METHOD FOR OPERATING A HYBRID VEHICLE

Description

[0001] The invention relates to a method for operating a hybrid vehicle, in particular a diesel hybrid vehicle, with an internal combustion engine and an electric drive, wherein a total power supplied for driving the vehicle is divided between the internal combustion engine and the electric drive according to a distribution ratio.

[0002] Regarding the state-of-charge management of a hybrid vehicle, real-time optimization-based operating strategies are known, implemented on an engine control unit or other vehicle control units. One example is the so-called "Equivalent Consumption Minimization Strategy (ECMS)." A key factor influencing the operating strategy is the carbon dioxide emissions of the combustion engine and, directly linked to this, its fuel consumption. The combustion engine's emissions behavior is modeled within the operating strategy using measurement data from steady-state combustion engine operation.

[0003] The emissions of an internal combustion engine, especially a diesel engine, can deviate significantly from steady-state operation during transient operation, such as during strong acceleration. In this case, the operating strategy may make suboptimal decisions regarding the power distribution between the internal combustion engine and the electric drive, resulting in a less than optimal load point for the internal combustion engine and consequently suboptimal fuel consumption.

[0004] DE 10 2005 044 828 A1 discloses a method for operating a hybrid vehicle in which operating point data are determined in a first step using a stored characteristic curve field. In a second step, the operating point data are optimized taking into account the dynamic behavior of the vehicle components.

[0005] The scientific publication "Method for real-time simulation of dynamic emission profiles of selected pollutants from gasoline engines" by Jurij Schmidgal describes another method in which the dynamic behavior of vehicle components is taken into account when determining the operating point data.

[0006] The object of the invention is therefore to provide a method in which an optimal load point of the combustion engine is selected even in transient operation of the combustion engine.

[0007] The problem is solved according to the invention by a method according to claim 1. Since dynamic quantities are used instead of static quantities to determine the optimal load point, altered performance characteristics during the transient operation of the internal combustion engine can be taken into account, which cannot be considered when using static quantities. By taking these altered performance characteristics into account, the load point of the internal combustion engine can be better adjusted and thus fuel consumption can be reduced.

[0008] In transient operation, the efficiency of the combustion engine differs from its steady-state efficiency, particularly due to boost pressure and lambda shifts. By taking this difference into account, the optimal load point of the combustion engine can be more accurately determined. Specifically, a load point shift is also determined, meaning a shift in the combustion engine's load point caused by torque exerted by the electric drive.

[0009] The transient efficiency of the combustion engine is determined using a linear approximation. This reduces the computational effort and thus simplifies and speeds up the determination of the transient efficiency.

[0010] According to the invention, the optimal load point of the internal combustion engine is determined by extremizing a Hamiltonian function, in particular by maximizing or minimizing it. This is done in particular by means of Pontryagin's extremal principle (depending on the sign convention, the maximum principle or the minimum principle).

[0011] The Hamiltonian function encompasses the time derivative of the dynamic fuel consumption of the combustion engine. Dynamic fuel consumption is particularly well-suited as a target value for the procedure, as fuel consumption is directly linked to the carbon dioxide emissions of the hybrid vehicle, which should generally be kept below a certain level. Furthermore, fuel consumption is measured in modern vehicles anyway, for example, for display on an on-board computer, and is therefore available without any additional measurement effort.

[0012] According to the invention, the Hamiltonian function comprises at least one constraint whereby the battery must supply a difference between the total drive power and the combustion engine power, as well as any electrical losses that occur. Additionally, one or more further conditions that the hybrid vehicle must fulfill during operation can be taken into account.

[0013] One aspect involves determining the current transient efficiency at certain intervals during the hybrid vehicle's operation. This allows the combustion engine's load point to be continuously adapted to the currently required load. As a result, an optimal load point for the combustion engine is maintained even under varying loads, thus optimizing fuel consumption.

[0014] According to one embodiment of the invention, the dynamic fuel consumption of the internal combustion engine is modeled using a second-degree polynomial in the engine power. This provides higher accuracy than a linear approximation. Furthermore, fuel consumption is to be optimized, i.e., minimized. Near an extremum of a function, a quadratic polynomial with appropriately adjusted coefficients always provides a good approximation.

