An electric drive assembly efficiency evaluation method associated with a user

Abstract

The application particularly relates to an electric drive assembly efficiency evaluation method associated with a user, comprising the following steps: acquiring user actual driving working condition data of various road types; performing data processing and analysis on the user actual driving working condition data of various road types to generate user working condition distribution under full life mileage; determining user commonly used vehicle speed based on the user working condition distribution under full life mileage, and then calculating comprehensive efficiency corresponding to the user commonly used vehicle speed; calculating electric drive assembly high efficiency area utilization rate under full life mileage based on the user working condition distribution under full life mileage; and realizing comprehensive evaluation of electric drive assembly efficiency through the comprehensive efficiency corresponding to the user commonly used vehicle speed and the electric drive assembly high efficiency area utilization rate under full life mileage. The application can realize the correlation between electric drive assembly efficiency and user actual driving working condition, and can accurately reflect the user actual driving working condition, so that the comprehensiveness and accuracy of electric drive assembly efficiency evaluation can be improved.

Description

Technical Field

[0001] This invention relates to the field of electric drive assembly efficiency evaluation technology, and more specifically to a user-related electric drive assembly efficiency evaluation method. Background Technology

[0002] Currently, with the continuous development of society and economy, new energy pure electric vehicles are attracting more and more attention. The electric drive system is one of the core components of pure electric vehicles, consisting of a motor controller, drive motor, and reducer. Its operating efficiency is a key concern for many companies and users, as the efficiency of the electric drive system directly affects the economy of the vehicle.

[0003] In the prior art, efficiency MAP diagrams of electric drive assemblies in drive mode and power generation mode are generally obtained through efficiency bench tests, and the proportion of high efficiency zone is used as an evaluation index of electric drive assembly efficiency.

[0004] However, in actual driving, user actions cause frequent starts / stops, accelerations / decelerations, and power / energy recovery changes in the electric drive system, resulting in highly complex operating conditions. Furthermore, different traffic and road conditions lead to varying percentages of time the electric drive system operates within its high-efficiency zone—problems that efficiency bench tests cannot reflect. In other words, existing methods that rely solely on bench tests to calculate the percentage of the high-efficiency zone as an indicator of electric drive system efficiency fail to consider actual user driving conditions (i.e., user actions, traffic conditions, and road conditions), resulting in poor comprehensiveness and accuracy in assessing electric drive system efficiency.

[0005] Therefore, designing an electric drive system efficiency evaluation method that can be correlated with actual user driving conditions is an urgent technical problem that needs to be solved. Summary of the Invention

[0006] To address the shortcomings of the prior art, the technical problem to be solved by this invention is: how to provide a user-related method for evaluating the efficiency of an electric drive assembly, which can correlate the efficiency of the electric drive assembly with the user's actual driving conditions and accurately reflect the user's actual driving conditions, thereby improving the comprehensiveness and accuracy of the evaluation of the efficiency of the electric drive assembly and providing a new approach to the evaluation of the efficiency of the electric drive assembly.

[0007] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0008] A user-related method for evaluating the efficiency of an electric drive assembly, comprising:

[0009] S1: Obtain actual driving condition data of users for various road types;

[0010] S2: Process and analyze the actual driving conditions of users on various road types to generate the user condition distribution over the entire lifespan mileage;

[0011] S3: Determine the user's commonly used vehicle speed based on the user's operating condition distribution over the entire life cycle mileage, and then calculate the comprehensive efficiency corresponding to the user's commonly used vehicle speed;

[0012] S4: Calculate the high-efficiency zone utilization rate of the electric drive assembly based on the user operating condition distribution over the entire lifespan;

[0013] S5: A comprehensive evaluation of the efficiency of the electric drive system is achieved by considering the overall efficiency at the user's usual vehicle speed and the utilization rate of the high-efficiency zone of the electric drive system over its entire lifespan.

[0014] Preferably, in step S1, user actual driving condition data within a certain mileage is collected on various road types according to the corresponding proportion, and the consistency of sampling start time, sampling end time and sampling frequency is ensured when collecting user actual driving condition data.

[0015] Road types include urban roads, expressways, general roads, and rough roads;

[0016] User's actual driving condition data includes the torque, speed, and operating mode of the electric drive assembly, including drive mode and generator mode.

[0017] Preferably, in step S2, the user operating condition distribution over the entire lifespan is generated through the following steps:

[0018] S201: Perform singularity removal and signal offset correction on the actual driving condition data of users for various road types to obtain user condition samples for various road types.

[0019] S202: Extract torque and speed data from user operating condition samples of various road types, and generate speed-torque frequency statistical matrices for various road types through frequency statistical analysis;

[0020] S203: The probability density distribution of the speed-torque frequency statistics matrix of various road types is fitted by a non-parametric two-dimensional kernel density estimation algorithm to generate mathematical models of probability density distribution for various road types.

[0021] S204: Extrapolate the speed-torque frequency statistical matrix of various road types by combining the Monte Carlo simulation method with the corresponding probability density distribution mathematical model, and then superimpose the extrapolated speed-torque frequency statistical matrices of various road types to generate the user operating condition distribution under the whole life mileage.

[0022] Preferably, in step S202, torque data {T} is extracted from the user's operating condition sample. 驱动1 T 驱动2 ...T 驱动i} and rotational speed data {n 驱动1 n 驱动2 ...n 驱动i Construct the corresponding speed-torque dataset M = {(T 驱动1 ,n 驱动1 (T) 驱动2 ,n 驱动2 )……(T 驱动i ,n 驱动i Then, frequency statistical analysis is performed on the speed-torque dataset M to generate the corresponding speed-torque frequency statistical matrix.

