EP3750089A1 - Verfahren und system zur analyse wenigstens einer einrichtung einer einheit, welche eine mehrzahl an verschiedenen einrichtungen aufweist - Google Patents
Verfahren und system zur analyse wenigstens einer einrichtung einer einheit, welche eine mehrzahl an verschiedenen einrichtungen aufweistInfo
- Publication number
- EP3750089A1 EP3750089A1 EP19709362.8A EP19709362A EP3750089A1 EP 3750089 A1 EP3750089 A1 EP 3750089A1 EP 19709362 A EP19709362 A EP 19709362A EP 3750089 A1 EP3750089 A1 EP 3750089A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- data
- unit
- transformation model
- units
- devices
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0808—Diagnosing performance data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/15—Vehicle, aircraft or watercraft design
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Definitions
- the invention relates to a method for analyzing at least one device of a unit which has a plurality of different devices, in particular a vehicle or an engine. Furthermore, the invention relates to a corresponding method for training an artificial neural network and corresponding systems for analyzing at least one device and for training an artificial neural network.
- quality standards are prescribed by the legislator, which are binding for the manufacturer.
- motor vehicle manufacturers must be able to guarantee that all vehicles of a vehicle type supplied by the manufacturer comply with predefined emission values over their useful life.
- Another example of the quality of a product is the probability of failure of individual devices, in particular components / components, of a vehicle. This probability of default has a direct impact on the total cost of ownership of a vehicle, which is very important to the end user. Also in this regard, a vehicle of a vehicle type should not exceed a predetermined value. In order to be able to ensure the function of a product even after its manufacture and delivery to the customer, it is therefore of great interest of the manufacturer to be able to determine the characteristics of the product, in the case of units with multiple devices, even individual units of the unit, at any time , Such properties may be, for example, tolerances of devices of the units, but also, for example, properties which change as a result of wear or other properties.
- Knowing the characteristics of a product or devices that are part of this product is generally more important the more individual devices make up the product or unit. For such products, for example, it is often difficult to conclude, or even identify which feature of the device causes the symptom, from a symptom that suggests a malfunction, to the device underlying the malfunction.
- DE 10 2015 205 205 A1 relates to a data analysis system for unknown error data having an analysis means configured to analyze error data received from a vehicle to identify error data corresponding to one of a plurality of known error patterns; wherein the analyzing means is configured to classify the received error data as unknown error data if the received error data can not be identified as corresponding to one of the known error patterns.
- sources of error in units, in particular faulty devices, should be identified or even predicted.
- a corresponding model should be formed.
- a first aspect of the invention relates to a, in particular computer-aided, method for analyzing at least one device of a unit to be tested, a category of units having a plurality of different devices, in particular a vehicle or motor, preferably comprising the following steps: simulating a Operating a unit of the genus in a defined
- Observable quantities can be determined for the characteristics used for simulating a property of the devices to be analyzed, in particular by means of a statistical method;
- Operating cycle recorded data set to issue an expression of at least one device to be analyzed Operating cycle recorded data set to issue an expression of at least one device to be analyzed.
- a second aspect of the invention relates to a corresponding system for analyzing at least one device of a unit to be tested of a class of units comprising a plurality of different devices, in particular a vehicle or motor, the system preferably comprising:
- Simulating deployed characteristics of a property of the device to be analyzed can be determined, in particular by means of a statistical method
- Means for performing a balancing calculation based on the selected data regions and the correlative expression Means for performing a balancing calculation based on the selected data regions and the correlative expression
- a third aspect of the invention relates to a, in particular computer-based, method for training a transformation model, in particular an artificial neural network, for characterizing units of a class of units which have a plurality of different devices, in particular a vehicle or engine, preferably the following Working steps comprising:
- Simulating used characteristics of a property of the at least one device is detectable, in particular by means of a statistical method; and / or - reading the selected data areas and the correlating expressions into a compensation calculation, which forms the basis for the transformation model.
- the transformation model is generated on the basis of the compensation calculation.