[0015] Preferably, the constraint further includes one or more of the following conditions: electrical losses, required total torque, required total power, maximum nitrogen oxide emissions, maximum carbon dioxide emissions, and maximum fuel consumption. In particular, the constraints depend on the operating parameters of the hybrid vehicle. By incorporating the constraints, the combustion engine and the electric drive can be controlled in such a way that they exhibit the optimal power distribution between the combustion engine and the electric drive in the respective situation. For example, at speeds below or around 50 km / h, the combustion engine and the electric drive could be controlled in such a way as to reduce nitrogen oxide emissions. Since it can be assumed that the hybrid vehicle is located in a built-up area, more power would therefore be required from the electric drive in this case to reduce emissions.

[0016] Further advantages and features of the invention will become apparent from the following description and the drawings, to which reference is made. These show: Fig. 1 a schematic representation of a possible topology of a drive system of a hybrid vehicle; Fig. 2 Corresponding diagrams, each plotting: (a) fuel consumption versus internal combustion engine power; (b) fuel consumption versus time; (c) internal combustion engine power versus time; and (d) dynamic efficiency of the internal combustion engine versus time; Fig. 3 a diagram of the dynamic efficiency of the internal combustion engine plotted against the torque of the internal combustion engine; Fig. 4 a linear approximation of the dynamic efficiency of Figure 3 ; and Fig. 5 A diagram of static fuel consumption and dynamic fuel consumption, each plotted against internal combustion engine power.

[0017] In Figure 1 A possible topology of a drive system 10 of a hybrid vehicle is shown. The drive system 10 comprises a Internal combustion engine 12 and an electric drive 14, which is powered by a battery 16. If it is a gasoline hybrid vehicle, the combustion engine 12 is a gasoline engine. If, on the other hand, it is a diesel hybrid vehicle, the combustion engine 12 is a diesel engine.

[0018] The combustion engine 12 and the electric drive 14 are each designed to drive wheels 18 of the hybrid vehicle. The total drive power of the hybrid vehicle is therefore the sum of the respective power outputs of the combustion engine 12 and the electric drive 14. A control unit 20 is configured to adjust the distribution of the total drive power between the combustion engine 12 and the electric drive 14 to an optimal ratio.

[0019] The following describes a state-of-the-art operating strategy for a hybrid vehicle.

[0020] At a given rotational speed n, optimal performance is achieved in the static operation of the hybrid vehicle (i.e., especially during no acceleration). P VM,opt of the internal combustion engine 12 is sought. For fuel consumption ṁ n as a function of the combustion engine power P V M The quadratic approximation is used. m ˙ n P VM = α 2 P VM 2 + α 1 P VM + α 0 Fuel consumption is therefore a second-degree polynomial in terms of combustion engine power with coefficients α 0 , α 1 and α 2. These coefficients can be determined by measurements at different rotational speeds and then used to control the operating strategy. In addition, the power delivered by battery 16 is also taken into account. P bat The following quadratic approximation was used: P batt = β 2 P FW t − P VM 2 + β 1 P FW t − P VM + β 0 . P batThis represents a constraint for the drive system 10: The battery 16 must be able to store the difference between the total drive power P FW and feed in the combustion engine power as well as any electrical losses that occur. P bat is also a second-degree polynomial in the combustion engine power. The coefficients β i These values ​​can, in turn, be determined by measurement at different rotational speeds and then used to control the operating strategy. A Hamiltonian function H is then given by: H P VM λ = m ˙ P VM + λ P batt .

[0021] This is λ a Lagrange multiplier for the constraint condition P bat. According to Pontryagin's extremal principle (depending on the sign convention, the minimum principle or the maximum principle), the optimal combustion engine power can then be determined by extremizing the Hamiltonian function (minimizing or maximizing it, depending on the sign convention). Therefore, the equations must be... ∇ P VM , λ H P VM λ = 2 α 2 P VM + α 1 − 2 λ β 2 P FW − P VM − λ β 1 β 2 P FW − P VM 2 + β 1 P FW − P VM + β 0 = 0 0 to be solved. From these two equations, the Lagrange multiplier can then be determined. λ as well as the optimal load point P VM,opt determine.

