Methods for monitoring an energy supply in a motor vehicle

Abstract

Method for monitoring an energy supply in a motor vehicle, wherein in a partial electrical system (30, 40) at least one energy storage device (32) supplies energy to several preferably safety-relevant consumers (36, 46), wherein at least one measured variable (Ubatt, Ibatt; Uv, Iv) of an energy storage device (32) and / or at least one consumer (36, 46) is recorded, wherein at least one wiring harness model (91) is provided which represents the partial electrical system (30, 40), and wherein a parameter estimator (104) is provided which estimates at least one characteristic variable (Rb12) of the wiring harness model (91) using the measured variable (Ubatt, Ibatt; Uv, Iv), characterized in that the parameter estimator (104) updates the characteristic variable (Rb12) using the measured variable (Ubatt, Ibatt) and / or the previous characteristic variable (Rb12).

Description

[0001] The invention relates to a method for monitoring an energy supply in a motor vehicle, in particular with automated driving operation, according to the preamble of the independent claim. State of the art

[0002] The vehicle's electrical system is responsible for supplying power to the electrical consumers. If the power supply fails due to a fault or aging in the electrical system or a component of it in modern vehicles, important functions such as power steering will be lost. Since the vehicle's steering ability is not affected, but only becomes difficult, the failure of the electrical system in modern production vehicles is generally accepted because the driver serves as a backup. To increase availability, dual-channel electrical system structures, such as those proposed in WO 2015 / 135729 A1, have been suggested. These are required to provide fault-tolerant power to systems for highly or fully automated driving.

[0003] A method for monitoring an electrical vehicle electrical system is already known from DE 102015221725 A1. This method comprises at least one channel with at least one component and a power supply, wherein a terminal voltage at one component is measured and compared with a voltage provided by the power supply. Furthermore, another measurement is carried out in a different state with a different current draw and compared with the first measurement. Taking into account the current flowing into the component in the different state, a contact resistance to this component is determined.

[0004] To determine the supply line resistance, the necessary measurements (for example, the battery voltage, another voltage in the supply line, and the battery current) only need to be recorded once. The resistance is then calculated in a second step using Ohm's law.

[0005] If the network is expanded, for example by adding further circuits for additional energy sources and consumers, the number of unknowns (line resistances of the individual circuits) and the number of boundary conditions (mesh equations) increase. In this case, the system is underdetermined, meaning that the number of unknowns (e.g., the line resistances) is one greater than the number of mesh equations. Therefore, the described direct approach for determining the individual line resistances cannot be applied.

[0006] From WO 2013 / 104 423 A1, an energy distribution network is known. The distribution network has at least one automation device subordinate to the distribution network control device, which is assigned to a subnetwork of the distribution network, wherein at least one measuring device connected to the subnetwork for determining measured values ​​relating to the subnetwork is connected to the automation device.

[0007] From DE 10 2016 221 249 A1, a method for operating an on-board electrical system is known in which, during the operation of the vehicle, model-based electrical simulations are carried out to simulate a future voltage profile of the on-board electrical system as a function of predicted loads in order to determine a state of the on-board electrical system, whereby this state is taken into account during the operation of the on-board electrical system and the vehicle.

[0008] The invention is based on the objective of further increasing the reliability of monitoring. This objective is achieved by the features of the independent claim. Disclosure of the invention

[0009] By employing a parameter estimation method to determine the relevant parameters of the vehicle electrical system, particularly the contact resistance, the reliability of the monitoring can be further increased. Using multiple measurements at different times (time profile), the parameters of even an underdetermined system can be determined because the temporal correlation (the parameters of the vehicle electrical system, such as the line resistances, remain approximately constant) leads to further equations for determining the parameters or resistances. According to the invention, the combination of multiple measurements at different times using parameter estimation methods is employed to determine specific parameters for the vehicle electrical system.Especially in complex electrical systems, utilizing a temporal correlation can provide sufficient information for the simultaneous estimation of several parameters of the electrical system (contact resistances), even in the absence of measuring points.

