DE102021003990A1 - Method for determining a driving power distribution - Google Patents

Abstract

The invention relates to a method for determining drive energy distributions, a computer program product for carrying out the method and a hybrid vehicle. The method for determining a drive energy distribution comprises the steps:S1: providing a multiplicity I of particles i, each particle i- a first particle component K1i ( 2) which describes a first randomly selected operating point in a first characteristics map (3) of a first drive component (4) of a drive train by means of a first parameter n1i and a second parameter τ1i, - a second particle component K2i (5) which describes a second randomly selected describes the operating point in a second characteristic map (6) of a second drive component (7) of the drive train by means of a third parameter n2i and a fourth parameter τ2i,- and comprises a particle quality parameter wi;S2: forming a first collection of first particle components K1i (2) in the first Map (3) and a z Wide accumulation of second particle components K2i (5) in the second characteristic map (6) by repeated shifting of the first and second particle components K1i, K2i (2, 5) by means of a particle swarm optimization (Particle Swarm Optimization; PSO) based on the particle quality parameters wi of the particles i;S3: determining a drive energy distribution by evaluating the first and second collection of the first and second particle components K1i (2), K2i (5).

Description

The invention relates to a method for determining drive energy distributions for a drive train of a hybrid vehicle, a computer program product for carrying out the method and a hybrid vehicle.

From the CN104408257A a method for optimizing motor vehicle parameters of hybrid vehicles using a particle swarm algorithm is known.

• S1: Bereitstellen einer Vielzahl I von Partikeln i, wobei jeder Partikel i
  • - eine erste Partikelkomponente K1i, die einen ersten zufällig ausgewählten Betriebspunkt in einem ersten Kennfeld einer ersten Antriebskomponente eines Antriebsstrangs mittels eines ersten Parameters n1i und eines zweiten Parameters τ1i beschreibt,
  • - eine zweite Partikelkomponente K2i, die einen zweiten zufällig ausgewählten Betriebspunkt in einem zweiten Kennfeld einer zweiten Antriebskomponente des Antriebsstrangs mittels eines dritten Parameters n2i und eines vierten Parameters τ2i beschreibt, wobei sich der erste Parameter n1i und der dritte Parameter n2i zu einem ersten vorgegebenen Fahrparameter n zusammensetzen und/oder wobei sich der zweite Parameter τ1i und der vierte Parameter τ2i zu einem zweiten vorgegebenen Fahrparameter τ zusammensetzen,
  • - und einen Partikelgüteparameter wi umfasst;
• S2: Bilden einer ersten Ansammlung von ersten Partikelkomponenten K1 i im ersten Kennfeld und einer zweiten Ansammlung von zweiten Partikelkomponenten K2i im zweiten Kennfeld durch wiederholtes Verschieben der ersten und zweiten Partikelkomponenten K1i, K2i mittels einer Partikelschwarmoptimierung (Particle Swarm Optimization; PSO) anhand der Partikelgüteparameter wi der Partikel i;
• S3: Bestimmen einer ersten Antriebsenergieverteilung V1 i durch Auswertung der ersten Ansammlung der ersten Partikelkomponenten K1i und zweiten Ansammlung der zweiten Partikelkomponenten K2i mittels einer Wahrscheinlichkeitsdichtefunktion.

The method according to the invention for determining a drive energy distribution for a drive train of a hybrid vehicle comprises the steps:

• S1: Providing a multiplicity I of particles i, each particle i
  • - a first particle component K1i, which describes a first randomly selected operating point in a first characteristic map of a first drive component of a drive train by means of a first parameter n1i and a second parameter τ1i,
  • - A second particle component K2i, which describes a second randomly selected operating point in a second map of a second drive component of the drive train by means of a third parameter n2i and a fourth parameter τ2i, with the first parameter n1i and the third parameter n2i becoming a first specified driving parameter n combine and/or wherein the second parameter τ1i and the fourth parameter τ2i combine to form a second specified driving parameter τ,
  • - and a particle quality parameter wi;
• S2: forming a first collection of first particle components K1 i in the first map and a second collection of second particle components K2i in the second map by repeatedly shifting the first and second particle components K1i, K2i by means of a particle swarm optimization (Particle Swarm Optimization; PSO) based on the particle quality parameters wi the particle i;
• S3: Determining a first drive energy distribution V1 i by evaluating the first collection of the first particle components K1i and the second collection of the second particle components K2i using a probability density function.

