Charging pile safety control method and system based on multi-source data fusion

Abstract

The application discloses a charging pile safety control method and system based on multi-source data fusion, and belongs to the technical field of data analysis. The method first collects temperature of temperature measuring points, environmental temperature and insulation resistance in the charging process and carries out pretreatment. Secondly, the heat propagation delay between the temperature measuring points is calculated by using the cross-correlation function, and a heat propagation topology graph is constructed to identify the root heat source and the subordinate heat source. Then, the temperature data is decomposed into the self-heating component and the propagation heat component based on the topological sorting, and the heat contribution weight and the charging safety index of each temperature measuring point are calculated. Finally, the charging current adjustment amount is calculated according to the deviation of the charging safety index and the safety threshold, and the output current of the charging pile is dynamically adjusted. The application realizes the closed-loop safety control of the charging process, and can improve the safety and reliability of the charging pile operation.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a method and system for safety control of charging piles based on multi-source data fusion. Background Technology

[0002] As a core component supporting industrial operations, the operational safety of charging infrastructure is a major concern. During power transmission, the continuous action of high current generates significant heat accumulation at power supply equipment, transmission cables, and connection points, creating a complex thermal field distribution. If this heat cannot be dissipated in time, it can easily lead to safety accidents such as insulation aging, overheating, or even fires. Therefore, real-time and accurate thermal safety monitoring and dynamic control of the charging process are crucial for ensuring the stable operation of facilities. The industry typically deploys multiple types of sensors in key heat-generating areas of equipment to acquire multi-dimensional temperature data and attempts to assess the overall thermal safety status by fusing multi-source data, thereby implementing corresponding control strategies.

[0003] However, existing technologies have significant limitations when integrating multi-source temperature data for thermal safety assessment. Traditional methods typically process data using direct timestamp alignment, which not only ignores the inherent heat propagation delay between different physical nodes but also fails to effectively eliminate the impact of synchronous temperature rise caused by fluctuations in ambient temperature. Due to these shortcomings, heat received by downstream monitoring points from upstream sources is often incorrectly attributed to the monitoring point's own heat generation. This misjudgment of heat source attributes causes safety assessment indicators to deviate significantly from the actual thermal distribution of the equipment, leading to false alarms from the system regarding normal heat conduction, masking real localized overheating hazards, and resulting in missed detections. This reduces the accuracy of safety assessments and the effectiveness of control strategies. Summary of the Invention

[0004] The purpose of this invention is to provide a charging pile safety control method and system based on multi-source data fusion to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides a charging pile safety control method based on multi-source data fusion, comprising:

[0006] The charging pile collects multi-source operational data throughout the charging process through its internal data acquisition module. The charging pile is equipped with different types of hardware modules, including a data acquisition module, a storage module, a communication module, and a main control MCU microcontroller.

[0007] The data acquisition module continuously collects multi-source operating data during the charging process at a set sampling frequency and transmits the raw data to the storage module for caching. The main control MCU microcontroller reads the data from the storage module for processing and calculation. The communication module is responsible for exchanging and transmitting the processing results and control commands of the main control MCU microcontroller with the external system. Each module is physically integrated in the internal control board of the charging pile and interconnects with the data through an internal bus or communication interfaces such as SPI / I2C.

[0008] The data acquisition module includes multiple temperature sensors and insulation resistance detection circuits, used to collect multi-source operating data. This multi-source operating data includes temperature data from multiple measurement points, ambient temperature, and insulation resistance.

[0009] The preferred temperature sensor is an NTC thermistor or a PT100 platinum resistance temperature sensor, with a measurement accuracy of ±0.5℃ and a response time of less than 1 second, which can meet the real-time monitoring requirements for rapid temperature changes during charging.

[0010] The insulation resistance detection circuit measures the leakage current by applying a preset test voltage between the charging pile output circuit and ground, and calculates the insulation resistance value accordingly. This is used to determine whether there is any deterioration or breakdown risk in the internal electrical insulation of the charging pile.

[0011] Ambient temperature sensors are typically placed near the ventilation openings of the charging pile casing or in non-heat-generating areas inside, serving as a reference for environmental calibration of temperature data.

[0012] Multiple temperature sensors are respectively placed at the charging gun connector, the charging cable sheath, and the power module heat sink fins to collect the connector temperature, cable temperature, and module temperature as temperature data.

[0013] The charging gun connector is the part most prone to poor contact, oxidation and wear, and overheating during the charging process. The temperature at this point directly reflects the health of the contact condition.

[0014] The surface temperature of a charging cable reflects its current-carrying capacity and heat dissipation capabilities. The temperature at the heat sink fins of a power module reflects the junction temperature of the power semiconductor device.

[0015] The three temperature measurement points are distributed along the path of "power module → cable → charging gun connector", which corresponds to the power transmission path of the charging current flowing from the inside of the charging pile to the electric vehicle, providing a reasonable physical topology basis for heat propagation analysis.

[0016] The storage module is used to store multi-source operating data and cached data from the main control MCU microcontroller.

[0017] Preprocessing of multi-source operational data includes:

[0018] The raw sampling data from each temperature sensor were resampled to a uniform sampling frequency, and outliers were removed from the resampled data using the Laida criterion. The specific process is as follows:

[0019] For each temperature sensor, the mean and standard deviation of the resampled data sequence within the time window are calculated. If the absolute value of the deviation of a sampling point from the mean exceeds 3 times the standard deviation, the sampling point is determined to be an outlier and is removed.

[0020] In real-world charging environments, ambient temperature fluctuates over time due to factors such as changes in sunlight and ventilation conditions. These fluctuations synchronously affect the temperature readings of all temperature measurement points through heat conduction and heat radiation. Without environmental correction, subsequent cross-correlation analysis will misinterpret this synchronous change as a heat propagation relationship, thus interfering with heat source identification.

[0021] Therefore, it is necessary to subtract the ambient temperature at the same time from the temperature data of each temperature sensor at each moment after removing outliers to obtain the ambient temperature correction data. This temperature data only reflects the net temperature rise of the measurement point relative to the environment, eliminating the interference of the common-mode component of the ambient temperature.

[0022] Extract the temperature time series of any two temperature measurement points within a preset time window from the preprocessed temperature data, and denote them as the first temperature series. Second temperature sequence .in, , The sequence length is given.

[0023] The setting of the preset time window needs to balance the requirements of calculation accuracy and real-time performance:

[0024] A time window that is too short will result in insufficient data points for cross-correlation calculations, making the estimated heat propagation delay unstable. A time window that is too long will introduce too much historical data, reducing the response speed to the current heat propagation state.

[0025] Calculate the normalized cross-correlation coefficients of the first and second temperature series at different time offsets, for the time offsets. Normalized cross-correlation coefficients The calculation formula is:

[0026] ;

[0027] Normalized cross-correlation function is used to measure the first temperature series. With the second temperature sequence At different time offsets The degree of similarity of the waveforms.

[0028] Normalization processing makes The value range is limited to [-1, 1], which eliminates the influence of the absolute amplitude difference of temperature data on the similarity measurement and makes the cross-correlation coefficients between different temperature measurement point pairs comparable.

[0029] when A value close to 1 indicates that the variation pattern of the second temperature sequence differs from that of the first temperature sequence in terms of delay. The consistent patterns of change over time indicate that the delay in heat transfer from the first temperature measurement point to the second temperature measurement point is approximately [missing information]. .

[0030] By traversing different Value and search The maximum value can be used to determine the heat propagation delay between the two temperature measurement points.

[0031] In the formula, This is the mean of the first temperature sequence; This is the mean of the second temperature sequence;

[0032] and These are the arithmetic mean values ​​of the two temperature series, used for mean removal to eliminate the influence of the DC component, so that the cross-correlation coefficient reflects only the fluctuation similarity of temperature changes rather than the similarity of absolute temperature levels.

[0033] The numerator in the formula is the cross-covariance term (the upper limit of the summation is...). Because when the second sequence shifts The number of sample points that can be paired later is reduced. (number), in the denominator and The square roots of the variances of the first and second temperature sequences are used to normalize the molecules to the [-1, 1] interval.

[0034] This is the time offset, and its value range is... . This is the preset maximum heat propagation delay limit; The setting is based on the maximum possible heat transfer time between any two temperature measuring points in the physical structure of the charging pile:

[0035] In actual charging pile structures, the heat conduction path from the power module heat sink fins to the charging gun connector via the metal structural components is typically between 0.5 and 2 meters. The heat conduction rate of metal is on the order of tens of millimeters per second, so the maximum heat propagation delay is typically in the range of 10 to 60 seconds.

