Method for testing overall performance of electro-tricycle

Abstract

The invention discloses an electro-tricycle whole vehicle performance test method, and relates to the technical field of vehicle performance test, and the method comprises the steps: carrying out the data mining according to the multivariate operation scene factors of an electro-tricycle, and constructing Q test scenes; performing multi-dimensional performance expectation prediction to obtain Q scene expectation power performance, Q scene expectation electrical performance and Q scene expectation comfort performance; performing dynamic characteristic detection on the electro-tricycle to obtain a dynamic characteristic detection result; performing electrical characteristic detection on the electric tricycle to obtain an electrical characteristic detection result; carrying out comfort feature detection on the electro-tricycle to obtain a comfort feature detection result; and according to the dynamic feature detection result, the electrical feature detection result and the comfort feature detection result, constructing a whole vehicle performance visual report of the electro-tricycle. The technical problem that the multi-dimensional performance of the electro-tricycle cannot be comprehensively evaluated in the prior art is solved, and the technical effect of comprehensively evaluating the multi-dimensional performance of the electro-tricycle is achieved.

Description

Technical Field

[0001] The invention relates to the technical field of vehicle performance testing, and in particular to a method for testing the performance of an electric tricycle. Background Art

[0002] With the widespread use of electric tricycles, especially in urban transportation, logistics and short-distance travel, the requirements for the performance of the whole vehicle are getting higher and higher. The performance of electric tricycles includes not only traditional power performance, but also electrical performance and comfort performance. At present, the performance evaluation of electric tricycles mainly relies on the test method in a single scenario, and lacks a comprehensive evaluation system for multiple complex operating scenarios. Summary of the invention

[0003] The present application provides a whole-vehicle performance testing method for an electric tricycle, which is used to solve the technical problem that the prior art is unable to comprehensively evaluate the multi-dimensional performance of an electric tricycle.

[0004] In view of the above problems, the present application provides a method for testing the performance of an electric tricycle.

[0005] The present application provides a method for testing the performance of an electric tricycle, the method comprising: Data mining is performed based on the multivariate operating scenario factors of the electric tricycle to construct Q test scenarios, wherein the multivariate operating scenario factors include the operating road state, the operating environment state and the operating load state, and Q is a positive integer greater than 1; a multidimensional performance expectation prediction is performed on the electric tricycle based on the Q test scenarios to obtain Q scenario expected power performance, Q scenario expected electrical performance and Q scenario expected comfort performance; based on the Q test scenarios, a power characteristic detection is performed on the electric tricycle based on the Q scenario expected power performance to obtain a power characteristic detection result; based on the Q test scenarios, an electrical characteristic detection is performed on the electric tricycle based on the Q scenario expected electrical performance to obtain an electrical characteristic detection result; based on the Q test scenarios, a comfort characteristic detection is performed on the electric tricycle based on the Q scenario expected comfort performance to obtain a comfort characteristic detection result; based on the power characteristic detection result, the electrical characteristic detection result and the comfort characteristic detection result, a visual report on the whole vehicle performance of the electric tricycle is constructed.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: The present application conducts data mining based on the multivariate operating scenario factors of an electric tricycle to construct Q test scenarios, wherein the multivariate operating scenario factors include the operating road state, the operating environment state and the operating load state, and Q is a positive integer greater than 1; a multidimensional performance expectation prediction is performed on the electric tricycle based on the Q test scenarios to obtain the expected power performance of Q scenarios, the expected electrical performance of Q scenarios and the expected comfort performance of Q scenarios; based on the Q test scenarios, a power characteristic detection is performed on the electric tricycle based on the expected power performance of the Q scenarios to obtain a power characteristic detection result; based on the Q test scenarios, an electrical characteristic detection is performed on the electric tricycle based on the expected electrical performance of the Q scenarios to obtain an electrical characteristic detection result; based on the Q test scenarios, a comfort characteristic detection is performed on the electric tricycle based on the expected comfort performance of the Q scenarios to obtain a comfort characteristic detection result; based on the power characteristic detection results, the electrical characteristic detection results and the comfort characteristic detection results, a visual report on the whole vehicle performance of the electric tricycle is constructed. The present invention solves the technical problem that the prior art cannot comprehensively evaluate the multi-dimensional performance of an electric tricycle. By testing based on multiple operating scenario factors, Q test scenarios are constructed, and data mining technology is combined to predict the power, electrical and comfort performance of an electric tricycle in different scenarios. The power, electrical and comfort characteristics are respectively detected through a feature detection model to obtain various performance data. The test results are summarized and visualized to form a vehicle performance report, thereby achieving the technical effect of comprehensively evaluating the multi-dimensional performance of an electric tricycle. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0008] Figure 1 A schematic flow chart of a method for testing the performance of an electric tricycle provided in an embodiment of the present application.

[0009] Figure 2 A schematic diagram of the process of obtaining power characteristic detection results in the electric tricycle whole vehicle performance testing method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0010] The present application provides a whole vehicle performance testing method for an electric tricycle, which is used to solve the technical problem that the prior art cannot comprehensively evaluate the multi-dimensional performance of an electric tricycle. By testing based on multiple operating scenario factors, Q test scenarios are constructed, and data mining technology is combined to predict the power, electrical and comfort performance of the electric tricycle in different scenarios. The power, electrical and comfort characteristics are respectively tested through a feature detection model to obtain various performance data. The test results are summarized and visualized to form a whole vehicle performance report, thereby achieving the technical effect of comprehensively evaluating the multi-dimensional performance of an electric tricycle.

