Electric vehicle load spectrum compilation method based on strong user association and related device

Abstract

The invention discloses an electric vehicle load spectrum compilation method based on strong user association and a related device. Preprocessed online operation data is divided into equal-mileage fragment data, characteristic parameters of the equal-mileage fragment data are calculated, and an initial database is constructed by using the characteristic parameters of the equal-mileage fragment data; the characteristic parameters of the equal-mileage fragment data in the initial database are subjected to multiple times of optimization pairing iteration, a typical data set is obtained through screening, and the deviation ratio between the characteristic parameters of the equal-mileage fragment data in the typical data set and the user vehicle model is within a set range; setting constraint conditions, and orderly splicing the equal-mileage fragment data in the typical data set to obtain spliced online operation data; and converting the spliced online operation data into an online load spectrum, and fusing the online load spectrum with the obtained offline load spectrum to obtain an electric vehicle driving system load spectrum. The objective of the invention is to solve the problems of difficult determination of typical data and poor fatigue damage consistency of online low-frequency data in the existing load spectrum compilation, and to realize effective association and damage equivalence of the load spectrum and the user characteristics, thereby more accurately reflecting the load damage condition of the electric vehicle driving system in actual use.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle load spectrum compilation, and in particular to an electric vehicle load spectrum compilation method based on strong user association and a related device. Background Art

[0002] With the continuous development of electric vehicle technology, the improvement of motor torque has significantly increased the power output of the vehicle, providing users with a more powerful power experience. However, this progress has also brought new problems: the risk of failure of electric vehicle drive systems has increased. Therefore, reliability testing of electric vehicle drive systems to ensure their stable operation under various working conditions has become a key link in the research and development and production process. In actual operation, since electric vehicles will face a variety of complex scenarios in actual use, such as different road conditions, driving habits, load conditions, etc., these will cause different load damage to the drive system.

[0003] Current reliability tests often fail to fully cover the damage intensity of these user loads, resulting in deviations between test results and actual usage. Specifically, although traditional real-car road tests can simulate the actual usage of users to a certain extent, the data sample size is usually small and lacks representativeness, making it difficult to fully reflect the characteristics of the user group. In addition, real-car road tests are also limited by factors such as time and cost, making it difficult to conduct large-scale tests. On the other hand, although online big data can provide rich vehicle operation data, it is often easy to lose important peak load information due to the low frequency of data collection. This information is crucial for accurately calculating load damage. Once lost, it will affect the accuracy of load spectrum compilation, and thus affect the results of drive system reliability assessment. Summary of the invention

[0004] In view of the problems existing in the prior art, the present invention provides a method and related devices for compiling an electric vehicle load spectrum based on strong user association, which aims to solve the problems of difficulty in determining typical data and poor consistency of fatigue damage of online low-frequency data in the existing load spectrum compilation, and to achieve effective association and damage equivalence between the load spectrum and user characteristics, so as to more accurately reflect the load damage situation of the electric vehicle drive system in actual use.

[0005] In order to solve the above technical problems, the present invention is implemented by the following technical solutions: According to a first aspect of the present invention, a method for compiling an electric vehicle load spectrum based on strong user association is provided, comprising: Obtain online operation data of user vehicles under different working conditions; Preprocessing the online operation data to obtain preprocessed online operation data, wherein the preprocessing includes resampling the online operation data to improve and unify the sampling frequency of the online operation data; Dividing the preprocessed online operation data into equal mileage segment data, calculating characteristic parameters of the equal mileage segment data, and constructing an initial database using the characteristic parameters of the equal mileage segment data; Perform multiple optimization pairing iterations on the feature parameters of the equal mileage segment data in the initial database to screen out a typical data set, wherein the deviation rate between the feature parameters of the equal mileage segment data in the typical data set and the user vehicle model is within a set range; Set constraint conditions, and sequentially splice the equal-mileage fragment data in the typical data set to obtain spliced ​​online operation data; The spliced ​​online operation data is converted into an online load spectrum and fused with the acquired offline load spectrum to obtain the load spectrum of the electric vehicle drive system.

[0006] In a possible implementation manner of the first aspect, the characteristic parameters of the equal mileage segment data include load state distribution parameters, driving speed distribution parameters, motor torque distribution parameters, driving style distribution parameters and road type distribution parameters.

[0007] In a possible implementation manner of the first aspect, a method for calculating the load state distribution parameter, the driving speed distribution parameter, or the motor torque distribution parameter is:

[0008] In the formula, represents the sampling time; Represents the monthly mileage of the user's vehicle; When calculating the load state distribution parameters, m Represents the full load / half load status serial number; j The sequence number representing the full load / half load state segment; Represents full load / half load status m Corresponding mileage ratio; Represents full load / half load status fragment j The corresponding vehicle speed; Represents the full load / half load status of the user's vehicle in the monthly data m Total number of; When calculating the driving speed distribution parameters, m Represents the driving speed state sequence number; j The serial number representing the driving speed state segment; Represents the driving speed status m Corresponding mileage ratio; Represents the driving speed status fragment j The corresponding vehicle speed; Represents the driving speed status of the user's vehicle in the monthly data m Total number of; When calculating the motor torque distribution parameters, m Represents the motor torque state number; j The serial number representing the motor torque state segment; Represents the motor torque state m Corresponding mileage ratio; Represents the motor torque state fragment j The corresponding vehicle speed; Represents the motor torque status in the user's vehicle monthly data m The total number of .