[0023] Preferably, in step S203, the probability density distribution mathematical model is generated through the following steps:

[0024] 1) Calculate the kernel density function and initial bandwidth using the following formula:

[0025]

[0026] h = 2.4σn -1 / 5 ;

[0027] σ=min(σ x ,σ y );

[0028] In the formula: K represents the kernel density function; x and y represent the speed and torque in the speed-torque frequency statistics matrix, respectively; h represents the initial bandwidth; n represents the size of the user operating condition sample; σ represents the standard deviation of the user operating condition sample. x σ represents the standard deviation of rotational speed. y The standard deviation of torque;

[0029] 2) Calculate the adaptive bandwidth using the following formula:

[0030]

[0031]

[0032] In the formula: P(x) represents the adaptive bandwidth; h represents the initial bandwidth; i ,y i ) represents the speed-torque operating point (x) in the speed-torque frequency statistics matrix. i ,y i The ratio of the number of occurrences to the total number of speed-torque operating points; α represents the sensitivity coefficient, with a value of 0.5; n represents the capacity of the user operating condition sample.

[0033] 3) Calculate the mathematical model of the probability density distribution using the following formula:

[0034]

[0035] In the formula: f(x,y) represents the mathematical model of the probability density distribution of the speed-torque frequency statistical matrix; represents the adaptive bandwidth; K represents the kernel density function; x and y represent the speed and torque in the speed-torque frequency statistics matrix, respectively.

[0036] Preferably, in step S204, the speed-torque operating points are randomly placed according to the extrapolation coefficients of the corresponding mileage using the Monte Carlo simulation method, thereby realizing the bidirectional extrapolation of the speed-torque frequency statistical matrix in terms of amplitude and frequency; the extrapolation coefficients are obtained by dividing the target mileage value of each road type by the measured mileage.

[0037] Preferably, in step S3, the overall efficiency corresponding to the user's commonly used vehicle speed is calculated through the following steps:

[0038] S301: An efficiency prediction model based on a BP neural network, with torque and speed as inputs and efficiency as output;

[0039] S302: Predict the efficiency at various speed-torque operating points using an efficiency prediction model;

[0040] η (i,k) =G(i,k);

[0041] In the formula: η (i,k) G represents the efficiency corresponding to the operating point with torque i and speed k, which is the output of the efficiency prediction model; G represents the efficiency prediction model; torque i and speed k are the inputs of the efficiency prediction model.

[0042] S303: The frequency of vehicle speed occurrence is calculated by combining the user's working condition distribution over the entire lifespan mileage with the following formula, and then the vehicle speed with the highest frequency of occurrence is selected as the user's commonly used vehicle speed.

[0043]

[0044] In the formula: Indicates vehicle speed V k The frequency percentage of occurrence; A(k,j) represents the frequency of occurrence of the speed-torque operating point with torque j and speed k in the user operating condition distribution under the whole life mileage; Q represents the total number of speed-torque operating points in the user operating condition distribution under the whole life mileage.

[0045] Among them, vehicle speed V k This indicates the vehicle speed when the rotational speed of the electric drive assembly is k.

[0046]

[0047] In the formula: r represents the wheel radius; i g Indicates the reduction ratio of the electric drive assembly;

[0048] S304: The comprehensive efficiency corresponding to the user's commonly used vehicle speed is calculated by combining the user's operating condition distribution under the whole life mileage and the efficiency corresponding to each speed-torque operating point with the following formula.

[0049]

[0050] In the formula: The user's commonly used vehicle speed is represented by vehicle speed V. k Overall efficiency at that time; The torque is represented by T. max Efficiency at rotational speed k; A(k,T) max The torque T represents the torque distribution under the user's operating conditions over the entire lifespan. max The frequency of the speed-torque operating point at speed k; Q represents the total number of speed-torque operating points in the user operating condition distribution over the entire lifespan.

[0051] Preferably, in step S301, the efficiency prediction model is constructed through the following steps:

[0052] S3011: Determine the BP neural network structure and training algorithm for the efficiency prediction model;

[0053] S3012: Based on the distribution of user operating conditions over the entire lifespan, determine the distribution principle of speed-torque operating points for testing, and then test the efficiency of each speed-torque operating point through an electric drive assembly efficiency bench test.

[0054] S3013: Select the corresponding speed-torque operating point and efficiency to construct a training dataset with speed and torque as input and efficiency as output, and divide the training dataset into training samples and test samples;

[0055] S3014: The neural network is trained and tested by combining training samples and test samples with the corresponding training algorithm to obtain the corresponding efficiency prediction model.

[0056] Preferably, in step S3012, the efficiency at the speed-torque operating point is calculated using the following formula:

[0057] 1) Efficiency calculation formula in drive mode:

[0058]

[0059] In the formula: η1 represents the efficiency of the electric drive assembly in drive mode; n1 represents the speed of the left dynamometer; T1 represents the torque of the left dynamometer; n2 represents the speed of the right dynamometer; T2 represents the torque of the right dynamometer; P i Indicates the input power of the electric drive assembly;

[0060] 2) Efficiency calculation formula under power generation mode:

[0061]

[0062] In the formula: η2 represents the efficiency of the electric drive assembly in generator mode; n1 represents the speed of the left dynamometer; T1 represents the torque of the left dynamometer; n2 represents the speed of the right dynamometer; T2 represents the torque of the right dynamometer; P i This indicates the input power of the electric drive assembly.

