Design of multiple frequency filters arranged in distributed manner in onboard energy network of vehicle

Abstract

The invention relates to a method (S1-S14) for designing a plurality of frequency filters (BP1, BP2, BPi) arranged in a distributed manner in an onboard energy network (EBN) of a vehicle (F), comprising at least the following steps: (a) establishing a simulation model of the onboard energy network (EBN), the airborne energy network has a predetermined number of frequency filters (BP1, BP2, BPi) (S1) at defined positions of the airborne energy network; (b) determining an initial set of filter parameters of the frequency filter (S2, S3); (c) calculating a simulation model in the time domain with the set of filter parameters (S5); (d) extracting a safety-critical trajectory from the calculated simulation model together with an interference function based on (S6); (e) transforming the security-critical trajectory from the time domain into the frequency domain (S7); (f) determining a transfer function in the mathematical s-domain by means of system identification (S8); (g) extracting a pair of residuals and poles from said transfer function in the s domain (S9); (h) determining an energy loss value (EL) and a voltage stability value (Qu) based on the extracted residuals and poles (S10); (i) changing the filter parameter set of the frequency filters (BP1, BP2, BPi) (S12); (j) repeating steps (c) to (h) with the set of filter parameters changed in step (i) until a termination criterion is reached; (k) after reaching a termination criterion: selecting, from the set of points having the energy loss value (EL) and the voltage stability value (Qu) determined in step (h), a point (p4) located on the Pareto front (PAR) of the set of points (S13, S14). The invention is particularly advantageously applicable to all-electrically driven vehicles.

Description

Technical Field

[0001] The invention relates to a method for designing a plurality of frequency filters which are arranged in a distributed manner in an onboard energy network of a vehicle, the method comprising at least the following steps: establishing a simulation model of the onboard energy network, the onboard energy network having a predetermined number of frequency filters at defined positions of the onboard energy network. The invention also relates to an onboard energy network of a vehicle, the onboard energy network having a plurality of distributed frequency filters, wherein the frequency filters are determined by means of the method. The invention also relates to a vehicle having such an onboard energy network. The invention can be applied particularly advantageously to vehicles with a fully electric drive. Background Art

[0002] The onboard energy network of a vehicle is composed of a locally extended cable bundle and connected components or loads. The resulting RLC network is vibrating due to the existing resonance. The excitation of the system causes voltage oscillations at the terminals of the connected components. A feasible solution to suppress critical voltage oscillations is to integrate a passive bandpass filter into the onboard energy network. The design of a single passive bandpass filter can be implemented algorithmically by parameterizing the filter parameters resistance R, inductance L, and capacitance C of the passive bandpass filter. In principle, in most cases, large capacitance provides improved voltage stability, but the power loss generated in the passive filter becomes larger as a result. In contrast, the parameterization of multiple bandpass filters present in the onboard energy network (for example, in different distributors) is more complicated to stack because the multiple bandpass filters may affect each other.

[0003] In order to reduce the effects of highly dynamic power fluctuations, multi-stage bandpass filters can be used. Thus, Martin Baumann, Ali Shoar Abouzari, Christoph Weissinger, Bjørn Gustavsen, Hans-Georg Herzog: "Passive Filter Design Algorithm for Transient Stabilization of Automotive Power Systems", 2021 IEEE 93rd Vehicular Technology Conference (VTC2021-Spring), April 25-28, 2021, Helsinki, Finland, describes that automotive onboard networks are increasingly expanded with highly dynamic power electronics. These components can cause malfunctions or failures of safety-related low-voltage components. This susceptibility to interference is often reduced by using oversized passive input electronics. In the publication, an alternative way of suppressing interference is proposed by introducing system-integrated adaptive passive filters. A method proposal is proposed for studying the suitability of potential access points within a complex network. An algorithmic method for parameterizing multiple switchable bandpass filter stages is described. Measurements in vehicles indicate the effectiveness of the determined size filter, which can reduce interference at 70kHz by more than 75%. The publication is fully incorporated into the present disclosure. Summary of the invention

[0004] The object of the present invention is to at least partially overcome the disadvantages of the prior art and in particular to provide an onboard energy network of a vehicle, which has a plurality of distributed frequency filters, wherein the frequency filters have particularly low power loss and achieve good voltage stability of the terminal voltage of safety-critical components of the onboard energy network.

[0005] This object is achieved according to the features of the independent claim. Preferred embodiments can be drawn in particular from the dependent claims.