[0036] set up The purpose is to limit the search range of time offset for cross-correlation calculations to avoid invalid searches for large delay values, while also preventing unrelated random fluctuations from being misjudged as long-delay heat propagation relationships.

[0037] The second temperature sequence in time offset Next The value at each sampling point.

[0038] Iterate through all values ​​within the range in ascending integer increments and calculate the corresponding normalized cross-correlation coefficients. Arranged in ascending order of time offset, they form a cross-correlation function curve.

[0039] Extract the time offset corresponding to the maximum value of the normalized cross-correlation coefficient in the cross-correlation function curve as the heat propagation delay between the two temperature measurement points, and construct the heat propagation delay matrix by traversing all temperature measurement point pairs.

[0040] The heat propagation delay matrix is ​​a × A square matrix, the first element in the matrix is... Line 1 The elements of the column represent the first... From the first temperature measurement point to the... The thermal propagation delay between temperature measurement points is represented by a diagonal element of 0, indicating that the delay of the temperature measurement point itself is zero.

[0041] Because heat propagation is directional, the delay matrix is ​​not necessarily symmetric.

[0042] Traversing all temperature measurement points means... Each temperature measurement point is paired up, and a total of [number] calculations are required. For each combination of temperature measurement points, a heat propagation delay value is determined based on the cross-correlation function.

[0043] Each temperature measurement point is used as a node in the heat propagation topology graph, and directed edges are established for temperature measurement point pairs whose heat propagation delay is greater than a preset delay threshold in the heat propagation delay matrix.

[0044] The direction of the directed edge is from the temperature measurement point with a smaller thermal propagation delay to the temperature measurement point with a larger thermal propagation delay. No directed edge is established for temperature measurement points with a thermal propagation delay less than or equal to a preset delay threshold.

[0045] The direction of a directed edge is determined based on the physical causality of heat propagation:

[0046] Heat always propagates from the source node where the temperature changes first to the downstream node where the temperature changes later. Therefore, the temperature measurement point with a smaller heat propagation delay is the heat source end, and the temperature measurement point with a larger heat propagation delay is the heat receiving end.

[0047] Directed edges point from the end with smaller delay to the end with larger delay. This directional convention ensures that the topology graph can correctly reflect the physical direction of heat conduction.

[0048] Calculate the in-degree and out-degree of each node in the heat propagation topology graph:

[0049] In a heat propagation topology graph, the in-degree of a node refers to the number of directed edges pointing to that node, indicating how many upstream heat sources propagate heat to that temperature measurement point.

[0050] The out-degree of a node refers to the number of directed edges originating from that node, indicating how many downstream nodes the temperature measuring point will propagate heat to.

[0051] A zero in-degree means that the node has no upstream heat source and is the starting point for heat propagation, while a large out-degree means that the node is the heat source for multiple downstream nodes and is located upstream in the heat propagation chain.

[0052] When there is a node with an in-degree of zero in the heat propagation topology graph, it means that there are no directed edges pointing to it in the heat propagation topology graph. This temperature measuring point is not affected by the heat conduction of any other temperature measuring point, and its temperature change is entirely caused by its own heat generation. Therefore, it is marked as a root heat source.

[0053] Nodes with an in-degree greater than zero are marked as subordinate heat sources, indicating that the temperature measurement point is affected by heat conduction from at least one upstream heat source, and its temperature change includes a component of propagated heat.

[0054] If the in-degree of the temperature measuring point on the power module heat sink fin is zero, while the in-degree of the temperature measuring point on the charging gun connector is greater than zero, then the power module heat sink fin is the root heat source, and the charging gun connector is the subordinate heat source.

[0055] When there are no nodes with an in-degree of zero in the heat propagation topology graph, it indicates that there is a loop structure in the graph, that is, all nodes have upstream nodes pointing to form a cyclic heat propagation relationship. In actual charging piles, this cyclic topology may be caused by the small difference in heat propagation delay between each temperature measurement point and the close cross-correlation coefficient, resulting in the formation of a closed loop by directed edges.

[0056] Since the root heat source cannot be directly identified by its zero in-degree, the principle of maximizing out-degree is adopted as an alternative strategy.

[0057] The node with the largest outgoing degree means that this temperature measuring point transmits heat to the most downstream nodes, and is most likely the root heat source in the upstream position of the heat propagation chain. Therefore, the node with the largest outgoing degree is marked as the root heat source, and the other nodes are marked as subordinate heat sources.

[0058] When there are multiple root heat sources in the heat propagation topology diagram, it indicates that there are multiple independent heat generation starting points inside the charging pile. At this time, it is necessary to identify the main heat source that has the greatest impact on overall thermal safety.

[0059] Calculate the sum of the heat propagation delays for all directed edges along the shortest propagation path from each root heat source to each subordinate heat source. The shortest propagation path is the path in the heat propagation topology graph that traverses the fewest directed edges or has the smallest sum of heat propagation delays between a root heat source node and a subordinate heat source node. The sum of the heat propagation delays for all directed edges along the path reflects the total time required for heat to propagate from the root heat source to the subordinate heat source.

[0060] The reciprocal of the sum of heat propagation delays is taken as the propagation intensity of the corresponding root heat source to the subordinate heat source, and the root heat source with the largest sum of propagation intensity is marked as the main heat source.

[0061] Preserve the directed edges from the main heat source to each subordinate heat source and their corresponding heat propagation delays, and remove the directed edges emanating from the remaining root heat sources to obtain the final heat propagation topology.

[0062] The complex topology of multiple heat sources is simplified into a tree structure with the main heat source as the only root node, so that the subsequent temperature component decomposition calculation has a clear propagation hierarchy and a unique starting point.

[0063] After removing directed edges originating from non-primary heat sources, the final heat propagation topology retains the complete propagation path from the primary heat source to each subordinate heat source, and each subordinate heat source retains only the directed edges originating from the primary heat source.

[0064] Analyze the temperature measurement point corresponding to the root heat source in the heat propagation topology diagram, take the temperature change of the measurement point at the current moment relative to the charging start moment as the self-heating component, and set the propagation heat component of the measurement point to zero.

[0065] For a temperature measurement point corresponding to a subordinate heat source, obtain all upstream nodes of that temperature measurement point in the heat propagation topology diagram. For each upstream node... Read the upstream node At a historical moment The self-heating component value has been calculated and stored.

[0066] Historical moment For the current moment Retrospective heat propagation delay At the time obtained, the upstream node At a historical moment The self-heating component value at that location has been calculated in the historical control cycle corresponding to that historical moment.

[0067] Multiply the self-heating component value by the propagation attenuation coefficient The summation over all upstream nodes gives the temperature reading at the current time for that temperature measurement point. The heat component of propagation. The formula for calculating the propagation attenuation coefficient is:

[0068] ;

[0069] Propagation attenuation coefficient Used to quantify heat from upstream nodes Propagation to downstream nodes The degree of attenuation over time, that is, what proportion of the self-heating component of the upstream node can be conducted to the downstream node to form a propagating heat component.

[0070] Combining factors from two dimensions:

[0071] First, the statistical correlation of temperature changes between the two nodes is determined by the maximum value of the normalized cross-correlation coefficient. It is reflected in the range of [0,1]. The larger the value, the closer the thermal coupling between the two points and the higher the heat transfer efficiency.

[0072] Second, heat dissipates and decays naturally over time and distance along the propagation path, as indicated by the exponential decay term. This value reflects that... An increase or decrease indicates that the longer the propagation delay, the more heat is dissipated.

[0073] The product of the two factors allows the propagation attenuation coefficient to comprehensively reflect the coupling strength and dissipation characteristics of heat propagation, which is more in line with the dissipation law of heat when it actually propagates along the physical path compared to existing methods that only use the correlation coefficient as the propagation weight.

[0074] In the formula, upstream node Arrive at this temperature measurement point The propagation attenuation coefficient; the larger the value, the higher the upstream node. The heat is transmitted to the temperature measuring point. The higher the efficiency, the smaller the attenuation.

[0075] For this temperature measurement point With upstream nodes The maximum value of the normalized cross-correlation coefficient between the two is used as a benchmark factor for the propagation attenuation coefficient to reflect the conduction capacity of the heat propagation path.

[0076] For this temperature measurement point With upstream nodes The delay in heat propagation between them; This represents the sampling interval for temperature data. Converting the delay from the number of sampling points to actual physical time makes the attenuation calculation consistent with physical dimensions.

[0077] The preset heat dissipation coefficient is a constant greater than zero; it reflects the rate of heat dissipation per unit time as heat propagates in the physical structure of the charging pile.

