Method for operating a hybrid energy storage system

Abstract

Method for determining an optimized system behavior of a hybrid energy storage system (1) with at least a first energy storage device (2) and a second energy storage device (3) by determining an optimal adjoint (λ*), comprising: - Executing a first iteration loop (10) in several passes, where each successive pass is associated with a possible adjoint (λ), - wherein, for each iteration of the first iteration loop (10), a second iteration loop (20) is executed, which includes the steps: • Determining an optimal control variable (u*) for each successive point in time within a given period, where the optimal control variable (u*) is a quantity that describes the energy absorbed or released by the second energy storage device and is determined based on a minimum of a Hamiltonian function (H), where the Hamiltonian function (H) belongs to the energy storage system (1) and depends on a modeled system parameter (L) m ), the possible adjoint (λ) and a storage state (SoC2) of the second energy storage (3) is, where the modeled system parameter (L m ) is a parameter of the hybrid energy storage system (1) to be optimized, and is determined according to a predefined performance profile which is traversed by the energy storage system (1) in the predefined period, and • Calculating a final memory state (SOC) 2E ) of the second energy storage device (3), which the second energy storage device (3) exhibits after the specified period, if the energy absorbed or released by the second energy storage device (3) during the specified period was controlled according to the respective optimal control variable (u*), - Detect whether the final memory state (SOC) calculated using the second iteration loops (20) 2E ) of the second energy storage (3) lies within a specified interval and provide an optimal adjoint (λ*) which corresponds to the possible adjoint (λ) of the first iteration loop (10) for which it was detected that the final storage state (SOC) 2E ) lies within the specified interval.

Description

State of the art

[0001] The present invention relates to a method for operating a hybrid energy storage system.

[0002] Technical systems whose energy needs are met by electrical energy storage devices place varying performance and energy demands on these storage devices. Examples of energy storage devices include primary batteries, secondary batteries, fuel cells, and double-layer capacitors.

[0003] In fuel cell vehicles, an additional electrical storage device is often used to operate the fuel cell efficiently, enable energy recuperation, and increase performance. For example, passive interconnections of batteries with different energy and power characteristics are known to enhance performance. Furthermore, systems are known in which two electrical storage devices are coupled via a DC / DC converter, thus enabling different voltage levels and more independent operation.

[0004] These systems, when they combine, for example, a fuel cell and a battery or a battery and a double-layer capacitor, are called hybrid energy storage systems. In the special case where the system combines two secondary batteries, it is also called a hybrid battery storage system.

[0005] Current systems with a hybrid energy storage system mostly use a simple control system, the real-time implementation of which can be achieved with minimal effort. However, this control system leads to suboptimal results with regard to potential goals over the operating lifetime, such as energy efficiency.

[0006] For example, systems are known in which two storage devices are connected to a DC / DC converter and a simple current controller is used for load sharing. Hybrid battery control systems for use in plug-in hybrid vehicles are also known, which are controlled by a rule-based controller or a filter.

[0007] For hybrid vehicles, approaches have already been presented that aim to minimize fuel consumption during driving using optimization methods. These are based on Pontryagin's minimum principle and are often derived in current vehicle applications through simplifications and assumptions leading to the so-called Equivalent Consumption Minimization Strategy (ECMS).

[0008] Furthermore, an approach is known which, also based on Pontryagin's minimum principle, leads to the control of a hybrid energy storage system in a vehicle, consisting of a fuel cell and a double-layer capacitor bank. This strategy minimizes hydrogen consumption and maintains the charge level of the double-layer capacitor within a defined range.

[0009] German patent DE102013014667A1 discloses a control system for a hybrid vehicle. It describes a coupling of an internal combustion engine and an electric machine, with the electric machine being powered by the battery. This design utilizes Pontryagin's principle of minimum. Disclosure of the invention

[0010] The inventive method for determining optimized system behavior of a hybrid energy storage system with at least one first energy storage device and a second energy storage device by determining an optimal adjoint comprises executing a first iteration loop in several passes, wherein each successive pass corresponds to a possible adjoint, and wherein a second iteration loop is executed with each pass of the first iteration loop. The second iteration loop comprises determining an optimal control variable for successive time points within a predetermined period, wherein the optimal control variable is a quantity that describes energy absorbed or released by the second energy storage device and is determined based on a minimum of a Hamiltonian function.wherein the Hamiltonian is associated with the energy storage system and depends on a modeled system parameter, the possible adjoint, and a storage state of the second energy storage device, wherein the modeled system parameter is a parameter of the hybrid energy storage system to be optimized, and is determined according to a predefined power profile that the energy storage system traverses within the predefined period, and a calculation of a final storage state of the second energy storage device that the second energy storage device exhibits after the predefined period if the energy absorbed or released by the second energy storage device within the predefined period was controlled according to the respective optimal control variable. The method further includes detection,whether the final storage state of the second energy storage device, calculated using the second iteration loops, lies within a specified interval, and providing an optimal adjoint that corresponds to the possible adjoint of the first iteration loop iteration for which it was detected that the final storage state lies within the specified interval.

[0011] The inventive method for controlling a hybrid energy storage system, comprising at least a first energy storage device and a second energy storage device, comprises acquiring a first input value and a second input value, wherein the first input value describes a pre-calculated value of an optimal adjoint and the second input value describes a deviation of a storage state of the second energy storage device from a reference storage state, adjusting the optimal adjoint for a current storage state of the second energy storage device based on the first and second input values ​​to determine an adapted adjoint, providing an energy target value, which describes energy to be jointly provided by the first energy storage device and the second energy storage device, and determining an optimal control variable based on the adapted adjoint and the energy target value.wherein the optimal control variable is a quantity by which energy absorbed or released by the second energy storage device is described, and the optimal control variable is determined according to a Hamiltonian function associated with the energy storage system, wherein the Hamiltonian function depends on a calculated system parameter, the fitted adjoint, and the current storage state of the second energy storage device.

[0012] Thus, the inventive method for determining the optimized system behavior of a hybrid energy storage system determines an optimized value for an adjoint for a specific energy storage system. This value is then applied to control this specific energy storage system using the inventive method for controlling the hybrid energy storage system. Although each of these methods is advantageous in itself, optimal control is only achieved through their interaction. Combining the inventive methods creates optimal control. A mathematical term for an adjoint is Lagrange multiplier. A storage state is a state of the storage system and thus describes a property of the respective energy storage system, which, in particular, influences the system parameter to be optimized if it changes.

[0013] Optimal control ensures energy-optimized power distribution in a hybrid energy storage system, particularly a hybrid battery control system. Furthermore, the system can react to a rapidly depleting high-energy storage section, formed by one of the energy storage units, and maintain capacity within that section.

[0014] Although the real-time optimal control according to the invention is particularly advantageous for battery systems, it can be applied to any type of hybrid energy storage system. It is not essential that one or both of the energy storage devices be electrical or electrochemical. An energy storage device can also be a mechanical energy storage device, for example, a flywheel. Another conceivable energy storage device as part of a hybrid energy storage system is a fuel cell.

[0015] The operation of a hybrid energy storage system requires a stored control system. The methods according to the invention enable energy-efficient control without relying on simple filtering or control approaches. Furthermore, the determination of a control value, i.e., a control parameter, according to the invention allows for broad performance maintenance of the hybrid energy storage system during operation.