[0063] Preferably, in step S4, the high-efficiency zone utilization rate of the electric drive assembly over its entire lifespan is calculated using the following formula:

[0064]

[0065] In the formula: D represents the high-efficiency zone utilization rate of the electric drive assembly under the whole life mileage; W represents the number of efficiencies greater than the expected efficiency M% among all torque-speed operating points in the user operating condition distribution under the whole life mileage; Q represents the total number of speed-torque operating points in the user operating condition distribution under the whole life mileage.

[0066] Compared with existing technologies, the user-related electric drive assembly efficiency evaluation method in this invention has the following advantages:

[0067] This invention generates a user operating condition distribution over the entire lifespan by acquiring actual user driving condition data for various road types. Then, it calculates the comprehensive efficiency corresponding to the user's commonly used vehicle speed and the utilization rate of the high-efficiency zone of the electric drive assembly over the entire lifespan based on the user operating condition distribution over the entire lifespan, thereby achieving a comprehensive evaluation of the efficiency of the electric drive assembly. On the one hand, this invention calculates the overall efficiency and the high-efficiency zone utilization rate of the electric drive system by analyzing the user's operating condition distribution over the entire lifespan. This allows for a correlation between the efficiency of the electric drive system and the user's actual driving conditions. Furthermore, the method of generating the user's operating condition distribution over the entire lifespan using actual driving condition data accurately reflects the user's actual driving conditions, fully considering the impact of the user's driving operations, traffic conditions, and road conditions on the electric drive system's operating conditions, thereby improving the accuracy of the electric drive system efficiency assessment. On the other hand, this invention achieves a comprehensive evaluation of the electric drive system efficiency by analyzing the overall efficiency corresponding to the user's commonly used vehicle speed and the high-efficiency zone utilization rate of the electric drive system over the entire lifespan. The high-efficiency zone utilization rate can reflect the operating status of the electric drive system within its high-efficiency zone to a certain extent, providing guidance for optimizing the efficiency characteristics of the electric drive system. Moreover, both the overall efficiency and the high-efficiency zone utilization rate are correlated with the user's actual driving conditions. Compared to using efficiency bench tests to calculate the proportion of the high-efficiency zone as an evaluation indicator for the electric drive system efficiency, this method provides a more comprehensive and accurate assessment of the electric drive system efficiency, thus improving the overall comprehensiveness of the electric drive system efficiency evaluation.

[0068] Therefore, the present invention can correlate the efficiency of the electric drive assembly with the user's actual driving conditions, and can accurately reflect the user's actual driving conditions, thereby improving the comprehensiveness and accuracy of the electric drive assembly efficiency assessment, and providing a new approach to the assessment of electric drive assembly efficiency. Attached Figure Description

[0069] To make the objectives, technical solutions, and advantages of the invention clearer, the invention will now be described in further detail with reference to the accompanying drawings, wherein:

[0070] Figure 1 A logic block diagram for a user-related method for evaluating the efficiency of electric drive assemblies;

[0071] Figure 2 This is a BP neural network topology;

[0072] Figure 3 This is a distribution diagram of test points for the efficiency bench test of the electric drive assembly.

[0073] Figure 4 and Figure 5 Efficiency MAPs for drive mode and power generation mode;

[0074] Figure 6The error convergence curve of the BP neural network;

[0075] Figure 7 The fitting effect between the actual values ​​and predicted values ​​in the training set;

[0076] Figure 8 To test the fit between the actual and predicted values ​​in the test set;

[0077] Figure 9 The relative error between predicted and actual values ​​in the training set

[0078] Figure 10 This represents the relative error between the predicted and actual values ​​for the test set. Detailed Implementation

[0079] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but only to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0080] It should be noted that similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the figures, or the orientation or positional relationship commonly used when the product is in use. They are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third," etc., are only used to distinguish descriptions and should not be construed as indicating or implying relative importance. In addition, the terms "horizontal," "vertical," etc., do not mean that the component is required to be absolutely horizontal or suspended, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted. In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0081] The following detailed explanation illustrates the specific implementation methods:

[0082] Example:

[0083] This embodiment discloses a user-associated method for evaluating the efficiency of an electric drive assembly.

[0084] like Figure 1 As shown, the user-related electric drive assembly efficiency evaluation method includes:

[0085] S1: Obtain actual driving condition data of users for various road types;

[0086] S2: Process and analyze the actual driving conditions of users on various road types to generate the user condition distribution over the entire lifespan mileage;

[0087] S3: Determine the user's commonly used vehicle speed based on the user's operating condition distribution over the entire life cycle mileage, and then calculate the comprehensive efficiency corresponding to the user's commonly used vehicle speed;

[0088] S4: Calculate the high-efficiency zone utilization rate of the electric drive assembly based on the user operating condition distribution over the entire lifespan;

[0089] S5: A comprehensive evaluation of the efficiency of the electric drive system is achieved by considering the overall efficiency at the user's usual vehicle speed and the utilization rate of the high-efficiency zone of the electric drive system over its entire lifespan.

[0090] In this embodiment, the overall efficiency at the user's commonly used vehicle speed and the high-efficiency zone utilization rate of the electric drive assembly over its entire lifespan are used as two evaluation indicators to comprehensively assess the efficiency of the electric drive assembly (e.g., each indicator accounts for 50% of the importance). The specific efficiency evaluation methods are existing mature technologies and will not be elaborated here. Furthermore, since the electric drive assembly has two operating modes—drive mode and power generation mode—it is necessary to calculate the corresponding overall efficiency at the user's commonly used vehicle speed and the high-efficiency zone utilization rate of the electric drive assembly over its entire lifespan separately for each operating mode to achieve the efficiency evaluation of the electric drive assembly under the corresponding operating mode.

