A method for quickly determining the operating mode of hybrid vehicles based on three-dimensional convex hull

Abstract

The present invention discloses a method for quickly determining the working mode of a hybrid vehicle based on a three-dimensional convex hull, comprising: obtaining the vehicle required torque, vehicle speed and battery SOC value corresponding to each working point under different driving conditions as the three-dimensional coordinates of the working point, and establishing a working point data set under each working condition; classifying the working point data set in the working condition to be processed according to the working mode, and storing the three-dimensional coordinates of each working point in each working mode to obtain a three-dimensional scatter plot of the working point; performing a three-dimensional convex hull solution on the working points in each working mode to obtain an initial three-dimensional convex hull model corresponding to each working mode; determining the priority of the initial three-dimensional convex hull model corresponding to each working mode, updating the initial three-dimensional convex hull model to obtain a final three-dimensional convex hull model corresponding to each working mode; obtaining the three-dimensional coordinates of the working point under the current driving condition, judging the final three-dimensional convex hull model to which the working point belongs, and determining the working mode of the working point.

Description

Technical Field

[0001] The present invention belongs to the technical field of vehicle energy management, and in particular relates to a method for quickly determining an operating mode of a hybrid vehicle based on a three-dimensional convex hull. Background Art

[0002] Hybrid vehicles can both offset the range limitations of pure electric vehicles and reduce fuel consumption and exhaust emissions compared to conventional vehicles. Developing hybrid vehicle energy management strategies to maximize their advantages is a critical issue. Dynamic programming energy management across a given operating range can calculate theoretical minimum fuel consumption, determine the optimal distribution of vehicle operating mode regions, and other global optimal solutions. However, dynamic programming is computationally intensive and time-consuming, making it difficult to apply to real-time energy management. Therefore, summarizing the patterns in dynamic programming results and applying them to real-time energy management is crucial.

[0003] For hybrid vehicles, the different operating conditions of the motor and engine in the powertrain result in different operating modes. Selecting the operating mode based on information such as vehicle speed, required power, required torque, and battery state of charge (SOC) is a crucial step in energy management. It determines whether the motor and engine can meet the vehicle's actual power requirements while minimizing energy consumption at the end of the entire operating cycle. Existing operating mode switching rules are typically based on two-dimensional information, such as vehicle speed and required power, or speed and required torque. However, operating conditions are diverse, and at some operating points, different operating modes may be selected at the same two-dimensional coordinate values ​​due to varying SOCs, resulting in overlapping operating points in the two-dimensional coordinate plane. Furthermore, the boundaries between operating modes are difficult to uniformly solve and express, requiring different equations or even different methods for solution and expression. Given this, a key challenge is how to rapidly determine the boundaries between operating modes with global energy management rules using a unified approach, represent the switching conditions in a unified manner within the operating mode switching rules, and develop a vehicle operating mode controller for real-time energy management. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull. The three-dimensional convex hull is used to cover the specific existence area of ​​each operating mode, and can adapt to the irregular area in the three-dimensional space formed by the operating points of each mode, so that the boundaries of each operating mode are clearer. There is no need to manually determine different mode switching curve equations for different operating conditions. Only the three-dimensional convex hulls need to be solved uniformly to automatically form the switching conditions of each mode.

[0005] The technical solution provided by the present invention is:

[0006] A method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull comprises the following steps:

[0007] Step 1: Obtain the vehicle required torque, vehicle speed, and battery SOC value corresponding to each operating point under different driving conditions as the three-dimensional coordinates of the operating point, and establish a data set of operating points under each operating condition;

[0008] Step 2: Classify the working points in the working condition to be processed according to the working mode, and store the three-dimensional coordinates of each working point in each working mode to obtain a three-dimensional scatter plot of the working points;

[0009] Step 3: Solve the three-dimensional convex hull for each operating point in each operating mode to obtain an initial three-dimensional convex hull model corresponding to each operating mode;

[0010] Step 4: Determine the priority of the initial three-dimensional convex hull model corresponding to each of the working modes, and update the initial three-dimensional convex hull model in descending order of priority to obtain a final three-dimensional convex hull model corresponding to each of the working modes;

[0011] Step 5: Obtain the three-dimensional coordinates of the operating point under the current driving condition, determine the final three-dimensional convex hull model to which the operating point belongs, and determine the working mode of the operating point;

[0012] The working mode corresponding to the final three-dimensional convex hull model to which the working point belongs is the working mode of the working point.

[0013] Preferably, the working modes include: pure electric driving mode, engine-only driving mode, combined driving mode and driving charging mode.

[0014] Preferably, in step three, the incremental method is used to solve the three-dimensional convex hull to obtain the initial three-dimensional convex hull model, which includes the following steps:

[0015] Step 1: Randomly select four non-coplanar working points under the same working mode to form the first three-dimensional convex hull;

[0016] Step 2: Randomly select a working point pi from the remaining working points in the working mode, and determine whether the first three-dimensional convex hull has an invisible area for the working point pi;

[0017] If the invisible area exists, the operating point pi, each operating point on the horizon, and the invisible area form a second three-dimensional convex hull;

[0018] The horizon is a closed polyline formed by the intersection of the visible area of ​​the first three-dimensional convex hull to the working point pi and the invisible area of ​​the first three-dimensional convex hull to the working point pi;

[0019] If the invisible area does not exist, the operating point pi is ignored;

[0020] Step 3: According to the method in step 2, the remaining operating points in the working mode are judged one by one, and the three-dimensional convex hull is iteratively updated; until all the operating points in the working mode are judged, the initial three-dimensional convex hull model corresponding to the working mode is obtained.

