AI-based fire early warning system for new energy vehicle charging stations

Abstract

This invention relates to the field of fire early warning technology for charging stations, specifically disclosing a fire early warning system for new energy vehicle charging stations based on AI algorithms. Through a multimodal sensor array deployed within the charging station and the battery management system built into each charging pile in use, the system collects dynamic parameters of each charging pile in real time during the charging process. Based on these dynamic parameters, a three-dimensional dynamic thermodynamic model of each charging pile in use is constructed. This model then identifies the potential voltage overload risk coefficient of each charging pile in use. Finally, based on the potential voltage overload risk coefficient of each charging pile in use, a graded early warning signal is generated. This intelligent early warning system based on real-time monitoring and analysis not only improves the safety and reliability of charging piles but also provides users with a more convenient and safer charging experience, meeting the urgent needs for safety and intelligence in the current electric vehicle charging field.

Description

Technical Field

[0001] This invention belongs to the field of fire early warning technology for charging stations, and relates to a fire early warning system for new energy vehicle charging stations based on AI algorithms. Background Technology

[0002] The importance and necessity of the Battery Management System (BMS) in addressing current / voltage feedback lag in charging stations is self-evident. As a key component of electric vehicle charging stations, the BMS undertakes the crucial task of monitoring and managing battery status. However, the importance of monitoring current / voltage feedback is increasingly highlighted because it can lead to safety issues such as overload or voltage breakdown. Firstly, timely and accurate current / voltage feedback information helps the BMS monitor the battery charging status in real time, preventing overload. Overload can lead to battery performance degradation, shortened lifespan, and even safety hazards; therefore, monitoring and feedback of current / voltage are essential. Secondly, voltage breakdown is a serious safety issue that can cause severe damage to the battery and charging station, and even lead to dangerous situations such as fires. Therefore, real-time monitoring and feedback of voltage is crucial to ensuring the safe and stable operation of charging stations.

[0003] However, current battery management systems (BMS) suffer from several drawbacks due to the lag in monitoring the output current / voltage of charging stations. First, some traditional BMS systems employ monitoring technologies and algorithms that may exhibit lag, failing to achieve real-time monitoring and feedback of current / voltage, thus hindering the timely identification and response to overload or voltage breakdown risks. Second, some BMS systems have limited monitoring accuracy and response speed, failing to accurately capture sudden changes in current / voltage, significantly reducing the effectiveness of overload or voltage breakdown warnings. Summary of the Invention

[0004] In view of the problems existing in the prior art, the present invention provides a fire early warning system for new energy vehicle charging stations based on AI algorithms to solve the above-mentioned technical problems.

[0005] To achieve the above and other objectives, the technical solution adopted by the present invention is as follows:

[0006] This invention provides a fire early warning system for new energy vehicle charging stations based on AI algorithms. The system includes a dynamic parameter acquisition module, a dynamic parameter parsing module, and a multi-level linkage early warning module. The above modules are connected by wired and / or wireless connections to realize data transmission between the modules.

[0007] Dynamic parameter acquisition module: Through a multimodal sensor array deployed in the charging station and the battery management system built into each charging pile in use, the module collects the dynamic parameters of each charging pile in use in real time during the charging process; and constructs a three-dimensional dynamic thermodynamic model of each charging pile in use based on the dynamic parameters.

[0008] Dynamic parameter analysis module: Based on the three-dimensional dynamic thermodynamic model of each charging pile in use, identify the potential risk coefficient of voltage overload for each charging pile in use;

[0009] Multi-level linkage early warning module: Generates graded early warning signals based on the potential risk coefficient of voltage overload of each charging pile in use.

[0010] For example, the dynamic parameters of each charging pile during the charging process include the charging current, voltage, battery temperature, ambient temperature, ambient humidity, and insulation status of the internal cables of the charging pile.

[0011] For example, a three-dimensional dynamic thermodynamic model of the charging station is constructed based on the aforementioned dynamic parameters. The specific construction logic is as follows:

[0012] Step S11: Synchronously collect charging current, voltage and corresponding battery temperature data of each charging pile in use, and establish a mapping relationship between charging equipment and spatial coordinates;

[0013] Step S12: Calculate the real-time heat generation power of each charging pile in use based on the charging current data, and generate a three-dimensional heat source distribution map of the charging equipment.

[0014] Step S13: Combine the outer shell temperature and ambient temperature data of each charging pile in use to quantify the thermal resistance coefficient of the air medium and superimpose the thermal resistance coefficient of the air medium onto the three-dimensional heat source distribution map.