[0091] This invention generates a user operating condition distribution over the entire lifespan by acquiring actual user driving condition data for various road types. Then, it calculates the comprehensive efficiency corresponding to the user's commonly used vehicle speed and the utilization rate of the high-efficiency zone of the electric drive assembly over the entire lifespan based on the user operating condition distribution over the entire lifespan, thereby achieving a comprehensive evaluation of the efficiency of the electric drive assembly. On the one hand, this invention calculates the overall efficiency and the high-efficiency zone utilization rate of the electric drive system by analyzing the user's operating condition distribution over the entire lifespan. This allows for a correlation between the efficiency of the electric drive system and the user's actual driving conditions. Furthermore, the method of generating the user's operating condition distribution over the entire lifespan using actual driving condition data accurately reflects the user's actual driving conditions, fully considering the impact of the user's driving operations, traffic conditions, and road conditions on the electric drive system's operating conditions, thereby improving the accuracy of the electric drive system efficiency assessment. On the other hand, this invention achieves a comprehensive evaluation of the electric drive system efficiency by analyzing the overall efficiency corresponding to the user's commonly used vehicle speed and the high-efficiency zone utilization rate of the electric drive system over the entire lifespan. The high-efficiency zone utilization rate can reflect the operating status of the electric drive system within its high-efficiency zone to a certain extent, providing guidance for optimizing the efficiency characteristics of the electric drive system. Moreover, both the overall efficiency and the high-efficiency zone utilization rate are correlated with the user's actual driving conditions. Compared to using efficiency bench tests to calculate the proportion of the high-efficiency zone as an evaluation indicator for the electric drive system efficiency, this method provides a more comprehensive and accurate assessment of the electric drive system efficiency, thus improving the overall comprehensiveness of the electric drive system efficiency evaluation.

[0092] Therefore, the present invention can correlate the efficiency of the electric drive assembly with the user's actual driving conditions, and can accurately reflect the user's actual driving conditions, thereby improving the comprehensiveness and accuracy of the electric drive assembly efficiency assessment, and providing a new approach to the assessment of electric drive assembly efficiency.

[0093] In the specific implementation process, actual driving condition data of users is collected through devices such as GPS, eDAQ data acquisition devices, and data acquisition lines. As shown in Table 1, based on the driving condition survey results of 90% of users provided by a certain company, the road types of the collected roads are divided into urban roads, expressways, general roads, and severe roads. Among them, general roads are a general term for national highways, provincial highways, and rural roads. Actual driving conditions of users within a certain mileage are collected according to the corresponding proportion. The actual driving conditions of users include electric drive assembly torque (N·m), electric drive assembly speed (r / min), and electric drive assembly operating mode; among which, the electric drive assembly operating mode includes drive mode and generator mode.

[0094] The electric drive assembly torque (N·m), electric drive assembly speed (r / min), and electric drive assembly operating mode can be obtained by reading the vehicle's CAN signal. The COM board provided by eDAQ supports CAN bus acquisition. The vehicle's CAN bus OBD interface has corresponding pins that connect to the data acquisition lines. By connecting the lines according to the corresponding relationships and finally importing the appropriate DBC file into the data acquisition system, the reading and acquisition of the vehicle's CAN signal can be achieved. To ensure that the acquired data corresponds one-to-one, the consistency of the sampling start time, sampling end time, and sampling frequency must be guaranteed when acquiring this operating condition information. Finally, based on the acquired motor operating mode, the user's operating condition is further divided into drive mode and energy recovery mode.

[0095] Table 1 User Road Data Collection Ratio

[0096]

[0097] The present invention can effectively obtain user actual driving condition data for various road types through the above-described method, thereby realizing the correlation between the efficiency of the electric drive assembly and the user's actual driving conditions. It can also fully consider the impact of the user's driving operation, traffic conditions and road conditions on the operating conditions of the electric drive assembly during actual driving, thereby further improving the accuracy of the electric drive assembly efficiency assessment.

[0098] In the specific implementation process, the user operating condition distribution over the entire life cycle mileage is generated through the following steps:

[0099] S201: Singular value removal and signal offset correction are performed on the actual driving condition data of users for various road types to obtain user condition samples for various road types; since the decimals in torque and speed have a negligible impact on the efficiency value of electric drive assembly, the present invention rounds the decimals in the data and finally retains only the integers.

[0100] S202: Extract torque and speed data from user operating condition samples of various road types, and generate speed-torque frequency statistical matrices for various road types through frequency statistical analysis;

[0101] In this embodiment, taking the urban road driving mode as an example, the collected user operating condition samples include torque data {T}. 驱动1 T 驱动2 ...T 驱动i} and rotational speed data {n 驱动1 n 驱动2 ...n 驱动i}, T 驱动i n 驱动i These represent the torque and speed of the i-th electric drive assembly over time.

[0102] Construct a speed-torque dataset M = {(T 驱动1 ,n 驱动1 (T) 驱动2 ,n 驱动2 )……(T 驱动i ,n 驱动i The torque and speed in parentheses correspond one-to-one over time.

[0103] Frequency statistical analysis was performed on the speed-torque dataset M, and the speed-torque frequency statistical matrix shown in Table 2 was generated.

[0104] Where T min n min T max n max These represent the minimum torque, minimum speed, maximum torque, and maximum speed of the electric drive assembly, respectively; A(k,j) represents the frequency of the operating point with speed k and torque j in the dataset M.

