Three-phase relay health assessment method based on multi-dimensional feature fusion and electric energy meter

Abstract

The application provides a three-phase relay health assessment method based on multi-dimensional fusion and an electric energy meter, fully reuses the high-speed data processing capability of the built-in metering chip and MCU of the electric energy meter, can synchronously complete the real-time accurate extraction of the arc energy integration operation / mechanical bounce characteristics in the millisecond short time window of the relay opening / closing action, introduces the thermal compensation coefficient and the thermal compensation factor, accurately quantifies the catalytic acceleration effect of the thermal degradation effect on the electrical ablation and mechanical impact damage, constructs a "thermal-mechanical-electrical" multi-dimensional fusion damage model, and effectively improves the accuracy and scientificity of the relay health state assessment. In addition, the independent and real-time health monitoring system of the single-phase relay is established according to the structural characteristics of the three-phase split type relay, effectively guarantees the long-term, stable and safe operation of the new split type built-in relay electric meter, and greatly improves the operation reliability and service life of the smart electric meter.

Description

Technical Field

[0001] This invention relates to the field of electricity meter testing technology, and in particular to a three-phase relay health assessment method based on multi-dimensional fusion and an electricity meter. Background Technology

[0002] Smart meters, with their built-in relays, perform critical functions such as remote control and anti-theft. Their operational reliability directly impacts the safe and stable operation of low-voltage power grids. In recent years, the load on low-voltage users across various regions has been increasing year by year, with a particularly significant increase during the high-temperature summer months. This has led to the development of smart meters with built-in switches and a maximum rated current of 100A. To enhance the switching performance of the built-in switches under high-current conditions and ensure the stable and reliable operation of the smart meters, several stringent performance evaluation indicators need to be set for the supporting built-in relays: (1) Apply 1.5 times the maximum current After 8 hours of continuous operation, the electricity meter was undamaged and the relay's opening and closing functions were normal. (2) Apply 2.5 times the maximum current for a short time After 3 seconds, the electricity meter was undamaged, and the relay could open and close normally. (3) Apply 1.2 times the maximum current. The circuit breaker was operated in cycles of 2 seconds on and 2 seconds off for at least 50 cycles. The energy meter did not reset or freeze during the entire test. After the test, the maximum operating current of the circuit was tested and the average impedance of the current circuit was not greater than 1.2mΩ. (4) The tripping and closing action of the relay shall not exceed 1 second.

[0003] High-current thermal stress, mechanical opening and closing impact, and arc erosion are the three main factors causing relay contact wear and failure. The above tests assessed the relay's ultimate load-bearing performance from three dimensions: resistance to continuous thermal aging, resistance to instantaneous high-current impact, and resistance to cyclic arc erosion.

[0004] Three-phase built-in energy meters can also adopt a separate architecture of three single-phase relays. While this design improves design flexibility, it also introduces new operational risks: the separate three-phase relays lack a mechanical linkage interlocking structure. During long-term operation, considering the uneven load on each phase relay, differences in heat dissipation environment, and individual material deviations, the aging rate of the relays will gradually diverge. If the inconsistent aging of the relays causes a deviation in the timing of opening and closing actions, an asynchronous arc superposition effect is likely to occur at the moment of disconnection. That is, if the arc of one phase is not completely extinguished, a new arc will be generated in another phase, which can easily cause interphase plasma conduction short circuits, significantly increasing the probability of contact welding failure.

[0005] As can be seen from the above, existing relay life testing methods have obvious technical shortcomings: the traditional method of relying on the statistics of the number of opening and closing cycles and the judgment of a single sensor signal cannot accurately identify the hidden faults of individual components; the contact capacitance detection and early warning scheme requires the addition of a dedicated detection circuit, and the built-in precision components are prone to insulation interference problems, and the hardware investment cost is relatively high; the computational load of deep learning evaluation algorithms such as GWO-BiLSTM is large and it is difficult to adapt to the real-time computing scenarios of the meter's main control MCU.

[0006] Therefore, there is an urgent need for a multi-dimensional feature fusion relay health status assessment solution for split three-phase relays that can directly reuse electrical energy, has hardware resources, requires no additional sensors, and can be built with a local MCU. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide a three-phase relay health assessment method and energy meter based on multi-dimensional fusion, reuse existing hardware resources, realize the comprehensive assessment of the health status of relays with multi-dimensional feature fusion at low cost, and effectively ensure the long-term reliable operation of split three-phase relays.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A three-phase relay health assessment method based on multi-dimensional fusion includes: S1: After receiving the opening / closing command, the MCU first configures the metering chip to high-speed sampling mode, and then sends the opening / closing command to the target relay. At the same time as sending the opening / closing command, the DMA is triggered to collect the instantaneous voltage values ​​on both sides of the target relay port and the instantaneous current values ​​in the target relay circuit at high speed. If it is an opening command, the start time of the acquisition and the arc extinction time will also be recorded. If it is a closing command, the start time of the acquisition and the closing completion time will also be recorded. S2: If it is a closing command, calculate the quantified mechanical impact damage degree of the target relay based on the DMA acquisition data corresponding to this closing command and then store it; If it is a tripping command, calculate the quantized arc erosion damage degree of the target relay based on the DMA acquisition data corresponding to this tripping command; and obtain the quantized mechanical impact damage degree calculated by the most recent closing command corresponding to the target relay, and the quantized thermal degradation loss degree calculated by the most recent thermal balance high current condition; execute S3. S3: The quantified arc ablation damage degree is used as an electrical loss feature, the quantified mechanical impact damage degree is used as a mechanical damage feature, and the quantified thermal degradation loss degree is used as a thermal degradation feature. These are input into the pre-constructed SOH fusion model, and the current health status value of the target relay is output. The expression for the SOH fusion model is: ; in, To quantify the degree of arc ablation damage as a characteristic of electrical loss; To quantify the degree of mechanical impact damage as a characteristic of mechanical damage; This is used to quantify the thermal degradation loss as a characteristic of thermal degradation. This is the thermal compensation coefficient; This is the thermal compensation factor; This represents the current health status value. This represents past health status. S4: Based on the current health status value of the target relay, determine whether to trigger the corresponding warning for the target relay.

[0009] Optionally, the calculation of the quantified arc ablation damage degree of the target relay in S2 includes: S2a1: Calculate the arc energy based on the arc energy integral formula; The formula for the integral of arc energy is: ; in, For the energy of the combustion arc; The instantaneous voltage difference across the target relay port; The instantaneous current value of the target relay circuit; The sampling time interval; This is the starting time point for data collection; This refers to the time point when the arc is extinguished. S2a2: Calculate the quantitative arc ablation damage degree based on the quantitative arc ablation damage degree calculation formula; The formula for calculating the quantitative arc ablation damage degree is as follows: ; in, This is the electrical damage weighting coefficient; This is the nonlinear damage coefficient.

[0010] Optionally, the calibration of the electrical damage weighting coefficient and the nonlinear damage coefficient includes: Several relay samples from the same batch were selected and grouped together and different gradient currents were applied. Under constant temperature and the same rated voltage conditions, opening and closing commands are cyclically sent to each relay sample until the relay fails. Record the total number of failures for each group of relay samples; Calculate the arc energy corresponding to a single trip command for each relay sample, and statistically obtain the total arc energy for each group of relay samples. The current health status value when the relay fails The total number of failures is the dependent variable, and the total arcing energy is the independent variable. A logarithmic transformation is performed on the simplified SOH fusion model of electrical losses, followed by linear regression, to obtain the calibrated electrical damage weighting coefficient and nonlinear damage coefficient. The simplified SOH fusion model of electrical losses is as follows: .