[0015] Step S14: Based on the internal cable insulation state parameters, calculate the additional heat generation rate caused by the impedance change of each charging pile cable in use, and superimpose the additional heat generation rate onto the three-dimensional heat source distribution map;

[0016] Step S15: Correct the thermal conductivity of the air medium based on the ambient humidity data, and construct a three-dimensional spatial temperature field model through iterative calculation method;

[0017] Step S16: Vector superposition of the three-dimensional heat source distribution map and the three-dimensional spatial temperature field model to construct a three-dimensional dynamic thermodynamic model of each charging pile in use.

[0018] For example, the specific implementation of step S13 includes:

[0019] Step S131: Scan the surface temperature field of each charging pile in use with an infrared thermal imager, extract the temperature gradient distribution features and mark local hot spots;

[0020] Step S132: Calculate the thermal resistance coefficients of natural convection and forced heat dissipation by combining data from ambient temperature and humidity sensors;

[0021] Step S133: Analyze the heat conduction path based on the principle of heat flow continuity and quantify the heat accumulation rate of key components;

[0022] Step S134: Calculate the thermal resistance coefficient of the air medium based on the product of the thermal resistance coefficient and the thermal accumulation rate.

[0023] For example, the identification logic for determining the potential risk factor of voltage overload for each charging station in use is as follows:

[0024] Step S21: Extract key thermal characteristic values ​​of each charging pile in use in real time from the three-dimensional dynamic thermodynamic model, including the core temperature of the battery compartment, the temperature gradient of the heat dissipation surface of the charging module, and the peak temperature of the cable channel.

[0025] Step S22: Calculate the conductor resistance increment based on the peak temperature distribution of the cable channel, and convert the resistance increment into an equivalent voltage drop compensation coefficient;

[0026] Step S23: Analyze the core temperature change curve of the battery compartment and establish a mapping relationship table between electrolyte activity decay and terminal voltage change;

[0027] Step S24: Based on the temperature gradient data of the charging pile heat dissipation surface, calculate the junction temperature of the power device and correlate it with the voltage ripple increment caused by switching losses;

[0028] Step S25: Based on the comprehensive voltage drop compensation coefficient, terminal voltage change mapping value, and voltage ripple increment, generate the voltage overload potential risk coefficient for each charging pile in use through a dynamic weight allocation algorithm combined with a weighted summation algorithm.

[0029] For example, the operation logic of step S22 is as follows:

[0030] Step S221: In the three-dimensional model of the charging station, delineate the three-dimensional coordinate range of the cable channel corresponding to each charging pile in use, and use the regional extreme value search algorithm to automatically identify the highest temperature point in the channel; take the peak point as the center, expand along the three-dimensional coordinate axis to form a detection cube, and calculate the arithmetic mean of all temperature measurement points in the cube as the representative temperature value;

[0031] Step S222: Obtain the original design parameters of the cable, including the standard temperature resistivity corresponding to the conductor material type and the resistance value per unit length of the cable at the standard ambient temperature; based on the actual total length of the cable, use the product of the resistance value per unit length and the total length as the reference resistance value;

[0032] Step S223: Subtract the standard ambient temperature from the average temperature of the current test cube to obtain the actual temperature difference value; multiply the reference resistance value by the temperature difference, and then multiply by the material temperature resistivity to obtain the resistance increment per unit length; superimpose the actual total length of the cable, and the total resistance increment is equal to the product of the resistance increment per unit length and the actual total length of the cable.

[0033] Step S224: Obtain the current charging current value passing through the cable in real time; multiply the total resistance increment by the charging current value to obtain the theoretical increase in voltage drop; divide the voltage drop increase by the rated output voltage of the charging pile, and then multiply by 100% to convert it into a percentage, thereby obtaining the equivalent voltage drop compensation coefficient.

[0034] For example, the operation logic of step S24 is as follows:

[0035] Based on the thermocouple sensors deployed on the heat dissipation surface of the charging pile, an X-axis coordinate system is established along the fin direction of the heat sink, and a Y-axis coordinate system is established in the vertical direction. After identifying the coordinates of the highest temperature point, a junction temperature inversion model is established through the thermal resistance parameters of the packaging structure. The highest measurement point temperature is multiplied by the material thermal conductivity correction factor, and the product of the thermal resistance parameter and the real-time power is superimposed to obtain the estimated junction temperature data value of the power device.