[0011] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0012] It should be noted that any variations of the terms "include" and "have" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0013] Examples, such as Figure 1 As shown, the present application provides a method for testing the performance of an electric tricycle, the method comprising: Step S100: Perform data mining based on the multivariate operating scenario factors of the electric tricycle to construct Q test scenarios, wherein the multivariate operating scenario factors include operating road conditions, operating environment conditions, and operating load conditions, and Q is a positive integer greater than 1.

[0014] In the embodiment of the present application, the multivariate operation scenario factors include the operation road state, the operation environment state and the operation load state. The operation road state refers to the road type, road condition, slope, traffic condition, etc. on which the tricycle is traveling. The operation environment state refers to the external environmental conditions, such as weather factors such as temperature, humidity, wind speed, etc. The operation load state refers to the performance of the tricycle under different load conditions.

[0015] When conducting data mining based on the multivariate operating scenario factors of the electric tricycle, data mining techniques such as cluster analysis are used to identify and construct multiple test scenarios. Specifically, historical data of the multivariate operating scenario factors are first obtained from the historical database, and representative scenario patterns are identified using data mining techniques such as cluster analysis.

[0016] Next, based on the multivariate factor data obtained through mining, multiple possible test scenario sets are constructed. Each test scenario contains different combinations of roads, environments, and loads. Subsequently, a pairwise similarity evaluation is performed on these test scenarios, and the similarity between different scenarios is determined by calculating the similarity coefficient.

[0017] Finally, according to the pre-set similarity threshold, Q test scenes with similarity below the threshold are screened out. Furthermore, the method provided in the embodiment of the application performs data mining based on the multivariate operating scenario factors of the electric tricycle to construct Q test scenarios, and also includes: Based on the multivariate operating scenario factors, a test scenario is set for the electric tricycle to obtain a test scenario solution set; pairwise similarity evaluation is performed based on the test scenario solution set to obtain a test scenario similarity coefficient set; based on the test scenario similarity coefficient set, the test scenario solution set is screened according to a test scenario similarity threshold to obtain the Q test scenarios that are less than the test scenario similarity threshold.

[0018] In an embodiment of the present application, the test scenario is first set according to the multivariate operating scenario factors of the electric tricycle. Each test scenario solution contains different combinations of these three factors, that is, each test scenario solution is jointly determined by multiple factors (such as road conditions, environmental conditions, and load conditions). Specifically, first, the operation data of the electric tricycle under different conditions are collected from the historical database, including multi-dimensional indicators such as vehicle speed, acceleration, power consumption, and battery status. Subsequently, these collected data are analyzed using data mining techniques, such as cluster analysis (such as K-means clustering). Through cluster analysis, test scenarios with similar characteristics are classified into the same category according to factors such as road conditions, environmental conditions, and load conditions, and each category represents a specific test scenario solution. Therefore, a plurality of test scenario solution sets containing different combinations of roads, environments, and load conditions are finally obtained.

[0019] Next, based on the test scenario set, pairwise similarity evaluation is performed. The similarity between scenarios is calculated by using cosine similarity. Specifically, cosine similarity measures the similarity between test scenarios by calculating the angle between the factor vectors. The higher the similarity, the more similar the two test scenarios are in terms of road, environment or load conditions, and vice versa, it means that they have significant differences. Through calculation, the test scenario similarity coefficient set is obtained.

[0020] Afterwards, in order to remove redundancy and ensure the representativeness of the test scenarios, screening is performed based on the obtained test scenario similarity coefficient set. The test scenario similarity coefficient set is compared with the preset test scenario similarity threshold, and the test scenario solution set is screened to obtain Q test scenarios that are less than the test scenario similarity threshold. Specifically, for each test scenario solution, the similarity coefficient of the test scenario solution with other scenarios is compared with the preset test scenario similarity threshold. If the similarity with other scenarios is lower than the test scenario similarity threshold, the scenario is considered to be independent and representative. Finally, Q test scenarios are screened out.

[0021] Step S200: Perform multi-dimensional performance expectation prediction on the electric tricycle according to the Q test scenarios to obtain Q scenarios' expected power performance, Q scenarios' expected electrical performance and Q scenarios' expected comfort performance.

[0022] In an embodiment of the present application, the process of predicting the expected multi-dimensional performance of an electric tricycle includes the prediction of power performance, electrical performance, and comfort performance. Specifically, first obtain the vehicle production process information of the electric tricycle, such as motor parameters, drive system configuration, battery technology, etc. Then, based on Q test scenarios and vehicle production process information, the expected predictions of power performance, electrical performance, and comfort performance are made respectively. In the power performance expectation prediction, the expected maximum vehicle speed, expected maximum climbing angle, and expected braking distance of each scenario are predicted based on the road, environment, and load conditions of the test scenario. In the electrical performance expectation prediction, the expected cruising range, motor efficiency, and motor temperature rise are predicted based on the same scenario conditions. In the comfort performance expectation prediction, the expected driving noise, driving vibration amplitude, and vibration frequency in each scenario are predicted.

[0023] Through these steps, we can finally obtain the expected power performance of Q scenarios, the expected electrical performance of Q scenarios and the expected comfort performance of Q scenarios.

