DE102017107284A1 - METHOD AND CONTROL DEVICE FOR MONITORING A PORTION NET OF A VEHICLE - Google Patents

Abstract

A method is presented for monitoring a vehicle electrical system (200), wherein the vehicle electrical system (200) has at least one distributor (207; 704) and a load (206) which are connected via a line (210; L21, H1) , with the following steps:
Reading a time series of data for a number of parameters, the parameters representing information about an operation of the vehicle by a driver and / or a state of the vehicle and / or a state of the driver and / or a driving environment;
Classifying the data as a normal value or an error value, wherein the normal values per parameter lie within a state space that is separated from the error values by means of a discrimination limit; and
Evaluating the data classified as an error value, wherein an error is detected when the data classified as an error value meets a criterion and / or evaluating the data classified as normal value, the data classified as normal value are statistically evaluated to determine a stochastic parameter, and wherein an error information is determined when the stochastic parameter exceeds a threshold value.

Description

Technical area

The present invention relates to a method for monitoring a vehicle electrical system of a vehicle and to a control device for monitoring a vehicle electrical system of a vehicle.

State of the art

The electrical system of a vehicle connects an energy source via distributors and subdistributors with the various loads or consumers. So you can understand among the components of the electrical system distributors, sub-distributors and loads, but also these connecting lines. The cables are connected via plug connections with interfaces of the distributors, sub-distributors or loads. The connectors are examples of inputs and outputs of the components mentioned. In addition to a total failure of a line, it comes during the life of a vehicle to degradation processes of the cables and the connectors, which lead to a gradual deterioration of the affected components of the electrical system, to failure of the components.

With the electrification of functions, the number of consumers in the on-board network continues to increase. The protection of the on-board network as an energy and information technology networked system is becoming increasingly complex, as possible interactions of the components are becoming increasingly confusing. At the same time, a diagnostic cover has to be implemented in the electrical system in order to meet the requirements of functional safety.

In today's vehicles safety-related functions already exist, such as the automated steering of the parking function. These functions must be designed fail-safe, which means that when an error is detected, the safe state is assumed "off" and the driver is informed via display instruments. The consideration of the functional safety of these functions therefore often ends at the plug of the control unit, as an interruption or a failure of the supply leads directly to the safe state "off". A security concept at the supply level can therefore be dispensed with.

By contrast, autonomous driving systems must be designed to be fail-operational with an error state transition to the safe state "on". The failure of the supply of energy (supply line) or communication (bus line) for this function can lead to a direct hazard to people. This safe condition "on" must be maintained until the vehicle can be stopped in a safe place or until the driver can take control.

DE 10 2015 209 588 B3 discloses a simple and reliable apparatus for detecting faults in a vehicle electrical system. The device comprises a receiving device that is configured to receive a first and a second temporal sequence of measured values, wherein the measured values of the first sequence at a first measuring point of the vehicle electrical system and the measured values of the second sequence at a second measuring point of the vehicle electrical system differ from the first And an evaluation device that is configured to detect the occurrence of a disturbance based on a comparison of a first scatter of the measured values of the first sequence and a second scatter of the measured values of the second sequence.

DE 10 2014 018 640 B3 discloses a method for determining an electrical resistance in a motor vehicle. In a first measurement phase in which a first device receives a load current from a voltage source and all other devices receive a quiescent current from the voltage source, the first device determines a first voltage value and a load current value and a second device determines a second voltage value. In a second measurement phase in which each of the devices receives a respective quiescent current from the voltage source, the second device determines a third voltage value and the first device determines a fourth voltage value. An analysis device determines a connection resistance of a connection line of the first device on the basis of the first voltage value, the second voltage value, the third voltage value, the fourth voltage value and the load current value.

DE 10 2014 004 791 B3 describes a method for checking a connection between a battery supplying a low-voltage network of a motor vehicle and the low-voltage network, wherein a DC-DC converter connects a high-voltage network to the low-voltage network, wherein the DC voltage converter modulates the output voltage into the low-voltage network with a modulation frequency, the current impressed thereby and the voltage impressed thereby is measured by the battery and a resistance value is determined and evaluated with respect to at least one connection criterion, according to which a connection failure is determined in at least one unfulfilled connection criterion.

Description of the invention

The invention is based on the object of detecting errors that are imminent (degradation processes) simply and reliably at an early stage by means of a diagnosis.

The object is solved by the subject matters of the independent claims. Advantageous developments of the invention are specified in the dependent claims, the description and the accompanying figures.

The method according to the invention for monitoring an on-board network of a vehicle has a step of reading in, a step of classifying and a step of evaluating. The electrical system comprises at least one distributor and one load, which are connected via a line. The distributor is connected to a power source or another distributor. In the step of reading a time sequence of data is read. The data includes at least one parameter or a plurality of parameters. Each parameter represents information about an operation of the vehicle by a driver and additionally or alternatively a state of the vehicle and additionally or alternatively a state of the driver and additionally or alternatively a driving environment of the vehicle. Thus, a parameter may represent measured data or parameters derived therefrom, but also information from another source such as a navigation device or position data such as GPS data, environmental data such as temperature, humidity, road condition, precipitation, etc. In the step of classifying the read data as Normal value or error value classified. For this purpose, a state space is used which defines a discrimination boundary between a valid range for normal values and the remaining value ranges. The state space has a dimension corresponding to the number of parameters. If a number n of parameters Pn are monitored and thus read in, the state space has the dimension n corresponding to the number n of the linked parameters. Optionally, a plurality of parameters are combined into a value derived therefrom, and this derived value is represented as a parameter in the state space. However, the derived value can also be determined before reading in and the derived value can be read in as a parameter. The normal values lie within the range of values defined by the discrimination limit or within the state space. The limit of discrimination is also multidimensional according to the dimension of the state space. In the step of the evaluation, the data classified as error value and additionally or alternatively the data classified as normal value are evaluated. In this case, an error information is determined if the data classified as an error value satisfy a criterion. A criterion can be understood as a rule that is mathematically describable. For example, the criterion can be defined algebraically or statistically. Alternatively or additionally, the data classified as normal value are evaluated stochastically in order to determine a stochastic parameter. A stochastic parameter may be understood to mean an average (algebraic, geometric,...), A standard deviation, a variance, a modal value, a dispersion, a correlation coefficient, or the like. In this case, an error information is recognized as an error if the stochastic parameter exceeds a threshold value. The threshold value may be, for example, a fixed quantity or a percentage deviation. Thus, a slow drift of classified as normal value data can be detected and a degradation can be detected.

