New energy vehicle control strategy threshold analysis method based on monte carlo simulation

Abstract

This invention relates to the field of new energy vehicle technology and discloses a threshold analysis method for new energy vehicle control strategies based on Monte Carlo simulation, comprising the following steps: Step 1: Collect initial data; the initial data is a predetermined threshold of the vehicle's basic indicators used by the automobile manufacturer; Step 2: Preprocess the initial data to obtain a basic sample set; Step 3: Perform several random samplings on the basic sample set to obtain a sampled sample set; Step 4: Repeat Step 3 using the Monte Carlo simulation method to obtain an array of sampled samples and obtain the K value from it, where the K value is the quantile eigenvector of the Monte Carlo simulation; Step 5: Perform confidence interval analysis on the K value to obtain an estimated threshold; Step 6: Based on the estimated threshold, evaluate and determine the target threshold. This invention can provide reliable data reference for threshold setting and evaluation, and helps to improve the accuracy of new energy vehicle safety performance evaluation.

Description

Technical Field

[0001] This invention relates to the field of new energy vehicle technology, and specifically to a threshold analysis method for new energy vehicle control strategies based on Monte Carlo simulation. Background Technology

[0002] With the continuous advancement of the green energy revolution in the automotive industry, my country's production, sales, and ownership of new energy vehicles have also increased significantly. However, correspondingly, the operational safety issues of new energy vehicles have become increasingly prominent, causing significant negative impacts on traffic safety and personal safety. Therefore, it is currently urgent to strengthen the inspection and evaluation of various operational safety performance indicators of new energy vehicles, such as electrical safety and power battery safety.

[0003] The rapid development of the new energy vehicle industry, coupled with the large number of manufacturers and models, has led to inconsistent standards for setting thresholds and evaluation criteria for various indicators. This makes it difficult to objectively and systematically assess the safety performance of vehicles. Furthermore, considering practical applications, current threshold analysis methods often rely on fixed quantiles to determine thresholds for a fixed sample set. However, this fixed sample set cannot represent all vehicle models, and the number of models continues to increase, resulting in low reliability. Direct quantile analysis cannot compensate for this reliability deficiency. Therefore, quantile analysis based on a fixed sample set is insufficient for accurately assessing the safety performance of new energy vehicles from different manufacturers and models. Summary of the Invention

[0004] The present invention aims to provide a threshold analysis method for control strategies of new energy vehicles based on Monte Carlo simulation, which can provide reliable data reference for threshold setting and evaluation, and help improve the authenticity of safety performance evaluation of new energy vehicles.

[0005] The basic solution provided by this invention is: a threshold analysis method for new energy vehicle control strategies based on Monte Carlo simulation, comprising the following steps:

[0006] Step 1: Collect initial data; the initial data refers to the predetermined threshold values ​​of the basic vehicle indicators used by the car manufacturer.

[0007] Step 2: Preprocess the initial data to obtain the basic sample set;

[0008] Step 3: Perform several random samplings on the basic sample set to obtain the sampled sample set;

[0009] Step 4: Using the Monte Carlo simulation method, repeat step 3 to obtain an array of sampled samples and obtain the K value from it. The K value is the quantile feature vector of the Monte Carlo simulation.

[0010] Step 5: Perform confidence interval analysis on the K value and obtain the estimated threshold;

[0011] Step 6: Based on the estimated threshold, evaluate and determine the target threshold.

[0012] The working principle and advantages of this invention are as follows: First and foremost, the threshold analysis object of this scheme is the predetermined threshold set by each automobile manufacturer and actually used for evaluation. Compared with conventional threshold analysis schemes, which often limit the analysis perspective to a certain basic indicator of the vehicle and determine the threshold by analyzing a fixed sample set of that basic indicator, the threshold obtained by this analysis is actually limited by the fixed samples of each manufacturer and is also affected to some extent by human factors (different manufacturers use different analysis and value taking methods). This makes the reliability of the new energy vehicle performance results obtained based on this threshold poor, and it is difficult to obtain unified and accurate performance test results. This scheme breaks away from single indicator analysis and breaks through by using predetermined thresholds as the object of threshold analysis. Based on the predetermined thresholds, it determines a more reliable and objective target threshold that comprehensively considers all predetermined thresholds. It can provide reliable data reference for threshold setting and evaluation, and helps to improve the authenticity and objectivity of new energy vehicle safety performance evaluation.