[0021] Preferably, in the step 4, determining the priority of the initial three-dimensional convex hull model comprises the following steps:

[0022] Step A: determining the three-dimensional coordinates of the center of gravity of the initial three-dimensional convex hull model;

[0023] Step B: calculating the electric mode deviation value of the initial three-dimensional convex hull model according to the three-dimensional coordinates of the center of gravity;

[0024]

[0025] Among them, Cx and Cy are the vehicle speed and required torque corresponding to the three-dimensional coordinates of the center of gravity respectively;

[0026] Step C: Prioritize the electric mode deviation values ​​in descending order;

[0027] Among them, the initial three-dimensional convex hull model with the largest electric mode deviation value has the highest priority.

[0028] Preferably, in step A, the method for calculating the three-dimensional coordinates of the center of gravity includes the following steps:

[0029] Step A1: Normalize the three-dimensional coordinate variables of each working point:

[0030]

[0031]

[0032]

[0033] Where V represents vehicle speed, T represents required torque, and SOC represents battery SOC; V max Indicates the maximum vehicle speed, T max Indicates the maximum required torque, SOC max and SOC min Represent the maximum and minimum values ​​of the battery SOC respectively; V' is the value obtained after V normalization, T' is the value obtained after T normalization, and SOC' is the value obtained after SOC normalization;

[0034] Step A2: taking any operating point e on the storage surface of the initial three-dimensional convex hull model, forming a tetrahedron with the operating point e and the storage surface of the initial three-dimensional convex hull model that does not contain the operating point e, and dividing the initial three-dimensional convex hull model into multiple tetrahedrons;

[0035] Step A3: Calculate the three-dimensional coordinates of the center of gravity of the initial three-dimensional convex hull model according to the following formula:

[0036]

[0037]

[0038]

[0039] Among them, Cx, Cy, and Cz are the three-dimensional coordinates of the center of gravity of the initial three-dimensional convex hull model, corresponding to V', T', and SOC' respectively; Cx i , Cy i , Cz i is the three-dimensional coordinate of the center of mass of the i-th tetrahedron, n is the number of tetrahedrons; v i is the volume of the i-th small tetrahedron.

[0040] Preferably, in step A3, the three-dimensional coordinates of the center of gravity of each tetrahedron are calculated by the following formula:

[0041]

[0042]

[0043]

[0044] In the formula, (x i1 ,y i1 , z i1 ), (x i2 ,y i2 , z i2 )、(x i3 ,y i3 , z i3 ) and (x e ,y e , z e ) are the three-dimensional coordinates of the vertices of the i-th tetrahedron.

[0045] Preferably, in step 4, updating the initial three-dimensional convex hull model comprises the following steps:

[0046] Step a: determining an intersection D of operating points of an initial three-dimensional convex hull model with a first priority and an initial three-dimensional convex hull model with a second priority;

[0047] Step b: connecting the center of gravity of the initial three-dimensional convex hull model with the first priority to the vertices of multiple storage surfaces of the initial three-dimensional convex hull model with the first priority to form multiple judgment convex hulls;

[0048] Step c, calculating the number m of operating points of the initial three-dimensional convex hull model with the second priority in each judgment convex hull; and sorting the judgment convex hulls in descending order according to m;

[0049] Step d, deleting the vertices of the storage surface corresponding to the judgment convex hull with the largest m from the initial three-dimensional convex hull model with the first priority, to obtain an updated initial three-dimensional convex hull model with the first priority;

[0050] Step e: determining the operating point intersection E of the updated initial three-dimensional convex hull model with the first priority and the initial three-dimensional convex hull model with the second priority; and calculating the reduction amount and reduction ratio of the operating point of the intersection E relative to the intersection D. If the update end condition is met, the update of the initial three-dimensional convex hull model with the first priority is completed;

[0051] The update end condition is as follows: the reduction amount of the operating point is above the reduction amount threshold, and the operating point ratio is below the ratio threshold;

[0052] Step f: Update other initial three-dimensional convex hull models in sequence according to steps a to d until all initial three-dimensional convex hull models are updated.

[0053] Preferably, the method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull further includes:

[0054] In the step e, if the reduction amount of the operating point is less than the reduction amount threshold, the vertex of the storage surface corresponding to the judgment convex hull with the largest m is added back to the initial three-dimensional convex hull model with the first priority;

[0055] The second-ranked judgment convex hull is processed according to steps d and e. If the update end condition is met, the update of the initial three-dimensional convex hull model with the first priority is completed; otherwise, the judgment convex hulls are sorted and the other judgment convex hulls are processed one by one according to steps d and e until the update end condition is met.

[0056] Preferably, in step a, the convex hulls of the operating points are determined one by one, and the operating points belonging to the initial three-dimensional convex hull model with the first priority and the initial three-dimensional convex hull model with the second priority form an intersection D;

[0057] The method for determining whether a working point belongs to a certain initial three-dimensional convex hull model is as follows:

[0058] The volume of the initial three-dimensional convex hull model is calculated, the point to be determined is added to the set of operating points constituting the initial three-dimensional convex hull model, and the initial three-dimensional convex hull model and its volume are re-determined; if the volumes of the initial three-dimensional convex hull models calculated twice are equal, the point to be determined belongs to the initial three-dimensional convex hull model; otherwise, the point to be determined does not belong to the initial three-dimensional convex hull model.

