Battery life detection method and system, intelligent terminal and storage medium

Abstract

The invention relates to the technical field of battery detection, in particular to a battery life detection method and system, an intelligent terminal and a storage medium, and the method comprises the steps: collecting operation data in real time through a multi-source sensor array disposed in a charging pile; the multi-source sensor array comprises a multi-point temperature sensor, a current waveform sensor and a displacement sensor; performing dynamic feature extraction on the collected operation data to generate a dynamic feature set; inputting the dynamic feature set into a multi-parameter coupling health analysis model to generate a whole machine health state; the health state of the whole machine is recognized, and when it is recognized that the health state is attenuated, a maintenance instruction is output; the maintenance instruction includes a faulty component identifier and a remaining life predictor. The method has the effects of dynamically and quantitatively sensing the performance state of the key internal components of the charging pile, positioning the fault component and predicting the residual service life of the charging pile.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of battery detection, in particular to a battery life detection method and system, an intelligent terminal and a storage medium. BACKGROUND

[0002] With the popularization of electric vehicles, as a core infrastructure, the operation reliability and service life of charging piles are crucial. In the long-term use process, internal components such as power modules, relays, connectors and heat dissipation systems of charging piles will have performance degradation or failure due to electrical stress, mechanical wear and tear, environmental factors, etc. The electric vehicle charging pile life dynamic detection technology aims to actively identify potential faults and performance degradation trends through real-time monitoring and analysis of the operating state of the charging pile, so as to carry out predictive maintenance, thereby avoiding sudden shutdown, prolonging the service life of the equipment, reducing operation and maintenance costs and ensuring charging safety.

[0003] In related technologies, some technical solutions have attempted to evaluate and predict the health status of charging piles. A Chinese patent application with publication number CN119377868A proposes an analysis method based on historical operating data (mainly temperature and charging quantity data) of charging piles. The method preprocesses the historical data, extracts characteristic parameters such as "temperature deviation" and "charging deviation", and compares them with pre-set static thresholds, finally classifying the state of the charging pile into "healthy", "sub-healthy" or "unhealthy" levels. This method can reflect the overall performance trend of the charging pile to some extent.

[0004] For the related technologies in the above, the scheme based on comparison of historical operating parameters with static thresholds mainly stays at the macro, lagging judgment level of the overall performance of the charging pile, such as "sub-healthy", but cannot deeply and dynamically perceive and quantify the specific performance degradation state of key internal components, such as relay contacts, heat dissipation fan bearings and electrical connection points, such as mechanical action delay, increased contact resistance and decreased heat dissipation efficiency, making it more difficult to accurately locate the specific components that will fail, and also unable to predict their remaining effective service life, resulting in lack of pertinence in operation and maintenance decision-making, and only being able to carry out passive maintenance or overall repair after the device performance has significantly decreased or failure occurs, making it difficult to achieve predictive maintenance and resource optimization. SUMMARY

[0005] In order to dynamically quantify the performance state of key internal components of the charging pile, locate the faulty components and predict the remaining service life of the charging pile, the application provides an electric vehicle charging pile life dynamic detection method, system, intelligent terminal and storage medium.

[0006] In a first aspect, the application provides a battery life detection method, which adopts the following technical solution: A battery life detection method, comprising: Real-time collection of operation data through a multi-source sensor array deployed inside the charging pile; The multi-source sensor array includes multi-point temperature sensors, current waveform sensors, and displacement sensors; Dynamic feature extraction is performed on the collected operation data to generate a dynamic feature set; The dynamic feature set includes a temperature rise feature set, a mechanical smoothness feature set, and an electrical smoothness feature set; The dynamic feature set is input into a multi-parameter coupled health analysis model to generate a whole-machine health state; The whole-machine health state is identified, and when the health state decays, a maintenance instruction is output; The maintenance instruction includes a faulty component identifier and a remaining life prediction value.

[0007] By using the above technical solution, multi-point temperature, current waveform, and displacement sensor arrays are deployed to collect data in real time, and dynamic extraction of temperature rise, mechanical smoothness, and electrical smoothness feature sets is performed to deeply quantify the performance decay state of key components such as relay contact action delay, contact resistance increase, and heat dissipation efficiency decrease, and accurately locate the components that are about to fail. Combined with a multi-parameter coupled health analysis model, not only can the whole-machine health state be identified, but also the remaining useful life of key components can be predicted, realizing the transition from passive maintenance to predictive maintenance. By timely outputting a maintenance instruction containing a faulty component identifier and a remaining life prediction value, targeted maintenance can be implemented, effectively avoiding sudden shutdown, significantly improving operation reliability, and prolonging the life of the whole machine. At the same time, by dynamically quantifying and monitoring key safety indicators, safety risks are reduced, operation and maintenance resources are optimized, overall maintenance costs are reduced, and accurate support is provided for the whole life cycle management of the charging pile.

[0008] Optionally, the method for dynamically extracting features from the collected operation data to generate a temperature rise feature set includes: Distribute multi-point temperature sensors according to the heat dissipation path topology of the charging pile to obtain time-series temperature data of the multi-point temperature sensors on the heat dissipation path; Calculate the gradient temperature between each key node on the heat dissipation path through the time-series temperature data; Based on the gradient temperature and the rate of change of the gradient temperature over time, calculate the temperature difference decay rate of the heat dissipation path; Real-time monitoring of charging pile power data, when detecting that the power change per unit time exceeds the set threshold, identifying it as a load mutation event and recording the mutation time point; Based on the mutation time point, extract temperature response data within a preset time window before and after the load mutation from the time-series temperature data; Calculate the load step temperature rise rate by linearly regressing the temperature-time curve slope of the temperature response data; According to the temperature difference decay rate of the heat dissipation path and the load step temperature rise rate, a temperature rise feature set is generated.

[0009] By adopting the above technical solution, based on the gradient temperature monitoring and temperature difference decay rate calculation of the heat dissipation path topology, the thermal conduction efficiency decline caused by aging or dust accumulation of the heat dissipation medium (such as thermal conductive silicone grease and heat sink) is dynamically captured, the limitations of traditional single-point temperature control are broken through, and the spatial dynamic quantitative evaluation of the performance of the heat dissipation system is realized. Through the load mutation triggering mechanism and the step temperature rise rate analysis, the thermal response delay or capacity decline of the heat dissipation system (such as fan bearing wear and heat dissipation fin blockage) under dynamic load is accurately identified, and the transient overheating risk is warned in advance. Combined with the correlation analysis of the gradient temperature abnormal points and the step temperature rise rate mutation points, the key nodes (such as specific heat dissipation modules or fans) of performance decline in the heat dissipation path are accurately located, which provides a basis for targeted maintenance. By fusing the features of temperature difference decay rate and step temperature rise rate, a multi-dimensional heat dissipation health portrait is constructed, the prediction accuracy of the remaining life of the heat dissipation system is improved, and the damage of the power module caused by heat dissipation failure is avoided.