[0009] In a possible implementation manner of the first aspect, a method for calculating the driving style distribution parameter is:

[0010]

[0011]

[0012]

[0013]

[0014] Where: is the sequence number of the acceleration point; is the total number of acceleration points; is the mean absolute value of acceleration; For the The acceleration of an acceleration point; is the standard deviation of the absolute value of acceleration; For the The jerk of each acceleration point, that is, the rate of change of acceleration; is the mean absolute value of jerk; is the standard deviation of the absolute value of jerk; The acceleration absolute mean, acceleration absolute standard deviation, jerk absolute mean and jerk absolute standard deviation of all equal mileage segments are used to form a characteristic parameter matrix, and the K-Means clustering algorithm is used to perform clustering operation on the characteristic parameter matrix. The cluster center is determined through multiple iterative calculations, and the driving styles are divided into three types: aggressive, general and conservative according to the distance relationship between the equal mileage segments and the cluster center. For each user's vehicle monthly operation data, the equal mileage segments are extracted and their driving style distribution parameters are calculated. The Euclidean distance between the driving style distribution parameters of each equal mileage segment and the three cluster centers is calculated, and the driving style of the equal mileage segments is determined based on the principle of the shortest distance. Finally, the mileage of each driving style is counted to obtain the mileage proportion of the user's vehicle monthly data driving style.

[0015] In a possible implementation manner of the first aspect, a method for calculating the road type distribution parameter is: The short-trip method is used to divide the user's monthly data into short-trip segments. The user's driving road type is determined based on the GPS range library and the user's data latitude and longitude. The mileage of each road type is counted to obtain the mileage ratio of the user's monthly data road type.

[0016] In a possible implementation of the first aspect, the feature parameters of the equal-mileage segment data in the initial database are subjected to multiple optimization pairing iterations to screen out a typical data set, specifically: The initial database n The data of equal mileage segments are represented as X i =( , ,…, ), i =1,2,…, n ,in - Respectively represent the road type mileage ratio, - Indicates the driving style mileage ratio, - Indicates the ratio of full and half load mileage. - Indicates the driving speed-mileage ratio, - Indicates the motor torque mileage ratio; set the target point G, traverse and calculate the initial database X i With the rest of the data X j Euclidean distance between the paired points and the target point G , the calculation formula is:

[0017] in, j = 1,2,…,n , j ≠ i ; ∈G) Record Minimum value of L imin and its corresponding X j , find L imin The minimum value of X i , X j , remove the two from the initial database and pair them to get data P=(P 1 , P 2 , …, P 15 ), where P k =(X ik + X jk ) / 2,P k ∈P, k = 1,2, …, 15); Update the calculation of the remaining data of the initial database , L imin , repeat the above pairing and updating process, obtain the pairing results in sequence, until the database pairing is completed; accumulate the paired data in sequence, and calculate the deviation rate between the accumulated data mean and the target point and average deviation rate , the calculation formula is:

[0018]

[0019] h Cumulative number of paired data; k =1, 2, …, 15; repeat the above process with the paired data as the new initial database, and take the accumulated data mileage as the target mileage MPE ≤0.05 and max ( PE )≤0.15 as the termination condition for data matching iteration, and finally screen out a typical data set that meets the user's usage characteristics.

[0020] In a possible implementation manner of the first aspect, setting the constraint condition to sequentially splice the equal-mileage segment data in the typical data set is specifically: For the two equal-length segments involved in the splicing, let the value and slope of the tail of the previous equal-length segment be T 1 and v 1 , the value and slope of the first part of the next equal mileage segment are T 2 andv 2 , set the splicing constraints as: | T 1 - T 2 | and · All are within the set range; For any equal mileage segment in the selected typical data set, traverse the remaining equal mileage segments according to the splicing constraints, and form a set of equal mileage segments that can be spliced ​​before it fi , the subsequent equal-mileage segments can be spliced ​​to form a set ri ; Select the corresponding set fi The equal mileage segment with the least equal mileage segments is taken as the first equal mileage segment to be spliced. If there are multiple equal mileage segments, the corresponding set is taken. ri The one with the most equal-mileage segments; in the set of the first equal-mileage segment ri In the dialog box, select the corresponding set ri The one with the most equal-mileage segments is used as the second equal-mileage segment to be spliced, and so on; when the splicing is interrupted at the nth equal-mileage segment, if n is the total number of equal-mileage segments in the screened data set, the splicing is completed; otherwise, the numerical restrictions or slope restrictions after the splicing of the nth equal-mileage segment are relaxed until the splicing is completed, and the online running data after splicing is obtained.

[0021] 8. The method for compiling an electric vehicle load spectrum based on strong user association according to claim 1, characterized in that the offline load spectrum is obtained by: Obtain offline operation data of user vehicles under different working conditions through real-vehicle road collection solutions; The offline operation data is extrapolated, firstly rain flow counting is performed on the offline operation data to obtain a rain flow matrix, and then the distribution of the rain flow matrix is ​​estimated by using a kernel density estimation method to obtain a load cycle distribution probability density function; According to the probability density function of the load cycle distribution of each working condition and the mileage proportion of each working condition, the composite load data of multiple working conditions are superimposed to obtain the offline load spectrum.