[0078] The value of is related to the heat dissipation conditions of the charging pile, the thermal conductivity of the structural materials, and other physical properties. The larger the value, the better the heat dissipation and the faster the heat is dissipated. The value can be obtained through experimental calibration.

[0079] Subtract the temperature change of the corresponding temperature measuring point at the current moment from the temperature change at the start of charging from the temperature change at the current moment of the heat source. The heat component of the propagation is used to obtain the temperature at the current time of the temperature measurement point. The self-heating component.

[0080] According to the topological sorting of the heat propagation topology diagram, the temperature data of each temperature measurement point is decomposed into self-heating component and propagation heat component.

[0081] Calculate the rate of change of the self-heating component at each temperature measurement point at the current moment, denoted as the rate of change of self-heating. Substitute this into the formula to calculate the heat contribution weight of each temperature measurement point:

[0082] ;

[0083] Heat contribution weight Used to quantify the The contribution ratio of the self-heating component of each temperature measuring point to the overall charging safety assessment is taken in the range of (0,1), and the sum of the thermal contribution weights of all temperature measuring points is 1.

[0084] By using the absolute amplitude of the self-heating component as a benchmark and introducing a topology depth penalty factor, temperature measurement points with larger self-heating amplitudes and closer proximity to the root heat source are given higher weight in the safety assessment.

[0085] In the formula, For the first The heat contribution weight of each temperature measurement point; For the first At the current moment, each temperature measurement point... The self-heating component; Let be the topological depth of the j-th temperature measurement point in the heat propagation topology map.

[0086] The topology depth is the number of intermediate temperature measurement points along the shortest path from the root heat source to the temperature measurement point in the heat propagation topology graph. The topology depth of the temperature measurement point corresponding to the root heat source is zero.

[0087] The penalty factor when the topological depth of the temperature measurement point corresponding to the root heat source is zero. That is, no decay, the penalty factor for a direct downstream node topology depth of 1 is . And so on.

[0088] is the depth penalty coefficient, a preset constant greater than zero; used to control the decay rate of weights with topology depth. The greater the depth, the stronger the punishment.

[0089] This represents the total number of temperature measurement points. This is the summation index variable, used to iterate through all... The calculation of the self-heating contribution of all temperature measuring points in the denominator achieves weight normalization.

[0090] It is a preset small constant greater than zero, used to ensure that each temperature measuring point receives an equal heat contribution weight when the self-heating component of all temperature measuring points is zero.

[0091] Calculate the charging safety index The formula is:

[0092] ;

[0093] Charging safety index The core indicator used to comprehensively quantify the current thermal safety status of charging piles.

[0094] The self-heating and propagation heat components of all temperature measurement points are weighted and summed according to their respective heat contribution weights, and a dynamic amplification factor is introduced to nonlinearly adjust the rate of change of self-heating. The higher the value, the higher the overall thermal safety risk of the charging station.

[0095] In the formula, For charging safety index; For the first At the current moment, each temperature measurement point... The heat transfer component reflects the temperature change conducted from the upstream heat source, for the root heat source. .

[0096] For the first The rate of change of self-heating at each temperature measurement point at the current moment; and The weighting coefficients are preset and satisfy: .

[0097] in, This is the weighting coefficient for the self-heating component, used to make the safety index more sensitive to the self-heating component. This is the weighting coefficient for the propagation heat component, used to adjust the contribution of the propagation heat component to the safety index.

[0098] The self-heating component, which directly reflects the abnormal heating level of the temperature measuring point itself, is the primary focus of safety risk assessment, while the propagated heat component, which is the result of normal heat conduction, has a relatively lower risk level.

[0099] This is the dynamic sensitivity coefficient, a preset constant greater than zero; it is used to control the sensitivity of the dynamic amplification factor to the rate of change of self-heating. The higher the safety index, the more sensitive it is to rapid self-heating and the easier it is to trigger protective actions in the early stages of temperature rise.

[0100] The dynamic amplification factor is 1 when the rate of change of self-heating is zero, and greater than 1 when the rate of change of self-heating is greater than zero.

[0101] In actual operation of charging piles, the development of local overheating faults usually goes through a process of "slow heating - accelerated heating - rapid heating".

[0102] During the slow heating phase, the absolute temperature value has not yet exceeded the traditional fixed threshold, but the rate of change of self-heating has begun to increase. Traditional monitoring methods based on fixed thresholds cannot respond in time during this phase.

[0103] The dynamic amplification factor of this invention can capture abnormal increases in the rate of self-heating change. A significant increase occurs at this stage, triggering the current reduction protection in advance and eliminating potential safety hazards at the nascent stage.

[0104] Read the insulation resistance at the current moment. When the insulation resistance is less than the preset insulation threshold, set the charging current adjustment to the negative value of the current charging current and cut off the output current of the charging pile.

[0105] Insulation resistance is a key parameter reflecting the electrical insulation status of a charging pile. When the insulation resistance is lower than the preset insulation threshold, it indicates that the insulation performance of the charging pile's internal or output circuit has been severely deteriorated, posing a direct risk of leakage or even short circuit.

[0106] At this point, thermal safety assessment is no longer the primary concern; the charging current must be immediately cut off to ensure personal and equipment safety. Therefore, insulation resistance assessment, as an independent safety condition, takes precedence over the charging safety index. The assessment and implementation.

[0107] When the insulation resistance is not less than the preset insulation threshold, a first safety threshold is set. Second security threshold And satisfy: .

[0108] When charging safety index Less than the first safety threshold At this time, the charging current adjustment is set to zero to maintain the current charging current unchanged.

[0109] When charging safety index Greater than or equal to the first safety threshold And less than the second safety threshold At that time, the charging current adjustment is calculated according to a non-linear ratio. This reduces the output current of the charging station.

[0110] when When the charging station is in the warning zone, it has already shown a certain degree of thermal safety risk. It is necessary to reduce the charging current to reduce the heat generation and suppress the temperature from rising further.

[0111] The specific calculation formula is as follows:

[0112] ;

[0113] according to The current reduction is calculated using a non-linear proportional method based on the relative position within the warning zone: The closer The greater the decrease in current, The closer The smaller the decrease in current.

[0114] Adjusted charging current for:

[0115] ;

[0116] In the formula, The current charging current; that is, the actual output current of the charging pile in the current control cycle, in... As a benchmark, the current adjustment is proportional to the current level to ensure that the current reduction ratio is consistent in different charging stages.

[0117] The preset curvature index is a positive integer greater than 1; it is used to control the nonlinearity of the current adjustment curve. The bigger Just over The smaller the drop in flow at that time, the closer to The greater the drop in flow rate at that time.

[0118] Nonlinear exponent Make the current adjustment curve concave in shape, Just over The decrease was relatively small at that time. near The drop in speed increases dramatically, achieving a balance between smooth transition and rapid response.

[0119] The charging safety index is recalculated in each control cycle. And update the charging current adjustment amount. The control cycle is the time interval between the main control MCU microcontroller executing a complete safety assessment and current adjustment process, typically ranging from 1 to 5 seconds.

[0120] The system recalculates within each control cycle. And according to Calculate the new current adjustment amount within the safe range.

[0121] When charging safety index Falling back to the first safety threshold At this point, the charging current is gradually restored to improve charging efficiency. The amount of current restoration is calculated. Gradually restore the charging current to the rated charging current. The calculation formula is:

[0122] ;

[0123] When charging safety index Greater than or equal to the second safety threshold When the charging current adjustment is set to the negative value of the current charging current, the output current of the charging pile is cut off.

[0124] The charging pile's output current is dynamically adjusted based on the charging current adjustment amount to achieve closed-loop safety control of the charging process.

[0125] The present invention also provides a charging pile safety control system based on multi-source data fusion, including a data acquisition module, a storage module, a communication module and a main control MCU microcontroller.

[0126] The data acquisition module is responsible for acquiring data at the perception layer, the storage module is responsible for data caching and management at the data layer, the main control MCU microcontroller is responsible for data processing and decision-making at the computing layer, and the communication module is responsible for information transmission at the interaction layer.

[0127] The data acquisition module includes multiple temperature sensors and insulation resistance detection circuits for collecting multi-source operating data. The temperature sensors are located at the charging gun connector, the charging cable sheath, and the heat sink fins of the power module.

[0128] The storage module is used to store multi-source operating data and cached data from the main control MCU microcontroller.

[0129] The communication module is used to interact with the electric vehicle battery management system.

[0130] The main control MCU microcontroller is connected to the data acquisition module, storage module and communication module, and is used to perform cross-correlation function calculation, heat propagation topology construction, temperature component decomposition and dynamic adjustment of charging current.