[0010] Optionally, the method for dynamically extracting features from the collected operation data to generate a mechanical fluency feature set comprises: Through the current waveform sensor embedded in the relay control circuit, the transient current waveform data of the relay attraction / release action is collected in real time; Based on the transient current waveform data, the relay action feature parameters are extracted, including the attraction delay time, the contact bounce duration and the steady-state current fluctuation entropy value; Through the displacement sensor installed in the charging gun locking mechanism, the displacement stroke curve in the plugging process is obtained; Based on the displacement stroke curve, the maximum displacement deviation and the average movement speed of the charging gun locking mechanism are calculated; Based on the relay action feature parameters, the maximum displacement deviation and the average movement speed, a multi-dimensional mechanical state vector is constructed; Based on the multi-dimensional mechanical state vector, a mechanical fluency feature set is generated; Wherein, the multi-dimensional mechanical state vector is represented as: , is the attraction delay time, is the contact bounce duration, is the steady-state current fluctuation entropy value, is the maximum displacement deviation, is the average movement speed.

[0011] By adopting the technical scheme, the transient current waveform of the relay attraction and release action is collected, the attraction delay time, the contact bounce duration and the steady-state current fluctuation entropy value are extracted, and the hidden troubles caused by the contact oxidation or mechanical wear, such as action delay and poor contact, are quantified. The displacement stroke curve of the charging gun locking mechanism is synchronously monitored, the maximum displacement deviation and the average movement speed are calculated, and the mechanical jamming or component deformation is accurately identified. The multi-dimensional mechanical state vector is constructed by fusing the above parameters, and the dynamic quantitative evaluation of the key mechanical performances such as the relay contact aging and the locking mechanism wear is realized. Based on the vector feature change trend, the mechanical action failure risk can be early warned and the specific fault components can be located, accurate basis is provided for targeted maintenance, and the charging interruption or connection safety accident caused by mechanical failure can be effectively avoided.

[0012] Optionally, the method for generating the electrical smoothness feature set by dynamically extracting the collected operation data includes: A detection current is injected at the DC bus terminal of the charging pile, and the current value and terminal voltage drop data are synchronously measured; Based on the current value and the voltage drop data, the dynamic contact resistance is calculated by the four-wire method in combination with the pre-stored inherent resistance value of the cable; Based on the displacement sensor trigger signal of the charging gun locking mechanism, the charging gun plug-in times are accumulated; A plug-in times-contact resistance growth model is established, and the non-linear wear coefficient of the contact resistance with the plug-in times is fitted; An electrical state vector is constructed according to the dynamic contact resistance and the non-linear wear coefficient; The electrical smoothness feature set is generated based on the electrical state vector; The electrical state vector is represented as: , is the dynamic contact resistance, is the non-linear wear coefficient, is the charging gun plug-in times.

[0013] By adopting the above technical scheme, the detection current is injected at the DC bus terminal and the dynamic contact resistance is calculated by the four-wire method, the cable resistance interference is eliminated, and the wear degree of the terminal electrical contact interface is accurately quantified; the displacement sensor synchronously accumulates the charging gun plug-in times, establishes the plug-in times-contact resistance growth model and fits the non-linear wear coefficient, and dynamically captures the acceleration trend of the contact material wear; the electrical state vector of the dynamic contact resistance, the non-linear wear coefficient and the plug-in times is fused, the real-time degradation evaluation of the electrical connection performance is realized; based on the abnormal change of the non-linear wear coefficient, the life inflection point of the contact resistance can be predicted, the connection failure risk can be early warned, the accurate maintenance opportunity judgment for the charging gun terminal cleaning or replacement is provided, and the overheating accident caused by poor contact is effectively prevented.

[0014] Optional methods for constructing a multi-parameter coupled health analysis model include: Based on the coupling relationship between the temperature rise feature set, the mechanical fluency feature set, and the electrical fluency feature set, a coupling failure logic is established; Based on coupled failure logic, a cross-domain failure rule library is established; Based on the cross-domain failure rule library, a two-layer analysis architecture is constructed. The two-layer analysis architecture includes a real-time diagnosis layer and a life prediction layer. The real-time diagnosis layer outputs the failure risk level, and the life prediction layer outputs the remaining life prediction value. Synchronous configuration model collaborative decision-making mechanism, when the fault risk level conflicts with the remaining life prediction value, generates the final health state according to the preset priority strategy; Based on the final health status, output maintenance instructions containing the fault component identifier and the remaining life prediction value; Among them, the coupling failure logic is: When the load step temperature rise rate in the temperature rise feature set is greater than a preset reference ratio, and the fan vibration energy in the mechanical smoothness feature set is greater than a preset vibration threshold, the fault component identifier is defined as cooling fan bearing wear; When the dynamic contact resistance in the electrical fluency feature set is greater than a preset resistance threshold and the contact bounce duration in the mechanical fluency feature set is greater than a preset time threshold, the fault component identifier is defined as relay contact burnout.

[0015] By adopting the above technical solution, a cross-domain failure rule library is formed by establishing coupled failure logic between temperature rise, mechanical and electrical feature sets, breaking through the limitations of single parameter analysis; based on a two-layer architecture, the real-time diagnosis layer outputs the fault risk level and the life prediction layer outputs the remaining life prediction value; a collaborative decision-making mechanism is configured to generate the final health state according to the preset priority strategy when the results conflict; complex fault scenarios are accurately located through coupling rules, such as determining cooling fan bearing wear when the load step temperature rise rate exceeds the benchmark and the fan vibration exceeds the threshold, or identifying relay contact erosion when the contact resistance exceeds the limit and the contact bounce time is abnormal; and finally outputting maintenance instructions containing the fault component identifier and the remaining life prediction value, realizing multi-dimensional fault precise location and dynamic life prediction, significantly improving the accuracy and timeliness of operation and maintenance decisions.

[0016] Optional methods for configuring the model collaborative decision-making mechanism include: Real-time monitoring of failure risk level and remaining life prediction value; When a conflict is detected between the fault risk level and the remaining life prediction value, it is determined whether the fault risk level is ≥ a preset high risk threshold and the remaining life prediction value is > a preset safe life value; If yes, the real-time diagnostic layer is output as the final health status; If no, judge whether the failure risk level is ≤ a preset low risk threshold value and the remaining life prediction value is ≤ a preset risk life value; If yes, extract the fluctuation standard deviation of the load step temperature rise rate , synchronously extract the fluctuation standard deviation of the dynamic contact resistance , and output the weighted calculation value of and as the final health state.

[0017] By adopting the above technical solutions, the three-level conflict decision mechanism ensures the reliability of health state evaluation. When the failure risk level reaches the risk threshold value and the remaining life prediction value is higher than the safe life, the real-time diagnosis result is preferentially adopted. When the risk level is lower than the low risk threshold value and the remaining life prediction value reaches the risk life, the life prediction layer conclusion is adopted. For the intermediate conflict state, the fluctuation standard deviation of the load step temperature rise rate and the dynamic contact resistance is extracted for weighted arbitration. The three-level conflict decision mechanism effectively solves the misjudgment problem of traditional single model, such as early wear of the cooling fan bearing which has a long remaining life but a sudden increase in real-time risk, or slight ablation of the relay contact which has a low risk level but a life close to the end, so as to accurately give an early warning or timely intervention, and finally eliminate the evaluation blind area through dynamic weighted arbitration to improve the decision accuracy in the complex failure scenario.

