Terminal contact state diagnosis method for multi-circuit intelligent electric meter

Abstract

The application relates to a terminal contact state diagnosis method of a multi-loop intelligent electric meter, which comprises the following steps: collecting a first temperature value output by a contact area temperature sensor on each terminal seat and a second temperature value output by a reference area temperature sensor; calculating a temperature gradient value by subtracting the first temperature value from the second temperature value; obtaining current sampling values of each loop, obtaining a normalized temperature coefficient based on the temperature gradient value and the current sampling value; forming a sample set by the normalized temperature coefficients of the loops in the same functional position and calculating statistical characteristic quantities, detecting whether the sample set presents a systematic deviation mode related to the physical arrangement position of the terminal seat, applying compensation correction to the temperature gradient value and recalculating; comparing the normalized temperature coefficient with the statistical characteristic quantity, and marking the corresponding terminal seat as an abnormal contact state if the deviation exceeds a deviation threshold value. The application eliminates environmental interference and thermal crosstalk through double temperature measurement point gradient difference, and realizes terminal fault self-diagnosis without external reference.

Description

Technical Field

[0001] This application relates to the field of smart meters, and more particularly to a method for diagnosing the terminal contact status of a multi-circuit smart meter. Background Technology

[0002] In power distribution systems, the terminal blocks of smart meters are critical connection nodes in the current loop. Over long-term operation, the contact surface of the terminal blocks experiences a gradual increase in contact resistance due to oxidation, corrosion, and mechanical loosening. This increased contact resistance causes additional Joule heat to be generated when the terminal block carries current, leading to a sustained rise in localized temperature. In severe cases, this can cause terminal erosion, insulation degradation, and even fires. Therefore, online monitoring and early warning of the terminal block contact status are fundamental requirements for ensuring the safe operation of power distribution systems.

[0003] Existing terminal temperature monitoring solutions typically place a temperature sensor near each terminal block to collect the absolute temperature value of the terminal block. An alarm is triggered when the absolute temperature value exceeds a pre-set fixed threshold. This single-point temperature measurement plus fixed threshold solution can meet basic overheat protection requirements in scenarios with a small number of terminal blocks and large spacing between them. However, as power distribution systems develop towards intensification and modularization, multi-circuit smart meters are widely used in distribution cabinets. Multi-circuit smart meters integrate multiple independent metering circuits within the same housing, each circuit equipped with an independent terminal block, with these terminal blocks densely arranged along the width of the meter. This densely arranged structure exposes the aforementioned single-point temperature measurement plus fixed threshold solution to a systemic failure risk.

[0004] Specifically, in multi-circuit metering meters, the physical spacing between multiple terminal blocks is small, resulting in short heat conduction paths, low thermal resistance, and significant thermal crosstalk between adjacent circuits. When a circuit is under heavy load, the heat generated by its terminal blocks can diffuse through the housing structure and air conduction to the terminal block areas of adjacent circuits, causing the temperature readings of the adjacent circuits' terminal blocks to rise. Under these conditions, single-point temperature sensors cannot distinguish whether the temperature rise of the terminal block comes from its own increased contact resistance or from thermal crosstalk between adjacent circuits, thus generating false alarms or missed alarms. Simultaneously, changes in ambient temperature and fluctuations in ventilation conditions within the distribution cabinet can also cause an overall shift in the temperature of all terminal blocks, further reducing the signal-to-noise ratio of single-point temperature measurement.

[0005] Furthermore, the load levels of each circuit in a multi-circuit meter are typically different. Even if the terminal contacts of a heavily loaded circuit are perfectly normal, its terminal temperature will be significantly higher than that of a lightly loaded circuit. If a uniform fixed temperature threshold is applied to all circuits, normal temperature rises in heavily loaded circuits may be misjudged as abnormal, while terminals in lightly loaded circuits that have begun to deteriorate but whose absolute temperature rise has not yet reached the threshold will be missed. Increasing the threshold increases the false alarm rate, while decreasing the threshold increases the false alarm rate; a fixed threshold cannot simultaneously accommodate circuits with different load levels.

[0006] The more fundamental problem is that the aforementioned thermal crosstalk interference and load differences are not independent of each other—they are intertwined and superimposed in the densely arranged multi-loop physical structure, jointly affecting the temperature reading of each terminal block. Eliminating only one of the interference factors is insufficient to obtain a reliable diagnostic conclusion. For example, even if the effect of thermal crosstalk is compensated for in some way, it is still impossible to directly compare temperature values ​​between different load loops to determine which terminal has a worse contact condition; conversely, even if the temperature values ​​are load normalized, the normalized input data itself is already distorted if thermal crosstalk is not eliminated. Summary of the Invention

[0007] In order to eliminate thermal crosstalk interference and load difference effects in a densely packed environment of multi-circuit smart meters, and to autonomously determine the terminal contact status without the need for external reference equipment, this application provides a terminal contact status diagnosis method for multi-circuit smart meters.

[0008] This application provides a terminal contact status diagnostic method for multi-circuit smart meters, which adopts the following technical solution:

[0009] A method for diagnosing the terminal contact status of a multi-circuit smart meter, wherein the multi-circuit smart meter includes multiple sets of terminal blocks, each set of terminal blocks is provided with a contact area temperature sensor and a reference area temperature sensor, each terminal block includes a conductor, the contact area temperature sensor is disposed at the terminal contact surface of the conductor, and the reference area temperature sensor is disposed at a predetermined distance offset from the contact area temperature sensor along the extension direction of the conductor. The terminal contact status diagnosis method includes the following steps:

[0010] S1. Collect the first temperature value output by the contact area temperature sensor and the second temperature value output by the reference area temperature sensor for each terminal block;

[0011] S2. For each terminal block, calculate the difference between the first temperature value and the second temperature value to obtain the temperature gradient value;

[0012] S3. Obtain the current sampling value of each circuit of the multi-circuit smart meter, and obtain the normalized temperature coefficient based on the ratio of the temperature gradient value of each terminal block to the square of the current sampling value of the corresponding circuit.

[0013] S4. Form a sample set of the normalized temperature coefficients in the same functional position of the multiple loops, and calculate the statistical characteristic quantity of the sample set; detect whether the sample set presents a systematic deviation pattern related to the physical arrangement position of the terminal block, and in response to the systematic deviation pattern, apply compensation correction to the temperature gradient value and re-execute S3 and the calculation of the statistical characteristic quantity based on the compensated temperature gradient value until the systematic deviation pattern is eliminated or the preset iteration termination condition is reached;

[0014] S5. In response to the absence of the systematic deviation mode or the completion of the compensation correction, each normalized temperature coefficient is compared with the statistical feature quantity. In response to the normalized temperature coefficient deviating from the statistical feature quantity by more than a deviation threshold, the corresponding terminal block is marked as a contact abnormality state.