[0011] Optionally, calculating the quantified mechanical impact damage degree of the target relay in S2 includes: S2b1: Calculate the voltage difference between the two sides of the target relay port based on the instantaneous voltage values ​​on both sides of the target relay port in the DMA acquisition data corresponding to this closing command. S2b2: Mark the voltage instantaneous value difference that is greater than the preset voltage threshold; S2b3: Determine the jump time point corresponding to the difference between each of the markers and its adjacent unmarked voltage instantaneous value; S2b4: If the time interval between two adjacent switching time points is less than a preset first time interval, and the time interval between the two switching time points and the closing time point is less than a preset second time interval, then the cumulative mechanical bounce value is incremented by one. S2b5: The quantitative mechanical impact damage degree is calculated based on the cumulative mechanical bounce value using the quantitative mechanical impact damage degree formula. The formula for quantifying mechanical impact damage is: ; in, The mechanical damage coefficient; This represents the cumulative value of mechanical bounce.

[0012] Optionally, the calibration of the mechanical damage coefficient includes: Select relay samples from the same batch, and under constant temperature, rated voltage and no-load conditions, cyclically send opening and closing commands to the relay samples until the relay fails. Record the cumulative mechanical bounce value of the relay sample for each closing command; Based on the simplified SOH fusion model of mechanical impact damage and the total number of closing commands, the calibrated mechanical damage coefficient is calculated using the mechanical damage coefficient formula. The simplified SOH fusion model for mechanical impact damage is as follows: The formula for the mechanical damage coefficient is: .

[0013] Optionally, the target relay calculates the quantified thermal degradation loss under thermal equilibrium high current conditions, including: S2c1: Under thermal equilibrium high current conditions, the MCU calculates the measured contact resistance of each relay at the current moment according to the dynamic contact resistance calculation formula at a preset time interval. The formula for calculating the dynamic contact resistance is as follows: ; in, This represents the measured contact resistance at the current moment. This represents the instantaneous voltage value at the relay input port at the current moment; This represents the instantaneous voltage value at the relay output port at the current moment. This represents the instantaneous current value of the relay circuit at the current moment; S2c2: Calculate the gain coefficient at the current moment based on the gain coefficient formula; The formula for the gain coefficient is: ; in, The covariance of the previous time step; This represents the instantaneous current value of the target relay circuit at the current moment; is the forgetting factor; n is the current time. S2c3: Calculate the prediction error at the current moment based on the prediction error formula; The prediction error formula is: ; in, This represents the prediction error at the current moment; This represents the instantaneous voltage difference between the two sides of the target relay port at the current moment. This is the contact resistance value at the previous moment; S2c4: Calculate the contact resistance value at the current moment based on the contact resistance value calculation formula; The formula for calculating the contact resistance value is: ; in, This represents the contact resistance value at the current moment; S2c5: Calculate the covariance at the current time based on the covariance formula; The covariance formula is: ; S2c6: The MCU executes steps S2c1 to S2c5 repeatedly at preset time intervals, iteratively calculates the contact resistance value at the current moment, and marks it as the smooth contact resistance value at the current moment. S2c7: Calculate the quantitative thermal degradation loss based on the quantitative thermal degradation loss calculation formula; The formula for calculating the quantified thermal degradation loss is as follows: ; in, This represents the smooth contact resistance value at the current moment; This represents the initial contact resistance value.

[0014] Optionally, the calibration of the thermal compensation coefficient includes: Several relay samples from the same batch were selected and divided into a control group and a thermal degradation group; A large current is first applied to the thermally degraded group until its contact resistance rises to a preset resistance value; then, under constant temperature environment and the same rated voltage conditions, opening and closing commands are cyclically issued to each relay sample of the control group and the thermally degraded group until the relay fails. The quantitative thermal degradation loss of each relay sample in the control group and the thermal degradation group was calculated respectively. Based on the simplified SOH fusion model of thermal degradation loss and the quantified thermal degradation loss degree of each relay sample, the thermal compensation coefficient corresponding to each relay sample is calculated. The simplified SOH fusion model for thermal degradation loss is as follows: ; The thermal compensation coefficients corresponding to each relay sample are: ; The thermal compensation coefficients of all relay samples were fitted to obtain the calibrated thermal compensation coefficients.

[0015] Optionally, it also includes: S5: Verify the SOH fusion model; S5 specifically includes: S51: Select several representative sets of DMA acquisition data corresponding to closing and opening commands, and calculate the corresponding quantitative arc ablation damage degree and quantitative mechanical impact damage degree. S52: Obtain the smooth contact resistance value calculated by the relay corresponding to the closing command and opening command in S51 under the thermal balance high current condition, and calculate the quantified thermal degradation loss degree based on the smooth contact resistance value. S53: Based on the quantitative arc ablation damage degree and quantitative mechanical impact damage degree calculated in S51, and the quantitative thermal degradation loss degree calculated in S52, construct a feature dataset including electrical loss characteristics, mechanical damage characteristics and thermal degradation characteristics. S54: Input the feature dataset into the SOH fusion model to obtain the predicted SOH health level value; S55: Obtain the measured contact resistance value of the relays corresponding to the closing and opening commands mentioned in S51 under thermal balance high current conditions, and calculate the true SOH health value at time i using the true SOH health value calculation formula. The formula for calculating the actual SOH health level value is as follows: ; in, Let be the measured contact resistance value at time i; This represents the initial contact resistance value. This is the preset failure contact resistance value; S56: Calculate the fitness between the actual SOH health status value and the predicted SOH health status value output by the SOH fusion model based on the fitness function; The fitness function is: ; in, Let i be the predicted SOH health status value at time i; Let i be the actual SOH health status value at time i; M be the time series length; S57: Initialize a preset number of particles, and randomly assign a set of global parameters to each particle. ; S58: For each particle, using its global parameters, perform iterative calculations based on the feature dataset and the SOH fusion model to obtain the fitness set of each particle corresponding to the time series length M; S59: Based on the fitness set of each particle, obtain the optimal global parameters of each particle and the optimal global parameters of the population corresponding to all particles. S510: For each particle, adjust the corresponding global parameters based on its own optimal global parameters and the population's optimal global parameters; S511: Iterate through S58 to S510 until the preset number of iterations is reached, obtain the current optimal global parameters of the population, and mark them as the calibrated global parameters.

[0016] Optionally, S4 specifically includes: If the current health level value is higher than or equal to the health level threshold, the target relay is determined to be in good working condition and has a safe and stable disconnection capability. If the current health level value is higher than or equal to the warning level threshold, but lower than the health level threshold, the target relay is determined to have entered the mid-term degradation stage and is in a warning state. The frequency of high current surge operations is limited, and a status prompt is triggered. If the current health status value is lower than the warning level threshold, the target relay is determined to have entered the late degradation stage and is in a dangerous state. High current surge operations are prohibited, and an emergency alarm is triggered.

[0017] Another technical solution provided by this invention is: An electricity meter includes an MCU, a metering chip, and a non-volatile memory; the non-volatile memory stores a computer program; the computer program, after being processed by the MCU, can implement the above-mentioned three-phase relay health assessment method based on multi-dimensional fusion.

