an alarm is triggered when the temperature at the monitoring point exceeds the set value.

[0004] However, this static monitoring method based on fixed thresholds has significant technical limitations. Specifically, the internal temperature of the distribution box is affected by both load current and ambient temperature. Under high load conditions in summer, even if the equipment connections are normal, the conductor temperature may approach the alarm threshold, leading to frequent false alarms and increasing the workload of maintenance personnel for ineffective inspections. Conversely, in winter or under low load conditions, even if the connection points have experienced severe aging or loosening, increasing contact resistance, the absolute temperature may still be below the alarm threshold due to the relatively small amount of Joule heat generated, causing the system to miss alarms until the equipment is completely burned out. In addition, due to the thermal capacity characteristics of metallic conductors, temperature changes exhibit a significant thermal lag relative to current changes, and directly comparing real-time current and temperature cannot accurately reflect the current contact status.

[0005] Therefore, how to eliminate the interference of the environment and load and accurately extract the essential characteristics reflecting the contact state of the conductor from the thermal hysteresis data is an urgent problem to be solved in the field of online monitoring of distribution boxes. Summary of the Invention

[0006] To address the technical problem that existing technologies, under complex operating conditions, are affected by load fluctuations, ambient temperature interference, and thermal hysteresis, making it impossible to accurately characterize the conductor contact state and thus causing false alarms and missed alarms in the monitoring system, this invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides an online monitoring method for the status of a comprehensive distribution box based on artificial intelligence, 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-series alignment processing on the terminal temperature, load current, and ambient temperature to obtain aligned terminal temperature sequences, load current sequences, and ambient temperature sequences; calculating the real-time heat dissipation power of the monitoring point based on the aligned terminal temperature sequences and ambient temperature sequences, and calculating the load heat source intensity of the monitoring point based on the aligned load current sequence; acquiring the equivalent electrothermal impedance modulus of the monitoring point based on the real-time heat dissipation power, the load heat source intensity, and the terminal temperature sequences; constructing a sliding time window based on the equivalent electrothermal impedance modulus, calculating the contact health within the sliding time window, and determining the operating status of the distribution box based on the contact health.

[0008] 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.

[0009] Preferably, the time-series alignment processing of terminal temperature, load current, and ambient temperature includes: setting a sampling period and acquiring a data sequence of multiple consecutive sampling points; performing linear interpolation and timestamp calibration on the data sequence 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.

[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, wherein the temperature difference is the terminal temperature at the previous moment minus the ambient temperature at the previous moment.

[0011] This invention quantifies the heat dissipation power of a conductor to the environment at the current moment by constructing a formula for calculating real-time heat dissipation power and introducing a 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 increased internal heat generation, providing accurate correction values ​​for accurately removing environmental interference and restoring the true internal heat generation situation.

[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 Constantly monitor 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 point. 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 where the terminal is located, 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. The load heat source intensity of the monitoring point is equal to the product of the square of the load current at the current moment and the sampling period. Based on the real-time heat dissipation power, the load heat source intensity, and the terminal temperature sequence, the equivalent electrothermal impedance modulus of the monitoring point is obtained. 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 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. A sliding time window is constructed based on the equivalent electrothermal impedance modulus. The contact health within the sliding time window is calculated. The contact health 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. 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 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.

6. The method for online monitoring of the status of an integrated distribution box based on artificial intelligence according to claim 5, 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.

7. 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 as described in any one of claims 1-6.

Images