[0013] Secondly, this scheme constructs a new sample set, namely the sampling sample set, by repeatedly randomly sampling the base sample set containing the predetermined threshold. Compared with the base sample set, the sampling sample set does not impose any human intervention on the sampling target during subsampling. This randomness in the sampling process effectively reduces the threshold determination differences caused by different manufacturers and vehicle models in conventional threshold determination, effectively reducing the influence of human factors on threshold determination, and thus helping to obtain a more objective and reliable target threshold. Furthermore, based on the geometric quantity and geometric characteristics of the predetermined threshold, this scheme uses the Monte Carlo simulation method to conduct sufficient digital simulation experiments using the sampling sample set (corresponding to step 4), and combines interval estimation to finally evaluate and obtain the target threshold. The overall threshold analysis method has a relatively simple operating logic and is highly efficient; it can obtain a more reliable target threshold.

[0014] Furthermore, in step 1, the vehicle's basic indicators include temperature, individual cell voltage, inter-cell voltage difference, SOC value, and SOC jump percentage.

[0015] Beneficial effects: This solution can analyze a wide variety of threshold items, and has universality and applicability.

[0016] Furthermore, in step 2, the preprocessing operation includes: taking the column vector corresponding to each predetermined threshold as a one-dimensional sample set; performing data cleaning based on the one-dimensional sample set to obtain a basic sample set.

[0017] Beneficial effects: Data cleaning can effectively remove useless data from the initial data, providing a reliable data foundation for subsequent threshold analysis.

[0018] Furthermore, data cleaning removes null values ​​and non-numerical data.

[0019] Beneficial effect: Cleaning and removing null values ​​and non-numerical data can effectively prevent such data from affecting the accuracy of K-value confirmation.

[0020] Furthermore, in step 3, the number of random samplings is equal to the length of the basic sample set; the random sampling is equal-length random sampling.

[0021] Beneficial effects: Using equal-length random sampling, the sampled samples are more uniform; the number of samplings is the same as the length of the basic sample set, which can obtain a sufficient number of sampled samples.

[0022] Furthermore, in step 4, K = [k i ], i = 1, 2, 3…1000;

[0023] Where, k i =S pi The sample set is S, S pi S is the sample set obtained by the i-th random sampling. i p-quantiles.

[0024] Beneficial effect: The reliable selection of quantile feature vectors can ensure the reliability of subsequent threshold estimation.

[0025] Furthermore, in step 5, the confidence level of the confidence interval is 95%.

[0026] Beneficial effect: A higher confidence level requires a higher accuracy in estimating the threshold. With this setting, although an estimation is used, the accuracy of the final estimated threshold can be guaranteed.

[0027] Furthermore, in step 5, the estimated threshold is calculated according to the following formula:

[0028]

[0029] Where μ is the estimated threshold, and S td These are the sample mean and sample variance, respectively; t α (n-1) is obtained from the t-distribution table.

[0030] Beneficial effect: The reliable confirmation of the estimated threshold can ensure the reliable confirmation of the subsequent target threshold. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the method flow of Embodiment 1 of the threshold analysis method for new energy vehicle control strategy based on Monte Carlo simulation of the present invention. Detailed Implementation

[0032] The following detailed explanation illustrates the specific implementation methods:

[0033] The basic implementation examples are as follows: Figure 1 As shown: A threshold analysis method for new energy vehicle control strategies based on Monte Carlo simulation, including the following steps:

[0034] Step 1: Collect initial data; the initial data refers to the predetermined threshold values ​​of the basic vehicle indicators used by the car manufacturer.

[0035] The vehicle's basic parameters include temperature, individual cell voltage, inter-cell voltage difference, SOC value, SOC jump percentage, insulation resistance, etc. Specifically, in this embodiment, temperature includes the highest temperature of the probe sensor, the lowest temperature of the probe sensor, and the temperature difference between the highest and lowest temperatures of the probe sensor. The probe sensor refers to temperature probe sensors installed at multiple key locations in the power battery system. At any given time, each temperature probe sensor collects a temperature value; the highest temperature is the maximum value among all collected temperature values, and the lowest temperature is the minimum value among all collected temperature values. Other parameters include the drive motor controller temperature, drive motor alarm temperature, and DC-DC controller temperature. Individual cell voltage includes the normal maximum voltage of ternary lithium batteries, the maximum voltage of lithium iron phosphate batteries, the lowest voltage of ternary lithium batteries, and the lowest voltage of lithium iron phosphate batteries. Inter-cell voltage difference includes the inter-cell voltage difference of ternary lithium batteries and the inter-cell voltage difference of lithium iron phosphate batteries. SOC value includes the minimum SOC value of the power battery system.

[0036] Step 2: Preprocess the initial data to obtain the basic sample set.

[0037] The preprocessing operation includes: taking the column vector corresponding to each predetermined threshold as a one-dimensional sample set; performing data cleaning based on the one-dimensional sample set to obtain the basic sample set X. The data cleaning removes null values ​​and non-numerical data.

[0038] Step 3: Perform several random samplings on the basic sample set to obtain the sampled sample set.

[0039] Specifically, the number of random samplings is equal to the length of the base sample set; the random sampling is equal-length random sampling. Let the length of the base sample set X be n, and perform n random samplings on the base sample set X to obtain a sampled sample set S of length n.