[0018] The beneficial effects of this invention are as follows: This invention requires no additional hardware, fully utilizing the high-speed data processing capabilities of the built-in metering chip and MCU of the electricity meter. Within a millisecond-level short time window of the relay's opening / closing action, it can simultaneously complete the real-time and accurate extraction of arc energy integration calculations and mechanical bounce characteristics, constructing an embedded "sensing-analysis-decision" integrated closed-loop monitoring mechanism. Simultaneously, by introducing thermal compensation coefficients and thermal compensation factors, the catalytic acceleration effect of thermal degradation on electrical ablation and mechanical impact damage is accurately quantified. The constructed "thermal-mechanical-electrical" multi-dimensional fusion damage model highly closely matches the actual physical process of relay material failure, effectively improving the accuracy and scientific rigor of relay health status assessment. Furthermore, considering the structural characteristics of three-phase split relays, this invention establishes an independent and real-time health monitoring system for single-phase relays. This system can accurately identify potential timing discrepancies caused by inconsistent aging levels in each phase relay, avoid severe faults such as phase-to-phase arcing and phase-to-phase short circuits, and completely solve the technical shortcomings of traditional overall monitoring methods that cannot distinguish the differentiated degradation of single-phase relays. This effectively ensures the long-term, stable, and safe operation of the new split-type built-in relay meter, significantly improving the operational reliability and service life of the smart meter. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the system hardware structure of the three-phase meter described in this embodiment of the invention; Figure 2 A schematic diagram of the basic process of a three-phase relay health assessment method based on multi-dimensional fusion provided for an embodiment of the present invention; Figure 3 This is a schematic diagram showing the relationship between the electrical loss characteristics of the DMA-acquired data and the sampling time in a specific embodiment of the present invention; Figure 4 This is a schematic diagram showing the relationship between the mechanical damage characteristics and sampling time of the DMA acquisition data described in a specific embodiment of the present invention. Figure 5 This is a schematic diagram illustrating the process of obtaining the quantified thermal degradation loss degree of the target relay under thermal equilibrium high current conditions, as described in a specific embodiment of the present invention. Detailed Implementation

[0020] To explain in detail the technical principles, specific implementable solutions, possible application scenarios, and achievable objectives and effects of the present invention, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. It is understood that the embodiments described herein and the embodiments shown in the accompanying drawings are only used to more clearly illustrate the technical solutions of the present invention, and are therefore only examples intended to explain the present invention, and should not be construed as limiting the present invention. It should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0021] Explanation of technical terms involved in this invention: DMA acquisition: Data is transferred directly without CPU. DMA automatically stores the data acquired by the ADC into memory without occupying the CPU. It has an extremely fast sampling speed and no missed points. It is a hardware automatic transfer.

[0022] Quantitative arc erosion damage: a physical quantity characterizing the material transfer, ablation, and surface morphology deterioration of contacts caused by the thermal effect and electrical erosion generated by arcing between contacts during the relay opening process.

[0023] Quantitative mechanical shock damage: characterizes the degree of physical damage caused by mechanical collision, bouncing, and mechanical fatigue of the moving and stationary contacts during the closing process of the relay.

[0024] Quantitative thermal degradation loss: a physical quantity characterizing the oxidation of relay contact material and thermal stress relaxation of spring material under long-term energized temperature rise and alternating ambient temperature.

[0025] Thermal compensation coefficient: a physical constant characterizing the resistivity of relay contact material as a function of temperature, or a corrective balance constant used to characterize the rate of temperature rise evolution of contacts under different rated current loads.

[0026] Thermal compensation factor: A dimensionless coefficient used to correct the current rates of electrical ablation and mechanical damage, reflecting the nonlinear accelerating catalytic effect of thermal degradation on traditional electromechanical damage.

[0027] Electrical damage weighting coefficient: a physical scale factor characterizing the metal ablation of contact materials and the health of the relay under a unit reference arc energy, i.e., a dimension used to correct the health of the relay and the arc energy.

[0028] Nonlinear damage coefficient: An exponential characteristic constant used to correct the nonlinear ablation acceleration effect caused by changes in arc energy density, concentration of microscopic heat flow, and the evolution of geometric irregularities on the contact surface.

[0029] Mechanical damage coefficient: characterizes the baseline fatigue damage caused to the contact surface, surface coating, and mechanical spring by each standard mechanical collision and bounce of the moving and stationary contacts.

[0030] This invention provides a multi-dimensional fusion-based health assessment method for three-phase relays built into three-phase electricity meters. It fully reuses the existing metering chip and main control chip (MCU), without relying on additional sensors or requiring any additional hardware. Figure 1 As shown, the three-phase meter includes a main control chip (MCU), a metering chip, and a non-volatile memory; the metering chip and the non-volatile memory are respectively communicatively connected to the main control chip (MCU).

[0031] In some specific embodiments, the system hardware structure of the three-phase meter is as follows: Figure 1 As shown, it includes a metering chip and a main control chip (MCU); it also includes a dual-mode communication module connected to the main control chip (MCU), as well as a voltage divider network, a manganese-copper voltage divider network, and a contact resistance sampling network, which are respectively connected to the metering chip.

[0032] The main control chip MCU, voltage divider network, manganese-copper voltage divider network and contact resistance sampling network constitute a metering module, which is used to collect instantaneous voltage and current of the line and transmit data to the main control chip MCU at high speed through the SPI interface.

[0033] The main control chip MCU is equipped with a DMA channel, which allows the sampled data to be directly written to the buffer of non-volatile memory, interrupt-triggered feature extraction calculation, and real-time evaluation of the relay health status.

[0034] The non-volatile memory, which can be EEPROM or off-chip Flash, supports power-down retention and is used to update and record the quantified arc erosion damage, quantified mechanical impact damage, quantified thermal degradation loss, current health status, and historical health status after each relay operation. Preferably, a corresponding health record is set for each phase of the split-type built-in relay to store the above-mentioned data.

[0035] The dual-mode communication module integrates broadband carrier HPLC and broadband low-power wireless HRF; it is used for electricity meters to connect to the concentrator or remote electricity information collection system via a "wired + wireless" dual-channel network communication, and has the function of actively reporting data.

[0036] Please see Figures 1 to 5 This invention provides a method for health assessment of three-phase relays based on multi-dimensional fusion. The subject executing the method can be, for example, […]. Figure 1 The three-phase electricity meter shown. Combined with... Figure 2 To understand this, the method specifically includes: S1: After receiving the tripping or closing command, the MCU first configures the metering chip to high-speed sampling mode, and then sends the command to the target relay. At the same time as sending the tripping or closing command, the DMA will be triggered to collect the instantaneous voltage value on both sides of the target relay port and the instantaneous current value of the target relay circuit at high speed. If it is a tripping command, the start time point and the arc extinguishing time point will also be recorded. If it is a closing command, the start time point and the closing completion time point will also be recorded.

[0037] For ease of description, the term "opening command or closing command" will be simplified to "opening / closing command" in the following text.

[0038] The opening / closing command can be issued by the host computer and received by the main control chip MCU of the energy meter through the dual-mode communication module HPLC; it can also be issued by the main control chip MCU based on its own logic; or it can be output from the local measurement and control unit to the main control chip MCU of the energy meter.

[0039] For example, the metering chip with model number HT7032L has a maximum sampling rate of 1.84MHz. Assuming the sampling power grid frequency is 50Hz, the number of sampling points in one signal cycle is approximately 36,800.