[0059] Preferably, the method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull further comprises: if a discrete point exists in a certain operating mode, ignoring the discrete point when constructing the initial three-dimensional convex hull model;

[0060] The discrete points are operating points that are far away from the center of the initial three-dimensional convex hull model corresponding to the working mode and close to the initial three-dimensional convex hull models corresponding to other working modes.

[0061] The beneficial effects of the present invention are as follows: the method provided by the present invention for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull divides the data obtained after dynamic programming according to factors in three dimensions, making the data classification more detailed and accurate, and avoiding partial data coverage caused by different operating points falling on the same position in a two-dimensional plane; the three-dimensional convex hull is used to cover the specific existence area of ​​each operating mode, and the three-dimensional convex hull is an irregular shape, which can adapt to the irregular area in the three-dimensional space formed by the operating points of each mode; there is no need to manually determine different mode switching curve equations for different operating conditions, and only the three-dimensional convex hulls need to be solved uniformly to automatically form the switching conditions of each mode. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a flow chart of the method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to the present invention.

[0063] Figure 2 This is a schematic diagram of the results of classifying the data obtained by dynamic programming in the low-speed zone of the WLTC operating condition according to the present invention according to the working mode.

[0064] Figure 3 This is a schematic diagram of the results of classifying the data obtained by the dynamic programming of the WLTC complete working condition according to the present invention according to the working mode.

[0065] Figure 4 Schematic diagram of the visible and invisible areas of the three-dimensional convex hull according to the present invention.

[0066] Figure 5 This is a flowchart of determining the convex hull priority according to the present invention.

[0067] Figure 6 Schematic diagram of the optimized relationship between the three-dimensional convex hulls of the present invention.

[0068] Figure 7 Schematic diagram of the three-dimensional convex hull optimization and updating method according to the present invention. DETAILED DESCRIPTION

[0069] The present invention will be described in further detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0070] The present invention provides a method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull, which is applicable to traditional hybrid vehicles and plug-in hybrid vehicles; the data obtained after dynamic programming is divided according to factors in three dimensions, so that the classification of the data is more detailed and accurate, and partial data coverage caused by different operating points falling on the same position of the two-dimensional plane is avoided; the specific existence area of ​​each operating mode is covered by the three-dimensional convex hull, and the three-dimensional convex hull is an irregular shape, which can adapt to the irregular area in the three-dimensional space formed by the operating points of each mode; there is no need to manually determine different mode switching curve equations for different operating conditions, and each mode switching condition can be automatically formed by uniformly solving each three-dimensional convex hull. The operating conditions applicable to the present invention include but are not limited to NEDC (New European Driving Cycle), CLTC (China Light Vehicle Test Cycle), EPA (Environmental Protection Agency), and WLTC operating conditions specified by WLTP (World Light Vehicle Test Procedure).

[0071] like Figure 1 As shown, the method for quickly determining a vehicle's operating mode based on a three-dimensional convex hull includes three parts: classifying operating points according to operating mode, solving the three-dimensional convex hull, and establishing a vehicle operating mode controller. The main process is as follows: classifying operating points according to operating mode and storing the three-dimensional data of each operating point to lay the foundation for subsequent regularity extraction; solving the three-dimensional convex hull for the area where each operating mode is located, including all irregular areas, and automatically forming the boundaries of each mode; the three-dimensional convex hulls formed by each operating mode have repeated intersections, and each three-dimensional convex hull is optimized and modified, and the geometric feature parameters of each convex hull are calculated. The priority of the convex hull modification calculation is determined based on the geometric feature parameters, and the convex hull is updated in descending order to make the boundaries between the operating modes clearer. The operating mode switching rules are determined, and the vehicle operating mode controller is established.

[0072] Dividing data based on three-dimensional factors makes the classification more detailed and accurate, avoiding data overlap caused by different working points falling on the same position in the two-dimensional plane; using the three-dimensional convex hull to cover the specific existence area of ​​each working mode, the three-dimensional convex hull is an irregular shape that can adapt to the irregular area in the three-dimensional space formed by the working points of each mode; there is no need to manually determine different mode switching curve equations for different working conditions, only the three-dimensional convex hull needs to be solved to automatically form the switching conditions of each mode, saving labor costs; it can quickly determine whether the point is within the three-dimensional convex hull, ensuring the real-time determination of the working mode during actual energy management.

[0073] The specific implementation process of the method for quickly determining the vehicle operating mode based on the three-dimensional convex hull is as follows:

[0074] 1. Classify working points according to working mode

[0075] The vehicle speed corresponding to each operating point under different driving conditions is acquired through sensors. The required speed at each operating point is calculated based on the vehicle speed, and the battery state of charge (SOC) value is determined as the three-dimensional coordinates of the operating point. A data set of operating points for each operating condition is then established. The operating point data set for the condition being processed is categorized by operating mode, and the three-dimensional coordinates of each operating point in each operating mode are stored to generate a three-dimensional scatter plot of the operating points.