[0040] Step 4: Using the Monte Carlo simulation method, repeat step 3 to obtain an array of sampled samples; and obtain the K value from it, where the K value is the quantile feature vector of the Monte Carlo simulation.

[0041] Specifically, step 3 is repeated N times, resulting in N sample sets, namely S1, S2… N In this embodiment, N = 1000; this setting allows for a larger number of sampling sample sets, and a richer sampling sample set helps to improve the adaptability and robustness of the target threshold obtained subsequently.

[0042] The specific value of K is: K = [k i ], i = 1, 2, 3…1000;

[0043] Where, k i =S pi The sample set is S, S pi S is the sample set obtained by the i-th random sampling. i p-quantiles.

[0044] Step 5: Perform confidence interval analysis on the K value and obtain the estimated threshold; the confidence level of the confidence interval is 95%.

[0045] Step 6: Based on the estimated threshold, evaluate and determine the target threshold.

[0046] Specifically, the estimated threshold is calculated according to the following formula:

[0047]

[0048] Where μ is the estimated threshold, and S td These are the sample mean and sample variance, respectively; t α (n-1) is obtained from the t-distribution table; since the confidence level of the confidence interval is 95%, the corresponding α = 0.05.

[0049] After calculating the estimated threshold using the above formula, the target threshold is further evaluated and determined. The evaluation of the target threshold also considers the actual physical meaning of various vehicle fundamental indicators. For example, if the estimated threshold for a certain highest temperature indicator is confirmed to be 50.8001936℃, then the target threshold is further confirmed to be 50℃. If the estimated threshold for the lowest ternary single-cell voltage indicator is confirmed to be 2.67795851V, then the target threshold is further confirmed to be 2.68V.

[0050] This embodiment provides a threshold analysis method for new energy vehicle control strategies based on Monte Carlo simulation, offering a novel analytical perspective for threshold analysis of new energy vehicles. It provides reliable data references for threshold standardization analysis and new energy vehicle state standardization analysis. This scheme obtains predicted estimated thresholds through Monte Carlo simulation analysis of sample data (including a basic sample set with predetermined thresholds) using random sampling of a specified length. Then, by combining the actual physical meaning of various indicators, a more reliable and objective target threshold that comprehensively considers all predetermined thresholds can be determined. This provides reliable data references for threshold setting and evaluation, helping to improve the authenticity and objectivity of new energy vehicle safety performance assessments.

[0051] This scheme constructs a new sampling sample set by performing multiple random samplings of a specified length on a given sample (basic sample set). This randomness effectively reduces the differences in threshold determination among manufacturers and vehicle models, contributing to an objective evaluation of vehicle fundamental indicators and leading to more realistic safety performance assessments of new energy vehicles. Furthermore, this scheme utilizes the geometric quantity and characteristics of a predetermined threshold, employing mathematical methods to conduct numerous numerical simulations (nearly a thousand times). Based on a probability model and combined with interval estimation, it obtains an approximate solution for the target threshold. This makes the target threshold obtained based on the Monte Carlo simulation method more robust and adaptable, and more valuable for practical reference. Although this scheme obtains the target threshold through threshold estimation, the accuracy of the estimated threshold is high (corresponding to a 95% confidence level). It fully leverages the simplicity of Monte Carlo simulation while maintaining high numerical confirmation accuracy through confidence interval control.

[0052] Example 2:

[0053] The threshold analysis method for control strategies of new energy vehicles based on Monte Carlo simulation adds a preparatory step to the method described in Example 1; the preparatory step is located between step 1 and step 2.

[0054] Specifically, the preparatory steps are as follows: process the collected initial data according to a preset processing strategy.

[0055] The preset processing strategy includes the following sub-steps:

[0056] S1: Classify the collected initial data according to the vehicle's basic indicator type;

[0057] S2: Retrieve all corresponding basic data of vehicle basic indicators that are susceptible to environmental factors; in this embodiment, vehicle basic indicators that are susceptible to environmental factors include temperature (specifically including the highest temperature of the probe sensor, the lowest temperature of the probe sensor, the temperature difference between the highest and lowest temperatures of the probe sensor, the temperature of the drive motor controller, the alarm temperature of the drive motor, the temperature of the DC-DC controller, etc.).

[0058] S3: Configure reliability labels for all basic data; the reliability labels are comprehensively confirmed based on the testing standards and testing environment adopted by the manufacturer providing the basic data (with a predetermined threshold). If the adopted testing standards or testing environment do not conform to the corresponding national standards, the reliability is reduced by 15%. The reliability label is the percentage value of the reliability. Basic data with a reliability below 90% undergoes additional screening.