[0015] By adopting the above technical solution, the contact area temperature sensor collects the temperature at the terminal contact surface, while the reference area temperature sensor collects the temperature at a distance offset along the conductor. When the terminal contact resistance increases, Joule heat is concentrated at the contact surface and decays along the conductor; therefore, the temperature difference between the two sensors directly reflects the local heat source intensity at the contact surface. Ambient temperature changes and far-field thermal crosstalk between adjacent circuits have approximately the same effect on two sensors with finite spacing on the same conductor, which are eliminated in the difference calculation. Based on this, the normalized temperature coefficient is obtained by dividing the temperature gradient value by the square of the corresponding circuit current, thus eliminating the load differences between different circuits and allowing the normalized temperature coefficient to directly characterize the contact resistance characteristics. The normalized temperature coefficients of the same functional positions in multiple circuits are used to form a sample set for statistical comparison. Multiple circuits with the same structure within the same device are used as a natural control group, and the statistical characteristics of the sample set itself are used as a dynamic benchmark, eliminating the need for a fixed temperature threshold or external benchmark equipment. When the statistical comparison detects a systematic deviation pattern related to the physical arrangement position, compensation correction is triggered and recalculation is performed, allowing the diagnostic results to drive the improvement of decoupling accuracy and forming a closed-loop diagnostic mechanism.

[0016] Optionally, the conductor is a conductive copper busbar, the contact area temperature sensor and the reference area temperature sensor are mounted on the conductive copper busbar, the reference area temperature sensor is disposed in the non-contact bearing area of ​​the conductive copper busbar, and the distance between the contact area temperature sensor and the reference area temperature sensor along the extension direction of the conductive copper busbar is 5 mm to 20 mm; the multiple sets of terminal blocks are arranged in an array along the width direction of the multi-circuit smart meter, and a heat insulation structure is provided between adjacent terminal blocks.

[0017] By adopting the above technical solution, the sensor is directly mounted on the surface of the conductive copper busbar. The temperature signal is transmitted through the high thermal conductivity path of the copper busbar, resulting in a fast response speed and immunity to air convection disturbances. The reference temperature sensor is placed in a non-contact load-bearing area that does not bear the wire crimping force, avoiding interference from crimping stress on the reference temperature measurement. The spacing is limited to 5 mm to 20 mm, within which the signal-to-noise ratio between the gradient signal amplitude and the sensor measurement error is within an engineerable range. A thermal insulation structure is provided between adjacent terminal blocks to reduce thermal coupling at the physical level and alleviate the computational burden of compensation and correction at the algorithm level.

[0018] Optionally, the compensation correction in S4 includes the following sub-steps:

[0019] S41. Construct a thermal coupling coefficient matrix between loops, wherein the thermal coupling coefficient matrix between loops is a strip-shaped sparse matrix based on the arrangement topology of the terminal blocks, and only physically adjacent terminal blocks have non-zero coupling coefficients.

[0020] S42. Based on the thermal coupling coefficient matrix between the circuits, the temperature gradient value of each terminal block is corrected twice to obtain the compensated temperature gradient value;

[0021] S43. Based on the compensated temperature gradient value, S3 is re-executed and the statistical characteristic quantity is recalculated;

[0022] S44. In response to the deviation amplitude of the systematic deviation mode falling below a preset convergence threshold, the systematic deviation mode is confirmed to be eliminated, or in response to the preset iteration termination condition being met when the number of iterations reaches a preset upper limit, the compensation correction is confirmed to be completed.

[0023] By adopting the above technical solution, the structure of the strip-shaped sparse matrix corresponds to the physical characteristics of the linear arrangement of terminal blocks, retaining only the non-zero coupling coefficient between adjacent terminal blocks. The matrix computation complexity is linearly related to the number of loops, making it suitable for real-time execution on meter-level microcontrollers. Using temperature gradient values ​​instead of absolute temperature values ​​as matrix inputs, since gradient difference has already completed one coarse decoupling, the matrix only needs to compensate for the residual near-field asymmetric crosstalk components. The condition number is better than that of the full decoupling matrix acting on the original temperature, resulting in higher numerical stability. The dual termination conditions of convergence threshold and iteration upper limit stop iteration in time when the deviation is eliminated, preventing infinite recursion when the matrix accuracy is insufficient.

[0024] Optionally, the method further includes an online recalibration step for the inter-loop thermal coupling coefficient matrix:

[0025] In response to a single circuit in the multi-circuit smart meter having a current change amplitude exceeding a preset proportion of the rated current and a duration exceeding a preset minimum duration, while the current change rate of the other circuits is lower than a preset stability threshold, the single circuit is marked as an excitation circuit.

[0026] Candidate coupling coefficients are calculated based on the ratio between the temperature gradient change of each terminal block in the excitation loop and the temperature gradient change of each physically adjacent terminal block.

[0027] In response to the candidate coupling coefficient falling within a preset consistency interval centered on the coupling coefficient at the corresponding position in the inter-loop thermal coupling coefficient matrix, the coupling coefficient at the corresponding position is updated with the candidate coupling coefficient;

[0028] In response to the candidate coupling coefficient exceeding the consistency range, the candidate coupling coefficient is discarded.

[0029] By employing the above technical solution, and utilizing the statistical independence of load changes in each circuit over time, online recalibration of the coupling coefficient matrix is ​​achieved without applying external excitation. Dual threshold screening based on the amplitude and duration of current changes eliminates contamination of the calibration by minor load fluctuations and transient spikes. Candidate coupling coefficients are checked for consistency with their corresponding positions in the matrix; only those falling within a preset range are accepted for updating, while those exceeding the range are discarded. This allows the matrix to track changes in thermal coupling paths caused by long-term factors such as dust accumulation and oxidation, while preventing contamination of the matrix by single abnormal events.

[0030] Optionally, the following steps may also be included:

[0031] Monitor the current change rate of each circuit. In response to the current change rate of a single circuit in the multi-circuit smart meter exceeding a preset change rate threshold and the current change rate of the remaining circuits being lower than a preset stability threshold, mark the time period that meets the conditions as a diagnostic window and mark the single circuit as an excitation circuit.

[0032] Within the diagnostic window, the temperature gradient values ​​of each terminal block in each of the other circuits besides the excitation circuit are collected over time, and the thermal time constant is extracted.

[0033] Record the multiple thermal time constants extracted from each terminal block in multiple diagnostic windows corresponding to different excitation circuits;

[0034] Cross-compare multiple thermal time constants of the same terminal block. If a thermal time constant exceeding a preset proportion is lower than a preset time constant reference value, it is confirmed that the contact resistance of the corresponding terminal block has increased.