[0018] Optionally, the method for outputting the weighted calculation value of and as the final health state comprises: obtaining a failure risk level quantization value and a remaining life prediction value normalizing the failure risk level quantization value and outputting a failure risk normalized value normalizing the remaining life prediction value and outputting a remaining life normalized value the fluctuation standard deviation of the load step temperature rise rate is the fluctuation standard deviation of the dynamic contact resistance is comparing the weighted health index H and a preset value range of H, when H≤ , outputting the final health state as an emergency failure, when <H≤ , outputting the final health state as serious degradation, and when H> , outputting the final health state as sub-health. ​​​​​​​​​

[0019] By adopting the technical solution, the normalized fault risk level and the remaining life prediction value are processed to eliminate the dimensional difference. The weight coefficient is distributed based on the dynamic ratio of the load step temperature rise rate fluctuation standard deviation and the reference stability of the heat dissipation system, so that the temperature rise abnormality weight is automatically increased in a harsh environment such as high temperature and high humidity. The weighted health index fuses the two-dimensional information of real-time risk and life degradation to accurately quantify the compound health state. According to the index threshold, the conclusions of emergency failure, serious degradation or sub-health are output, for example, when the heat dissipation fan suddenly loses stability, the emergency failure response is triggered even if the remaining life is long, or the serious degradation is accurately determined when the relay contact is gradually ablated but the risk is controllable. Finally, the adaptive arbitration in the conflict scene is realized, and the robustness of health evaluation and the operability of maintenance instructions in complex working conditions are significantly improved.

[0020] In a second aspect, the application provides a battery life detection system, which adopts the following technical solution: A battery life detection system comprises: An acquisition module is configured to acquire operation data, a dynamic feature set, and a whole-machine health state. A memory is configured to store a program of a control method of the battery life detection method according to any one of claims 1 to 7. A processor, the program in the memory can be loaded and executed by the processor, and the control method of the battery life detection method according to the first aspect is implemented.

[0021] In a third aspect, the application provides an intelligent terminal, which adopts the following technical solution: An intelligent terminal comprises a memory and a processor, and the memory stores a computer program that can be loaded and executed by the processor, and the computer program is the battery life detection method according to the first aspect.

[0022] In a fourth aspect, the application provides a computer storage medium capable of storing a corresponding program, which adopts the following technical solution: A computer readable storage medium stores a computer program that can be loaded and executed by a processor, and the computer program is any one of the battery life detection methods.

[0023] In summary, the application has at least one of the following beneficial technical effects: The temperature rise characteristics, mechanical smoothness and electrical smoothness feature set are dynamically extracted by the multi-source sensor array, which breaks through the limitations of traditional macro health evaluation; the micro degradation parameters such as relay contact bounce delay, heat dissipation path temperature difference decay rate, and contact resistance nonlinear wear are accurately quantified by combining the multi-parameter coupling model; the compound failure components are located by synchronously constructing the cross-domain failure rule library, and the remaining life prediction value is output. The operation and maintenance are upgraded from passive response to active intervention, precise maintenance is implemented before the shutdown caused by hidden failures such as heat dissipation fan bearing micro-wear and contact oxidation, and the charging interruption caused by sudden failure is avoided; Real-time monitoring of dynamic contact resistance and wear coefficient based on electrical flow characteristics, early warning of terminal overheating risk; Capture transient heat dissipation failure through step temperature rise rate analysis triggered by load mutation; Use mechanical state vector to identify displacement deviation of charging gun locking mechanism or abnormal relay action. The triple protection mechanism is deeply integrated to realize early diagnosis of core safety hazards such as poor electrical contact, mechanical jamming, and heat dissipation recession, and to improve the charging safety level; Relying on the double-layer analysis architecture and conflict arbitration mechanism, dynamically coordinate fault risk level and remaining life prediction value, and output graded maintenance instructions; Predict the contact material life inflection point through the number of insertions-wear coefficient model, and predict the fan replacement cycle combined with the heat dissipation health portrait. Convert overall maintenance into component-level precise maintenance, such as replacing only the worn heat dissipation module instead of the entire machine, reducing invalid operation and maintenance; Optimize spare parts inventory based on life prediction value, reduce idle cost, achieve the dual goals of reducing operation and maintenance cost and improving equipment availability, and ensure efficient operation of the charging network. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 is a flowchart of a battery life detection method according to an embodiment of the present application.

[0025] Figure 2 is a flowchart of a method for dynamically extracting features from collected operation data to generate a temperature rise feature set according to an embodiment of the present application.

[0026] Figure 3 is a flowchart of a method for dynamically extracting features from collected operation data to generate a mechanical flow feature set according to an embodiment of the present application.

[0027] Figure 4 is a flowchart of a method for dynamically extracting features from collected operation data to generate an electrical flow feature set according to an embodiment of the present application.

[0028] Figure 5 is a flowchart of a method for constructing a multi-parameter coupled health analysis model according to an embodiment of the present application.

[0029] Figure 6 is a flowchart of a method for configuring a model collaborative decision mechanism according to an embodiment of the present application.

[0030] Figure 7 is a flowchart of a method for outputting the weighted calculation value of and as the final health state according to an embodiment of the present application.

[0031] Figure 8 is a module diagram of a battery life detection method according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] The present application is further described in detail below with reference to the accompanying drawings.

[0033] This specific embodiment is merely an explanation of the present application and is not a limitation of the present application. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed, but as long as they are within the scope of the claims of the present application, they are protected by the patent law.

[0034] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the following will be combined with the appended examples of the embodiments of the present application. Figures 1-8 , clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of this application.

[0035] The present application embodiment discloses a battery life detection method. Figure 1 , the battery life detection methods include: Step 100: Collect operating data in real time through a multi-source sensor array deployed inside the charging pile.

[0036] A multi-source sensor array refers to a combination of sensor devices deployed in key components of a charging pile, used to synchronously monitor power module temperature, relay operating current, and charging gun displacement timing.

[0037] The acquisition method is to trigger the synchronous sampling instruction based on the central control unit and transmit the original data stream through the anti-interference digital bus at a frequency of 200Hz.

[0038] For example, in a 30kW DC fast charging pile, three PT1000 temperature sensors are arranged on the power module IGBT heat dissipation substrate, the main relay K1 control circuit integrates a 100kHz sampling current Hall sensor, and the charging gun locking mechanism is installed with a 0.1mm resolution laser displacement sensor.

[0039] Step 101: A multi-source sensor array includes multiple temperature sensors, current waveform sensors, and displacement sensors.

[0040] Multi-point temperature sensors target and monitor key nodes in the heat dissipation path, covering areas on the power module surface with a range greater than 150 degrees Celsius and areas on the heat sink inlet and outlet with a range of -40 to 120 degrees Celsius.