[0010] Furthermore, using such estimation methods offers the advantage that they not only provide the estimated value itself, but also an estimate of its accuracy (for example, the variance of the estimate), allowing the reliability of the determined parameters to be assessed. This reliability can be used as an additional criterion, particularly for enabling safety-relevant functions, such as in automated driving systems.

[0011] Furthermore, the parameters to be estimated, in particular the contact resistances or resistances, can be determined simultaneously and independently of each other. The only requirement is that a corresponding excitation is present in the respective string.

[0012] In particular, adding further, additional measurements – for example, through the use of central power distribution units – allows this additional information to be used to increase the statistical reliability of the results, for instance, by selecting the most accurate sensors or by fusing the measurement data. Furthermore, the additional redundancy can reduce the probability of system failure.

[0013] Parameter estimation is particularly useful when it can be ensured that the current is sufficient to excite the respective circuit. For this purpose, a pre-filter device, such as a logic circuit, could be incorporated to evaluate the measured current and compare it to a threshold value. If the measured current of a circuit exceeds the threshold, the parameter estimation of the corresponding characteristic can be initiated. This further improves the accuracy of the diagnostic result. If the threshold is not exceeded, the update of the estimation value is suspended for that time step. This further enhances the reliability of the system.

[0014] A specially optimized excitation can be particularly useful for improving the estimation accuracy. This is especially advantageous if it does not impair the operation of the vehicle electrical system. A redundant channel or branch of the electrical system is particularly preferred for this purpose.

[0015] In a suitable further development process, the measured quantities used to determine the characteristic value are synchronized, in particular by a filter, preferably a fractional delay filter, a Lagrange filter, a Farrow filter, or similar methods. This reduces the distortion of a diagnostic result due to possible asynchrony of the measured quantities, for example, due to non-synchronous clock sources.

[0016] A filter is particularly useful for determining a delay factor to synchronize the measured quantities for all quantities to be delayed. The delay factor of the filter is also particularly useful when determined by a parameter estimator. This allows for high signal quality of the delayed signals with minimal computational effort for estimating the delay factor.

[0017] Further appropriate training courses result from other dependent requirements and from the description. Brief description of the drawings

[0018] They show Fig. 1. A possible on-board electrical system for a vehicle for automated driving, Fig. 2 a supply and data structure for two safety-relevant channels for on-board network diagnostics, Fig. 3. An equivalent circuit diagram or model of an energy storage device, Fig. 4 the structural design of the parameter estimator as well as Fig. 5. The structural design of the filter structure for time synchronization. Embodiments of the invention

[0019] The invention is schematically illustrated with reference to embodiments in the drawings and is described in detail below with reference to the drawings.

[0020] The exemplary embodiment describes a battery or accumulator as a possible energy storage device. However, other energy storage devices suitable for this task, such as inductive or capacitive systems, fuel cells, capacitors, or similar devices, can also be used.

[0021] Fig. Figure 1 shows a possible topology of an energy supply system consisting of a basic electrical system 22, which supplies at least one basic consumer 24, shown as an example. Alternatively, the basic electrical system 22 could also include an energy storage device or a battery with an associated (battery) sensor and / or a starter, and / or several non-safety-related comfort consumers, which could be protected or controlled by an electrical load distribution system. The basic electrical system 22 has a lower voltage level than a high-voltage electrical system 10; for example, it could be a 14 V electrical system. A DC-DC converter 20 is arranged between the basic electrical system 22 and the high-voltage electrical system 10. The high-voltage electrical system 10 includes, for example, a high-voltage energy storage device 16, such as a high-voltage battery, possibly with an integrated battery management system, and, for example, a non-safety-related load 18.Comfort consumers such as an air conditioning system supplied with a higher voltage level, etc., as well as an electric motor 12. The energy storage device 16 can be connected to supply the high-voltage electrical system 10 via a switching device 14. In this context, high voltage is understood to be a voltage level that is higher than the voltage level of the basic electrical system 22. For example, it could be a 48-volt electrical system. Alternatively, especially in vehicles with electric drives, even higher voltage levels could be used. Alternatively, the high-voltage electrical system 10 could be omitted entirely, in which case components such as the starter, generator, and energy storage device would be assigned to the basic electrical system.