[0076] As a preferred approach, dynamic programming energy management is used to determine the battery SOC value. Dynamic programming is a global optimization method that determines the minimum energy consumption of a vehicle over a given operating cycle, as well as the optimal sequence of variables for achieving this minimum energy consumption. Applying the resulting patterns to formulate energy management strategies can ensure that vehicle operating behavior and ultimate fuel consumption are aligned with the globally optimized results.

[0077] In one embodiment, the dynamic programming process is as follows:

[0078] Knowing the vehicle speed per second, the transmission input torque and input shaft speed can be calculated through the driving resistance model and transmission system model.

[0079] (1) Variable selection

[0080] The battery SOC and gear state g are state variables, and the ratio of motor and transmission input torque q and gear change d are control variables. That is, x = [SOC, g] T ,u=[q,d] T .

[0081] (2) Objective function establishment

[0082] The objective function is fuel consumption J. The optimization goal is to minimize J. The calculation formula of J is as follows:

[0083]

[0084] Where L is the fuel consumption in the kth stage.

[0085] (3) Establishment of state transfer equation

[0086] The state transfer equation establishes the relationship between the state variables and the control variables. The formula is as follows:

[0087]

[0088] In the above formula, T m is the motor torque; n m is the motor speed.

[0089] It should be noted that the motor speed n m It is related to the vehicle configuration and can be calculated. Taking a hybrid vehicle with P2 configuration as an example, its motor speed is equal to the known transmission input shaft speed when the motor is working.

[0090] (4) Reverse recursion process

[0091] Discretize a complete operating cycle into N stages. Set the final value of the state variable SOC to 0.3 and the final value of g to the gear corresponding to the parking state. Starting at time k = N - 1, perform a recursive calculation backwards to obtain the optimal control variables, optimal state variables, and optimal target values ​​for each stage, until k = 0, when the recursive calculation ends.

[0092] (5) Forward optimization process

[0093] The SOC starting value is set to 0.8, and the initial value of g is the gear corresponding to the parking state. According to the optimal target value obtained in (3) and its corresponding optimal control variables in each stage, the optimal state SOC trajectory can be obtained. Specifically: According to the selected control variables q and d in each stage, the motor output torque T in each stage can be calculated. m Since the initial SOC is known, the optimal sequence value of SOC can be obtained by calculation (3).

[0094] So far, the results of dynamic programming energy management have been obtained.

[0095] Dynamic programming energy management is carried out for the working conditions to be processed. The results are classified according to the working modes, and the three-dimensional information of each working condition point is stored to obtain a three-dimensional scatter plot of the working condition points. Taking a hybrid vehicle in P2 configuration as an example for the specific classification, its working modes are usually divided into the following five types: pure electric drive mode, engine single drive mode, combined drive mode, driving and charging mode, and braking energy recovery mode. When the motor works and the engine does not work, it is the pure electric drive mode at this time; when the motor does not work and the engine works, it is the engine single drive mode at this time; when the motor and the engine work simultaneously, it is the combined drive mode at this time; when the engine provides the driving energy for the whole vehicle and charges the battery at the same time, and the motor generates electricity, it is the driving and charging mode at this time; when the engine does not work and the motor generates electricity, it is the braking energy recovery mode at this time. The braking energy recovery mode generally occurs when the driver brakes, so the working area of this mode is not considered in the present invention.

[0096] According to the optimal sequence of the ratio q of the motor torque to the transmission input torque obtained by dynamic programming energy management, the working mode category of each working condition point can be determined. q = 0 belongs to the engine single drive mode; q = 1 belongs to the pure electric drive mode; 0 < q < 1 belongs to the combined drive mode; q < 0 belongs to the driving and charging mode or the braking energy recovery mode. To further determine the mode when q < 0, it can be judged according to the transmission input torque. When the transmission input torque is negative, it belongs to the braking energy recovery mode; when the transmission input torque is positive, it belongs to the driving and charging mode.

[0097] The data is classified according to the above four working modes, and the three-dimensional information of each working condition point under each mode is stored, including the vehicle demand torque T, vehicle speed V, and battery SOC value, and these are used as the three-dimensional coordinates of the working condition points.

[0098] Different cycle working condition characteristics, different parameter settings, etc. will all lead to diverse dynamic programming results. Therefore, there must be some working conditions where when different working condition points fall at the same position on the two-dimensional plane, it will cause partial data coverage. For example, the low-speed area of the WLTC working condition is repeated twice to increase the working condition point data, and dynamic programming energy management is carried out on it, and the obtained results are classified. The final classification results are as Figure 2 shown. The entire cycle working condition is in a low-speed state, which usually occurs in the case of urban congestion. At this time, the vehicle power demand is always in a small state, and the combined drive mode does not appear. From Figure 2 it can be seen that in the three-dimensional coordinates, for the part with a larger speed and a larger demand torque, when the SOC is higher, it is the scatter point corresponding to the driving and charging mode, and when the SOC is lower, it is the scatter point corresponding to the pure electric drive mode. Since there is a position with a higher SOC value in the driving and charging mode, it usually covers the working condition scatter points of the pure electric drive mode, resulting in partial information loss, which shows the importance and necessity of the three-dimensional information.