[0035] By adopting the above technical solution, a dynamic diagnostic dimension of thermal time constant is introduced in addition to the steady-state diagnosis of the normalized temperature coefficient. When the contact resistance increases, heat concentrates on the contact surface, the equivalent heat capacity decreases, and the thermal time constant shortens—this, based on a different physical mechanism than the normalized temperature coefficient, provides a mutually corroborating criterion for the same fault mode. The diagnostic window uses the "one change, multiple stability" condition to capture natural load transient events. The same terminal block is repeatedly diagnosed in multiple diagnostic windows corresponding to different excitation circuits. Cross-comparison eliminates misjudgments caused by single, occasional interferences, and the diagnostic reliability increases with the accumulation of the number of windows.

[0036] Optionally, the time constant reference value is determined based on the following method:

[0037] During the initial operation phase of the multi-circuit smart meter, the thermal time constant of each terminal block in multiple diagnostic windows is collected, and the statistical mean of the thermal time constant of each terminal block is used as the reference value of the time constant and stored.

[0038] By adopting the above technical solution, the statistical mean of the thermal time constant of each terminal block during the initial operation phase is used as the benchmark value. The contact resistance is at its lowest level since the factory design during the initial phase, and the statistical mean eliminates random errors from single measurements, enabling subsequent dynamic diagnostics to detect the degree of degradation relative to the factory condition.

[0039] Optionally, in S3, in response to the current sample value of the circuit being lower than a preset minimum current threshold, the calculation of the normalized temperature coefficient of the corresponding terminal block is skipped.

[0040] The temperature gradient value of the corresponding terminal block is compared with the historical temperature gradient baseline value of the corresponding terminal block. In response to the temperature gradient value exceeding a preset multiple of the historical temperature gradient baseline value, the corresponding terminal block is marked as having an abnormal contact state.

[0041] By adopting the above technical solution, when the loop current is below the minimum current threshold, the normalization calculation is skipped, avoiding the divergence of the normalized temperature coefficient value caused by the square of the current approaching zero. Instead, the temperature gradient value is compared longitudinally with the historical temperature gradient baseline value of the terminal block itself. The longitudinal comparison does not depend on the lateral comparison between loops, and it can still work independently under low load conditions, ensuring full operating condition coverage.

[0042] Optionally, the statistical features include the mean and standard deviation of the sample set, and the deviation threshold is the sum of the mean and N times the standard deviation, where N is a preset positive number;

[0043] The determination condition for the normalized temperature coefficient deviating from the statistical feature quantity by more than the deviation threshold in S5 includes: the normalized temperature coefficient deviating from the statistical feature quantity by more than the deviation threshold by K times within M consecutive sampling periods, where M and K are preset positive integers and K is less than or equal to M.

[0044] The step S5 further includes: recording the time series of the normalized temperature coefficient of each terminal block over multiple consecutive monitoring periods, and generating a deterioration warning for the corresponding terminal block in response to the time series showing a monotonically increasing trend and the rate of increase exceeding a preset trend threshold.

[0045] By adopting the above technical solution, the deviation threshold dynamically changes with the sample set distribution as the sum of the mean and N times the standard deviation. When most loops are in the same state, the threshold tightens and the sensitivity increases; when normal dispersion exists in each loop, the threshold is moderately relaxed. The sliding window confirmation mechanism, which verifies K deviations within M consecutive periods, upgrades single judgment to multiple confirmations, filtering transient interference. The monotonically increasing trend detection of the time series generates an early warning for terminal blocks that have not yet triggered the threshold but are gradually deteriorating, extending the diagnosis from post-event alarm to pre-event warning.

[0046] In summary, this application includes at least one of the following beneficial technical effects:

[0047] 1. By eliminating environmental interference and far-field thermal crosstalk through gradient difference at dual temperature measurement points, and by stripping load differences through current square normalization, and by using multi-loop circuits with the same structure as a natural control group for horizontal statistical comparison, autonomous diagnosis of terminal contact status without external reference equipment and fixed temperature threshold is achieved. Furthermore, a closed-loop diagnostic mechanism is formed by systematic deviation mode detection-driven compensation correction.

[0048] 2. By using a two-level architecture of coarse decoupling and fine correction of a striped sparse matrix, a screening and verification mechanism of online recalibration, and a dual iterative termination condition, the compensation correction can be performed in real time under the computing power constraint of the meter-level microcontroller, and the matrix can track the drift changes of the thermal coupling path during long-term operation of the equipment.

[0049] 3. By capturing natural load transient events and introducing a dynamic diagnostic dimension of thermal time constant through multi-window cross-comparison, combined with low load degradation processing, dynamic adaptive threshold, sliding window multiple confirmation and time series trend early warning, it covers the full operating condition range from full load to no load and the full failure evolution stage from sudden failure to gradual degradation. Attached Figure Description

[0050] Figure 1 This is a flowchart illustrating the terminal contact status diagnosis method for a multi-circuit smart meter in one embodiment of this application.

[0051] Figure 2This is a top view schematic diagram of the overall layout of a multi-circuit smart meter in one embodiment of this application.

[0052] Figure 3 This is a side view of the detailed structure of the terminal block in one embodiment of this application.

[0053] Explanation of reference numerals in the attached figures:

[0054] 100. Multi-circuit smart meter; 110. Terminal block; 111. Conductive copper busbar; 112. Contact area temperature sensor; 113. Reference area temperature sensor; 114. Terminal contact surface; 116. Thermal insulation structure. Detailed Implementation

[0055] The present application will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the scope of the application.

[0056] This application discloses a terminal contact status diagnosis method for a multi-circuit smart meter 100. The multi-circuit smart meter 100 in this application refers to a smart energy meter that integrates multiple independent metering circuits within the same device, also known as a multi-circuit metering terminal. The multi-circuit smart meter 100 is installed in a distribution cabinet. Each device can integrate 8, 12, or 16 independent metering circuits. Each circuit is equipped with an independent terminal block 110, which includes an incoming terminal and an outgoing terminal. Figure 2 A top-view diagram of the overall layout of a multi-circuit smart meter 100 is shown. Taking a 12-circuit smart meter as an example, the 12 circuits supply power to different floors or areas in a commercial building, and the load type and load level of each circuit are different. For example, circuit 1 supplies power to the elevator, with a rated operating current of approximately 30A; circuit 5 supplies power to public area lighting, with an operating current of approximately 5A; circuit 9 is currently in an unloaded state, with only a weak standby current.