Thus, the read data can be considered as a point cloud or a cluster, the point cloud usually does not change their position substantially. A wandering or drifting of the point cloud can be diagnosed as an error or an emerging error, even if no value of a parameter has yet been classified as an error value. As a result of this procedure, emerging errors, which are caused, for example, by a degradation of a contact, can be detected early and eliminated as part of a predictive maintenance in order to increase the safety of the vehicle as a whole.

Advantageously, the method is suitable for monitoring a connection quality between components of the electrical system. In the step of reading, a resistance value of the line during operation of the vehicle can be read in a time-spaced manner and, additionally or alternatively, a number of parameters can be read in, from which the resistance value is determined. At least once a drive motor of the vehicle is turned off at least once in the course of reading. Thus, the method can diagnose a creeping fault during operation of the vehicle. Advantageously, the diagnostic options thus go beyond measurements and analyzes on a test bench. Thus, data can be collected, read in and stored over long periods of days, weeks, months or even years to ensure an evaluation of errors or pending or potential errors over this (long) period of time.

Components of the electrical system can be at least a distributor and a load, which are connected via a line. In the step of reading in a sub-step of determining a resistance value of the line are determined several times spaced. In this case, the resistance value is measured when the vehicle electrical system is active, that is during operation of the vehicle. In this case, a drive motor of the vehicle is switched off at least once between the determinations. In the step of classifying, in a substep of the determination, an error information can be obtained by using the resistance value and a predetermined comparison information.

As described in the introduction, the functional safety of the loads must be ensured for parts of the vehicle electrical system for certain applications such as autonomous driving. Even if corresponding parts should be redundant, the functionality must be ensured. For this, the connection quality between components is monitored. Since degradation processes are observed, the determination of the resistance value takes place over a longer period of time. Such a degradation process may initially be gradual over a period of months. Thus, the observation period over several trips or driving cycles is selected. In this case, a driving cycle is limited by at least one time switching off the drive motor and a standstill of the vehicle. Thus, a driving cycle can be understood as meaning a time interval in which the vehicle is permanently in operation, for example the drive motor is switched on. In this case, short interruptions due to a start-stop system can be ignored in the consideration, that is, the driving cycle then takes over this time. Thus, a driving cycle may be limited by a service interruption of the vehicle. Thus, the resistance value of the line is determined in at least two different driving cycles. In this case, the resistance value represents a value corresponding to the resistance of the line.

There is an expected value for the resistance of the line with a tolerance range or tolerance ranges surrounding this expectation value. Thus, a tolerance range for different degrees of error or degrees of degradation of the connection can be defined. The tolerance range defined in this way or the tolerance ranges defined in this way can then serve as the basis for the comparison information and thus make it possible to determine the error information in conjunction with the determined resistance value. For example, in one embodiment, doubling the resistance of the line may be defined as an error case.

Furthermore, in the step of reading in, a voltage value representing a voltage drop across the line can be read in. The resistance value can be determined in a sub-step of the determination using the voltage value. In an advantageous embodiment, the influence of a variable load can be eliminated or at least kept so low via the stochastic distribution, the choice of the time interval between two read-in voltage values or the time of reading in the voltage values that it has no appreciable influence on the resistance value to be determined , If the influence of the load is not considered, the current can be considered constant. In this case, the voltage drop correlates with the resistance value. In the electrical system, it can be assumed that the load is moving in a certain load range, which means that a certain bandwidth for the current can be assumed as the expected range. Consequently, the voltage drop across the line is expected in a corresponding voltage range.

It is also favorable if in the step of reading in a current value is read in, which represents a current between the output of the distributor and the load. Advantageously, the resistance value can then be determined using the current value. As already stated, an expected bandwidth for the current value results from the load and a load profile. In the case of a deviation, that is to say a current value outside a comparison range, a change in the resistance value can be concluded. In a favorable variant of the presented method, the resistance value can be determined directly by means of Ohm's law.

After the read-in step, in one embodiment, an average value can be formed over voltage values read in a predefined mean-time interval, and the mean value can be used as a voltage value in the following steps. This has the advantage that a possible influence of a variable or changing load on the voltage drop across the line is still negligible. The resistance values can be detected at a high sampling rate. A high sampling rate can be understood here as a sampling rate between 100 and 1000 Hz. In a favorable embodiment, a resistance value of the line with a time interval of a few milliseconds, in particular with averaging of a few seconds, that is at least three seconds, better more than seven seconds, can be determined. By means of the averaging, a data aggregation takes place. An advantage of a sampling rate in the single-digit millisecond range is that peaks are detected in the voltage curve and then have only a small influence on the result by averaging. A similar positive effect has long measurement intervals, whereby the influence of corresponding outliers on the result is also reduced. About the choice of the duration of the Mean time interval, the possible error in the determination of the resistance value can be influenced. In this case, a longer mean-time interval is to be preferred to a shorter mean-time interval. Thus, the values to be kept can be reduced and the influence of load peaks can be reduced. In a preferred embodiment, the mean time interval corresponds to the duration of a drive cycle.

It is also advantageous if, following the step of forming the mean value over an average time interval, a step of determining stochastic quantities follows. In the step of determining stochastic variables, a scattering of the average values over a plurality of driving cycles can be determined. For example, the scattering can be determined over 10 to 50 driving cycles, in particular over 15 to 30 driving cycles, particularly preferably over 20 to 25 driving cycles. In this case, the dispersion can be determined by sliding over the selected number of driving cycles. In a particular embodiment, the spread over all stored driving cycles, i. over which all stored values of all driving cycles are determined. However, this is expensive, so that advantageously a limitation of the number of driving cycles will be done as described.

In addition to the scattering, a standard deviation or a correlation coefficient can also be determined in the step of determining stochastic variables. Advantageously, it is thus possible to conclude a stochastic change in the resistance value or the contact resistance represented by the contact line monitored by it. Thus, the error information can be made using the determined stochastic quantities, for example by defining a limit value for the scattering or the correlation coefficient as comparison information.