[0040] In this embodiment, the metering chip is configured to high-speed sampling mode before the opening / closing command is issued, and DMA is synchronously started to acquire target data at high speed the instant the opening / closing command is issued. This acquisition method does not require CPU participation and can accurately and completely capture the transient voltage and current characteristic data of closing, ensuring the real-time performance of sampling and data integrity, and providing a reliable basis for subsequent accurate evaluation of relay closing performance.

[0041] In this embodiment, the built-in relay of the three-phase meter can be a separate built-in relay, that is, a combination of three separate single-phase relays. Under this architecture, the closing command corresponds to a single-phase relay. Of course, the built-in relay of the three-phase meter can also be an integrated built-in relay. Under this architecture, the opening / closing command corresponds to the entire three-phase relay.

[0042] For a single trip command, the corresponding DMA acquisition data specifically includes: all instantaneous voltage values ​​on both sides of the target relay port and all instantaneous current values ​​in the target relay circuit from the acquisition start time to the arc extinction time. The instantaneous voltage values ​​on both sides of the relay port refer to the instantaneous voltage values ​​at the relay input and output terminals.

[0043] For a single closing command, the corresponding DMA acquisition data specifically includes: all instantaneous voltage values ​​on both sides of the target relay port and all instantaneous current values ​​in the target relay circuit during the period from the acquisition start time to the closing completion time (i.e., the time when the target relay contacts close).

[0044] In some specific implementations, after the MCU records the start time and arc-extinguishing time of the current tripping command (or the start time and closing time of the current closing command), it immediately controls the metering chip and DMA to stop high-speed sampling in order to reduce system power consumption and release hardware resources.

[0045] S2: If the MCU receives a closing command, it calculates and stores the quantified mechanical impact damage degree of the target relay based on the DMA acquisition data corresponding to this closing command. If the MCU receives a tripping command, it calculates the quantized arc erosion damage degree of the target relay based on the DMA acquisition data corresponding to this tripping command; and obtains the quantized mechanical impact damage degree calculated for the most recent closing command of the target relay, as well as the quantized thermal degradation loss degree calculated for the most recent thermal balance high current condition; then executes S3.

[0046] The specific calculation methods for the quantitative arc ablation damage, quantitative mechanical impact damage, and quantitative thermal degradation loss are as follows: (1) Quantitative analysis of arc ablation damage degree and electrical loss characteristics: like Figure 3 The figure shows a schematic diagram illustrating the relationship between the electrical loss characteristics of the DMA-acquired data and the sampling time. The MCU will process the data acquired by the DMA in real time and record the start time of the current trip command acquisition. and the time of arc extinguishing Based on this, all instantaneous voltage and current values ​​during the period are obtained and marked as DMA acquisition data corresponding to this tripping command.

[0047] Wherein, the arc extinguishing time point The criterion for judgment is: during the process of waiting for the target relay contacts to separate, if continuously... If the instantaneous current value of the target relay circuit sampled within the time period is lower than the preset current threshold (e.g., 0.5A), the arc is determined to be extinguished, and the corresponding time point is recorded as the arc extinguishing time point. .

[0048] In some specific implementations, the arc extinction determination time The selectable range is 20μs to 50μs; preferably 30μs. This is because the average arcing time is generally between 2ms and 9ms. Considering electromagnetic interference, a viewing window of 20μs to 50μs is specifically set to ensure that the current does not temporarily drop due to noise.

[0049] Quantifying arc ablation damage The specific calculation process includes the following steps: S2a1: Calculate the arc energy according to the following arc energy integral formula. ; The formula for the integral of arc energy is: ; in, It is the difference between the instantaneous voltage values ​​on both sides of the target relay port, that is, the difference between the instantaneous voltage value at the input terminal of the target relay and the instantaneous voltage value at its output terminal. The instantaneous current value of the target relay circuit. The sampling time interval, i.e., the start time of data collection. Arrive at the time of arc extinction The time interval between; This is the starting time point for data collection; This is the time point when the arc is extinguished.

[0050] Here, through the analysis of and Performing absolute value processing ensures that the energy during reverse arc discharge (during contact rebound) is also included, thus avoiding underestimation of the total arc erosion.

[0051] S2a2: Calculate the quantitative arc ablation damage degree according to the following formula. ; The formula for calculating the quantitative arc ablation damage degree is as follows: ; in, This is the electrical damage weighting coefficient, used to correct the dimensions of relay health and arcing energy. It is a nonlinear damage coefficient used to amplify the weight of high-current arcing on relay life loss, thereby achieving adaptive evaluation.

[0052] (2) Quantitative mechanical impact damage degree - mechanical damage characteristics: like Figure 4 The figure shows a schematic diagram of the relationship between the mechanical damage characteristics and sampling time of DMA acquired data.

[0053] Quantifying mechanical impact damage The specific calculation process includes the following steps: S2b1: Calculate the voltage difference between the two sides of the target relay port based on the instantaneous voltage values ​​on both sides of the target relay port in the DMA acquisition data collected in the corresponding closing command.

[0054] The MCU will process the data acquired by DMA in real time, record the start time of acquisition and the completion time of this closing command, and obtain all instantaneous voltage and current values ​​during this period, marking them as the DMA acquisition data corresponding to this closing command.

[0055] S2b2: For voltages exceeding a preset threshold The voltage instantaneous value difference Mark it.

[0056] Optionally, voltage threshold Set to 10% of the rated phase voltage.

[0057] In some specific implementations, this step specifically employs binarization processing: Cluster the instantaneous voltage difference values The value is mapped to state "1", which means there is power; otherwise, it is mapped to state "0", which means there is no power.

[0058] S2b3: Determine the jump time point corresponding to the difference between each of the marked and its adjacent unmarked voltage instantaneous value.

[0059] In some of the specific implementations described above, this step specifically employs edge detection processing: Identify the transition edges from state "0" to state "1" in the instantaneous voltage difference cluster after mapping processing, calculate the time when each transition edge occurs, and mark it as the transition time point.

[0060] S2b4: If the time interval between two adjacent switching time points is less than a preset first time interval, and the time interval between the two switching time points and the closing time point is less than a preset second time interval, then the cumulative mechanical bounce value is incremented by one.

[0061] In some of the specific implementations described above, this step specifically employs bounce filtering: If the time interval between two consecutive transitions from state "0" to state "1" is less than a preset first time interval threshold. (Optional, default 1ms), and the time from the closing completion point is less than the preset second time interval. (Optional, default 20ms), then it is determined to be a mechanical bounce, and the cumulative value of the mechanical bounce is calculated. Increment the count by 1; otherwise, treat it as noise.

[0062] Under 50Hz power frequency conditions, the relay tripping and closing actions can typically be completed within one power frequency cycle (20ms), i.e., 1 / 50Hz = 20ms. Considering the impact of arcing, performing the tripping and closing at the voltage zero crossing point can minimize equipment losses. Under this condition, the longest allowable operating time is half a power frequency cycle, i.e., 10ms. Based on the above characteristics, the lower limit of the designed tripping and closing action time is 10ms, i.e. The minimum window time is 10ms, preferably 20ms; appropriately extending the window time can increase the effective sampling data and improve the accuracy of the detection results. Relay contact bounce mainly occurs in the initial stage of action, with the total number of bounces generally not exceeding 8. Statistically, the average interval between a single bounce is 1.25ms to 2.5ms. To reliably identify contact transition behavior, the time interval for transition determination must be less than 1.25ms. Ultimately, the first time interval threshold is set... Set to 1ms.