[0036] Based on this, a dynamic quantitative model of switching loss is constructed. Taking the test value of 25 degrees Celsius environment as the benchmark, a temperature compensation rule is set to increase the loss by 35% for every 100 degrees Celsius increase in junction temperature. The loss value is accumulated and calculated for the parallel operation of multiple devices.

[0037] Subsequently, ripple characteristic reconstruction is performed. The total switching loss is divided by the switching period to obtain the power fluctuation component. This component is then divided by the DC bus voltage value to convert it into an equivalent current ripple, and then multiplied by the total impedance of the charging circuit to obtain the voltage ripple increment.

[0038] For example, a graded early warning signal is generated based on the potential risk coefficient of voltage overload for each charging pile currently in use, including:

[0039] The voltage overload risk coefficient of each charging pile in use is compared with the preset safety assessment threshold range [S1,S2].

[0040] If the voltage overload risk factor of a charging pile in use is less than S1, then the warning level of the charging pile in use is marked as Level 1, and no warning is issued for the charging pile in use.

[0041] If the voltage overload risk factor of a charging pile in use is greater than S2, then the warning level of the charging pile in use will be marked as Level 3, the power supply to the charging pile in use will be automatically cut off, and the dry powder fire extinguishing device will be activated.

[0042] If the potential risk factor of a voltage overload of a charging station in use is within the safety assessment threshold range, the warning level of the charging station in use will be marked as Level 2, triggering the audible and visual alarm device and pushing the warning information through the mobile APP.

[0043] As described above, the AI-based fire early warning system for new energy vehicle charging stations provided by this invention has at least the following beneficial effects:

[0044] This invention provides an AI-based fire early warning system for new energy vehicle charging stations. By deploying a multi-modal sensor array and the charging pile's built-in battery management system, it collects dynamic parameters of the charging pile in real time during the charging process, constructs a three-dimensional dynamic thermodynamic model, identifies potential voltage overload risk coefficients, and generates graded early warning signals. This provides a comprehensive understanding of the state changes during charging, helping to promptly detect abnormalities and take corresponding measures to ensure the safe operation of the charging pile. Secondly, the three-dimensional dynamic thermodynamic model can more accurately simulate the working environment and thermal characteristics of the charging pile, helping to analyze the heat distribution and transfer during battery charging, providing a scientific basis for preventing battery overheating. Simultaneously, identifying potential voltage overload risk coefficients can effectively assess the voltage safety of the charging pile, promptly detect potential voltage overload risks, and avoid damage to the battery and charging equipment. Finally, based on the potential voltage overload risk coefficients, graded early warning signals can be generated, allowing for appropriate preventative measures to be taken according to the risk level, such as reducing charging power, stopping charging, or performing maintenance, thereby ensuring the safe operation of the charging pile and the electric vehicle. This intelligent early warning system based on real-time monitoring and analysis not only improves the safety and reliability of charging piles, but also provides users with a more convenient and safer charging experience, meeting the urgent needs of the current electric vehicle charging field for safety and intelligence. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a schematic diagram showing the connections of the various modules in the system of the present invention. Detailed Implementation

[0047] The following description, in conjunction with the implementation of this invention, is merely an example and illustration of the concept of this invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the inventive concept or exceed the scope defined in these claims, all of which should fall within the protection scope of this invention.

[0048] Example 1

[0049] Please see Figure 1 As shown, the fire early warning system for new energy vehicle charging stations based on AI algorithms includes a dynamic parameter acquisition module, a dynamic parameter parsing module, and a multi-level linkage early warning module. The above modules are connected by wired and / or wireless connections to realize data transmission between the modules.

[0050] Dynamic parameter acquisition module: Through a multimodal sensor array deployed in the charging station and the battery management system built into each charging pile in use, the module collects the dynamic parameters of each charging pile in use in real time during the charging process; and constructs a three-dimensional dynamic thermodynamic model of each charging pile in use based on the dynamic parameters.

[0051] The dynamic parameters of each charging pile during the charging process include the charging current, voltage, battery temperature, ambient temperature, ambient humidity, and insulation status of the internal cables of the charging pile.