[0022] The basic electrical system 22 is connected, for example, to two safety-related sub-systems 30 and 40. The first safety-related sub-system 30 is connected to the basic electrical system 22 via a isolating element 28. The second safety-related sub-system 40 is connected to the basic electrical system 22 via another isolating element 26. The first safety-related sub-system 30 can be supplied with energy via an energy storage device 32. The characteristic parameters of the energy storage device 32 are detected by a sensor 34. The sensor 34 is preferably located adjacent to the energy storage device 32. The first safety-related sub-system 30 supplies a safety-related consumer 36. This safety-related consumer 36 is shown only as an example. Depending on requirements, further safety-related consumers 36 are supplied via the sub-system 30, as will be explained in more detail below.

[0023] The additional safety-relevant sub-system 40 can also be supplied by another energy storage device 42. The parameters of the additional energy storage device 42 are recorded by another sensor 44. This additional sensor 44 is located adjacent to the additional energy storage device 42. The additional safety-relevant sub-system 40 supplies at least one additional safety-relevant consumer 46. Depending on requirements, further safety-relevant consumers 46 can also be supplied in the additional safety-relevant sub-system 40. A fault condition 50 is indicated in the additional safety-relevant sub-system 40. This could lead to an increased cable harness resistance 48.

[0024] As explained in more detail below, a cable harness diagnostic is performed using a parameter estimator 104, which reliably detects such fault cases 50.

[0025] The redundant, and in particular functionally redundant, safety-relevant consumers 36, 46, which can be supplied via the two safety-relevant sub-networks 30, 40, are those necessary to bring a vehicle from automated driving mode (no driver intervention required) to a safe state, for example, in critical fault situations. As described in more detail below, this can involve stopping the vehicle, either immediately, at the roadside, or at the next rest area, etc.

[0026] The following safety-relevant consumers 36 and 46 are provided as examples. These include at least one braking system 60. The braking functionality can be achieved either by a first component or consumer 62 (for example, an electronic stability program that can bring the vehicle to a standstill), which is supplied via the sub-network 30. Alternatively, redundant braking functionality is achieved via another component or consumer 64 (for example, an iBooster that generates brake pressure electromechanically). This additional consumer 64 is supplied via the other safety-relevant channel 40. Redundant power supply is possible via the two safety-relevant sub-networks 30 and 40, so that even if one of the sub-networks 30 or 40 fails, reliable braking of the vehicle is still possible.

[0027] As a further safety-relevant component, a steering system 66 is provided, for example. The steering system 66 consists of a first component or consumer 68, for example in the form of an electric power steering system, which is supplied by the partial electrical system 30. Furthermore, the steering system 66 includes another component or consumer 70 (also, for example, an electric power steering system), which can steer the vehicle as desired, independently of the first component 68. The further component or consumer 70 is supplied with energy by the further safety-relevant partial electrical system 40.

[0028] An automated driving function 78 is provided as a further safety-relevant and redundantly implemented component. A corresponding first processing unit 80 and a first sensor unit 74 are supplied with energy by the first channel 30. The automated driving function 78 can be implemented redundantly by a further processing unit 82 and one or more sensor units 76, which are supplied by the further safety-relevant channel 40. In the processing units 80 and 82, for example, trajectory planning with associated control values ​​for the required actuators takes place. The sensor units 74 and 76 each provide the current environmental information of the vehicle for trajectory planning.

[0029] Furthermore, in the exemplary embodiment according to Fig. Figure 2 also shows the data connections 54 and 56. The data connections 54 and 56 transmit the acquired data, in particular current and / or voltage values ​​at the isolating element 26 and 28 and / or at energy storage devices 32 and 42, as determined, for example, by sensors 34 and 44, and / or at the safety-relevant consumers 62, 64, 68, 70, 74, 76, 80, and 82, to an on-board network diagnostic system 58. For example, a first data connection 54 transmits the data from those components supplied by the sub-on-board network 30 to the on-board network diagnostic system 58. A further data transmission 56 transmits, for example, the data from those components supplied by the further sub-on-board network 40 to the on-board network diagnostic system 58. The data connections 54 and 56 can be suitable bus systems, such as a LIN or CAN bus.