[0022] According to a second aspect of the present invention, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for compiling an electric vehicle load spectrum based on strong user association is implemented.

[0023] According to a third aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for compiling an electric vehicle load spectrum based on strong user association is implemented.

[0024] Compared with the prior art, the present invention has at least the following beneficial effects: The present invention provides a method for compiling a load spectrum of an electric vehicle based on strong user association. By acquiring online operation data of a user's vehicle under different working conditions and performing comprehensive and meticulous processing, the method can accurately fit the actual use scenario of the user. Starting from the source collection of the operation data, a rich variety of working conditions are covered to ensure the comprehensiveness and authenticity of the data. After preprocessing and resampling to improve the quality and sampling frequency of the online operation data, the more subtle operation changes of the user's driving vehicle can be captured, so that the data can more accurately reflect the actual operation status of the vehicle. When constructing the initial database, the characteristic parameters of the equal mileage segment data are calculated to further mine the key information about the user's driving vehicle in the data. The typical data sets are screened out through multiple optimal matching iterations, ensuring that the data are highly matched with the user's vehicle model, and the deviation rate is controlled within a set range. The load spectrum finally formed can highly restore the actual operation of the user's vehicle, provide an accurate basis for the design and optimization of the electric vehicle drive system, effectively avoid the design deviation caused by the data not being consistent with the actual use scenario, and improve the reliability and applicability of the load spectrum. In the data processing process, the present invention divides the data into equal mileage segment data and calculates the characteristic parameters, which helps to structure and standardize the data for comparison and analysis. The method of multiple optimal pairing iterations to screen typical data sets can accurately find the most representative data from a large amount of data, avoid data redundancy and interference, and improve data utilization. Setting constraints to orderly splice the equal mileage fragment data in the typical data set ensures the coherence and logic of the data, so that the spliced ​​online operation data is more in line with the actual situation. These scientific and effective data processing methods ensure the accuracy of the load spectrum finally obtained, which is conducive to the performance evaluation and optimization of the electric vehicle drive system. The present invention realizes the effective fusion of online and offline data, gives full play to the advantages of the two types of data, and provides more comprehensive and accurate information for the compilation of the load spectrum of the electric vehicle drive system. The online operation data has a wide coverage and real-time performance, and can reflect the actual use of users under different working conditions; while the offline load spectrum may contain more accurate experimental data and detailed information under specific working conditions. By converting the spliced ​​online operation data into an online load spectrum and fusing it with the offline load spectrum, the advantages of the two types of data can be comprehensively utilized to make up for the shortcomings of a single data source. This fused load spectrum can more comprehensively reflect the stress conditions of the electric vehicle drive system under various actual working conditions, and provide a more reliable basis for the reliability design and durability evaluation of the drive system.

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the specific implementation modes of the present invention, the drawings required for use in the description of the specific implementation modes will be briefly introduced below. Obviously, the drawings described below are some implementation modes of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0027] Figure 1 : A flowchart of an electric vehicle load spectrum compilation method based on strong user association.

[0028] Figure 2 :Load spectrum data acquisition program flow chart; clarify the overall process of data acquisition and the relationship between each link.

[0029] Figure 3 :GPS range library partitioning flow chart: shows the steps of building a road type GPS range library.

[0030] Figure 4 :Flowchart of data screening method based on global optimal pairing Figure 5 : Constraint-based data splicing flow chart: clarify the data splicing process.

[0031] Figure 6 : Full life cycle load spectrum: reflects the load spectrum after the fusion of online and offline data.

[0032] Figure 7 : Logarithmic cumulative frequency curve: used to analyze the fatigue characteristics of the load spectrum.

[0033] Figure 8 : Load spectrum damage calculation result diagram. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] like Figure 1 and Figure 2 As shown, the embodiment of the present invention provides a method for compiling an electric vehicle load spectrum based on strong user association, which specifically includes the following steps: S1. Obtain online operation data of user vehicles under different working conditions.

[0036] In one possible implementation, online operation data of user vehicles under different working conditions is obtained from the intelligent network platform. For example, typical cities such as Beijing, Chongqing, Chengdu, and Xi'an are selected as data collection sources, and massive user vehicle operation data is fully extracted on the intelligent network platform, including precise time, driving speed, GPS positioning, seat sensor signals, drive motor speed and torque and other multi-dimensional information.

[0037] Specifically, by analyzing the impact of the number of user vehicle samples on the representativeness of the load spectrum, it is found that the mean value of load damage tends to be stable and the variance decreases as the number of user vehicles increases. Therefore, a big data platform (on the intelligent network platform) is selected to obtain typical data using an online data screening method based on user vehicle portraits.

[0038] For example, randomly select n user vehicles as independent variables, select n user vehicles based on the big data platform and extract load data of a certain length (10,000 kilometers), repeat 10 times to obtain 10 load damage values ​​under n user quantities, calculate the damage mean and variance, and then compare and analyze the corresponding relationship between user quantity and load damage. Calculate the mean and variance of the 10 load damage values ​​under each user sample, draw the relationship curve between the user quantity and the damage mean and variance, and by observing the trend of the curve, it is clear that when the number of user samples reaches a certain scale, the load damage mean tends to stabilize and the variance gradually decreases.