[0024] Furthermore, in the method provided in the embodiment of the application, a multi-dimensional performance expectation prediction is performed on the electric tricycle according to the Q test scenarios to obtain the expected power performance of Q scenarios, the expected electrical performance of Q scenarios and the expected comfort performance of Q scenarios, and further includes: Obtain vehicle production process information of the electric tricycle; predict the expected power performance of the electric tricycle based on the Q test scenarios and the vehicle production process information, and generate the expected power performance of the Q scenarios, wherein the expected power performance of each scenario includes the expected maximum vehicle speed, expected maximum climbing angle and expected braking distance corresponding to each test scenario; predict the expected electrical performance of the electric tricycle based on the Q test scenarios and the vehicle production process information, and generate the expected electrical performance of the Q scenarios, wherein the expected electrical performance of each scenario includes the expected cruising range, expected motor efficiency and expected motor temperature rise corresponding to each test scenario; predict the expected comfort performance of the electric tricycle based on the Q test scenarios and the vehicle production process information, and generate the expected comfort performance of the Q scenarios, wherein the expected comfort performance of each scenario includes the expected driving noise amount, expected driving vibration amplitude and expected driving vibration frequency corresponding to each test scenario.

[0025] In an embodiment of the present application, the vehicle production process information of the electric tricycle uploaded in advance is first obtained, including key technical parameters of the electric tricycle, such as motor power, drive system configuration, battery capacity, battery management system, braking system and frame design.

[0026] Next, based on the Q test scenarios and vehicle production process information, the expected power performance is predicted to generate the expected power performance in each test scenario. Specifically, the test scenario record, vehicle production process record and scenario standard power performance record of the electric tricycle are first loaded. The test scenario record includes road type, environmental conditions and load conditions, while the scenario standard power performance record provides the standard power performance of the electric tricycle in different test scenarios, such as the expected maximum vehicle speed, maximum climbing angle and braking distance. Then, supervised learning methods such as neural networks are used to train the test scenario record and vehicle production process information as input data and the scenario standard power performance record as the target output. In this process, the goal of the training is to minimize the expected prediction loss coefficient of the power performance, that is, to reduce the error between the prediction result and the actual standard power performance. After the training is completed, if the prediction loss coefficient is less than the preset power performance expected prediction loss threshold, the power performance expected prediction model is generated. At this time, the qth test scenario and vehicle production process information are input into the model to obtain the expected power performance of the test scenario and add it to the Q scenario expected power performance list. By inputting one of the Q test scenarios and the vehicle production process information into the model, Q expected power performances of the scenarios are obtained, wherein the expected power performance of each scenario includes the expected maximum vehicle speed, the expected maximum climbing angle and the expected braking distance corresponding to each test scenario.

[0027] Similarly, based on Q test scenarios and vehicle production process information, electrical performance expectations are predicted to generate expected electrical performance for Q scenarios. The specific process is similar to the power performance prediction. First, the test scenario records and vehicle production process information of the electric tricycle are loaded, and then the electrical performance prediction model is trained using a supervised learning method. The goal of the model is to predict indicators such as expected cruising range, motor efficiency, and motor temperature rise in each test scenario. Finally, the trained electrical performance expectation prediction model is also applied to each test scenario to obtain the electrical performance expectations of Q scenarios, and add them to the corresponding list. Among them, the expected electrical performance of each scenario includes the expected cruising range, expected motor efficiency, and expected motor temperature rise corresponding to each test scenario.

[0028] Finally, based on Q test scenarios and vehicle production process information, the comfort performance of the electric tricycle is expected to be predicted, and the expected comfort performance of Q scenarios is generated. This part of the prediction is similar to the prediction process of power performance and electrical performance. After loading the relevant data, the comfort performance expectation prediction model is trained using the supervised learning method. The goal of the model is to predict the expected driving noise, vibration amplitude, vibration frequency and other indicators under each test scenario. The trained comfort performance expectation prediction model is applied to each test scenario to obtain the expected comfort performance of Q scenarios. Among them, the expected comfort performance of each scenario includes the expected driving noise, expected driving vibration amplitude and expected driving vibration frequency corresponding to each test scenario.

[0029] Through the above steps, the corresponding expected power performance, expected electrical performance and expected comfort performance are generated for each test scenario.

[0030] Furthermore, in the method provided in the embodiment of the application, based on the Q test scenarios and the vehicle production process information, the expected power performance of the electric tricycle is predicted to generate the expected power performance of the Q scenarios, and further includes: Loading electric tricycle test scenario records, electric tricycle production process records and scenario standard power performance records; using the electric tricycle test scenario records and the electric tricycle production process records as input information and the scenario standard power performance records as output information, performing supervised training on a predetermined learning model to obtain an expected prediction loss coefficient of power performance; if the expected prediction loss coefficient of power performance is less than an expected prediction loss threshold of power performance, generating an expected power performance prediction model; inputting the qth test scenario and the vehicle production process information into the expected power performance prediction model to obtain the expected power performance of the qth scenario, and adding the expected power performance of the qth scenario to the expected power performance of the Q scenarios.