For determining the voltage drop across the line, an output voltage value representing an output voltage at the output of the distributor and an input voltage value representing an input voltage to the input of a sub-distributor or the load connected via the line to the output can be read. Many components of the vehicle electrical system already determine input or output voltages. Thus, without additional measurement effort advantageously the voltage drop across the line can be determined. Thus, additional functional safety monitoring can be implemented cost-effectively.

Advantageously, the resistance values, the voltage values, the current values, the output voltage values, the input voltage values, at least one stochastic parameter and / or the mean values are stored in a step of storing. Thus, the evaluation can not first detect an error, if this has occurred. Rather, it is possible through appropriate data analysis early to detect trends. Advantageously, an error can thus already be countered by predictive maintenance without having to exchange components cost-effectively without reason.

For data reduction, the values to be stored can only be stored in the saving step if they represent a new minimum or maximum. Thus, the memory to be stored in a vehicle can be made smaller. Also reduces the volume of data to be transferred when the values to be stored are sent to a cloud in the sense of vehicle-to-infrastructure (Car-to-X or C2X) and stored and evaluated there.

Furthermore, at least one further vehicle information may be stored in the step of storing. In this case, further vehicle information may be understood as meaning, for example, engine information, equipment information about the vehicle, geographical position of the vehicle during operation, climatic conditions such as temperature or air humidity, a road condition or a time of storing the stored values. Then, in the step of the evaluation, the error information can be determined using the at least one further vehicle information.

In a preferred embodiment, the method may additionally comprise a step of transmitting the values stored in the step of storing to a central infrastructure device. An infrastructure facility can also be understood as a cloud. The values to be stored can at least partially be transmitted, transmitted or received wirelessly. Thus, the corresponding data can be collected and evaluated by a plurality of vehicles in the central infrastructure facility. Due to the resulting large amount of data errors or better said trends for future errors can be detected even earlier and thus, for example, in regular maintenance intervals of individual vehicles remedy.

Advantageously, a learning system can be created in an evaluation in the cloud. Thus, a drift of the read data in the form of the point cloud or the cluster over a large number of vehicles can be observed, for example, to a missing parameter, such as environmental influences such as Outside temperature, close. Also, the data in the cloud can be further enriched via a connection to other data or information sources for evaluation.

Advantageously, at least individual steps of the presented method are carried out repeatedly. Thus, the steps of reading, forming, classifying, determining, evaluating or determining, providing, storing and / or transmitting may be repeatedly performed. Thus, in the step of evaluating, the error information may be determined using the stored values. As already stated above, a larger amount of data can be used to detect a faulty trend earlier.

• - Empfangen einer Vielzahl von Werten von einer Vielzahl von Fahrzeugen in einer Infrastruktureinrichtung;
• - Ablegen der Vielzahl von Werten in einer zentralen Speichereinrichtung der Infrastruktureinrichtung;
• - Bestimmen zumindest einer Fehlerinformation für zumindest ein Bordnetz eines der Fahrzeuge unter Verwendung der Vielzahl von Werten der Vielzahl von Fahrzeugen; und
• - Bereitstellen der Fehlerinformation von der Infrastruktureinrichtung an zumindest eines der Fahrzeuge und/oder eine Überwachungsfunktion für die Vielzahl von Fahrzeugen.

When using an infrastructure device, at least part of the method may be executed there or using the values stored in the infrastructure device. In this embodiment, the method may comprise further steps:

• Receiving a plurality of values from a plurality of vehicles in an infrastructure;
• - storing the plurality of values in a central storage device of the infrastructure device;
• Determining at least one fault information for at least one on-board network of one of the vehicles using the plurality of values of the plurality of vehicles; and
• Providing the error information from the infrastructure device to at least one of the vehicles and / or a monitoring function for the plurality of vehicles.

The method may include a step of establishing the comparison information using the temporal sequence of data for a number of parameters, the stochastic parameters, the voltage values, the output voltage values and the input voltage values, the averages, and / or the current values. Thus, the comparative information can be adapted to corresponding influences.

Advantageously, a center of gravity of the data is determined in the step of evaluating as a stochastic parameter. An error information representing a degradation or a predicted error is detected when the center of gravity moves over time. In this case, a movement of the center of gravity is determined and compared with a threshold value, or a derivative of the movement of the center of gravity is determined and compared with a threshold value. If the threshold value is exceeded, error information is provided.

Furthermore, in the step of evaluating as a criterion, a number of data classified as an error value can be monitored in a predetermined time interval. As a criterion for the output of an error information, reaching a predefined number of data classified as an error value can be determined. It can be defined here that the predefined number of data classified as error value is achieved in a sliding time interval. Alternatively, a distance to the discrimination boundary may be defined as error information. Thus, the predefined number can vary depending on the distance to the discrimination limit.

In an optional step of learning, a number of state spaces can be learned, wherein in a first sub-step of reading a time series of learning data for a number of parameters is read, which were previously categorized representing normal value, wherein in a second sub-step at least one envelope is determined as a discrimination limit for the learned state spaces. The time series of learning data read in the step of reading in the step of learning can be results of a simulation and additionally or alternatively measurement data from different driving tests such as summer test or winter test. The learning data includes at least one parameter or a plurality of parameters. These are the same parameters that are read in during the read-in step in productive use.

Optionally, in a step of deriving between the step of reading and the step of classifying, the at least one temporal sequence of data is derived. Then in the step of classifying the derived data are classified and evaluated in the step of evaluating the classified, evaluated data.

The above object is also achieved by means of a control device for monitoring a vehicle electrical system of at least one vehicle. In this case, components of the electrical system comprise at least one distributor and a load, which are connected via a line. The control unit comprises at least one device for reading in a chronological sequence of data for a number of parameters, wherein the parameters include information about an operation of the vehicle by a driver and / or a state of the vehicle and / or a state of the driver and / or Representing a driving environment, means for classifying the data as a normal value or an error value, the normal values per parameter lying within a state space obtained by means of a Discrimination limit is separated from the error values, and means for evaluating the data classified as an error value, wherein an error is detected when the data classified as an error value meet a criterion and / or evaluating the data classified as normal value, wherein the data classified as normal value statistically are evaluated to determine a statistical parameter and an error is detected if the statistical parameter exceeds a threshold.

The control unit mentioned here can be embodied, in particular, as a processor unit and / or an at least partially hardwired or logic circuit arrangement for the described method. Said control device can be or include any type of processor or computer or computer with correspondingly necessary peripherals (memory, input / output interfaces, input / output devices, etc.).