[0057] In this environment of densely arranged multi-circuit systems, the traditional single-point temperature measurement and fixed temperature threshold monitoring scheme faces three systemic problems. First, multiple sets of terminal blocks 110 are closely arranged along the width of the multi-circuit smart meter 100, with small physical spacing between adjacent terminal blocks 110. When a circuit is under heavy load, the heat generated by the terminal block 110 of that circuit is diffused to the terminal block 110 area of ​​adjacent circuits through the casing structure and air conduction. The single-point temperature sensor cannot distinguish whether the temperature rise originates from increased contact resistance within itself or from thermal crosstalk between adjacent circuits. Second, the load levels of each circuit are different. Even if the terminal contact condition of a heavy-load circuit is normal, its terminal temperature will be significantly higher than that of a light-load circuit. When a uniform fixed temperature threshold is applied to all circuits, the normal temperature rise of a heavy-load circuit may be misjudged as abnormal, while terminals in light-load circuits that have already begun to deteriorate may be missed. Third, the aforementioned thermal crosstalk and load differences are intertwined and superimposed in the densely arranged multi-circuit structure; eliminating only one of the interfering factors is insufficient to obtain a reliable diagnostic conclusion.

[0058] To address the aforementioned issues, the diagnostic method provided in this application forms a closed loop through five steps: acquiring raw temperature data from dual temperature measurement points; eliminating ambient temperature offset and far-field thermal crosstalk through gradient difference; eliminating load differences through current square normalization; achieving anomaly identification without external benchmarks through multi-loop lateral statistical comparison using circuits of the same structure within the same equipment as a natural control group; and driving compensation correction by detecting systematic deviation patterns related to physical arrangement positions in the statistical samples, thereby improving the decoupling accuracy of the diagnostic results feedback. The following is combined with… Figure 1 The overall flowchart shown provides a detailed explanation of each step.

[0059] Each terminal block 110 is equipped with two temperature sensors. For example... Figure 3 As shown, terminal block 110 includes a conductor that carries current and provides a crimped connection for the wire. A contact area temperature sensor 112 is disposed at the terminal contact surface 114 of the conductor, i.e., the area where the wire and conductor are physically crimped; this area is the source of Joule heat generated by contact resistance. A reference area temperature sensor 113 is offset from the contact area temperature sensor 112 by a preset distance along the extension direction of the conductor. The selection of this preset distance needs to consider both the amplitude of the gradient signal and the purity of the reference measurement.

[0060] In some embodiments, the conductor is a conductive copper busbar 111. The conductive copper busbar 111 has a flat rectangular cross-section, and the contact area temperature sensor 112 and the reference area temperature sensor 113 are directly mounted on the surface of the conductive copper busbar 111. The temperature signal is transmitted to the sensor through the high thermal conductivity path of the conductive copper busbar 111, resulting in a fast response speed and immunity to air convection disturbances. The reference area temperature sensor 113 is located in the non-contact bearing area of ​​the conductive copper busbar 111, i.e., the section that does not bear the wire crimping force. The crimping force can cause local deformation and stress concentration of the conductive copper busbar 111 near the crimping point. If the reference area sensor is located in this area, it will introduce temperature interference unrelated to contact resistance, reducing the accuracy of the gradient differential. The distance between the contact area temperature sensor 112 and the reference area temperature sensor 113 along the extension direction of the conductive copper busbar 111 is 5 mm to 20 mm. When the spacing is less than 5 mm, the difference in readings between the two sensors is less than the typical measurement accuracy of an NTC thermistor (±0.5℃), and the gradient signal is drowned out by sensor noise. When the spacing is greater than 20 mm, the reference area sensor is too far from the contact surface, and its temperature reading is increasingly affected by the heat dissipation conditions of the middle section of the conductive copper busbar 111 (such as whether it is adjacent to the heat insulation structure 116 or the location of the vent), and it is no longer a reference standard that purely reflects the temperature of the conductive copper busbar 111 itself. In this embodiment, the spacing is selected as 10 mm, and the cross-section of the conductive copper busbar 111 is 6 mm × 30 mm.

[0061] In other embodiments, the conductor can also be a conductive aluminum busbar. Compared to the conductive copper busbar 111, the thermal conductivity of the conductive aluminum busbar is approximately 60% of that of the conductive copper busbar 111. While the amplitude of the temperature gradient signal is slightly lower at the same spacing, it is also more cost-effective, making it suitable for cost-sensitive applications. The sensor type is not limited to NTC thermistors; platinum resistance temperature sensors or thermocouples can also be used. Platinum resistance sensors offer better long-term stability than NTC thermistors, making them suitable for high-precision applications sensitive to long-term drift; thermocouples offer faster response times, making them suitable for applications requiring dynamic capture of temperature changes.

[0062] Multiple terminal blocks 110 are arranged in an array along the width of the multi-circuit smart meter 100, and a thermal insulation structure 116 is provided between adjacent terminal blocks 110. The thermal insulation structure 116 reduces the thermal coupling between adjacent circuits at a physical level. The thermal insulation structure 116 can be an air gap, a thermal insulation pad, or a ceramic partition. An air gap utilizes the low thermal conductivity of air to achieve thermal insulation without increasing additional material costs, but its thermal insulation effect is affected by airflow disturbances inside the distribution cabinet; thermal insulation pads and ceramic partitions provide more stable thermal insulation and are suitable for compact devices with extremely small spacing between terminal blocks 110.

[0063] First, temperature data acquisition is performed. The first temperature value output by the contact area temperature sensor 112 and the second temperature value output by the reference area temperature sensor 113 of each terminal block 110 are acquired. Taking the incoming terminal of circuit 1 in a 12-circuit smart meter as an example, at a certain sampling moment, the reading of the contact area temperature sensor 112 is 62.3℃, and the reading of the reference area temperature sensor 113 is 58.1℃.

[0064] The temperature gradient value is then calculated by measuring the difference between the first and second temperature values ​​for each terminal block 110. The temperature gradient value for the incoming terminal of circuit 1 is 62.3℃ − 58.1℃ = 4.2℃. The temperature gradient value reflects the intensity of the local heat source at the terminal contact surface 114. When the terminal contact resistance increases, additional Joule heat is concentrated at the contact surface, conducted along the conductive copper busbar 111 to both sides and gradually attenuating. The temperature rise of the sensor in the contact area is greater than that of the sensor in the reference area, and the temperature gradient value increases.

[0065] Gradient differential measurement can eliminate the interference of ambient temperature shifts and far-field thermal crosstalk on measurement results. Assuming the overall ambient temperature of the distribution cabinet rises from 25℃ to 30℃, the readings of the contact area sensor and the reference area sensor simultaneously rise by approximately 5℃, becoming 67.3℃ and 63.1℃ respectively, with a difference of 4.2℃. Assuming the adjacent circuit 2 is operating at full load, the heat generated by its terminal block 110 diffuses to the area of ​​circuit 1, causing the readings of the two sensors in circuit 1 to simultaneously rise by approximately 2℃, becoming 64.3℃ and 60.1℃ respectively, with a difference of 4.2℃. However, if the terminal contact resistance of circuit 1 increases, Joule heat is concentrated at the contact surface, causing the temperature of the contact area sensor to rise by 3℃ while the reference area sensor only rises by 0.8℃, changing the temperature gradient from 4.2℃ to 6.4℃. Thus, the temperature gradient value isolates the interference from the environment and thermal crosstalk, retaining only the thermal signal of the terminal contact surface 114 itself.