[0006] The object is achieved by a method for designing a plurality of frequency filters which are arranged in a distributed manner in an onboard power grid of a vehicle, the method comprising at least the following steps:

[0007] (a) establishing a simulation model of an onboard power network having a predetermined number of frequency filters at defined locations or access points of the onboard power network;

[0008] (b) determining an initial set of filter parameters of the frequency filter;

[0009] (c) calculating a simulation model in the time domain using the filter parameter set;

[0010] (d) extracting safety-critical trajectories together with the underlying disturbance functions from the calculated simulation model;

[0011] (e) transforming safety-critical trajectories from the time domain to the frequency domain;

[0012] (f) determining the transfer function in the s-domain by means of system identification;

[0013] (g) extracting pairs of residuals and poles according to the transfer function in the s-domain;

[0014] (h) determining an energy loss value and a voltage stability value based on the extracted residuals and poles;

[0015] (i) changing a filter parameter set of a frequency filter;

[0016] (j) repeating steps (c) to (h) with the set of filter parameters changed in step (i) until a termination criterion is reached;

[0017] (k) After reaching the termination criterion: selecting a point on the Pareto front of the set of Pareto optimized points from the set of points having the energy loss value and the voltage stability value determined in step (h).

[0018] The method achieves the advantage that an optimized filter parameter set can be found for an onboard power system having a plurality of frequency filters, wherein cross-influences between the individual frequency filters are also taken into account.

[0019] The vehicle can be, for example, a vehicle with an internal combustion engine, a hybrid vehicle, or a fully electric vehicle. The vehicle can be, for example, a land vehicle, such as a passenger car, a motorcycle, a bus, a truck, etc., an air vehicle, such as an airplane, a helicopter, etc., or a water vehicle, such as a ship, etc. The onboard energy network is used for energy supply for the components or electrical appliances connected thereto. The components are in principle arbitrary and can include safety-related ("ASIL") components and / or non-safety-related comfort ("QM") components. Safety-related components can, for example, include an integrated braking system, an electric drive servo steering system, a windshield wiper motor, and comfort components can include, for example, an electric fan, a rear axle steering system, etc.

[0020] The frequency filter is used to reduce voltage oscillations outside the permitted frequency band that occur in the distributor. For this purpose, the frequency filter or its filter stage can be constructed to have or be constructed as any suitable frequency filter in principle, such as a bandpass, a bandstop, a lowpass and / or a highpass. In particular, the frequency filter or the filter stage is a passive filter, in particular a passive bandpass filter. "The frequency filter is a passive filter" in particular includes: its filter characteristics can be expressed or defined by filter parameters in the form of filter stage resistance, filter stage capacitance and filter stage inductance. The filter stage frequency corresponds in particular to the center frequency of the associated passband.

[0021] The frequency filter may include one or more filter stages. The determination of the filter parameters includes determining the filter parameters for each filter stage in the filter stage and can therefore also be understood as the determination of the filter stage parameters. The determination of the filter parameter group of the frequency filter includes determining the specific selection of the filter (stage) parameters of all filters (stages). A further improvement is that the filter stages of the multi-stage frequency filter are permanently connected. A further improvement is that the filter stages of the multi-stage frequency filter can be connected and disconnected via corresponding switches, particularly electronic switches.

[0022] The establishment of a simulation model for an electrical network, in particular an onboard power grid, according to step (a) is known in principle and is described, for example, in Chapter III in a publication by Martin Baumann et al. Here, the model of the onboard power network includes interference sources and interference sinks. Interference sources can be, for example, electrical consumers or energy sources, such as DC voltage converters, comfort components and / or highly dynamic safety-related (ASIL) components, which can cause highly dynamic power fluctuations in the onboard power network. Interference sinks can be, for example, highly sensitive and / or safety-related electrical consumers, such as integrated braking systems, electric drive servo steering systems, windshield wiper motors, sensors or computing devices. Interference sinks can also be current supply devices such as DC voltage converters. A component can be both an interference source and an interference sink. Interference sources, interference sinks and frequency filters are connected to various access points of the onboard network. The frequency filter is connected to the electrical path in order to reduce interference, in particular interference caused by interference sources at the interference sink (for example, at its connection terminals).

[0023] The calculation of the simulation model in the time domain with the filter parameter set or in the time domain comprises, in the first execution of step (c), the initial filter parameter set selected in step (b) and in the further executions, the set previously changed in step (j).

[0024] A “safety-critical” trajectory is understood in particular to be the profile of the terminal voltage u(t) at a highly sensitive and / or safety-critical component, which is a response to the profile of a disturbance function in the form of an interference current i(t) produced by a source of interference. q interference sources and n s If there are interference sinks, then calculate n t =(n s ·n g ) safety-critical trajectories. The extraction in step (d) corresponds to a calculation in the time domain using the model.

[0025] The transformation in step (e) may, for example, comprise a Fourier transform, such as a discrete Fourier transform, ie, DFT.