[0041] The current waveform sensor captures the transient current rising edge waveform of the relay's energizing and releasing state, which is less than 10 microseconds, and the peak current fluctuation of the cooling fan start and stop, with an accuracy of plus or minus 0.5 amps. The displacement sensor records the charging cable's full insertion and removal stroke, from 0 to 15 mm.

[0042] For example, in a 30kW fast charging pile, temperature sensor T1 monitors the chip, T2 monitors the heat dissipation substrate, and T3 monitors the heat dissipation fin, forming an axial heat dissipation path. The current sensor captures the typical inrush current waveform amplitude of K1 relay from 0 to 5 amperes for 20 milliseconds, and the displacement sensor outputs the timing data of the insertion stage from 0 to 15 millimeters lasting 1.2 seconds.

[0043] Step 102, dynamic feature extraction is performed on the collected operation data to generate a dynamic feature set.

[0044] The dynamic feature set is a set of indicators for quantifying the performance degradation of the component.

[0045] The extraction method includes calculating the heat dissipation path gradient temperature ΔT equal to the difference between T2 and T3, fitting the temperature rise rate dT / dt from power 30 kilowatts to 60 kilowatts through step load experiment, analyzing the inrush delay time of relay current waveform, that is, the time length from zero to 90% of steady-state current, FFT analyzing the vibration spectrum energy of the fan in the frequency band of 50 to 200 Hz, and using the four-wire method to inject 10 amperes of detection current at the bus terminal and simultaneously measure the voltage drop to calculate the dynamic contact resistance.

[0046] For example, the extraction results of a 30kW fast charging pile: the gradient temperature ΔT is 7 degrees Celsius, the reference value is less than or equal to 3 degrees Celsius, the temperature rise rate dT / dt is 1.5 degrees Celsius per second, the reference value is 0.8 degrees Celsius per second, the K1 relay inrush delay time is 15 milliseconds (initial 5 milliseconds), the fan vibration energy is 35 decibels, and the contact resistance of the charging gun after 2000 times of insertion and extraction is 1.7 milliohms (initial 0.5 milliohms).

[0047] Step 103, the dynamic feature set includes a temperature rise feature set, a mechanical smoothness feature set, and an electrical smoothness feature set.

[0048] The temperature rise feature set includes the heat dissipation path temperature gradient ΔT, the step temperature rise rate dT / dt, and the thermal time constant τ. The mechanical smoothness feature set includes the relay inrush release time, the driving motor action current peak value, and the vibration spectrum energy of the moving parts. The electrical smoothness feature set includes the contact resistance estimation value of the large current loop, the cumulative value N of the connector insertion and extraction times, and the resistance wear slope k.

[0049] For example, the dynamic feature set of a 30kW fast charging pile includes a temperature rise feature set with a gradient temperature of 7 degrees Celsius, a temperature rise rate of 1.2 degrees Celsius per second, and a thermal time constant of 28 seconds. The mechanical smoothness feature set includes an inrush delay time of 15 milliseconds, a current peak value of 4.8 amperes, and a vibration energy of 35 decibels. The electrical smoothness feature set includes a contact resistance of 1.8 milliohms, an insertion and extraction time of 2000 times, and a wear slope of 0.03 milliohms per time.

[0050] Step 104, input the dynamic feature set into the multi-parameter coupling health analysis model to generate the overall machine health status.

[0051] The multi-parameter coupling health analysis model is a decision engine for establishing temperature rise mechanical and electrical feature correlation rules. The core coupling relationships include: dynamic matching degree of fan speed and dT / dt, failure correlation between relay action time and contact resistance, and correlation between plug-in times N and wear slope k.

[0052] The generation method includes: real-time diagnosis layer matches cross-domain rule library, such as fan speed 3000 revolutions per minute, dT per dt greater than 1.0 degrees Celsius per second, triggering fan efficiency degradation fault code, life prediction layer calculates residual life based on degradation model, such as relay contact life equal to 100000 times divided by 1 plus 0.2 times attraction delay time minus 5, conflict arbitration executes three-level decision mechanism.

[0053] For example, after inputting the 30kW fast charging pile feature set, the diagnosis layer outputs the fan efficiency degradation fault code and the relay ablation fault code, and the prediction layer outputs the charging gun terminal residual life of 142 times.

[0054] Step 105, identify the overall machine health status, and output maintenance instructions when the health status decays.

[0055] Health status decay refers to the presence of advanced fault codes such as relay ablation fault codes, or the presence of two or more intermediate fault codes such as fan efficiency degradation fault codes, or residual life less than a safety threshold such as terminal life less than 200 times.

[0056] The identification method is to scan the diagnosis layer output code and the prediction layer life value in real time, and match the decay judgment rule table.

[0057] For example, the 30kW fast charging pile identifies the advanced fault code relay ablation and the intermediate fault code fan efficiency degradation, and the terminal residual life is 142 times, which is less than the threshold of 200 times, and the health status is determined to be degraded and the instruction output is triggered.

[0058] Step 106, the maintenance instruction includes the fault component identifier and the residual life prediction value.

[0059] The fault component identifier is encoded according to the rule of position, component type number, and the residual life prediction value is in units of hours or cycle times.

[0060] The output method is to encode the diagnosis result as a preset identifier, and the life value is rounded off.

[0061] For example, the 30kW fast charging pile outputs the instruction: the heat dissipation fan module A1 needs to be replaced within 720 hours, the power relay K1 needs to be repaired immediately, and the charging gun terminal has a residual plug-in life of 142 times.

[0062] Refer to Figure 2 The method for dynamic feature extraction on the collected operation data to generate the temperature rise feature set comprises: Step 200, distributing multi-point temperature sensors according to the heat dissipation path topology of the charging pile to obtain time-series temperature data of the multi-point temperature sensors on the heat dissipation path.

[0063] The heat dissipation path topology refers to the heat conduction spatial structure from the power module to the heat sink, including three key nodes of the chip surface, the heat-conducting substrate and the heat dissipation fin.

[0064] The acquisition method is to record the temperature values of each node based on a preset sampling frequency of 200 Hz to form a time-temperature sequence.

[0065] For example, in a 30kW fast charging pile, sensors are arranged along the heat dissipation path: T1 chip center point, T2 heat-conducting substrate middle section, T3 heat dissipation fin inlet, and time-series temperature data is collected at 0.5 second intervals.

[0066] Step 201, calculating the gradient temperature between each key node on the heat dissipation path through the time-series temperature data.

[0067] The gradient temperature refers to the temperature change per unit distance between adjacent nodes, reflecting the heat conduction efficiency.

[0068] The calculation method is to take the average value of T2 minus T1 in the continuous 10 sampling periods as the substrate gradient, and the average value of T3 minus T2 as the fin gradient.

[0069] For example, according to the data of T1=85℃, T2=78℃, T3=70℃, the substrate gradient temperature is calculated as 7 degrees Celsius per centimeter, and the fin gradient temperature is calculated as 7 degrees Celsius per centimeter.

[0070] Step 202, calculating the temperature difference decay rate of the heat dissipation path based on the gradient temperature and the change rate data of the gradient temperature over time.

[0071] The temperature difference decay rate refers to the decline rate of the gradient temperature per unit time, which represents the degree of decline of the heat dissipation performance.