[0066] Furthermore, the current sampling values ​​of each circuit of the multi-circuit smart meter 100 are acquired. The multi-circuit smart meter 100 itself has energy metering functionality; its current sampling module continuously collects the real-time current values ​​of each circuit. These current values ​​can be directly used for diagnostic calculations without the need for an additional current transformer. Based on the ratio of the temperature gradient value of each terminal block 110 to the square of the corresponding circuit's current sampling value, a normalized temperature coefficient is obtained. According to Joule's law, the power dissipation at the contact surface is equal to the product of the contact resistance and the square of the current. Under thermal equilibrium, the temperature gradient value is approximately proportional to the power dissipation at the contact surface; therefore, the normalized temperature coefficient is approximately proportional to the contact resistance itself.

[0067] Continuing with the aforementioned values, the current sampling value for loop 1 is 30A, the temperature gradient is 4.2℃, and the normalized temperature coefficient is 4.2 divided by 900, approximately equal to 0.00467℃ / A². The current sampling value for loop 5 is 5A, the temperature gradient is 0.15℃, and the normalized temperature coefficient is 0.15 divided by 25, equal to 0.006℃ / A². The absolute value of the temperature gradient in loop 5 is only about 3.6% of that in loop 1. If judged solely based on the absolute magnitude of the temperature gradient, loop 5 would not attract any attention. However, the normalized coefficient for loop 5 is actually higher than that for loop 1, meaning that loop 5 generates more heat on the contact surface per unit current square, and its contact resistance may have already begun to deteriorate. This difference is completely undetectable in single-point temperature measurement and fixed threshold schemes.

[0068] In some embodiments, when the current sample value of the circuit is lower than a preset minimum current threshold, the calculation of the normalized temperature coefficient of the corresponding terminal block 110 is skipped. Circuit 9 is currently in an unloaded state, with the current close to zero. If a division operation is forcibly performed, the denominator approaching zero will cause the normalized temperature coefficient to diverge, resulting in extremely large coefficient values ​​from small temperature fluctuations. The minimum current threshold can be set to 0.5A. When the circuit current is lower than this threshold, the temperature gradient value of the corresponding terminal block 110 is compared longitudinally with the historical temperature gradient baseline value of the terminal block 110. The historical temperature gradient baseline value is the average of the temperature gradient values ​​of the terminal block 110 under the same low-load conditions over the past 30 days. When the temperature gradient value exceeds a preset multiple (e.g., 1.5 times) of the historical temperature gradient baseline value, the corresponding terminal block 110 is marked as having an abnormal contact state. The longitudinal comparison does not rely on the lateral comparison between circuits and can operate independently even when only a few circuits are in a low-load state.

[0069] After calculating the normalized temperature coefficient of each terminal block 110, the normalized temperature coefficients of multiple circuits in the same functional position are combined into a sample set. The same functional position refers to the terminal block 110 position in each circuit that performs the same electrical function. For example, in a 12-circuit smart meter, the incoming terminals of the 12 circuits constitute one sample set, and the outgoing terminals of the 12 circuits constitute another sample set. Since the terminal blocks 110 of each circuit within the same device use the same structural design, the same manufacturing process, and the same conductive materials, under ideal conditions where the contact is normal and thermal crosstalk interference has been fully eliminated, the normalized temperature coefficients of each circuit within the same sample set should have good consistency. Their dispersion mainly reflects manufacturing tolerances and installation differences.

[0070] Next, the statistical characteristics of the sample set are calculated. In some embodiments, the statistical characteristics include the mean and standard deviation of the sample set. After excluding loops that skip the normalization calculation due to current below the threshold, taking the input terminal sample set as an example, assuming there are 10 valid samples, the mean is 0.0050℃ / A², and the standard deviation is 0.0008℃ / A². The deviation threshold is the sum of the mean and N times the standard deviation, where N is a preset positive number. When N equals 2, the deviation threshold is 0.0050 plus 2 multiplied by 0.0008, which equals 0.0066℃ / A². This threshold dynamically changes with the actual distribution of the sample set—when the contact condition of most loop terminals is good and consistent, the standard deviation is small, the threshold is tightened, and the sensitivity is improved; when there is normal manufacturing variability between loops, the standard deviation is moderately increased, the threshold is relaxed, and reasonable differences are tolerated.

[0071] Before comparing the normalized temperature coefficients with statistical features, it is necessary to detect whether the sample set exhibits a systematic bias pattern related to the physical arrangement of the terminal blocks 110 in the array. If the normalized temperature coefficient sequence and the position numbers of the terminal blocks 110 show a statistical correlation—for example, the loop coefficients located in the middle of the array are systematically high while the loop coefficients located at the ends of the array are systematically low—this is not a fault signal of individual terminals, but rather a fingerprint of the residual effects of near-field asymmetric thermal crosstalk that has not been completely eliminated after gradient differencing. The detection method can be to perform correlation analysis on the normalized temperature coefficient sequence and the position numbers of the terminal blocks 110, and determine the existence of a systematic bias pattern when the absolute value of the correlation coefficient exceeds a preset correlation threshold.

[0072] In response to the detection of a systematic deviation pattern, a compensation correction is applied to the temperature gradient value, and the calculation of the normalized temperature coefficient and statistical characteristics is re-executed based on the compensated temperature gradient value until the systematic deviation pattern is eliminated or the preset iteration termination condition is met. This closed-loop mechanism enables the feedback of diagnostic results to drive the improvement of decoupling accuracy—if the decoupling is insufficient, statistical comparison will expose location-related deviation patterns, thereby triggering more accurate compensation correction.

[0073] In some embodiments, the compensation correction is implemented as follows. First, a thermal coupling coefficient matrix between loops is constructed. Since multiple sets of terminal blocks 110 are linearly arranged along the width direction, each terminal block 110 has significant thermal coupling only with its left and right adjacent terminal blocks 110, while the thermal coupling with non-adjacent terminal blocks 110 approaches zero due to distance attenuation. The structure of the thermal coupling coefficient matrix between loops corresponds to this physical characteristic and is a strip-shaped sparse matrix based on the topology of the terminal block 110 arrangement. Only physically adjacent terminal blocks 110 have non-zero coupling coefficients, and the coupling coefficients at non-adjacent positions are set to zero. Taking a 12-loop incoming terminal as an example, the matrix is ​​a 12th-order square matrix, with non-zero elements only on the main diagonal and the two adjacent sub-diagonals, and zero elements at all other positions. This sparse structure makes the computational cost of matrix operations linear rather than quadratic with the number of loops, making it suitable for real-time execution on meter-level microcontrollers.