[0026] The determination of the transfer function in the mathematical s-domain by means of system identification and the extraction of the pairing of residuals and poles from the transfer function in the s-domain can be performed in an approximate manner, for example, in a manner similar to that described in the publication by Martin Baumann et al. in Chapter III. In general, for each approximated transfer function in the s-domain (where s is the Laplace variable), a plurality of pairs of residuals and poles are extracted or determined, see, for example, equation (1) in Chapter III of the publication by Martin Baumann et al., where, for example, N=4 pairs usually already approximate the original transfer function with high accuracy. In a similar manner to equation (1) in the publication by Martin Baumann et al., the approximated transfer function is in particular a sum of polynomial fractional terms.

[0027] In step (h), the generally multiple pairs of residuals and poles for all transfer functions and thus all safety-critical trajectories are used to determine or calculate a single energy loss value E for the current filter (stage) parameter set. L and a unique voltage stability value Q u This corresponds to a point in the following diagram with the energy loss E caused by the frequency filter L and the voltage stability Q of the terminal voltage of the considered safety-critical component u As an axis.

[0028] Changing the set of filter parameters of the frequency filter (stage) in step (i) comprises changing at least one filter (stage) parameter compared to the previous set.

[0029] The termination criterion in step (j) can in principle be selected arbitrarily and can include, for example, reaching a predefined number of repetitions of steps (c) to (h) and / or falling short of a distance between two successive points.

[0030] After reaching the termination criterion, in step (k), in particular the energy loss value E determined in each case in step (h) is determined. L and voltage stability value Q u A set of points, ie a point corresponds to a pair consisting of an energy loss value and a voltage stability value determined in the same execution. The number of sets of points corresponds to the number of repetitions plus the initial execution.

[0031] From the set of points, a Pareto front (also called a Pareto set) can be found with the aid of known Pareto optimization methods, which Pareto front includes a set of Pareto optimized points. A point can be selected from this set in order to specifically determine the filter (stage) parameters of the frequency filter (stage). This is possible because for (E L ;Q u ) space stores a corresponding set of filter parameters.

[0032] This configuration of the filter parameter set is modified by means of an optimization method. This achieves the advantage of a particularly fast calculation of the Pareto-optimized points.

[0033] One design solution is to change the filter parameter set by means of a particle swarm optimization method. Particle swarm optimization is particularly suitable for the current optimization problem, that is, it has fast convergence. Particle swarm optimization (Partikelschwarmoptimierung, PSO) refers to a known, nature-like optimization method that seeks solutions to optimization problems based on the behavior of biological groups. However, in principle, other multi-criteria optimization methods, such as memetic algorithms, Pareto search, global search, etc., can also be used.

[0034] One design scheme is that, in order to implement step (b), a solution space is determined in sub-step (b1), the solution space includes feasible values ​​of filter parameters of the frequency filter, which are based on the specific available values ​​of the filter resistance, filter capacitance and filter inductance, and in sub-step (b2), an initial filter parameter group is found from the solution space. Therefore, the advantage of a particularly practical design of the filter parameters is achieved. The solution space of the simulation model is determined by the maximum and minimum values ​​of passive components such as resistors, capacitors and inductors available for use in order to construct the filter. In addition, there are parasitic resistances of capacitors and inductors. The values ​​of the resistance, capacitance and inductance of the filter stage are not limited to the values ​​of the individual available components, but may also include series and / or parallel circuits of these elements.

[0035] A further development consists in that the passband frequency of the passive bandpass filter lies in a predetermined frequency range, in particular in a range between 10 kHz and 150 kHz.

[0036] One design solution is to select the initial filter parameter set based on a stability boundary condition with a corresponding maximum filter (stage) capacitance, a balance boundary condition with a minimum filter (stage) capacitance, or a loss boundary condition with a target filter (stage) capacitance between the stability boundary condition and the balance boundary condition. The stability boundary condition leads to relatively high losses in the case of particularly stable filtering; the balance boundary condition leads to relatively low losses in the case of high frequency selectivity; and the loss boundary condition leads to a compromise between these two extremes. Therefore, the selection of the stability boundary condition has the following advantages: a particularly high, in particular maximum stability Q is taken into account u The selection of a balanced boundary condition therefore results in the following advantages: it is thus possible to take into account a particularly low, in particular minimal, loss E L The loss boundary condition is advantageously chosen to be generated at a point between the above two points.

[0037] A further improvement is to find the initial filter parameter set according to one of the three boundary conditions mentioned above, and to select the next two variants according to the other two boundary conditions in step (i). L ; Q u ) space, which already form a very favorable starting situation for the selected optimization algorithm to converge quickly. This in turn facilitates the subsequent optimization of the filter parameters and shortens the calculation time for calculating other Pareto optimal points. It is therefore particularly advantageous to initially perform steps (c) to (h) with the three boundary conditions and to change the filter parameter set by means of the optimization method only from the fourth execution of steps (c) to (h).