[0072] The calculation method is to perform a first-order derivative operation on the gradient temperature sequence and take the absolute value of the decay slope.

[0073] For example, it is monitored that the substrate gradient temperature decreases from 7℃ / cm to 5℃ / cm in 60 seconds, and the temperature difference decay rate is calculated as 0.033 degrees Celsius per centimeter per second.

[0074] Step 203, real-time monitoring of the charging pile power data, when detecting that the power change per unit time exceeds the set threshold, identifying it as a load mutation event and recording the mutation time point.

[0075] Load jump event refers to a power change rate greater than 5 kW / s.

[0076] The identification method is to calculate the instantaneous power derivative using a sliding time window, and trigger the event marking when the threshold is exceeded.

[0077] For example, in a 30kW fast charging pile, the power jumps from 30kW to 60kW in 2 seconds, with a change rate of 15kW / s, and the jump time point T=2030-02-3020:30:00.000 is recorded.

[0078] Step 204, based on the mutation time point, extract the temperature response data within the preset time window before and after the load mutation from the time series temperature data.

[0079] The preset time window in this embodiment is defined as 10 seconds before the mutation to 60 seconds after the mutation.

[0080] The extraction method is to take a temperature sequence segment centered on the mutation time point.

[0081] For example, extract 70 seconds of data before and after T1 sensor at 20:30:00.000 to form a temperature response curve.

[0082] Step 205, calculate the load step temperature rise rate by linear regression fitting the temperature-time curve slope of the temperature response data.

[0083] The load step temperature rise rate refers to the temperature increment per unit time during the temperature rise phase after the mutation.

[0084] The calculation method is to perform least squares linear fitting on the temperature data from 0 to 30 seconds after the mutation, and take the absolute value of the slope.

[0085] For example, fitting the T1 temperature curve gives a slope of 1.5°C / s, i.e. the load step temperature rise rate.

[0086] Step 206, generate a temperature rise feature set according to the temperature difference decay rate of the heat dissipation path and the load step temperature rise rate.

[0087] The temperature rise feature set is a set of quantitative parameters integrating steady-state and transient thermal performance.

[0088] The generation method is to combine the gradient temperature, temperature difference decay rate, and step temperature rise rate into a feature vector.

[0089] For example, the 30kW fast charging pile output temperature rise feature set is: [ Gradient temperature: 7°C / cm for the substrate, 8°C / cm for the fins, Temperature difference decay rate: 0.033°C / cm / s, Step temperature rise rate: 1.5°C / s ].

[0090] Referring Figure 3 The method for extracting dynamic features from collected operation data to generate a mechanical fluency feature set includes: Step 300, through the current waveform sensor embedded in the relay control circuit, real-time acquisition of transient current waveform data of relay attraction / release action.

[0091] Transient current waveform data refers to the millisecond-level current change curve generated during the relay coil power-on / power-off process.

[0092] The acquisition method is to use a 100 kHz sampling rate current Hall sensor to synchronously trigger recording, covering the complete process from attraction start to steady state establishment.

[0093] For example, in the 30kW fast charging pile main relay K1 attraction action, the current waveform data from 0 ampere to 5 ampere for 20 milliseconds is collected.

[0094] Step 301, based on the transient current waveform data, extract the relay action feature parameters, including attraction delay time, contact bounce duration and steady state current fluctuation entropy value.

[0095] The attraction delay time refers to the time from the application of the driving voltage to the current reaching 90% of the steady state value. The contact bounce duration refers to the time from the current first reaching the steady state value to the fluctuation stabilizing within ±2%. The steady state current fluctuation entropy value refers to the information entropy of the steady state current waveform, reflecting the contact stability.

[0096] The extraction method is: delay time, identify the time point when the current rises from 0 to 4.5 ampere; bounce time, calculate the time when the current first enters and stays in the 4.9-5.1 ampere interval; entropy value, calculate the Shannon entropy of the steady state current sampling sequence.

[0097] For example, the 30kW fast charging pile K1 relay has an attraction delay time of 15 milliseconds, an initial 5 milliseconds, a bounce duration of 8 milliseconds, an initial 2 milliseconds, and a steady state entropy value of 1.2 bits, an initial 0.5 bits.

[0098] Step 302, through the displacement sensor installed in the charging gun locking mechanism, obtain the displacement stroke curve in the plugging process.

[0099] The displacement stroke curve refers to the trajectory of the displacement amount changing with time during the charging gun insertion / plugging process.

[0100] The acquisition method is to record the displacement-time sequence of 0-15mm stroke at a sampling rate of 500Hz.

[0101] For example, in the 2000th insertion and extraction of a 30kW fast charging pile, the displacement from 0mm to 15mm in the insertion stage takes 1.5 seconds, forming a stroke curve S-t.

[0102] In step 303, the maximum displacement deviation and average movement speed of the charging gun locking mechanism are calculated based on the displacement stroke curve.

[0103] The maximum displacement deviation refers to the maximum deviation of the actual stroke from the standard trajectory. The average movement speed refers to the ratio of the stroke distance to the time consumed.

[0104] The calculation method is: displacement deviation, taking the absolute value of the difference between the actual curve and the standard curve; average speed, the stroke endpoint displacement divided by the total action time.

[0105] For example, according to the curve S-t, the maximum displacement deviation is 0.8mm, which is less than the standard value of 0.3mm, and the average movement speed is 10mm / s, which is less than the standard value of 12mm / s.

[0106] In step 304, a multi-dimensional mechanical state vector is constructed based on the relay action characteristic parameters, the maximum displacement deviation and the average movement speed.

[0107] The multi-dimensional mechanical state vector refers to the mathematical representation of the performance parameters of the relay and mechanical mechanism.

[0108] The construction method is to arrange the five parameters in a fixed order (see step 306).

[0109] In step 305, a mechanical fluency feature set is generated based on the multi-dimensional mechanical state vector.

[0110] The mechanical fluency feature set is a feature set that quantifies the performance degradation of mechanical action.

[0111] The generation method is to directly output the vector parameters as feature set elements.

[0112] For example, the feature set of a 30kW fast charging pile is: { Attractive delay time: 15ms, Contact bounce time: 8ms, Current entropy value: 1.2bit, Maximum displacement deviation: 0.8mm, Average speed: 10mm / s } In step 306, the multi-dimensional mechanical state vector is represented as: , is the attractive delay time, is the contact bounce duration, is the steady-state current fluctuation entropy value, Max displacement deviation, Average motion speed.

[0113] Vector defines parameter order and physical unit, ensuring feature set standardization.

[0114] Unit: millisecond; Unit: millisecond; Unit: bit; Unit: millimeter; Unit: millimeter per second.

[0115] For example, 30kW fast charging pile vector: [15ms, 8ms, 1.2bit, 0.8mm, 10mm / s] Referring to Figure 4 The method for extracting dynamic features from collected operation data to generate an electrical smoothness feature set includes: Step 400, inject a detection current at the DC bus terminal of the charging pile, and simultaneously measure the current value and terminal voltage drop data.