[0031] In the embodiment of the present application, the electric tricycle test scenario record, the electric tricycle production process record and the scenario standard power performance record are first loaded from the historical database. The electric tricycle test scenario record includes different test conditions, such as road type, ambient temperature, humidity and vehicle load status, which directly affect the dynamic performance of the vehicle; the electric tricycle production process record provides the manufacturing parameters of the vehicle, including motor power, drive system configuration, battery type and capacity, brake system design, etc., which will significantly affect the power performance of the electric tricycle; the scenario standard power performance record contains the known standard power performance data of the electric tricycle in different test scenarios, such as the expected maximum speed, maximum climbing angle, braking distance, etc.

[0032] Once these input data are ready, the next step is to train them using a supervised learning model. Specifically, the electric tricycle test scene records and electric tricycle production process records are used as input features, and the scene standard power performance records are used as target outputs. The predetermined learning model is trained using methods such as support vector machine regression. The goal of supervised learning is to minimize the error between the predicted value and the actual value, that is, the model learns how to predict the power performance under a given test scenario and production process through the input features.

[0033] During the training process, the power performance expected prediction loss coefficient is calculated, which is an indicator to evaluate the model performance by comparing the difference between the predicted results and the actual standard power performance. The power performance expected prediction loss coefficient is obtained by calculating the error between the predicted value and the true value through the mean square error, which measures the difference between the model prediction results and the actual data.

[0034] If the power performance expected prediction loss coefficient after training is less than the power performance expected prediction loss threshold preset by technical experts, the model is considered to have achieved the required accuracy and can be used for actual prediction. At this time, using this trained power performance expected prediction model, the qth test scenario and the production process information of the electric tricycle are input into the model to obtain the expected power performance of the qth scenario, including the expected maximum vehicle speed, maximum climbing angle and braking distance, etc. Finally, the expected power performance of the qth scenario is added to the expected power performance of the Q scenarios.

[0035] Step S300: Based on the Q test scenarios, power characteristic detection is performed on the electric tricycle according to the expected power performance of the Q scenarios to obtain a power characteristic detection result.

[0036] In the embodiment of the present application, when the power characteristics of the electric tricycle are detected, the power performance confidence test is first performed to obtain the actual test power performance in each test scenario. Then, the test power performance of each scenario is compared with the corresponding expected power performance, the deviation is identified, and a power performance deviation matrix containing all scenarios is generated.

[0037] Subsequently, the power performance deviation matrix is ​​input into a power performance detection model, and the power performance detection coefficient is further calculated to quantify the power performance difference of each test scenario. The power performance stability is obtained by reciprocal calculation of the variance of these detection coefficients, which is an indicator to measure the consistency of power output of electric tricycles in different test scenarios. Finally, the power performance stability, test power performance and power performance detection coefficient are outputted comprehensively to form the power characteristic detection result.

[0038] Further, such as Figure 2 As shown, in the method provided in the embodiment of the application, the power characteristic detection of the electric tricycle is performed according to the expected power performance of the Q scenarios to obtain the power characteristic detection result, and also includes: According to the Q test scenarios, a power performance confidence test is performed on the electric tricycle to obtain Q scenario test power performances; based on the Q scenario expected power performances, deviations of the Q scenario test power performances are identified to generate Q scenario power performance deviation matrices; the Q scenario power performance deviation matrices are input into a power performance detection model to obtain Q scenario power performance detection coefficients; the inverse of the variance of the Q scenario power performance detection coefficients is taken as the power performance stability; the power performance stability, the Q scenario test power performances and the Q scenario power performance detection coefficients are output as the power feature detection results.

[0039] In the embodiment of the present application, firstly, the power performance confidence test is performed on the electric tricycle according to Q test scenarios. In this stage, the electric tricycle will be actually tested according to different test scenario conditions (such as load, road conditions, environment, etc.) to obtain the test power performance in each scenario. Through this process, Q scenario test power performance is obtained.

[0040] Next, based on the expected power performance of the Q scenarios, the deviations of the tested power performance of the Q scenarios are identified. That is, the expected power performance of the Q scenarios are compared with the tested power performance of the Q scenarios. By comparing the gap between the actual and expected, the deviations in each scenario are identified, and a power performance deviation matrix of the Q scenarios is generated.

[0041] Then, the power performance deviation matrices of Q scenarios are input into the power performance detection model to obtain the power performance detection coefficients of Q scenarios. The power performance detection model is pre-trained. When training, the historical test data is first obtained from the historical database, including the power performance deviation matrix of the historical scenarios and the corresponding power performance detection coefficients. The power performance deviation matrix of the historical scenarios is used as input data, and the power performance detection coefficient of each test scenario is used as output data. The multi-layer perceptron neural network is trained to obtain the power performance detection model.

[0042] Then, the power performance detection coefficients of the Q scenes are used to calculate the variance of the detection coefficients, and the inverse of the variance of the power performance detection coefficients of the Q scenes is taken as the power performance stability.

[0043] Finally, the power performance stability, Q scenario test power performance and Q scenario power performance detection coefficients are summarized as the final power characteristic detection result.

[0044] Furthermore, in the method provided in the embodiment of the application, according to the Q test scenarios, the power performance confidence test is performed on the electric tricycle to obtain the power performance of Q scenario tests, and further includes: According to the Q test scenarios, extract the qth test scenario, where q is a positive integer, 1≤q≤Q; perform multiple maximum speed tests on the electric tricycle according to the qth test scenario to obtain a maximum speed test data set; perform multiple maximum climbing angle tests on the electric tricycle according to the qth test scenario to obtain a maximum climbing angle test data set; perform a braking distance test on the electric tricycle according to the qth test scenario to obtain a braking distance test data set; perform mean calculations on the maximum speed test data set, the maximum climbing angle test data set and the braking distance test data set, respectively, to obtain the qth scenario confident maximum speed, the qth scenario confident maximum climbing angle and the qth scenario confident braking distance; output the qth scenario confident maximum speed, the qth scenario confident maximum climbing angle and the qth scenario confident braking distance as the qth scenario test power performance, and add the qth scenario test power performance to the Q scenario test power performance.