[0074] The temperature gradient values ​​of each terminal block 110 are corrected a second time based on the inter-loop thermal coupling coefficient matrix. The input for correction is the temperature gradient value rather than the absolute temperature value—since the gradient difference has already completed a coarse decoupling to eliminate environmental offset and far-field thermal crosstalk, the matrix only needs to compensate for the residual near-field asymmetric crosstalk component. The correction magnitude is much smaller than the original crosstalk amount, and the numerical stability of the matrix is ​​higher than that of the full decoupling matrix that directly acts on the original temperature. Taking terminal block A as an example, terminal block A is located in the middle of the array. The temperature gradient value after gradient difference is 5.1℃, which includes a thermal crosstalk contribution of 0.3℃ from the adjacent terminal block B and a thermal crosstalk contribution of 0.4℃ from the terminal block C. After matrix correction, the compensated temperature gradient value is 4.4℃.

[0075] Based on the compensated temperature gradient values, the normalized temperature coefficient is recalculated, and the statistical features are recalculated. The correlation between the normalized temperature coefficient sequence and the position number in the sample set is re-detected. In this embodiment, after the first iteration, the correlation coefficient decreases from 0.7 to 0.3, and after the second iteration, it decreases to 0.05, which is lower than the preset convergence threshold of 0.1, confirming that the systematic bias pattern has been eliminated and the compensation correction is complete. If the correlation coefficient still does not decrease below the convergence threshold after three iterations, the preset iteration limit (e.g., 3 times) is reached, and the compensation correction is also confirmed to be complete, proceeding to the subsequent judgment steps. The iteration limit prevents infinite recursion when the matrix precision is insufficient, and in most cases, convergence can be achieved in one or two iterations.

[0076] The initial coupling coefficients of the matrix can be calibrated at the factory by applying a known heat load to each terminal block 110 individually and recording the temperature response of adjacent terminal blocks 110. In other embodiments, the initial coupling coefficients can also be learned gradually during the initial operation of the multi-loop smart meter 100 by observing natural load change events in each loop and following the online recalibration method described later.

[0077] During long-term operation, factors such as dust accumulation on the surface of terminal blocks 110, changes in the oxide layer, and aging of the thermal insulation structure 116 can alter the heat conduction path between adjacent terminal blocks 110, causing the factory-calibrated coupling coefficient to gradually deviate from the actual value. Therefore, an online recalibration step is included to address this issue. Online recalibration utilizes the statistical independence of load changes in each circuit of the multi-circuit smart meter 100 over time—within any given time period, the start and stop times of loads in different circuits are usually asynchronous, and there will always be periods where the load of one circuit changes significantly while the loads of the others remain stable.

[0078] Specifically, the current changes in each circuit are continuously monitored. In response to a single circuit in the multi-circuit smart meter 100 where the current change exceeds a preset percentage (e.g., 30%) of the rated current and lasts for more than a preset minimum duration (e.g., 3 minutes), while the current change rate of the remaining circuits is below a preset stability threshold (e.g., 0.1 A / s), that single circuit is marked as the excitation circuit. The dual thresholds of current change amplitude and duration eliminate interference from minor load fluctuations and brief current spikes in the calibration process—changes with insufficient amplitude produce thermal effects that are too weak to be distinguished from sensor noise; changes with insufficient duration have not yet produced a measurable steady-state response in the thermal domain.

[0079] Taking loop 1 as an example, at 10:00 AM one morning, the elevator started from standby mode. The current in loop 1 jumped from 0.5A to 30A, a change of approximately 49% of the rated current of 60A, exceeding the preset proportion of 30%. This continued for more than 5 minutes, exceeding the preset minimum duration of 3 minutes. Meanwhile, the current change rate of the other 11 loops was all below the stable threshold of 0.1A / s. Loop 1 is marked as the excitation loop.

[0080] Record the temperature gradient changes of each terminal block 110 in the excitation loop and the temperature gradient changes of each terminal block 110 physically adjacent to the excitation loop. The temperature gradient change of terminal block 110 in loop 1 is 3.8℃, and the temperature gradient change of the adjacent terminal block 110 in loop 2 is 0.15℃. Calculate the candidate coupling coefficient based on the ratio of these two values; 0.15 divided by 3.8 is approximately 0.039. Perform a consistency check between the candidate coupling coefficient and the corresponding coupling coefficient in the inter-loop thermal coupling coefficient matrix. The coupling coefficient from loop 1 to loop 2 in the current matrix is ​​0.042. The preset consistency range is ±30% centered on the current value, i.e., 0.029 to 0.055. The candidate value 0.039 falls within this range, and the coupling coefficient at the corresponding position is updated using the candidate coupling coefficient. If, during a calibration event, the contact state of adjacent terminal blocks 110 suddenly deteriorates, causing an abnormally large change in the temperature gradient value, the calculated candidate coefficient will be much larger than the current value and exceed the consistency range. In this case, the candidate coupling coefficient should be discarded to prevent the matrix from being contaminated by the abnormal event.

[0081] If compensation correction is completed or no systematic deviation pattern is detected, each normalized temperature coefficient is compared with a statistical characteristic. In response to a normalized temperature coefficient deviating from the statistical characteristic by more than a deviation threshold, the corresponding terminal block 110 is marked as a contact abnormality.

[0082] Based on the aforementioned values, the mean of the incoming terminal sample set is 0.0050℃ / A², and the deviation threshold is 0.0066℃ / A². The normalized temperature coefficient of loop 5 is 0.006℃ / A², which is below the deviation threshold and does not trigger an anomaly flag. The normalized temperature coefficient of loop 7 is 0.0085℃ / A², which exceeds the deviation threshold.

[0083] In some embodiments, a sliding window confirmation mechanism is applied to the determination of anomaly marking. Specifically, the corresponding terminal block 110 is marked as having a contact anomaly only when the normalized temperature coefficient deviates from the statistical characteristic value by more than K times within M consecutive sampling periods, where M and K are preset positive integers and K is less than or equal to M. M is set to 5 and K to 3, and the sampling period is 10 minutes. If the normalized temperature coefficient of circuit 7 exceeds the deviation threshold 4 times within 5 consecutive sampling periods, satisfying the condition that K is greater than or equal to 3, the incoming terminal of circuit 7 is marked as having a contact anomaly. The sliding window confirmation mechanism upgrades single-judgment to multiple confirmations; single deviations caused by transient load changes or sensor noise will not trigger anomaly marking, only persistent anomalies will be captured.