[0105] Table 2. Schematic diagram of speed-torque frequency statistics matrix

[0106]

[0107] S203: The probability density distribution of the speed-torque frequency statistics matrix of various road types is fitted by a non-parametric two-dimensional kernel density estimation algorithm to generate mathematical models of probability density distribution for various road types.

[0108] In this embodiment, taking the urban road driving mode as an example, the speed-torque distribution in the user's driving condition is highly random due to its relationship with the user's actual operation, traffic conditions, road conditions, and other factors. Therefore, it is difficult to describe it using a probability density function of a specific distribution. Nonparametric two-dimensional kernel density estimation, however, does not require the samples to follow a specific distribution. Therefore, this invention uses nonparametric two-dimensional kernel density estimation to fit the probability density distribution of speed-torque in the user's driving condition. The steps of nonparametric two-dimensional kernel density estimation mainly consist of the following three parts:

[0109] 1) Calculate the kernel density function and initial bandwidth using the following formula:

[0110]

[0111] h = 2.4σn -1 / 5 ;

[0112] σ=min(σ x ,σ y );

[0113] In the formula: K represents the kernel density function; x and y represent the speed and torque in the speed-torque frequency statistics matrix, respectively; h represents the initial bandwidth; n represents the size of the user operating condition sample; σ represents the standard deviation of the user operating condition sample. x σ represents the standard deviation of rotational speed. y The standard deviation of torque;

[0114] 2) Calculate the adaptive bandwidth using the following formula:

[0115]

[0116]

[0117] In the formula: P(x) represents the adaptive bandwidth; h represents the initial bandwidth; i ,y i ) represents the speed-torque operating point (x) in the speed-torque frequency statistics matrix. i ,y i The ratio of the number of occurrences to the total number of speed-torque operating points; α represents the sensitivity coefficient, with a value of 0.5; n represents the capacity of the user operating condition sample.

[0118] 3) Calculate the mathematical model of the probability density distribution using the following formula:

[0119]

[0120] In the formula: f(x,y) represents the mathematical model of the probability density distribution of the speed-torque frequency statistical matrix; represents the adaptive bandwidth; K represents the kernel density function; x and y represent the speed and torque in the speed-torque frequency statistics matrix, respectively.

[0121] S204: Extrapolate the speed-torque frequency statistical matrix of various road types by combining the Monte Carlo simulation method with the corresponding probability density distribution mathematical model, and then superimpose the extrapolated speed-torque frequency statistical matrices of various road types to generate the user operating condition distribution under the whole life mileage.

[0122] In this embodiment, the speed-torque operating points are randomly placed according to the extrapolation coefficient of the corresponding mileage using the Monte Carlo simulation method, thereby realizing the bidirectional extrapolation of the speed-torque frequency statistical matrix in terms of amplitude and frequency, which can enrich the distribution area of ​​user operating conditions; the extrapolation coefficient is obtained by dividing the target mileage value of each road type by the measured mileage.

[0123] This invention obtains more accurate user operating condition samples by performing singular value removal and signal offset correction on user's actual driving condition data; it better describes the distribution of user operating conditions by fitting the probability density distribution of the speed-torque frequency statistical matrix for various road types using a nonparametric two-dimensional kernel density estimation algorithm; and it enriches the distribution area of ​​user operating conditions by extrapolating the speed-torque frequency statistical matrix for various road types using Monte Carlo simulation. In other words, this invention can accurately reflect the user's actual driving conditions, and fully consider the impact of the user's driving operation, traffic conditions, and road conditions on the operating conditions of the electric drive assembly, thereby further improving the accuracy of electric drive assembly efficiency assessment.

[0124] In the specific implementation process, the overall efficiency corresponding to the user's commonly used vehicle speed is calculated through the following steps:

[0125] S301: An efficiency prediction model based on a BP neural network, with torque and speed as inputs and efficiency as output;

[0126] In this embodiment, the BP neural network has a very strong nonlinear mapping capability, and the efficiency of the electric drive assembly has a nonlinear mapping relationship with torque and speed. Therefore, this step uses the BP neural network to establish an efficiency prediction model related to torque and speed based on measured bench efficiency data.

[0127] S302: Predict the efficiency at various speed-torque operating points using an efficiency prediction model;

[0128] η (i,k) =G(i,k);

[0129] In the formula: η (i,k)The output of the efficiency prediction model is represented by the torque i and the speed k; G represents the efficiency prediction model; and the torque i and speed k are the inputs of the efficiency prediction model.

[0130] S303: The frequency of vehicle speed occurrence is calculated by combining the user's working condition distribution over the entire lifespan mileage with the following formula, and then the vehicle speed with the highest frequency of occurrence is selected as the user's commonly used vehicle speed.

[0131]

[0132] In the formula: Indicates vehicle speed V k The frequency percentage of occurrence; A(k,j) represents the frequency of occurrence of the speed-torque operating point with torque j and speed k in the user operating condition distribution under the whole life mileage; Q represents the total number of speed-torque operating points in the user operating condition distribution under the whole life mileage.

[0133] Among them, vehicle speed V k This indicates the vehicle speed when the rotational speed of the electric drive assembly is k.

[0134]

[0135] In the formula: r represents the wheel radius; i g Indicates the reduction ratio of the electric drive assembly;

[0136] S304: The comprehensive efficiency corresponding to the user's commonly used vehicle speed is calculated by combining the user's operating condition distribution under the whole life mileage and the efficiency corresponding to each speed-torque operating point with the following formula.

[0137]

[0138] In the formula: The user's commonly used vehicle speed is represented by vehicle speed V. k Overall efficiency at that time; The torque is represented by T. max Efficiency at rotational speed k; A(k,T) max The torque T represents the torque distribution under the user's operating conditions over the entire lifespan. max The frequency of the speed-torque operating point at speed k; Q represents the total number of speed-torque operating points in the user operating condition distribution over the entire lifespan.