[0039] S2. Preprocessing the online operation data to obtain preprocessed online operation data, wherein the preprocessing includes resampling the online operation data to increase the sampling frequency of the online operation data.

[0040] In one achievable manner, the online operation data is resampled using a resampling method.

[0041] For example, a set of transmission shaft torque sequences with a duration of 1.5 hours and a sampling frequency of 1000 Hz is selected from the actual vehicle test data as the original data. The original data should cover the operation of the vehicle under various working conditions and be representative. The original data is resampled by data extraction sampling, and different sampling frequencies (such as 500 Hz, 250 Hz, 100 Hz, 50 Hz, 10 Hz, 5 Hz, 2 Hz, 1 Hz, 0.1 Hz, etc.) are set. Each time the data is resampled, a series of load data with different sampling frequencies is generated according to the corresponding extraction rules (such as when 1000 Hz data is resampled to 100 Hz, 1 point is extracted from every 10 points to form a new data sequence).

[0042] Tecware software is used to count the load cycles using the rain flow counting method for the load data at each sampling rate, and a rain flow matrix diagram of the data at each sampling rate is obtained to intuitively display the distribution of the load cycles.

[0043] To elaborate, the effect of sampling rate on the accuracy of load fatigue damage and reliability testing is explored. By resampling and comparing load cycle counts and fatigue damage at different sampling rates, it is determined that the low sampling rate data has poor consistency in load cycle counts and fatigue damage, and then high-frequency data from actual vehicle tests is collected.

[0044] Preferably, the preprocessing further includes processing outliers in the online operation data and eliminating idle data. Exemplarily, processing outlier points based on the Darai criterion and linear interpolation method and eliminating long-term idle data are performed to ensure the quality of the online operation data.

[0045] S3, dividing the preprocessed online operation data into equal mileage segment data, calculating characteristic parameters of the equal mileage segment data, and constructing an initial database using the characteristic parameters of the equal mileage segment data.

[0046] In one possible implementation, the pre-processed online operation data is divided by an equal mileage division method. For example, the pre-processed online operation data is divided into equal mileage segments by an equal mileage division method based on 10 km as the standard, ensuring that the mileage value of each equal mileage segment is consistent, and the characteristic parameters are superimposed during subsequent data pairing.

[0047] In one achievable manner, the characteristic parameters of the equal mileage segment data include a load state distribution parameter, a driving speed distribution parameter, a motor torque distribution parameter, a driving style distribution parameter and a road type distribution parameter.

[0048] Specifically, the load status distribution is to evaluate the vehicle's full / half-load status by sensing the number of passengers through seat sensors, and calculate the full / half-load mileage ratio.

[0049] Driving speed distribution is to divide driving speed states according to the driving speed cumulative mileage ratio curve, and calculate the mileage ratio of each driving speed state. Motor torque distribution is to divide motor torque states according to the motor torque cumulative mileage ratio curve, and calculate the mileage ratio of each motor torque state.

[0050] In one possible implementation, the calculation method of the load state distribution parameter, the driving speed distribution parameter or the motor torque distribution parameter is:

[0051] In the formula, represents the sampling time, s; Represents the monthly mileage of the user's vehicle, km; When calculating the load state distribution parameters, m Represents the full load / half load status serial number; j The sequence number representing the full load / half load state segment; Represents full load / half load status m Corresponding mileage ratio; Represents full load / half load status fragment j Corresponding vehicle speed, km / h; Represents the full load / half load status of the user's vehicle in the monthly data m The total number of .

[0052] For example, regarding the load status: high-precision seat sensors are installed on the vehicle seats to collect pressure change signals on the seats in real time to determine the number of passengers in the vehicle. The threshold for determining the full / half load status is set according to the maximum number of passengers in the vehicle. Suppose the maximum number of passengers in the vehicle is M, and the number of passengers in the vehicle obtained from the seat sensor signal is m. When m is greater than 0.5M, it is determined to be fully loaded, and when it is less than 0.5M, it is half loaded. Within a one-month period, the vehicle operation data is classified and counted according to the set judgment rules to determine the mileage under full and half load conditions.

[0053] When calculating the driving speed distribution parameters, m Represents the driving speed state sequence number; j The serial number representing the driving speed state segment; Represents the driving speed status m Corresponding mileage ratio; Represents the driving speed status fragment j The corresponding vehicle speed; Represents the driving speed status of the user's vehicle in the monthly data m The total number of .

[0054] For example, regarding the distribution of driving speed: for a certain electric vehicle model, a large number of users' driving data under different road conditions are collected, and the driving speed mileage ratio and cumulative mileage ratio are counted. According to the driving speed cumulative mileage ratio curve, the driving speed status is accurately divided into low speed (0 - 40km / h), medium speed (40 -70km / h), and high speed (greater than 70km / h) in a ratio of 4:4:2. Within a month, the driving speed of each user's vehicle is classified and counted according to the driving speed status, the mileage under each driving speed status is calculated, and then the mileage ratio corresponding to the driving speed status of the monthly data is calculated according to the formula.