[0016] The methods according to the invention enable the optimization of the energy storage system according to arbitrary system parameters. For example, the system parameter can describe a system loss, i.e., an energy loss, and the energy storage system can be operated in a particularly energy-efficient manner. In another example, the system parameter can describe system aging, and the energy storage system can be operated in a particularly long-lasting manner.

[0017] The inventive method for controlling a hybrid energy storage system is particularly advantageous because it can be executed in real time with low computing power. The inventive method for determining optimized system behavior of a hybrid energy storage system is particularly advantageous because it can also be executed without the physical presence of a corresponding energy storage system.

[0018] The dependent claims describe preferred embodiments of the invention.

[0019] It is advantageous if the optimal control variable describes a current to be delivered by the second energy storage device, and / or the modeled system parameter is a modeled system loss, and / or the storage state is a state of charge, where the final storage state is a final state of charge, and / or the interval is a charging interval. Such a selection makes it possible to adapt the method for determining optimized system behavior of a hybrid energy storage system to a battery system, whereby the method is particularly suitable for minimizing power losses in the battery system. A strategy that, in conjunction with a system solution, can minimize energy losses is especially advantageous for battery cells with high internal resistance. As a side effect, the method can lead to aging benefits in a hybrid energy storage system by reducing the current rate load on the high-energy component.The minimization of electrical losses, as can be achieved by the methods according to the invention, also leads to a reduced temperature development in the energy storage devices, which is caused, among other things, by ohmic losses.

[0020] It is advantageous for the first iteration loop to perform a bisection procedure to determine the possible adjoint. This minimizes the number of iterations required for the first loop. The possible adjoint is determined, in particular, based on an interval bisection or interval division procedure.

[0021] It is advantageous, when determining the optimized system behavior of a hybrid energy storage system, to determine the optimal control variable in each iteration of the second iteration loop by executing a third iteration loop. In this third iteration loop, each successive iteration corresponds to a possible control value. In each iteration of the third iteration loop, a value of the Hamiltonian function is calculated according to the possible control variable associated with that iteration. The possible control variable for which the Hamiltonian function has a minimum value is then determined as the optimal control variable. This approach limits the number of possible control values ​​to those that can actually be processed and applied by the energy storage system's controller.Determining the optimal control variable in this way is also easy to implement and requires little computing power per execution. Alternatively, it is advantageous to determine the optimal control variable analytically in each iteration of the second iteration loop. This allows for a particularly accurate value for the optimal control variable to be determined.

[0022] Furthermore, it is advantageous if, when determining the optimized system behavior of a hybrid energy storage system, the specified interval is an interval around an initial state of charge of the second energy storage device, as it appears at the beginning of the specified period. Specifically, the interval is a percentage deviation from the initial state of charge of the second energy storage device. This allows for the definition of the required accuracy when determining the adjoint.

[0023] It is also advantageous, when determining the optimized system behavior of a hybrid energy storage system, to calculate the modeled system parameter based on the DC resistance of the first energy storage device, the DC resistance of the second energy storage device, the power output of the energy storage system at any given time according to the specified power profile, and the efficiency of a DC / DC converter connecting the first and second energy storage devices within the system. This allows the modeled system loss to approximate the actual system loss of the energy storage system. This enables a particularly precise determination of the optimal adjoint. The system parameter is, in particular, a system loss of the energy storage system.

[0024] It is equally advantageous if, when determining the optimized system behavior of a hybrid energy storage system, the value of the Hamiltonian is calculated by adding the modeled system parameter to a derivative of the storage state of the second energy storage device, which is then multiplied by its possible adjoint. A Hamiltonian formulated in this way describes the system behavior of the hybrid energy storage system with particular precision.

[0025] It is advantageous if the storage state is a state of charge, and / or the energy target value is a current target value, and / or the optimal control variable describes a current to be delivered by the second energy storage device, and / or the calculated system parameter is a system loss of the hybrid energy storage system. Such a selection makes it possible to adapt the method for determining an optimized system behavior of a hybrid energy storage system to a battery system, whereby the method is particularly suitable for minimizing power losses of the battery system.

[0026] It is advantageous if, in the method for controlling a hybrid energy storage system, the calculated system parameter depends on the power loss of the first energy storage device and the power loss of the second energy storage device. In particular, it is advantageous if the calculated system parameter also depends on the power loss of an energy converter through which the first and second energy storage devices are connected in the energy storage system. The energy converter is preferably a DC / DC converter. The calculated system parameter can thus be determined with particular accuracy, leading to highly efficient control of the hybrid energy storage system. The method can therefore be optimized to minimize power loss. The system parameter is, in particular, a system loss of the energy storage system.

[0027] It is also advantageous if the optimal control parameter is calculated in the method for controlling a hybrid energy storage system. In this way, a value for the optimal control parameter is calculated and does not need to be provided in advance. Therefore, no pre-calculated values ​​need to be provided.

[0028] It is equally advantageous if the optimal control parameter for a hybrid energy storage system is determined tabularly, whereby a table is queried in which different combinations of values ​​of the adapted adjoint and values ​​of the current target value are assigned an optimal control parameter. In this way, a particularly responsive control of the hybrid energy storage system can be achieved.

[0029] Furthermore, it is advantageous if the adjustment of the optimal adjoint in the method for controlling a hybrid energy storage system is carried out using a PI controller. Such a component is readily available at a low cost and allows for rapid, real-time execution of the process.

[0030] It is particularly advantageous if the optimal adjoint, which is received as the first input value in the method for controlling a hybrid energy storage system, has been determined by the inventive method for determining the system behavior of a hybrid energy storage system. Thus, an optimal adjoint can, for example, be determined at the factory and transferred to a control device, which executes the method for controlling a hybrid energy storage system. This control device can, for example, be installed in a vehicle.

[0031] A device for generating an optimal adjoint, which includes a computing unit configured to execute the inventive method for determining an optimized system behavior of a hybrid energy storage system, has all the advantages of the method.

[0032] A device for controlling a hybrid energy storage system, comprising a control unit configured to execute the inventive method for controlling a hybrid energy storage system, has all the advantages of the method. Brief description of the drawings

[0033] Exemplary embodiments of the invention are described in detail below with reference to the accompanying drawing. The drawing shows: Fig. 1. A circuit diagram of an exemplary hybrid energy storage system, Fig. 2 a flowchart of a method according to the invention for determining an optimized system behavior of a hybrid energy storage system according to a preferred embodiment of the invention, Fig. 3 a block diagram of a device according to the invention for controlling a hybrid energy storage system according to a preferred embodiment of the invention, and Fig. 4 a device according to the invention for generating an optimal adjoint. Embodiments of the invention

[0034] Fig. Figure 1 shows a circuit diagram of an example hybrid energy storage system. The energy storage system is a battery storage system 1. The battery storage system 1 comprises a first battery 2 as a first energy storage device. The battery storage system 1 comprises a second battery 3 as a second energy storage device.

[0035] A positive terminal of the first battery 2 is connected to an input of a DC / DC converter 4. A positive terminal of the second battery 3 is connected to an output of the DC / DC converter 4. A negative terminal of the first battery 2 is connected to a negative terminal of the second battery 3 via a circuit ground. The DC / DC converter 4 is an energy converter.

[0036] The first and second batteries 2, 3 are of different designs. For example, the first battery 2 is a battery suitable for providing high power output for short periods. The second battery 3, for example, is a battery suitable for providing constant power output over a long period. In alternative hybrid energy storage systems, at least one of the batteries 2, 3 is replaced by another energy storage device, such as a capacitor.