- a fourth aspect of the invention relates to a corresponding system for training a transformation model, in particular an artificial neural network, for characterizing units of a class of units comprising a plurality of different devices, in particular a vehicle or engine, preferably comprising: means for simulating an operation of a unit of the genus in a defined
- a fifth aspect of the invention relates to a method, in particular computer-aided, for analyzing at least one device of a unit to be tested of a class of units comprising a plurality of different devices have, in particular a vehicle or engine, preferably the following steps:
- Characterization of the at least one device by means of a transformation model which has an assignment rule between observation variables and a property of devices to be analyzed, wherein the transformation model is based on a compensation calculation with respect to simulation results which consist of several operating modes.
- a sixth aspect of the invention relates to a corresponding system for analyzing at least one device of a unit to be tested of a class of units which have a plurality of different devices, in particular of a vehicle or engine, preferably comprising:
- Observational quantities in relation to the unit to be tested Means for determining, on the basis of the data set, an expression by means of a transformation model, which has an assignment rule between observation variables and a property for analyzing the device, wherein the transformation model is based on a compensation calculation with respect to simulation results which comprise a plurality of operating modes.
- a seventh and eighth aspect of the invention relate to a computer program and a computer-readable medium for carrying out the inventive methods according to the first, third and fifth aspects of the invention.
- a ninth aspect of the invention relates to a data sequence which characterizes a transformation model for analyzing at least one device of a unit to be tested of a class of units comprising a plurality of different devices, wherein the transformation model for a defined operating cycle of the class of units comprises an assignment rule Observation variables and a property of at least one device, wherein the transformation model is based on a compensation calculation of simulation results, which are mapped from several operating simulations of the defined operating cycle by means of a model in which at least analyzable devices as a virtual device, each with different configuration of the units result, each configuration being characterized by an expression of a property of at least one device.
- a class of units in the sense of the invention is preferably a set of units which agree in their essential characteristics and therefore more preferably identical.
- a motor as a unit, for example, preferably at least the essential components of the motors of a genus are identical.
- a particular unit is thus preferably a realization of the class of units.
- a transformation model in the sense of the invention preferably establishes a relationship between observation variables and target variables or a group of target variables.
- a transformation model establishes a relationship between observational quantities of a unit of the class of units and a property of at least one unit of the unit.
- the transformation model is not time resolved. More preferably, however, the transformation model itself is based on a time-resolved simulation of the operation of a plurality of units of the class of units each having a different configuration.
- Simulating an operation of a unit within the meaning of the invention is preferably a time-resolved or distance-resolved simulation of the operating behavior of a unit, in particular of a vehicle, of a drive train or of an engine, by means of a model. Operation of a unit is characterized in particular by observational variables such as speed, torque, mass flows, pressures, temperatures, fuel consumption, consumption of other resources, emissions, OBD values, speed, gear selection, etc.
- Simulation-based within the meaning of the invention preferably means that a transformation model based on performed or to be performed simulations is used.
- a unit in the sense of the invention is preferably a device whose operation can be characterized by observation variables.
- the observation variables are at least partially derived from measured variables or from measured variables, wherein the values of the measured variables can be based on an actual measurement or on a simulation.
- such measured variables are recorded with sensors.
- Observation variables are preferably also or alternatively Signals which serve to control and / or monitor the unit and / or its devices.
- the devices of the unit are at least partially electronically controlled. Further preferably, the unit is controlled electronically per se.
- Non-limiting examples of a unit are apparatus or land vehicles, in particular trucks or passenger cars, or watercraft, in particular ships, or aircraft, in particular rotorcraft or aircraft, or spacecraft, in particular space shuttle or missile.
- Further examples of a unit are robots, in particular military, industrial or personal robots, a test and / or development arrangement, in particular a process control, an automatic identification system (AIS) or a flexible manufacturing system (FMS).
- Further examples of a unit are controlled machines, in particular production machines, machine tools or construction machines, or data processing systems, in particular mobile devices or stationary computers.
- units may also be subsystems of such or other entities, for example a drive train or a component of a drive train, in particular a motor or a transmission, or a chassis, or an energy store, or a fuel cell, or an assistance system.