The above explanations regarding the method apply to the control unit accordingly and vice versa. The controller may be implemented in one component or distributed in multiple components.

The solution presented here further comprises a computer program product which can be loaded directly into a memory of a digital computer, comprising program code parts which are suitable for performing steps of the method described here.

The above-described characteristics, features, and advantages of this invention, as well as the manner in which they will be achieved, will become clearer and more clearly understood in connection with the following schematic description of exemplary embodiments which will be described in detail in conjunction with the drawings. In this case, the same or equivalent elements may be provided with the same reference numerals for clarity.

list of figures

• 1 ein Ablaufdiagramm des erfindungsgemäßen Verfahrens;
• 2 eine schematische Darstellung eines Teils eines Bordnetzes eines Fahrzeugs;
• 3 eine schematische Darstellung eines Steuergeräts zum Überwachen (einer Verbindungsqualität zwischen Komponenten) eines Bordnetzes zumindest eines Fahrzeugs während des Betriebs des Fahrzeugs;
• 4 in einem Diagramm Widerstandsänderungen von mehreren Verbindungen über die Zeit;
• 5 in einem Diagramm Spannungen an einem Verbraucher über die Zeit;
• 6 eine symbolische Darstellung des Spannungsabfalls über die Zeit;
• 7 eine schematische Darstellung eines Bordnetzes eines Fahrzeugs;
• 8 eine Mehrzahl von Fahrzeugen mit einem Steuergerät zum Überwachen der Verbindungsqualität zwischen Komponenten des Bordnetzes sowie eine Infrastruktureinrichtung;
• 9 eine schematische Darstellung eines Bordnetzes eines Fahrzeugs;
• 10 - 17 eine grafische Darstellung in kartesischen Koordinatensystemen von Zustandsräumen; und
• 18 eine schematische Darstellung eines Steuergeräts zum Überwachen eines Bordnetzes zumindest eines Fahrzeugs während des Betriebs des Fahrzeugs.

Hereinafter, an advantageous embodiment of the invention will be explained with reference to the accompanying figures. Show it:

• 1 a flow diagram of the method according to the invention;
• 2 a schematic representation of part of a vehicle electrical system of a vehicle;
• 3 a schematic representation of a control device for monitoring (a connection quality between components) of a vehicle electrical system of at least one vehicle during operation of the vehicle;
• 4 in a diagram resistance changes of several connections over time;
• 5 in a diagram voltages on a consumer over time;
• 6 a symbolic representation of the voltage drop over time;
• 7 a schematic representation of a vehicle electrical system of a vehicle;
• 8th a plurality of vehicles having a control unit for monitoring the connection quality between components of the electrical system and an infrastructure device;
• 9 a schematic representation of a vehicle electrical system of a vehicle;
• 10 - 17 a graphical representation in Cartesian coordinate systems of state spaces; and
• 18 a schematic representation of a control device for monitoring a vehicle electrical system of at least one vehicle during operation of the vehicle.

The figures are merely schematic representations and serve only to illustrate the invention. Identical or equivalent elements are consistently provided with the same reference numerals.

Detailed description

Modern vehicle functions have to be considered from the point of view of Functional Safety (FuSi). Systems of autonomous driving belong to the so-called Fusi-relevant functions.

The diagnosis is an integral part of a safety concept that complies with ISO26262. The redundant design of supply lines without diagnostics does not constitute a sufficient safety concept since the line reserved as redundant can not be available in the event of a fault due to a latent error for the acceptance of the supply. The diagnosis as an integral part of the safety concept first leads to extra effort; on the other hand, u.U. a single diagnosed supply path without redundancy would be sufficient for a coherent security concept. This would be the case if it succeeds in timely detection of impending errors (failure supply by degradation processes). A minimization of redundant lines in the wiring harness is absolutely desirable because the cable set already reaches its limits in terms of weight and space.

Degradation processes (eg degradation of a contact) are based on error growth processes. Based on the description of these processes, error models are developed.

Error growth processes are self-reinforcing and can basically be abstracted by exponential functions.

• mit S0: Anfangszustand
• c(t): Wachstumskonstante. Diese ist nicht konstant, da sich die Treiber der Störgröße je nach Zustand ändern.

For the growth of an error S, the following equation can be given: $$\text{S = S0 * exp}\left[\text{c}\left(\text{t}\right)\text{t}\right]$$

• with S0: initial state
• c (t): growth constant. This is not constant because the drivers of the disturbance variable change depending on the state.

This can be illustrated by an example of self-reinforcing growth of fretting corrosion in a contact. The fretting corrosion itself has a small growth constant C1, so the disturbance builds up over a long time (e.g., several months). The disturbance results in the change of the characteristic size of contact resistance. The increased contact resistance in turn results in an increase in the power loss in contact. If this additional power loss can no longer be adequately dissipated, the damage process starts with thermo-mechanical destruction. This has a much larger growth constant C2. The released heat leads to an increase in the contact resistance and this to a self-reinforcing further heating (Thermal Run-Away). This process means that when the point is reached, the contact suddenly fails and the supply of the safety-related function is interrupted.

The method described below monitors in a specific embodiment a change of the characteristic parameter contact resistance before the critical failure time is reached.

1 shows a flowchart of a method 100 for monitoring a vehicle electrical system of a vehicle according to an embodiment of the present invention. The procedure 100 includes in its most general way at least three steps: reading S1 , Classify S2 , Evaluate S3 ,

In step S1 reading in becomes a time sequence of data for a number n of parameters pn read. Here, n is a natural number greater than or equal to 1. The parameters represent information about operation of the vehicle by a driver, a state of the vehicle, a state of the driver, or a state of a driving environment. Thus, measured data combined with information from databases or navigation systems or weather information can be read in and processed in the following steps. In the following step S2 of classifying, the read in are discriminated against normal and error values. In this case, individual parameters are checked for a position in an assigned state space or the number of parameters is checked for an n-dimensional state space in order to differentiate the data into normal values and error values and to classify them accordingly. The normal values are separated from the error values by a discrimination limit. In step S3 the classified data are evaluated. In a first variant, the data classified as an error value are evaluated. Thus, an error is indicated when a predetermined number of data classified as an error value have been classified in a predetermined (sliding) time interval. In a further variant, all data or the data classified as normal values to form a cluster, ie a kind of point cloud, is summarized and considered as such. The data are stochastically evaluated in this variant to determine a stochastic parameter, such as center of gravity (geometric or algebraic center), standard deviation, variance, median,. Error information is provided when the stochastic parameter exceeds a threshold.