[0063] In some of the specific embodiments described above, the following are also included: Bouncing end determination: Exceeding the preset second time interval If no new jumps are added, the jump count is considered complete.

[0064] Mechanical bounce cumulative value This directly reflects the degree of surface roughness of the relay contacts and the fatigue of the contact springs. It represents the cumulative mechanical bounce during the closing process of a brand-new relay. Typically 0-2 times; as the contact gradually degenerates, the cumulative mechanical bounce value increases. It can be increased to 3-6 times; when the contact is close to failure, the cumulative mechanical bounce value... It can be done more than 8 times.

[0065] S2b5: The quantitative mechanical impact damage degree is calculated based on the cumulative mechanical bounce value using the following formula. ; The formula for quantifying mechanical impact damage is: ; in, The mechanical damage coefficient; This represents the cumulative value of mechanical bounce.

[0066] (3) Quantification of thermal degradation loss degree - thermal degradation characteristics: The quantified thermal degradation loss It was calculated using dynamic contact resistance analysis.

[0067] like Figure 5 As shown, the quantized thermal degradation loss degree The specific calculation process includes the following steps: S2c1: When the relay is under thermal equilibrium high current operation, the MCU calculates the dynamic contact resistance corresponding to the voltage drop across the relay according to a preset time interval (e.g., every minute) using the following dynamic contact resistance calculation formula. Also known as measured contact resistance.

[0068] The aforementioned thermal balance high current operating condition refers to a situation where the relay is in the conducting state, and the load current (also known as the operating current) reaches a preset proportion (such as 70%) of the rated current (generally 120A) and continues to operate for more than a preset time (such as 10 minutes). This can be considered as the relay being in the thermal balance high current operating condition.

[0069] The formula for calculating the dynamic contact resistance is as follows: ; in, This represents the measured contact resistance at the current moment. This represents the instantaneous voltage value at the relay input port at the current moment; This represents the instantaneous voltage value at the relay output port at the current moment; in the formula... and These are sampled values ​​within the same epoch; This represents the instantaneous current value of the relay circuit at the current moment.

[0070] Since fluctuations in the mains voltage can cause random errors in the relay input port voltage, this embodiment uses the recursive least squares (RLS) method to calculate the smooth contact resistance. Specifically, the steps include the following: S2c2: Initialization settings, setting the initial contact resistance value. Initial covariance and forgetting factor .

[0071] Among them, the initial covariance This can be any large positive number, representing the uncertainty about the initial value. The larger the value, the faster the algorithm converges in the early stages. Forgetting factor The selectable value range is 0.95-0.999; the larger the value, the longer the algorithm remembers the old data and the better the smoothing effect; the smaller the value, the easier it is for the algorithm to forget the old data and the more sensitive it is to noise.

[0072] Calculate the gain coefficient at the current moment using the following gain coefficient formula. ; The formula for the gain coefficient is: ; in, The covariance of the previous time step; This represents the instantaneous current value of the target relay circuit at the current moment; is the forgetting factor; n is the current time.

[0073] S2c3: Calculate the prediction error at the current time based on the following prediction error formula. ; The prediction error formula is: ; in, This represents the instantaneous voltage difference between the two sides of the target relay port at the current moment. This represents the instantaneous current value of the target relay circuit at the current moment; The contact resistance value at the previous moment is calculated using the dynamic contact resistance calculation formula.

[0074] S2c4: Calculate the contact resistance value at the current moment according to the following contact resistance calculation formula. ; The formula for calculating the contact resistance value is: .

[0075] S2c5: Calculate the covariance at the current time using the following covariance formula. ; The covariance formula is: .

[0076] S2c6: For each sampling point acquired at a preset time interval, the MCU repeats steps S2c1 to S2c5 above to iteratively calculate the contact resistance value at the current moment (i.e., the latest time point). And mark it as the smooth contact resistance value at the current moment. .

[0077] In some specific implementations, such as Figure 5 As shown, outliers in the sampled data will be filtered using the following inequality before step S2c2: ; in, The preset resistance value change threshold can be set to 0.5.

[0078] If the above inequality is satisfied, the corresponding sampling point is determined to be an impulse interference from the power grid and is filtered out.

[0079] S2c7: Calculate the quantized thermal degradation loss according to the following formula. ; The formula for calculating the quantified thermal degradation loss is as follows: ; in, This represents the smooth contact resistance value at the current moment; This represents the initial contact resistance value.

[0080] Based on the current smooth contact resistance value and the initial contact resistance value That is, the quantified thermal degradation loss degree This allows calculation of the relay contact resistance change rate and temperature rise consistency under the most recent high-current thermal equilibrium condition. A sudden increase in the contact resistance change rate indicates accelerated oxidation of the relay contacts, significantly increasing the risk of contact welding. The MCU should issue a timely warning signal to avoid permanent welding failures caused by high-current surges.

[0081] S3: Quantify the arc ablation damage degree As an electrical loss characteristic, the quantified mechanical impact damage degree As a characteristic of mechanical damage, the quantified thermal degradation loss degree As a thermal degradation feature, it is input into a pre-built SOH fusion model, and the output is the current health status value of the target relay.

[0082] The SOH fusion model is constructed based on the above three dimensions of features and is used to measure the health status of relays. The model expression is as follows: ; in, To quantify the degree of arc ablation damage as a characteristic of electrical loss; To quantify the degree of mechanical impact damage as a characteristic of mechanical damage; This is used to quantify the thermal degradation loss as a characteristic of thermal degradation. This is the thermal compensation coefficient; This is the thermal compensation factor; This represents the current health status value. This represents the past health status value; initially, it can be set to 1.

[0083] It is understandable that an increase in ambient temperature will lead to a decrease in the smoothing contact resistance value. Increase. Regarding arc energy And mechanical bounce cumulative value For some relays, high-temperature environments can exacerbate device damage. This embodiment introduces a thermal compensation factor, which may synergistically correct and amplify electrical and mechanical damage, improving the matching degree between the SOH fusion model and actual physical conditions.

[0084] S4: Based on the current health status value of the target relay, determine whether to trigger the corresponding warning for the target relay.

[0085] In some specific implementations, after each opening / closing command operation is completed, the MCU will update the health record (quantifying the degree of arc erosion damage) of each phase relay in the memory. Quantification of mechanical impact damage Quantifying thermal degradation loss and current health status value ), and then carry out the following health level classification and treatment: Health Level: Current health status value ≥Health level threshold, the target relay is in good working condition and has safe and stable disconnection capability; Warning level: Warning level threshold ≤ current health value When the health threshold is reached, the target relay enters the mid-term degradation stage and is in a warning state. The frequency of high current surge operations is limited, and a status prompt is triggered to remind maintenance personnel to check and maintain it. Danger Level: Current health status value <Warning level threshold: The target relay has entered the late degradation stage and is in a dangerous state. High current surge operations are prohibited. An emergency alarm is triggered and actively reported to remind maintenance personnel to replace it in time.

[0086] For example, the health level threshold is set to 70%; the warning level threshold is set to 40%.

[0087] Another embodiment of the present invention is a further extension of the above embodiments, and the specific calibration method of multidimensional feature parameters is described in detail.