[0052] A three-dimensional dynamic thermodynamic model of the charging station is constructed based on the aforementioned dynamic parameters. The specific construction logic is as follows:

[0053] Step S11: Synchronously collect charging current, voltage and corresponding battery temperature data of each charging pile in use, and establish a mapping relationship between charging equipment and spatial coordinates;

[0054] In practical implementation, a multimodal sensor array is installed within the charging station in a three-dimensional spatial grid layout. Temperature sensing units are installed on the top, sides, and base of each charging pile in use, and dual-mode current and voltage sensors are configured at the charging gun connection port. The charging current data includes, but is not limited to, instantaneous current values ​​and the effective root mean square value of a 15-second sliding window; the battery temperature data is the battery cell temperature gradient data, which can be obtained through multi-point distributed temperature measurement; then, through the association and matching of the GPS positioning module and the physical code of the charging pile, combined with the equipment coordinates in the charging station's BIM model, a three-dimensional coordinate label with elevation values ​​is established for each working charging pile. The spatial coordinate system is set with the center of the charging station ground as the origin, establishing orthogonal coordinate axes of X (east-west), Y (north-south), and Z (vertical).

[0055] Step S12: Calculate the real-time heat generation power of each charging pile in use based on the charging current data, and generate a three-dimensional heat source distribution map of the charging equipment.

[0056] In practice, the DC resistance parameters of the cables of different charging piles in use are obtained by measurement or calibration. The square of the real-time current value is multiplied by the resistance value to obtain the heat generation power per unit time. Further considering the switching losses of the power electronic devices inside the charging pile, a correction coefficient is added to finally generate the real-time heat generation power value for each charging pile in use. In the construction of the three-dimensional heat source distribution map, the real-time heat generation power values ​​of the spatial coordinates of each charging pile in use are imported into the three-dimensional modeling software, and a continuous three-dimensional heat source density field is generated through interpolation algorithms. For example, if two adjacent charging piles in a certain area generate 10W and 15W of heat respectively, the software will generate the transition heat source value of the area between them based on spatial distance weights and visualize it in the form of a heat map.

[0057] Step S13: Combine the outer shell temperature and ambient temperature data of each charging pile in use to quantify the thermal resistance coefficient of the air medium and superimpose the thermal resistance coefficient of the air medium onto the three-dimensional heat source distribution map.

[0058] The specific implementation of step S13 includes:

[0059] Step S131: Scan the surface temperature field of each charging pile in use with an infrared thermal imager, extract the temperature gradient distribution features and mark local hot spots;

[0060] Step S132: Calculate the thermal resistance coefficients of natural convection and forced heat dissipation by combining data from ambient temperature and humidity sensors;

[0061] Step S133: Analyze the heat conduction path based on the principle of heat flow continuity and quantify the heat accumulation rate of key components;

[0062] Step S134: Calculate the thermal resistance coefficient of the air medium based on the product of the thermal resistance coefficient and the thermal accumulation rate.

[0063] In practice, an infrared thermal imager is used to perform a panoramic scan of the charging pile casing at a sampling rate of 30 frames per second, generating a two-dimensional temperature distribution thermal map with a resolution of 640×480 pixels. Using the image processing algorithm built into the edge computing unit, the thermal image is divided into 5mm×5mm grid cells, and the temperature difference between adjacent grid cells is calculated as the gradient value. Based on a set gradient threshold, potential hotspot areas are automatically marked (for example, if three consecutive high-gradient grid cells appear near the charging gun plug, it is identified as a heat dissipation anomaly).

[0064] During implementation, environmental data was collected using temperature and humidity sensors. The natural convection thermal resistance and forced heat dissipation thermal resistance were calculated using a classical thermal resistance model.

[0065] Based on the correlation between Prandtl number and Grashof number, the temperature difference (shell temperature - ambient temperature) and characteristic length are substituted into the empirical formula to obtain the natural convection thermal resistance; the fan outlet wind speed is measured by an anemometer, the flow state is determined based on the Reynolds number, and the forced convection thermal resistance is calculated by combining the Nusselt number correlation.

[0066] The combined thermal resistance is calculated by connecting the two in parallel and then associated with the mesh nodes of the 3D model.

[0067] In the heat flow path analysis, the internal structure of the charging pile is divided into a heat generation core area, a heat transfer path area, and a heat dissipation terminal area. Heat flow sensors are used to measure the heat flow density between the interfaces of each area, and combined with the physical properties of the heat-conducting medium, the difference in heat flow at each node per unit time is calculated. Based on the heat capacity formula, the heat difference is divided by the material's heat capacity and mass to obtain the thermal deposition rate.

