Electric vehicle charging safety monitoring method and device

Abstract

The invention discloses an electric vehicle charging safety monitoring method and device. The method comprises the steps that historical charging data and real-time charging data of a to-be-monitored electric vehicle are collected; obtaining fault charging data of a fault of a rechargeable battery of an electric vehicle of the same type as the to-be-monitored electric vehicle through a big data platform; synchronizing the historical charging data with the fault charging data to obtain a fault case; establishing a fault prediction model according to the fault case, and calculating the fault probability of the rechargeable battery of the to-be-monitored electric vehicle based on the fault prediction model; determining charging parameters of the to-be-monitored electric vehicle according to the real-time charging data and a pre-established charging MAP table; extracting a health influence factor as a sample according to the fault charging data, training a convolutional neural network, and obtaining a life prediction model; extracting features from the historical charging data and inputting the features into a life prediction model to obtain predicted life; the method can perform fault prediction and life prediction on the electric vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle safety, and in particular to a method and device for monitoring the charging safety of an electric vehicle. Background Art

[0002] At present, environmental and energy problems are becoming increasingly severe. Traditional fuel vehicles cause serious environmental pollution and consume a lot of energy, which has caused tremendous pressure on my country's construction of an environmental resource-based society and has aroused more and more attention and concerns. Electric vehicles use "electricity" instead of "fuel" to provide power. Compared with traditional fuel vehicles, they have the advantages of less emission pollution, lower noise, and higher energy utilization. Vigorously developing electric vehicles is an effective way to alleviate environmental pollution and energy consumption pressure, an important measure to promote the sustainable development of the automobile industry, and a strategic need for my country's automobile industry to transform and upgrade, cultivate new economic growth points, achieve new leapfrog development, and enhance international competitiveness.

[0003] The rapid development of electric vehicles has driven the entire industry chain, especially the rapid development of the charging industry. As the main way to replenish energy for electric vehicles, the planning and construction of charging facilities must take into account both the progress of industrial development and economic efficiency. Therefore, the construction and development of the intelligent charging service network for electric vehicles is rapid, actively responding to charging needs, and realizing orderly charging through active coordination between chargers, battery management systems and power grids, forming a wide-area coverage service network with networking, intelligence and standardization characteristics.

[0004] However, in recent years, the number of electric vehicle safety accidents has increased, and the causes of charging failures are very complicated. Due to the low degree of vehicle-pile intelligence and information interconnection, the charging operation platform cannot analyze and process charging pile failures in real time, resulting in the inability of charging facility operation and maintenance personnel to carry out targeted inspections and improvements, which seriously restricts the healthy and rapid development of electric vehicles. Patent text CN (111942186A) discloses a method and system for early warning of safety risks in the charging process of electric vehicles, including: obtaining the insulation resistance value of the charging circuit of the charger and the battery pack interaction information and single cell interaction information of the electric vehicle and the charging pile monitored by the charging information monitoring terminal; identifying the safety risks of the charging process of the electric vehicle according to the battery pack interaction information and single cell interaction information of the electric vehicle and the charging pile; using the single cell interaction information of the electric vehicle and the charging pile to identify the abnormal risk of single cell consistency in the charging process of the electric vehicle; using the insulation resistance value of the charging circuit of the charger to identify the abnormal risk of insulation resistance in the charging process of the electric vehicle; based on the identification results of the above risks, issuing early warnings to the charging operation platform and users. However, this method can only warn of conventional risks of electric vehicles, and cannot predict some potential risks of electric vehicles. Summary of the invention

[0005] The present invention provides a method and device for monitoring the charging safety of an electric vehicle, which can perform fault prediction and life prediction on the electric vehicle.

[0006] A method for monitoring the charging safety of an electric vehicle, comprising:

[0007] Collect historical charging data and real-time charging data of the electric vehicles to be monitored;

[0008] Acquire fault charging data of a rechargeable battery failure of an electric vehicle of the same type as the electric vehicle to be monitored through a big data platform;

[0009] Synchronize the historical charging data with the fault charging data to obtain a fault case;

[0010] Establishing a fault prediction model according to the fault case, and calculating the failure probability of the rechargeable battery of the electric vehicle to be monitored based on the fault prediction model;

[0011] Determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and a pre-established charging MAP table;

[0012] Extracting health impact factors as samples based on the faulty charging data, training a convolutional neural network, and obtaining a life prediction model;

[0013] Features are extracted from the historical charging data of the electric vehicle to be monitored, and the features are input into a life prediction model to obtain a predicted life of the electric vehicle to be monitored.

[0014] Further, the historical charging data includes the usage time of the rechargeable battery of the electric vehicle to be monitored, and the fault charging data includes the usage time of the rechargeable battery of the electric vehicle of the same type that has a fault;

[0015] Synchronizing the historical charging data with the fault charging data to obtain a fault case, including:

[0016] The fault charging data corresponding to the same type of electric vehicle rechargeable battery with the same service life as the rechargeable battery of the electric vehicle to be monitored is selected as the fault case.

[0017] Furthermore, a fault prediction model is established according to the fault case, including:

[0018] Obtaining fault parameters and safety parameters corresponding to the rechargeable battery that fails in the fault case, and calculating an average failure rate according to the fault parameters and safety parameters;

[0019] Calculate the standard deviation of the differences between different safety parameters of each faulty rechargeable battery;

[0020] The fault prediction model is established according to the average failure rate and the standard deviation.