In a specific embodiment, the method monitors 100 a connection quality of a line between components of the electrical system of at least one vehicle during operation of the vehicle. In step S1 the reading is determined in a sub-step a plurality of times spaced resistance value of the line, and in step S2 of classifying the error information using the resistance value and a predetermined comparison information determined. Between the provisions, a drive motor of the vehicle was turned off at least once. This is how the reading takes place S1 and classify S2 Repeated and over a plurality of driving cycles, over a period of days, weeks or years. In step S3 All classified data are evaluated.

In other words, in one embodiment, the resistance value is determined at least once per drive cycle. Preferably, the resistance value is determined using averaging per drive cycle.

2 shows a schematic representation of part of a vehicle electrical system 200 of a vehicle. The electrical system 200 includes two sources of energy 202 , a distributor 204 as well as a load 206 , At the load 206 it is a function relevant to functional safety. An entrance 208 the load 206 is over a line 210 with an exit 212 of the distributor 204 connected. Furthermore, the load 206 via a communication connection 214 with the distributor 204 , or with a controller designed as a microcontroller .mu.C 216 connected.

In other words shows 2 a fusi-relevant function 206 over a line 210 from a redundant source arrangement 202 (eg two batteries 202 ) is supplied. The characteristic quantity with which damage to the supply channel is to be closed is the total resistance R _ik from the electronic distributor 204 to the board with the load 206 (Fusi function).

1. 1. Spannungsmessung U_e(t) im elektronischen Verteiler 204
2. 2. Strommessung I(t) im elektronischen Verteiler 206, z.B. als Bestandteil der elektronischen Absicherung
3. 3. Spannungsmessung U_a(t) in der Last 206

In a particular embodiment, the online diagnosis uses the following measured variables, which are measured, for example, every 30 ms:

1. 1. Voltage measurement U _{e (t)} in the electronic distributor 204
2. 2. Current measurement I (t) in the electronic distributor 206 , eg as part of electronic security
3. 3. Voltage measurement U _{a (t)} in the load 206

The instantaneous contact resistance is calculated to $${\text{R}}_{\text{IK}}\left(\text{t}\right)=\left[\text{Ue}\left(\text{t}\right)-\text{Ua}\left(\text{t}\right)\right]/\text{I}\left(\text{t}\right),$$

The calculation of R _IK may be in the electronic power distributor 204 respectively. For this, the voltage values Ua (t) in the Fusi-relevant load must be 206 over a serial bus 214 to the distributor 204 be communicated.

If the measurement in the load 206 asynchronous to the measurement in the distributor 204 Thus, Ue (t1) and Ua (t2) are measured with t1 <> t2. Especially for dynamic loads 206 Thus, a considerable error can occur if, for example, a tip is present at t1 due to an inductive voltage and t2 is no longer present.

In a particular embodiment, the asynchronous measurement signals Ue (t) and Ua (t)] are linked via a stochastic link.

One approach is eg $${\text{R}}_{\text{IK}}=\left\{\text{MW}{\left[\text{Ue}\left(\text{t}\right)\right]}_{\mathrm{\tau \; =}3\text{s}}-\text{MW}{\left[\text{Ua}\left(\text{t}\right)\right]}_{\mathrm{\tau \; =}3\text{s}}\right\}/\text{MW}{\left[\text{I}\left(\text{t}\right)\right]}_{\mathrm{\tau \; =}3\text{s}}$$

For the individual asynchronous measured variables, an averaging is performed in a time window (here 3s) and a determination of the contact resistance based on the average values. If only the average voltage MW [Ua (t)] _{τ = 3s has} to be transmitted, this can also considerably reduce the bus load.

Advantageously, a self-diagnosis of the electrical system for a conclusive FUSI safety concept is thus realized.

3 shows a schematic representation of a control device 216 for monitoring a connection quality between components of a vehicle electrical system of at least one vehicle during operation of the vehicle. The electrical system can be a variant of a part in 2 shown electrical system 200 act.

The control unit 216 has an interface 330 for reading in signals. Depending on the exemplary embodiment, the signals are analog signals which are then digitized or already digital signals which are transmitted, for example, via a bus. Furthermore, the control unit 216 An institution 332 for determining the resistance value 334 the line of the electrical system to be monitored and a device 336 for determining an error information 338 on. The resistance value 334 corresponds to the resistance Rik in 2 , At this time, the error information using the resistance value 334 determined. In this case, the resistance value is understood to be a signal or information that corresponds to the resistance R _ik the supervised line corresponds. The device 336 for determining the error information 338 is trained, the error information 338 using the resistance value 334 and a predetermined comparison information 339 to investigate.

In a preferred embodiment, the controller reads 216 continue via the interface 330 a voltage drop across the line representing voltage value .DELTA.U or an input current to the line Ue (t) and an output current Ua (t) from the line to the resistance value 334 to determine. Referring to the illustration in and description to 2 In this case, the input current into the line Ue (t) corresponds to an output current from the distributor 204 and the output current Ua (t) from the line to an input current of the load 206 , When the input current Ue (t) and the output current Ua (t) are read in, is optional in a device 338 the voltage drop .DELTA.U certainly.

About the interface 330 a current value l (t) is read in and the resistance value 334 as already stated above, using the current value I (t).

mit t = „Zeitspanne Fahrzyklus“ sehr gut funktioniert. So werden in einem Ausführungsbeispiel die Mittelwerte über einen kompletten Fahrzyklus gebildet und der Widerstandswert 334 mit diesen Mittelwerten bestimmt.In one embodiment, the controller includes 216 An institution 340 to form averages of the read voltage values. The mean values are formed either over a predefined mean-time interval of a few seconds or minutes, or in a preferred variant over a drive cycle. In a further variant model calculations show that the determination of the contact resistance $${\text{R}}_{\text{IK}}=\left\{\text{MW}{\left[\text{Ue}\left(\text{t}\right)\right]}_{\mathrm{\tau }}-\text{MW}{\left[\text{Ua}\left(\text{t}\right)\right]}_{\mathrm{\tau }}\right\}/\text{MW}{\left[\text{I}\left(\text{t}\right)\right]}_{\mathrm{\tau }}$$

with t = "time span driving cycle" works very well. Thus, in one embodiment, the averages are formed over a complete drive cycle and the resistance value 334 determined with these means.