[0045] In the embodiment of the present application, the qth test scene is first randomly extracted from Q test scenes, where q is a positive integer, 1≤q≤Q. Each test scene contains different environmental conditions, road conditions, load conditions, etc., and the performance of the electric tricycle will change according to these factors.

[0046] Next, according to the qth test scenario, a series of specific performance tests are performed on the electric tricycle. First, a maximum speed test is performed. This test records the maximum speed of the electric tricycle in this scenario through multiple tests. Through these test data, a maximum speed test data set is formed. Then, in the same test scenario, a maximum climbing angle test is performed. Through multiple tests, the maximum slope angle that the electric tricycle can climb is measured. Finally, the maximum slope angle climbed by the electric tricycle in multiple measurements is integrated to obtain the maximum climbing angle test data set. Next, according to the qth test scenario, a braking distance test is performed on the electric tricycle. Multiple braking tests are performed, and the braking distance required from the maximum speed to the stop of each test is recorded. The longest braking distance of each test is taken as the result of the test. Finally, these longest braking distance values ​​are integrated to obtain a braking distance test data set.

[0047] Then, the mean of the maximum vehicle speed test data set, the maximum climbing angle test data set and the braking distance test data set are calculated respectively. The mean of the maximum vehicle speed test data set is calculated to obtain the confident maximum vehicle speed of the qth scene; the mean of the maximum climbing angle test data set is calculated to obtain the confident maximum climbing angle of the qth scene; the mean of the braking distance test data set is calculated to obtain the confident braking distance of the qth scene.

[0048] Finally, the calculated confident maximum vehicle speed of the qth scenario, the confident maximum climbing angle of the qth scenario and the confident braking distance of the qth scenario are output as the qth scenario test power performance, and the qth scenario test power performance is added to the Q scenario test power performance.

[0049] In order to obtain the global test results, all Q test scenarios are traversed, and the above steps are repeated for the test results of each scenario, and finally the test power performance of Q scenarios is obtained.

[0050] Step S400: Based on the Q test scenarios, electrical characteristic detection is performed on the electric tricycle according to the expected electrical performance of the Q scenarios to obtain electrical characteristic detection results.

[0051] In an embodiment of the present application, based on Q test scenarios, the electrical performance factors of the electric tricycle are first set. These factors include key electrical performance indicators such as cruising range, motor efficiency, and motor temperature rise. Then, for each test scenario, an electrical performance confidence test is performed to obtain the test electrical performance data for each scenario. Then, by comparing the actual electrical performance of each test scenario with the expected electrical performance, deviation identification is performed to generate an electrical performance deviation matrix. This deviation matrix is ​​then input into a pre-trained electrical performance detection model to calculate the electrical performance detection coefficient for each scenario. Then, by calculating the inverse of the variance of these coefficients, the electrical performance stability is obtained. Finally, the electrical performance stability, the test electrical performance, and the electrical performance detection coefficient are integrated and output to generate an electrical feature detection result.

[0052] Furthermore, in the method provided in the embodiment of the application, based on the Q test scenarios, the electrical characteristics of the electric tricycle are detected according to the expected electrical performance of the Q scenarios to obtain the electrical characteristics detection results, and the method also includes: An electrical performance factor of the electric tricycle is set, wherein the electrical performance factor includes cruising range, motor efficiency and motor temperature rise; based on the electrical performance factor, an electrical performance confidence test is performed on the electric tricycle according to the Q test scenarios to obtain Q scenario test electrical performance; deviations are identified for the Q scenario test electrical performance according to the Q scenario expected electrical performance to generate Q scenario electrical performance deviation matrices; the Q scenario electrical performance deviation matrices are input into an electrical performance detection model to obtain Q scenario electrical performance detection coefficients; the inverse of the variance of the Q scenario electrical performance detection coefficients is calculated to obtain electrical performance stability; the electrical performance stability, the Q scenario test electrical performance and the Q scenario electrical performance detection coefficients are output as the electrical feature detection result.

[0053] In an embodiment of the present application, the electrical performance factor of the electric tricycle is first set, including the cruising range, motor efficiency and motor temperature rise. Next, based on the electrical performance factor, the electric tricycle is subjected to an electrical performance confidence test according to Q test scenarios. Specifically, the process of conducting an electrical performance confidence test is similar to the previous power performance confidence test. The electrical performance confidence test first randomly extracts the qth test scene from the Q test scenes, and then tests the electric tricycle multiple times in the randomly extracted scenes to obtain test electrical performance data under different scenarios, including indicators such as maximum cruising range, maximum motor efficiency and maximum motor temperature rise, which respectively constitute the maximum cruising range test data set, the maximum motor efficiency test data set and the maximum motor temperature rise test data set. Subsequently, the maximum cruising range test data set, the maximum motor efficiency test data set and the maximum motor temperature rise test data set are respectively averaged to obtain the maximum cruising range of the qth scene, the maximum motor efficiency of the qth scene and the maximum motor temperature rise of the qth scene. Then, the calculated maximum cruising range of the qth scenario, the maximum motor efficiency of the qth scenario, and the maximum motor temperature rise of the qth scenario are output as the qth scenario test electrical performance, and the qth scenario test electrical performance is added to the Q scenario test electrical performance. In order to obtain the global test results, all Q test scenarios are traversed, and the above steps are repeated for the test results of each scenario, and finally the Q scenario test electrical performance is obtained.