[0037] Furthermore, a contact terminal is arranged at the output of the DC / DC converter 4, via which the battery storage system 1 can be contacted in order to draw a total current I from it. ges to be extracted. The total current I ges consists of a converter current I supplied by the DC / DC converter 4 1d and a second current I2 supplied by the second battery 3. When the first battery 3 is discharged, the DC / DC converter 4 is supplied with a first current I1 from the first battery 2. When the first battery 3 is charged, the first battery 3 is supplied with the first current I1 by the DC / DC converter 4. The direction of flow of the first current I1 changes during this process.

[0038] To achieve efficient and gentle discharge of the battery storage system 1, it is operated using the methods according to the invention. The second current I2 is regulated according to an optimal control variable I2*, which defines a specific value for the second current I2. The second current I2 is regulated, for example, via the DC / DC converter 4. The first battery 2 has a first battery voltage U1. The second battery 3 has a second battery voltage U2.

[0039] The invention and the underlying mathematical principles are explained in more detail below.

[0040] The methods according to the invention enable energy-efficient control of a hybrid energy storage system, for example, the battery storage system 1, and each define one of two steps. In the first step, an optimal adjoint λ*, or a Lagrange multiplier, is found and provided offline, i.e., by calculating a priori known case examples. This can optionally be carried out for different case examples. In the second step, an associated adaptive control is then disclosed, which makes it possible to operate the energy storage system, for example, the battery storage system 1, in real time and to regulate the magnitude of the first current I1 and the second current I2.

[0041] The method for determining an optimized system behavior of a hybrid energy storage system solves an optimal control problem for the hybrid battery storage system.

[0042] First, a general optimization problem for optimal control problems is described. Subsequently, the optimal control problem specifically arising for the hybrid battery storage system 1 and its solution using Pontryagin's minimum principle are presented. It is described how the optimal adjoint λ* can be determined iteratively using a so-called bisection method according to the invention.

[0043] In an optimal control problem, a cost function J is to be minimized subject to given constraints. An example of such a problem is given below. The Lagrange goodness-of-fit measure L specifies the costs as a function of the state variables x(t) and the control variables u(t) for each point in time.

[0044] The optimal control problem can be formulated as follows: min{∀k: x_(t)∈X(t),u_(t)∈U(t)}J J :=∫tateL(x_(t),u_(t))dt st x_(t)∈X(t)⊆ℝn x_(t)∈U(t)⊆ℝm x_(ta)=x_0 x_(te)∈[x_te,min,x_te,max]

[0045] An optimal control variable u* then satisfies the following equation: U*=arg min{∀t: x(t)∈X(t),u(t)∈U(t)}(J(x_(t),u_(t)))

[0046] The optimization problem and its solution using Pontryagin's minimum principle will now be applied to the problem in Fig. The hybrid battery storage system 1 shown in Figure 1 is used as the basis for this procedure. The aim of the procedure is to minimize the losses in the hybrid battery storage system 1 over a given period [t]. a , t e ]. With electrical losses in the first battery P v,1 (t), electrical losses in the second battery P v,2 (t) and electrical losses in the DC / DC converter P v,dcdc (t) can determine the cost function J for the hybrid battery storage system 1 from Fig. 1. When minimizing the cost function J over the period [t] a , te The electrical losses in the battery storage system 1 shown are minimized: J=∫tate[Pv,1(t)+Pv,2(t)+Pv,dcdc(t)]dt

[0047] The total current I ges The following node equation is composed of the sum of the converter current I 1d on one of the sides of the DC / DC converter 4 facing the second battery 3 and the second current I2 of the second battery 3 together: Iges=I1d+I2

[0048] On one side of the first battery 2, in the case that the first battery 2 is discharged (this is the case in the battery storage system 1 shown, because the first current I1 is greater than 0 (I1 > 0)) with a DC / DC conversion efficiency η dcdc : I1=1ηdcdc⋅I1d⋅U2U1=ηdcdcz⋅I1d⋅U2U1 with z=−1 in the discharge case

[0049] Therefore, if the first battery 2 is being charged: I1=ηdcdc⋅I1d⋅U2U1=ηdcdcz⋅I1d⋅U2U1 with z=1 in the charging case

[0050] The state of charge of the first battery SoC1 and the state of charge of the second battery SoC2 can be determined based on the respective time integral of the current flowing through the respective battery 2, 3, the nominal charge (Q). nom,1 , or Q nom,2 ) as well as an initial state of charge (SoC) init,1 , SOC init,2 The calculations are shown in equations (12) and (13). The convention used here is that a positive current corresponds to the discharge of the respective battery 2, 3. SoC1(t)=SoCinit,1|t=ta−∫τ=tatI1(τ)Qnom,1dτ SoC2(t)=SoCinit,2|t=ta−∫τ=tatI2(τ)Qnom,2dτ

[0051] The preceding general form of an optimal control problem is now applied to the hybrid battery storage system 1. The state of charge of the second battery SoC2 is chosen as the state variable x (x = SoC2). The battery current of the second battery 3 represents the control variable, as can be seen from equation (14). u=I2

[0052] The aforementioned boundary and constraint conditions for the optimization parameters are formulated in equations (15) and (16) for application in the hybrid battery storage system 1. A starting state of charge SoC2(t a The state of charge (SoC2) of the second battery 3 is a fixed value. eThe state of charge of the second battery 3 should be in close proximity to this initial value, i.e., the initial state of charge SoC2(t1). For a battery whose state of charge ranges from 0% to 100%, values ​​around 50% are appropriate. The definition of the initial state of charge SoC2(ta) can also depend on other factors, such as the system configuration, voltage requirements, the choice of a connected inverter, or the characteristics of the energy storage device itself. SoC2(ta)=SoC2,taSoC2(te)=SoC2,te SoC2,min≤SoC2(t)≤SoC2,maxI2,min≤I2(t)≤I2,max

[0053] The solution to the optimization problem using Pontryagin's minimum principle is presented below. The Lagrange goodness-of-fit measure (see equation (2)) corresponds here to the sum of the electrical losses in batteries 2, 3 and the DC / DC converter 4 integrated in equation (8), as described by equation (17). L=Pv,1+Pv,2+Pv,dcdc

[0054] Equation (18) represents the defined Hamiltonian function H. Besides the Lagrange goodness-of-fit measure L, this includes another term, the product of a time-varying adjoint λ(t). T and a state differential equation f (x(t), u(t), t). H(x(t),u(t),λ(t),t)=L(x(t),u(t),t)+λ(t)T⋅f(x(t),u(t),t)

[0055] Pontryagin's minimum principle specifies a series of necessary conditions that must be met for optimality to hold. This is formulated in equations (19) to (23). Optimal trajectories and quantities are indicated here by the symbol (·)*. x˙∗(t)=∇λH|∗=herex˙2∗(t)=−u*Qnom,2=f(x2∗(t),u∗(t)) λ˙∗(t)=∇xH|∗see following explanations x(ta)=xa x(te)=xe

[0056] The time-dependent changes of the state x and the Lagrange multiplier (also called the adjoint λ) are described by the canonical differential equations (19) and (20). By choosing the optimal control variable u*, the value of the Hamiltonian H to be minimized according to equation (23) is always smaller than the value of the Hamiltonian H that results from a control variable that deviates from the optimal control variable u*. H(u(t),x∗(t),λ∗(t),t)≧H(u∗(t),x∗(t),λ∗(t),t),∀u(t)∈U(t),∀t∈[t0,te]

[0057] Now, over the period t ∈ [t a , t eAn optimal tractor u*(t) must be found from the range of permissible control values ​​u(t) that minimizes the Hamiltonian function H while adhering to the boundary and constraint conditions of the control and state variables. An extremum of the form dH / du=0 is not always necessary, as a minimum of the Hamiltonian function H may occur precisely at the boundary of the permissible control range where the derivative of the Hamiltonian function H is not yet 0. Therefore, the general requirement is u*=arg min H, where values ​​of the control variable u within the permissible control range are considered.