The random selection of operating points also includes a selection based on computer-generated random numbers.

“Drive components” are understood here to mean components used to drive vehicles, such as internal combustion engines, electric motors, fuel cells or traction batteries.

The term “hybrid vehicle” is understood to mean a vehicle that can be operated using at least two drive components. This includes electric vehicles with a combustion engine range extender. The term "vehicles" includes motor vehicles, road and land vehicles, ships and other water vehicles, aircraft and rail vehicles.

Depending on which drive components are considered in the method according to the invention, first, second, third or fourth parameters n1i, τ1i, n2i, τ2i can be, for example, engine speeds, engine torques, battery and/or fuel cell currents and/or voltages. Examples of first and/or second predefined parameters n, τ can be output torques, output rotational speeds and/or driving speeds predefined by a driver or by a perspective. Time-based and distance-based systems are particularly suitable for specifying parameters using a forecast horizon.

The particle quality parameters wi are characteristic values dependent on the first and/or second and/or third and/or fourth parameters n1i, τ1i, n2i, τ2i, which are used in the method according to the invention to evaluate a quality of the particles i using predetermined criteria. Particle quality parameters can be, for example, an overall energy efficiency of the drive components, CO ₂ emissions from the vehicle or a drive component temperature.

In step S2 of the method according to the invention, the particle components K1i, K2i are shifted by means of a particle swarm optimization (Particle Swarm Optimization; PSO). The PSO is a calculation method known from heuristics in mathematics for the approximate solution of optimization problems.

Because the first and second particle components K1i, K2i are repeatedly shifted by means of a particle swarm optimization based on the particle quality parameter wi, a drive energy distribution of the drive components can be calculated using the method according to the invention, which is optimized with regard to a criterion specified by the particle quality parameter.

Steps S1 to S3 of the method according to the invention are preferably repeated a plurality J of repetitions j, with each repetition j in step S1 a plurality I of particles i of a previous repetition j-1 being provided and with each repetition j in step 2 new ones first driving parameters n and/or new second driving parameters τ are specified by a forecast horizon. This makes it possible to determine drive energy distributions in a future that can be predicted using the forecast horizon.

Furthermore, the application of the PSO in step S2 enables the method according to the invention to be carried out in real time due to a low computing resource requirement compared to other known optimization methods. “Real time” is understood here to mean that the optimized drive energy distribution is determined within a predetermined period of time. This period of time is preferably shorter than 50 ms.

The first characteristic map and the second characteristic map are preferably restricted to operating parameter ranges that can be expected in future operation. This restriction prevents operating parameter ranges from being taken into account when determining the drive energy distribution, which ranges are out of the question for an upcoming driving operation. This reduces the computing effort and/or the calculation time when the method is carried out with the aid of a computer.

When executed by a computing unit, the computer program product according to the invention causes the computing unit to execute a method according to the invention for determining a drive energy distribution of a hybrid vehicle.

The hybrid vehicle according to the invention comprises a first drive component, a second drive component and a control unit with a computing unit which is designed and set up to execute a computer program product according to the invention, the control unit being designed and set up to control a drive energy distribution of the first drive component and the second drive component .

The dependent claims describe further advantageous embodiments of the invention.

• 1 ein Ausführungsbeispiel eines erfindungsgemäßen Hybridfahrzeugs,
• 2a ein Ausführungsbeispiel einer ersten Kennfelddarstellung einer ersten Antriebskomponente, umfassend zufällig verteilte erste Partikelkomponenten K1 i einer Vielzahl I von Partikeln i,
• 2b ein Ausführungsbeispiel einer zweiten Kennfelddarstellung einer zweiten Antriebskomponente, umfassend zufällig verteilte zweite Partikelkomponenten K2i einer Vielzahl I von Partikeln i,
• 2c ein Ausführungsbeispiel einer ersten Kennfelddarstellung einer ersten Antriebskomponente, umfassend eine Ansammlung erster Partikelkomponenten K1 i einer Vielzahl I von Partikeln i,
• 2d ein Ausführungsbeispiel einer zweiten Kennfelddarstellung einer zweiten Antriebskomponente, umfassend eine Ansammlung zweiter Partikelkomponenten K2i einer Vielzahl I von Partikeln i und
• 3 ein Ausführungsbeispiel eines erfindungsgemäßen Verfahrens zum Bestimmen einer Antriebsenergieverteilung eines Hybridfahrzeugs.