[0038] One design is that, in order to parameterize the frequency filters, initial values ​​of the filter parameters for each frequency filter are determined by means of an algorithmic calculation. The algorithmic determination of the initial parameters for each frequency filter can be implemented, for example, according to the method described in Chapter IV of the publication by Martin Baumann et al.

[0039] Another design is that the initial filter parameter group is selected arbitrarily from the solution space. This is advantageously particularly simple, but may increase the total time for implementing the method. A further improvement is to arbitrarily change at least one filter parameter group following the initial group. A further improvement is to change the group following a certain number of arbitrarily changed groups by means of an optimization method, that is, the arbitrarily changed groups only form a subset of each of the groups, and subsequent optimization is constructed on the subset. The certain number can be, for example, three.

[0040] The object is also achieved by an onboard energy network of a vehicle having a plurality of distributed frequency filters, wherein the filter parameters of the frequency filters are determined according to the method as described above. The onboard energy network can be constructed similarly to the method and vice versa, and the onboard energy network has the same advantages. In particular, the frequency filters are designed according to the following filter parameter groups, which are stored for the selected points located on the Pareto front in the steps by means of the method described above.

[0041] One design solution is that at least one of the frequency filters is integrated into a distributor. This has the following advantages: only a few frequency filters are required and the frequency filters can be integrated into the onboard power grid particularly easily. The distributor is used to connect a plurality of electrical components (loads, electrical consumers, etc.) to the onboard power grid and for this purpose the distributor has in particular corresponding current branches with corresponding interfaces or connection terminals. The connectable components are in principle arbitrary and can include safety-related (ASIL) components and / or non-safety-related comfort (QM) components.

[0042] The power distributor is in particular an “electronic” power distributor, so that at least some of the current branches, in particular all current branches, to which corresponding electrical components can be connected, are protected by corresponding electronic fuses.

[0043] In one embodiment, at least one of the frequency filters is a multi-stage filter. This has the advantage that resonances present in the onboard power grid can be filtered out particularly effectively (particularly with regard to safety-related components) and thus its terminal voltage can be stabilized, for example in the event of current / power transients.

[0044] In one embodiment, the onboard energy network is a low-voltage onboard energy network, in particular a low-voltage onboard energy sub-network. The low-voltage onboard energy network can, for example, have a nominal voltage between 12 V and 60 V. If the onboard energy network is a low-voltage onboard energy sub-network, it can be part of an overall onboard energy network, which also has a high-voltage onboard energy sub-network, which, for example, has a nominal voltage higher than the nominal voltage of the low-voltage onboard energy sub-network, for example, between 48 V and 800 V or even higher.

[0045] The object is also achieved by a vehicle having an onboard energy network as described above. The onboard energy network can be designed similarly to the onboard energy network and vice versa, and the onboard energy network has the same advantages. The vehicle can therefore have an overall onboard energy network, which has the low-voltage onboard energy sub-network and the high-voltage onboard energy sub-network, wherein, in a further development, the low-voltage onboard energy sub-network can be fed with electrical energy from the high-voltage onboard energy sub-network.

[0046] One embodiment is that the vehicle is an all-electric or battery-driven vehicle. The method and the onboard power grid are particularly advantageous for this purpose, since the stability of the onboard power grid is particularly important in all-electric vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The above-mentioned characteristics, features and advantages of the present invention and the implementation thereof will become clearer and easier to understand in conjunction with the following illustrative description of embodiments, which are explained in more detail in conjunction with the accompanying drawings.

[0048] FIG. 1 shows a detail of an equivalent circuit diagram of a possible onboard energy network of a vehicle with multiple frequency filters; and

[0049] Figure 2 An abstract equivalent circuit diagram showing a multi-stage frequency filter of the onboard energy network in FIG1 ;

[0050] Figure 3 A possible flow chart of a method for designing filter parameters of a frequency filter of an onboard energy network in FIG. 1 is shown;

[0051] Figure 4 A voltage oscillation diagram is shown in a (V, t) diagram; and

[0052] Figure 5 The (E L ;Q u ) point (E L ;Q u )picture. DETAILED DESCRIPTION

[0053] FIG. 1 shows, with its sub-images FIG. 1 (A) and FIG. 1 (B), a detail of an equivalent circuit diagram of an onboard energy network EBN of a vehicle F. The onboard energy network EBN has a low voltage (Niedervolt, NV) onboard subnetwork, which can be supplied from a high voltage (Hochvolt, HV) onboard subnetwork with a high voltage network voltage VHV via a plurality of electrically isolated DC voltage converters GSW1, GSW2, GSW3. The DC voltage converters GSW1, GSW2, GSW3 convert the higher HV network voltage VHV of the HV onboard subnetwork into the corresponding lower NV voltage V_CM, V_CS1 or V_CS2. The DC voltage converters GSW1, GSW2, GSW3 are connected to a communication channel of the vehicle F, here for example a CAN bus CAN, and can communicate with each other via said communication channel. In particular, the DC voltage converters GSW1 , GSW2 , GSW3 can be present in a master-slave arrangement, wherein, for example, the DC voltage converter GSW1 acts as a master and the DC voltage converters GSW2 , GSW3 act as slaves.