The microcontroller μC or the control unit in the load 206 Thus, the voltage Ua (t) averages over one drive cycle, the intelligent distributor 204 Voltage Ue (t) and current le (t). After completion of the driving cycle, these three data or the average value of the three data on the driving cycle are sent to the infrastructure, also referred to as the cloud.

If a current measurement on the input side of the line is not possible, then it is possible to conclude on the correlation of the voltage at the input Ue (t) and the output Ua (t) on a stochastic change in the contact resistance of the line.

The voltage at the input Ue (t) is again averaged over a drive cycle - in the following, the English notation U _in and for Ua (t) the notation U _{out is} used for Ue (t): MW1 (Uin). Likewise at the output of the line (MW1 (Uout)): These values are transferred to the cloud.

$${\partial }_{M{W}_{{}_2}\left(U_{out}\right)}=M{W}_1\left(U_{out}\right)-M{W}_2\left(M{W}_1\left(U_{out}\right)\right)$$

$$r_{U_{in}{U}_{out}}=\frac{M{W}_2\left({\partial }_{M{W}_2\left(U_{in}\right)}*{\partial }_{M{W}_2\left(U_{out}\right)}\right)}{\sqrt{M{W}_2\left({\partial }_{M{W}_2\left(U_{in}\right)}{}^2\right)*M{W}_2\left({\partial }_{M{W}_2\left(U_{out}\right)}{}^2\right)}}$$

Then a correlation coefficient can be calculated by calculating a further averaging MW2, for example, over the values of 20 driving cycles and thus the dispersion: $${\partial }_{M{W}_{{}_2}\left(U_{in}\right)}=M{W}_1\left(U_{in}\right)-M{W}_2\left(M{W}_1\left(U_{in}\right)\right)$$

$${\partial }_{M{W}_{{}_2}\left(U_{Out}\right)}=M{W}_1\left(U_{Out}\right)-M{W}_2\left(M{W}_1\left(U_{Out}\right)\right)$$

$$r_{U_{in}{U}_{Out}}=\frac{M{W}_2\left({\partial }_{M{W}_2\left(U_{in}\right)}*{\partial }_{M{W}_2\left(U_{Out}\right)}\right)}{\sqrt{M{W}_2\left({\partial }_{M{W}_2\left(U_{in}\right)}{}^2\right)*M{W}_2\left({\partial }_{M{W}_2\left(U_{Out}\right)}{}^2\right)}}$$

The correlation coefficient is then normally 1, since the profile of the two voltages is the same. If a stochastic change in Uout results from a fluctuating contact resistance of the line, then the correlation coefficient decreases to, for example, 0.98.

In the embodiment described here, the device 340 for forming average values, in addition to the averaging, it is also possible to determine the correlation coefficient and the scattering.

In a particularly preferred embodiment, the controller comprises 216 furthermore a memory device 342 that is designed to be read in and to store and provide specific values upon request. Advantageously, the averaged voltage values MW1 (Uin), MW1 (Uout) and, if present, the averaged current value MW1 (I) are stored. If possible from the storage location, the resistance values, the scattering and the correlation coefficient are additionally stored. The stored values become the device 336 provided for determining the error information.

For data reduction in the storage device 342 is provided in a variant that the values are stored only if they represent a new minimum and / or maximum. This check is then done in the maximum determination device 344 ,

The control unit 216 designed to ensure the functional safety of individual lines without redundancy, as potential errors are detected early. In order to be able to carry out a secure and possibly earlier diagnosis of degradation processes with even more data, in a particularly preferred embodiment the values determined or stored by the control unit are transmitted via an output interface 346 along with vehicle information 348 to an infrastructure facility 350 transfer. This is indicated in the description 8th further elaborated.

However, the cloud approach can already be realized within a vehicle if the data are combined in a central memory of the vehicle and then evaluated. Thus, the central evaluation device is capable of a plurality of lines or contacts between the lines and the distributors and To monitor and compare loads. Knowing correlations between individual loads in the vehicle, trends leading to errors can be detected very early on, and a corresponding warning message can be issued to a driver or a workshop.

4 shows in a diagram resistance changes of several connections over time. Shown is the increase in the contact resistance of a line due to a degradation process of one of the two contacts.

5 shows in a diagram voltages on a consumer over time. Shown here is the lowering of the load voltage based on a cyclically recurring dynamic load with a constant assumed supply voltage.

6 shows a symbolic representation of the voltage drop over time. The loads move in a concrete application or a specific vehicle configuration in a fixed area whose lower and upper limits are known and constant. It follows that the voltage drop moves within a predefined voltage band correlating to the load band. Without an error, the input voltage to the line should also be considered constant within a tolerance range. By sending this forward, one can expect the averaged voltage drop within an unfocused cloud. If this region of an expected value for the voltage drop, referred to here as a fuzzy cloud, migrates, an error can be concluded.

7 shows a schematic representation of a vehicle electrical system 200 of a vehicle. For clarity, the same components are not consistently provided with reference numerals. Compared to the in 2 illustrated electrical system 200 here are additional subdistributors 704 and more loads 206 shown.

In one embodiment, almost any controller measures 216 in the electrical system 200 its supply voltage. One approach of the inventive idea is that by means of a stochastic evaluation of all voltage measurement signals in the system, imminent errors are detected in the overall system. A strategy is used that keeps the additional load of the bus systems as low as possible.

A simultaneous, prolonged lowering of the voltage signals U11 and U12 with respect to the other voltage signals indicates a degradation process of the line H1 or the sub-distributor 704 shut down. A sustained lowering of only the voltage U21 indicates a problem of the line L21 or the load 206 shut down.

Each controller determines the probability distribution and spread of the voltage measurements. Raw data does not have to be transmitted. After a predetermined time interval, the scattering and probability distribution data are transmitted to a master. This recognizes shifts in the stochastic sizes. The predetermined time interval may be up to 120 minutes, which may be, for example, greater than 3 and less than 60 minutes. Thus, in one exemplary embodiment, the data for scattering and probability distribution per driving cycle are collected and transmitted every 10 minutes. In an alternative embodiment, one data item each is determined and transmitted for scattering and probability distribution for a complete drive cycle, for example, the time interval may be a duration of up to 8 or 24 hours.