[0030] Specifically, the data connection 54 transmits relevant current and voltage values ​​from the isolating or coupling device 28 (Udcdc, Idcdc). The sensor 34, in particular a battery sensor, detects the voltage Ubatt at the energy storage device 32 and / or the current Ibatt of the energy storage device 32 and / or other states, parameters of a model of the energy storage device 32. Furthermore, the currents Iv at the individual consumers 62, 64, 68, 70, 74, 76, 80, 82 and the current voltages Uv (terminal voltages) of the individual consumers 62, 64, 68, 70, 74, 76, 80, 82 at the respective terminals of these consumers 62, 64, 68, 70, 74, 76, 80, 82 are available for transmission to the on-board network diagnostics 58.

[0031] In Fig. Figure 3 shows, as an example, the wiring harness model 91 based on a typical partial wiring harness 30,40. Here, a partial wiring harness 30,40 is shown according to the Fig. 1 or 2 and with the relevant on-board network components, consisting of a DC-DC converter (isolating element 26, 28), the energy storage device 32, 42 with sensor 34, 44, cable harness branches with the corresponding positive-side supply resistances including line resistances 51 (positive-side resistances, index 1 Rx1: R11, R21, R31) and ground-side supply resistances including contact resistances 55 (ground-side resistances, index 2 Rx2: R12, R22, R32) as well as with various safety-relevant consumers 36, 46, characterized by the associated load currents Iv1, Iv2, Iv3 and the terminal voltages Uv1, Uv2, Uv3 for three exemplary consumers 36, 46. The consumers 36, 46 are controlled by a current distributor 52. This structure forms the basis for the wiring harness model 91.The cable harness resistances (positive-side line resistances including contact resistances 51, ground-side supply line resistances including contact resistances 55) are estimated using a parameter estimation method, as explained in more detail below. The necessary currents Iv and voltages Uv for calculating the cable harness resistances (positive-side line resistances including contact resistances 51, ground-side supply line resistances including contact resistances 55) are provided by components containing the measuring devices via the described communication network (data connections 54, 56).

[0032] The resistors Rb1, Rb2, etc. represent the supply resistances 51 including contact resistances 55 to the respective component. According to the invention, the positive-side (index 1) and ground-side (index 2) resistances Rxy, 51, 55 are estimated as a sum, i.e., for example, for the contact resistance of the energy storage device 32 on the positive side Rb1 and on the ground side Rb2 as follows: RB1,B2=RB1+RB2 Faults occurring in one strand can therefore be detected on both lines, but not distinguished. The same applies to the supply line resistances of the separating element 26 (Rdc1+Rdc2=Rdc1,dc2; R11+R12=R11,12; R21+R22=R21,22; R31+R32=R31,32). The resistances described above are corresponding parameters of the cable harness model 92 and are subsequently summarized in a state vector Xk.

[0033] Existing voltage and current measurements are incorporated into the equivalent circuit diagram as U BATT , I BATT etc. shown.

[0034] The following system of equations results, which describes the vehicle electrical system model or wiring harness model 91 of the respective sub-vehicle electrical system 30, 40: UBatt=UDCDC−IDCDC⋅RDC1,DC2+IBatt⋅RB1,B2UDCDC=UBatt−IBatt⋅RR1,R2+IDCDC⋅RDC1,DC2 UV1=UBatt+IBatt⋅RB1,B2−IV1⋅R11,12 UV2=UBatt+IBatt⋅RB1,B2−IV2⋅R21,22 UV3−UBatt+IBatt⋅RB1,B2−IV3⋅R31,32 for each of the three consumers 36, 46, energy storage devices 34, 44 and separating elements 26, 28 supplied by the power distributor 52 shown as examples, according to Fig. 3.

[0035] The above system of equations describing the sub-network 30, 40 can be expressed as a parameter vector x in the form of the relevant parameters of the sub-network 30, 40, namely the sum of the corresponding positive-side supply resistances including contact resistances 51 and ground-side supply resistances including contact resistances 55 in the respective branches. k describe the parameter vector x kUsing the parameter estimator 104 and referring to corresponding measured values ​​in the vector z, it is possible to k summarized, update for each new time step.