Preferred exemplary embodiments are explained in more detail with reference to the following figures. while showing

• 1 an embodiment of a hybrid vehicle according to the invention,
• 2a an embodiment of a first characteristic map representation of a first drive component, comprising randomly distributed first particle components K1 i of a plurality I of particles i,
• 2 B an embodiment of a second characteristic map representation of a second drive component, comprising randomly distributed second particle components K2i of a plurality I of particles i,
• 2c an embodiment of a first characteristic map representation of a first drive component, comprising a collection of first particle components K1 i of a plurality I of particles i,
• 2d an exemplary embodiment of a second characteristic diagram representation of a second drive component, comprising a collection of second particle components K2i of a plurality I of particles i and
• 3 an embodiment of a method according to the invention for determining a drive energy distribution of a hybrid vehicle.

This in 1 Hybrid vehicle 1 shown comprises a parallel hybrid drive train with a first drive component 4, which is designed as an electric engine, and a second drive component 7, which is designed as an internal combustion engine.

Furthermore, the hybrid vehicle 1 includes an internal data memory 10 in which models of the drive components 4, 7 are stored. The models include maps 3, 6 of operating ranges of the drive components 4, 7. 2a shows a representation of a first characteristics map 3 of the electric motor 4. The representation includes a plot of an electric motor speed n1 on a first axis 11 and a first electric motor torque τ1 on a second axis 12. Furthermore, FIG 2 B a representation of a second characteristic map 6 of the internal combustion engine 7. The representation includes a plot of an internal combustion engine speed n2 on a first axis 13 and an internal combustion engine torque τ2 on a second axis 14.

Furthermore, the hybrid vehicle 1 includes a control device 9 (Engine Control Unit, ECU), which is designed and set up to control a drive energy distribution of the electric motor 4 and the internal combustion engine 7 . The control unit 9 includes a computing unit 8 which is designed and set up to be a computer to execute a program product for determining a drive energy distribution of the hybrid vehicle 1 .

An exemplary embodiment of a method according to the invention for determining a drive energy distribution of hybrid vehicle 1 is described below. The steps of the procedure are in 3 shown schematically. In the example, data of a route ahead of the hybrid vehicle 1 is first recorded. This is done using a navigation device (not shown) and route planning by a driver (not shown) of the hybrid vehicle 1. The computing unit 8 is designed to create a forecast horizon based on the data of the route ahead. This includes a most probable path (MPP) as well as boundary conditions for the path, such as expected gradients and vehicle speeds on the route segments ahead.

In the exemplary embodiment, the computing unit 8 uses the forecast horizon to calculate the minimum and maximum expected electric engine speeds, electric engine torques, internal combustion engine speeds and internal combustion engine torques and restricts the first characteristic map 3 and the second characteristic map 6 to operating parameter ranges which lie between the minimum and maximum values to be expected. Furthermore, the computing unit 8 normalizes the characteristic fields 3, 6 to the respective expected maximum values of the corresponding torques or speeds. The limitation and standardization of the characteristic diagrams 3, 6 reduces the computing effort and the calculation time when the method is carried out with the aid of a computer.

In a first step of the method S1, the computing unit 8 provides a limited number I of particles i. In the exemplary embodiment, these are exactly 100 particles i. Each particle i includes a first particle component K1i, 2, which describes a first randomly selected operating point in the first characteristic map 3 of the electric motor 4 using a first parameter n1i and a second parameter τ1i. In the exemplary embodiment, the first parameter n1 i is an electric motor speed and the second parameter τ1i is an electric motor torque. Furthermore, each particle i includes a second particle component K2i, 5, which describes a second randomly selected operating point in a second characteristic diagram 6 of the internal combustion engine 7 by means of a third parameter n2i and a fourth parameter τ2i. In the exemplary embodiment, the third parameter n2i is an internal combustion engine speed and the fourth parameter τ2i is an internal combustion engine torque. The second parameter τ1i and the fourth parameter τ2i combine to form a second driving parameter τ specified by a driver. In the exemplary embodiment, τ is a total torque requested by the driver.

Furthermore, each particle i includes a particle quality parameter wi. In the exemplary embodiment, the particle quality parameter wi is an overall efficiency composed of an electric engine efficiency w1i and an internal combustion engine efficiency w2i. The particle quality parameter wi of each particle i depends on the parameters n1i, τ1i, n2i, τ2i of its particle components K1i, K2i, 2, 5. Since there are no particles with identical parameters n1i, τ1i, n2i, τ2i in the exemplary embodiment, the particle quality parameters wi all have particle i different values. Since the number of particles i considered is finite, a particle quality parameter wi of a single particle i has a maximum value wmax.