[0054] Here, the NV voltages V_CM, V_CS1 and V_CS2 are applied to the locally extended cable bundle KB of the NV sub-onboard network via the respective nodes A, B or C and more precisely via lines with the respective line impedances Z_L,CM, Z_L,CS1 and Z_L,CS2. Adjacent nodes A and B, B and C, etc. of the extended cable bundle KB are generally separated from one another by line sections with line impedances Z_L,AB, Z_L,BC, etc.

[0055] The battery BAT is also connected to the cable harness KB, here for example at the node A via a line having a line impedance Z_L,BAT.

[0056] Furthermore, a plurality of electrical components are connected to the cable harness KB of the NV onboard network, and more precisely via at least one (here two) power distribution units PD1, PD2. Currently, power distribution units PD1 and PD2 are connected to nodes B or C via line impedances Z_L, PD1 and Z_L, PD2, respectively. Not only safety-related (ASIL) components such as integrated brake system (IB), electric servo steering system (EPB), windshield wiper motor (WIP), etc., but also comfort (QM) components such as electric fan (ELF) and rear axle steering system (RAS) can be connected to power distribution units PD1, PD2 via separate current branches. The impedances present in the individual current branches are shown here as line impedances Z_MNH,P and as impedances Z_MN,P of the connected components, where "M" is the number of the distribution unit PD1, PD2, "N" is the name for the current branch, where "A" is used for ASIL components, "Q" is used for comfort components, and "P" is the number of the ASIL component or comfort component in the respective distribution unit PD1, PD2.

[0057] In order to avoid interference, in particular interference of safety-related components IB, EPB, WIP, a multi-stage bandpass filter BP1 or BP2 (having respective impedances Zf,PD1 and Zf,PD2) is provided in each of the power distribution units PD1, PD2, electrically in parallel with the respectively connected electrical components, each of which has a plurality of selectively switchable and disconnectable stages FS. j The filters BP1 and BP2 can be designed to be identical in structure or - for example - in the filter stage FS j in terms of the number and / or filtering characteristics of the

[0058] The multi-stage bandpass filters BP1, BP2 and the electrical components IB, EPB, WIP, ELF, RAS connected to the power distribution boxes PD1, PD2 are connected via corresponding electronic fuses or "electronic fuses" EFi (see Figure 2 )Protect.

[0059] In principle, one or more power distribution units PDi (i≥1) can be present in the NV onboard subnetwork, i.e. also more than two power distribution units PD1, PD2. In particular, at least one safety-related component IB, EPB, WIP and / or at least one comfort component ELF, RAS can be connected to each power distribution unit PDi in the power distribution unit.

[0060] In order to further reduce interference, in particular interference of (especially also safety-related) components IB, EPB, WIP, ELF, RAS, etc., additional frequency filters and / or electronic fuses (not shown) can be present in the associated individual current branches, which are provided specifically for protecting the corresponding components IB, EPB, WIP, ELF, RAS, etc. In the presence of filters BP1 and BP2, the additional frequency filters can advantageously be smaller than without filters BP1 and BP2.

[0061] Figure 2 The abstract equivalent circuit diagram shows a multi-stage passive bandpass filter BPi integrated in one of the power distribution boxes PDi of the onboard power grid EBN. The filter BPi is optionally protected by a low-impedance electronic fuse EFi, which can be switched off via a switch Q fuse The electronic fuse EFi is located in the same separate current branch as the bandpass filter BPi and is advantageously connected upstream of the bandpass filter and, if necessary, installed in a common module with the bandpass filter BPi.

[0062] The multi-stage filter BPi includes k=1,...,k max (Here k max >2) parallel bandpass filter stages FS n , the bandpass filter stages have their own fixed parameterization by the ohmic resistor R n 、Inductance L k and capacitor C k The combination of , whose value can be regarded as the filter parameter. Bandpass filter stage FS k can optionally be switched via a corresponding electronic switch Q k Selectively connect to or disconnect from the grid. Electronic switch Q k Here it is designed as a low-impedance MOSFET.

[0063] Filter stage FS n It is protected against short-duration voltage pulses by the bidirectional suppression diode D1 inside the multistage filter BPi.