[0068] Step S14: Based on the internal cable insulation state parameters, calculate the additional heat generation rate caused by the impedance change of each charging pile cable in use, and superimpose the additional heat generation rate onto the three-dimensional heat source distribution map;

[0069] Cable insulation aging leads to increased impedance, thereby increasing Joule heat loss. In this step, the equivalent impedance value of the internal cables of each charging pile in use is first monitored online using an impedance spectrum analyzer and compared with the initial calibration value to calculate the impedance change ratio. The additional heat generation rate is obtained by multiplying the square of the current by the impedance increment. The additional heat generation rate is incrementally added to the corresponding nodes of each charging pile in use in the original three-dimensional heat source distribution map. At the same time, the distribution of adjacent heat source nodes is interpolated and smoothed to avoid model distortion caused by local numerical abrupt changes.

[0070] Step S15: Correct the thermal conductivity of the air medium based on the ambient humidity data, and construct a three-dimensional spatial temperature field model through iterative calculation method;

[0071] During implementation, capacitive humidity sensors deployed within charging stations are used to monitor air humidity in real time. Based on the nonlinear relationship between humidity and air thermal conductivity, the air's thermal conductivity coefficient is dynamically corrected. Specifically, an empirical formula is used to map humidity values ​​to a correction factor for thermal conductivity. When constructing the three-dimensional temperature field model, the physical space of each area where a charging pile is in use is discretized into three-dimensional grid cells. Based on the corrected air thermal conductivity coefficient, the heat exchange between each grid cell and the environment is calculated. The iterative process uses the Gauss-Seidel algorithm to gradually approximate the steady-state temperature distribution: assuming the initial temperature of all grid cells is the ambient temperature, the temperature value of each grid cell is updated successively based on the temperature difference between adjacent grid cells and the corrected thermal conductivity coefficient, until the maximum temperature change of the entire grid in two consecutive iterations is less than a set threshold, at which point convergence is considered achieved.

[0072] Step S16: Vector superposition of the three-dimensional heat source distribution map and the three-dimensional spatial temperature field model to construct a three-dimensional dynamic thermodynamic model of each charging pile in use.

[0073] In the simulation platform, the three-dimensional heat source distribution map and the three-dimensional spatial temperature field model are spatially vector-superimposed: for each grid cell, the combined effect of the heat source vector direction and the temperature gradient vector is calculated, and the dynamic heat conduction process is simulated through a time-stepping algorithm. Finally, a three-dimensional dynamic thermodynamic model of each charging pile in use is generated.

[0074] Dynamic parameter analysis module: Based on the three-dimensional dynamic thermodynamic model of each charging pile in use, identify the potential risk coefficient of voltage overload for each charging pile in use;

[0075] The logic for identifying the potential risk factor of voltage overload for each charging station in use is as follows:

[0076] Step S21: Extract key thermal characteristic values ​​of each charging pile in use in real time from the three-dimensional dynamic thermodynamic model, including the core temperature of the battery compartment, the temperature gradient of the heat dissipation surface of the charging module, and the peak temperature of the cable channel.

[0077] Step S22: Calculate the conductor resistance increment based on the peak temperature distribution of the cable channel, and convert the resistance increment into an equivalent voltage drop compensation coefficient;

[0078] Step S23: Analyze the core temperature change curve of the battery compartment and establish a mapping relationship table between electrolyte activity decay and terminal voltage change;

[0079] The operation logic of step S23 is as follows:

[0080] First, a three-dimensional temperature monitoring system is established in the battery compartment. The battery pack is divided into three layers: upper, middle and lower. Three embedded thermocouple sensors are arranged in each layer. Through continuous monitoring, it is determined that when the temperature of a certain middle layer exceeds the average temperature of other layers by two degrees Celsius for ten minutes, it is identified as the hot core area.

[0081] Based on this, an electrolyte activity decay model is constructed, setting 25-30 degrees Celsius as the baseline temperature range corresponding to 100% ion migration activity. When the monitored temperature exceeds this range, basic decay calculations are performed according to the rule that the activity decreases linearly by 0.6% for every 1 degree Celsius increase. For high-temperature conditions exceeding 45 degrees Celsius, an additional exponential amplification factor is introduced, and an irreversible activity loss of 0.2% per hour is accumulated. When converting the activity decay into terminal voltage drop, a direct mapping relationship between the percentage of activity decay and the voltage value is established, where each 1% activity decay corresponds to a base voltage drop of 0.08 volts. For high-temperature regions above 40 degrees Celsius, a compensation of 0.005 volts is added for every 1 degree Celsius increase. When the battery cycle count exceeds 500 cycles, a 1.2x aging compensation factor is activated. Based on the above rules, a three-dimensional parameter mapping table is generated, which includes five dimensions: core temperature, duration, activity decay, theoretical voltage drop, and aging compensation coefficient. The table is calculated by nearest neighbor interpolation by matching the current temperature-time combination in real time.