[0021] Furthermore, the real-time charging data includes the SOC of the charging battery; the charging MAP table includes a mapping relationship between the charging temperature, SOC and the charging current;

[0022] Determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and the pre-established charging MAP table, including:

[0023] Detecting the temperature of each battery cell in the rechargeable battery of the electric vehicle to be monitored;

[0024] Select the minimum temperature among all cells as the initial charging temperature;

[0025] Determining the charging current of the next charging cycle according to the initial charging temperature, the rechargeable battery SOC and the charging MAP table;

[0026] During the charging process, the real-time temperature of each battery cell of the rechargeable battery of the monitored electric vehicle is collected, and the minimum real-time temperature is selected as the current temperature. The charging current of the next charging cycle is determined based on the current temperature, the rechargeable battery SOC and the charging MAP table.

[0027] Further, the fault charging data includes battery capacity and battery operating parameters;

[0028] According to the fault charging data, health impact factors are extracted as samples, and a convolutional neural network is trained to obtain a life prediction model, including:

[0029] constructing an initial correlation model for battery capacity according to the fault charging data;

[0030] Performing matrix transformation on the initial correlation model to obtain a final correlation model;

[0031] Obtaining first correlation coefficients between battery operating parameters according to the final correlation model;

[0032] Taking the battery operating parameters and the battery capacity as indicators, obtaining the weights of the indicators based on the hierarchical analysis method, establishing a weight matrix according to the weights of the indicators, and obtaining a second correlation coefficient according to the weight matrix;

[0033] Selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient;

[0034] Selecting an initial correlation influencing factor whose first correlation coefficient is greater than a first preset threshold as a health influencing factor;

[0035] The health impact factors and the corresponding battery life are marked to obtain samples, and the samples are input into a convolutional neural network for training to obtain a life prediction model.

[0036] Furthermore, the fault charging data includes the battery capacity corresponding to each operating parameter in different time periods;

[0037] The initial correlation model is as follows:

[0038]

[0039] In the formula, Q i (k) represents the historical battery capacity corresponding to the battery operating parameter i in the kth time period, Q j (k-1) represents the historical battery capacity corresponding to the battery operating parameter j in the k-1th time period, α ji represents the degree of influence of the historical battery capacity data corresponding to the battery operating parameter j in the k-1th time period on the historical battery capacity data corresponding to the battery operating parameter i in the kth time period, β i is a constant, γ i (k) represents the error.

[0040] Furthermore, the first correlation coefficient is determined by the following formula:

[0041]

[0042] In the formula, r ij represents the correlation coefficient of battery operating parameter i to battery operating parameter j, λ is an adjustable parameter, and n is the number of time periods; ρ i (t) represents the proportion of the battery capacity corresponding to the battery operating parameter i in the tth time period to the original total battery capacity.

[0043] Further, selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient includes:

[0044] The battery operation parameter corresponding to the second correlation coefficient greater than the second preset threshold is used as the initial correlation influencing factor.

[0045] Further, selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient includes:

[0046] The second correlation coefficients are sorted in descending order, and battery operating parameters corresponding to a preset number of second correlation coefficients that are ranked first are selected as initial correlation influencing factors.

[0047] An electric vehicle charging safety monitoring device, comprising:

[0048] A collection module, used to collect historical charging data and real-time charging data of the electric vehicle to be monitored;

[0049] A data acquisition module, used to acquire fault charging data of a rechargeable battery failure of an electric vehicle of the same type as the electric vehicle to be monitored through a big data platform;

[0050] A synchronization module, used to synchronize the historical charging data with the fault charging data to obtain a fault case;

[0051] A fault prediction module, used to establish a fault prediction model according to the fault case, and calculate the fault probability of the rechargeable battery of the electric vehicle to be monitored based on the fault prediction model;

[0052] A charging control module, used to determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and a pre-established charging MAP table;

[0053] A life prediction model building module is used to extract health impact factors as samples based on the fault charging data, train a convolutional neural network, and obtain a life prediction model;

[0054] The life prediction module is used to extract features from the historical charging data of the electric vehicle to be monitored, input the features into the life prediction model, and obtain the predicted life of the electric vehicle to be monitored.

[0055] Further, the historical charging data includes the usage time of the rechargeable battery of the electric vehicle to be monitored, and the fault charging data includes the usage time of the rechargeable battery of the electric vehicle of the same type that has a fault;

[0056] The synchronization module synchronizes the historical charging data with the fault charging data to obtain a fault case, including:

[0057] The fault charging data corresponding to the same type of electric vehicle rechargeable battery with the same service life as the rechargeable battery of the electric vehicle to be monitored is selected as the fault case.

[0058] Furthermore, the fault prediction module establishes a fault prediction model according to the fault case, including:

[0059] Obtaining fault parameters and safety parameters corresponding to the rechargeable battery that fails in the fault case, and calculating an average failure rate according to the fault parameters and safety parameters;

[0060] Calculate the standard deviation of the differences between different safety parameters of each faulty rechargeable battery;

[0061] The fault prediction model is established according to the average failure rate and the standard deviation.