[0099] Since the complete WLTC operating condition includes not only the low-speed zone, but also the medium-speed zone, the high-speed zone, and the ultra-high-speed zone, a combined driving mode appears in the dynamic programming result. Figure 3 As shown, the data obtained by dynamic planning of the complete WLTC working condition is classified according to the working mode.

[0100] 2.3D convex hull solution

[0101] A three-dimensional convex hull was calculated for the data of the pure electric drive mode, engine-only drive mode, on-the-road charging mode, and combined drive mode to represent the operating range and boundaries of each mode. If a particular mode had fewer than four data points, a three-dimensional convex hull could not be formed. Since the small number of points indicates a low probability of occurrence, this mode was not considered. For example, the WLTC operating condition results show that the engine drive mode has three operating points, so the relevant issues of this mode are not considered here.

[0102] As a preferred method, the incremental method is used to solve the three-dimensional convex hull to improve the accuracy of the solution. The idea of ​​using the incremental method to solve the three-dimensional convex hull is: first, use four non-coplanar points to form a convex hull, and store the three points of each face in the plane set H in turn. Then judge the remaining points in turn, regard the judgment point as a light source, and the light source shines on the original convex hull. When there is an area that cannot be illuminated, that is, an invisible area, it means that the judgment point does not belong to the original convex hull. At this time, it needs to be added to the original convex hull to form a new convex hull; when the entire original convex hull is within the illumination range of the judgment point, that is, it is all in the visible area, then the point exists in the original convex hull, and the point is ignored. If Figure 4 The figure shows a schematic diagram of the visible and invisible areas of the three-dimensional convex hull. The closed polyline enclosed by the visible and invisible areas of the three-dimensional convex hull is called the horizon. The entire horizon is composed of some points of the original convex hull.

[0103] Assume that the set of n points to be found is P, and the specific algorithm process of the incremental method is as follows:

[0104] (1) Select four non-coplanar points in the point set P, denoted as p1, p2, p3, and p4, and connect them in sequence to construct the convex hull.

[0105] (2) Randomly select a point pi from the remaining points in the point set P and determine whether there is an invisible area of ​​the original convex hull for pi. If so, then

[0106] Pi and each convex hull point on the horizon form a partial plane of the new convex hull, and form a new convex hull with the invisible area; otherwise, the point is ignored.

[0107] (3) Repeat (2) until the entire point set P is judged.

[0108] Through the above process, the three-dimensional convex hull of each working mode is solved. The storage mode of the three-dimensional convex hull is an m×3 matrix composed of the sequence numbers of the three points of the m convex hull planes.

[0109] 3. Establishing the vehicle working mode controller

[0110] Solving the 3D convex hull for each working mode means enclosing all points of each working mode in several closed 3D figures. However, the distribution of working points is not completely regular. There are often some discrete points far away from the center point of the mode they belong to, but close to the center point of another working mode. This will cause the 3D convex hull of the working mode to which they originally belong to be larger, and contain more points of the other working mode, and the two modes have a large overlap. For example Figure 3 The discrete point marked in the figure belongs to the driving charging mode. There are many pure electric driving mode operating points in the region between the nearest operating point in the same mode. When solving the 3D convex hull for the driving charging mode operating points, since all driving charging mode operating points must be included, all pure electric driving mode operating points in the region between them will also be included. To minimize errors, this discrete point needs to be ignored when solving the 3D convex hull for the driving charging mode. The next step is to identify and remove similar points as much as possible to optimize the convex hull shape.

[0111] Convex hull shape optimization involves optimizing the relationship between the two convex hulls, reducing overlap and minimizing the number of points in each convex hull representing other operating modes. To this end, optimization of one convex hull must be performed relative to the other. Once the shape of one convex hull is determined, the other convex hulls no longer need to be optimized relative to the optimized convex hull. Therefore, convex hull shape optimization should follow a specific order to determine which convex hull to optimize first. The method chosen here prioritizes the convex hulls to be optimized based on a metric. The final three-dimensional convex hulls are updated and saved sequentially, and used to determine the operating mode in real time.

[0112] Before optimizing the convex hull shape, it is also necessary to determine the priority of the convex hull. As a preferred method, parameters related to the center of gravity of the convex hull are selected as the basis for determining the priority of the convex hull. Figure 5 As shown in , the process of determining the convex hull priority includes:

[0113] 1. Calculate the geometric feature parameters of the convex hull

[0114] After solving the three-dimensional convex hull for each operating mode and determining the convex hull boundary, since the dimensions of the three-dimensional coordinate axis variables (vehicle speed V, required torque T, and battery SOC) are very different, each variable must be normalized before calculating the center of gravity to eliminate the impact of the dimension on subsequent work. The process of calculating the center of gravity is as follows:

[0115] (1) Normalization of three-dimensional coordinate values

[0116] With reference to the working condition under study, the minimum vehicle speed is 0. Since the braking energy recovery mode is not considered, the minimum required torque is also 0. Select the maximum vehicle speed V max , Maximum required torque T max And the maximum value of battery SOC SOC max , minimum SOC min As a normalization standard, all three-dimensional data of the operating point are limited to 0 to 1. Normalize each variable:

[0117]

[0118]

[0119]

[0120] Wherein, V' is the value obtained after normalization of V, T' is the value obtained after normalization of T, and SOC' is the value obtained after normalization of SOC.