[0082] Step S24: Based on the temperature gradient data of the charging pile heat dissipation surface, calculate the junction temperature of the power device and correlate it with the voltage ripple increment caused by switching losses;

[0083] Step S25: Based on the comprehensive voltage drop compensation coefficient, terminal voltage change mapping value, and voltage ripple increment, generate the voltage overload potential risk coefficient for each charging pile in use through a dynamic weight allocation algorithm combined with a weighted summation algorithm.

[0084] The operation logic of step S22 is as follows:

[0085] Step S221: In the three-dimensional model of the charging station, delineate the three-dimensional coordinate range of the cable channel corresponding to each charging pile in use, and use the regional extreme value search algorithm to automatically identify the highest temperature point in the channel; take the peak point as the center, expand along the three-dimensional coordinate axis to form a detection cube, and calculate the arithmetic mean of all temperature measurement points in the cube as the representative temperature value;

[0086] Step S222: Obtain the original design parameters of the cable, including the standard temperature resistivity corresponding to the conductor material type and the resistance value per unit length of the cable at the standard ambient temperature; based on the actual total length of the cable, use the product of the resistance value per unit length and the total length as the reference resistance value;

[0087] Step S223: Subtract the standard ambient temperature from the average temperature of the current test cube to obtain the actual temperature difference value; multiply the reference resistance value by the temperature difference, and then multiply by the material temperature resistivity to obtain the resistance increment per unit length; superimpose the actual total length of the cable, and the total resistance increment is equal to the product of the resistance increment per unit length and the actual total length of the cable.

[0088] Step S224: Obtain the current charging current value passing through the cable in real time; multiply the total resistance increment by the charging current value to obtain the theoretical increase in voltage drop; divide the voltage drop increase by the rated output voltage of the charging pile, and then multiply by 100% to convert it into a percentage, thereby obtaining the equivalent voltage drop compensation coefficient.

[0089] In its specific implementation, this invention first delineates the three-dimensional coordinate range of each charging pile cable channel in use within the three-dimensional model of the charging station. Temperature monitoring areas are formed by extending detection cubes along the X / Y / Z axes and setting their side lengths. An extreme value search algorithm based on octree spatial partitioning is used to locate the highest temperature point. A cube detection body with a set threshold side length is established centered on this point. The arithmetic mean of all temperature measurement nodes within this cube is calculated to obtain the representative temperature value of the current cable channel. In the stage of obtaining conductor material parameters, the original design parameters from the cable's factory parameter database are extracted: the reference resistance value per unit length of the cable at 20 degrees Celsius is retrieved, and this reference value is multiplied by the actual total length of the cable laying path to obtain the theoretical basic resistance value of the complete cable. When performing dynamic resistance compensation calculations, the average temperature within the current detection cube is first subtracted from the standard ambient temperature of 20 degrees Celsius to obtain the actual temperature rise value; for example, the temperature rise is 45 degrees Celsius when the detection temperature is 65 degrees Celsius. This temperature rise value is then multiplied by the reference resistance value and the material temperature coefficient to obtain the resistance increment per unit length of the cable. Finally, the total resistance change is calculated based on the actual cable length. During the voltage drop dynamic simulation phase, the charging pile's operating current data is collected in real time, including instantaneous current, effective root mean square value, and high-frequency ripple component. The effective current value within the current 0.1-second sliding window is taken, and the product of this current value and the total resistance increment is used as the theoretical voltage drop. Finally, this voltage drop value is divided by the charging pile's rated output voltage and converted into an equivalent voltage drop compensation coefficient.

[0090] The operation logic of step S24 is as follows:

[0091] Based on the thermocouple sensors deployed on the heat dissipation surface of the charging pile, an X-axis coordinate system is established along the fin direction of the heat sink, and a Y-axis coordinate system is established in the vertical direction. After identifying the coordinates of the highest temperature point, a junction temperature inversion model is established through the thermal resistance parameters of the packaging structure. The highest measurement point temperature is multiplied by the material thermal conductivity correction factor, and the product of the thermal resistance parameter and the real-time power is superimposed to obtain the estimated junction temperature data value of the power device.

[0092] Based on this, a dynamic quantitative model of switching loss is constructed. Taking the test value of 25 degrees Celsius environment as the benchmark, a temperature compensation rule is set to increase the loss by 35% for every 100 degrees Celsius increase in junction temperature. The loss value is accumulated and calculated for the parallel operation of multiple devices.