[0131] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0132] Compared to traditional fixed threshold temperature monitoring methods, this method separates propagated heat from self-heating by decomposing temperature components. This allows for the differentiation between normal temperature rise caused by charging current load and localized temperature rise caused by abnormal contact, thus reducing the false alarm rate.

[0133] By introducing a preset delay threshold to filter temperature measurement point pairs with excessively small delays, the structure of the heat propagation topology map is made to better reflect the actual physical heat propagation path. At the same time, the cyclic topology is handled by the node with the largest out-degree, ensuring the robustness of root heat source identification.

[0134] Compared to existing methods that use only the correlation coefficient as the propagation weight, the propagation attenuation coefficient formula... By combining correlation with delay decay, the longer the propagation path, the greater the decay, making the calculation of the propagation heat component more consistent with the dissipation characteristics of heat during actual propagation along the physical path, thus improving the accuracy of temperature component decomposition.

[0135] Compared to existing methods that assign weights solely based on the ratio of change rates, the thermal contribution weighting formula... By using the amplitude of the self-heating component as a benchmark and introducing a topology depth penalty factor, the weights can be correctly allocated under both steady-state and transient conditions, making the temperature measurement points near the root heat source dominate the safety index calculation and improving the sensitivity of locating abnormal heat sources.

[0136] Compared to existing methods that simply use linear weighted summation to construct the safety index, the charging safety index formula... By introducing a dynamic amplification factor, the charging safety index is nonlinearly amplified when the self-heating component rises rapidly, which can trigger current regulation earlier and improve the timeliness of safety response.

[0137] By prioritizing insulation resistance as an independent safety judgment condition over the charging safety index, an immediate response to insulation faults can be achieved, which can cover more fault types compared to safety control methods based solely on temperature.

[0138] Compared to the traditional step-current cutoff strategy, the nonlinear proportional adjustment strategy achieves a smooth reduction and recovery of the charging current, which reduces unnecessary charging interruptions and improves charging continuity while ensuring safety. Attached Figure Description

[0139] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0140] Figure 1 This is a flowchart illustrating the charging pile safety control method based on multi-source data fusion according to the present invention.

[0141] Figure 2 This is a schematic diagram of the temperature change curves of each temperature measuring point as a function of charging time in this invention;

[0142] Figure 3 This is a schematic diagram of the dynamic change curve of the charging safety index and charging current of the present invention. Detailed Implementation

[0143] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0144] Please see Figure 1 This invention provides a charging pile safety control method based on multi-source data fusion, comprising:

[0145] S100 collects multi-source operating data throughout the charging process through the data acquisition module inside the charging pile. The charging pile is equipped with different types of hardware modules, including a data acquisition module, a storage module, a communication module, and a main control MCU microcontroller.

[0146] The data acquisition module continuously collects multi-source operating data during the charging process at a set sampling frequency and transmits the raw data to the storage module for caching. The main control MCU microcontroller reads the data from the storage module for processing and calculation (including core steps such as cross-correlation analysis, heat propagation topology construction, temperature component decomposition, and safety index calculation). The communication module is responsible for exchanging and transmitting the processing results and control commands of the main control MCU microcontroller with the external system. Each module is physically integrated in the internal control board of the charging pile and interconnects with the data through an internal bus or communication interfaces such as SPI / I2C.

[0147] The data acquisition module includes multiple temperature sensors and insulation resistance detection circuits, used to collect multi-source operating data. This multi-source operating data includes temperature data from multiple measurement points, ambient temperature, and insulation resistance.

[0148] The preferred temperature sensor is an NTC thermistor or a PT100 platinum resistance temperature sensor, with a measurement accuracy of ±0.5℃ and a response time of less than 1 second, which can meet the real-time monitoring requirements for rapid temperature changes during charging.

[0149] The insulation resistance detection circuit measures the leakage current by applying a preset test voltage (such as DC 500V) between the charging pile output circuit and ground, and calculates the insulation resistance value accordingly. This is used to determine whether there is any deterioration or breakdown risk in the internal electrical insulation of the charging pile.

[0150] Ambient temperature sensors are typically placed near the ventilation openings of the charging pile casing or in non-heat-generating areas inside, serving as a reference for environmental calibration of temperature data.

[0151] Multiple temperature sensors are respectively placed at the charging gun connector, the charging cable sheath, and the power module heat sink fins to collect the connector temperature, cable temperature, and module temperature as temperature data.

[0152] The charging gun connector is the part most prone to poor contact, oxidation and wear, and overheating during the charging process. The temperature at this point directly reflects the health of the contact condition.

[0153] The surface temperature of a charging cable reflects its current-carrying capacity and heat dissipation capabilities. The temperature at the heat sink fins of a power module reflects the junction temperature of the power semiconductor device.

[0154] The three temperature measurement points are distributed along the path of "power module → cable → charging gun connector", which corresponds to the power transmission path of the charging current flowing from the inside of the charging pile to the electric vehicle, providing a reasonable physical topology basis for heat propagation analysis.

[0155] The storage module is used to store multi-source operating data and cached data from the main control MCU microcontroller.

[0156] Preprocessing of multi-source operational data includes:

[0157] The raw sampling data from each temperature sensor were resampled to a uniform sampling frequency, and outliers were removed from the resampled data using the Laida criterion. The specific process is as follows:

[0158] For each temperature sensor, the mean and standard deviation of the resampled data sequence within the time window are calculated. If the absolute value of the deviation of a sampling point from the mean exceeds 3 times the standard deviation, the sampling point is determined to be an outlier and is removed.

[0159] The data at the location of outliers is replaced by linear interpolation of adjacent normal sampling points. For example, if the charging gun connector temperature sensor reads a jump value of 120°C due to electromagnetic interference at a certain moment, while the normal temperature before and after that moment is about 45°C, the outlier will be identified and removed.

[0160] In real-world charging environments, ambient temperature fluctuates over time due to factors such as changes in sunlight and ventilation conditions. These fluctuations synchronously affect the temperature readings of all temperature measurement points through heat conduction and heat radiation. Without environmental correction, subsequent cross-correlation analysis will misinterpret this synchronous change as a heat propagation relationship, thus interfering with heat source identification.

[0161] Therefore, it is necessary to subtract the ambient temperature at the same time from the temperature data of each temperature sensor at each moment after removing outliers to obtain the ambient temperature correction data. This temperature data only reflects the net temperature rise of the measurement point relative to the environment, eliminating the interference of the common-mode component of the ambient temperature.

[0162] S200. Extract the temperature time series of any two temperature measurement points within a preset time window from the preprocessed temperature data, and record them as the first temperature series. Second temperature sequence .in, , The sequence length is given.

[0163] The setting of the preset time window needs to balance the requirements of calculation accuracy and real-time performance:

[0164] A time window that is too short will result in insufficient data points for cross-correlation calculations, making the estimated heat propagation delay unstable. A time window that is too long will introduce too much historical data, reducing the response speed to the current heat propagation state.

[0165] In practical applications, the preset time window is typically set to 60 to 120 seconds, corresponding to the sequence length. It depends on the sampling frequency. For example, at a sampling frequency of 1 Hz, the sequence length L = 120 corresponds to a 120-second window.

[0166] Calculate the normalized cross-correlation coefficients of the first and second temperature series at different time offsets, for the time offsets. Normalized cross-correlation coefficients The calculation formula is:

[0167] ;

[0168] Normalized cross-correlation function is used to measure the first temperature series. With the second temperature sequence At different time offsets The degree of similarity of the waveforms.

[0169] Normalization processing makes The value range is limited to [-1, 1], which eliminates the influence of the absolute amplitude difference of temperature data on the similarity measurement and makes the cross-correlation coefficients between different temperature measurement point pairs comparable.

[0170] when A value close to 1 indicates that the variation pattern of the second temperature sequence differs from that of the first temperature sequence in terms of delay. The consistent patterns of change over time indicate that the delay in heat transfer from the first temperature measurement point to the second temperature measurement point is approximately [missing information]. .

[0171] By traversing different Value and search The maximum value can be used to determine the heat propagation delay between the two temperature measurement points.

[0172] In the formula, This is the mean of the first temperature sequence; This is the mean of the second temperature sequence;

[0173] and These are the arithmetic mean values ​​of the two temperature series, used for mean removal to eliminate the influence of the DC component, so that the cross-correlation coefficient reflects only the fluctuation similarity of temperature changes rather than the similarity of absolute temperature levels.

[0174] The numerator in the formula is the cross-covariance term (the upper limit of the summation is...). Because when the second sequence shifts The number of sample points that can be paired later is reduced. (number), in the denominator and The square roots of the variances of the first and second temperature sequences are used to normalize the molecules to the [-1, 1] interval.