[0058] The solution of the two-point boundary value problem, consisting of state differential equation (19) and adjoint differential equation (20), can be solved under the assumptions described below.

[0059] The losses of the DC / DC converter 4 depend essentially on the input and output voltage, i.e., the first and second battery voltages U1, U2, as well as a transferred power P. v,dcdc dependent. Assuming that the battery voltage U1 of the second battery 3 remains constant during operation within a partial voltage range and thus within a small state of charge range x2(t) = SoC2(t) ≈ constant, the following simplification can be made in equation (24). ∂∂SoC2Pv,dcdc(U1,U2(SoC2(t)),I1d)≈U2≈constant0

[0060] For the partial derivation of a power loss of the second battery P v,2 With the assumption made, equation (25) applies. ∂∂SoC2Pv,2(Rdc,2(SoC2(t),I2))=∂∂SoC2(Rdc,2⋅I22)≈SoC2≈constant0

[0061] Power loss of the first battery P v,1is independent of the state of charge of the second battery SoC2 and therefore equation (26) applies to the partial derivative of this power loss. ∂∂SoC2Pv,1(Rdc,1,I1)=0

[0062] Furthermore, for the state differential equation with SoC2(t) ≈ constant, equation (27) applies. So˙C2(t)≈SoC2≈constant0

[0063] With the assumption made and the resulting partial differential equations (24) to (27), the necessary condition in equation (20) can be solved. It yields 0. It also follows that the adjoint λ is constant, as can be seen from equation (28). λ(t)=−δH(SoC2(t),u(t),λ(t),t)δSoCPB≈0⇒λ(t)=λ=constant

[0064] The losses in the battery parts, i.e. in the first battery 2 and the second battery 3, can be determined via the DC resistance of the respective battery part according to equations (29) and (30). Pv,1=Rdc,1⋅I12 Pv,2=Rdc,2⋅I22

[0065] Using equations (29) and (30), the node equation (9), the relationship u = I2 and the current efficiency equations (10) and (11) at the DC / DC converter 4, the electrical losses of the first battery P can be calculated. v,1 Rearrange and express according to equation (31). Here, z = -1 for the discharge case of the first battery 3 and z = +1 for the charging case. The current of the first battery I1 is then formulated according to equations (10) and (11). Pv,1(u)=Rdc,1⋅I12 =Rdc,1⋅(I1d⋅ηdcdcz⋅U2U1)2 =Rdc,1⋅((Iges−u)⋅ηdcdcz⋅U2U1)2

[0066] The output power P out,dcdc The DC / DC converter 4 can be used with a current direction-dependent factor k as a function of a discharge power P. 1d the first battery during a discharge, as also represented by equation (32). Pout,dcdc=P1d⋅k, with k={1Discharging battery part 1,I1>01ηdcdcz Charging battery part 1, I1<0

[0067] The power loss of the DC / DC converter 4 can now be represented as a function of the control variable u according to equation (33). The discharge power P can be determined in this context. 1d The following are displayed in the case of discharge of the first battery 3: P 1d = U2 · (I ges = u). In the case of charging the first battery 3, the following applies: P 1d = U2 · (I ges = u) · 1 / η dcdc . Based on these relationships, the power loss of the DC / DC converter P is calculated as follows: v,dcdc the following equation (33). Pv,dcdc=U2⋅(Iges−u)⋅k⋅(1−ηdcdc)

[0068] Using equations (17), (19), (28), (32) and (33), the Hamiltonian function H can be represented as a function of the control variable u = I2, as has been done in equation (34). H=L+λ⋅So˙C=Pv,2+Pv,1+Pv,dcdc−λ⋅uQnom,2=Rdc,2⋅u2+Rdc,1⋅((Iges−u)⋅ηdcdcz⋅U2U1)2 +U2(Iges−u)⋅k⋅(1−ηdcdc)−λ⋅uQnom,2

[0069] Pontryagin's minimum principle implies that the Hamiltonian function H at every time t ∈ [t a , t eThe Hamiltonian H has a global minimum within the given time period. Finding this minimum is equivalent to solving an optimization problem. An extremum can first be determined by differentiating the Hamiltonian H with respect to the control variable u and setting the result to zero. This fulfills a necessary condition for the validity of the extremum. It must then be verified whether the control variable u represents a global extremum at time t within the control range u ∈ U(t). Equation (35) shows the Hamiltonian H differentiated with respect to the control variable u, set to zero (∂H / ∂u = 0), and solved for the control variable u = I². The values ​​of the batteries 2, 3 (voltages, DC internal resistances) can be calculated at any given time as a function of the state variable and currents, or they can be fixed, like the nominal charge. Similarly, the efficiency of the DC / DC converter 4 can also be determined for the respective operating point.Only the optimal adjoint λ* needs to be determined. This is achieved accordingly by the inventive method for determining an optimized system behavior of a hybrid energy storage system. u=U2⋅k⋅(1−ηdcdc)+2⋅Iges⋅Rdc.1(ndcdcz⋅U2U1)2+λQnom.22⋅(Rdc,2+Rdc,1(ηdcdcz⋅U2U1)2)

[0070] The method for determining an optimized system behavior of a hybrid energy storage system therefore describes a method for determining the optimal constant adjoint λ* using a bisection method.

[0071] With the given equation (35) it is possible to determine, depending on the current I required by the battery storage system 1 ges The goal is to calculate the optimal control variable u*, or the optimal current I2*. However, the optimal adjoint λ*, for which the constraint and boundary conditions are satisfied, is still missing for the calculation.

[0072] The method for determining optimized system behavior of a hybrid energy storage system is an iterative loop method, the flow diagram of which is shown in Fig. 2 is shown. By means of the procedure for determining an optimized system behavior of a hybrid energy storage system, it is possible to determine the system behavior for a given period [t] over the specified period. a , t e ] a priori given performance profile P ges (t) the optimal (time constant) adjoint λ* can be determined. This is based on the assumption that the influence of the state of charge of the second battery 3 (SoC2) on the Hamiltonian function H is small.

[0073] At the beginning of the procedure, 40 initial values ​​for a possible adjoint λ are determined by initialization. m and model sizes of the hybrid battery storage system 1, in particular an initial state of charge of the first battery SoC1(t a) at the beginning of the specified period, an initial state of charge of the second battery SoC2(t a ) at the beginning of the specified period and a zero current at the beginning of the specified period, specified.

[0074] After initialization, a first iteration loop 10 is started. This loop has several iterations, with each successive iteration considering a possible adjoint λ. m is associated with this. Thus, when the first iteration loop is executed, a number of iterations are carried out in which the same procedure steps are processed. The same procedure steps are performed for varying values ​​of the possible adjoint λ. m The optimal adjoint λ* is then selected from the multitude of possible adjoints λ. mThis is determined using a so-called bisection method, also known as interval nesting. In the described embodiment, an interval halving method is used as an example.