[0116] The detection current refers to a constant 10 ampere test current dedicated to contact resistance measurement.

[0117] The injection method is to apply through an independent current source during the charging interval to avoid interference with the working current.

[0118] For example, 30kW fast charging pile injects a 10 ampere DC detection current into the charging gun terminal for 200 milliseconds after 5 seconds of charging is completed.

[0119] Step 401, based on the current value and voltage drop data, combined with the pre-stored inherent resistance value of the cable, calculate the dynamic contact resistance by four-wire method.

[0120] The dynamic contact resistance refers to the pure contact interface resistance after eliminating the influence of the cable resistance.

[0121] The calculation method (four-wire method formula) is: contact resistance = voltage drop / current value minus cable inherent resistance.

[0122] For example, the measured current is 10.005 amperes, the voltage drop is 18.02 millivolts, the cable inherent resistance is 0.1 milliohm, and the calculated dynamic contact resistance is 1.8 milliohm minus 0.1 milliohm, which equals 1.7 milliohm.

[0123] Step 402, based on the displacement sensor trigger signal of the charging gun locking mechanism, accumulate the number of charging gun insertion and extraction times.

[0124] The insertion and extraction times accumulation refers to automatically adding 1 when the displacement sensor detects a complete insertion and release cycle.

[0125] The cumulative mode is to capture the complete waveform of displacement from 0 to 15 mm and then return to zero by the edge detection circuit.

[0126] For example, the 30kW fast charging pile displacement sensor records the 2000th complete plug-in and plug-out action.

[0127] Step 403, establish the number of plug-in and plug-out times-contact resistance growth model, fit the non-linear wear coefficient of contact resistance with the number of plug-in and plug-out times.

[0128] The non-linear wear coefficient refers to the resistance growth caused by unit plug-in and plug-out times.

[0129] The fitting mode is to take the last 100 measurement data and calculate the slope by piecewise linear regression.

[0130] For example, according to the data fitting curve of 1950-2000 times, the wear coefficient k=0.03 mΩ / time is obtained.

[0131] Step 404, construct the electrical state vector according to the dynamic contact resistance and the non-linear wear coefficient.

[0132] The electrical state vector refers to a mathematical vector representing the degradation of electrical contact performance.

[0133] The construction mode is to combine parameters in a fixed order (see step 406).

[0134] Step 405, generate the electrical fluency feature set based on the electrical state vector.

[0135] The electrical fluency feature set is a feature set that quantifies the reliability of electrical connection.

[0136] The generation mode is to directly output the vector parameters as feature set elements.

[0137] For example, the 30kW fast charging pile feature set: { Dynamic contact resistance: 1.7 mΩ, Wear coefficient: 0.03 mΩ / time, Plug-in and plug-out times: 2000 } Step 406, wherein the electrical state vector is represented as: , is the dynamic contact resistance, is the non-linear wear coefficient, is the number of plug-in and plug-out times of the charging gun.

[0138] The vector definition clearly defines the parameter order and physical units, ensuring the standardization of the feature set. The unit is milliohm, The unit is milliohm per time, For the dimensionless cumulative value.

[0139] For example, 30kW fast charging pile vector [1.7mΩ, 0.03mΩ / once, 2000].

[0140] Referring to Figure 5 The method for constructing the multi-parameter coupling health analysis model comprises the following steps: Step 500, based on the coupling relationship between the temperature rise feature set, the mechanical fluency feature set and the electrical fluency feature set, a coupling failure logic is established.

[0141] The coupling failure logic refers to the decision rule of parameter association triggering failure across feature domains.

[0142] The establishment mode is to define a composite condition. Cooling fan bearing wear: load step temperature rise rate greater than reference value 30% and fan vibration energy greater than threshold value 30 decibels; relay contact ablation: dynamic contact resistance greater than threshold value 1.5 millimeter, contact bounce duration greater than threshold value 10 milliseconds.

[0143] Step 501, based on the coupling failure logic, a cross-domain failure rule library is established.

[0144] The cross-domain failure rule library is a database for storing coupling logic, and each rule contains a fault code, a trigger condition and a component identifier.

[0145] The establishment mode is to encode the logic entries into a queryable structure.

[0146] For example, the rule entry is: Rule ID: F001 Fault code: FAN_BRG_WEAR Trigger condition: dT / dt>1.3℃ / s AND Vib>30dB Component identifier: COOLING_FAN_BEARING_A1 Step 502, based on the cross-domain failure rule library, a double-layer analysis architecture is constructed, which comprises a real-time diagnosis layer and a life prediction layer, and the real-time diagnosis layer outputs a fault risk level and the life prediction layer outputs a remaining life prediction value.

[0147] The real-time diagnosis layer matches the rule library to output a fault risk level (low level 0-1, medium level 2-3, high level 4-5); the life prediction layer calculates the remaining life based on a degradation model.

[0148] The construction mode is: diagnosis layer: real-time scanning feature set matches rule library, outputs risk level. Prediction layer: according to feature value, call life equation, such as terminal life=(2.0-R_contact) / k.

[0149] For example, the 30kW fast charging pile diagnostic layer detects dT / dt = 1.5℃ / s > 1.3℃ / s and Vib = 35dB > 30dB, and outputs COOLING_FAN_BEARING_A1; the prediction layer calculates the terminal life 142 times.

[0150] Step 503, the synchronous configuration model cooperates with the decision mechanism, and when the fault risk level and the residual life prediction value conflict, a final health state is generated according to a preset priority strategy.

[0151] The cooperative decision mechanism arbitrates according to a three-level strategy: the risk level is greater than or equal to 4 and the residual life is greater than 200 times, the diagnostic layer result is adopted; the risk level is less than or equal to 2 and the residual life is less than or equal to 50 times, the prediction layer result is adopted; for the rest of the conflict, the weighted value of the standard deviation of the temperature rise fluctuation and the standard deviation of the contact resistance fluctuation is taken.

[0152] For example, the 30kW fast charging pile conflict scenario: the risk level is 4 (high) but the terminal life is 142 times > 50 times, the diagnostic layer result is adopted and high-level failure is output.

[0153] Step 504, based on the final health state, a maintenance instruction containing a fault component identifier and a residual life prediction value is output.

[0154] The output mode is to encode according to a preset template: [component identifier]; [action instruction]: [residual life value].

[0155] For example, the 30kW fast charging pile output: COOLING_FAN_BEARING_A1; REPLACE: 720h RELAY_MAIN_K1; REPAIR: IMMEDIATE.

[0156] Step 505, wherein the coupling failure logic is: When the load step temperature rise rate in the temperature rise feature set is greater than a preset reference ratio, and the fan vibration energy in the mechanical smoothness feature set is greater than a preset vibration threshold, the fault component identifier is defined as a cooling fan bearing wear; When the dynamic contact resistance in the electrical smoothness feature set is greater than a preset resistance threshold, and the contact bounce duration in the mechanical smoothness feature set is greater than a preset time threshold, the fault component identifier is defined as a relay contact ablation.