[0054] Next, the deviations of the Q scenario test electrical performance are identified based on the expected electrical performance of the Q scenarios, that is, the gaps between the actual test data and the expected values ​​in each scenario are calculated, and based on these deviations, the electrical performance deviation matrix of the Q scenarios is generated.

[0055] Then, the electrical performance deviation matrices of Q scenarios are input into the electrical performance detection model to obtain the electrical performance detection coefficients of Q scenarios. The electrical performance detection model is pre-trained. When training the electrical performance detection model, the historical electrical performance deviation matrix set and the corresponding electrical performance detection coefficient set are first obtained from the historical database. The historical electrical performance deviation matrix set is used as input data and the electrical performance detection coefficient set is used as output data to train the multi-layer perceptron to obtain the electrical performance detection model.

[0056] Then, the inverse of the variance of the electrical performance detection coefficients of the Q scenarios is calculated to obtain the electrical performance stability. The electrical performance stability, the Q scenario test electrical performance and the Q scenario electrical performance detection coefficients are output as electrical feature detection results.

[0057] Step S500: Based on the Q test scenarios, comfort feature detection is performed on the electric tricycle according to the expected comfort performance of the Q scenarios to obtain comfort feature detection results.

[0058] In an embodiment of the present application, the comfort performance factor of the electric tricycle is first set, including the driving noise amount, driving vibration amplitude and driving vibration frequency. Then, in each test scenario, a comfort performance confidence test is performed based on the comfort performance factor to obtain the test comfort performance data of Q scenarios. Next, the test comfort performance of each scenario is compared with the expected value, the deviation is identified and a comfort performance deviation matrix is ​​generated. The deviation matrix is ​​input into a pre-trained comfort performance detection model to calculate and obtain the comfort performance detection coefficient of each scenario. Finally, the comfort performance stability is obtained by calculating the inverse of the variance of the detection coefficient, and the stability, test comfort performance and detection coefficient are output as the final comfort feature detection result.

[0059] Furthermore, in the method provided in the embodiment of the application, based on the Q test scenarios, the comfort feature detection of the electric tricycle is performed according to the expected comfort performance of the Q scenarios to obtain the comfort feature detection result, and further includes: A comfort performance factor of the electric tricycle is set, wherein the comfort performance factor includes driving noise, driving vibration amplitude and driving vibration frequency; based on the comfort performance factor, a comfort performance confidence test is performed on the electric tricycle according to the Q test scenarios respectively to obtain Q scenario test comfort performance; deviation identification is performed on the Q scenario test comfort performance according to the Q scenario expected comfort performance to generate a Q scenario comfort performance deviation matrix; the Q scenario comfort performance deviation matrix is ​​input into a comfort performance detection model to obtain Q scenario comfort performance detection coefficients; the inverse of the variance of the Q scenario comfort performance detection coefficients is taken as the comfort performance stability; the comfort performance stability, the Q scenario test comfort performance and the Q scenario comfort performance detection coefficients are output as the comfort feature detection result.

[0060] In the embodiment of the present application, the comfort performance factor of the electric tricycle is first set, wherein the comfort performance factor includes the driving noise amount, the driving vibration amplitude and the driving vibration frequency.

[0061] Next, based on the comfort performance factor, the comfort performance confidence test is performed on the electric tricycle according to the Q test scenarios. Specifically, the qth scene is first randomly extracted from the Q test scenes, and the electric tricycle is tested multiple times in this scene, and the driving noise, driving vibration amplitude, and driving vibration frequency in each test are recorded respectively. In each test, the maximum value is taken to form a test data set for each indicator. Next, the test data set of each indicator is averaged to obtain the test data of the maximum driving noise of the qth scene, the maximum driving vibration amplitude of the qth scene, and the maximum driving vibration frequency of the qth scene. After that, the calculated maximum driving noise of the qth scene, the maximum driving vibration amplitude of the qth scene, and the maximum driving vibration frequency of the qth scene are output as the comfort performance of the qth scene test, and the comfort performance of the qth scene test is added to the comfort performance of the Q scenes test. In order to obtain the global test results, all Q test scenes are traversed, and the above steps are repeated for the test results of each scene, and finally the comfort performance of the Q scenes test is obtained.

[0062] The comfort performance data of the Q scenarios were then compared with the expected comfort performance values ​​corresponding to each scenario to identify the comfort performance deviation of each scenario. This deviation reflects the difference between the actual test results and the expected values, and the comfort performance deviation matrix of the Q scenarios was generated through comparison.