[0064] Electronic switch Q kIt is controlled by means of a gate driver GT of the power distribution unit PDi, which in turn is set by means of a control unit ECU (e.g. in the form of a microcontroller) of the power distribution unit PDi. The control unit ECU receives changes in the switching state of the NV sub-onboard network via the CAN bus CAN. The control unit ECU compares the switching state of the NV sub-onboard network or the change in the actual switching state at the time with a database, in particular a lookup table stored therein, in which a corresponding discrete (switching) configuration of the filter BPi is assigned to each possible switching state of the NV sub-onboard network. The discrete configuration of the filter BPi is particularly understood as the discrete configuration of each electronic switch Q j For this purpose, the configuration of the filter BPi stored for a specific switching state of the NV sub-onboard network is the most suitable discrete configuration of the filter BPi for suppressing the resonance in the NV sub-onboard network in this switching state of the NV sub-onboard network. Thus, for each switching state of the NV sub-onboard network there is (in particular, exactly one) of the BPi filters through the various stages FS k In other words, the control unit ECU interprets or ascertains the electronic switch Q of the filter BPi based on the information about the switching state of the NV sub-onboard network transmitted via the CAN bus CAN. k The configuration to which it belongs.

[0065] The discrete switching states of the filter BPi found by the control unit ECU are further transmitted to the gate driver GT, which switches the electronic switches Q accordingly. k .

[0066] System voltage V s Corresponds to the voltage detected at the power distribution unit PDi between the nodes B, C, etc. of said power distribution unit and the local reference potential (ground).

[0067] Figure 3 The filter parameters R used to design the frequency filters BP1, BP2 or BPi are shown. n , L k and C k A feasible process.

[0068] In step S1 , a simulation model of an onboard power grid, for example an onboard power grid EBN, is created (for example in the programming language PYTHON) with a predetermined number of frequency filters BP1 , BP2 or BPi at defined locations or access points of the onboard power grid.

[0069] In step S2, a solution space is determined, which includes the frequency filters BP1, BP2 or BPi or the filter stages FS1, FS2, FS k The filter parameters R n , L k and C k feasible values ​​based on the specific available values ​​of the filter resistance, filter capacitance and filter inductance.

[0070] In step S3, filter levels FS1, FS2, FS3 for all frequency filters BP1, BP2 or BPi or the frequency filters are found from the solution space. k The initial set of filter parameters R n , L k and C k , for example, based on stability boundary conditions, equilibrium boundary conditions or loss boundary conditions.

[0071] In step S4, the initial set of filter parameters R n , L k and C k The incorporation into the simulation model can also be referred to as parameterization of the simulation model.

[0072] In step S5 , the parameterized simulation model is prepared (eg compiled) for use in a simulation computing environment (eg Modelica, an object-oriented modeling language for physical models) and then calculated in the time domain in the simulation computing environment in step S6 .

[0073] The calculation in step S6 results in a safety-critical trajectory which is extracted and transformed or converted together with the underlying disturbance function into the frequency domain to form a complex safety-critical trajectory in step S7 .

[0074] In step S8 , the transformed trajectory is used to determine or calculate a corresponding transfer function having a Laplace variable s in the mathematical s-domain by means of a system identification (which is implemented in MATLAB Runtime, for example).

[0075] In step S9, at least one pair, generally multiple pairs, of residuals r (referred to as “c n ”) and the extreme point p (called “a n ”), and more precisely in particular by using an approximate transfer function in the s-domain.

[0076] In step S10, the energy loss E is calculated from all the extracted residuals and pole pairs. L The value and voltage stability Q u The value of (E L ;Q u ) The coordinate values ​​of the points in the figure are obtained. Each point corresponds to a specific set of filter parameters R for all filters BPi or filter stages FSi of the considered airborne energy network EBN n , L k and C k .

[0077] This can be achieved in particular taking into account the following considerations:

[0078] i) The on-board energy network EBN has a favourable voltage stability if high load fluctuations at the interference source only influence the terminal voltage at the interference sink to a small extent. Figure 4 The concept for this definition of voltage stability is shown. For example, a step of, for example, 1 A at a source / load of interference results in a voltage oscillation u at the interface of the considered interference sink that decays over time t. TF (t). The voltage oscillation u TF (t) has a certain amplitude and the duration of the transition state is t s and voltage-time-integral Q u,t The voltage time integral Q u,t is used to quantify voltage stability. Therefore, not only the step response amplitude but also the transition state duration t s All are taken into consideration.

[0079] ii) To this end, the complex pole p found in step S9 is investigated with respect to its oscillation tendency. For a single pair consisting of a complex pole p and its corresponding complex residual (Rest), the second-order transfer function G is established as follows PF (s):

[0080]

[0081] Among them, s is the Laplace coordinate, r is the complex residual, p is the complex pole, is the complex conjugate residual, and are complex conjugate poles.