[0062] Furthermore, the real-time charging data includes the SOC of the charging battery; the charging MAP table includes a mapping relationship between the charging temperature, SOC and the charging current;

[0063] The charging control module determines the charging parameters of the electric vehicle to be monitored according to the real-time charging data and the pre-established charging MAP table, including:

[0064] Detecting the temperature of each battery cell in the rechargeable battery of the electric vehicle to be monitored;

[0065] Select the minimum temperature among all cells as the initial charging temperature;

[0066] Determining the charging current of the next charging cycle according to the initial charging temperature, the rechargeable battery SOC and the charging MAP table;

[0067] During the charging process, the real-time temperature of each battery cell of the rechargeable battery of the monitored electric vehicle is collected, and the minimum real-time temperature is selected as the current temperature. The charging current of the next charging cycle is determined based on the current temperature, the rechargeable battery SOC and the charging MAP table.

[0068] Further, the fault charging data includes battery capacity and battery operating parameters;

[0069] The life prediction model building module extracts health impact factors as samples according to the fault charging data, trains the convolutional neural network, and obtains a life prediction model, including:

[0070] constructing an initial correlation model for battery capacity according to the fault charging data;

[0071] Performing matrix transformation on the initial correlation model to obtain a final correlation model;

[0072] Obtaining first correlation coefficients between battery operating parameters according to the final correlation model;

[0073] Taking the battery operating parameters and the battery capacity as indicators, obtaining the weights of the indicators based on the hierarchical analysis method, establishing a weight matrix according to the weights of the indicators, and obtaining a second correlation coefficient according to the weight matrix;

[0074] Selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient;

[0075] Selecting an initial correlation influencing factor whose first correlation coefficient is greater than a first preset threshold as a health influencing factor;

[0076] The health impact factors and the corresponding battery life are marked to obtain samples, and the samples are input into a convolutional neural network for training to obtain a life prediction model.

[0077] Furthermore, the fault charging data includes the battery capacity corresponding to each operating parameter in different time periods;

[0078] The initial correlation model is as follows:

[0079]

[0080] In the formula, Q i (k) represents the historical battery capacity corresponding to the battery operating parameter i in the kth time period, Q j (k-1) represents the historical battery capacity corresponding to the battery operating parameter j in the k-1th time period, α ji represents the degree of influence of the historical battery capacity data corresponding to the battery operating parameter j in the k-1th time period on the historical battery capacity data corresponding to the battery operating parameter i in the kth time period, β i is a constant, γ i (k) represents the error.

[0081] Furthermore, the first correlation coefficient is determined by the following formula:

[0082]

[0083] In the formula, r ij represents the correlation coefficient of battery operating parameter i to battery operating parameter j, λ is an adjustable parameter, and n is the number of time periods; ρ i (t) represents the proportion of the battery capacity corresponding to the battery operating parameter i in the tth time period to the original total battery capacity.

[0084] Furthermore, the life prediction model building module selects relevant battery operating parameters as initial correlation influencing factors according to the second correlation coefficient, including:

[0085] The battery operation parameter corresponding to the second correlation coefficient greater than the second preset threshold is used as the initial correlation influencing factor.

[0086] Furthermore, the life prediction model building module selects relevant battery operating parameters as initial correlation influencing factors according to the second correlation coefficient, including:

[0087] The second correlation coefficients are sorted in descending order, and battery operating parameters corresponding to a preset number of second correlation coefficients that are ranked first are selected as initial correlation influencing factors.

[0088] The electric vehicle charging safety monitoring method and device provided by the present invention at least include the following

[0089] Beneficial effects:

[0090] (1) Analyze the failure cases of the same type of electric vehicle rechargeable batteries obtained through the big data platform to obtain the failure probability of the electric vehicle to be monitored, which can provide reliable early warning for electric vehicle failures and indicate the potential risks of electric vehicles;

[0091] (2) During the charging process, the charging current is adjusted according to the real-time charging data to further ensure the safety of charging;

[0092] (3) The faulty charging data is analyzed and modeled. The life prediction model obtained can predict the life of the electric vehicle to be monitored, prompt the potential risks of the electric vehicle, and ensure the safe operation of the electric vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 The present invention provides a flow chart of an embodiment of the electric vehicle charging safety monitoring method.

[0094] Figure 2 The present invention provides a flowchart of an embodiment of establishing a fault prediction model in the electric vehicle charging safety monitoring method provided by the present invention.

[0095] Figure 3 The present invention provides a flowchart of an embodiment of determining charging parameters of an electric vehicle to be monitored in the electric vehicle charging safety monitoring method provided by the present invention.

[0096] Figure 4 A flowchart of an embodiment of obtaining a life prediction model in the electric vehicle charging safety monitoring method provided by the present invention.

[0097] Figure 5 The present invention provides a flow chart of an embodiment of the electric vehicle charging safety monitoring device. DETAILED DESCRIPTION

[0098] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0099] refer to Figure 1 In some embodiments, a method for monitoring charging safety of an electric vehicle is provided, comprising:

[0100] S1. Collect historical charging data and real-time charging data of the electric vehicle to be monitored;

[0101] S2. Acquire, through a big data platform, fault charging data of a rechargeable battery failure of an electric vehicle of the same type as the electric vehicle to be monitored;

[0102] S3, synchronizing the historical charging data with the fault charging data to obtain a fault case;

[0103] S4, establishing a fault prediction model according to the fault case, and calculating the failure probability of the rechargeable battery of the electric vehicle to be monitored based on the fault prediction model;

[0104] S5. Determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and the pre-established charging MAP table;

[0105] S6. Extracting health impact factors as samples based on the faulty charging data, training a convolutional neural network, and obtaining a life prediction model;

[0106] S7. Extracting features from the historical charging data of the electric vehicle to be monitored, and inputting the features into a life prediction model to obtain a predicted life of the electric vehicle to be monitored.