[0139] This invention predicts efficiency at various speed-torque operating points using an efficiency prediction model. It leverages the strong nonlinear mapping capability of BP neural networks to achieve a nonlinear mapping between efficiency, torque, and speed. By calculating the frequency of vehicle speed occurrences based on the user's operating condition distribution over the entire lifespan, and then combining this with the efficiency corresponding to each speed-torque operating point, it calculates the comprehensive efficiency corresponding to the user's commonly used vehicle speed. This allows for the calculation of comprehensive efficiency based on the user's operating condition distribution over the entire lifespan, thereby establishing a correlation between the efficiency of the electric drive system and the user's actual driving conditions, and further improving the accuracy of electric drive system efficiency assessment.

[0140] The efficiency prediction model is constructed using the following steps:

[0141] S3011: Determine the BP neural network structure and training algorithm for the efficiency prediction model;

[0142] This embodiment specifically includes:

[0143] 1) Determine the number of network layers

[0144] like Figure 2 As shown, the basic structure of a BP neural network consists of one output layer, one input layer, and one or more hidden layers. The number of hidden layers needs to be determined. More hidden layers result in higher prediction accuracy, but also increase training time and computational requirements. Generally, a two-hidden-layer structure is sufficient for modeling most nonlinear mapping problems. Considering all factors, this embodiment uses a two-hidden-layer structure.

[0145] 2) Determine the number of input and output layer nodes.

[0146] The number of nodes in the input and output layers is determined by the dimensions of the input and output vectors, respectively. Based on the efficiency testing principles of the electric drive assembly, the number of nodes in the input layer is determined to be 2, and the number of nodes in the output layer is determined to be 1.

[0147] 3) Determining the number of hidden layer nodes

[0148] The number of nodes in a hidden layer affects the prediction accuracy of a neural network. Generally, more hidden layer nodes result in higher training accuracy. However, more hidden layer nodes are not always better; excessive hidden layer nodes can lead to excessively long training times and overfitting. The number of hidden layer nodes can be calculated using the following empirical formula.

[0149]

[0150] In the formula: n and m represent the number of nodes in the input layer and the output layer, respectively. a is generally an integer between [0, 10].

[0151] 4) Selection of excitation function

[0152] The activation function for the hidden layers uses the sigmoid function, as shown in the formula below. The activation function for the output layer generally uses a linear function, which can approximate any rational number.

[0153]

[0154] 5) Determining the training algorithm

[0155] Different backpropagation (BP) neural network training algorithms have different weight and threshold update principles. Currently, there are 13 commonly used BP neural network training algorithms. This embodiment uses trainlm (LM (Levenberg-Marquardt)), which has the advantages of high training accuracy and relatively fast training speed.

[0156] S3012: Based on the distribution of user operating conditions over the entire lifespan, determine the distribution principle of speed-torque operating points for testing, and then test the efficiency of each speed-torque operating point through an electric drive assembly efficiency bench test.

[0157] like Figure 3 As shown, the efficiency bench test of the electric drive assembly typically involves determining test operating points at equal intervals and then testing the efficiency at each operating point. In this embodiment, the number of test operating points in areas where the electric drive assembly frequently occurs at speed-torque operating points is appropriately increased based on the user's operating condition distribution over the entire lifespan, in order to increase the efficiency prediction accuracy of the subsequent neural network model under common operating conditions.

[0158] In this embodiment, the efficiency at the speed-torque operating point is calculated using the following formula:

[0159] 1) Efficiency calculation formula in drive mode:

[0160]

[0161] In the formula: η1 represents the efficiency of the electric drive assembly in drive mode; n1 represents the speed of the left dynamometer (r / min); T1 represents the torque of the left dynamometer (N·m); n2 represents the speed of the right dynamometer (r / min); T2 represents the torque of the right dynamometer (N·m); P i This indicates the input power (kW) of the electric drive assembly.

[0162] 2) Efficiency calculation formula under power generation mode:

[0163]

[0164] In the formula: η2 represents the efficiency of the electric drive assembly in generator mode; n1 represents the speed of the left dynamometer (r / min); T1 represents the torque of the left dynamometer (N·m); n2 represents the speed of the right dynamometer (r / min); T2 represents the torque of the right dynamometer (N·m); P i This indicates the input power (kW) of the electric drive assembly.

[0165] After the experiment was completed, data processing was used to obtain the following results: Figure 4 and Figure 5 The diagram shows the efficiency MAP of the electric drive assembly in drive mode and generator mode.

[0166] S3013: Select the corresponding speed-torque operating point and efficiency to construct a training dataset with speed and torque as input and efficiency as output, and divide the training dataset into training samples and test samples;

[0167] In this embodiment, the efficiency of each operating point under the drive mode is defined as η. D1 η D2 ...η Di The corresponding torque and speed are w1 = (T D1 n D1 w2 = (T) D2 n D2 )……w i =(T Di n Di ), where η Di T Di n Di Let η represent the efficiency, torque, and speed at the i-th operating point, respectively; similarly, define the efficiency at each operating point under the power generation mode as η. B1 η B2 ...η Bi The corresponding torque and speed are h1 = (T B1 n B1 h2=(T) B2 n B2 )……h i =(T Bi n Bi ), where η Bi T Bi n Bi These represent the efficiency, torque, and speed at the k-th operating point, respectively.