[0088] The SOH fusion model contains multiple variables with complex coupling relationships. To achieve accurate calibration of the model parameters, this embodiment employs a combined calibration method of control variable method and multiple regression analysis to solve for and calibrate the parameters. Specifically, the steps are as follows: 1) Initialization: Define the initial health level value =1.0; Define the relay failure criteria as follows: 1. Smooth contact resistance value ≥ Failure contact resistance value Or the current health status value ≤ Failure level threshold (e.g., 30%) 2. No normal movement.

[0089] 2) Mechanical damage coefficient Calibration: 2-1. Select relay samples from the same batch, and under constant temperature, rated voltage and no-load conditions, cyclically send opening and closing commands to the relay samples until the relay fails. 2-2. Record the cumulative mechanical bounce value of the relay sample for each closing command. ; 2-3. Based on the simplified SOH fusion model of mechanical impact damage and the total number of closing commands, the calibrated mechanical damage coefficient is calculated using the mechanical damage coefficient formula; Here, since electrical ablation and thermal degradation interference are not considered, and only mechanical damage is taken into account, the SOH fusion model can be simplified to a mechanical impact damage simplified SOH fusion model, as shown in the following expression: ; Combining this with the total lifespan, i.e., the total number of closing commands, we can obtain the formula for the mechanical damage coefficient: .

[0090] 3) Electrical damage weighting coefficient and nonlinear damage coefficient Calibration: 3-1. Select several relay samples from the same batch and apply different gradient currents to them in groups; for example, different gradient currents such as 50A, 100A, and 150A; 3-2. Under constant temperature and the same rated voltage conditions, cyclically send opening and closing commands to each relay sample until the relay fails; 3-3. Record the total number of failures for each group of relay samples; 3-4. Calculate the arc energy corresponding to each trip command for each relay sample, and statistically obtain the total arc energy for each group of relay samples; 3-5. Define the current health status value when the relay fails. Using the total number of failures as the dependent variable and the total arcing energy as the independent variable, the simplified SOH fusion model of electrical loss is logarithmically transformed and then linearly regressed to obtain the calibrated electrical damage weight coefficient and the calibrated nonlinear damage coefficient. Here, since mechanical damage and thermal degradation interference are not considered, and only electrical ablation is taken into account, the SOH fusion model can be simplified to an electrical loss simplified SOH fusion model, as follows: .

[0091] 4) Thermal compensation coefficient Calibration: 4-1. Select several relay samples from the same batch and divide them into a control group and a thermal degradation group; 4-2. Apply a large current to the thermally degraded group until its contact resistance rises from the initial value to a preset resistance value; for example, the large current is 150A; the preset resistance value is 1.2mΩ; Subsequently, under constant temperature and the same rated voltage conditions, tripping and closing commands were cyclically issued to each relay sample in the control group and the thermal degradation group until the relay failed. 4-3. Calculate the quantitative thermal degradation loss of each relay sample in the control group and the thermal degradation group respectively; 4-5. Based on the simplified SOH fusion model of thermal degradation loss and the quantified thermal degradation loss degree of each relay sample, the thermal compensation coefficient corresponding to each relay sample is calculated. It is understandable that, under the same electrical operation, the two sets of relay samples will have different degrees of SOH consumption due to their different contact resistances. Therefore, the following simplified SOH fusion model for thermal degradation loss can be obtained: ; Therefore, the thermal compensation coefficients corresponding to each relay sample can be derived as follows: ; 4-6. Fit the thermal compensation coefficients of all relay samples to obtain the calibrated thermal compensation coefficients.

[0092] This embodiment employs differentiated calibration methods tailored to the damage mechanisms of the multi-dimensional feature parameters in the SOH fusion model. This allows for precise matching of different failure characteristics, including arc ablation damage, mechanical impact damage, and thermal degradation loss, avoiding parameter mismatches caused by a single calibration method. Consequently, it effectively improves the accuracy of multi-dimensional damage quantification, enhances the environmental adaptability and assessment accuracy of the SOH fusion model, and ensures the reliability of relay health status detection results.

[0093] Another embodiment of the present invention further extends any of the above embodiments, providing a method for verifying the comprehensive prediction performance of the SOH fusion model.

[0094] To verify the comprehensive predictive capability of the SOH fusion model, this embodiment simulates random operating conditions of the relay and selects multiple sets of representative experimental data. On one hand, multi-dimensional features are extracted and a feature dataset is constructed based on the method provided in the above embodiment. The predicted SOH health level value is then obtained using the SOH fusion model. On the other hand, based on the measured contact resistance under the above random operating conditions, the actual health level value at the corresponding time is calculated using the formula for calculating the actual SOH health level value. This effectively verifies the comprehensive predictive capability of the SOH fusion model, ensures the model's generalization ability and prediction accuracy under complex operating scenarios, and provides reliable support for relay life assessment.

[0095] Specifically, the three-phase relay health assessment method based on multi-dimensional fusion provided in this embodiment further includes: S5: Verify and optimize the SOH fusion model. This specifically includes the following steps S51-S511.

[0096] S51: Select several representative sets of DMA acquisition data corresponding to closing and opening commands respectively, and calculate the quantized arc ablation damage degree corresponding to each opening command. And the quantified mechanical impact damage degree corresponding to each closing command. .

[0097] The predicted SOH health level value is obtained according to the SOH fusion model provided by this invention through the following steps S52-S54: S52: Obtain the smooth contact resistance value calculated by the relay corresponding to the closing command and opening command in S51 under the thermal balance high current condition, and calculate the quantified thermal degradation loss degree based on the smooth contact resistance value.

[0098] The specific calculations for the smooth contact resistance value and the quantified thermal degradation loss are described in the above embodiments and will not be repeated here.

[0099] S53: Based on the quantized arc ablation damage degree and quantized mechanical impact damage degree calculated in S51, and the quantized thermal degradation loss degree calculated in S52, construct a feature dataset including electrical loss characteristics, mechanical damage characteristics and thermal degradation characteristics.

[0100] S54: Input the feature dataset into the SOH fusion model to obtain the predicted SOH health level value.

[0101] The true SOH health level value is obtained using the formula for calculating the true SOH health level value, as described in S55 below: S55: Obtain the measured contact resistance value of the relays corresponding to the closing and opening commands mentioned in S51 under thermal balance high current conditions, and calculate the true SOH health value at time i using the following formula for calculating the true SOH health value. The formula for calculating the actual SOH health level value is as follows: ; in, Let be the measured contact resistance at time i; This represents the initial contact resistance value. This is the preset failure contact resistance value.

[0102] Through steps S56-S511 below, the Particle Swarm Optimization (PSO) algorithm is used to optimize the global parameters in the SOH fusion model, with the actual SOH health level as the objective. Optimization will be implemented in three stages: Phase 1, Initialization: S56: Limit the parameter search boundary based on actual calibration conditions, such as the nonlinear damage coefficient. Limited to the range [1,2]; the thermal compensation coefficient Limited to the range [0, 0.5]; to minimize the sum of squared residuals between the predicted SOH curve and the actual SOH curve of the SOH fusion model, a fitness function is defined: ; in, Let i be the predicted SOH health status value at time i; Let i be the actual SOH health status value at time i; M be the time series length; The fitness between the actual SOH health value and the predicted SOH health value output by the SOH fusion model is calculated based on the fitness function described above.

[0103] Phase Two: Iterative Calculation S57: Initialize a preset number of particles, and randomly assign a set of global parameters to each particle. For example, the number of particles in a swarm is 30-50.

[0104] S58: For each particle, using its global parameters, based on the feature dataset, substitute it into the SOH fusion model for iterative calculation to obtain the fitness set of each particle corresponding to the time series length M.