[0075] For this purpose, a range of values ​​is defined which describes values ​​that represent the possible adjoint λ. m This range of values ​​can be assumed. It begins with an initial value λ. u and extends to a final value λ o This range of values ​​initially has a lowest starting value λ. u,0 and a highest final value λ o,0 on.

[0076] During the first iteration of the first iteration loop 10, a value of the possible adjoint λ is determined. m In selection step 14, a value is set which is between the initial value λ. u,0 and the highest final value λ o,0 lies. The value of the possible adjoint λ mThe first possible adjoint λ of the first iteration loop is selected. m,1 denoted. The first possible adjoint λ m,1 This results from the lowest initial value λ u,0 and the highest final value λ o,0 to: λ m,1 = (λ u,0 + λ o,0 ) / 2.

[0077] In subsequent iterations of the first iteration loop 10, the value of the possible adjoint λ is m newly defined. This involves determining whether a value of the optimal adjoint λ* is higher or lower than the value of the possible adjoint λ. m of the current iteration of the first iteration loop 10. This is done after the first iteration of the first iteration loop 10, for example, by checking whether the following criterion is met: [SoC2(te,λm,1)−SoC2,Ref]*[SoC2(te,λ0,1)−SoC2,Ref]<0

[0078] The variable SoC2(t) describes e , λ m,1) a charge state that results when applying the possible adjoint λ associated with this pass m , in the first iteration of the first iteration loop 10, i.e., when applying the first possible adjoint λ m,1 This results in a charge state of one of the respective possible adjoints λ. m associated final state of charge SoC2(t e ), which is for a final time step t e results. Accordingly, the variable SoC2(t) describes e ,λ 0,1 ) a charge state that results when the final value λ associated with this pass is applied o,0 The first iteration of the first iteration loop yields 10 if its value is used as the value for the possible adjoint λ. m is selected. This charge state is a value corresponding to the respective final value λ. o,0 associated final state of charge SoC2(t e ), which is for the last time step t e results.

[0079] Depending on whether this condition is met, the initial value λ will be u and the final value λ o The value for the next iteration of the first iteration loop 10 is set. If the condition is met, the initial value λ is set after the first iteration of the first iteration loop 10. u for the second iteration of the first iteration loop 10 set to λ u,1 = λ m,1 The final value λ o,0 The value of the first iteration loop 10 remains unchanged for the second iteration and thus results in: λ o,1 = λ o,0 If the condition is not met, the final value λ is used. o,0 For the second iteration of the first iteration loop, 10 is set to: λ o,1 = λ m,1 The initial value λ u,0 In this case, the value of the first iteration loop 10 remains unchanged for the second iteration and thus results in: λ u,1 = λ u,0 .

[0080] In each subsequent iteration of the first iteration loop 10, the possible adjoint λ is m The initial value λ is set to a new value. u and the final value λ o The value chosen is the one to which they were set in the previous iteration of loop 10. Accordingly, the possible adjoint λ is selected. m for each iteration of the first iteration loop 10 from the corresponding initial value λ u and the corresponding final value λ o determined. For the second iteration of the first iteration loop 10, the possible adjoint λ is obtained. m Thus, as follows: λ m,2 = (λ u,1 + λ o,1 ) / 2.

[0081] Accordingly, during the second iteration of the first iteration loop 10, it was checked whether the following criterion was met: [SoC2(te,λm,2)−SoC2,Ref]*[SoC2(te,λ0,2)−SoC2,Ref]<0

[0082] Thus, for each iteration of the first iteration loop 10, an initial value λ is assigned. u , a final value λ o and a related possible adjoint λ m determined.

[0083] The first iteration loop 10 is repeated until a termination criterion is met, which is detected in a check step 13 that is also performed during each iteration of the first iteration loop.

[0084] Each iteration of the first iteration loop (10) is followed by a second iteration loop (20). The second iteration loop (20) is executed multiple times, with different values ​​for the possible adjoint λ assigned to the process steps performed in the second iteration loop (20). m underlie.

[0085] The second iteration loop 20 has several iterations, with each successive iteration taking a time value t from a given period [ta , t e ] is assigned. Thus, when the second iteration loop is executed, a number of iterations are carried out in which the same process steps are processed. The same process steps are executed for varying values ​​of the time value t.

[0086] For this, the specified time period [t a , t e ] through an initial time t a and an end time t e defined. During the first iteration of the second iteration loop 20, a value of the time value t is equal to the starting time t. a The value is set. At the beginning of each subsequent iteration of the second iteration loop 20, this value is incremented in a time-value increment step 21. The second iteration loop 20 is executed until the value of the time value t equals the end time t. eThe second iteration loop 20 determines the behavior of the battery storage system 1 over the specified period [t]. a , t e ] simulated across. The time value t describes a point in time currently considered in the respective iteration of the second iteration loop 20 from the given period [t a , t e ].

[0087] In each iteration of the second iteration loop 20, an optimal control variable u* is determined for successive time points within the specified period. Thus, for each iteration of the second iteration loop 20, a corresponding optimal control variable u* is determined. This is selected for each iteration of the second iteration loop 20 from a multitude of possible control variables u. m determined.

[0088] The optimal control variable u* describes a current I2 to be delivered by the second battery 3. The optimal control variable thus describes an energy absorbed or delivered by the second battery 3. For example, a value of the optimal control variable u* is equal to a value to which the second current I2 is set for the time t considered in the respective iteration of the second iteration loop 20 within the specified period [t a , t e ] to be regulated in order to enable the lowest possible energy-efficient operation of the battery storage system 1 with the DC / DC converter.

[0089] The optimal control variable u* is determined and stored in each iteration of the second iteration loop 20. The optimal control variable u* is determined based on a minimum of the Hamiltonian function H. This determination of the optimal control variable u* is performed in each iteration of the second iteration loop 20 by executing a third iteration loop 30.

[0090] The third iteration loop 30 has several iterations, with each successive iteration representing a possible control value u. m is associated with this. Thus, when the third iteration loop is executed, a number of iterations are carried out in which the same procedural steps are processed. The same procedural steps are used for changing values ​​of the possible control value u. m executed.

[0091] For this purpose, a control value range is defined, which describes the values ​​that the possible control value can represent. m This control value range has a lowest initial value u and a highest final value u. During the first iteration of the third iteration loop 30, and during each iteration of the second iteration loop 20, a value of the possible control value u is determined. m The initial value of u is set. At the beginning of each subsequent iteration of the third iteration loop 10, this value is incremented in a u-increment step 31. The third iteration loop 30 is executed until the value of the possible control value u is reached. m is equal to the final value of u. The third iteration loop 30 is therefore performed for a number of possible control values ​​u. m through which the increase in the value of the possible tax value u m in the u-increment step 31 results.

[0092] In each iteration of the third iteration loop 30, a value of the Hamiltonian function H, as represented in formula (34), is calculated in a calculation step 11 according to the value of the possible control variable u associated with the respective iteration of the third iteration loop 30. m calculated. Therefore, when calculating the value of the Hamiltonian function H, a modeled system loss L is added. m with a derivation of the state of charge SoC2 of the second battery 3, which is combined with the possible adjoint λ m was multiplied. The modeled system loss L m is a parameter that describes the behavior of the energy storage system, i.e., battery storage system 1, and is therefore a modeled system parameter. In the Hamiltonian function H, the modeled system loss L is thus represented. mThe parameter to be optimized in the hybrid battery storage system 1, which is to be minimized, is also the state of charge (SoC2) of the second battery 3. This state is a storage state of the second energy storage system.