- exemplary subsystems of a test arrangement are a consumption measurement technique, a
- Emission measurement technology a power measurement or conditioning equipment, or components of these subsystems.
- a device according to the invention is preferably a component of a unit, in particular a component or an assembly.
- devices depending on what is considered as a unit, are mechanical elements, in particular a pipe, a turbocharger, a heat exchanger or a machine element, or mechatronic elements, in particular an actuator, an injection system or its components, a variable turbine geometry, an electrical charger of an intake system, a sensor, or electronic elements, in particular power electronics, an electronic component or sensors, or electrical elements, in particular an electric drive or generator, a heater, a voltage converter or a cooler, or electro-chemical elements, in particular a battery, an exhaust aftertreatment device or a fuel cell.
- different devices or systems can be classified either as a unit or as a device, depending on whether the device or system is to be analyzed within a higher-level unit or, in turn, its subordinate devices are to be analyzed.
- Properties in the sense of the invention are preferably features of a device which are suitable for characterizing them.
- Non-limiting examples of properties include dimensions, manufacturing tolerances, installation tolerances, aging changes or drift changes.
- a vehicle according to the invention is any type of land, water, air or spacecraft, in particular a motor vehicle.
- Recording in the sense of the invention is preferably a, at least temporary, storage of data.
- a compensation calculation in the sense of the invention is preferably a regression method or a pattern recognition method, in particular based on an artificial neural network or an artificial neural network, and / or a combination of both, in particular in the sense of statistical teaching.
- Output according to the invention is preferably a provision.
- the provision preferably takes place via a data interface.
- a configuration in the sense of the invention preferably corresponds to a realization of the unit, which by an expression of at least one property at least one of the devices of the unit is characterized.
- a configuration and / or an expression is preferably characterized by a component specification, a calibration parameter, an accuracy and / or an aging state.
- An operating cycle in the sense of the invention is preferably a time-resolved and / or distance-resolved sequence of operating points of an operation of a unit.
- a model in the sense of the invention is preferably a pure software model or even a model which consists of hardware, for example an emulator, connected to a software model.
- a model can model the unit or parts of the unit.
- a driving state in the sense of the invention is preferably defined by one or more values of an observation variable or constellation or several constellations of values of several observation variables, depending on whether the driving state is considered situationally (for example the presence of a cornering) or whether a driving state is itself only from the time course of parameters results (for example, the presence of a tip-ins).
- a driving state within the meaning of the invention is particularly the driving dynamics of the vehicle again.
- Driving conditions are in particular gliding at constant speed, acceleration, cornering, parking, straight ahead, idling (roll-ride), tip-in (sudden gas), let-off (sudden gas take-off), constant driving, switching, standstill, ascent, descent, Electric driving, braking by recuperation, mechanical braking or even a combination of at least two of these driving conditions.
- the driving dynamics are also determined by the type of drive or by the operating state of vehicle components.
- three different tip-in driving conditions are possible, a tip-in, which is driven by the internal combustion engine, a tip-in, which is driven by the electric machine and a tip-in, in which the electric machine as additional Electric boost is used.
- a means in the sense of the present invention may preferably be formed by hardware and / or software technology, in particular a data or signal-connected, preferably digital, processing, in particular microprocessor unit (CPU) and / or preferably with a memory and / or bus system one or more programs or program modules.
- the microprocessor unit may be configured to execute instructions implemented as a program stored in a memory system, to capture input signals from a data bus, and / or to output output signals to a data bus.
- a storage system may comprise one or more, in particular different, storage media, in particular optical, magnetic, solid state and / or other non-volatile media.
- the program may be such that it is capable of embodying or executing the methods described herein so that the microprocessor unit may perform the steps of such methods and thus, in particular, analyze at least one device or train a transformation model.
- the invention is based in particular on the knowledge that there are correlations or correlations between the data generated in an operation of a unit, which are available, for example, in the CAN network of a vehicle, and the configurations of units of the unit which influence the operation that are difficult to identify using traditional data processing techniques.
- the data amounts are generally too large, too complex, too fast-paced and / or too weakly structured to be able to recognize these relationships, in particular completely.