[0055] When calculating the motor torque distribution parameters, m Represents the motor torque state number; j The serial number representing the motor torque state segment; Represents the motor torque state m Corresponding mileage ratio; Represents the motor torque state fragment j The corresponding vehicle speed; Represents the motor torque status in the user's vehicle monthly data m The total number of .

[0056] For example, regarding the distribution of motor torque: for specific models, the torque data of the driving motors of many user vehicles are collected, and the mileage percentage and cumulative mileage percentage of the absolute value of the motor torque are counted. According to the cumulative mileage percentage curve, the motor torque state is clearly divided into small torque (absolute value less than or equal to 20N・m), medium torque (20-70N・m), and large torque (greater than 70N・m) in a ratio of 4:4:2. Within a month, the user vehicle data is classified and counted by motor torque state, the mileage under each motor torque state is calculated, and the mileage percentage corresponding to the motor torque state of the monthly data is calculated.

[0057] Specifically, the driving style distribution is based on clustering algorithms. The driving styles are classified according to vehicle speed, acceleration and other data, and the relevant characteristic parameters are calculated. The driving style is determined through the K-Means clustering algorithm, and the mileage percentage of each driving style is counted.

[0058] In one possible implementation, the driving style distribution parameter is calculated as follows:

[0059]

[0060]

[0061]

[0062]

[0063] Where: is the sequence number of the acceleration point; is the total number of acceleration points; is the mean absolute value of acceleration, m / s 2 ; For the The acceleration of an acceleration point; is the standard deviation of the absolute value of acceleration, m / s 2 ; For the The jerk of each acceleration point, that is, the rate of change of acceleration; is the mean absolute value of jerk, m / s 3 ; is the standard deviation of the absolute value of jerk, m / s 3 ; The acceleration absolute mean, acceleration absolute standard deviation, jerk absolute mean and jerk absolute standard deviation of all equal mileage segments are used to form a characteristic parameter matrix, and the K-Means clustering algorithm is used to perform clustering operation on the characteristic parameter matrix. The cluster center is determined through multiple iterative calculations, and the driving styles are divided into three types: aggressive, general and conservative according to the distance relationship between the equal mileage segments and the cluster center. For each user's vehicle monthly operation data, the equal mileage segments are extracted and their driving style distribution parameters are calculated. The Euclidean distance between the driving style distribution parameters of each equal mileage segment and the three cluster centers is calculated, and the driving style of the equal mileage segments is determined based on the principle of the shortest distance. Finally, the mileage of each driving style is counted to obtain the mileage proportion of the user's vehicle monthly data driving style.

[0064] For example, the vehicle operation data of multiple users are randomly selected, and after removing long-term idling and abnormal data, equal mileage segments are extracted as the basic database for driving style clustering. Among them, equal mileage segments are segments where the vehicle speed starts from 0 to the next 0 and the average vehicle speed is greater than 5 km / h.

[0065] Specifically, the road type distribution is based on the OpenStreetMap open source vector spatial dataset and ArcGIS software to build a GPS range library of typical urban road types, determine the driving road type according to the user's GPS data, and count the mileage proportion of each road type.

[0066] For example, Figure 3 As shown in the figure, taking a typical city (such as Beijing) as an example, relevant data is imported into ArcMap software, and the city boundary and road vector data are obtained by using vector data editing methods. Based on OpenStreetMap data, urban expressways are separated, measurement points are constructed to calculate longitude and latitude, and the expressway GPS database is output. DEM raster data is downloaded from the geospatial data cloud, and terrain data is processed to obtain the mountain range boundary. The suburban area is determined and the boundary line is drawn, measurement points are constructed to obtain longitude and latitude, and a road classification GPS range library is built.

[0067] In one possible implementation, the road type distribution parameter is calculated as follows: The short-trip method is used to divide the user's monthly data into short-trip segments. The type of road traveled is determined based on the GPS range library and the user's data latitude and longitude. The mileage of each road type is counted to obtain the mileage ratio of the user's monthly data road type.

[0068] S4. Perform multiple optimization pairing iterations on the feature parameters of the equal mileage segment data in the initial database to screen out a typical data set, wherein the deviation rate between the feature parameters of the equal mileage segment data in the typical data set and the user vehicle model is within a set range.

[0069] In one possible implementation, the feature parameters of the equal-mileage segment data in the initial database are subjected to multiple optimization pairing iterations to screen out a typical data set, specifically: The initial database n The data of equal mileage segments are represented as X i =( , ,…, ), i =1,2,…, n ,in - Respectively represent the road type mileage ratio, - Indicates the driving style mileage ratio, - Indicates the ratio of full and half load mileage. - Indicates the driving speed-mileage ratio, - Indicates the motor torque-mileage ratio. Set the target point G=(0.268, 0.397, 0.138, 0.179, 0.164, 0.427, 0.409, 0.376, 0.624, 0.393,0.453, 0.154, 0.376, 0.444, 0.180). Traverse and calculate the initial database in X i With the rest of the data X j Euclidean distance between the paired points and the target point G , the calculation formula is:

[0070] where j = 1, 2, …, n , j≠i; ∈G; Record Minimum value of L imin and its corresponding X j , find L imin The minimum value of X i , X j , remove the two from the initial database and pair them to get data P=(P 1 , P 2 , …, P15 ), where P k =(X ik + X jk ) / 2,P k ∈P, k = 1,2, …, 15); Update the calculation of the remaining data of the initial database , L imin , repeat the above pairing and updating process, obtain the pairing results in sequence, until the database pairing is completed; accumulate the paired data in sequence, and calculate the deviation rate between the accumulated data mean and the target point and average deviation rate , the calculation formula is

[0071]

[0072] h Cumulative number of paired data; k =1, 2, …, 15; repeat the above process with the paired data as the new initial database, and take the accumulated data mileage as the target mileage (e.g. 120,000 kilometers) MPE ≤0.05 and max ( PE )≤0.15 is the termination condition for data matching iteration, and finally a typical user data set that meets the user usage characteristics is selected. The specific process is as follows Figure 4 shown.

[0073] S5. Set constraint conditions, and sequentially splice the equal-mileage segment data in the typical data set to obtain spliced ​​online operation data. Sequential splicing avoids large-value load cycles and abnormal changes.

[0074] In an implementable manner, the setting of constraint conditions to sequentially splice the equal-mileage segment data in the typical data set is specifically as follows: For the two equal mileage segments involved in the splicing, the torque characteristics of the user's vehicle are used as an example of data splicing. Assume that the value and slope of the tail of the previous equal mileage segment are T 1 and v 1 , the value and slope of the first part of the next equal mileage segment are T 2 and v 2 , set the splicing constraints as: | T 1 - T 2 |≤10 N·m; · ≥0.

[0075] For any equal mileage segment in the filtered data set, traverse the remaining equal mileage segments according to the splicing constraints, and form a set of equal mileage segments that can be spliced ​​before it. fi , the subsequent equal-mileage segments can be spliced ​​to form a set ri . Select the corresponding set fi The equal-mileage segment with the least segments is taken as the first equal-mileage segment to be spliced. If there are multiple equal-mileage segments, the corresponding set is taken. ri The one with the most equal-mileage segments. In the set of the first equal-mileage segment ri In the dialog box, select the corresponding set ri The one with the most equal mileage segments is used as the second equal mileage segment to be spliced, and so on. When the splicing is interrupted at the nth equal mileage segment, if n is the total number of equal mileage segments in the screening data set, the splicing is completed; otherwise, the numerical restrictions (such as relaxing the numerical difference restrictions to 15N・m) or slope restrictions (such as allowing the slope product to be between -0.5 and 0) after the nth equal mileage segment is appropriately relaxed until the splicing is completed, and the online user data load spectrum after orderly splicing is obtained, that is, the online operation data after splicing. The specific process is as follows Figure 5 shown.

[0076] S6. Convert the spliced ​​online operation data into an online load spectrum, and merge it with the acquired offline load spectrum to obtain the load spectrum of the electric vehicle drive system.

[0077] In one possible implementation, the obtaining of the offline load spectrum is specifically as follows: First, obtain the offline operation data of the user's vehicle under different working conditions.

[0078] For example, based on the user portrait results, a real vehicle road data collection plan is formulated from the aspects of road type and load, and the offline operation data of the user's vehicle under different working conditions is determined.

[0079] Next, the offline operation data is extrapolated to obtain a load cycle distribution probability density function.

[0080] Exemplarily, the rain flow counting method and the kernel density estimation method are used to process the offline operation data to obtain the load cycle distribution probability density function.

[0081] More specifically, the offline operation data collected from the actual vehicle under various working conditions are counted by fatigue analysis software (such as Tecware, etc.) using the rain flow counting method to obtain the from-to type rain flow matrix, which intuitively displays the amplitude and frequency distribution of the load cycle. The kernel density estimation method is used to estimate the distribution of the rain flow matrix, and the Gaussian kernel is selected as the kernel function. The two-dimensional kernel density estimation expression is:

[0082] in, is the total number of load cycles; , For the i from and to values ​​of each load cycle; h is the bandwidth of the kernel density estimate, and the optimal bandwidth is calculated using the rule of thumb, as follows:

[0083] in, d is the dimension of kernel density estimation, which is 2; σ is the standard deviation of the two-dimensional data sample, and the probability density function of the load cycle distribution is obtained.

[0084] Then, according to the probability density function of the load cycle distribution of each working condition and the mileage proportion of each working condition, the composite load data of multiple working conditions are superimposed to obtain the offline load spectrum.

[0085] Specifically, the calculation formula is:

[0086] in, is the target total mileage; , , and Respectively i The total mileage, mileage proportion, total number of load cycles and load cycle probability density function of the various working condition test data are calculated, and finally the load spectrum of the offline real vehicle data is obtained.

[0087] Finally, the load spectra of online user data and offline real vehicle data are superimposed to obtain the full life cycle load spectrum of the vehicle.