[0168] Taking the driving mode as an example, the training dataset for the efficiency prediction model is as follows:

[0169] {η D1 ,w1}、{η D2 ,w2}……{η Dj ,w Dj}……{ηDq ,w Dq};w Dj =(T Dj n Dj ) represents the j-th sample input vector, and η Dj Let q be the output vector of the j-th sample, q be the number of training samples, and the remaining iq samples be used as test samples.

[0170] S3014: The neural network is trained and tested by combining training samples and test samples with the corresponding training algorithm to obtain the corresponding efficiency prediction model.

[0171] In this embodiment, 9 / 10 of the data is used as training samples and 1 / 10 of the data is used as test samples.

[0172] Specifically, programming is performed in MATLAB to continuously train and test the BP neural network, such as... Figure 6 The process continues until the predicted error is less than the set error. Finally, the trained network is saved, and the efficiency prediction model in the driving mode is obtained. The same method can be used to obtain the efficiency prediction model in the power generation mode. The efficiency prediction model is denoted in the form of the following formula.

[0173] η = G(T, n).

[0174] Figure 7 and Figure 8 These represent the fitting effect between the actual efficiency values ​​and the efficiency values ​​predicted by the BP neural network in the training and test sets, respectively. Figure 9 The error represents the relative error between the actual efficiency value of the training set and the predicted efficiency value of the BP neural network. As can be seen from the error graph, the maximum relative error does not exceed 0.8%, and most of the relative errors are concentrated below 0.2%. Figure 10 The error plot shows the relative error between the actual efficiency value of the test set and the predicted efficiency value of the BP neural network. As can be seen from the error plot, all relative errors are below 0.2%.

[0175] As can be seen from the fitting effect and relative error change curves of the test set and training set, the electric drive assembly efficiency prediction model established by this invention has very high prediction accuracy. It can accurately reflect the mapping relationship between the electric drive assembly efficiency and its torque and speed, and can quickly and accurately predict the efficiency under actual driving conditions.

[0176] This invention determines the speed-torque operating point for testing by analyzing the user operating condition distribution over the entire lifespan of the vehicle. Then, it tests the efficiency of each speed-torque operating point through an electric drive assembly efficiency bench test and constructs a corresponding training dataset. This enables accurate and effective training of the efficiency prediction model, and further utilizes the strong nonlinear mapping capability of the BP neural network to achieve a nonlinear mapping between efficiency and torque and speed.

[0177] In practice, the utilization rate of the high-efficiency zone of the electric drive assembly over its entire lifespan is calculated using the following formula:

[0178]

[0179] In the formula: D represents the high-efficiency zone utilization rate of the electric drive assembly under the whole life mileage; W represents the number of efficiencies greater than the expected efficiency M% among all torque-speed operating points in the user operating condition distribution under the whole life mileage; Q represents the total number of speed-torque operating points in the user operating condition distribution under the whole life mileage.

[0180] This invention calculates the high-efficiency zone utilization rate of the electric drive assembly by analyzing the user's operating conditions over the entire lifespan. This enables the correlation between the efficiency of the electric drive assembly and the user's actual driving conditions. Furthermore, the correlation between the high-efficiency zone utilization rate of the electric drive assembly and the user's actual driving conditions is more comprehensive and accurate than the method of using efficiency bench tests to calculate the proportion of the high-efficiency zone as an evaluation indicator for the efficiency of the electric drive assembly.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the technical solutions. Those skilled in the art should understand that any modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. A method for evaluating the efficiency of an electric drive assembly associated with a user, the method comprising: Comprise: S1: obtaining user actual driving condition data of various road types; S2: data processing and analysis of user actual driving condition data of various road types, generating user condition distribution under full life mileage; S3: determining user common speed based on user condition distribution under full life mileage, and then calculating comprehensive efficiency corresponding to user common speed; In step S3, the comprehensive efficiency corresponding to the user common speed is calculated by the following steps: S301: constructing an efficiency prediction model with torque and speed as input and efficiency as output based on BP neural network; S302: predicting the efficiency of each speed-torque condition point through the efficiency prediction model; η (i,k) = G(i, k); wherein: η (i,k) represents the efficiency corresponding to the operating point of torque i and speed k, i.e. the output of the efficiency prediction model; G represents the efficiency prediction model; torque i and speed k are the inputs of the efficiency prediction model; S303: calculating the frequency ratio of vehicle speed through the following formula combined with the user condition distribution under full life mileage, and then selecting the speed with the largest frequency ratio as the user common speed; In the formula: represents the vehicle speed V k ; A(k,j) represents the frequency of occurrence of the speed-torque condition point with torque j and speed k in the user working condition distribution under the total life mileage; Q represents the total number of speed-torque condition points in the user working condition distribution under the total life mileage; wherein the vehicle speed V k represents the vehicle speed when the rotational speed of the electric drive assembly is k. where: r represents the wheel radius; i g represents the electric drive assembly reduction ratio; S304: calculating the comprehensive efficiency corresponding to the user common speed through the following formula combined with the user condition distribution under full life mileage and the efficiency corresponding to each speed-torque condition point; In the formula: represents the overall efficiency when the user's common vehicle speed is vehicle speed V k ; represents the efficiency when the torque is T max and the rotational speed is k; A(k, T max ) represents the frequency of occurrence of the rotational speed-torque operating point when the torque is T max and the rotational speed is k in the user's operating condition distribution under the total life mileage; Q represents the total number of rotational speed-torque operating points in the user's operating condition distribution under the total life mileage; S4: calculating the high-efficiency area utilization rate of the electric drive assembly under full life mileage based on the user condition distribution under full life mileage; S5: realizing comprehensive evaluation of electric drive assembly efficiency through comprehensive efficiency corresponding to user common speed and high-efficiency area utilization rate of electric drive assembly under full life mileage.