- the inventors have developed a method and a system with which such relationships can be mapped by means of a compensation calculation, in particular with artificial neural networks.
- correlations that are determined for a defined operating cycle in certain data areas of the recorded operation can be used here. These correlations are again used as the basis for training a transformation model.
- This training phase can be carried out without the use or operation of a real unit.
- simulations of the unit are used by means of a model.
- properties of one or more devices of a unit can be analyzed and determined on the basis of an operation of a real unit in the defined operating cycle.
- the variability of the operating behavior of a unit to be tested can be explained unambiguously by means of a transformation model purposefully trained according to the invention. Gaps or inaccurate measurements can thus be excluded completely according to the invention. In this way, even weak effects of varied factors are detectable and thus included in the transformation model.
- the component specification of a device can be determined without having to separately analyze the device separately.
- a device with its component specifications moves outside of a predetermined target range, the device can be replaced as a non-non-ok part even before it fails.
- learning parameters in a control unit can be adapted as a function of the component properties of the respectively analyzed device. This allows optimized operation of the analyzed device and thus the unit. For example, a learning parameter can adapt the control by a control unit even for cases in which a device is outside the allowable component tolerances. Also can be counteracted or compensated with a learning parameter in this way an aging or wear of a component.
- a so-called tolerance image can also be created after installation of the individual devices of a unit.
- the change in the tolerances before installation in the unit and after installation in the unit is determined by means of the method according to the invention. This allows the effects of the installation to be determined on the respective tolerances.
- component specifications of the individual device analyzed can be adapted, for example for optimization.
- statistics can be generated, for example with regard to installation tolerances. Basically, it is intended to perform a simulation of the operation of the unit on a test bench, partly on a test bench and partly model-based or preferably purely model-based and to first create a transformation model based on this simulation.
- the assignment rule in particular exclusively, is valid in those areas of the defined operating cycle which correspond to the selected data areas.
- the selection of the data area is performed by means of a feature selection method, preferably a supervised feature selection method, more preferably a reliefF method.
- the methods according to the invention furthermore have the following working steps:
- Differential records are stored; selecting data areas from the data sets and / or the average data sets and / or the differential data sets.
- the respective moving average value and the value of the difference quotient are preferably calculated for each individual value of the recorded data records, which correspond in each case to a characteristic of a property of a device. These calculated values then form the respective average data record and the respective difference data record.
- the average data sets are preferably combined in an overall average data record, the differential data sets are preferably combined in an overall difference data record.
- the systems according to the invention have: means for calculating a moving average of recorded
- the simulation is a hardware-in-the-loop simulation, wherein at least one real control, in particular a motor control, of the unit to be tested is integrated in the simulation.
- the operation of a unit to be tested is indeed simulated, but the control of the unit to be tested per se is not simulated.
- the real control algorithm is integrated in the simulation or even the entire hardware of the controller including the control algorithm is integrated.
- the systems according to the invention have:
- the methods according to the invention furthermore have the following working step: splitting the recorded data records into a conditioning section and a
- Identification section based on at least one criterion, wherein selected data areas are located exclusively in the identification section.
- a conditioning section it can be ensured that defined initial conditions are created for selecting the data areas, in particular for identifying features in the data records in a feature selection method.
- One criterion could be, for example, the loading state of a particulate filter. For example, it may be required that the particulate filter should be in a regenerated state at the beginning of the identification section.
- the systems according to the invention have:
- a real operating cycle in the sense of the invention is preferably present in a real environment, in particular in a real drive in real traffic.
- the systems have:
- the methods according to the invention furthermore have the following working steps:
- the simulation of the operation of a plurality of units of the type can likewise be carried out by means of the real operating cycle as a defined operating cycle.
- the transformation model is then generated based on this simulation of the real operating cycle.
- a transformation model is created, by means of which the unit to be tested can be analyzed, in hindsight, by applying the transformation model to the recorded data set of the unit to be tested.
- the methods according to the invention furthermore have the following working step:
- the characteristic and the transformation model can be output for further use in data processing via a data interface.
- the characteristics preferably provide information about the particular configuration of the unit to be tested.