In a second step of method S2, a first accumulation of first particle components K1i, 2 is formed in first characteristic diagram 3 and a second accumulation of second particle components K2i, 5 in second characteristic diagram 6 (see 2c) , i.e )). For this purpose, the computing unit 8 shifts the first and second particle components K1i, K2i, 2, 5 by means of particle swarm optimization (PSO).

In the exemplary embodiment, the computing unit 8 shifts the first particle components K1i, 2 in the first characteristic map 3 by the shift values Δn1i and Δr1i and the second particle components K2i, 5 in the second characteristic map 6 by the shift values Δn2i and Δτ2i using the PSO. The displacement values Δn1i, Δτ1i, Δn2i, Δτ2i each include a random part and a part which is dependent on the parameters of the particle with the maximum particle quality parameter wimax. As a result, when shifting, the particle components K1i, K2i, 2, 5 move on average towards areas in the characteristic diagrams 3, 6 which have a high overall efficiency.

In the exemplary embodiment, the computing unit 8 repeats the shifting of the first and second particle components K1i, K2i, 2, 5 in step S2 100 times. After each displacement of the first and second particle components K1i, K2i, 2, 5 of each particle i, the computing unit 8 recalculates its particle quality parameter wi. As a result of the repetition, the areas in the first and second characteristics map 3, 6, in which the particle components K1i, K2i, 2, 5 accumulate, are reduced and the density of the particle component accumulation in the respective areas is increased. This leads to an increase in accuracy when determining an optimized drive energy distribution.

In the exemplary embodiment, after each shifting of the first and second particle components K1i, K2i, 2, 5 of a particle i, the computing unit 8 checks in a step 21 of the method whether the operating points of the electric motor 4 and the Internal combustion engine 7 are within physically possible operating ranges.

In the event that a first or second particle component K1i, K2i, 2, 5 of a particle i describes an operating point which is outside a physically possible operating range, the computing unit 8 overwrites the respective particle component K1i, K2i, 2, 5 in a step S22 with a new particle component K1i, K2i, 2, 5, which is within a physically possible operating range.

As a result of the overwriting, a calculation speed can be reduced when the method is carried out in a computer-implemented manner and the accuracy in determining an optimized drive energy distribution can be increased.

In a third step of the method S3, the computing unit 8 determines the drive energy distribution by evaluating the first collection of the first particle components K1i, 2 and the second collection of the second particle components K2i, 5 using probability density functions. In the exemplary embodiment, the probability density functions are Gaussian distributions. These are each calculated separately for the first and second particle accumulations K1i, K2i, 2, 5. Maximum values of the two Gaussian distributions reflect optimized operating points of the electric engine 4 and the internal combustion engine 7 and thus an optimized drive energy distribution.

In the exemplary embodiment, the arithmetic unit 8 repeats the steps S1 to S3 a predetermined plurality J of repetitions j. For each repetition j, the arithmetic unit 8 provides a multiplicity I of particles i of a previous repetition j-1 in step S1. The particle components K1i, K2i, 2, 5 of the particles i are already accumulated in areas in the first and second characteristics map 3, 6 when they are provided by the repeated shifting in step S2 of the previous repetition j-1 (see Fig 2c , 2d ). In step S2 of each repetition j, the look-ahead horizon specifies a new, second driving parameter τ, which corresponds to an expected torque request from the driver on a likely traveled route section of the most probable driving path. Due to the new second driving parameters τ, the particles i are shifted into new areas of the characteristic diagrams 3, 6 in step S2 of each repetition j by means of the computing unit 8. In step S3 of each repetition j, the computing unit 8 calculates a new, optimized drive energy distribution based on the particle component accumulations in the characteristic diagrams 3, 6. This enables drive energy distributions to be determined on route sections of the most probable driving path ahead in a future that can be predicted based on the forecast horizon.

The exemplary embodiment of the method according to the invention described above is used in 1 Hybrid vehicle 1 shown used to determine a drive energy distribution. For this purpose, the computing unit 8 of the hybrid vehicle 1 executes a computer program product which is designed and set up, when executed by the computing unit 8, to cause the latter to 3 to carry out the method shown for determining a drive energy distribution of the hybrid vehicle 1 .