[0093] The derivation of the state of charge SoC2 of the second battery 3 is carried out using the value of the possible adjoint λ chosen for the present iteration of the first iteration loop 10. m , the value of the possible control variable u chosen for the present iteration of the third iteration loop 30 m and the nominal charge Q nom,2 calculated for the second battery 3 for the currently considered time.

[0094] The modeled system loss L m and the nominal charge Q nom,2The values ​​of the second battery 3 for the currently considered time are determined in a modeling step 32, which is executed before the calculation step 11 in the third iteration loop 30.

[0095] In modeling step 32, the modeled system loss L is m calculated. Within this procedure, this is referred to as "modeled" because the described method for determining optimized system behavior of a hybrid energy storage system can be executed by software without the battery storage system 1 being physically available. In formula (34), this is represented by the system loss L. The modeled system loss L m results from the electrical loss in the first battery P v,1 (t), the electrical losses in the second battery P v,2 (t) and the electrical losses in the DC / DC converter P v,dcdc(t). To calculate these electrical losses, it is necessary that the total current I required by the battery storage system 1 be known. ges The current required for the currently considered time is known. Therefore, in modeling step 32, the total current I required by battery storage system 1 is determined. ges according to a predefined performance profile P ges (t) determined. In the performance profile P ges (t) is for each point in time of the given period [t a , t e A required power output is stored, which is to be provided by the battery storage system 1. This required power output is read out by a query 22, which takes place within the second iteration loop 20, for the time point currently considered in the respective iteration of the second iteration loop 20 and made available for calculation step 11. Based on the required power output, the required total current I is calculated. gescalculated for the currently considered point in time. The other values ​​required for this are stored during initialization 40. The modeled system loss L m is thus carried out according to a predefined performance profile P ges (t) determines which of the battery storage system 1 in the specified period [t a , t e ] is traversed. The modeled system loss L m This is based on the DC resistance R dc,1 the first battery 2, the DC resistance R dc,2 the second battery 3, a power P to be delivered by the battery storage system 1 at the respective time according to the specified power profile P ges (t) and the efficiency η dcdc of the DC / DC converter 4.

[0096] Furthermore, in modeling step 32, the nominal charge Q is nom,2The state of charge (SoC) of the second battery 3 is calculated for the currently considered time. For this purpose, the state of charge (SoC2) of the second battery 3, which was stored during a previous iteration of the second iteration loop 20, is accessed. This state of charge of the second battery (SoC2) is updated according to formula (13) and made available for calculation step 11. Furthermore, this state of charge of the second battery (SoC2) is stored so that it can be accessed in a subsequent iteration of the second iteration loop 20. During the first iteration of the second iteration loop 20, if no state of charge of the second battery (SoC2) from a previous iteration of the second iteration loop 20 is available, the initial state of charge of the second battery (SoC2) provided during initialization 40 is used. a ) accessed.

[0097] For the calculation of the Hamiltonian function H in calculation step 11, an adjoint λ is also required, as can be seen from formula (34). The possible adjoint λ is used as the adjoint λ. m The function H, which corresponds to the current iteration of the first iteration loop 10, is used for the calculation. Formula (34) shows that the Hamiltonian H depends on a modeled system loss L. m , of the possible adjoints λ m and the state of charge of the second battery SoC2. The Hamiltonian function H belongs to battery storage system 1, since it is formulated by values ​​that describe battery storage system 1.

[0098] For each iteration of the third iteration loop, the value of the Hamiltonian function H obtained during calculation step 11 is stored. In calculation step 11, the smallest of the stored values ​​is determined. The smallest of the stored values, obtained in the final iteration of the third iteration loop, is the minimum of the Hamiltonian function H. The possible control variable u m The smallest of the stored values ​​is generated during the third iteration loop (30), and this value is determined as the optimal control variable u*. Thus, the possible control variable u is... m The optimal control variable u* is determined, at which the value of the Hamiltonian function H has a minimum value.

[0099] In each execution of the second iteration loop 20, a final state of charge (SoC) of the second battery is calculated. 2E (also known as SoC2(t)e ) denoted), which the second battery 3 after an expiry of the specified period [t a , t e ] exhibits when the second current I2 supplied by the second battery 3 is within the specified period [t a , t e ] was controlled according to the respective optimal control parameter u*. For this purpose, the state of charge of the second battery SoC2, stored in the last iteration of the second iteration loop 20 in modeling step 32, is accessed, which results when the possible control parameter u is used in the third iteration loop 30. m equal to the optimal control variable u*.

[0100] Once the second iteration loop 20 has finished executing, i.e., all iterations of the second iteration loop 20 have been completed, the first iteration loop 10 detects whether the final state of charge (SOC) of the second battery calculated using the second iteration loops 20 is correct. 2Ewithin a given charging interval and providing an optimal adjoint λ*, which corresponds to the possible adjoint λ of the first iteration loop 10 iteration for which it was detected that the final state of charge SOC 2E within the specified charging interval. The specified charging interval is an interval around an initial state of charge (SOC) of the second energy storage device. 2A , i.e., to determine the initial state of charge of the second battery SoC2(t a In this embodiment, it is checked whether the final state of charge (SOC) is reached. 2E the second battery 3 in an interval around the initial state of charge of the second battery SoC2(t) chosen during initialization 40 a ). This checks whether the final state of charge (SOC) is reached. 2E the second battery 3 for each iteration of the first iteration loop 10 less than 1% of the initial state of charge of the second battery SoC2(t a) deviates. If this is the case, the value of the possible adjoint λ corresponding to the respective iteration of the first iteration loop 10 is used. m as the value for the optimal adjoint λ*.

[0101] In the procedure for determining the optimized system behavior of a hybrid energy storage system, the solution of an optimization problem using an iterative method involves discretization both temporally and in terms of the model dynamics. After initialization in the first time step, the evolution of the model variables (states of charge in the batteries) can be updated for the subsequent time step. Then, for a predefined set of control values ​​u = I2, in the "u = I2 loop," i.e., the third iteration loop 30, each corresponding value of the Hamiltonian H for the given adjoint λ is calculated using equation (34). The set of control values ​​corresponds to the possible current values ​​of the first battery 3.

[0102] Subsequently, in the block “min(H)”, the current value for the second current I2 and thus the optimal control variable u* is determined, which minimizes the Hamiltonian function H, while adhering to the constraints.

[0103] Then, in test step 12, it is checked whether the entire performance profile L(t) up to the final time step t is maintained. e The calculation of the optimal current setpoint is performed for the entire problem and the possible adjoint λ specified in the step.

[0104] Is the calculation in the last time step t e Once arrived, it is checked whether the boundary condition that the final state of charge SoC2(t) is met. e ) from the second battery 3 the initial state of charge SoC2(t a ) corresponds to the given possible adjoint λ mThe limit was met. Since the model is discretized, a limit of, for example, SoC2(t2) + / - 1% can be chosen. If this is not the case, the procedure described below is followed.

[0105] In a further iteration of the first iteration loop, the procedure now selects further possible adjoints λ by iteratively approximating an optimal value for the optimal adjoint λ*. m from which, according to the described scheme, are incorporated into the calculation process in selection step 14.