8th shows a plurality of vehicles 860 each with at least one control unit 216 for monitoring the electrical system 200 as well as an infrastructure facility 350 , As already in the description too 3 executed, is the control unit 216 trained in the control unit 216 present data to an infrastructure facility 350 transferred to.

In a preferred embodiment, the data is from a plurality of controllers 216 of the electrical system 200 transmitted to a higher-level control unit in the vehicle, which parts of the tasks of processing and storing and in particular the transfer to the infrastructure 350 takes over.

The infrastructure facility 350 includes a central storage device 862 as well as a device 864 an interface for receiving the plurality of values from the plurality of vehicles 860 a memory device for storing the plurality of values in the central storage device 862 a device for determining at least one error information for at least one electrical system 200 one of the vehicles 860 using the plurality of values and an interface to provide the error information to at least one of the vehicles 860 or to a monitoring function, for example in the form of a control center.

In the illustrated example, the error information may concern the affected vehicles 860 which include the connection or loads in the on-board network and an error probability or trend for the fault.

The above-described method or controller will be described with an illustrative concrete example of monitoring a Connection quality between a current source and a current sink and determining a resistance of the connecting line described. However, this can generally be used in a vehicle and also monitor individual sensors, loads or other components in the electrical system or components connected to the vehicle electrical system. The inventive arrangement also relates to a learning system which is able to accommodate the allowed operating states (allowed state space) in a teach-in process. The teach-in process can be based on a vehicle electrical system simulation. The same can be the learning process during test drives, such as summer test and winter test, made. The result of the teach-in process is a state space of dimension n corresponding to the number of linked parameters pn , Lying during vehicle use individual parameters pn outside the allowed state space, an error or at least a degradation is detected. This is then provided as error information.

By error injection, it must be checked whether the state space is suitable for reliably detecting faults. There must be no false positive or false negative detections. If this is the case, then the parameter space may have to be increased by one parameter and thus by one dimension, such as the parameter ambient temperature or driving dynamics.

Thus, a standardized procedure for the diagnosis of all functions of the vehicle is created. The number of parameters pn can be adjusted if necessary. From the original state space P1, P2, ... Pn, functional mappings can be made into new state spaces, eg dP1 / dt, dP2 / dt, ... dPn / dt.

The functional mappings can also be done in state spaces with reduced dimensionality up to dimension 1 P1 or dimension 0 with the statement good / bad. The approach allows a reduction in the volume of data for remote online diagnostics for Connected Vehicles fleets.

Another application is in 9 shown. Given is an arrangement with a source, an electronic distributor with current and voltage measurement, as well as a consumer. As a source of error in the arrangement according to 9 here assumed a parallel arc. The load chosen here 206 is eg an air conditioning compressor in the 48 V electrical system, which can be operated in three power levels 1 kW, 2 kW and 3 kW.

The consumer 206 is accepted for performance. With fluctuations in the vehicle electrical system voltage U (t) also varies its input current I (t) corresponding to P = U * I.

The following figures show read data from electricity I (t) via voltage U (t) or the values derived from it. In a teach-in process, current and voltage in the electronic power distributor are measured, for example, every 10 ms and in a state space 1000 recorded. So shows 10 the two-dimensional state space 1000 I over U. In this, the three performance levels of the compressor are formed. Analog here shows 11 a state space derived via the functions dU / dt and dl / dt 1000 , This refers to the time changes of the parameters, including the switching of the power levels.

12 and 13 show that envelopes of the "good" state spaces 1000 can be defined. This "good" envelope, referred to as discrimination limit (s) 1200, in the state spaces 1000 thus allow a classification of the load.

If the measured values of the online measurement in the vehicle move out of the "good" envelope, this indicates an error.

14 and 15 show how the measured values fall out of the "good" envelope with the occurrence of a parallel arc. Criteria for error detection can now be defined, eg errors, if at least 5 values outside the "good" envelope occur in 50ms and distance less than 20k and with 2 values outside per 10 ms, if the distance to the "good" envelope is> 20k ,

It may be that the state space shown above only at one temperature, ie at e.g. 25 ° C applies. If this is the case, the state space is extended by the dimension "temperature".

The prerequisite for a successful execution of the presented method is that a classification of the load according to 12 / 13 is possible.

The procedure allows a reduction in the amount of data for remote online diagnostics for fleets of connected vehicles. For this purpose, the data is only communicated "upwards" if it moves into its center of gravity in the state space or moves towards the boundary of the "good" envelope (pre-warning subspace).

16 shows the original state space 1000 a compressor that is infinitely variable. Since a large room is raised, the delineation of "bad" states is difficult. On the other hand, the rate of change of the differential Current value used - see 17 , so in this example, the controller characteristic of the controlled load is formed. In the case of a fault (arc), condition points outside the controller behavior of the load occur. With this classification, therefore, a recognition of the error would be possible.

18 shows a further embodiment of a control device 216 for monitoring a vehicle electrical system. The control unit 216 includes at least one device 330 for reading a time series of data for a number of parameters, the parameters representing information about an operation of the vehicle by a driver and / or a state of the vehicle and / or a state of the driver and / or a driving environment, a device 336 for classifying the data as a normal value or an error value, the normal values per parameter lying within a state space separated from the error values by means of a discrimination limit, and a device 346 for evaluating the data classified as an error value, an error being detected when the data classified as error value satisfy a criterion and / or evaluating the data classified as normal value, the data classified as normal being statistically evaluated to determine a statistical parameter and an error is detected if the statistical parameter exceeds a threshold.