[0121] (2) Take any point in the convex hull

[0122] Before calculating the volume, we need to select a point within the convex hull and form several smaller convex hulls with the faces of the convex hull. If the point is outside the convex hull, the calculated volume will be larger than the actual volume. To ensure that the selected point is within the convex hull, we select a point e from the convex hull, that is, any point in the m×3 matrix of the three-dimensional convex hull storage surface.

[0123] (3) Find the volume of the three-dimensional convex hull

[0124] Delete the plane containing point e from the m planes. The remaining planes form small convex hulls with point e, that is, tetrahedrons. Calculate the volume of each of them.

[0125] The volume of a three-dimensional convex hull is equal to the sum of the volumes of all the small convex hulls that make it up. Therefore, the formula for finding the volume of a three-dimensional convex hull is as follows:

[0126]

[0127] Where v is the volume of the three-dimensional convex hull, n is the number of all small convex hulls that constitute the three-dimensional convex hull, and v i is the volume of the i-th small convex hull.

[0128] (4) Find the centroid of the three-dimensional convex hull

[0129] According to (3), the three-dimensional convex hull has been divided into several convex tetrahedrons. The centroid of each small tetrahedron is calculated as follows:

[0130]

[0131]

[0132]

[0133] Where Cx i , Cy i , Cz i is the three-dimensional coordinate of the center of mass of the i-th tetrahedron, x i1 、x i2 、x i3 are the x-axis coordinates of the three vertices of the i-th convex hull, y i1 、y i2 、y i3 are the coordinates of the three vertices of the i-th convex hull in the y-axis direction, z i1 、z i2 、z i3 are the coordinates of the three vertices of the i-th convex hull in the z-axis direction, (x e 、y e 、z e ) are the three-dimensional coordinates of point e.

[0134] The final three-dimensional convex hull centroid calculation formula is as follows:

[0135]

[0136]

[0137]

[0138] Where Cx, Cy, and Cz are the three-dimensional coordinates of the centroid of the final three-dimensional convex hull.

[0139] Based on the above process, the centroid coordinates of the three-dimensional convex hull can be calculated. Next, we need to calculate the priority index based on the centroid coordinates.

[0140] 2. Determine the convex hull priority

[0141] According to the centroid of the three-dimensional convex hull, the formula for calculating the sum of the squares of the two-dimensional coordinates of the centroid r is as follows:

[0142]

[0143] The reason for choosing r as the indicator is that Cx and Cy are the vehicle speed V and the required torque T respectively. Usually, the one with the smallest r value is the pure electric mode, which has the most points and the widest coverage. Therefore, it is easy for the obtained convex hull to surround part of the convex hull of other modes. If you want to reduce the points that contain other modes through optimization, you will lose a lot of pure electric mode working condition point data. Therefore, consider sorting the convex hull in order of r value from large to small, and optimize the convex hull in turn. For example, the convex hulls of the three working modes A, B, and C are represented by the design. The r obtained after calculation is a >r b >r c , then the convex hull is sorted by priority as A>B>C.

[0144] 3. Update the convex hull

[0145] After determining the convex hull priority, the convex hull update optimization should be performed based on the convex hull priority. Taking the convex hulls A, B, and C in step 2 as an example, the optimization relationship between the convex hulls is as follows: Figure 6 As shown in the figure, first A is optimized with respect to B and C respectively. Then, A is kept fixed and B is optimized with respect to C, and the optimization is completed.

[0146] Taking A optimizing B as an example, the specific optimization update method is as follows: Figure 7 As shown, it contains the following contents:

[0147] (1) Find the point set D where A and B intersect.

[0148] Convex hull A corresponds to the number of scattered points in the working mode (the points that make up convex hull A and the scattered points contained in it) that exist in convex hull B. Convex hull B corresponds to the number of scattered points in the working mode (the points that make up convex hull B and the scattered points contained in it) that exist in convex hull A. The two constitute point set D. The method for determining whether a point is in a specified convex hull is as follows:

[0149] Find the volume of the specified convex hull, the result is v. Add the point to be determined to the set of points that constitute the specified convex hull, and re-solve the three-dimensional convex hull and its volume, the result is v'. If v = v', the point to be determined is in the specified convex hull; otherwise, it is not. Using this method, we can obtain the point set D.

[0150] (2) The centroid C1 of the convex hull A and the storage surfaces of the convex hull A in the point set D (referring to the surface contour of the convex hull) form a tetrahedron F.

[0151] The centroid of the convex hull A is obtained according to the method in step 1 and is recorded as C1. C1 and the storage surfaces of the convex hull A in the point set D respectively form a tetrahedron F.

[0152] (3) Calculate the number of scattered points in the working mode corresponding to the convex hull B in each tetrahedron F (the points that constitute the convex hull B and the scattered points contained in it). Here, the method of determining whether a point is in the specified convex hull in (1) is used to calculate the number of scattered points in the working mode corresponding to the convex hull B in each tetrahedron F (the points that constitute the convex hull B and the scattered points contained in it).

[0153] (4) Sort the tetrahedron F from high to low according to the number of points, and record it as F1>F2>…>Fn.

[0154] (5) Delete the points on the storage surface A in tetrahedron F1.

[0155] (6) Resolve the three-dimensional convex hull A.

[0156] The three-dimensional convex hull A is re-solved for the point set after the point is deleted, and the working mode point set contained in the convex hull A is also updated accordingly.