2. The method of claim 1, wherein: In step S1, user actual driving condition data within a certain mileage is collected on various road types in a corresponding proportion, and the consistency of sampling start time, sampling end time and sampling frequency is ensured when collecting user actual driving condition data; The road types include urban roads, highways, general roads and bad roads; The user actual driving condition data includes torque, speed and working mode of the electric drive assembly, and the working mode includes driving mode and power generation mode.

3. The method of claim 1, wherein the method further comprises: In step S2, the user condition distribution under full life mileage is generated by the following steps: S201: singular value rejection and signal offset correction are performed on the user actual driving condition data of various road types to obtain user condition samples of various road types; S202: torque data and speed data in the user condition samples of various road types are extracted, and frequency statistical analysis is performed to generate speed-torque frequency statistical matrix of various road types; S203: the probability density distribution fitting of the speed-torque frequency statistical matrix of various road types is performed by non-parametric two-dimensional kernel density estimation algorithm, and the probability density distribution mathematical model of various road types is generated; S204: the speed-torque frequency statistical matrix of various road types is extrapolated by Monte Carlo simulation method combined with the corresponding probability density distribution mathematical model, and then the extrapolated speed-torque frequency statistical matrix of various road types is superimposed to generate the user condition distribution under full life mileage.

4. The method of claim 3, wherein: In step S202, torque data {T 驱动1 , 驱动2 , 驱动i} and speed data {n 驱动1 , 驱动2 , 驱动i} are extracted from the user working condition sample to construct a corresponding speed-torque data set M={(T 驱动1 , n 驱动1 ),(T 驱动2 , n 驱动2 ),…,(T 驱动i , n 驱动i )}; and then frequency statistical analysis is performed on the speed-torque data set M to generate a corresponding speed-torque frequency statistical matrix.

5. The method of claim 3, wherein the efficiency of the electric drive assembly associated with the user is evaluated by: In step S203, the probability density distribution mathematical model is generated by the following steps: 1) the kernel density function and the initial bandwidth are calculated by the following formula: h = 2.4 σn -1 / 5 ; σ = min(σ x ,σ y ); In the formula, K represents a kernel density function; x and y represent the rotating speed and torque in the rotating speed-torque frequency statistical matrix respectively; h represents an initial bandwidth; n represents the capacity of the user working condition sample; σ represents the standard deviation of the user working condition sample, σ x represents the standard deviation of the rotating speed, σ y represents the standard deviation of the torque; 2) the adaptive bandwidth is calculated by the following formula: In the formula: represents the adaptive bandwidth; h represents the initial bandwidth; P(x i ,y i ) represents the ratio of the number of occurrences of the speed-torque working point (x i ,y i ) in the speed-torque frequency statistical moment matrix to the total number of speed-torque working points; α represents a sensitivity coefficient, and the value is 0.5; n represents the capacity of the user working condition sample; 3) the probability density distribution mathematical model is calculated by the following formula: In the formula, f(x, y) represents a probability density distribution mathematical model of the rotating speed-torque frequency statistical matrix; represents an adaptive bandwidth; K represents a kernel density function; and x and y respectively represent rotating speed and torque in the rotating speed-torque frequency statistical matrix.

6. The method of claim 3, wherein the efficiency of the electric drive assembly associated with the user is evaluated by: In step S204, the speed-torque condition points are randomly placed according to the extrapolation coefficient of the corresponding mileage by the Monte Carlo simulation method, and then the bidirectional extrapolation of the speed-torque frequency statistical matrix in the amplitude and frequency is realized; the extrapolation coefficient is obtained by dividing the target mileage of each road type by the measured mileage.

7. The method of claim 1, wherein: In step S301, the efficiency prediction model is constructed by the following steps: S3011: determining the BP neural network structure and training algorithm of the efficiency prediction model; S3012: determining the speed-torque condition point distribution principle for testing based on the user condition distribution under the full life mileage, and then testing the efficiency of each speed-torque condition point through the electric drive assembly efficiency bench test; S3013: selecting the corresponding speed-torque condition points and efficiency to construct the training data set with speed and torque as input and efficiency as output, and dividing the training data set into training samples and test samples; S3014: training and testing the neural network by the training samples and test samples combined with the corresponding training algorithm, so as to obtain the corresponding efficiency prediction model.

8. The method of claim 7, wherein the method further comprises: In step S3012, the efficiency of the speed-torque condition point is calculated by the following formula: 1) Efficiency calculation formula in driving mode: wherein: η1 represents the efficiency of the electric drive assembly in the drive mode; n1 represents the left motor speed; T1 represents the left motor torque; n2 represents the right motor speed; T2 represents the right motor torque; P i represents the electric drive assembly input power; 2) Efficiency calculation formula in power generation mode: wherein: η2 represents the efficiency of the electric drive assembly in power generation mode; n1 represents the left side power machine rotational speed; T1 represents the left side power machine torque; n2 represents the right side power machine rotational speed; T2 represents the right side power machine torque; P i represents the electric drive assembly input power.

9. The method of claim 1, wherein: In step S4, the electric drive assembly high efficiency area utilization rate under the full life mileage is calculated by the following formula: In the formula, D represents the electric drive assembly high efficiency area utilization rate under the full life mileage; W represents the number of efficiencies greater than the expected efficiency M% corresponding to all torque-speed condition points in the user condition distribution under the full life mileage; Q represents the total number of speed-torque condition points in the user condition distribution under the full life mileage.

Images