[0107] Specifically, in step S1, a communication connection can be established between the portable detection terminal and the onboard BMS, dedicated charging pile, etc. of the electric vehicle to be monitored to capture the historical charging data of the electric vehicle to be monitored, and the real-time charging data can be collected through the charging pile during the charging process.

[0108] Further, in some embodiments, the historical charging data includes the usage time of the rechargeable battery of the electric vehicle to be monitored, and in step S2, the fault charging data includes the usage time of the rechargeable battery of the electric vehicle of the same type that has a fault;

[0109] Specifically, in step S3, the historical charging data is synchronized with the fault charging data to obtain a fault case, including:

[0110] The fault charging data corresponding to the same type of electric vehicle rechargeable battery with the same service life as the rechargeable battery of the electric vehicle to be monitored is selected as the fault case.

[0111] Further, refer to Figure 2 In step S4, a fault prediction model is established according to the fault case, including:

[0112] S41, obtaining fault parameters and safety parameters corresponding to the rechargeable battery having a fault in the fault case, and calculating an average failure rate according to the fault parameters and safety parameters;

[0113] S42, calculating the standard deviation of the differences between different safety parameters of the rechargeable batteries with various faults;

[0114] S43: Establish the fault prediction model according to the average failure rate and the standard deviation.

[0115] Specifically, the fault charging data includes the relevant charging data of all faulty rechargeable batteries. The charging data related to the fault of each rechargeable battery may be different. Therefore, the charging data related to each faulty rechargeable battery is used as a fault parameter, and all fault-related data is used as safety parameters. In step S41, the ratio of the number of fault parameters to the number of safety parameters can be used as the average failure rate.

[0116] Furthermore, in step S42, all security parameters are regarded as sample data, and the standard deviation is the difference between multiple samples in the sample data. Therefore, the standard deviation is calculated by the following formula:

[0117]

[0118] Where β represents the standard deviation of the differences between different safety parameters, Xi represents the i-th safety parameter, and m represents the number of safety parameters. Indicates the arithmetic mean of the safety parameter.

[0119] As a statistic, standard deviation reflects the degree of deviation of each safety parameter value from the average value of these parameters, thereby evaluating the consistency and variability of safety parameters of different rechargeable batteries.

[0120] Further, in step S43, the fault prediction model is as follows:

[0121]

[0122] Where n represents the number of failure cases, γ represents the failure probability of the rechargeable battery of the electric vehicle to be monitored, T represents the usage frequency of the rechargeable battery within the preset time, which is the ratio of the usage time of the rechargeable battery within the preset time to the preset time; α represents the average failure rate in the failure case, and β represents the standard deviation of the differences between different safety parameters.

[0123] Specifically, the usage frequency T of the rechargeable battery within a preset time can be regarded as a time scale in the model, which affects the probability of failure occurrence.

[0124] In the method provided in the above embodiment, by analyzing the faults of electric vehicles of the same type, the failure probability of the electric vehicle to be monitored can be predicted based on the usage time of the rechargeable battery of the electric vehicle to be monitored, thereby achieving more comprehensive monitoring of the electric vehicle and ensuring the safe operation of the electric vehicle.

[0125] Further, in step S5, the real-time charging data includes the SOC of the charging battery; the charging MAP table includes a mapping relationship between the charging temperature, the SOC and the charging current;

[0126] refer to Figure 3 In step S51, the charging parameters of the electric vehicle to be monitored are determined according to the real-time charging data and the pre-established charging MAP table, including:

[0127] S51, detecting the temperature of each battery cell in the rechargeable battery of the electric vehicle to be monitored;

[0128] S52, selecting the minimum temperature among all the battery cells as the initial charging temperature;

[0129] S53, determining the charging current of the next charging cycle according to the initial charging temperature, the SOC of the rechargeable battery and the charging MAP table;

[0130] S54, during the charging process, the real-time temperature of each battery cell of the rechargeable battery of the monitored electric vehicle is collected, and the minimum real-time temperature is selected as the current temperature, and the charging current of the next charging cycle is determined according to the current temperature, the rechargeable battery SOC and the charging MAP table.

[0131] Specifically, the charging MAP table is created in advance, and the charging MAP table used in the general vehicle battery charging process can be referred to, which is usually provided by the battery manufacturer. The table includes information on the charging temperature segment, the SOC of the rechargeable battery, and the charging current. At the initial moment, different positions can be set to the same initial charging temperature, so that the corresponding charging current of the next charging cycle can be found according to the initial charging temperature and SOC and the pre-created MAP table, and the size of the charging current further causes changes in the charging temperature and SOC.

[0132] Furthermore, by selecting the lowest temperature in each battery cell as the characteristic temperature and obtaining the charging current for the next charging cycle, the corresponding charging current can be made more consistent with the safety of the rechargeable battery, thereby ensuring the charging safety of the monitored electric vehicle.

[0133] In some embodiments, the real-time charging data may also include single cell voltage, single cell temperature difference, vehicle current measurement value, vehicle voltage measurement value, vehicle voltage difference value, vehicle current difference value, etc.