[0036] The corresponding quantities are summarized below in the corresponding vectors or matrices as described: xk=[RB1,B2RDC1,DC2R11,12R21,22R31,32],f(x^k−1,uk)=x^k−1, zk=[UBattUDCDCUV1UV2UV3] h(x^k−,0)=[UDCDC−IDCDC⋅RDC1,DC2+IBatt⋅RB1,B2UBatt−IBatt⋅RB1,B2+IDCDC⋅RDC1,DC2 UBatt+IBatt⋅RB1,B2−IV1⋅R11,12 UBatt+IBatt⋅RB1,B2−IV2⋅R21,22 UBatt+IBatt⋅RB1,B2−IV3⋅R31,32]

[0037] For the recursive solution of the system of equations, a so-called Extended Kalman Filter (EKF) is used as parameter estimator 104 in the exemplary embodiment, as shown in Fig. 4 shown. Alternatively, other parameter estimators 104 such as (recursive) least-squares methods, or other state estimators such as standard / unscented Kalman filters, particle filters or similar estimation / optimization methods could be used.

[0038] The parameter estimator 104 according to the embodiment shown in Fig. 4 includes at least one prediction 106. The input variables for prediction 106 are an initial estimate x̂. k-1 , the state variable x k , as well as an initial estimate P k-1 the error covariance P k Furthermore, in the steady state, the currently determined initial values ​​from a correction 108, namely the current parameter vector x, are obtained. k as well as the current error covariance P k , as input variables to prediction 106.

[0039] Prediction 106 involves an update, specifically a time update. This involves predicting the state of the state variable x. k in the form of the equation x^k−=f(x^k−1,uk)

[0040] Furthermore, prediction 106 includes a prediction of the error covariance matrix P. k in the form of the equation Pk−=AkPk−1AkT+WkQk−1WkT

[0041] Where Ak: Jacobian matrix of f(x̂ k-1 , u k ) Wk: System matrix of system noise Q k-1 : Variance of system noise

[0042] The initial variables x^k− as well as Pk− The input variables arrive at block Correction 108. In Correction 108, the estimated values ​​are updated based on the measurement(s). First, the so-called Kalman gain K is calculated. k with the formula Kk=Pk−HkT(HkPk−HkT+VkRkVkT)−1

[0043] Where Hk: Jacobi matrix of h(x^k−,0) Vk: System matrix of measurement noise

[0044] Subsequently, in correction 108, the estimate update is carried out based on the measurement according to the formula: x^k=x^k−+Kk[zk−h(x^k−,0)]

[0045] Finally, in correction 108, the error covariance Pk is updated according to the formula: Pk=(I−KkHk)Pk−

[0046] Where I corresponds to the identity matrix.

[0047] This results in an estimate of the expected value and covariance of the parameters x̂. kIn the first step of the filtering process, the previously obtained estimate of the state dynamics is applied to obtain a prediction for the current time. These predictions are then corrected in block 108 using the new information from the current measurements, resulting in the desired current estimate. Measurement data fusion / selection

[0048] By adding further, additional measured variables, for example through the use of (additional) central power distribution units 52, this additional information can be used to increase the statistical reliability of the results, e.g., by selecting the most accurate measuring sensors or by fusing the measurement data. Furthermore, the additional redundancy can reduce the probability of system failure. The parameter estimation can be adjusted by including the additional measured variables in h(x⌢k−,0) and z k take place. Selective estimation method

[0049] A prerequisite for a good diagnostic result is that the excitation current has sufficient amplitude. Therefore, it is advantageous to perform parameter estimation 104 only when it can be ensured that the current Iv has sufficient magnitude to excite the respective circuit. This can be achieved, for example, by installing a pre-filter device such as a logic circuit that evaluates the measured current Iv and compares it with a threshold value (e.g., 15 A). If the measured current Iv of a circuit is, for example, 20 A and thus exceeds the threshold value, parameter estimation 104 of the corresponding conductor resistance 51 can be started. Otherwise, the estimation of the resistance 51 for the current measurement sample is suspended. For this purpose, the matrices of the Extended Kalman Filter for the corresponding measurement equation are not updated in this time step.