[0175] This is the time offset, and its value range is... . This is the preset maximum heat propagation delay limit; The setting is based on the maximum possible heat transfer time between any two temperature measuring points in the physical structure of the charging pile:

[0176] In actual charging pile structures, the heat conduction path from the power module heat sink fins to the charging gun connector via the metal structural components is typically between 0.5 and 2 meters. The heat conduction rate of metal is on the order of tens of millimeters per second, so the maximum heat propagation delay is typically in the range of 10 to 60 seconds.

[0177] set up The purpose is to limit the search range of time offset for cross-correlation calculations to avoid invalid searches for large delay values, while also preventing unrelated random fluctuations from being misjudged as long-delay heat propagation relationships.

[0178] The second temperature sequence in time offset Next The value at each sampling point.

[0179] Iterate through all values ​​within the range in ascending integer increments and calculate the corresponding normalized cross-correlation coefficients. Arranged in ascending order of time offset, they form a cross-correlation function curve.

[0180] Extract the time offset corresponding to the maximum value of the normalized cross-correlation coefficient in the cross-correlation function curve as the heat propagation delay between the two temperature measurement points, and construct the heat propagation delay matrix by traversing all temperature measurement point pairs.

[0181] The heat propagation delay matrix is ​​a × square array ( (total number of temperature measurement points), the first in the matrix Line 1 The elements of the column represent the first... From the first temperature measurement point to the... The thermal propagation delay between temperature measurement points is represented by a diagonal element of 0, indicating that the delay of the temperature measurement point itself is zero.

[0182] Because heat propagation is directional (from high-temperature nodes to low-temperature nodes), the delay matrix is ​​not necessarily symmetric.

[0183] Traversing all temperature measurement points means... Each temperature measurement point is paired up, and a total of [number] calculations are required. For each combination of temperature measurement points, a heat propagation delay value is determined based on the cross-correlation function.

[0184] Each temperature measurement point is used as a node in the heat propagation topology graph, and directed edges are established for temperature measurement point pairs whose heat propagation delay is greater than a preset delay threshold in the heat propagation delay matrix.

[0185] The direction of the directed edge is from the temperature measurement point with a smaller thermal propagation delay to the temperature measurement point with a larger thermal propagation delay. No directed edge is established for temperature measurement points with a thermal propagation delay less than or equal to a preset delay threshold.

[0186] The direction of a directed edge is determined based on the physical causality of heat propagation:

[0187] Heat always propagates from the source node where the temperature changes first to the downstream node where the temperature changes later. Therefore, the temperature measurement point with a smaller heat propagation delay (i.e., the temperature changes first) is the heat source end, and the temperature measurement point with a larger heat propagation delay (i.e., the temperature changes later) is the heat receiving end.

[0188] Directed edges point from the end with smaller delay to the end with larger delay. This directional convention ensures that the topology graph can correctly reflect the physical direction of heat conduction.

[0189] Calculate the in-degree and out-degree of each node in the heat propagation topology graph:

[0190] In a heat propagation topology graph, the in-degree of a node refers to the number of directed edges pointing to that node, indicating how many upstream heat sources propagate heat to that temperature measurement point.

[0191] The out-degree of a node refers to the number of directed edges originating from that node, indicating how many downstream nodes the temperature measuring point will propagate heat to.

[0192] A zero in-degree means that the node has no upstream heat source and is the starting point for heat propagation, while a large out-degree means that the node is the heat source for multiple downstream nodes and is located upstream in the heat propagation chain.

[0193] When there is a node with an in-degree of zero in the heat propagation topology graph, it means that there are no directed edges pointing to it in the heat propagation topology graph. This temperature measuring point is not affected by the heat conduction of any other temperature measuring point, and its temperature change is entirely caused by its own heat generation. Therefore, it is marked as a root heat source.

[0194] Nodes with an in-degree greater than zero are marked as subordinate heat sources, indicating that the temperature measurement point is affected by heat conduction from at least one upstream heat source, and its temperature change includes a component of propagated heat.

[0195] If the in-degree of the temperature measuring point on the power module heat sink fin is zero, while the in-degree of the temperature measuring point on the charging gun connector is greater than zero (there is a directed edge from the power module to the charging gun connector), then the power module heat sink fin is the root heat source, and the charging gun connector is the subordinate heat source.

[0196] When there are no nodes with an in-degree of zero in the heat propagation topology graph, it indicates that there is a loop structure in the graph, that is, all nodes have upstream nodes pointing to form a cyclic heat propagation relationship. In actual charging piles, this cyclic topology may be caused by the small difference in heat propagation delay between each temperature measurement point and the close cross-correlation coefficient, resulting in the formation of a closed loop by directed edges.

[0197] Since the root heat source cannot be directly identified by its zero in-degree, the principle of maximizing out-degree is adopted as an alternative strategy.

[0198] The node with the largest outgoing degree means that this temperature measuring point transmits heat to the most downstream nodes, and is most likely the root heat source in the upstream position of the heat propagation chain. Therefore, the node with the largest outgoing degree is marked as the root heat source, and the other nodes are marked as subordinate heat sources.

[0199] When there are multiple root heat sources in the heat propagation topology diagram, it indicates that there are multiple independent heat generation starting points inside the charging pile (such as the power module and charging gun connector may exist as independent heat sources at the same time). In this case, it is necessary to identify the main heat source that has the greatest impact on overall thermal safety.

[0200] Calculate the sum of the heat propagation delays for all directed edges along the shortest propagation path from each root heat source to each subordinate heat source. The shortest propagation path is the path in the heat propagation topology graph that traverses the fewest directed edges or has the smallest sum of heat propagation delays between a root heat source node and a subordinate heat source node. The sum of the heat propagation delays for all directed edges along the path reflects the total time required for heat to propagate from the root heat source to the subordinate heat source.

[0201] The reciprocal of the sum of heat propagation delays is taken as the propagation intensity of the corresponding root heat source to the subordinate heat source, and the root heat source with the largest sum of propagation intensity is marked as the main heat source.

[0202] Preserve the directed edges from the main heat source to each subordinate heat source and their corresponding heat propagation delays, and remove the directed edges emanating from the remaining root heat sources to obtain the final heat propagation topology.

[0203] The complex topology of multiple heat sources is simplified into a tree structure with the main heat source as the only root node, so that the subsequent temperature component decomposition calculation has a clear propagation hierarchy and a unique starting point.

[0204] After removing directed edges originating from non-primary heat sources, the final heat propagation topology retains the complete propagation path from the primary heat source to each subordinate heat source, and each subordinate heat source retains only the directed edges originating from the primary heat source.

[0205] S300. Analyze the temperature measurement point corresponding to the root heat source in the heat propagation topology diagram, take the temperature change of the temperature measurement point at the current time relative to the charging start time as the self-heating component, and set the propagation heat component of the temperature measurement point to zero.

[0206] For a temperature measurement point corresponding to a subordinate heat source, obtain all upstream nodes of that temperature measurement point in the heat propagation topology diagram. For each upstream node... Read the upstream node At a historical moment The self-heating component value has been calculated and stored.

[0207] Historical moment For the current moment Retrospective heat propagation delay At the time obtained, the upstream node At a historical moment The self-heating component value at that location has been calculated in the historical control cycle corresponding to that historical moment.

[0208] Multiply the self-heating component value by the propagation attenuation coefficient The summation over all upstream nodes gives the temperature reading at the current time for that temperature measurement point. The heat component of propagation. The formula for calculating the propagation attenuation coefficient is:

[0209] ;

[0210] Propagation attenuation coefficient Used to quantify heat from upstream nodes Propagation to downstream nodes The degree of attenuation over time, that is, what proportion of the self-heating component of the upstream node can be conducted to the downstream node to form a propagating heat component.

[0211] Combining factors from two dimensions:

[0212] First, the statistical correlation of temperature changes between the two nodes is determined by the maximum value of the normalized cross-correlation coefficient. It is reflected in the range of [0,1]. The larger the value, the closer the thermal coupling between the two points and the higher the heat transfer efficiency.

[0213] Second, heat dissipates and decays naturally over time and distance along the propagation path, as indicated by the exponential decay term. This value reflects that... An increase or decrease indicates that the longer the propagation delay, the more heat is dissipated.

[0214] The product of the two factors allows the propagation attenuation coefficient to comprehensively reflect the coupling strength and dissipation characteristics of heat propagation, which is more in line with the dissipation law of heat when it actually propagates along the physical path compared to existing methods that only use the correlation coefficient as the propagation weight.

[0215] In the formula, upstream node Arrive at this temperature measurement point The propagation attenuation coefficient; the larger the value, the higher the upstream node. The heat is transmitted to the temperature measuring point. The higher the efficiency, the smaller the attenuation.