[0134] During the charging process, when the single cell voltage exceeds the preset single cell voltage maximum value (4.25V) or the growth rate of the single cell voltage exceeds the preset growth rate maximum value (0.007575V / min), an early warning is issued;

[0135] When the temperature difference of a single battery exceeds the preset temperature difference range (2-6°C), an early warning is issued;

[0136] When the waveform frequency of the vehicle current measurement value in the early stage of charging exceeds the first preset frequency maximum value or does not oscillate, the average value in the middle stage of charging exceeds the first preset vehicle current range (much greater than or much less than 77.8A), and the current value monitored in real time in the late stage of charging exceeds the second preset vehicle current range (much greater than or much less than 47.4A) or is unstable, an early warning is issued;

[0137] When the vehicle voltage measurement value exceeds the preset vehicle voltage range (much less than 359V or much greater than 407V) or the growth rate of the vehicle voltage measurement value exceeds the preset growth range (much greater than or much less than 0.7272V / min), an early warning is issued;

[0138] When the rate of decrease of the vehicle voltage difference exceeds a preset decrease range (much greater than or much less than 0.7424 V / min), an early warning is issued;

[0139] When the waveform frequency of the vehicle current difference in the early stage of charging exceeds the second preset frequency maximum value or does not oscillate, the average value in the middle stage of charging exceeds the first preset difference range, and the vehicle current difference monitored in real time in the late stage of charging exceeds the second preset difference range or is unstable, an early warning is issued.

[0140] In some embodiments, the real-time charging data also includes: electric vehicle measured voltage and electric vehicle measured current.

[0141] Assume that the charging voltage required by the electric vehicle is V r , the electric vehicle measures a voltage of V m , the maximum allowable total charging voltage V is obtained during the parameter configuration phase of the electric vehicle charging process b , then the voltage safety factor of the electric vehicle charging process is V f The calculation formula is:

[0142]

[0143] Warning voltage safety factor threshold V fm The calculation formula is:

[0144]

[0145] Assume that the electric vehicle charging demand current is A r , the electric vehicle measures current in A m , the maximum allowable total charging current A is obtained during the parameter configuration stage of the electric vehicle charging process b , then the current safety factor A of the electric vehicle charging process f The calculation formula is:

[0146]

[0147] Warning current safety factor threshold A fm The calculation formula is:

[0148]

[0149] According to the specific vehicle model, the voltage safety factor V of each SOC stage is obtained. soc , current safety factor A soc , then the warning threshold alarm formula for quantitative scoring is:

[0150] V soc ≥V fm ; (7)

[0151] A soc ≥A fm ; (8)

[0152] If the calculation result of a certain SOC stage is V soc ≥V fm or A soc ≥A fm , then a charging safety warning is issued;

[0153] If the calculated result of a certain SOC stage is 0.75*V fm ≤V soc <V fm or 0.75*A fm ≤A soc fm , the battery status is judged as passing, no warning is issued, and the user is reminded to pay attention to battery maintenance;

[0154] If the calculation result of a certain SOC stage is 0.5*V fm ≤V soc <0.75*V fm or 0.5*A fm ≤A soc <0.75*A fm , the battery status is judged as medium, no warning is issued, and the user is reminded to pay attention to battery maintenance;

[0155] If the calculated result of a certain SOC stage is 0.25*V fm ≤V soc <0.5*V fm or 0.25*A fm ≤A soc <0.5*A fm , the battery status is judged to be good and no warning is issued;

[0156] If the calculation result of a certain SOC stage is 0≤V soc <0.25*V fm ​or 0≤A soc <0.25*A fm , the battery status is judged to be excellent and no warning is issued.

[0157] Further, in step S6, the fault charging data includes battery capacity and battery operating parameters;

[0158] refer to Figure 4 In step S6, the health impact factors are extracted as samples according to the faulty charging data, and the convolutional neural network is trained to obtain a life prediction model, including:

[0159] S61, constructing an initial correlation model for battery capacity according to the fault charging data;

[0160] S62, performing matrix conversion on the initial correlation model to obtain a final correlation model;

[0161] S63, obtaining a first correlation coefficient between two battery operating parameters according to the final correlation model;

[0162] S64, using the battery operating parameters and the battery capacity as indicators, obtaining weights of the indicators based on the hierarchical analysis method, establishing a weight matrix according to the weights of the indicators, and obtaining a second correlation coefficient according to the weight matrix;

[0163] S65, selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient;

[0164] S66, selecting an initial correlation influencing factor whose first correlation coefficient is greater than a first preset threshold as a health influencing factor;

[0165] S67. Mark the health impact factor and the corresponding battery life to obtain a sample, input the sample into a convolutional neural network for training, and obtain a life prediction model.

[0166] Further, in step S61, the fault charging data includes the battery capacity corresponding to each battery operating parameter in different time periods; the battery operating parameters may include power battery voltage, current, temperature, temperature rise rate, charging time, etc.

[0167] The initial correlation model is as follows:

[0168]

[0169] In the formula, Q i (k) represents the historical battery capacity corresponding to the battery operating parameter i in the kth time period, Q j (k-1) represents the historical battery capacity corresponding to the battery operating parameter j in the k-1th time period, αji represents the degree of influence of the historical battery capacity data corresponding to the battery operating parameter j in the k-1th time period on the historical battery capacity data corresponding to the battery operating parameter i in the kth time period, β i is a constant, γ i (k) represents the error.