[0082] iii) The transfer function G PF (s) is transformed from the mathematical s domain to the time domain, and the voltage oscillation u is obtained according to the following formula TF (t):

[0083]

[0084] in,

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091] The voltage oscillation u TF (t) to obtain its envelope u env (t):

[0092]

[0093] In which u is replaced by the constant K TF The trigonometric expression in (t) describes an oscillation with a certain frequency and amplitude. The amplitude corresponds to the maximum value of the expression and at the time t=t pk Appear, among which:

[0094]

[0095]

[0096] For a particular pair of residual r and pole p, the voltage-time-integral Q u,t Corresponding to the envelope u env (t) and Figure 4 The dashed line shows the value u for steady state ss The (V, t)-integral between the two is calculated until the predetermined termination time criterion ε is reached. v The termination criterion may be, for example, the envelope u env (t) is a fraction, for example 1%, of the maximum amplitude at t = 0. This can be determined, for example, from ε v =1% of [u env (t=0)-u ss ] to achieve it.

[0097] iv) Now for all (oscillatory) (r, p) pairs, we have the envelope u env (t) and the value u for steady state ss Add and from this calculate the track-specific voltage-time-integral Q as described above u,p, the (r, p) pairing belongs to a specific approximate transfer function and thus to a specific safety-critical trajectory. This holds for all n t =(n s ·n q ) safety-critical trajectory execution.

[0098] v) In this embodiment, the voltage stability Q of the airborne energy network EBN as a whole u Including all n t =(n s ·n q ) Trajectory-specific instantaneous voltage-time-integral Q of a safety-critical trajectory u,p Then it can be defined as voltage-time-integral Q according to the following formula: u,p The sum of:

[0099] .

[0100] v) The energy loss E generated in the on-board energy network EBN by the current filter parameter set can be defined, for example, as follows L :

[0101]

[0102] Here, E L is the energy loss in the entire onboard energy network EBN through the filter or filter stage. As described here, it can be calculated as the squared current i through the filter stage f,s (t)(filter index f, filter stage index s) multiplied by the predetermined filter stage resistance R f,s The integral of the product of over the transition state duration ts. The filter stage FS flowing through the filter BPf s The current i f,s (t) can be calculated from the above equation, for example by noting that the voltage oscillation u TF (t) case.

[0103] Now the integral is applied to all n f All n filters f,s The energy loss E is obtained by adding the filter stages L .

[0104] In step S11 it is checked whether a certain termination criterion is met. If this is the case ("N"), the filter parameter R is changed in step S12 n , L k and C k The change can be made, for example, as follows: for the initial filter parameter R n , Lk and C k The following values ​​are used, which satisfy the stability boundary condition; for the second and third executions of steps S4 to S9, the following values ​​are used, which correspond to the equilibrium boundary condition or the loss boundary condition, and the subsequent changes are made in step S12 based on the particle swarm optimization method. However, if the situation is yes ("Y"), then in step S13 (E L ;Q u ) The set of points determines the Pareto front.

[0105] In step S14, a point on the Pareto front is selected. Therefore, the filter parameter R based on this point is also found. n , L k and C k and can then be used, for example in step S15, to construct an onboard energy network EBN accordingly.

[0106] Figure 5 A method that can be used in step S13 is shown. L ;Q u ) diagram, which is the energy loss E L (Unit: µWs) and voltage stability Q u (Unit: µVs) with multiple (E L ;Q u ) points. The points on the Pareto front PAR correspond to the optimal parameterization of the frequency filter. Here, point p1 corresponds to the parameter set for the stability boundary condition, point p2 corresponds to the parameter set for the equilibrium boundary condition, and point p3 corresponds to the parameter set for the loss boundary condition. The parameter sets for the other points are determined by particle swarm optimization. Point p4 corresponds to a point on the Pareto front PAR and can be selected, for example, as a point for a specific parameterization of the onboard energy network EBN.

[0107] Of course, the invention is not limited to the exemplary embodiments shown.

[0108] Generally speaking, “a”, “an” etc. can be understood as singular or plural, especially in the sense of “at least one” or “one or more” etc., as long as this is not explicitly excluded, for example by expressing “exactly one” etc.

[0109] Unless expressly excluded, the numerical values ​​given may also include the exact numerical value given as well as the general tolerance ranges.