[0216] For this temperature measurement point With upstream nodes The maximum value of the normalized cross-correlation coefficient between the two is used as a benchmark factor for the propagation attenuation coefficient to reflect the conduction capacity of the heat propagation path.

[0217] For this temperature measurement point With upstream nodes The delay in heat propagation between them; This represents the sampling interval for temperature data. Converting the delay from the number of sampling points to actual physical time (seconds) ensures that the attenuation calculation has consistent physical dimensions (e.g., if...). Each sampling point The actual delay time is 8 seconds (if the delay time is 1 second).

[0218] The preset heat dissipation coefficient is a constant greater than zero; it reflects the rate of heat dissipation per unit time as heat propagates in the physical structure of the charging pile.

[0219] The value of is related to the heat dissipation conditions of the charging pile, the thermal conductivity of the structural materials, and other physical properties. The larger the value, the better the heat dissipation and the faster the heat is dissipated. The value can be obtained through experimental calibration.

[0220] Subtract the temperature change of the corresponding temperature measuring point at the current moment from the temperature change at the start of charging from the temperature change at the current moment of the heat source. The heat component of the propagation is used to obtain the temperature at the current time of the temperature measurement point. The self-heating component.

[0221] According to the topological sorting of the heat propagation topology diagram, the temperature data of each temperature measurement point is decomposed into self-heating component and propagation heat component.

[0222] Calculate the rate of change of the self-heating component at each temperature measurement point at the current moment, denoted as the rate of change of self-heating. Substitute this into the formula to calculate the heat contribution weight of each temperature measurement point:

[0223] ;

[0224] Heat contribution weight Used to quantify the The contribution ratio of the self-heating component of each temperature measuring point to the overall charging safety assessment is taken in the range of (0,1), and the sum of the thermal contribution weights of all temperature measuring points is 1.

[0225] By using the absolute amplitude of the self-heating component as a benchmark and introducing a topology depth penalty factor, temperature measurement points with larger self-heating amplitudes and closer proximity to the root heat source (smaller topology depth) receive higher weight in the safety assessment.

[0226] In the formula, For the first The heat contribution weight of each temperature measurement point; For the first At the current moment, each temperature measurement point... The self-heating component; Let be the topological depth of the j-th temperature measurement point in the heat propagation topology map.

[0227] The topology depth is the number of intermediate temperature measurement points along the shortest path from the root heat source to the temperature measurement point in the heat propagation topology graph. The topology depth of the temperature measurement point corresponding to the root heat source is zero.

[0228] The penalty factor when the topological depth of the temperature measurement point corresponding to the root heat source is zero. That is, no decay, the penalty factor for a direct downstream node topology depth of 1 is . And so on.

[0229] is the depth penalty coefficient, a preset constant greater than zero; used to control the decay rate of weights with topology depth. The greater the depth, the stronger the punishment.

[0230] This represents the total number of temperature measurement points. This is the summation index variable, used to iterate through all... The calculation of the self-heating contribution of all temperature measuring points in the denominator achieves weight normalization.

[0231] It is a preset small constant greater than zero, used to ensure that each temperature measuring point receives an equal heat contribution weight when the self-heating component of all temperature measuring points is zero.

[0232] Calculate the charging safety index The formula is:

[0233] ;

[0234] Charging safety index The core indicator used to comprehensively quantify the current thermal safety status of charging piles.

[0235] The self-heating and propagation heat components of all temperature measurement points are weighted and summed according to their respective heat contribution weights, and a dynamic amplification factor is introduced to nonlinearly adjust the rate of change of self-heating. The higher the value, the higher the overall thermal safety risk of the charging station.

[0236] In the formula, For charging safety index; For the first At the current moment, each temperature measurement point... The heat transfer component reflects the temperature change conducted from the upstream heat source, for the root heat source. .

[0237] For the first The rate of change of self-heating at each temperature measurement point at the current moment; and The weighting coefficients are preset and satisfy: .

[0238] in, This is the weighting coefficient for the self-heating component, used to make the safety index more sensitive to the self-heating component. This is the weighting coefficient for the propagation heat component, used to adjust the contribution of the propagation heat component to the safety index.

[0239] The self-heating component, which directly reflects the abnormal heating level of the temperature measuring point itself, is the primary focus of safety risk assessment, while the propagated heat component, which is the result of normal heat conduction, has a relatively lower risk level.

[0240] This is the dynamic sensitivity coefficient, a preset constant greater than zero; it is used to control the sensitivity of the dynamic amplification factor to the rate of change of self-heating. The higher the safety index, the more sensitive it is to rapid self-heating and the easier it is to trigger protective actions in the early stages of temperature rise.

[0241] The dynamic amplification factor is 1 when the rate of change of self-heating is zero, and greater than 1 when the rate of change of self-heating is greater than zero.

[0242] In actual operation of charging piles, the development of local overheating faults (such as a sudden increase in the contact resistance of the charging gun connector) usually goes through a process of "slow heating - accelerated heating - rapid heating".

[0243] During the slow heating phase, the absolute temperature value has not yet exceeded the traditional fixed threshold, but the rate of change of self-heating has begun to increase. Traditional monitoring methods based on fixed thresholds cannot respond in time during this phase.

[0244] The dynamic amplification factor of this invention can capture abnormal increases in the rate of self-heating change. A significant increase occurs at this stage, triggering the current reduction protection in advance and eliminating potential safety hazards at the nascent stage.

[0245] For example, when the self-heating component at a certain temperature measuring point increases at a rate of 0.5℃ per second and At this time: the dynamic amplification factor is 2, and the safety contribution of this temperature measurement point is amplified by one time, significantly improving it. Sensitivity to early temperature rise anomalies.

[0246] S400: Read the insulation resistance at the current moment. When the insulation resistance is less than the preset insulation threshold, set the charging current adjustment to the negative value of the current charging current and cut off the output current of the charging pile.

[0247] Insulation resistance is a key parameter reflecting the electrical insulation status of a charging pile. When the insulation resistance is lower than the preset insulation threshold, it indicates that the insulation performance of the charging pile's internal or output circuit has been severely deteriorated, posing a direct risk of leakage or even short circuit.

[0248] At this point, thermal safety assessment is no longer the primary concern; the charging current must be immediately cut off to ensure personal and equipment safety. Therefore, insulation resistance assessment, as an independent safety condition, takes precedence over the charging safety index. The assessment and implementation.

[0249] When the insulation resistance is not less than the preset insulation threshold, a first safety threshold is set. Second security threshold And satisfy: .

[0250] When charging safety index Less than the first safety threshold At this time, the charging current adjustment is set to zero to maintain the current charging current unchanged.

[0251] When charging safety index Greater than or equal to the first safety threshold And less than the second safety threshold At that time, the charging current adjustment is calculated according to a non-linear ratio. This reduces the output current of the charging station.

[0252] when When the charging station is in the warning zone, it has already shown a certain degree of thermal safety risk. It is necessary to reduce the charging current to reduce the heat generation and suppress the temperature from rising further.

[0253] The specific calculation formula is as follows:

[0254] ;

[0255] according to The current reduction is calculated using a non-linear proportional method based on the relative position within the warning zone: The closer (That is, the higher the risk) the greater the decrease in current. The closer (That is, the lower the risk) the smaller the decrease in current.

[0256] Adjusted charging current for:

[0257] ;

[0258] In the formula, The current charging current; that is, the actual output current of the charging pile in the current control cycle, in... As a benchmark, the current adjustment is proportional to the current level to ensure that the current reduction ratio is consistent in different charging stages.

[0259] The preset curvature index is a positive integer greater than 1 (usually taken as...). ); used to control the nonlinearity of the current adjustment curve, The bigger Just over The smaller the drop in flow at that time, the closer to The greater the drop in flow rate at that time.

[0260] Nonlinear exponent Make the current adjustment curve concave in shape, Just over The price drop was relatively small (allowing for some tolerance), in near The rate of decrease increases dramatically (rapidly suppressing risk escalation), achieving a balance between a smooth transition and a rapid response.

[0261] The charging safety index is recalculated in each control cycle. And update the charging current adjustment amount. The control cycle is the time interval between the main control MCU microcontroller executing a complete safety assessment and current adjustment process, typically ranging from 1 to 5 seconds.

[0262] The system recalculates within each control cycle. And according to Calculate the new current adjustment amount within the safe range.

[0263] When charging safety index Falling back to the first safety threshold At this point, the charging current is gradually restored to improve charging efficiency. The amount of current restoration is calculated. Gradually restore the charging current to the rated charging current. The calculation formula is:

[0264] ;

[0265] When charging safety index Greater than or equal to the second safety threshold When the charging current adjustment is set to the negative value of the current charging current, the output current of the charging pile is cut off.