[0063] Next, the comfort performance deviation matrices of Q scenes are input into the comfort performance detection model. The model calculates the comfort performance detection coefficient of each scene by analyzing the pattern of the deviation matrix, and obtains the comfort performance detection coefficients of Q scenes. The comfort performance detection model is pre-trained. When training the comfort performance detection model, the historical electrical performance deviation matrix set and the corresponding electrical performance detection coefficient set are first obtained from the historical database. The historical comfort performance deviation matrix set is used as input data and the comfort performance detection coefficient set is used as output data to train the multi-layer perceptron to obtain the comfort performance detection model. After training, the comfort performance detection model can predict the corresponding comfort performance detection coefficient according to the comfort performance deviation matrix of each test scene.

[0064] Next, the variance of the comfort performance detection coefficients of the Q scenes is calculated, and the inverse of the variance of the comfort performance detection coefficients of the Q scenes is taken as the comfort performance stability.

[0065] Finally, the comfort performance stability, Q scene test comfort performance and Q scene comfort performance detection coefficients are output as comfort feature detection results.

[0066] Step S600: constructing a visual report of the whole vehicle performance of the electric tricycle according to the power characteristic detection result, the electrical characteristic detection result and the comfort characteristic detection result.

[0067] In an embodiment of the present application, when constructing a visual report of the vehicle performance of an electric tricycle based on the power feature detection results, the electrical feature detection results, and the comfort feature detection results, the power performance, electrical performance, and comfort performance data obtained based on each test scenario are first directly sorted and displayed. By presenting these test results in the form of charts, the report can clearly show the actual performance of each performance in different test scenarios. In addition, through the display of deviation analysis and detection coefficients, the report reflects the difference between each performance indicator and the expected value. Finally, all this information is collected to form a comprehensive report to complete the construction of the visual report of the vehicle performance of the electric tricycle.

[0068] In the embodiments of the present application, in summary, the embodiments of the present application have at least the following technical effects: The present application conducts data mining based on the multivariate operating scenario factors of an electric tricycle to construct Q test scenarios, wherein the multivariate operating scenario factors include the operating road state, the operating environment state and the operating load state, and Q is a positive integer greater than 1; a multidimensional performance expectation prediction is performed on the electric tricycle based on the Q test scenarios to obtain the expected power performance of Q scenarios, the expected electrical performance of Q scenarios and the expected comfort performance of Q scenarios; based on the Q test scenarios, a power characteristic detection is performed on the electric tricycle based on the expected power performance of the Q scenarios to obtain a power characteristic detection result; based on the Q test scenarios, an electrical characteristic detection is performed on the electric tricycle based on the expected electrical performance of the Q scenarios to obtain an electrical characteristic detection result; based on the Q test scenarios, a comfort characteristic detection is performed on the electric tricycle based on the expected comfort performance of the Q scenarios to obtain a comfort characteristic detection result; based on the power characteristic detection results, the electrical characteristic detection results and the comfort characteristic detection results, a visual report on the whole vehicle performance of the electric tricycle is constructed. The present invention solves the technical problem that the prior art cannot comprehensively evaluate the multi-dimensional performance of an electric tricycle. By testing based on multiple operating scenario factors, Q test scenarios are constructed, and data mining technology is combined to predict the power, electrical and comfort performance of an electric tricycle in different scenarios. The power, electrical and comfort characteristics are respectively detected through a feature detection model to obtain various performance data. The test results are summarized and visualized to form a vehicle performance report, thereby achieving the technical effect of comprehensively evaluating the multi-dimensional performance of an electric tricycle.

[0069] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0070] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0071] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. The electric tricycle vehicle performance test method is characterized by: The method comprises: Data mining is performed based on the multivariate operating scenario factors of the electric tricycle to construct Q test scenarios, wherein the multivariate operating scenario factors include operating road conditions, operating environment conditions, and operating load conditions, and Q is a positive integer greater than 1; Perform multi-dimensional performance expectation prediction on the electric tricycle according to the Q test scenarios to obtain Q scenario expected power performance, Q scenario expected electrical performance and Q scenario expected comfort performance; Based on the Q test scenarios, performing power characteristic detection on the electric tricycle according to the expected power performance of the Q scenarios to obtain a power characteristic detection result; Based on the Q test scenarios, performing electrical characteristic detection on the electric tricycle according to the expected electrical performance of the Q scenarios to obtain an electrical characteristic detection result; Based on the Q test scenarios, performing comfort feature detection on the electric tricycle according to the expected comfort performance of the Q scenarios to obtain a comfort feature detection result; A visual report on the overall vehicle performance of the electric tricycle is constructed based on the power feature detection results, the electrical feature detection results and the comfort feature detection results.

2. The method according to claim 1, characterized in that Based on the Q test scenarios, power characteristics detection is performed on the electric tricycle according to the expected power performance of the Q scenarios to obtain a power characteristics detection result, including: According to the Q test scenarios, a power performance confidence test is performed on the electric tricycle to obtain Q scenario test power performance; Based on the Q expected power performances of the Q scenarios, deviations of the Q tested power performances are identified to generate a power performance deviation matrix for the Q scenarios; Inputting the Q scene power performance deviation matrices into a power performance detection model to obtain Q scene power performance detection coefficients; The inverse of the variance of the power performance detection coefficients of the Q scenarios is taken as the power performance stability; The power performance stability, the Q scene test power performance and the Q scene power performance detection coefficients are output as the power characteristic detection result.