[0106] It will be, as in Fig. 2, the process is repeated until the criterion SoC2(t) is met in test step 13. a ) ≈ SoC2(t2) is satisfied. Then the optimal adjoint λ* is determined. This criterion is the termination criterion. The optimal adjoint λ* is the possible adjoint λ m, which underlies the first iteration loop 10, in which the termination criterion is met.

[0107] In alternative embodiments of the method for determining optimized system behavior of a hybrid energy storage system (1), the optimal control variable (u*) is determined in each iteration of the second iteration loop (20) by analytical calculation. This involves a computational minimization of the Hamiltonian function given in equation (34). This can be done using algorithms that can, for example, also run in a control unit. In this case, the general optimization problem u* = arg min H applies. Thus, the requirement dH / du = 0 is changed to the more general requirement u* = arg min H. This is advantageous because the minimum of the Hamiltonian function may occur at the boundary of the limited range of values ​​of the control variable u. In this case, the condition dH / du = 0 does not apply. Therefore, the more general form is preferable.

[0108] The following refers to Fig. 3 explains the inventive method for controlling a hybrid energy storage system 1. In this method, the optimal adjoint λ* is adapted for real-time adaptive control.

[0109] The offline-determinable solution, described in the preceding procedure for determining optimized system behavior of a hybrid energy storage system, is now being implemented in a real-time capable control system. The goal is therefore to determine the appropriate response for any time-dependent power demand I. ges (t) to enable optimal control over a period of time. This results in causal energy management that does not require information from a prediction time horizon. This enables an adaptive process that operates based on the offline solutions predetermined by the method for determining optimized system behavior.

[0110] The optimal adjoint λ* calculated for a known problem can now be used, in conjunction with the relationships defined in equation (35), to control the current distribution between the current of the first battery I and the current of the other battery I, which must be determined at each time step. 1d (t) and the current of the second battery I2(t). For each time step, the current of the second battery I2(t) is calculated, which corresponds to a control variable u(t). Therefore: I2(t) = u(t).

[0111] If a control problem is solved that is similar to, but not exactly the same as, the off-line solution process, the final value criterion of a fixed charge state will not always be met.

[0112] Likewise, the end time t e unknown. To address this problem, an adaptive control system is introduced that compensates for deviations from a reference state of charge (SoC). 2,Sollcounteracts the second battery 3. This is achieved with a PI controller 51, which reacts to the difference between the target and actual state of charge of the second battery 3. The reference state of charge (SoC) 2,Soll This corresponds in particular to the reference state of charge (SoC). 2,Ref , which was chosen when determining the optimal first adjoint λ*.

[0113] The equation of the PI controller 51 is shown in formula (36). Depending on the control deviation between the target and actual state of charge of the second battery 3, excessively rapid discharging or charging is prevented. The state of charge remains within the range of the reference state of charge (SOC), depending on the controller parameters provided to the controller: a proportional value Kp and an integral value Ki. 2,Soll The magnitudes of the proportional value Kp and the integral value Ki can be found via a parameter study. λ˜(t)=λ*+Kp(SoC2,target−SoC2(t))+Ki∫tat(SoC2,target−SoC2(τ))dτ

[0114] In Fig. Figure 3 shows the controller structure of the hybrid battery storage system 1 embedded in the energy management system. This begins with a control difference between the target and actual state of charge of the second battery 3, with the specified optimal adjoint λ* and with the controller parameters.

[0115] This shows Fig. 3 a device 50 for controlling the hybrid battery storage system 1, comprising a control unit 54 which is configured to execute the method for controlling the hybrid battery storage system 1.

[0116] The device 50 comprises the PI controller 51, which has two inputs. A first input value and a second input value are acquired via these two inputs, wherein the first input value describes a pre-calculated value of the optimal adjoint λ* and the second input value represents a deviation of a state of charge SoC2 of the second battery 3, i.e., a storage state of the second battery 3, from a reference state of charge SOC. 2,Soll The value for the optimal adjoint λ* is provided, for example, by a memory module in which the optimal adjoint λ*, determined according to the procedure for determining optimized system behavior of a hybrid energy storage system, is stored. The reference state of charge (SOC) is also stored in this memory module. 2,Soll The value is stored and made available as a second input value for acquisition by the PL controller 51. The reference state of charge (SOC) 2,Sollis, for example, equal to a value of the initial state of charge of the first battery SoC1(t) chosen in the procedure for determining an optimized system behavior of a hybrid energy storage system. a ) at the beginning of the specified period [t a , t e ] was elected.

[0117] The PI controller 51 adjusts the optimal adjoint λ* for the current state of charge of the second battery SoC2 based on the first and second input values, i.e., based on the optimal adjoint λ* and the reference state of charge SOC. 2,Soll , to obtain a fitted adjoint λ a to determine this. First, the current state of charge of the second battery SOC2 is determined. This is done using at least one measurement taken from the second battery 2. The adapted adjoint λ a is determined according to formula (36) and provided at an output of the PI controller 51.

[0118] The adapted adjugate λ a The determined value is transmitted by the PI controller 51 to a control unit 52. The control unit 52 also has an input to which a target current value is provided. This target current describes the current to be supplied jointly by the first battery 2 and the second battery 3. Thus, the target current value is an energy target value, as it describes the energy to be supplied by the first battery and the second battery 3. This target current value is provided, for example, by external electronics that determine the current required by a load connected to the battery storage system 1. The target current value describes the required total current I. ges .

[0119] The control unit 52 determines a value of an optimal control variable I2* according to formula (35) (u = I2*), which describes the current to be delivered by the second battery 3. The optimal control variable I2* thus describes the energy absorbed or delivered by the second battery 3. The adjusted adjoint λ is used in this calculation. a The adjoint λ is used to calculate the optimal control variable I2* according to formula (35). The optimal control variable I2* is determined by storing formula (35) in the control unit 52 and solving it computationally. This can be done, for example, using a digital processing unit. Alternatively, it is advantageous for the optimal control variable I2* to be determined from a table by the control unit 52, whereby a table is queried in which different combinations of values ​​the adjusted adjoint λ is calculated. a and values ​​of the current target value I gesEach is assigned an optimal control variable I2*. This means that for all possible combinations of values ​​for the adjusted adjoint λ a and the current target value has already been calculated with a corresponding optimal control parameter I2*, and these have been stored in the control unit 52.

[0120] In both cases, the determination of an optimal control variable I2* is based on the adapted adjoint λ. a and the current target value I ges Regardless of whether the values ​​of the optimal control variable I2* were pre-calculated and are now selected, or whether they are calculated in real time, the optimal control variable I2* is determined according to a Hamiltonian function H associated with the battery storage system 1, since this is determined based on formula (35), which is obtained by solving the Hamiltonian function H described in formula (34). As previously described, this Hamiltonian function H depends on a calculated system loss L.b , the adapted adjoint λ a and the current state of charge (SoC2) of the second energy storage device. The calculated system loss L is... b In formula (35) the system loss L is to be understood as the system loss L. Thus, the system loss L is here referred to as the calculated system loss L b This is designated as such because it is calculated for the actually existing battery storage system 1.

[0121] As can also be seen from formula (34), the calculated system loss L b depending on the power loss of the first battery P v,1 , the power loss of the second battery P v,2 and the power loss of the DC / DC converter P v,dcdc .