[0050] Active diagnostics with specially optimized suggestions to improve the accuracy of the estimation are also conceivable, provided they do not impair the operation of the vehicle electrical system. A redundant channel or sub-network 30, 40 is ideally suited for this purpose. Online estimation and compensation of systematic measurement errors

[0051] Measurement errors, i.e., errors in the actual measurement process, are fundamentally composed of systematic (epistemic) and random (aleatory) errors. The latter originate from random physical processes and cannot be influenced without changing the physical measurement principle. Systematic errors, however, are based on deterministic relationships. If it is possible to estimate these relationships online, i.e., during operation, this source of error can be eliminated.

[0052] The existing estimation model can be extended to include a so-called disturbance model. This disturbance model describes the influence of unknown conditions on the measurement result, in this case, the systematic measurement error. To compensate for the systematic measurement error, these unknown parameters must be estimated in addition to the line resistances 51. This is possible with a sufficient number of measurement points (observability).

[0053] A commonly used model in practice for systematic measurement errors is the linear relationship. m⌢=a⋅m+b.

[0054] Here, m represents the physical quantity to be measured; in the electrical system, for example, current Iv or voltage Uv. Accordingly, m̂ is the generally deviating measurement. The parameters a and b correspond to gain and bias (a constant superimposed quantity). A perfect sensor would have a = 1 and b = 0. If a and b can be estimated (also called calibration), the systematic measurement error can be compensated for.

[0055] The vehicle electrical system typically contains several sensors 34, 44 and corresponding systematic measurement errors. However, for estimating the supply and contact resistances 51, 55, only the cumulative influence of these errors on the resistance estimates 51, 55 is relevant. Therefore, the additional parameters to be estimated can be significantly reduced, which is usually what makes calibration possible in the first place. For example, estimating three resistances 51, 55 and corresponding measuring points yields the voltage U1 in the first phase as follows: U1=U2,m−I2,mR2+I1,mR1.

[0056] If we now assume the above model for systematic measurement errors, the following results: U1=α1U2+b1−(a2I2+b2)R2+(a3I1+b3)R1 =a1U2−a2I2R2+a3I1R1+(b1−b2R2+b3R1)︸b⌢1

[0057] This means that instead of estimating the biases b1, b2, and b3, only the cumulative bias b̂1 needs to be estimated to compensate for the bias error. The same approach is possible for the gain error, depending on the availability of measurement points. The same procedure applies here. Furthermore, a combination of the two calibration approaches is possible, provided that suitable excitations are available.

[0058] Furthermore, estimating the systematic measurement error can be used to diagnose measurement points, allowing them to be identified as faulty if large errors are present. This additional information can be essential for achieving adequate ASIL qualification, which also considers the diagnosability of the measurement points.

[0059] If an estimation method is used that also takes random measurement errors into account, the measurement equation can be extended to include random variables ε, e.g., to m̂ = (a · m + b) + ε. The parameter estimation can be adjusted by including the additional parameters in h(x⌢k−,0) and x̂ k take place.

[0060] The estimation method, e.g. a Kalman filter, then provides, if possible, an estimate of the lead and contact resistances 51, 55 and the possibly cumulative parameters of the measurement equations with minimal variance, i.e. with the highest possible estimation accuracy. Time synchronization:

[0061] The structural design of the filter structure for time synchronization is described in Fig. Figure 5 shows that an input quantity u(k) is fed to a parameter estimator 116 for a delay D, to a physical system 112, and to a filter 114, specifically a fractional delay filter. The measured quantities z(k) are fed to both the parameter estimator 104 and the parameter estimator 116 for the delay D. The delay D determined by the parameter estimator 116 is also fed to the filter 114. The filter 114 determines time-synchronous output quantities u(k + D * ts), which are fed to the parameter estimator 104. The parameter estimator 104 determines the estimated parameters Rb12 of the wiring harness model 91 from the supplied time-synchronous input quantities u(k + D * ts) and the measured quantities z(k). These parameters Rb12, estimated by the parameter estimator 104, are also fed as input quantities to the parameter estimator 116 for the delay D.