[0170] Furthermore, in step S62, the process of performing matrix conversion on the initial correlation model is as follows:

[0171] Q in (k) = Q im (k)·α i +γ

[0172] =[Q i (k-n+1)Q i (k-n+2)…Q i (k-1)] T ; (10)

[0173]

[0174] α i =[α 0i α 1i …α Ni ] T ; (12)

[0175] γ=[γ1γ2…γ n-1 ] T ; (13)

[0176] The final correlation model is shown in equations (10)-(13).

[0177] Further, in step S63, the first correlation coefficient is determined by the following formula:

[0178]

[0179] In the formula, r ij represents the correlation coefficient of battery operating parameter i to battery operating parameter j, λ is an adjustable parameter, and n is the number of time periods; ρ i (t) represents the proportion of the battery capacity corresponding to the battery operating parameter i in the tth time period to the original total battery capacity.

[0180] In step S64, the battery operating parameters and battery capacity are taken as indicators, the weights of various indicators are obtained based on the hierarchical analysis method, a weight matrix is ​​established according to the weights of various indicators, and the second correlation coefficient is obtained according to the weight matrix, which specifically includes: taking various battery operating parameters and battery capacity as indicators, constructing a judgment matrix according to the importance of the correlation between various indicators, calculating the weights of various indicators according to the judgment matrix and performing a consistency check, establishing a weight matrix according to the weights that pass the consistency check, and each weight in the weight matrix is ​​used as the second correlation coefficient.

[0181] Furthermore, in step S65, selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient includes:

[0182] The battery operation parameter corresponding to the second correlation coefficient greater than the second preset threshold is used as the initial correlation influencing factor; or,

[0183] The second correlation coefficients are sorted in descending order, and battery operating parameters corresponding to a preset number of second correlation coefficients that are ranked first are selected as initial correlation influencing factors.

[0184] Specifically, when the second correlation coefficient is greater than the second preset threshold, or is ranked high, it means that the corresponding battery operating parameter is strongly correlated with the battery capacity, so it is used as the initial correlation influencing factor.

[0185] Furthermore, the initial correlation influencing factors are further screened by the second correlation coefficient, that is, the initial correlation influencing factors whose first correlation coefficients are greater than the first preset threshold are taken as health influencing factors.

[0186] In step S67, the health impact factor and the corresponding battery life are marked to obtain samples, and the samples are input into a convolutional neural network for training until the convolutional neural network converges to obtain a life prediction model.

[0187] Furthermore, in step S7, features are extracted from the historical charging data of the electric vehicle to be monitored, that is, relevant battery operating parameters in the historical charging data are extracted and input into a life prediction model, so as to obtain the predicted life of the electric vehicle to be monitored.

[0188] The above embodiment provides a method for monitoring the charging safety of an electric vehicle, which has at least the following beneficial effects:

[0189] (1) Analyze the failure cases of the same type of electric vehicle rechargeable batteries obtained through the big data platform to obtain the failure probability of the electric vehicle to be monitored, which can provide reliable early warning for electric vehicle failures and indicate the potential risks of electric vehicles;

[0190] (2) During the charging process, the charging current is adjusted according to the real-time charging data to further ensure the safety of charging;

[0191] (3) The faulty charging data is analyzed and modeled. The life prediction model obtained can predict the life of the electric vehicle to be monitored, prompt the potential risks of the electric vehicle, and ensure the safe operation of the electric vehicle.

[0192] refer to Figure 5 In some embodiments, a device for monitoring the charging safety of an electric vehicle is also provided, comprising:

[0193] The collection module 201 is used to collect historical charging data and real-time charging data of the electric vehicle to be monitored;

[0194] A data acquisition module 202 is used to acquire fault charging data of a rechargeable battery failure of an electric vehicle of the same type as the electric vehicle to be monitored through a big data platform;

[0195] A synchronization module 203, used to synchronize the historical charging data with the fault charging data to obtain a fault case;

[0196] A fault prediction module 204 is used to establish a fault prediction model according to the fault case, and calculate the fault probability of the rechargeable battery of the electric vehicle to be monitored based on the fault prediction model;

[0197] The charging control module 205 is used to determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and the pre-established charging MAP table;

[0198] A life prediction model building module 206 is used to extract health impact factors as samples according to the fault charging data, train a convolutional neural network, and obtain a life prediction model;

[0199] The life prediction module 207 is used to extract features from the historical charging data of the electric vehicle to be monitored, input the features into the life prediction model, and obtain the predicted life of the electric vehicle to be monitored.

[0200] Further, the historical charging data includes the usage time of the rechargeable battery of the electric vehicle to be monitored, and the fault charging data includes the usage time of the rechargeable battery of the electric vehicle of the same type that has a fault;

[0201] The synchronization module 203 synchronizes the historical charging data with the fault charging data to obtain a fault case, including:

[0202] The fault charging data corresponding to the same type of electric vehicle rechargeable battery with the same service life as the rechargeable battery of the electric vehicle to be monitored is selected as the fault case.