[0088] For example, the load spectrum of online user data and the load spectrum of offline real vehicle data are accurately superimposed to obtain the load spectrum of a certain model over the entire life cycle of 240,000 kilometers. The Goodman equation is used to perform average stress correction on the two-dimensional load spectrum to obtain the logarithmic cumulative frequency curves of the online and offline data load spectra. The fatigue damage calculation theory (such as Miner's linear cumulative damage theory) is used to calculate the damage value of each load spectrum, and the characteristics of online and offline data (such as the difference between the numerical range of the online data load spectrum, damage value, and the frequency of each load amplitude and offline data) are compared and analyzed to verify the effectiveness of the method of fusing online and offline, high- and low-frequency data in enhancing the damage equivalence of the full life cycle load spectrum, thereby forming the final fused load spectrum, which provides an accurate basis for the reliability test of the electric vehicle drive system. The full life cycle load spectrum, logarithmic cumulative frequency curve, and load spectrum damage calculation results are as follows: Figure 6 , Figure 7 , Figure 8 shown.

[0089] It has been verified from the damage per unit mileage that the method of fusing online and offline sampling and high- and low-frequency data enhances the damage equivalence of the full life cycle load spectrum.

[0090] Preferably, when new user vehicle data is uploaded to the cloud platform, the monthly characteristic value and monthly mileage of the new data are calculated in time, and the updated calculation formula is used:

[0091] In the formula, , are the eigenvalues ​​before and after the update respectively; is the mileage corresponding to the feature value before updating; a is the number of new data users; For new users i The number of months; , For new users i No. j The overall characteristic values ​​of users are efficiently updated and calculated, and the characteristic values ​​are updated and iterated in real time on a monthly and multi-user basis. Finally, a comprehensive and accurate user usage model is constructed, which describes the vehicle usage characteristics of users in the target market in detail from five dimensions: road type, driving style, load, speed, and torque.

[0092] In another embodiment of the present invention, a computer device is provided, the computer device including a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions in a computer storage medium to implement the corresponding method flow or corresponding function; the processor described in the embodiment of the present invention can be used for the operation of a method for compiling an electric vehicle load spectrum based on strong user association.

[0093] In another embodiment of the present invention, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in a computer device and, of course, extended storage media supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the above-mentioned embodiment regarding a method for compiling an electric vehicle load spectrum based on strong user association.

[0094] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0095] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0096] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0097] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0098] The present invention also provides a computer program product, which is used to execute any of the above-mentioned methods for compiling electric vehicle load spectrum based on strong user association. Since the computer program product provided by the present invention and the above-mentioned method for compiling electric vehicle load spectrum based on strong user association belong to the same inventive concept, the computer program product provided by the present invention has all the advantages of the above-mentioned method for compiling electric vehicle load spectrum based on strong user association, so the beneficial effects of the computer program product provided by the present invention will not be described one by one here.

[0099] In the present invention, the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" etc. mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the described specific features, structures, materials or characteristics may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0100] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for compiling electric vehicle load spectrum based on strong user association, characterized in that: include: Obtain online operation data of user vehicles under different working conditions; Preprocessing the online operation data to obtain preprocessed online operation data, wherein the preprocessing includes resampling the online operation data to improve and unify the sampling frequency of the online operation data; Dividing the preprocessed online operation data into equal mileage segment data, calculating characteristic parameters of the equal mileage segment data, and constructing an initial database using the characteristic parameters of the equal mileage segment data; Perform multiple optimization pairing iterations on the feature parameters of the equal mileage segment data in the initial database to screen out a typical data set, wherein the deviation rate between the feature parameters of the equal mileage segment data in the typical data set and the user vehicle model is within a set range; Setting constraint conditions, sequentially splicing the equal-mileage fragment data in the typical data set, and obtaining spliced ​​online operation data; The spliced ​​online operation data is converted into an online load spectrum and fused with the acquired offline load spectrum to obtain the load spectrum of the electric vehicle drive system.

2. The method for compiling electric vehicle load spectrum based on strong user association according to claim 1 is characterized in that: The characteristic parameters of the equal mileage segment data include load state distribution parameters, driving speed distribution parameters, motor torque distribution parameters, driving style distribution parameters and road type distribution parameters.

3. The method for compiling electric vehicle load spectrum based on strong user association according to claim 2 is characterized in that: The calculation method of the load state distribution parameter, the driving speed distribution parameter or the motor torque distribution parameter is: In the formula, represents the sampling time; Represents the monthly mileage of the user's vehicle; When calculating the load state distribution parameters, m Represents the full load / half load status serial number; j The sequence number representing the full load / half load state segment; Represents full load / half load status m Corresponding mileage ratio; Represents full load / half load status fragment j The corresponding vehicle speed; Represents the full load / half load status of the user's vehicle in the monthly data m Total number of; When calculating the driving speed distribution parameters, m Represents the driving speed state sequence number; j The serial number representing the driving speed state segment; Represents the driving speed status m Corresponding mileage ratio; Represents the driving speed status fragment j The corresponding vehicle speed; Represents the driving speed status of the user's vehicle in the monthly data m Total number of; When calculating the motor torque distribution parameters, m Represents the motor torque state number; j The serial number representing the motor torque state segment; Represents the motor torque state m Corresponding mileage ratio; Represents the motor torque state fragment j The corresponding vehicle speed; Represents the motor torque status in the user's vehicle monthly data m The total number of .