LIST OF REFERENCE NUMBERS

100

method

S1

Step of reading in

S2

Step of classifying

S3

Step of the evaluation

200

board network

202

energy

204

distributor

206

load

208

Input of the load

210

management

212

Output of the distributor

214

Communication connection, bus

216

control unit

U _e (t)

input voltage

U _a (t)

output voltage

I (t)

electricity

R _ik

Resistance (of the line)

330

Interface, device for reading

332

Device for determining

334

resistance

336

Device for determining, device for classifying

338

error information

339

Comparative information

.DELTA.U

Voltage value, voltage drop

340

Device for forming mean values

342

memory device

344

Maxima determiner

346

Output interface

348

vehicle information

350

infrastructure facility

U ₁₁ , U ₁₂

voltage signals

H ₁

management

704

sub-distribution

U ₂₁

tension

L ₂₁

management

860

vehicle

862

central storage device

864

contraption

pn

parameter

1000

state space

1200

discrimination border

QUOTES INCLUDE IN THE DESCRIPTION

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

Cited patent literature

• DE 102015209588 B3 [0006]
• DE 102014018640 B3 [0007]
• DE 102014004791 B3 [0008]

Claims (19)

A method of monitoring a vehicle electrical system (200), the on-board network (200) having at least one distributor (207; 704) and a load (206) connected via a line (210; L21, H1) to the following ones steps: a. Reading in (S1) a chronological sequence of data for a number of parameters, wherein the parameters represent information about an operation of the vehicle by a driver and / or a state of the vehicle and / or a state of the driver and / or a driving environment; b. Classifying (S2) the data as a normal value or an error value, the normal values per parameter lying within a state space separated from the error values by means of a discrimination limit; c. Evaluating (S3) the data classified as the error value, wherein an error is detected when the data classified as error value satisfy a criterion and / or evaluating the data classified as normal value, wherein the data classified as normal value are statistically evaluated to a stochastic parameter determine and wherein an error information is determined when the stochastic parameter exceeds a threshold. Method according to Claim 1 in which, in the step (S3) of the evaluation, a connection quality is monitored between components of the on-board network (200) in which, in the reading-in step, a resistance value (Rik; 334) of the line (210; L ₂₁ , H ₁ ) during the Operation of the vehicle (860) is read and / or a number of parameters is read from which the resistance value is determined, wherein between the provisions of a drive motor of the vehicle (860) has been turned off at least once. Method (100) according to Claim 2 in which, in step (S1) of reading, a voltage value (ΔU) representing a voltage drop across the line (210; L21, H1) is read in as the first parameter, and wherein the resistance value (Rik; 334) is determined using the voltage value (ΔU) is determined. Method (100) according to Claim 3 in which, after the step (S1) of reading in, an average value is formed via voltage values (ΔU) read in at a predefined mean time interval and the mean value is used as voltage value (ΔU) in the following steps. Method (100) according to Claim 4 in which, using the average formed over the mean time interval and a plurality of mean values formed over previous average time intervals, a dispersion is determined and in the step of determining the error information (338) is determined using the dispersion. Method (100) according to one of Claims 3 to 5 wherein, to determine the voltage drop (ΔU) across the line (210; L21, H1), an output voltage value (Ue (t)) representing an output voltage at the output (212) of the distributor (204) and an input voltage thereto via the line (210; L21, H1) input (208) connected to the output (212) of a sub-distributor (704) or input voltage value (Ua (t)) representing the load (206) is read. Method (100) according to one of the preceding claims, in which, in the reading-in step, a current value (I (t)) which is a current between an output (212) of the distributor (204) and an input (208) of the load (206) is read in, and wherein the resistance value (Rik; 334) is determined using the current value (I (t)). A method (100) according to any one of the preceding claims, comprising a step of storing data for a number of parameters, the data being the resistance values (Rik; 334), the voltage values (ΔU), the current values (I (t)), the output voltage values (Ue (t)), the input voltage values (Ua (t)), the average values and / or at least one stochastic parameter. Method (100) according to Claim 8 in which, in the step of saving, the values to be stored are only stored if they represent a new minimum and / or maximum. Method (100) according to Claim 8 or 9 in which, in the step of saving, the values to be stored are only stored if they have a minimum distance from a previously stored value. Method (100) according to one of Claims 8 to 10 in which in the step of storing at least one further vehicle information is stored, and in the step of determining the error information is determined using the at least one further vehicle information. Method (100) according to one of Claims 8 to 11 , comprising a step of transmitting the values stored in the step of storing to a central infrastructure device (350). Method (100) according to one of the preceding claims, characterized in that the steps of reading, forming, classifying, determining, evaluating or determining, providing, storing and / or transmitting are carried out repeatedly, in particular, wherein in the step of determining the error information (338) is determined using the stored values. Method (100) according to one of the preceding claims with at least one of the following steps: a. Receiving a plurality of values from a plurality of vehicles (860) in an infrastructure (350); b. Storing the plurality of values in a central storage device (862) of the infrastructure device (350); c. Determining at least one fault information (338) for at least one on-board network (200) of one of the vehicles (860) using the plurality of values of the plurality of vehicles (860); and d. Providing the fault information (338) from the infrastructure device (350) to at least one of the vehicles (860) and / or a monitoring function for the plurality of vehicles (860). Method (100) according to one of the preceding claims, in which a center of gravity of the data is determined in the step (S3) of the evaluation as a statistical parameter, and a degradation and / or an error is predicted if a movement of the center of gravity increases over time Threshold is detected. Method (100) according to one of the preceding claims, wherein in the step (S3) of the evaluation as criterion a number of data classified as error value in a predetermined time interval exceeds a criterion limit value. Method (100) according to one of the preceding claims, comprising a learning step in which a state space is learned, wherein in a first sub-step of reading in a time sequence of learning data is read for a number of parameters previously categorized representing normal value , wherein in a second sub-step at least one envelope is determined as a discrimination limit for the learned state space. Method (100) according to one of the preceding claims, comprising a step of deriving between the step of reading in and the step of classifying, wherein the at least one chronological sequence of data is derived, and classifying the derived data in the step of classifying and in the step of Evaluating the classified, evaluated data are evaluated. Control unit for monitoring a vehicle electrical system (200), wherein the vehicle electrical system (200) has at least one distributor (207; 704) and a load (206) which are connected via a line (210; L21, H1) to the following ones Features: a. a device for reading in a chronological sequence of data for a number of parameters, wherein the parameters represent information about an operation of the vehicle by a driver and / or a state of the vehicle and / or a state of the driver and / or a driving environment; b. means for classifying the data as a normal value or an error value, the normal values per parameter lying within a state space separated from the error values by means of a discrimination limit; c. means for evaluating the data classified as an error value, wherein an error is detected when the data classified as an error value meets a criterion and / or evaluating the data classified as normal value, the data classified as normal being statistically evaluated to a statistical parameter and an error is detected if the statistical parameter exceeds a threshold.

Images