- the systems according to the invention have a data interface for outputting the characteristic and / or the transformation model, in particular a regression model.
- the methods according to the invention furthermore have the following working steps:
- the systems according to the invention have:
- the method according to the invention also has the following working step: checking at least one trigger condition with respect to the data record of the unit to be tested recorded during the defined operating cycle, in particular whether at least one observation parameter assumes a predefined value or a predefined value combination, which corresponds at least to a section, preferably a driving state, of a data area for which there is a correlation to the characteristics of a property of the at least one device; wherein the operations of applying the transformation model and / or recording and / or determining an expression are performed only when the at least one trigger condition is met.
- the course of the defined operating cycle preferably corresponds to a certain operating state of the unit.
- the systems according to the invention have means for checking at least one trigger condition with respect to the data set of the unit to be tested recorded during the defined operating cycle, in particular whether at least one observation variable assumes a predefined value or a predefined value combination, preferably an operating condition which at least corresponds to a section of a data area for which a correlation exists with the characteristics of a property of the at least one device, wherein the systems execute the application of the transformation model and / or the recording and / or the determination of an expression only if the at least one trigger condition is met.
- the application of a transformation model or the operation of the unit under test in particular exclusively, in sections of the defined operating cycle, in particular in operating states, performed in which a correlation of observation variables to the characteristics of a property of the at least one device consists.
- the operating cycle when determining occurrences on statistically relevant parts can be restricted.
- a shortened operating cycle can be run through with the unit to be tested.
- test times and simulation capacities can be saved.
- the methods according to the invention also have the following working step: Checking whether an OBD event occurs, wherein the operations of applying and / or recording and / or determining are performed only when an OBD event occurs, in particular over a defined period of time before and / or after the OBD event ,
- OBD events can be evaluated. In particular, it can be determined whether an actual OBD event, a so-called validated OBD event, is present, or whether, for example, the sensor responsible for determining the OBD event outputs an incorrect value.
- learning parameters of the OBD system can then be adapted.
- the systems according to the invention have means for checking whether an OBD event occurs, wherein the operations of applying and / or recording and / or determining are executed only when an OBD event occurs, in particular over a defined period of time before and / or after the OBD event.
- the methods according to the invention furthermore have at least one of the following working steps, which are preferably carried out periodically or use-specifically:
- Facilities of the unit are taken into account, which lead to error message, but actually do not affect the function of the respective device. Determining a replacement date of a component of a device or the
- the systems according to the invention have:
- a plurality of devices of the unit to be tested is analyzed.
- the methods furthermore preferably have the following working step: Determining a faulty or functioning device of the plurality of devices, wherein a device is faulty in particular if the output characteristic of the device with respect to at least one property is outside a predetermined range.
- the systems according to the invention comprise means for determining a faulty or functioning device of the plurality of devices, wherein a device is faulty in particular if the output characteristic of the device is outside a predetermined range with respect to at least one property.
- the methods according to the invention also have the following working steps: transmission of the data record of observation variables of the unit to be tested to a central computer; wherein a data processing, in particular wherein the operations of reading into a compensation calculation, the application of the transformation model and / or the determination of an expression on the
- a system for data processing in the respective unit to be tested can be equipped with low computing power.
- the system may be simplistic since software updates are not necessarily required.
- the elaborate evaluation of the data sets takes place on the central computer, where also preferably the transformation model is deposited or even formed.
- the systems according to the invention have:
- the systems according to the invention also have a central computer in this case.
- the defined operating cycle is a conformity of production cycle or an end of line cycle in which the unit to be tested, in particular immediately after production, is operated from outside, in particular in towing mode.
- characteristics for example tolerances, of equipment can already be determined after production. These data can be evaluated by quality assurance and any countermeasures taken in the event of faulty equipment.
- the compensation calculation and / or the transformation model is based on a regression model, in particular an artificial neural network.