[0022] This operating strategy sets the load point of the combustion engine 12 based on coefficients measured during steady-state operation of the hybrid vehicle. However, the emissions of the combustion engine 12 can deviate significantly from steady-state operation during transient operation, such as during strong acceleration, which is why the operating strategy may set a suboptimal load point for the combustion engine 12. This is particularly relevant in Figures 2 and 5 To recognize: The dynamic fuel consumption curve Your bra deviates significantly from the static fuel consumption curve Bra size away.

[0023] The following will be based on the Figures 1 to 5a method according to the invention for operating a hybrid vehicle is explained, in which an optimal load point of the combustion engine 12 is also selected in the transient operation of the hybrid vehicle.

[0024] In Figure 2 (a) are a static fuel consumption curve Bra size as well as a dynamic fuel consumption curve Your bra each as a function of the combustion engine power P V M at constant speed n shown. Figures 2(b), (c) and (d) show the corresponding time dependencies of the Fuel consumption ṁ fuel, the combustion engine power P V M or the dynamic efficiency yours of the internal combustion engine 12. In this process yours in units of static efficiency the state specified, that is, that at yours = 1 applies η dyn = η stat .

[0025] Up to a certain point in time (1), the hybrid vehicle operates at a constant speed and constant load. Fuel consumption up to this point is calculated at a power output of... P Bass given by Bass. At time 1, a load change occurs, for example, due to strong acceleration of the hybrid vehicle. As in Figures 2 (b) to (d) As can be seen, fuel consumption rises almost instantly to the load point consumption. ṁ Lsp However, in this transient operating state, the combustion engine 12 does not achieve the static efficiency, due among other things to boost pressure deviation and lambda deviation. the state, but only a reduced, dynamic efficiency the woman. This means that the combustion engine, with the same fuel consumption, ṁ Lsp a performance P you exhibits a power that is smaller than the corresponding static power P state.

[0026] At time 2, the efficiency of the combustion engine 12 is not yet at a steady-state level, but is approaching it. At time 3, the load change is complete, so that the following now applies again. η dyn = 1 and Pdyn = Pstat. The combustion engine 12 then operates in a quasi-stationary state again. In other words, the dynamic fuel consumption curve approaches zero. Your bra between time points 2 and 3 of the static fuel consumption curve Bra size to.

[0027] To select the optimal load point of the internal combustion engine 12, the dynamic efficiency of the internal combustion engine 12 must therefore be taken into account. This will be demonstrated below using the Figures 3 and 4 explained.

[0028] In Figure 3 This is an example of a possible dependency of the transient efficiency. yours shown from the height of the load jump. In the example shown, the load jump occurs (upper arrow in Figure 3) from an internal combustion engine torque of 50 Nm with yours (50 Nm) = 1 to an internal combustion engine torque of 150 Nm with yours (150 Nm) ≈ 0.6. However, if the electric drive 14 supports the load change with a torque of 50 Nm, for example, it shifts the load point of the combustion engine 12 to 100 Nm (lower arrow in Figure 3 ), so the transient efficiency of the combustion engine 12 only drops to yours (150 Nm) ≈ 0.88. In other words, the combustion engine 12 then operates more efficiently in the transient range. The optimal load point of the combustion engine 12 is determined using the method described below.

[0029] Since the dependence of the transient efficiency on the load point displacement is generally complex, a linear approach has proven to be expedient. This is known as yours, yours in Figure 4This is illustrated. In this approach, the transient efficiency is determined from the following equations: η dyn P VM = k P VM + d , k = η dyn t − 1 − η dyn t P FW − P LPV t − 1 − P FW − P LPV t , d = η dyn t − 1 − k P FW − P LPV t − 1 .

[0030] The indices "t" and "t - The "1" indicates that the current transient efficiency is determined at certain time intervals. The linearized relationship between the efficiency of the combustion engine 12 and the combustion engine torque is determined, in particular, by including the current values ​​and the values ​​from the last time step.

[0031] The following equation is now used for dynamic fuel consumption (see also Figure 5 ): m ˙ dyn P VMdyn = m ˙ stat P VM = α 2 P VM 2 + α 1 P VM + α 0 .