In the exemplary embodiment, the computing unit 8 and the computer program product are designed and set up to calculate the drive energy distribution of the hybrid vehicle 1 in real time. “Real time” means that the optimized drive energy distribution is determined within a period of 50 ms. The calculation in real time is made possible by the use of the PSO, which, in comparison to other known calculation methods for the approximate solution of optimization problems, requires little computing resources, and a high data processing speed of the computing unit 8 .

In further exemplary embodiments of the method according to the invention (not shown), more than 2 drive components are included in the determination of a drive energy distribution. In addition to the overall efficiency of the drive components, battery temperatures and/or emissions of CO2, NOx or noise during operation of the hybrid vehicle can also be taken into account as particle quality parameters wi.

In a further exemplary embodiment of the method according to the invention (not shown), a computing unit calculates a drive energy distribution for a serial hybrid drive train of a hybrid vehicle, comprising a first drive component and a second drive component. In step S1 of the method, the computing unit provides a multiplicity I of particles i, each particle i comprising a first particle component K1i, which describes a first randomly selected operating point in a first characteristic map of the first drive component using a first parameter n1i and a second parameter τ1i , and where each particle i has a second particle compo nent K2i, which describes a second randomly selected operating point in a second map of the second drive component by means of a third parameter n2i and a fourth parameter τ2i. The first parameters n1 i describe speeds of possible operating points of the first drive component. The third parameters n2i describe speeds of possible operating points of the second drive component, which combine with the first parameters n1i to form a specified speed n.

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• CN 104408257A [0002]

Claims (10)

Method for determining drive energy distributions for a drive train of a hybrid vehicle, comprising the steps: S1: Providing a multiplicity I of particles i, each particle i - a first particle component K1i (2), which describes a first randomly selected operating point in a first characteristic map (3) of a first drive component (4) of a drive train by means of a first parameter n1i and a second parameter τ1i, - a second particle component K2i (5), which describes a second randomly selected operating point in a second characteristic diagram (6) of a second drive component (7) of the drive train by means of a third parameter n2i and a fourth parameter τ2i, with the first parameter n1i and the combine third parameters n2i to form a first specified driving parameter n and/or where the second parameter τ1i and the fourth parameter τ2i are combined to form a second specified driving parameter τ, - and a particle quality parameter wi; S2: Forming a first accumulation of first particle components K1i (2) in the first characteristic map (3) and a second accumulation of second particle components K2i (5) in the second characteristic map (6) by repeatedly shifting the first and second particle components K1i, K2i (2, 5) using Particle Swarm Optimization (PSO) based on the particle quality parameters wi of particles i; S3: Determining a drive energy distribution by evaluating the first collection of the first particle components K1i (2) and the second collection of the second particle components K2i (5) using a probability density function. procedure after claim 1 , where steps S1 to S3 are repeated a plurality J of repetitions j, with each repetition j in step S1 a plurality I of particles i of a previous repetition j-1 being provided and with each repetition j in step 2 new first driving parameters n and/or new second driving parameters τ are specified by a forecast horizon. procedure after claim 1 or 2 , wherein in step S2 the first and/or second particle components K1i, K2i (2, 5) are shifted by a shift value that includes at least a random part. Method according to one of the preceding claims, in which the characteristic diagrams (3, 6) are restricted to operating parameter ranges which can be expected in future driving operations. Method according to one of the preceding claims, in which the characteristic diagrams (3, 6) are normalized to expected operating parameter ranges. Method according to one of the preceding claims, comprising the step: S21: Checking whether a first and/or second operating point described by a first and/or second particle component K1i, K2i (2, 5) lies within a physically possible operating range after a shift. procedure after claim 6 , comprising the step: S22: overwriting a first and/or second particle component K1i, K2i (2, 5), which describes a first and/or second operating point that lies outside a physically possible operating range, with a first and/or second particle component K1i, K2i (2, 5), which describes an operating point that lies within a physically possible operating range. Computer program product which, when executed by a computing unit (8), causes the computing unit (8) to carry out a method according to one of Claims 1 until 7 to execute. Hybrid vehicle (1), comprising a first drive component (4), a second drive component (7) and a control unit (9) with a computing unit (8), which is designed and set up, according to a computer program product claim 8 execute, wherein the control unit (9) is designed and set up to control a drive energy distribution of the first drive component (4) and the second drive component (7). Hybrid vehicle (1) according to claim 9 , wherein the computing unit (8) is designed to determine a drive energy distribution in real time.

Images