[0157] Cooling fan bearing wear: the load step temperature rise rate is greater than the reference ratio by 30% and the fan vibration energy is greater than the preset vibration threshold by 30 decibels; Relay contact ablation: the dynamic contact resistance is greater than the preset resistance threshold by 1.5 milliohms and the contact bounce duration is greater than the preset time threshold by 10 milliseconds; For example, the dT / dt reference value in a 30kW fast charging pile is 1.0℃ / s, and the actual value is 1.5℃ / s, which is 50% higher than the reference value. =1.7mΩ>1.5mΩ, =8ms<10ms, only triggering fan bearing wear.

[0158] Referring to Figure 6 The method for configuring the model collaborative decision mechanism comprises the following steps: Step 600, real-time monitoring of the fault risk level and the remaining life prediction value.

[0159] The fault risk level refers to the 0-5 risk quantization value output by the real-time diagnosis layer, and the highest risk is level 5. The remaining life prediction value refers to the component operation time or number of times output by the life prediction layer.

[0160] The monitoring method is to synchronously read the diagnosis layer and prediction layer output results through the data bus.

[0161] For example, the 30kW fast charging pile monitors that the cooling fan bearing risk level is 4 and the terminal remaining life is 142 times.

[0162] Step 601, when the fault risk level and the remaining life prediction value are detected to be in conflict, determining whether the fault risk level is ≥ a preset high risk threshold and the remaining life prediction value is > a preset safe life value.

[0163] The conflict refers to that the remaining life of a high-risk component is long or the life of a low-risk component is about to be exhausted. The high-risk threshold is defined as level 4, and the safe life value is defined as 200 times for the terminal and 1000 hours for the fan.

[0164] The determination method is to call the priority strategy table to compare the threshold.

[0165] For example, it is detected that the cooling fan risk level is 4 ≥ the threshold 4, and the remaining life is 720 hours > the safe value 200 hours, which satisfies the first level condition.

[0166] Step 602, if yes, the real-time diagnosis layer output is adopted as the final health state.

[0167] The final health state adopts the diagnosis layer conclusion.

[0168] The execution method is to override the prediction layer result and directly output the diagnosis layer fault code and risk level.

[0169] For example, the 30kW fast charging pile satisfies the first level condition, and the final state adopts: the cooling fan bearing is worn out, and the risk level is 4.

[0170] Step 603, if no, determining whether the fault risk level is ≤ a preset low risk threshold and the remaining life prediction value is ≤ a preset risk life value.

[0171] Low risk threshold is defined as level 2, and risk life value is defined as 50 times of terminal 50 times, and fan 200 hours.

[0172] The judgment method is to match the secondary conditions.

[0173] For example, assuming that the terminal risk level is 1≤2 level, and the remaining life is 40 times≤50 times, the second level condition is met.

[0174] Step 604, if yes, output the life prediction layer as the final health state.

[0175] The execution method is to adopt the life conclusion of the prediction layer and ignore the low risk result of the diagnosis layer.

[0176] For example, in the terminal scenario, output “the remaining life of the charging gun terminal is 40 times, which needs to be replaced immediately”.

[0177] Step 605, if no, extract the fluctuation standard deviation of the load step temperature rise rate , and synchronously extract the fluctuation standard deviation of the dynamic contact resistance , and output the weighted calculation value of and as the final health state.

[0178] The fluctuation standard deviation refers to the dispersion degree of the characteristic parameter in the last 24 hours.

[0179] The calculation method is : take 20 step temperature rise rate samples to calculate the standard deviation; : take 50 dynamic contact resistance samples to calculate the standard deviation; the weighted value formula is: health index=0.6× +0.4× .

[0180] For example, in the intermediate conflict scenario of the 30kW fast charging pile: =0.25℃ / s (20 times of dT / dt fluctuation); =0.15mΩ (50 times of contact resistance fluctuation); health index=0.6×0.25+0.4×0.15=0.21 , according to the threshold table (0-0.2 emergency failure, 0.2-0.5 serious recession, >0.5 sub-health) to output “serious recession” state.

[0181] Referring to Figure 7 , the method of outputting the weighted calculation value of and as the final health state includes: Step 700, obtaining the failure risk level quantization value and the remaining life prediction value , quantitative value of fault risk level Perform normalization and output the normalized value of fault risk , the remaining life prediction value Perform normalization and output the normalized value of remaining life .

[0182] Failure risk normalized value Refers to the mapping of 0-5 risk levels to a standardized value in the range of 0-1. Refers to converting the life value into a decay degree in the range of 0-1 according to the component type.

[0183] The normalization method is: = / 5(S_diag Range 0-5); =1-min( / ,1)( is the maximum design life of the component).

[0184] For example, a 30kW fast charging pile cooling fan: =4→ =4 / 5=0.8; =720 hours, =8000 hours → =1-min(720 / 8000,1)=0.91.

[0185] Step 701: assign dynamic weight coefficients based on the environmental state. , , is the standard deviation of the load step temperature rise rate fluctuation, It is the benchmark stability value of the cooling system.

[0186] The dynamic weight coefficient refers to the weight of risk and life adjusted according to the severity of the environment, and the weight of temperature rise fluctuation is automatically increased in high temperature environment ( ), increase the resistance fluctuation weight in high humidity environment ( ).

[0187] The allocation method is: ( Take 0.2℃ / s); ( Take 0.1mΩ).

[0188] For example, in high temperature environments: =0.25℃ / s, =0.25 / (0.25+0.2)=0.556; =0.15mΩ, = 0.15 / (0.15 + 0.1) = 0.6 Step 702, the weighted health index is calculated as .

[0189] The weighted health index H is a comprehensive index that integrates real-time risk and life decline; The calculation method is:

[0190] For example, a 30kW fast charging pile: H = (0.8 x 0.556) + (0.91 x 0.6) = 0.4448 + 0.546 = 0.9908 Step 703, compare the weighted health index H and the preset value range of H, when H≤ , output the final health state as emergency failure, when ≤ H ≤ , output the final health state as serious decline, when H > 0.5, output the final health state as sub-health.

[0191] The health state classification is determined according to the index threshold, emergency failure: H≤0.2; serious decline: 0.2

[0192] The output method is to generate a text state code matching the threshold interval.

[0193] For example, the 30kW fast charging pile H = 0.9908 > 0.5, the output final health state is sub-health.

[0194] Based on the same inventive concept, the embodiments of the present application provide a battery life detection system.

[0195] Referring to Figure 8 , a battery life detection system includes: An acquisition module for acquiring operating data, a dynamic feature set, and a whole machine health state; A memory for storing the program of the control method of the battery life detection method; The program in the memory can be loaded and executed by the processor, and the control method of the battery life detection method is implemented.

[0196] ​Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0197] The embodiment of the present application provides a computer readable storage medium, which stores a computer program capable of being loaded by a processor and executing a battery life detection method.

[0198] The computer storage medium includes, for example, a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media capable of storing program codes.

[0199] Based on the same inventive concept, the embodiment of the present application provides an intelligent terminal, which comprises a memory and a processor, and the memory stores a computer program capable of being loaded by the processor and executing a battery life detection method.