3. The method according to claim 2, characterized in that According to the Q test scenarios, a power performance confidence test is performed on the electric tricycle to obtain Q scenario test power performance, including: Extracting the qth test scene according to the Q test scenes, where q is a positive integer, 1≤q≤Q; Performing multiple maximum speed tests on the electric tricycle according to the qth test scenario to obtain a maximum speed test data set; Performing multiple maximum climbing angle tests on the electric tricycle according to the qth test scenario to obtain a maximum climbing angle test data set; Performing a braking distance test on the electric tricycle according to the qth test scenario to obtain a braking distance test data set; Performing mean calculation on the maximum vehicle speed test data set, the maximum climbing angle test data set, and the braking distance test data set, respectively, to obtain the qth scene confident maximum vehicle speed, the qth scene confident maximum climbing angle, and the qth scene confident braking distance; The qth scenario's confident maximum vehicle speed, the qth scenario's confident maximum climbing angle, and the qth scenario's confident braking distance are output as the qth scenario's test power performance, and the qth scenario's test power performance is added to the Q scenario's test power performance.

4. The method according to claim 1, characterized in that Based on the Q test scenarios, electrical characteristic detection is performed on the electric tricycle according to the expected electrical performance of the Q scenarios to obtain electrical characteristic detection results, including: Setting the electrical performance factor of the electric tricycle, wherein the electrical performance factor includes cruising range, motor efficiency and motor temperature rise; Based on the electrical performance factor, performing electrical performance confidence tests on the electric tricycle according to the Q test scenarios respectively to obtain electrical performance of Q scenario tests; According to the expected electrical performance of the Q scenarios, deviations of the tested electrical performance of the Q scenarios are identified to generate an electrical performance deviation matrix of the Q scenarios; Inputting the Q scene electrical performance deviation matrices into an electrical performance detection model to obtain Q scene electrical performance detection coefficients; Calculating the inverse of the variance of the electrical performance detection coefficients of the Q scenes to obtain the electrical performance stability; The electrical performance stability, the Q scenario test electrical performance and the Q scenario electrical performance detection coefficients are output as the electrical characteristic detection result.

5. The method according to claim 1, characterized in that Based on the Q test scenarios, a comfort feature detection is performed on the electric tricycle according to the expected comfort performance of the Q scenarios to obtain a comfort feature detection result, including: Setting a comfort performance factor of the electric tricycle, wherein the comfort performance factor includes a running noise amount, a running vibration amplitude and a running vibration frequency; Based on the comfort performance factor, performing comfort performance confidence tests on the electric tricycle according to the Q test scenarios respectively to obtain Q scenario test comfort performances; According to the Q scenes' expected comfort performance, deviations of the Q scenes' tested comfort performance are identified to generate a Q scenes' comfort performance deviation matrix; Inputting the Q scene comfort performance deviation matrices into a comfort performance detection model to obtain Q scene comfort performance detection coefficients; The inverse of the variance of the comfort performance detection coefficients of the Q scenes is taken as the comfort performance stability; The comfort performance stability, the Q scene test comfort performance and the Q scene comfort performance detection coefficients are output as the comfort feature detection result.

6. The method according to claim 1, characterized in that According to the Q test scenarios, a multi-dimensional performance expectation prediction is performed on the electric tricycle to obtain Q scenario expected power performance, Q scenario expected electrical performance and Q scenario expected comfort performance, including: Obtaining vehicle production process information of the electric tricycle; Based on the Q test scenarios and the vehicle production process information, the expected power performance of the electric tricycle is predicted to generate the Q expected power performance scenarios, wherein the expected power performance of each scenario includes the expected maximum vehicle speed, the expected maximum climbing angle and the expected braking distance corresponding to each test scenario; Based on the Q test scenarios and the vehicle production process information, an expected electrical performance prediction is performed on the electric tricycle to generate the Q scenario expected electrical performance, wherein the expected electrical performance of each scenario includes the expected cruising range, expected motor efficiency and expected motor temperature rise corresponding to each test scenario; Based on the Q test scenarios and the vehicle production process information, the expected comfort performance of the electric tricycle is predicted to generate the Q expected comfort performance scenarios, wherein the expected comfort performance of each scenario includes the expected driving noise amount, expected driving vibration amplitude and expected driving vibration frequency corresponding to each test scenario.

7. The method according to claim 6, characterized in that Based on the Q test scenarios and the vehicle production process information, the expected power performance of the electric tricycle is predicted to generate the expected power performance of the Q scenarios, including: Load the electric tricycle test scenario records, electric tricycle production process records and scenario standard power performance records; Using the electric tricycle test scene record and the electric tricycle production process record as input information and the scene standard power performance record as output information, supervised training is performed on a predetermined learning model to obtain an expected predicted loss coefficient of power performance; If the power performance expected prediction loss coefficient is less than the power performance expected prediction loss threshold, generating a power performance expected prediction model; The qth test scenario and the vehicle production process information are input into the power performance expectation prediction model to obtain the qth scenario expected power performance, and the qth scenario expected power performance is added to the Q scenario expected power performance.

8. The method according to claim 1, characterized in that Data mining is performed based on the multi-scenario factors of electric tricycles to construct Q test scenarios, including: Performing test scenario settings on the electric tricycle based on the multiple operating scenario factors to obtain a test scenario solution set; Perform pairwise similarity evaluation based on the test scenario solution set to obtain a test scenario similarity coefficient set; Based on the test scenario similarity coefficient set, the test scenario solution set is screened according to a test scenario similarity threshold to obtain the Q test scenarios that are less than the test scenario similarity threshold.

Images