[0093] Subsequently, ripple characteristic reconstruction is performed. The total switching loss is divided by the switching period to obtain the power fluctuation component. This component is then divided by the DC bus voltage value to convert it into an equivalent current ripple, and then multiplied by the total impedance of the charging circuit to obtain the voltage ripple increment.

[0094] Multi-level linkage early warning module: Generates graded early warning signals based on the potential risk coefficient of voltage overload of each charging pile in use.

[0095] Based on the potential risk coefficient of voltage overload at each charging station in use, a graded early warning signal is generated, including:

[0096] The voltage overload risk coefficient of each charging pile in use is compared with the preset safety assessment threshold range [S1,S2].

[0097] If the voltage overload risk factor of a charging pile in use is less than S1, then the warning level of the charging pile in use is marked as Level 1, and no warning is issued for the charging pile in use.

[0098] If the voltage overload risk factor of a charging pile in use is greater than S2, then the warning level of the charging pile in use will be marked as Level 3, the power supply to the charging pile in use will be automatically cut off, and the dry powder fire extinguishing device will be activated.

[0099] If the potential risk factor of a voltage overload of a charging station in use is within the safety assessment threshold range, the warning level of the charging station in use will be marked as Level 2, triggering the audible and visual alarm device and pushing the warning information through the mobile APP.

[0100] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0101] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0102] It should be understood that determining B based on A does not mean determining B solely based on A; it also means determining B based on A and / or other information.

[0103] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0104] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. 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 fire early warning system for new energy vehicle charging stations based on AI algorithms, characterized in that, include: Dynamic parameter acquisition module: Through a multimodal sensor array deployed in the charging station and the battery management system built into each charging pile in use, the module collects the dynamic parameters of each charging pile in use in real time during the charging process; and constructs a three-dimensional dynamic thermodynamic model of each charging pile in use based on the dynamic parameters. Dynamic parameter analysis module: Based on the three-dimensional dynamic thermodynamic model of each charging pile in use, it identifies the potential risk coefficient of voltage overload for each charging pile in use; the specific identification logic is as follows: Step S21: Extract key thermal characteristic values ​​of each charging pile in use in real time from the three-dimensional dynamic thermodynamic model, including the core temperature of the battery compartment, the temperature gradient of the heat dissipation surface of the charging module, and the peak temperature of the cable channel. Step S22: Calculate the conductor resistance increment based on the peak temperature distribution of the cable channel, and convert the resistance increment into an equivalent voltage drop compensation coefficient; Step S23: Analyze the core temperature change curve of the battery compartment and establish a mapping relationship table between electrolyte activity decay and terminal voltage change; Step S24: Based on the temperature gradient data of the charging pile heat dissipation surface, calculate the junction temperature of the power device and correlate it with the voltage ripple increment caused by switching losses; Step S25: Based on the comprehensive voltage drop compensation coefficient, terminal voltage change mapping value, and voltage ripple increment, generate the voltage overload potential risk coefficient for each charging pile in use through a dynamic weight allocation algorithm combined with a weighted summation algorithm. The operation logic of step S22 is as follows: Step S221: In the three-dimensional model of the charging station, delineate the three-dimensional coordinate range of the cable channel corresponding to each charging pile in use, and use the regional extreme value search algorithm to automatically identify the highest temperature point in the channel; take the peak point as the center, expand along the three-dimensional coordinate axis to form a detection cube, and calculate the arithmetic mean of all temperature measurement points in the cube as the representative temperature value; Step S222: Obtain the original design parameters of the cable, including the standard temperature resistivity corresponding to the conductor material type and the resistance per unit length of the cable at standard ambient temperature; Based on the actual total length of the cable, the product of the resistance value per unit length and the total length is used as the reference resistance value; Step S223: Subtract the standard ambient temperature from the current average temperature of the tested cube to obtain the actual temperature difference value; multiply the reference resistance value by the temperature difference, and then multiply by the material temperature resistivity to obtain the resistance increment per unit length. The total resistance increment is equal to the product of the resistance increment per unit length and the actual total length of the cable. Step S224: Obtain the current charging current value through the cable in real time; multiply the total resistance increment by the charging current value to obtain the theoretical increase in voltage drop; divide the voltage drop increase by the rated output voltage of the charging pile, and then multiply by 100% to convert it into a percentage, thereby obtaining the equivalent voltage drop compensation coefficient. The operation logic of step S24 is as follows: Based on the thermocouple sensors deployed on the heat dissipation surface of the charging pile, an X-axis coordinate system is established along the fin direction of the heat sink, and a Y-axis coordinate system is established in the vertical direction. After identifying the coordinates of the highest temperature point, a junction temperature inversion model is established through the thermal resistance parameters of the packaging structure. The highest measurement point temperature is multiplied by the material thermal conductivity correction factor, and the product of the thermal resistance parameter and the real-time power is superimposed to obtain the estimated junction temperature data value of the power device. Based on this, a dynamic quantitative model of switching loss is constructed. Taking the test value of 25 degrees Celsius environment as the benchmark, a temperature compensation rule is set to increase the loss by 35% for every 100 degrees Celsius increase in junction temperature. The loss value is accumulated and calculated for the parallel operation of multiple devices. Then, ripple characteristic reconstruction is performed. The total switching loss is divided by the switching period to obtain the power fluctuation component. This component is divided by the DC bus voltage value to convert it into the equivalent current ripple. Then, it is multiplied by the total impedance of the charging circuit to obtain the voltage ripple increment. Multi-level linkage early warning module: Generates graded early warning signals based on the potential risk coefficient of voltage overload of each charging pile in use.