[0110] Reference numerals list

[0111] A-Node

[0112] ASIL safety-related components

[0113] B-Node

[0114] BPi i-th multi-stage bandpass filter

[0115] C-Node

[0116] C j Capacitor of the jth filter stage

[0117] D1 suppression diode

[0118] EBN Airborne Energy Network

[0119] ECU control device

[0120] EFi electronic fuse for the ith distributor

[0121] ELF electric fan

[0122] E L Energy loss

[0123] EPB electric drive servo steering system

[0124] ε v Termination criteria

[0125] FTransportation

[0126] FS j The jth filter stage

[0127] GSWi the i-th DC voltage converter

[0128] GT Gate Driver

[0129] IB integrated braking system

[0130] L j Inductance of the jth filter stage

[0131] PAR Pareto frontier

[0132] PDi the i-th distributor

[0133] Points p1-p4

[0134] QM comfort components

[0135] Q fuse Electronic fuse switch

[0136] Q j Switching of the jth filter stage

[0137] Q u Voltage stability value

[0138] Q u,t Voltage-time-integral for residual-pole-pairs

[0139] Q u,p Trace-specific voltage-time-integral

[0140] R j The resistance of the jth filter stage

[0141] RAS rear axle steering system

[0142] S1-S14 Method Steps

[0143] u ss The voltage in steady state

[0144] t time

[0145] V HV High voltage network voltage

[0146] V Voltage

[0147] WIP Windshield Wiper Motor

[0148] Z resistance

[0149] Z l Line resistance

Claims

1. A method (S1-S14) for designing a plurality of frequency filters (BP1, BP2, BPi) which are arranged in a distributed manner in an onboard energy network (EBN) of a vehicle (F), the method comprising at least the following steps: (a) establishing a simulation model of an onboard power network (EBN) having a predetermined number of frequency filters (BP1, BP2, BPi) at defined locations of the EBN (S1); (b) determining an initial filter parameter set of the frequency filter (S2, S3); (c) calculating a simulation model in the time domain using the filter parameter set (S5); (d) extracting safety-critical trajectories together with the underlying disturbance functions from the calculated simulation model (S6); (e) transforming the safety-critical trajectory from the time domain to the frequency domain (S7); (f) determining the transfer function in the mathematical s-domain by means of system identification (S8); (g) extracting pairs of residuals and poles according to the transfer function in the s-domain (S9); (h) Determine the energy loss value (E) based on the extracted residuals and poles L ) and voltage stability value (Q u )(S10); (i) changing the filter parameter set (S12) of the frequency filter (BP1, BP2, BPi); (j) repeating steps (c) to (h) with the set of filter parameters changed in step (i) until a termination criterion is reached; (k) After reaching the termination criterion: from the energy loss value (E L ) and voltage stability value (Q u ) selects a point (p4) located on the Pareto front (PAR) of the set of points (S13, S14).

2. The method (S1-S14) according to claim 1, wherein: In step (j), the filter parameter set is modified by means of an optimization method.

3. The method (S1-S14) according to claim 2, wherein: In step (j), the filter parameter set is modified by means of a particle swarm optimization method.

4. The method (S1-S14) according to one of the preceding claims, wherein: To implement step (b), first - determining in sub-step (b1) a solution space which comprises feasible values ​​of filter parameters of the frequency filter based on the specific values ​​to be used of the filter resistance, the filter capacitance and the filter inductance (S2), and - In sub-step (b2), find the initial set of filter parameters from the solution space (S3).

5. The method (S1-S14) according to claim 4, wherein: The initial filter parameter set (S2, S3) is selected according to a stability boundary condition with a maximum filter stage capacitance, a balance boundary condition with a minimum filter stage capacitance, or a loss boundary condition with a target filter stage capacitance between the stability boundary condition and the balance boundary condition (S2, S3).

6. The method (S1-S14) according to claim 4, wherein: The initial filter parameter set is arbitrarily selected from the solution space (S2, S3).

7. An onboard energy network (EBN) of a vehicle (F) having a plurality of distributed frequency filters (BP1, BP2, BPi), wherein: The frequency filters (BP1, BP2, BPi) are designed according to a filter parameter set which is stored for the selected points located on the Pareto front in step (k) by means of a method (S1-S14) according to one of the preceding claims.

8. The on-board energy network (EBN) according to claim 7, wherein: At least one of the frequency filters (BP1, BP2, BPi) is integrated into an electronic power distributor (PD1, PD2).

9. An on-board energy network (EBN) according to any one of claims 7 to 8, wherein: At least one of the frequency filters (BP1, BP2, BPi) is a multi-stage filter.

10. On-board energy network (EBN) according to any one of claims 7 to 9, wherein: The onboard power network (EBN) is a low-voltage onboard power network, in particular a low-voltage onboard power subnetwork.

11. Vehicle (F) comprising an on-board energy network (EBN) according to one of claims 7 to 10.

12. The vehicle (F) according to claim 11, wherein: The vehicle (F) is a fully electric vehicle.

Images