4. The method for compiling electric vehicle load spectrum based on strong user association according to claim 2 is characterized in that: The driving style distribution parameter is calculated as follows: Where: is the sequence number of the acceleration point; is the total number of acceleration points; is the mean absolute value of acceleration; For the The acceleration of an acceleration point; is the standard deviation of the absolute value of acceleration; For the The jerk of each acceleration point, that is, the rate of change of acceleration; is the mean absolute value of jerk; is the standard deviation of the absolute value of jerk; The acceleration absolute mean, acceleration absolute standard deviation, jerk absolute mean and jerk absolute standard deviation of all equal mileage segments are used to form a characteristic parameter matrix, and the K-Means clustering algorithm is used to perform clustering operation on the characteristic parameter matrix. The cluster center is determined through multiple iterative calculations, and the driving styles are divided into three types: aggressive, general and conservative according to the distance relationship between the equal mileage segments and the cluster center. For each user's vehicle monthly operation data, the equal mileage segments are extracted and their driving style distribution parameters are calculated. The Euclidean distance between the driving style distribution parameters of each equal mileage segment and the three cluster centers is calculated, and the driving style of the equal mileage segments is determined based on the principle of the shortest distance. Finally, the mileage of each driving style is counted to obtain the mileage proportion of the user's vehicle monthly data driving style.

5. The method for compiling electric vehicle load spectrum based on strong user association according to claim 2 is characterized in that: The calculation method of the road type distribution parameter is: The short-trip method is used to divide the user's monthly data into short-trip segments. The user's driving road type is determined based on the GPS range library and the user's data latitude and longitude. The mileage of each road type is counted to obtain the mileage ratio of the user's monthly data road type.

6. The method for compiling electric vehicle load spectrum based on strong user association according to claim 1 is characterized in that: The characteristic parameters of the equal-mileage segment data in the initial database are subjected to multiple optimization pairing iterations to screen out a typical data set, specifically: The initial database n The data of equal mileage segments are represented as X i =( , ,…, ), i =1,2,…, n ,in - Respectively represent the road type mileage ratio, - Indicates the driving style mileage ratio, - Indicates the ratio of full and half load mileage. - Indicates the driving speed-mileage ratio, - Indicates the motor torque-mileage ratio; Set the target point G and traverse to calculate the X in the initial database i With the rest of the data X j Euclidean distance between the paired points and the target point G , the calculation formula is: in, j = 1,2,…, n , j ≠ i ; ∈G) Record Minimum value of L imin and its corresponding X j , find L imin The minimum value of X i , X j , remove the two from the initial database and pair them to get data P=(P1, P2, …, P 15 ), where P k =(X ik + X jk ) / 2,P k ∈P, k = 1,2, …, 15); Update the calculation of the remaining data of the initial database , L imin , repeat the above pairing and updating process, obtain the pairing results in sequence, until the database pairing is completed; accumulate the paired data in sequence, and calculate the deviation rate between the accumulated data mean and the target point and average deviation rate , the calculation formula is: h Cumulative number of paired data; k =1, 2, …, 15; repeat the above process with the paired data as the new initial database, and take the accumulated data mileage as the target mileage MPE ≤0.05 and max ( PE )≤0.15 as the termination condition for data matching iteration, and finally screen out a typical data set that meets the user's usage characteristics.

7. The method for compiling electric vehicle load spectrum based on strong user association according to claim 1 is characterized in that: The constraint conditions are set to sequentially splice the equal-mileage segment data in the typical data set, specifically: For the two equal-length segments involved in the splicing, let the value and slope of the tail of the previous equal-length segment be T 1 and v 1, the value and slope of the first part of the next equal mileage segment are T 2 and v 2. Set the splicing constraints as: | T 1- T 2 | and · All are within the set range; For any equal mileage segment in the selected typical data set, traverse the remaining equal mileage segments according to the splicing constraints, and form a set of equal mileage segments that can be spliced ​​before it fi , the subsequent equal-mileage segments can be spliced ​​to form a set ri ; Select the corresponding set fi The equal mileage segment with the least equal mileage segments is taken as the first equal mileage segment to be spliced. If there are multiple equal mileage segments, the corresponding set is taken. ri The one with the most equal-mileage segments; in the set of the first equal-mileage segment ri In the dialog box, select the corresponding set ri The one with the most equal-mileage segments is used as the second equal-mileage segment to be spliced, and so on; when the splicing is interrupted at the nth equal-mileage segment, if n is the total number of equal-mileage segments in the screened data set, the splicing is completed; otherwise, the numerical restrictions or slope restrictions after the splicing of the nth equal-mileage segment are relaxed until the splicing is completed, and the online running data after splicing is obtained.

8. The method for compiling electric vehicle load spectrum based on strong user association according to claim 1 is characterized in that: The obtaining of the offline load spectrum is specifically as follows: Obtain offline operation data of user vehicles under different working conditions through real-vehicle road collection solutions; The offline operation data is extrapolated, firstly rain flow counting is performed on the offline operation data to obtain a rain flow matrix, and then the distribution of the rain flow matrix is ​​estimated by using a kernel density estimation method to obtain a load cycle distribution probability density function; According to the probability density function of the load cycle distribution of each working condition and the mileage proportion of each working condition, the composite load data of multiple working conditions are superimposed to obtain the offline load spectrum.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for compiling an electric vehicle load spectrum based on strong user association as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for compiling an electric vehicle load spectrum based on strong user association as described in any one of claims 1 to 8 is implemented.

Images