[0157] (7) Determine whether the number of scattered points in the working mode corresponding to the convex hull A and the convex hull B (the points constituting the convex hull B and the scattered points contained therein) is reduced by K1.

[0158] K1 needs to be set manually based on a lot of experience. If the number of points reduced is less than K1, go to (8); if the number of points reduced reaches K1, go to (9).

[0159] (8) Re-add the stored pastry in A in F1 and go to (10).

[0160] (9) Continue to determine whether the ratio of the number of scattered points in A corresponding to the working mode of the convex hull B (the points constituting the convex hull B and the scattered points contained therein) is less than or equal to K2.

[0161] K2 needs to be set manually based on a lot of experience. If the ratio is greater than K2, go to (10); if the ratio is less than or equal to K2, go to (11).

[0162] (10) F2 is upgraded to F1, F3 is upgraded to F2, and so on. Then the optimization is continued from (5).

[0163] (11) The convex hull A is updated to A', and the optimization update of the convex hull A is completed.

[0164] After the convex hull A is updated, the convex hull B is optimized and updated according to the above method. After the convex hull B is optimized and updated, the remaining convex hull C is also optimized and updated, and the entire optimization and update process is completed.

[0165] 4. Determine the rules for judging working modes

[0166] Once the three-dimensional convex hulls of each operating mode are updated and determined, the basic work of establishing the controller is completed. It is necessary to further determine the application rules of the controller, that is, the rules for real-time determination of the operating mode.

[0167] The controller's role is to input the three-dimensional information of the operating point, including the vehicle's required torque T, vehicle speed V, and battery SOC, sequentially determine whether the operating point is within each convex hull, and quickly output the operating mode to which the operating point belongs. The order in which the controller determines the positional relationship between the operating point and the convex hulls should be consistent with the convex hull priority order determined in step 2. This is because a high-priority convex hull is often partially surrounded by a lower-priority convex hull. Therefore, the operating point should first be determined to be within the higher-priority convex hull. If not, the controller should then determine whether it is within the lower-priority convex hull.

[0168] Taking convex hulls A, B, and C (priority: A>B>C) as an example, the controller sequentially determines whether the input point is within these convex hulls. Once the input point is determined to be within a convex hull, the controller stops determining the operating mode corresponding to that convex hull and directly outputs the operating mode corresponding to that convex hull. Because input points may exist outside all convex hulls, a supplementary operating mode determination module is required to determine the operating mode corresponding to operating points outside the convex hulls.

[0169] The supplementary module stipulates that when the operating point is input into the controller and the input point is judged to be outside the convex hull area, the engine is kept operating along the optimal fuel curve. In this way, the operating mode can be determined for each input point.

[0170] In order to make the input points as much as possible within the convex hull, the number of working condition points participating in the dynamic programming working condition cycle can be appropriately increased to increase the coverage of each convex hull.

[0171] The working mode controller formed based on the above method can quickly determine the working mode of the input point, ensure the real-time performance of actual energy management, and improve the global suboptimality of real-time energy management.

[0172] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the description and implementation methods. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull, characterized in that: The steps include: Step 1: Obtain the vehicle required torque, vehicle speed, and battery SOC value corresponding to each operating point under different driving conditions as the three-dimensional coordinates of the operating point, and establish a data set of operating points under each operating condition; Step 2: Classify the working point data set in the working condition to be processed according to the working mode, and store the three-dimensional coordinates of each working point in each working mode to obtain a three-dimensional scatter plot of the working points; Step 3: Solve the three-dimensional convex hull for each operating point in each operating mode to obtain an initial three-dimensional convex hull model corresponding to each operating mode; Step 4: Determine the priority of the initial three-dimensional convex hull model corresponding to each of the working modes, and update the initial three-dimensional convex hull model in descending order of priority to obtain a final three-dimensional convex hull model corresponding to each of the working modes; Step 5: Obtain the three-dimensional coordinates of the operating point under the current driving condition, determine the final three-dimensional convex hull model to which the operating point belongs, and determine the working mode of the operating point; The working mode corresponding to the final three-dimensional convex hull model to which the working point belongs is the working mode of the working point.

2. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 1, characterized in that: The working modes include: pure electric driving mode, engine-only driving mode, combined driving mode and driving charging mode.

3. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 2, characterized in that: In step three, the incremental method is used to solve the three-dimensional convex hull to obtain the initial three-dimensional convex hull model, which includes the following steps: Step 1: Randomly select four non-coplanar working points under the same working mode to form the first three-dimensional convex hull; Step 2: Randomly select a working point pi from the remaining working points in the working mode, and determine whether the first three-dimensional convex hull has an invisible area for the working point pi; If the invisible area exists, the operating point pi, each operating point on the horizon, and the invisible area form a second three-dimensional convex hull; The horizon is a closed polyline formed by the intersection of the visible area of ​​the first three-dimensional convex hull to the working point pi and the invisible area of ​​the first three-dimensional convex hull to the working point pi; If the invisible area does not exist, the operating point pi is ignored; Step 3: According to the method in step 2, the remaining operating points in the working mode are judged one by one, and the three-dimensional convex hull is iteratively updated; until all the operating points in the working mode are judged, the initial three-dimensional convex hull model corresponding to the working mode is obtained.

4. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 3, characterized in that: In the step 4, determining the priority of the initial three-dimensional convex hull model includes the following steps: Step A: determining the three-dimensional coordinates of the center of gravity of the initial three-dimensional convex hull model; Step B: calculating the electric mode deviation value of the initial three-dimensional convex hull model according to the three-dimensional coordinates of the center of gravity; Among them, Cx and Cy are the vehicle speed and required torque corresponding to the three-dimensional coordinates of the center of gravity respectively; Step C: Prioritize the electric mode deviation values ​​in descending order; Among them, the initial three-dimensional convex hull model with the largest electric mode deviation value has the highest priority.

5. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 4, characterized in that: In step A, the method for calculating the three-dimensional coordinates of the center of gravity includes the following steps: Step A1: Normalize the three-dimensional coordinate variables of each working point: Where V represents vehicle speed, T represents required torque, and SOC represents battery SOC; V max Indicates the maximum vehicle speed, T max Indicates the maximum required torque, SOC max and SOC min Represent the maximum and minimum values ​​of the battery SOC respectively; V' is the value obtained after V normalization, T' is the value obtained after T normalization, and SOC' is the value obtained after SOC normalization; Step A2: taking any operating point e on the storage surface of the initial three-dimensional convex hull model, forming a tetrahedron with the operating point e and the storage surface of the initial three-dimensional convex hull model that does not contain the operating point e, and dividing the initial three-dimensional convex hull model into multiple tetrahedrons; Step A3: Calculate the three-dimensional coordinates of the center of gravity of the initial three-dimensional convex hull model according to the following formula: Among them, Cx, Cy, and Cz are the three-dimensional coordinates of the center of gravity of the initial three-dimensional convex hull model, corresponding to V', T', and SOC' respectively; Cx i , Cy i , Cz i is the three-dimensional coordinate of the center of mass of the i-th tetrahedron, n is the number of tetrahedrons; v i is the volume of the i-th tetrahedron.

6. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 5, characterized in that: In step A3, the three-dimensional coordinates of the center of gravity of each tetrahedron are calculated using the following formula: In the formula, (x i1 ,y i1 , z i1 ), (x i2 ,y i2 , z i2 )、(x i3 ,y i3 , z i3 ) and (x e ,y e , z e ) are the three-dimensional coordinates of the vertices of the i-th tetrahedron.

7. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 4, 5 or 6, characterized in that: In the step 4, the initial three-dimensional convex hull model is updated, including the following steps: Step a: determining an intersection D of operating points of an initial three-dimensional convex hull model with a first priority and an initial three-dimensional convex hull model with a second priority; Step b: connecting the center of gravity of the initial three-dimensional convex hull model with the first priority to the vertices of multiple storage surfaces of the initial three-dimensional convex hull model with the first priority to form multiple judgment convex hulls; Step c, calculating the number m of operating points of the initial three-dimensional convex hull model with the second priority in each judgment convex hull; and sorting the judgment convex hulls in descending order according to m; Step d, deleting the vertices of the storage surface corresponding to the judgment convex hull with the largest m from the initial three-dimensional convex hull model with the first priority, to obtain an updated initial three-dimensional convex hull model with the first priority; Step e: determining the operating point intersection E of the updated initial three-dimensional convex hull model with the first priority and the initial three-dimensional convex hull model with the second priority; and calculating the reduction amount and reduction ratio of the operating point of the intersection E relative to the intersection D. If the update end condition is met, the update of the initial three-dimensional convex hull model with the first priority is completed; The update end condition is as follows: the reduction amount of the operating point is above the reduction amount threshold, and the reduction ratio of the operating point is below the ratio threshold; Step f: Update other initial three-dimensional convex hull models in sequence according to steps a to d until all initial three-dimensional convex hull models are updated.

8. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 7, characterized in that: Also includes: In the step e, if the reduction amount of the operating point is less than the reduction amount threshold, the vertex of the storage surface corresponding to the judgment convex hull with the largest m is added back to the initial three-dimensional convex hull model with the first priority; The second-ranked judgment convex hull is processed according to steps d and e. If the update end condition is met, the update of the initial three-dimensional convex hull model with the first priority is completed; otherwise, the judgment convex hulls are sorted and the other judgment convex hulls are processed one by one according to steps d and e until the update end condition is met.

9. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 8, characterized in that: In step a, the convex hulls of the operating points are determined one by one, and the operating points belonging to the initial three-dimensional convex hull model with the first priority and the initial three-dimensional convex hull model with the second priority form an intersection D; The method for determining whether a working point belongs to a certain initial three-dimensional convex hull model is as follows: The volume of the initial three-dimensional convex hull model is calculated, the point to be determined is added to the set of operating points constituting the initial three-dimensional convex hull model, and the initial three-dimensional convex hull model and its volume are re-determined; if the volumes of the initial three-dimensional convex hull models calculated twice are equal, the point to be determined belongs to the initial three-dimensional convex hull model; otherwise, the point to be determined does not belong to the initial three-dimensional convex hull model.

10. The method for quickly determining the operating mode of a hybrid vehicle based on a three-dimensional convex hull according to claim 9, characterized in that: Also includes: If there are discrete points in a certain working mode, these discrete points will be ignored when constructing the initial 3D convex hull model; The discrete points are operating points that are far away from the center of the initial three-dimensional convex hull model corresponding to the working mode and close to the initial three-dimensional convex hull models corresponding to other working modes.

Images