[0200] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0201] The above are preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in the specification (including the abstract and drawings) can be replaced by other equivalent or similar features unless specifically described. That is, each feature is only an example of a series of equivalent or similar features unless specifically described.

Claims

1. A battery life detection method, characterized in that: include: Real-time operation data is collected through a multi-source sensor array deployed inside the charging pile; The multi-source sensor array includes multi-point temperature sensors, current waveform sensors and displacement sensors; Perform dynamic feature extraction on the collected operating data to generate a dynamic feature set; The dynamic feature set includes a temperature rise feature set, a mechanical fluency feature set, and an electrical fluency feature set; The dynamic feature set is input into the multi-parameter coupled health analysis model to generate the health status of the entire machine; Identify the health status of the entire machine and output maintenance instructions when the health status is identified to be deteriorating; The maintenance order includes the faulty component identifier and the remaining life prediction.

2. A battery life detection method according to claim 1, characterized in that: The method of extracting dynamic features from the collected operating data and generating a temperature rise feature set includes: Distribute multiple temperature sensors according to the topology of the heat dissipation path of the charging pile to obtain time-series temperature data of the multiple temperature sensors on the heat dissipation path; Calculate the gradient temperature between key nodes on the heat dissipation path through time series temperature data; Based on the gradient temperature and the rate of change of the gradient temperature over time, the temperature difference attenuation rate of the heat dissipation path is calculated; Real-time monitoring of charging pile power data. When the power change per unit time exceeds the set threshold, it is identified as a load mutation event and the mutation time point is recorded; Based on the mutation time point, the temperature response data within the preset time window before and after the load mutation is extracted from the time series temperature data; The load step temperature rise rate is calculated by linear regression fitting the slope of the temperature-time curve of the temperature response data; A temperature rise feature set is generated based on the temperature difference attenuation rate of the heat dissipation path and the load step temperature rise rate.

3. The battery life detection method according to claim 2, characterized in that: The method for dynamically extracting features from the collected operating data and generating a mechanical smoothness feature set includes: Through the current waveform sensor embedded in the relay control circuit, the transient current waveform data of the relay's closing / release action is collected in real time; Based on the transient current waveform data, the relay action characteristic parameters are extracted, including the pick-up delay time, contact bounce duration and steady-state current fluctuation entropy value; The displacement curve during the plugging and unplugging process is obtained by installing a displacement sensor on the locking mechanism of the charging gun; Based on the displacement curve, calculate the maximum displacement deviation and average movement speed of the charging gun locking mechanism; Construct a multi-dimensional mechanical state vector based on the relay action characteristic parameters, maximum displacement deviation and average motion speed; Generate a mechanical fluency feature set based on the multi-dimensional mechanical state vector; Among them, the multidimensional mechanical state vector is expressed as: , is the pull-in delay time, is the contact bounce duration, is the entropy value of steady-state current fluctuation, is the maximum displacement deviation, is the average movement speed.

4. The battery life detection method according to claim 3, characterized in that: The method for dynamically extracting features from the collected operating data and generating an electrical fluency feature set includes: Inject a test current at the DC bus terminal of the charging pile and simultaneously measure the current value and terminal voltage drop data; Based on the current value and voltage drop data, combined with the pre-stored cable inherent resistance value, the dynamic contact resistance is calculated using the four-wire method; Based on the displacement sensor trigger signal of the charging gun locking mechanism, the number of charging gun plug-in and unplugging times is accumulated; A plugging and unplugging times-contact resistance growth model was established to fit the nonlinear wear coefficient of contact resistance with plugging and unplugging times; Construct the electrical state vector based on the dynamic contact resistance and nonlinear wear coefficient; Based on the electrical state vector, an electrical fluency feature set is generated; Among them, the electrical state vector is expressed as: , is the dynamic contact resistance, is the nonlinear wear coefficient, The number of times the charging gun is plugged in and out.

5. The battery life detection method according to claim 4, characterized in that: The construction method of the multi-parameter coupled health analysis model includes: Based on the coupling relationship between the temperature rise feature set, the mechanical fluency feature set, and the electrical fluency feature set, a coupling failure logic is established; Based on coupled failure logic, a cross-domain failure rule library is established; Based on the cross-domain failure rule library, a two-layer analysis architecture is constructed. The two-layer analysis architecture includes a real-time diagnosis layer and a life prediction layer. The real-time diagnosis layer outputs the failure risk level, and the life prediction layer outputs the remaining life prediction value. Synchronous configuration model collaborative decision-making mechanism, when the fault risk level conflicts with the remaining life prediction value, generates the final health state according to the preset priority strategy; Based on the final health status, output maintenance instructions containing the fault component identifier and the remaining life prediction value; Among them, the coupling failure logic is: When the load step temperature rise rate in the temperature rise feature set is greater than a preset reference ratio, and the fan vibration energy in the mechanical smoothness feature set is greater than a preset vibration threshold, the fault component identifier is defined as cooling fan bearing wear; When the dynamic contact resistance in the electrical fluency feature set is greater than a preset resistance threshold and the contact bounce duration in the mechanical fluency feature set is greater than a preset time threshold, the fault component identifier is defined as relay contact burnout.

6. A battery life detection method according to claim 5, characterized in that: Methods for configuring the model collaborative decision-making mechanism include: Real-time monitoring of failure risk level and remaining life prediction value; When a conflict is detected between the fault risk level and the remaining life prediction value, it is determined whether the fault risk level is ≥ a preset high risk threshold and the remaining life prediction value is > a preset safe life value; If yes, the real-time diagnostic layer is output as the final health status; If not, determine whether the fault risk level is ≤ the preset low risk threshold and the remaining life prediction value is ≤ the preset risk life value; If yes, the lifespan prediction layer is output as the final health state; If not, extract the standard deviation of the load step temperature rise rate fluctuation , synchronously extract the standard deviation of the dynamic contact resistance fluctuation ,Will and The weighted calculated value of is output as the final health state.

7. A battery life detection method according to claim 6, characterized in that: Will and The weighted calculation value output as the final health state method includes: Obtaining a quantitative value of the fault risk level and remaining life prediction , quantitative value of fault risk level Perform normalization and output the normalized value of fault risk , the remaining life prediction value Perform normalization and output the normalized value of remaining life ; Assign dynamic weight coefficients based on the environment state, , , is the standard deviation of the load step temperature rise rate fluctuation, is the benchmark stability value of the cooling system; The weighted health index is calculated as ; Compare the weighted health index H with the preset value range of H, and when H≤ When the final health status of the output is emergency failure, <H≤ When H> When , the final health status output is sub-healthy.

8. A battery life detection system, characterized in that: include: The acquisition module is used to obtain operating data, dynamic feature sets, and the health status of the entire machine; A memory for storing a program of a control method of a battery life detection method according to any one of claims 1 to 7; The program in the memory can be loaded and executed by the processor to implement the control method of the battery life detection method according to any one of claims 1 to 7.

9. An intelligent terminal, characterized in that: The device comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute the battery life detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The device stores a computer program capable of being loaded by a processor and executing the battery life detection method according to any one of claims 1 to 7.

Images