[0266] The charging pile's output current is dynamically adjusted based on the charging current adjustment amount to achieve closed-loop safety control of the charging process.

[0267] The present invention also provides a charging pile safety control system based on multi-source data fusion, including a data acquisition module, a storage module, a communication module and a main control MCU microcontroller.

[0268] The data acquisition module is responsible for acquiring data at the perception layer, the storage module is responsible for data caching and management at the data layer, the main control MCU microcontroller is responsible for data processing and decision-making at the computing layer, and the communication module is responsible for information transmission at the interaction layer.

[0269] The data acquisition module includes multiple temperature sensors and insulation resistance detection circuits for collecting multi-source operating data. The temperature sensors are located at the charging gun connector, the charging cable sheath, and the heat sink fins of the power module.

[0270] The storage module is used to store multi-source operating data and cached data from the main control MCU microcontroller.

[0271] The communication module is used to interact with the electric vehicle battery management system.

[0272] The main control MCU microcontroller is connected to the data acquisition module, storage module and communication module, and is used to perform cross-correlation function calculation, heat propagation topology construction, temperature component decomposition and dynamic adjustment of charging current.

[0273] Example 1: This example uses a DC charging pile with a rated power of 120 kW as the experimental platform, with a rated output voltage of 500 volts and a rated output current of 240 amps.

[0274] The charging pile's internal control board integrates a data acquisition module, a storage module, a communication module, and a main control microcontroller.

[0275] The data acquisition module includes three NTC thermistor temperature sensors with a measurement accuracy of ±0.5℃ and a response time of less than 1 second. These sensors are located on the heat sink fins of the power module, the outer sheath of the charging cable, and the charging gun connector, respectively. An ambient temperature sensor is also located at the vent of the casing, along with an insulation resistance detection circuit.

[0276] The sampling frequency is set to 1 Hz, the control period is 2 seconds, the time window length is 60 seconds, the maximum heat propagation delay is set to 40 seconds, the delay threshold is set to 5 seconds, the heat dissipation coefficient is 0.04 per second, the depth penalty coefficient is 0.3, the first safety threshold is 0.4, the second safety threshold is 0.8, and the weighting coefficient is... , The dynamic sensitivity coefficient is set to 2.0, the curvature index is set to 2, the insulation threshold is set to 1 megohm, and the reference temperature rise is set to 100℃ for safety index normalization.

[0277] The experiment analyzed the entire fast charging process of an electric vehicle, with an ambient temperature of approximately 25°C. After charging commenced, the current climbed to 240 amps within 30 seconds and maintained a constant current. At 300 seconds of charging, all temperature measurement points entered a thermal steady state.

[0278] The power module temperature is approximately 55°C, the cable temperature is approximately 48°C, the connector temperature is approximately 52°C, the ambient temperature is approximately 25.2°C, and the insulation resistance is approximately 160 megohms. At this temperature, the charging safety index is approximately 0.23, maintaining the rated current output.

[0279] At 310 seconds of charging, an increased contact resistance fault was introduced at the charging gun connector to analyze for poor contact. Afterward, the connector temperature rose sharply, while the cable and module temperatures showed a delayed increase due to hysteresis in heat conduction.

[0280] Taking a time window of 310 to 370 seconds, the normalized cross-correlation functions of the three temperature series, after outlier removal using the Laida criterion and subtraction of ambient temperature, were calculated pairwise. The results are as follows:

[0281] The thermal propagation delay from connector to cable is 18 seconds with a maximum cross-correlation coefficient of 0.85; the delay from connector to module is 30 seconds with a coefficient of 0.62; and the delay from cable to module is 14 seconds with a coefficient of 0.78.

[0282] All of the above delays are greater than the 5-second threshold, and a directional heat propagation topology is constructed accordingly.

[0283] Topology analysis shows that the joint node with an in-degree of zero and an out-degree of two is marked as the root heat source; the cable node with an in-degree of one and an out-degree of one, and the module node with an in-degree of two and an out-degree of zero are all marked as subordinate heat sources; temperature component decomposition is performed according to topology sorting.

[0284] Taking the 370-second mark as an example, compared to the start of charging, the total temperature rise of the connector is 50°C. Since it is a root heat source, its self-heating component is 50°C and its heat transfer component is zero.

[0285] The total temperature rise of the cable is 28°C, of ​​which the heat propagation component from the connector is about 21°C, corresponding to a propagation attenuation coefficient of 0.42, and the self-heating component is about 7°C; the total temperature rise of the module is 26°C, of ​​which the heat propagation component is about 13°C and the self-heating component is about 13°C.

[0286] Based on this, the heat contribution weights of each temperature measuring point are calculated to be 0.78 for the connector, 0.09 for the cable, and 0.13 for the module. The heat source connector occupies the dominant weight due to its large self-heating amplitude and zero topology depth.

[0287] Combining the self-heating rate of change, which is approximately 0.8°C per second at the connector, the dynamic amplification factor produces a significant nonlinear amplification on the connector term. The final calculated charging safety index is approximately 0.74, which is in the warning zone between the first and second safety thresholds.

[0288] Based on this, the current adjustment was calculated using a non-linear ratio, reducing the charging current from 240 amps to approximately 145 amps. After the current reduction, the connector heating power decreased, and the temperature gradually dropped.

[0289] At 450 seconds, the joint temperature reached approximately 66°C, and the safety index dropped below 0.28. Following the recovery formula, the current was gradually increased back to approximately 160 amps. Throughout the entire process, the insulation resistance remained greater than 1 megohm, and the insulation protection was not triggered.

[0290] No false alarms occurred during the entire experiment due to normal heat conduction. Local overheating at the joint was detected and current reduction was triggered even when the absolute temperature was still below the traditional fixed threshold of 80°C, verifying the invention's ability to distinguish heat source attributes and its early safety response capability. The temperature changes and safety index at each measuring point, along with the dynamic current adjustment process, are described below. Figure 2 and Figure 3 As shown.

[0291] Please see Figure 2 The horizontal axis represents charging time in seconds, and the vertical axis represents temperature in degrees Celsius (°C). The graph shows four curves: ambient temperature, power module temperature, charging cable temperature, and charging gun connector temperature.

[0292] The temperature of the four circuits remained stable for 310 seconds, with the ambient temperature maintained at around 25°C. The other three circuits reached a steady state as the charging power increased.

[0293] After 310 seconds, the connector temperature rose sharply due to abnormal contact. The cable temperature followed suit about 18 seconds later. The module temperature changed the most gradually, reflecting the delayed characteristics of heat propagation along the path from the connector to the cable to the module.

[0294] After the 370-second downflow was initiated, the temperature at each point dropped successively.

[0295] Please see Figure 3 The left vertical axis represents the charging safety index, and the right vertical axis represents the charging current, in amperes.

[0296] 310 seconds ago, the safety index was stable at around 0.2; 310 seconds later, due to the rapid rise in self-heating of the connector, the safety index climbed to a peak of 0.74 in about 60 seconds, entered the warning zone and triggered current reduction, with the current dropping from 240 amps to 145 amps;

[0297] Subsequently, the safety index gradually dropped below 0.3, and the current slowly recovered. The horizontal dashed lines in the figure indicate the first safety threshold of 0.4 and the second safety threshold of 0.8, respectively, and the dark area represents the current reduction protection interval.

[0298] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0299] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A charging pile safety control method based on multi-source data fusion, characterized in that: The method includes: S100: Collects multi-source operation data throughout the charging process through the data acquisition module inside the charging pile. The multi-source operation data includes temperature data from multiple temperature measurement points, ambient temperature, and insulation resistance. The multi-source operation data is then preprocessed. S200. Calculate the heat propagation delay between the temperature time series of any two temperature measurement points using the cross-correlation function, construct a heat propagation topology map based on the heat propagation delay, and identify the root heat source and dependent heat sources in the heat propagation topology map. S300: Based on the heat propagation topology map, the temperature data of each temperature measuring point is decomposed into self-heating component and propagation heat component in order of topology sorting, so as to calculate the heat contribution weight of each temperature measuring point and the charging safety index. S400: Set a safety threshold and calculate the charging current adjustment amount based on the deviation between the charging safety index and the safety threshold; dynamically adjust the output current of the charging pile based on the charging current adjustment amount to achieve closed-loop safety control of the charging process.