[0084] Furthermore, after anomaly marking, the normalized temperature coefficient of each terminal block 110 is recorded over multiple consecutive monitoring periods. In response to a monotonically increasing time series with a rate exceeding a preset trend threshold, a degradation warning is generated for the corresponding terminal block 110. The normalized temperature coefficient of the incoming terminal of loop 3 gradually increased from 0.0048℃ / A² to 0.0058℃ / A² over the past 30 days, never exceeding the deviation threshold of 0.0066℃ / A², and was not marked as abnormal in any horizontal comparison during each sampling period. However, the rate of increase of this coefficient was approximately 0.0003℃ / A² per month, exceeding the preset trend threshold of 0.0002℃ / A² per month, generating a degradation warning. The degradation warning detects terminal blocks 110 that have not yet triggered the deviation threshold but are gradually deteriorating—the oxide layer on the terminal contact surface 114 is slowly thickening, causing a slight increase in contact resistance each month. Without warning, this terminal may be close to the danger zone when it exceeds the threshold several months later. Anomaly marking and degradation early warning complement each other, covering both sudden failure and gradual degradation failure evolution modes.

[0085] In addition to the steady-state diagnostics based on the normalized temperature coefficient, this application also provides a dynamic diagnostic dimension based on the thermal time constant. The normalized temperature coefficient reflects the static characteristics of the contact resistance of terminal block 110 under thermal equilibrium conditions, and its accuracy depends on the assumption that the temperature gradient value has stabilized. In actual operation, the load on each circuit continuously changes, and the temperature gradient value always follows the load change, rarely reaching strict thermal equilibrium. The thermal time constant reflects the dynamic response speed of terminal block 110 to load changes, does not depend on the thermal equilibrium assumption, and, based on different physical mechanisms than the normalized temperature coefficient, provides mutually corroborating criteria for the same fault mode.

[0086] The thermal time constant is the physical meaning of the time required for the temperature gradient value of terminal block 110 to change from its initial value before the load change to 63.2% of its new steady-state value. The thermal time constant of a normal terminal block 110 depends on the heat capacity and heat dissipation area of ​​the conductive copper busbar 111. When the terminal contact resistance increases, heat is concentrated at the contact surface. The equivalent heat capacity of the contact surface area is smaller than that of the entire conductive copper busbar 111, causing the temperature gradient value to accumulate at the contact surface more quickly and tend towards a steady state, thus shortening the thermal time constant.

[0087] The dynamic diagnosis is performed as follows:

[0088] 1. Monitor the current change rate of each circuit. If the current change rate of a single circuit in the multi-circuit smart meter 100 exceeds a preset change rate threshold while the current change rates of the remaining circuits are below a preset stability threshold, mark the time period meeting the conditions as a diagnostic window and designate that single circuit as the excitation circuit. The conditions used for the diagnostic window are the same as those for online recalibration—both require "one circuit to change while the others remain stable." The same natural load change event can simultaneously serve two functions: online recalibration observes the thermal impact of the excitation circuit on adjacent terminal blocks 110 to update the matrix coupling coefficients; dynamic diagnosis observes the response characteristics of the non-excitation circuit's own terminal block 110 under thermal background changes to extract the thermal time constant.

[0089] 2. Within the diagnostic window, collect data on the temperature gradient values ​​of each terminal block 110 in each of the loops other than the excitation loop over time. Changes in the load of the excitation loop cause changes in the heat generated at its terminal block 110. A portion of this heat diffuses through heat conduction to the terminal block 110 areas of adjacent loops, resulting in slight changes in the temperature gradient values ​​of adjacent terminal blocks 110. Extract the thermal time constant from this change data; that is, the time required for the temperature gradient value to move from the initial point of change to 63.2% of its new steady-state value.

[0090] 3. Record multiple thermal time constants extracted from each terminal block 110 in multiple diagnostic windows corresponding to different excitation loops. Different diagnostic windows correspond to different excitation loops—the first diagnostic window may use loop 4 as the excitation source, the second diagnostic window may use loop 8 as the excitation source, and the third diagnostic window may use loop 1 as the excitation source. In each diagnostic window, the same terminal block 110 is observed under different thermal background conditions. If the thermal time constants extracted multiple times consistently point to an anomaly, the reliability of the conclusion is higher than that of a single observation result.

[0091] 4. Cross-compare multiple thermal time constants of the same terminal block 110. If a thermal time constant exceeding a preset proportion is lower than a preset time constant reference value, an increase in the contact resistance of the corresponding terminal block 110 is confirmed. The preset proportion can be set to 70%, meaning an anomaly is confirmed only when more than 70% of the thermal time constants extracted from the diagnostic windows are lower than the reference value. Cross-comparison eliminates misjudgments caused by single, occasional interference—random changes in environmental conditions within a single diagnostic window may cause the time constant to be lower, but if observations under multiple different conditions consistently indicate a shortening, random factors can be ruled out. The reliability of the diagnostic conclusion increases with the accumulation of diagnostic windows; multi-loop smart meters 100 with more loops have a higher probability of generating diagnostic windows, a statistical advantage that single-loop meters do not possess.

[0092] In some embodiments, the time constant reference value is determined based on the following method: During the initial operation phase of the multi-loop smart meter 100, the contact state of each terminal block 110 is in its factory state, and the contact resistance is at its lowest level. During the initial operation phase, the thermal time constant of each terminal block 110 is collected in multiple diagnostic windows, and the statistical mean of the thermal time constants of each terminal block 110 is used as the time constant reference value and stored. Using the statistical mean rather than the single measurement value as the reference eliminates random errors caused by load fluctuations and environmental disturbances in a single measurement. The reference value is continuously used in subsequent operations after being stored, enabling dynamic diagnostics to detect the degree of degradation relative to the factory state. In other embodiments, the method for extracting the time constant is not limited to exponential fitting, but can also be estimated by linear interpolation of the temperature gradient values ​​at specific time points (such as 30 seconds and 60 seconds) after a load change.

[0093] The multi-loop smart meter 100 includes multiple sets of terminal blocks 110, a current sampling module, and a processor at the hardware level. The multiple sets of terminal blocks 110 are arranged in an array along the width direction. Each set of terminal blocks 110 has a contact area temperature sensor 112 and a reference area temperature sensor 113 mounted on its conductive copper busbar 111. The current sampling module is configured to collect the current values ​​of each loop. The processor is configured to execute the aforementioned terminal contact state diagnostic method. Temperature data acquisition is accomplished by the contact area temperature sensor 112 and the reference area temperature sensor 113 converting analog temperature signals into digital values ​​via the processor's analog-to-digital converter interface. The current sampling values ​​required for normalization calculations are provided by the current sampling module, which serves both energy metering and diagnostic functions, eliminating the need for an additional current sensor for the diagnostic function. All computational logic, including gradient differential, normalization, statistical comparison, compensation correction iteration, sliding window confirmation, trend detection, thermal time constant extraction, and cross-comparison, is executed by the processor.