[0203] Furthermore, the fault prediction module 204 establishes a fault prediction model according to the fault case, including:

[0204] Obtaining fault parameters and safety parameters corresponding to the rechargeable battery that fails in the fault case, and calculating an average failure rate according to the fault parameters and safety parameters;

[0205] Calculate the standard deviation of the differences between different safety parameters of each faulty rechargeable battery;

[0206] The fault prediction model is established according to the average failure rate and the standard deviation.

[0207] Furthermore, the fault prediction model is as follows:

[0208]

[0209] Where n represents the number of failure cases, γ represents the failure probability of the rechargeable battery of the electric vehicle to be monitored, T represents the frequency of use of the rechargeable battery within a preset time; α represents the average failure rate in the failure cases, and β represents the standard deviation of the differences between different safety parameters.

[0210] Furthermore, the real-time charging data includes the SOC of the charging battery; the charging MAP table includes a mapping relationship between the charging temperature, SOC and the charging current;

[0211] The charging control module 205 determines the charging parameters of the electric vehicle to be monitored according to the real-time charging data and the pre-established charging MAP table, including:

[0212] Detecting the temperature of each battery cell in the rechargeable battery of the electric vehicle to be monitored;

[0213] Select the minimum temperature among all cells as the initial charging temperature;

[0214] Determining the charging current of the next charging cycle according to the initial charging temperature, the rechargeable battery SOC and the charging MAP table;

[0215] During the charging process, the real-time temperature of each battery cell of the rechargeable battery of the monitored electric vehicle is collected, and the minimum real-time temperature is selected as the current temperature. The charging current of the next charging cycle is determined based on the current temperature, the rechargeable battery SOC and the charging MAP table.

[0216] Further, the fault charging data includes battery capacity and battery operating parameters;

[0217] The life prediction model building module 206 extracts health impact factors as samples according to the faulty charging data, trains the convolutional neural network, and obtains a life prediction model, including:

[0218] constructing an initial correlation model for battery capacity according to the fault charging data;

[0219] Performing matrix transformation on the initial correlation model to obtain a final correlation model;

[0220] Obtaining first correlation coefficients between battery operating parameters according to the final correlation model;

[0221] Taking the battery operating parameters and the battery capacity as indicators, obtaining the weights of the indicators based on the hierarchical analysis method, establishing a weight matrix according to the weights of the indicators, and obtaining a second correlation coefficient between the battery operating parameters and the battery capacity according to the weight matrix;

[0222] Selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient;

[0223] Selecting an initial correlation influencing factor whose first correlation coefficient is greater than a first preset threshold as a health influencing factor;

[0224] The health impact factors and the corresponding battery life are marked to obtain samples, and the samples are input into a convolutional neural network for training to obtain a life prediction model.

[0225] Furthermore, the fault charging data includes the battery capacity corresponding to each operating parameter in different time periods;

[0226] The initial correlation model is as follows:

[0227]

[0228] In the formula, Q i (k) represents the historical battery capacity corresponding to the battery operating parameter i in the kth time period, Q j (k-1) represents the historical battery capacity corresponding to the battery operating parameter j in the k-1th time period, α ji represents the degree of influence of the historical battery capacity data corresponding to the battery operating parameter j in the k-1th time period on the historical battery capacity data corresponding to the battery operating parameter i in the kth time period, β i is a constant, γ i (k) represents the error.

[0229] Furthermore, the first correlation coefficient is determined by the following formula:

[0230]

[0231] In the formula, r ij represents the correlation coefficient of battery operating parameter i to battery operating parameter j, λ is an adjustable parameter, and n is the number of time periods; ρ i (t) represents the proportion of the battery capacity corresponding to the battery operating parameter i in the tth time period to the original total battery capacity.

[0232] Furthermore, the life prediction model building module 206 selects relevant battery operating parameters as initial correlation influencing factors according to the second correlation coefficient, including:

[0233] The battery operation parameter corresponding to the second correlation coefficient greater than the second preset threshold is used as the initial correlation influencing factor; or,

[0234] The second correlation coefficients are sorted in descending order, and battery operating parameters corresponding to a preset number of second correlation coefficients that are ranked first are selected as initial correlation influencing factors.

[0235] Please refer to the method embodiment for the specific working principle, which will not be repeated here.

[0236] The above embodiment provides an electric vehicle charging safety monitoring device, which has at least the following beneficial effects:

[0237] (1) Analyze the failure cases of the same type of electric vehicle rechargeable batteries obtained through the big data platform to obtain the failure probability of the electric vehicle to be monitored, which can provide reliable early warning for electric vehicle failures and indicate the potential risks of electric vehicles;

[0238] (2) During the charging process, the charging current is adjusted according to the real-time charging data to further ensure the safety of charging;

[0239] (3) The faulty charging data is analyzed and modeled. The life prediction model obtained can predict the life of the electric vehicle to be monitored, prompt the potential risks of the electric vehicle, and ensure the safe operation of the electric vehicle.

[0240] Although preferred embodiments of the present invention have been described, additional changes and modifications may be made to these embodiments by those skilled in the art once the basic inventive concepts are known. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for monitoring the charging safety of an electric vehicle, characterized in that: include: Collect historical charging data and real-time charging data of the electric vehicles to be monitored; Acquire fault charging data of a rechargeable battery failure of an electric vehicle of the same type as the electric vehicle to be monitored through a big data platform; Synchronize the historical charging data with the fault charging data to obtain a fault case; Establishing a fault prediction model according to the fault case, and calculating the failure probability of the rechargeable battery of the electric vehicle to be monitored based on the fault prediction model; Determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and a pre-established charging MAP table; Extracting health impact factors as samples based on the faulty charging data, training a convolutional neural network, and obtaining a life prediction model; Features are extracted from the historical charging data of the electric vehicle to be monitored, and the features are input into a life prediction model to obtain a predicted life of the electric vehicle to be monitored.