2. The charging pile safety control method based on multi-source data fusion according to claim 1, characterized in that: The S100 charging pile contains different types of hardware modules, including a data acquisition module, a storage module, a communication module, and a main control MCU microcontroller. The data acquisition module includes multiple temperature sensors and insulation resistance detection circuits, used to collect multi-source operating data; Multiple temperature sensors are respectively placed at the charging gun connector, the charging cable sheath and the power module heat sink fins to collect the connector temperature, cable temperature and module temperature as temperature data. The storage module is used to store multi-source operating data and cached data from the main control MCU microcontroller; the specific preprocessing method is as follows: The raw sampling data of each temperature sensor is resampled to the same sampling frequency, and outliers are removed from the resampled data using the Laida criterion. After removing outliers, the ambient temperature at each time step is subtracted from the temperature data of each temperature sensor at each time step to obtain the ambient temperature correction data.

3. The charging pile safety control method based on multi-source data fusion according to claim 1, characterized in that: S200 includes: S201. Extract the temperature time series of any two temperature measurement points within a preset time window from the preprocessed temperature data, and record them as the first temperature series and the second temperature series respectively. S202. Calculate the normalized cross-correlation coefficients of the first temperature sequence and the second temperature sequence at different time offsets, and arrange them in ascending order of time offset to form cross-correlation function curves. S203. Extract the time offset corresponding to the maximum value of the normalized cross-correlation coefficient in the cross-correlation function curve as the heat propagation delay between the two temperature measurement points. Traverse all temperature measurement point pairs and construct the heat propagation delay matrix.

4. The charging pile safety control method based on multi-source data fusion according to claim 3, characterized in that: In S202, the calculation of the normalized cross-correlation coefficient includes: Let the first temperature sequence be The second temperature sequence is ;in: , For sequence length; for time offset Normalized cross-correlation coefficients The calculation formula is: ； In the formula, This is the mean of the first temperature sequence; This is the mean of the second temperature sequence; This is the time offset, and its value range is... ; This is the preset maximum heat propagation delay limit; The second temperature sequence in time offset Next The value at each sampling point; Iterate through all values ​​within the range in ascending integer increments and calculate the corresponding normalized cross-correlation coefficients. .

5. The charging pile safety control method based on multi-source data fusion according to claim 3, characterized in that: In S203, constructing the heat propagation topology map based on the heat propagation delay matrix includes: S2031. Using each temperature measurement point as a node in the heat propagation topology graph, establish directed edges for temperature measurement point pairs whose heat propagation delay is greater than a preset delay threshold in the heat propagation delay matrix. S2032. The direction of the directed edge is from the temperature measurement point with a smaller thermal propagation delay to the temperature measurement point with a larger thermal propagation delay. No directed edge is established for temperature measurement points with a thermal propagation delay less than or equal to a preset delay threshold. S2033. Calculate the in-degree and out-degree of each node in the heat propagation topology graph: When there are nodes with an in-degree of zero in the heat propagation topology graph, the nodes with an in-degree of zero are marked as root heat sources, and the nodes with an in-degree greater than zero are marked as subordinate heat sources. When there are no nodes with an in-degree of zero in the heat propagation topology graph, the node with the largest out-degree is marked as the root heat source, and the remaining nodes are marked as subordinate heat sources. S2034. When there are multiple root heat sources in the heat propagation topology graph, calculate the sum of the heat propagation delays corresponding to all directed edges on the shortest propagation path from each root heat source to each subordinate heat source. S2035. The reciprocal of the sum of heat propagation delays is taken as the propagation intensity of the corresponding root heat source to the subordinate heat source, and the root heat source with the largest sum of propagation intensity is marked as the main heat source. S2036. Retain the directed edges from the main heat source to each subordinate heat source and their corresponding heat propagation delays, remove the directed edges from the remaining root heat sources, and obtain the final heat propagation topology graph.

6. The charging pile safety control method based on multi-source data fusion according to claim 1, characterized in that: In S300, the temperature data at each temperature measurement point is decomposed into a self-heating component and a propagating heat component, including: S301. For the temperature measurement point corresponding to the root heat source, the temperature change of the temperature measurement point at the current moment relative to the charging start moment is taken as the self-heating component, and the heat propagation component of the temperature measurement point is set to zero. S302. For the temperature measurement point corresponding to the subordinate heat source, obtain all upstream nodes of the temperature measurement point in the heat propagation topology diagram. For each upstream node Read the upstream node At a historical moment The self-heating component value at that location is multiplied by the propagation attenuation coefficient. The summation over all upstream nodes gives the temperature reading at the current time for that temperature measurement point. The propagation heat component; the formula for calculating the propagation attenuation coefficient is: ； In the formula, upstream node Arrive at this temperature measurement point The propagation attenuation coefficient; For this temperature measurement point With upstream nodes The maximum value of the normalized cross-correlation coefficient between them; For this temperature measurement point With upstream nodes The delay in heat propagation between them; The sampling interval for temperature data; The preset heat dissipation coefficient; S303, Subtract the temperature change of the temperature measuring point at the current moment from the temperature change at the current moment relative to the start of charging from the temperature change of the temperature measuring point corresponding to the subordinate heat source at the current moment. The heat component of the propagation is used to obtain the temperature at the current time of the temperature measurement point. The self-heating component; S304. Perform all temperature measurement points sequentially according to the topological order of the heat propagation topology diagram to complete the decomposition of the self-heating component and the propagation heat component.

7. The charging pile safety control method based on multi-source data fusion according to claim 6, characterized in that: In the S300, the calculation of the thermal contribution weight of each temperature measurement point and the charging safety index includes: Calculate the rate of change of the self-heating component at each temperature measurement point at the current moment, denoted as the rate of change of self-heating; substitute it into the formula to calculate the heat contribution weight of each temperature measurement point: ； In the formula, For the first The heat contribution weight of each temperature measurement point; For the first At the current moment, each temperature measurement point... The self-heating component; Let be the topological depth of the j-th temperature measurement point in the heat propagation topology map; The topology depth is the number of intermediate temperature measurement points along the shortest path from the root heat source to the temperature measurement point in the heat propagation topology graph. The topology depth of the temperature measurement point corresponding to the root heat source is zero. This is the depth penalty coefficient; This represents the total number of temperature measurement points. For summation index variables; It is a preset small constant that is greater than zero; Charging safety index The calculation formula is: ； In the formula, For charging safety index; For the first At the current moment, each temperature measurement point... The heat component of propagation; For the first The rate of change of self-heating at each temperature measurement point at the current moment; and The weighting coefficients are preset and satisfy: ; is the dynamic sensitivity coefficient, which is a preset constant greater than zero.

8. The charging pile safety control method based on multi-source data fusion according to claim 1, characterized in that: The S400 includes: S401. Read the insulation resistance at the current moment. When the insulation resistance is less than the preset insulation threshold, set the charging current adjustment to the negative value of the current charging current and cut off the output current of the charging pile. S402. When the insulation resistance is not less than the preset insulation threshold, set the first safety threshold. Second security threshold And satisfy: ; When charging safety index Less than the first safety threshold At this time, the charging current adjustment is set to zero to maintain the current charging current unchanged; When charging safety index Greater than or equal to the first safety threshold And less than the second safety threshold At the same time, the charging current adjustment is calculated according to a non-linear ratio to reduce the output current of the charging pile; When charging safety index Greater than or equal to the second safety threshold When the charging current adjustment is set to the negative value of the current charging current, the output current of the charging pile is cut off.

9. The charging pile safety control method based on multi-source data fusion according to claim 8, characterized in that: The charging current adjustment calculated using a non-linear proportional method includes: Charging current adjustment The calculation formula is: ； Adjusted charging current for: ； In the formula, This is the current charging current; The preset curvature index is used; the charging safety index is recalculated in each control cycle. And update the charging current adjustment amount. ; When charging safety index Falling back to the first safety threshold The following steps involve substituting the values ​​into the formula to calculate the current recovery amount. Gradually restore the charging current to the rated charging current. The calculation formula is: 。 10. A charging pile safety control system based on multi-source data fusion, characterized in that: The system includes a data acquisition module, a storage module, a communication module, and a main control MCU microcontroller, used to implement the charging pile safety control method based on multi-source data fusion as described in any one of claims 1 to 9; The data acquisition module includes multiple temperature sensors and insulation resistance detection circuits for collecting multi-source operating data; the temperature sensors are respectively located at the charging gun connector, the charging cable sheath, and the heat sink fins of the power module. The storage module is used to store multi-source operating data and cached data from the main control MCU microcontroller. The communication module is used to interact with the electric vehicle battery management system; The main control MCU microcontroller is connected to the data acquisition module, storage module and communication module, and is used to perform cross-correlation function calculation, heat propagation topology construction, temperature component decomposition and dynamic adjustment of charging current.

Images