[0062] The sensor fusion and measurement error calibration methods listed above require synchronous measurement data. In this context, synchronous means that the start time of the measurement is the same for all measured variables. (Uv(k),Iv(k),UBATT(k),IBATT(k)etc).

[0063] However, this is not the case in decentralized systems such as the vehicle's electrical system. The measurements are distributed across several subcomponents, each using its own internal, independent clock. Therefore, the diagnostic result is strongly influenced by the asynchrony of the measured variables.

[0064] However, if the asynchrony can be determined and eliminated, this has a positive effect on the diagnostic result. The following synchronization concept assumes discrete-time measurements that are available at the same sampling rate but exhibit a time shift relative to each other, D · t. s is (t sis the sample time of the sampling, D represents the delay factor between the signals as a linear factor related to the sample time).

[0065] Therefore, the following measured values ​​are available at sampling time k for the above measured quantities (assuming that the time shift is identical for each measured quantity of a specific component): Uv(k),Iv(k),UBATT(k+D⋅ts),IBATT(k+D⋅ts)

[0066] To determine the delay factor D, all measured quantities u(k) to be delayed are filtered, resulting in a time shift of an initial factor D. This can be achieved using a so-called fractional delay filter. The implementation using a first-order Lagrange filter is described below. u(k+D⋅ts)=(u(k)⋅(1−D)+u(k−1)⋅D)

[0067] This yields the estimated output quantity h(k + D · t) according to the procedure described above. s) calculated.

[0068] The estimate of the new delay factor can then be calculated as follows: D(k)=D(k−1)+[z(k)−h(k+D⋅ts)]⋅[h(k−1+D⋅ts)−h(k+D⋅ts)]−1 k : sampling time z: measured output quantity h: initial quantity estimated from u via model equation

[0069] Implementing this measurement equation in a parameter estimator 116, the delay D can be estimated using the noisy measured quantities u(k) and z(k).

[0070] To estimate the parameters of the cable harness's supply and contact resistances, the already calculated value for u(k + D · t) can then be used. s) can be used. Alternatively, using the calculated factor D, an additional, higher-quality filter can be implemented (for example, a higher-order Lagrange filter), the outputs of which are then fed into the parameter estimation. This has the advantage that a high signal quality of the delayed signals can be achieved with minimal computational effort for estimating the factor D.

[0071] The parameter estimator 104 can, for example, be found in the Fig. The on-board network diagnostics block 58 shown in section 2 is located there, or may be part of the on-board network diagnostics 58. The corresponding measured values ​​are transmitted according to... Fig. 2 to this on-board network diagnostic 58.

[0072] The described method for determining relevant parameters of a partial electrical system 30, 40 in a motor vehicle is particularly suitable for systems where high predictive accuracy is required, such as in autonomous driving. In any case, appropriate countermeasures must be initiated quickly and reliably in the event of a potential fault. The proposed parameter estimator 104 enables rapid and reliable diagnosis, making it particularly suitable for this purpose. The determined parameters Rb12 and xk are thus provided to a higher-level electrical system diagnostic function to ensure a reliable power supply to selected systems (e.g., braking system, steering, environmental perception, control computer). If there are significant deviations in the parameters of the wiring harness model 91, a fault message is generated. This fault message is then provided to the higher-level electrical system diagnostic function.If this error information is present, countermeasures such as bringing the vehicle to a safe stop or similar actions may be initiated. Alternatively, if the vehicle is not operating autonomously, such operation could be prevented. Further information regarding maintenance or diagnostics may also be initiated. If necessary, the determined parameters can be corrected or determined based on certain systematic measurement errors, in particular an error covariance Pk, a gain a, or a bias b, and / or an error message can be generated depending on the systematic measurement error.