2. The method according to claim 1, characterized in that The historical charging data includes the usage time of the rechargeable battery of the electric vehicle to be monitored, and the fault charging data includes the usage time of the rechargeable battery of the electric vehicle of the same type that has a fault; Synchronizing the historical charging data with the fault charging data to obtain a fault case, including: The fault charging data corresponding to the same type of electric vehicle rechargeable battery with the same service life as the rechargeable battery of the electric vehicle to be monitored is selected as the fault case.

3. The method according to claim 2, characterized in that A fault prediction model is established according to the fault case, including: Obtaining fault parameters and safety parameters corresponding to the rechargeable battery that fails in the fault case, and calculating an average failure rate according to the fault parameters and safety parameters; Calculate the standard deviation of the differences between different safety parameters of each faulty rechargeable battery; The fault prediction model is established according to the average failure rate and the standard deviation.

4. The method according to claim 1, characterized in that: The real-time charging data includes the charging battery SOC; the charging MAP table includes the mapping relationship between the charging temperature, SOC and the charging current; Determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and the pre-established charging MAP table, including: Detecting the temperature of each battery cell in the rechargeable battery of the electric vehicle to be monitored; Select the minimum temperature among all cells as the initial charging temperature; Determining the charging current of the next charging cycle according to the initial charging temperature, the rechargeable battery SOC and the charging MAP table; During the charging process, the real-time temperature of each battery cell of the rechargeable battery of the monitored electric vehicle is collected, and the minimum real-time temperature is selected as the current temperature. The charging current of the next charging cycle is determined based on the current temperature, the rechargeable battery SOC and the charging MAP table.

5. The method according to claim 1, characterized in that The fault charging data includes battery capacity and battery operating parameters; According to the fault charging data, health impact factors are extracted as samples, and a convolutional neural network is trained to obtain a life prediction model, including: constructing an initial correlation model for battery capacity according to the fault charging data; Performing matrix transformation on the initial correlation model to obtain a final correlation model; Obtaining first correlation coefficients between battery operating parameters according to the final correlation model; Taking the battery operating parameters and the battery capacity as indicators, obtaining the weights of the indicators based on the hierarchical analysis method, establishing a weight matrix according to the weights of the indicators, and obtaining a second correlation coefficient according to the weight matrix; Selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient; Selecting an initial correlation influencing factor whose first correlation coefficient is greater than a first preset threshold as a health influencing factor; The health impact factors and the corresponding battery life are marked to obtain samples, and the samples are input into a convolutional neural network for training to obtain a life prediction model.

6. The method according to claim 5, characterized in that The fault charging data includes the battery capacity corresponding to each operating parameter in different time periods; The initial correlation model is as follows: In the formula, Q i (k) represents the historical battery capacity corresponding to the battery operating parameter i in the kth time period, Q j (k-1) represents the historical battery capacity corresponding to the battery operating parameter j in the k-1th time period, α ji represents the degree of influence of the historical battery capacity data corresponding to the battery operating parameter j in the k-1th time period on the historical battery capacity data corresponding to the battery operating parameter i in the kth time period, β i is a constant, γ i (k) represents the error.

7. The method according to claim 6, characterized in that The first correlation coefficient is determined by the following formula: In the formula, r ij represents the correlation coefficient of battery operating parameter i to battery operating parameter j, λ is an adjustable parameter, and n is the number of time periods; ρ i (t) represents the proportion of the battery capacity corresponding to the battery operating parameter i in the tth time period to the original total battery capacity.

8. The method according to claim 6, characterized in that Selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient includes: The battery operation parameter corresponding to the second correlation coefficient greater than the second preset threshold is used as the initial correlation influencing factor.

9. The method according to claim 6, characterized in that Selecting a relevant battery operating parameter as an initial correlation influencing factor according to the second correlation coefficient includes: The second correlation coefficients are sorted in descending order, and battery operating parameters corresponding to a preset number of second correlation coefficients that are ranked first are selected as initial correlation influencing factors.

10. An electric vehicle charging safety monitoring device, characterized in that: include: A collection module, used to collect historical charging data and real-time charging data of the electric vehicle to be monitored; A data acquisition module, used to acquire fault charging data of a rechargeable battery failure of an electric vehicle of the same type as the electric vehicle to be monitored through a big data platform; A synchronization module, used to synchronize the historical charging data with the fault charging data to obtain a fault case; A fault prediction module, used to establish a fault prediction model according to the fault case, and calculate the fault probability of the rechargeable battery of the electric vehicle to be monitored based on the fault prediction model; A charging control module, used to determine the charging parameters of the electric vehicle to be monitored according to the real-time charging data and a pre-established charging MAP table; A life prediction model building module is used to extract health impact factors as samples based on the fault charging data, train a convolutional neural network, and obtain a life prediction model; The life prediction module is used to extract features from the historical charging data of the electric vehicle to be monitored, input the features into the life prediction model, and obtain the predicted life of the electric vehicle to be monitored.

Images