[0122] The optimal control variable I2* determined by control unit 52 is provided to regulate the second current I2, i.e., the current supplied by the second battery 3. To regulate the first current I1, or in the case of the... Fig. 1 known battery storage system 1 the current at the output of the DC / DC converter 4, i.e. the converter current I 1d To control the current, a difference between the optimal control variable I2* and the target current value is determined. This is done, for example, using a differentiator 53.

[0123] Fig. Figure 4 shows a device 60 according to the invention for generating the optimal adjoint λ*. The device 60 for generating the optimal adjoint λ* comprises a computing unit 61, for example a processor, which is configured to execute the previously described method for determining an optimized system behavior of the hybrid battery storage system 1. The device 60 has an input 62 via which the values ​​necessary to define the battery storage system 1 can be provided. The device 60 has an output 63 via which the optimal adjoint λ* is provided.

[0124] It is further explicitly pointed out that the methods according to the invention can be carried out in a corresponding manner if the optimal control variable u* in the method for determining an optimized system behavior of the hybrid energy storage system 1 and the optimal control variable I2* in the method for controlling the hybrid energy storage system 1 describe the current to be supplied by the first battery storage system.

[0125] In addition to the above revelation, explicit reference is made to the revelation of Fig. 1 to 4 referred.

Claims

[1] Method for determining an optimized system behavior of a hybrid energy storage system (1) with at least a first energy storage device (2) and a second energy storage device (3) by determining an optimal adjoint (λ*), comprising: - Executing a first iteration loop (10) in several passes, where each successive pass is associated with a possible adjoint (λ), - wherein, for each iteration of the first iteration loop (10), a second iteration loop (20) is executed, which includes the steps: • Determining an optimal control variable (u*) for each successive point in time within a given period, where the optimal control variable (u*) is a quantity that describes the energy absorbed or released by the second energy storage device and is determined based on a minimum of a Hamiltonian function (H), where the Hamiltonian function (H) belongs to the energy storage system (1) and depends on a modeled system parameter (L) m ), the possible adjoint (λ) and a storage state (SoC2) of the second energy storage (3) is, where the modeled system parameter (L m ) is a parameter of the hybrid energy storage system (1) to be optimized, and is determined according to a predefined performance profile which is traversed by the energy storage system (1) in the predefined period, and • Calculating a final memory state (SOC) 2E ) of the second energy storage device (3), which the second energy storage device (3) exhibits after the specified period, if the energy absorbed or released by the second energy storage device (3) during the specified period was controlled according to the respective optimal control variable (u*), - Detect whether the final memory state (SOC) calculated using the second iteration loops (20) 2E ) of the second energy storage (3) lies within a specified interval and provide an optimal adjoint (λ*) which corresponds to the possible adjoint (λ) of the first iteration loop (10) for which it was detected that the final storage state (SOC) 2E ) lies within the specified interval. [2] Method according to claim 1, characterized by , that the optimal control variable (u*) describes a current to be delivered by the second energy storage device (3), and / or the modeled system parameter (L m ) is a modeled system loss, and / or the memory state is a load state, where the final memory state (SOC) 2E ) is a final charge state, and / or the interval is a charging interval. [3] Method according to claim 1, characterized by, that the first iteration loop (10) performs a bisection procedure to determine the possible adjoint (λ). [4] Method according to claim 1, characterized by , that - Determining the optimal control variable (u*) in each iteration of the second iteration loop (20) is accomplished by executing a third iteration loop (30), wherein successive iterations in the third iteration loop (30) each correspond to a possible control value (u), wherein in each iteration of the third iteration loop (30) a value of the Hamiltonian function (H) is calculated according to the possible control variable (u) associated with the respective iteration of the third iteration loop (30), and the possible control variable (u) is determined as the optimal control variable (u*) at which the value of the Hamiltonian function (H) has a minimum value, or - the determination of the optimal control variable (u*) in each iteration of the second iteration loop (20) is carried out by analytical calculation. [5] Method according to any one of the preceding claims, characterized by , that the specified interval is an interval around an initial memory state (SOC) 2A ) of the second energy storage device (3), which the second energy storage device (3) has at the beginning of the specified period. [6] Method according to any one of the preceding claims, characterized by , that the modeled system parameter (L m) is determined based on a DC resistance of the first energy storage device (2), a DC resistance of the second energy storage device (3), a power (P) to be delivered by the energy storage system (1) at the respective time according to the specified power profile and an efficiency of a DC / DC converter (4) via which the first energy storage device (2) and the second energy storage device (3) are connected in the energy storage system (1). [7] Method according to any of the preceding claims, characterized by , that when calculating the value of the Hamiltonian function (H) an addition of the modeled system parameter (L) m ) with a derivative of the storage state (SoC2) of the second energy storage (3), which was multiplied by the possible adjoint (λ). [8] Method for controlling a hybrid energy storage system (1) comprising at least a first energy storage device (2) and a second energy storage device (3), comprising: - Acquiring a first input value and a second input value, wherein the first input value describes a pre-calculated value of an optimal adjoint (λ*) and the second input value a deviation of a storage state (SoC2) of the second energy storage (3) from a reference storage state (SOC) 2,Soll ) describes, - Adjusting the optimal adjoint (λ*) for a current storage state (SoC2) of the second energy storage (3) based on the first and second input values ​​to obtain a fitted adjoint (λ a ) to determine, - Providing an energy target value (I ges ), which describes an energy that is to be jointly provided by the first energy storage device (2) and the second energy storage device (3), - Determining an optimal control variable (I2*) based on the fitted adjoint (λ) a ) and the energy target value (I ges ), where the optimal control variable (I2*) is a quantity that describes energy absorbed or released by the second energy storage device, and the optimal control variable (I2*) is determined according to a Hamiltonian function (H) associated with the energy storage system (1), wherein the Hamiltonian function (H) depends on a calculated system parameter (L) b ) of the hybrid energy storage system (1), the adapted adjoint (λ a ) and the current storage state (SoC2) of the second energy storage device. [9] Method according to claim 8, characterized by , that the storage state is a state of charge, and / or the energy target value (I ges) is a target current value, and / or the optimal control variable (I2*) describes a current to be delivered by the second energy storage device (3), and / or the calculated system parameter (L b ) a system loss of the hybrid energy storage system (1). [10] Method according to one of claims 8 or 9, characterized by , that the calculated system parameter (L b ) is dependent on a power loss of the first energy storage device (2) and a power loss of the second energy storage device (3) and in particular is further dependent on a power loss of an energy converter (4) via which the first energy storage device (2) and the second energy storage device (3) are connected in the energy storage system (1). [11] Method according to any one of claims 8 to 10 above, characterized by, that the optimal control variable (I2') is determined computationally, or the optimal control variable (I2') is determined in tabular form, whereby in tabular determination a table is queried in which different combinations of values ​​of the adjusted adjoint (λ) a ) and values ​​of the current target value (I ges ) each is assigned an optimal control variable (I2*). [12] Method according to any one of claims 8 to 11 above, characterized by , that the adjustment of the optimal adjoint (λ*) is carried out using a Pl controller (7). [13] Method according to any one of claims 8 to 12 above, characterized by , that the optimal adjoint (λ*) is generated by the method according to one of claims 1 to 6. [14] Device for generating an optimal adjoint (λ*), comprising a computing unit which is configured to perform the method for determining an optimized system behavior of a hybrid energy storage system (1) according to any one of claims 1 to 7. [15] Device (50) for controlling a hybrid energy storage system, comprising a control unit (54) which is configured to perform the method for controlling a hybrid energy storage system according to any one of claims 8 to 13.

Images