2. The fire early warning system for new energy vehicle charging stations based on AI algorithms according to claim 1, characterized in that, The dynamic parameters of each charging pile during the charging process include the charging current, voltage, battery temperature, ambient temperature, ambient humidity, and insulation status of the internal cables of the charging pile.

3. The fire early warning system for new energy vehicle charging stations based on AI algorithms according to claim 2, characterized in that, A three-dimensional dynamic thermodynamic model of the charging station is constructed based on the aforementioned dynamic parameters. The specific construction logic is as follows: Step S11: Synchronously collect charging current, voltage and corresponding battery temperature data of each charging pile in use, and establish a mapping relationship between charging equipment and spatial coordinates; Step S12: Calculate the real-time heat generation power of each charging pile in use based on the charging current data, and generate a three-dimensional heat source distribution map of the charging equipment. Step S13: Combine the outer shell temperature and ambient temperature data of each charging pile in use to quantify the thermal resistance coefficient of the air medium and superimpose the thermal resistance coefficient of the air medium onto the three-dimensional heat source distribution map. Step S14: Based on the internal cable insulation state parameters, calculate the additional heat generation rate caused by the impedance change of each charging pile cable in use, and superimpose the additional heat generation rate onto the three-dimensional heat source distribution map; Step S15: Correct the thermal conductivity of the air medium based on the ambient humidity data, and construct a three-dimensional spatial temperature field model through iterative calculation method; Step S16: Vector superposition of the three-dimensional heat source distribution map and the three-dimensional spatial temperature field model to construct a three-dimensional dynamic thermodynamic model of each charging pile in use.

4. The fire early warning system for new energy vehicle charging stations based on AI algorithms according to claim 3, characterized in that, The specific implementation of step S13 includes: Step S131: Scan the surface temperature field of each charging pile in use with an infrared thermal imager, extract the temperature gradient distribution features and mark local hot spots; Step S132: Calculate the thermal resistance coefficients of natural convection and forced heat dissipation by combining data from ambient temperature and humidity sensors; Step S133: Analyze the heat conduction path based on the principle of heat flow continuity and quantify the heat accumulation rate of key components; Step S134: Calculate the thermal resistance coefficient of the air medium based on the product of the thermal resistance coefficient and the thermal accumulation rate.

5. The fire early warning system for new energy vehicle charging stations based on AI algorithms according to claim 1, characterized in that, Based on the potential risk coefficient of voltage overload at each charging station in use, a graded early warning signal is generated, including: The potential risk factor of voltage overload for each charging station in use is compared with the preset safety assessment threshold range. Perform a comparison; If the voltage overload risk factor of a charging pile in use is less than S1, then the warning level of the charging pile in use is marked as Level 1, and no warning is issued for the charging pile in use. If the voltage overload risk factor of a charging pile in use is greater than S2, then the warning level of the charging pile in use will be marked as Level 3, the power supply to the charging pile in use will be automatically cut off, and the dry powder fire extinguishing device will be activated. If the potential risk factor of a voltage overload of a charging station in use is within the safety assessment threshold range, the warning level of the charging station in use will be marked as Level 2, triggering the audible and visual alarm device and pushing the warning information through the mobile APP.

Images