[0094] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0095] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0096] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for diagnosing the terminal contact status of a multi-loop smart meter, wherein the multi-loop smart meter includes multiple sets of terminal blocks, each set of terminal blocks being provided with a contact area temperature sensor and a reference area temperature sensor, characterized in that, Includes the following steps: S1. Collect the first temperature value output by the contact area temperature sensor and the second temperature value output by the reference area temperature sensor for each terminal block; S2. For each terminal block, calculate the difference between the first temperature value and the second temperature value to obtain the temperature gradient value; S3. Obtain the current sampling value of each circuit of the multi-circuit smart meter, and obtain the normalized temperature coefficient based on the ratio of the temperature gradient value of each terminal block to the square of the current sampling value of the corresponding circuit. S4. Form a sample set from the normalized temperature coefficients of the multiple loops that are in the same functional position, and calculate the statistical characteristics of the sample set; Detect whether the sample set exhibits a systematic deviation pattern related to the physical arrangement of the terminal blocks. In response to the systematic deviation pattern, apply compensation correction to the temperature gradient value and re-execute the calculation of S3 and the statistical feature based on the compensated temperature gradient value until the systematic deviation pattern is eliminated or the preset iteration termination condition is reached. S5. In response to the absence of the systematic deviation mode or the completion of the compensation correction, each normalized temperature coefficient is compared with the statistical feature quantity. In response to the normalized temperature coefficient deviating from the statistical feature quantity by more than the deviation threshold, the corresponding terminal block is marked as a contact abnormality state. The compensation correction in S4 includes the following sub-steps: S41. Construct a thermal coupling coefficient matrix between loops, wherein the thermal coupling coefficient matrix between loops is a strip-shaped sparse matrix based on the arrangement topology of the terminal blocks, and only physically adjacent terminal blocks have non-zero coupling coefficients. S42. Based on the thermal coupling coefficient matrix between the circuits, the temperature gradient value of each terminal block is compensated and corrected to obtain the compensated temperature gradient value; S43. Based on the compensated temperature gradient value, S3 is re-executed and the statistical characteristic quantity is recalculated; S44. In response to the deviation amplitude of the systematic deviation mode falling below a preset convergence threshold, the systematic deviation mode is confirmed to be eliminated, or in response to the preset iteration termination condition being met when the number of iterations reaches a preset upper limit, the compensation correction is confirmed to be completed.

2. The terminal contact status diagnosis method for multi-circuit smart meters according to claim 1, characterized in that, The terminal block includes a conductor, the contact area temperature sensor is disposed at the terminal contact surface of the conductor, and the reference area temperature sensor is disposed at a preset distance offset from the contact area temperature sensor along the extension direction of the conductor.

3. The terminal contact status diagnosis method for multi-circuit smart meters according to claim 2, characterized in that, The conductor is a conductive copper busbar. The contact area temperature sensor and the reference area temperature sensor are mounted on the conductive copper busbar. The reference area temperature sensor is located in the non-contact bearing area of ​​the conductive copper busbar. The distance between the contact area temperature sensor and the reference area temperature sensor along the extension direction of the conductive copper busbar is 5 mm to 20 mm. The multiple sets of terminal blocks are arranged in an array along the width direction of the multi-circuit smart meter, and a heat insulation structure is provided between adjacent terminal blocks.

4. The terminal contact status diagnosis method for multi-circuit smart meters according to claim 1, characterized in that, It also includes the step of online recalibrating the thermal coupling coefficient matrix between the loops: In response to a single circuit in the multi-circuit smart meter having a current change amplitude exceeding a preset proportion of the rated current and a duration exceeding a preset minimum duration, while the current change rate of the other circuits is lower than a preset stability threshold, the single circuit is marked as an excitation circuit. Candidate coupling coefficients are calculated based on the ratio between the temperature gradient change of each terminal block in the excitation loop and the temperature gradient change of each physically adjacent terminal block. In response to the candidate coupling coefficient falling within a preset consistency interval centered on the coupling coefficient at the corresponding position in the inter-loop thermal coupling coefficient matrix, the coupling coefficient at the corresponding position is updated with the candidate coupling coefficient; In response to the candidate coupling coefficient exceeding the consistency range, the candidate coupling coefficient is discarded.

5. The terminal contact status diagnosis method for a multi-circuit smart meter according to claim 1, characterized in that, It also includes the following steps: Monitor the current change rate of each circuit. In response to the current change rate of a single circuit in the multi-circuit smart meter exceeding a preset change rate threshold and the current change rate of the remaining circuits being lower than a preset stability threshold, mark the time period that meets the conditions as a diagnostic window and mark the single circuit as an excitation circuit. Within the diagnostic window, the temperature gradient values ​​of each terminal block in each of the other circuits besides the excitation circuit are collected over time, and the thermal time constant is extracted. Record the multiple thermal time constants extracted from each terminal block in multiple diagnostic windows corresponding to different excitation circuits; Cross-compare multiple thermal time constants of the same terminal block. If a thermal time constant exceeding a preset proportion is lower than a preset time constant reference value, it is confirmed that the contact resistance of the corresponding terminal block has increased.

6. The terminal contact status diagnosis method for a multi-circuit smart meter according to claim 5, characterized in that, The time constant reference value is determined based on the following method: During the initial operation phase of the multi-circuit smart meter, the thermal time constant of each terminal block in multiple diagnostic windows is collected, and the statistical mean of the thermal time constant of each terminal block is used as the reference value of the time constant and stored.

7. The terminal contact status diagnosis method for a multi-circuit smart meter according to claim 1, characterized in that, In S3, in response to the current sample value of the circuit being lower than a preset minimum current threshold, the calculation of the normalized temperature coefficient of the corresponding terminal block is skipped. The temperature gradient value of the corresponding terminal block is compared with the historical temperature gradient baseline value of the corresponding terminal block. In response to the temperature gradient value exceeding a preset multiple of the historical temperature gradient baseline value, the corresponding terminal block is marked as having an abnormal contact state.

8. The terminal contact status diagnosis method for a multi-circuit smart meter according to claim 1, characterized in that, The statistical features include the mean and standard deviation of the sample set, and the deviation threshold is the sum of the mean and N times the standard deviation, where N is a preset positive number. The determination condition for the normalized temperature coefficient deviating from the statistical feature quantity by more than the deviation threshold in S5 includes: the normalized temperature coefficient deviating from the statistical feature quantity by more than the deviation threshold by K times within M consecutive sampling periods, where M and K are preset positive integers and K is less than or equal to M. The step S5 further includes: recording the time series of the normalized temperature coefficient of each terminal block over multiple consecutive monitoring periods, and generating a deterioration warning for the corresponding terminal block in response to the time series showing a monotonically increasing trend and the rate of increase exceeding a preset trend threshold.

Images