Claims

[1] Method for monitoring an energy supply in a motor vehicle, wherein in a partial electrical system (30, 40) at least one energy storage device (32) supplies energy to several preferably safety-relevant consumers (36, 46), wherein at least one measured variable (Ubatt, Ibatt; Uv, Iv) of an energy storage device (32) and / or at least one consumer (36, 46) is recorded, wherein at least one wiring harness model (91) is provided which represents the partial electrical system (30, 40), and wherein a parameter estimator (104) is provided which estimates at least one characteristic variable (Rb12) of the wiring harness model (91) using the measured variable (Ubatt, Ibatt; Uv, Iv). characterized by , that the parameter estimator (104) updates the characteristic (Rb12) using the measured quantity (Ubatt, Ibatt) and / or the previous characteristic (Rb12). [2] Method according to claim 1, characterized by, that the characteristic parameter (Rb12) of the cable harness model (91) is used to generate an error information, for example in case of deviation of the characteristic parameter (Rb12) from a limit value and / or in case of deviation of other characteristic parameters, which are determined as a function of the characteristic parameter of the cable harness model (41), from a limit value. [3] Method according to any one of the preceding claims, characterized by , that depending on the parameter (Rb12) or the fault information, the vehicle is transferred to a safe operating state or autonomous driving operation is prevented. [4] Method according to any one of the preceding claims, characterized by , that the parameter estimator (104) includes at least one prediction (106) and / or at least one correction (108). [5] Method according to any one of the preceding claims, characterized by, that in the sub-network (30,40) at least one load distribution (52) is provided, wherein the load distribution (52) is supplied with energy by the energy storage device (32), wherein the load distribution (52) supplies energy to several consumers (36,46), in particular those required for autonomous driving operation, wherein a measured quantity (Uv, Iv) is recorded at the load distribution (52) and supplied to the parameter estimator (104) to determine the characteristic value (Rb12)). [6] Method according to any one of the preceding claims, characterized by , that in the partial on-board network (30,40) at least one isolating element (26), in particular a DC voltage converter, is provided, wherein a measured quantity (Udcdc, Idcdc) is recorded at the isolating element (26) and supplied to the parameter estimator (104) to determine the characteristic value (Rb12). [7] Method according to any one of the preceding claims, characterized by, that the parameter estimator (104) determines a new characteristic value (Rb12) at least when a current for exciting the sub-network (30) has a sufficient magnitude. [8] Method according to one of the preceding claims, characterized in that a contact resistance, in particular between energy storage (32) and load distribution (52), is used as a characteristic parameter (Rb12) and / or that a state vector is used as a characteristic parameter which comprises several contact resistances and / or supply line resistances (Rb12, Rdc1,dc2; R11,12) in different branches of the partial on-board network (30). [9] Method according to any one of the preceding claims, characterized by, that the parameter estimator (104) determines a systematic measurement error, in particular an error covariance (Pk), a gain (a) or a bias (b), and generates error information depending on the systematic measurement error, and / or corrects the parameter (Rb12) depending on the systematic measurement error. [10] Method according to any one of the preceding claims, characterized by , that as a characteristic parameter (Rb12) of the cable harness model (91) a positive-side contact resistance (Rb1) and a ground-side contact resistance (Rb2) are estimated in sum (Rb12). [11] Method according to any one of the preceding claims, characterized by , that the measured quantities (Ubatt, Ibatt; Uv, Iv) used to determine the characteristic quantity (Rb12) are synchronized, in particular by a filter (114), preferably a fractional delay filter, a Lagrange filter, or a Farrow filter. [12] Method according to any one of the preceding claims, characterized by , that for time synchronization of the measured quantities (Ubatt, Ibatt; Uv, Iv) a filter (114) determines a delay factor (D) for all measured quantities to be delayed (Ubatt, Ibatt; Uv, Iv). [13] Method according to any one of the preceding claims, characterized by , that the delay factor (D) of the filter is determined by a parameter estimator (116). [14] Method according to any one of the preceding claims, characterized by , that a state estimator, in particular a Kalman filter, especially preferably an Extended Kalman Filter (EKF) or Unscented Kalman Filter, in particular recursive Least Squares methods or particle filters, is used as the parameter estimator (104, 116). [15] Method according to any one of the preceding claims, characterized by , that a predictive diagnosis and / or maintenance recommendation and / or an operational strategy optimization is carried out using a future parameter (Rb12).

Images