[0059] An impact label is configured for the basic data retrieved in S2. This impact label is comprehensively confirmed by combining the geographical location of the manufacturer providing the basic data (predetermined threshold) and the time the basic data was determined. In this embodiment, taking temperature as an example, for manufacturers located in high-altitude areas, an impact label is set for their provided predetermined threshold. The difference between their predetermined threshold and the predetermined thresholds provided by other manufacturers located in normal-altitude areas is compared. If the difference is greater than a preset difference range (the lower limit of the preset difference range is set to 10℃), it is determined that the predetermined threshold is affected by altitude (due to altitude, the direct temperature value differs from that at normal altitude; if it is unconverted data, the data differs from the actual temperature value). An impact label is then configured for it, and the predetermined threshold is converted based on the manufacturer's altitude. The impact labels are then removed for predetermined thresholds that are not affected by altitude and for those that have been converted.

[0060] The determination time of the basic data is specifically the determination time of the sample data referenced when confirming the established threshold. If the determination time is concentrated in June-September or January-March, the external environmental temperature has a significant impact during these periods (high temperature or cold weather), which may affect the sample values. For this type of basic data, an impact label is configured, and this type of basic data is re-verified by comparing its established threshold with the established thresholds provided by manufacturers whose determination time is in other time periods. If the difference is greater than the preset difference range (the lower limit of the preset difference range is set to 10℃ here), it is determined that the established threshold is affected by the environment, and the impact label is removed for established thresholds that are not clearly affected by the environment. The established threshold data with the impact label can be selected to participate in S4 according to the actual threshold analysis needs. This setting allows for a more detailed determination of the reliability of the basic data.

[0061] S4: The basic data processed through the above steps will then proceed to step 2.

[0062] This embodiment provides a threshold analysis method for new energy vehicle control strategies based on Monte Carlo simulation. Compared with Embodiment 1, it additionally verifies the basic data, combines the actual evaluation conditions of the basic data provider, and the environmental factors that certain vehicle indicators are susceptible to, to verify the reliability of the basic data itself, and adjusts some affected data, which can effectively improve the overall reliability of the basic data, thereby helping to improve the accuracy of subsequent threshold analysis.

[0063] The above descriptions are merely embodiments of the present invention. Commonly known structures and characteristics of the solutions are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent.

Claims

1. A method for analyzing threshold values of a control strategy of a new energy vehicle based on Monte Carlo simulation, characterized in that, The method comprises the following steps: Step 1: collecting initial data; the initial data is a fixed threshold of a vehicle basic index used by an automobile manufacturer; Step 2: preprocessing the initial data to obtain a basic sample set; Step 3: performing random sampling on the basic sample set for several times to obtain a sample sample set; Step 4: repeatedly performing step 3 by using a Monte Carlo simulation method to obtain a plurality of sample sample sets, and obtaining a K value from the sample sample sets, the K value being a quantile feature vector of the Monte Carlo simulation; Step 5: performing confidence interval analysis on the K value to obtain an estimated threshold; Step 6: determining a target threshold based on the estimated threshold. 2.The new energy vehicle control strategy threshold analysis method based on Monte Carlo simulation according to claim 1, wherein, In step 1, the vehicle basic index includes temperature, single body voltage, inter-single body pressure difference, SOC value and SOC jump percentage. 3.The new energy vehicle control strategy threshold analysis method based on Monte Carlo simulation according to claim 1, wherein, In step 2, the preprocessing operation includes: taking a column vector corresponding to each fixed threshold as a one-dimensional sample set; and performing data cleaning based on the one-dimensional sample set to obtain the basic sample set. 4.The method of analyzing threshold values of a new energy vehicle control strategy based on Monte Carlo simulation according to claim 3, characterized in that, The data cleaning removal objects include null values and non-numeric data.

5. The new energy vehicle control strategy threshold analysis method based on Monte Carlo simulation according to claim 1, characterized in that, In step 3, the number of random samplings is equal to the length of the basic sample set; and the random sampling is equal-length random sampling. 6.The method of analyzing threshold values of a new energy vehicle control strategy based on Monte Carlo simulation according to claim 5, characterized in that, The step 3 is repeated N times, and N sample sets are obtained, respectively ; In step 4, , ; wherein, ; is the p-quantile of the sample set obtained in the i-th random sampling. is the p-quantile of the sample set obtained in the i-th random sampling. 7.The new energy vehicle control strategy threshold analysis method based on Monte Carlo simulation according to claim 1, wherein, In step 5, the confidence level of the confidence interval is 95%. 8.The method of analyzing threshold values of a new energy vehicle control strategy based on Monte Carlo simulation according to claim 7, characterized in that, In step 5, the estimated threshold is calculated according to the sample sample mean and the sample sample variance. In step 5, the confidence level of the confidence interval is 95%. In step 5, the estimated threshold is calculated according to the sample sample mean and the sample sample variance.

Images