[0105] Phase 3, Update Optimal: S59: Based on the fitness set of each particle, obtain the optimal global parameters of each particle and the optimal global parameters of the population corresponding to all particles. S510: For each particle, adjust the corresponding global parameters based on its own optimal global parameters and the population's optimal global parameters; S511: Iterate through S58 to S510 until the preset number of iterations is reached, obtain the current optimal global parameters of the population, and mark them as the calibrated global parameters.

[0106] Preferably, the calibrated global parameters It is embedded into the ROM area of ​​the MCU.

[0107] Specifically, this embodiment employs the PSO algorithm to further optimize and calibrate the global parameters of the SOH fusion model. By efficiently optimizing in the parameter space, it avoids the dependence of linear regression on model assumptions, thereby improving the calibration accuracy and robustness of the global parameters. Through cluster iterative search, it quickly converges to the global optimum, effectively reducing the impact of experimental data errors and significantly improving the reliability of the SOH fusion model.

[0108] Based on any of the above embodiments, the present invention also provides an electricity meter.

[0109] like Figure 1 As shown, the energy meter in this embodiment includes a main control chip (MCU), a metering chip, and a non-volatile memory; the non-volatile memory stores a computer program; after being processed by the MCU, the computer program can implement the three-phase relay health assessment method based on multi-dimensional fusion described in any of the above embodiments.

[0110] The steps of the three-phase relay health assessment method based on multi-dimensional fusion will not be repeated here; please refer to the description in the above embodiments for details.

[0111] The main control chip MCU in the energy meter provided in this embodiment is configured with a DMA channel, which enables the sampled data to be directly written into the buffer of the non-volatile memory. The interrupt triggers the three-phase relay health assessment method based on multi-dimensional fusion described in any of the above embodiments to assess the health status of the relay in real time.

[0112] In some specific implementations, for split-type built-in relays, such as Figure 1 As shown, the MCU independently assesses the health status of the three single-phase relays, constructing corresponding health records for phase A, phase B, and phase C, to independently store the quantified arc erosion damage of each relay. Quantification of mechanical impact damage Quantification of thermal degradation loss and current health status value Preferably, the health record is stored in non-volatile memory.

[0113] In some specific implementations, such as Figure 1 As shown, the energy meter also includes a dual-mode communication module connected to the main control chip MCU, as well as a voltage divider network, a manganese-copper voltage divider network, and a contact resistance sampling network connected to the metering chip respectively.

[0114] The electricity meter provided in this embodiment has a real-time health status assessment function for its built-in relays based on a multi-dimensional fusion of thermal, mechanical, and electrical factors. Without requiring any additional hardware, it fully utilizes the high-speed data processing capabilities of the electricity meter's built-in metering chip and MCU. Within a millisecond-level short-time window of the relay's opening and closing actions, it accurately extracts the electrical loss characteristics, mechanical damage characteristics, and thermal degradation characteristics. Then, through a pre-built SOH fusion model, it assesses the health status of the target relay in real time and proactively reports warnings when necessary. This effectively ensures the long-term, stable, and safe operation of the electricity meter's built-in relays, significantly improving the meter's operational reliability and service life.

[0115] Those skilled in the art will understand that all or part of the processes in the above technical solutions can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the methods described above. After being executed by a processor, the program can also achieve the beneficial effects of the corresponding methods.

[0116] The storage medium can be a disk, optical disc, read-only memory (ROM), or random access memory (RAM), etc.

[0117] Although preferred embodiments of the invention have been described, those skilled in the art, upon understanding the basic inventive concept, can make other changes and modifications to the embodiments. Therefore, the appended claims are intended to cover the preferred embodiments and all equivalent modifications falling within the scope of protection defined by the claims and their equivalents. Clearly, those skilled in the art can make various alterations and variations to the invention without departing from its spirit and scope. If such modifications and variations fall within the scope of protection defined by the claims and their equivalents, the invention also intends to include them.

[0118] It should be noted that any reference signs placed between parentheses in the claims should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware.

[0119] In this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "multiple" means two or more, unless otherwise explicitly specified. The term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone.

[0120] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0121] In this specification, the use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refers to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the present invention. The illustrative expressions of the above terms in this specification should not be construed as necessarily referring to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0122] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent modifications made based on the content of the present invention specification and drawings, or direct or indirect applications in related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A three-phase relay health assessment method based on multi-dimensional fusion, characterized in that, include: S1: After receiving the opening / closing command, the MCU first configures the metering chip to high-speed sampling mode, and then sends the opening / closing command to the target relay. At the same time as sending the opening / closing command, the DMA is triggered to collect the instantaneous voltage values ​​on both sides of the target relay port and the instantaneous current values ​​in the target relay circuit at high speed. If it is an opening command, the start time of the acquisition and the arc extinction time will also be recorded. If it is a closing command, the start time of the acquisition and the closing completion time will also be recorded. S2: If it is a closing command, calculate the quantified mechanical impact damage degree of the target relay based on the DMA acquisition data corresponding to this closing command and then store it; If it is a tripping command, calculate the quantized arc erosion damage degree of the target relay based on the DMA acquisition data corresponding to this tripping command; And obtain the quantified mechanical shock damage degree calculated according to the most recent closing command for the target relay, and the quantified thermal degradation loss degree calculated according to the most recent thermal balance high current condition; execute S3; S3: The quantified arc ablation damage degree is used as an electrical loss feature, the quantified mechanical impact damage degree is used as a mechanical damage feature, and the quantified thermal degradation loss degree is used as a thermal degradation feature. These are input into the pre-constructed SOH fusion model, and the current health status value of the target relay is output. The expression for the SOH fusion model is: ； in, To quantify the degree of arc ablation damage as a characteristic of electrical loss; To quantify the degree of mechanical impact damage as a characteristic of mechanical damage; This is used to quantify the thermal degradation loss as a characteristic of thermal degradation. This is the thermal compensation coefficient; This is the thermal compensation factor; This represents the current health status value. This represents past health status. S4: Based on the current health status value of the target relay, determine whether to trigger the corresponding warning for the target relay.

2. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 1, characterized in that, The calculation of the quantified arc ablation damage degree of the target relay in S2 includes: S2a1: Calculate the arc energy based on the arc energy integral formula; The formula for the integral of arc energy is: ； in, For the energy of the combustion arc; The instantaneous voltage difference across the target relay port; The instantaneous current value of the target relay circuit; The sampling time interval; This is the starting time point for data collection; This refers to the time point when the arc is extinguished. S2a2: Calculate the quantitative arc ablation damage degree based on the quantitative arc ablation damage degree calculation formula; The formula for calculating the quantitative arc ablation damage degree is as follows: ； in, This is the electrical damage weighting coefficient; This is the nonlinear damage coefficient.

3. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 2, characterized in that, The calibration of the electrical damage weighting coefficient and the nonlinear damage coefficient includes: Several relay samples from the same batch were selected and grouped together and different gradient currents were applied. Under constant temperature and the same rated voltage conditions, opening and closing commands are cyclically sent to each relay sample until the relay fails. Record the total number of failures for each group of relay samples; Calculate the arc energy corresponding to a single trip command for each relay sample, and statistically obtain the total arc energy for each group of relay samples. The current health status value when the relay fails The total number of failures is the dependent variable, and the total arcing energy is the independent variable. A logarithmic transformation is performed on the simplified SOH fusion model of electrical losses, followed by linear regression, to obtain the calibrated electrical damage weighting coefficient and nonlinear damage coefficient. The simplified SOH fusion model of electrical losses is as follows: 。 4. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 1, characterized in that, The calculation of the quantified mechanical impact damage degree of the target relay in S2 includes: S2b1: Calculate the voltage difference between the two sides of the target relay port based on the instantaneous voltage values ​​on both sides of the target relay port in the DMA acquisition data corresponding to this closing command. S2b2: Mark the voltage instantaneous value difference that is greater than the preset voltage threshold; S2b3: Determine the jump time point corresponding to the difference between each of the markers and its adjacent unmarked voltage instantaneous value; S2b4: If the time interval between two adjacent switching time points is less than a preset first time interval, and the time interval between the two switching time points and the closing time point is less than a preset second time interval, then the cumulative mechanical bounce value is incremented by one. S2b5: The quantitative mechanical impact damage degree is calculated based on the cumulative mechanical bounce value using the quantitative mechanical impact damage degree formula. The formula for quantifying mechanical impact damage is: ； in, The mechanical damage coefficient; This represents the cumulative value of mechanical bounce.

5. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 4, characterized in that, The calibration of the mechanical damage coefficient includes: Select relay samples from the same batch, and under constant temperature, rated voltage and no-load conditions, cyclically send opening and closing commands to the relay samples until the relay fails. Record the cumulative mechanical bounce value of the relay sample for each closing command; Based on the simplified SOH fusion model of mechanical impact damage and the total number of closing commands, the calibrated mechanical damage coefficient is calculated using the mechanical damage coefficient formula. The simplified SOH fusion model for mechanical impact damage is as follows: The formula for the mechanical damage coefficient is: 。 6. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 1, characterized in that, The target relay was calculated to have a quantified thermal degradation loss under thermal equilibrium high-current conditions, including: S2c1: Under thermal equilibrium high current conditions, the MCU calculates the measured contact resistance of each relay at the current moment according to the dynamic contact resistance calculation formula at a preset time interval. The formula for calculating the dynamic contact resistance is as follows: ； in, This represents the measured contact resistance at the current moment. This represents the instantaneous voltage value at the relay input port at the current moment; This represents the instantaneous voltage value at the relay output port at the current moment. This represents the instantaneous current value of the relay circuit at the current moment; S2c2: Calculate the gain coefficient at the current moment based on the gain coefficient formula; The formula for the gain coefficient is: ； in, The covariance of the previous time step; This represents the instantaneous current value of the target relay circuit at the current moment; is the forgetting factor; n is the current time. S2c3: Calculate the prediction error at the current moment based on the prediction error formula; The prediction error formula is: ； in, This represents the prediction error at the current moment; This represents the instantaneous voltage difference between the two sides of the target relay port at the current moment. This is the contact resistance value at the previous moment; S2c4: Calculate the contact resistance value at the current moment based on the contact resistance value calculation formula; The formula for calculating the contact resistance value is: ； in, This represents the contact resistance value at the current moment; S2c5: Calculate the covariance at the current time based on the covariance formula; The covariance formula is: ； S2c6: The MCU executes steps S2c1 to S2c5 repeatedly at preset time intervals, iteratively calculates the contact resistance value at the current moment, and marks it as the smooth contact resistance value at the current moment. S2c7: Calculate the quantitative thermal degradation loss based on the quantitative thermal degradation loss calculation formula; The formula for calculating the quantified thermal degradation loss is as follows: ； in, This represents the smooth contact resistance value at the current moment; This represents the initial contact resistance value.

7. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 1, characterized in that, The calibration of the thermal compensation coefficient includes: Several relay samples from the same batch were selected and divided into a control group and a thermal degradation group; A large current is first applied to the thermally degraded group until its contact resistance rises to a preset resistance value; then, under constant temperature environment and the same rated voltage conditions, opening and closing commands are cyclically issued to each relay sample of the control group and the thermally degraded group until the relay fails. The quantitative thermal degradation loss of each relay sample in the control group and the thermal degradation group was calculated respectively. Based on the simplified SOH fusion model of thermal degradation loss and the quantified thermal degradation loss degree of each relay sample, the thermal compensation coefficient corresponding to each relay sample is calculated. The simplified SOH fusion model for thermal degradation loss is as follows: ； The thermal compensation coefficients corresponding to each relay sample are: ； The thermal compensation coefficients of all relay samples were fitted to obtain the calibrated thermal compensation coefficients.

8. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 1, characterized in that, Also includes: S5: Verify the SOH fusion model; S5 specifically includes: S51: Select several sets of representative closing and opening commands corresponding to DMA acquisition data, and calculate the corresponding quantitative arc ablation damage degree and quantitative mechanical impact damage degree. S52: Obtain the smooth contact resistance value calculated by the relay corresponding to the closing command and opening command in S51 under the thermal balance high current condition, and calculate the quantified thermal degradation loss degree based on the smooth contact resistance value. S53: Based on the quantitative arc ablation damage degree and quantitative mechanical impact damage degree calculated in S51, and the quantitative thermal degradation loss degree calculated in S52, construct a feature dataset including electrical loss characteristics, mechanical damage characteristics and thermal degradation characteristics. S54: Input the feature dataset into the SOH fusion model to obtain the predicted SOH health level value; S55: Obtain the measured contact resistance value of the relays corresponding to the closing and opening commands mentioned in S51 under thermal balance high current conditions, and calculate the true SOH health value at time i using the true SOH health value calculation formula. The formula for calculating the actual SOH health level value is as follows: ； in, Let be the measured contact resistance value at time i; This represents the initial contact resistance value. This is the preset failure contact resistance value; S56: Calculate the fitness between the actual SOH health status value and the predicted SOH health status value output by the SOH fusion model based on the fitness function; The fitness function is: ； in, Let i be the predicted SOH health status value at time i; Let i be the actual SOH health status value at time i; M be the time series length; S57: Initialize a preset number of particles, and randomly assign a set of global parameters to each particle. ; S58: For each particle, using its global parameters, perform iterative calculations based on the feature dataset and the SOH fusion model to obtain the fitness set of each particle corresponding to the time series length M; S59: Based on the fitness set of each particle, obtain the optimal global parameters of each particle and the optimal global parameters of the population corresponding to all particles. S510: For each particle, adjust the corresponding global parameters based on its own optimal global parameters and the population's optimal global parameters; S511: Iterate through S58 to S510 until the preset number of iterations is reached, obtain the current optimal global parameters of the population, and mark them as the calibrated global parameters.

9. The three-phase relay health assessment method based on multi-dimensional fusion as described in claim 1, characterized in that, S4 specifically includes: If the current health level value is higher than or equal to the health level threshold, the target relay is determined to be in good working condition and has a safe and stable disconnection capability. If the current health level value is higher than or equal to the warning level threshold, but lower than the health level threshold, the target relay is determined to have entered the mid-term degradation stage and is in a warning state. The frequency of high current surge operations is limited, and a status prompt is triggered. If the current health status value is lower than the warning level threshold, the target relay is determined to have entered the late degradation stage and is in a dangerous state. High current surge operations are prohibited, and an emergency alarm is triggered.

10. An electricity meter, characterized in that, It includes an MCU, a metering chip, and a non-volatile memory; the non-volatile memory stores a computer program; the computer program, after being processed by the MCU, can implement the three-phase relay health assessment method based on multi-dimensional fusion as described in any one of claims 1 to 9.

Images