[0032] The following three equations apply: P VMdyn = P Bas + P VM − P Bas η dyn = = 1 k 2 d + k P VM k P VM − 1 + d 2 + 1 , η dyn = k P VM + d , η dyn P Bas = 1 und damit P Bas = 1 − d k .

[0033] To determine the optimal load point of the internal combustion engine, the following Hamiltonian function H is used: H P VMdyn , λ = m ˙ dyn P VMdyn P VM + λ P batt .

[0034] The Hamilton functionThis includes the time derivative of the dynamic fuel consumption of the internal combustion engine 12 as well as a Lagrange multiplier. λ and, in analogy to the stationary case discussed at the beginning, the coercive condition P bat. The Hamilton function is therefore a second-degree polynomial in the dynamic combustion engine power.

[0035] According to Pontryagin's extremal principle (depending on the sign convention, the minimum principle or the maximum principle), the optimal combustion engine power is determined by extremizing the Hamiltonian function (minimizing or maximizing it, depending on the sign convention). Thus, the equations are... (iv) dH P VMdyn P VM , λ dP VM = dH P VMdyn P VM dP VMdyn dP VMdyn dP VM = 0 , (v) dH P VMdyn P VM , λ dλ = 0 , solved. From the two equations (iv) and (v), the Lagrange multiplier is then calculated using equations (i) to (iii). λ as well as the optimal load point P VM,opt determined. The optimal load point is calculated as follows: PVM,opt=−b±b2−4ac2a, with a = 4 λβ 2 k, b = 2( α 2 + λβ 2 (2 d - 2 k P FW - 1) - λβ 1 k ) and c = α 1 + 2 λ β 2 P FW (1 - 2d) - 2 λ β 1 d + λ β 1 . It should be mentioned that this applies to the borderline case of stationary operation, i.e., for η dyn = 1, k The solution for steady-state operation is reproduced when d = 0 and d = 1. The coefficients α i and β i are known from stationary operation and are therefore available for the procedure.

[0036] With appropriate modifications, alternative or additional constraints can be implemented in the process. These alternative or additional constraints could be, for example, a required total torque, a required total power output, a maximum nitrogen oxide emission, a maximum carbon dioxide emission, and / or a maximum fuel consumption. The additional constraints are then formulated using an equation based on the Lagrange multiplier method. for ( P VM, cj ) shown ( cj (here denotes any further variables and parameters) and with a Lagrange multiplier λ i incorporated into the Hamiltonian function. For a natural number N The constraints then result in the Hamiltonian function H P VMdyn λ i = m ˙ dyn P VMdyn P VM + ∑ i = 1 N λ i f i P VM c j .

[0037] The optimal load point of the internal combustion engine is then determined analogously to the procedure described above by extremizing the Hamilton function.

[0038] The described procedure can be carried out fully automatically using the control unit 20.

Claims

1. Method for operating a hybrid vehicle with an internal combustion engine (12) and an electric drive (14), wherein the electric drive (14) is fed by a battery (16), wherein a total power applied for driving the vehicle is divided between the internal combustion engine (12) and the electric drive (14) according to a distribution ratio, wherein at least in a transient operation of the internal combustion engine (12) dynamic variables are used to determine an optimal load point of the internal combustion engine (12), wherein a transient efficiency of the internal combustion engine (12) is used for determining the optimal load point of the internal combustion engine (12), characterised in that the transient efficiency of the internal combustion engine (12) is determined by means of a linear approximation, wherein the optimal load point of the internal combustion engine (12) is determined by extremising a Hamilton function, wherein the Hamilton function comprises the time derivative of the dynamic fuel consumption of the internal combustion engine (12), and wherein the Hamilton function comprises at least one constraint condition, according to which the battery (16) must feed in a difference from a total drive power and an internal combustion engine power as well as occurring electrical losses.

2. Method according to claim 1, wherein during the operation of the hybrid vehicle the current transient efficiency is determined at certain time intervals.

3. Method according to one of the preceding claims, wherein the dynamic fuel consumption of the internal combustion engine (12) is modelled by means of a polynomial of the second degree in the internal combustion engine power.

4. Method according to one of the preceding claims, wherein the constraint condition additionally comprises one or more of the following conditions: electrical losses, demanded total torque, demanded total power, maximum nitrogen oxide emission, maximum carbon dioxide emission and maximum fuel consumption.