- Fig. 1 is a functional block diagram of an embodiment of a system for analyzing at least one device of a unit to be tested of a class of units according to the sixth aspect of the invention for carrying out a method according to the fifth aspect of the invention;
- Fig. 2 is a functional block diagram of an embodiment of a part of a
- 3 is a functional block diagram of an embodiment of another
- Fig. 4 is a diagram of a measurement of an engine speed and the sliding
- Fig. 9 is a table of an example of data selection
- Fig. 11 is a graph of measured turbocharger speed versus time in which the selected data areas of Fig. 9 are identified;
- FIG. 12 is a diagram illustrating the measurement accuracy predicted by the method according to the invention and the actual measurement accuracy of a boost pressure sensor
- FIG. 13 shows a temporal speed curve of an operating cycle during which the
- Fig. 15 is a block diagram of a method for training a
- Transformation model according to the third aspect of the invention is a block diagram of a method for analyzing at least one device of a unit-of-type unit under test according to the fifth aspect of the invention.
- the invention will be described below in terms of a motor or a prime mover as units X.
- a vehicle 1, in particular a motor vehicle as a whole or an entire drive train of such a vehicle 1 may also be considered as units X.
- FIG. 1 shows a functional block diagram for a system 30 for analyzing at least one device j of a unit x to be checked of a class of units X, which is set up to execute a corresponding method 300.
- the individual means 31, 32, 33, 34 are shown as separate blocks. However, it will be understood by those skilled in the art that the individual means may be part of one or more modules that summarize individual means.
- the motor x belongs to a genus of engines X, which are each at least substantially identical.
- the motors X have the same facilities.
- a motor x to be tested is operated 301 with means 31 in an operating cycle B.
- the means 31 for operating the engine x are preferably a test bench on which a driving cycle is run as a defined operating cycle B.
- the operation 301 of the engine x can also take place in a ferry operation of a vehicle 1, for example on a chassis dynamometer.
- an operation 301 can also be carried out in real ferry operation on a travel route, which then corresponds to the defined driving cycle B.
- the operation 301 of the engine to be tested x is therefore preferably a real operation in the vehicle 1 or on the test bench.
- individual devices j of the unit to be tested x or also the entire unit to be tested x can be simulated as a simulation model with an unknown configuration x x or with respect to the units j with unknown characteristics x nj of the properties X.
- the defined operating cycles B used can preferably also be a conformity-of-production cycle or an end-of-line cycle.
- motors to be tested x can also be operated from the outside in towing mode, more preferably, for example, immediately after production for purposes of quality control.
- a drive machine control unit in particular the engine control unit (ECU)
- ECU engine control unit
- data which is available at a drive machine control unit, in particular the engine control unit (ECU) is recorded 304 in a data memory 32 as a data set Y x .
- a real drive cycle B is used as the defined operating cycle, it is preferably additionally a history of this real driving cycle is recorded during a route traveled by a vehicle 1 303.
- Such a recorded real Operating cycle can, as will be explained later, in particular serve as a defined operating cycle B for generating a transformation model TM.
- the data set Y x is preferably a data matrix in which the various recorded tracks of the individual measurement or control data available on the engine control unit are contained, as shown in FIG. 1.
- This data set Y x can, for example, be data tracks over the time of the rotational speed, of the torque, of the mass flows, of pressures and temperatures, of the
- the recorded data matrix Y x is subsequently provided to a means 33 for determining 307 at least one characteristic x nj of at least one property X, of at least one device j of the engine to be tested x.
- a transformation model TM is used, which is a
- a transformation model TM is based on an artificial neural network KNN, as will be explained in more detail below.
- the observation variables r preferably correspond to individual control unit channels of the engine control unit. These observation variables r are preferably normalized to values between 0 and 1 for use in the method 300 according to the invention.
- the characteristic or characteristics x nj of the at least one device X, of the motor x is preferably output by means of a data interface 34.
- the expression x nj is a value or a combination of values of a property X of the device j and indicates the functional state of the device j. If a plurality of occurrences x nj of several devices j or even of all devices j of the motor x is determined by means of a suitable driving cycle B, a configuration x x of the motor x can be derived. Such a configuration x x characterizes a functional state of the engine x and has the characteristics x nj of the characteristics X of the devices j of the engine x. As shown in FIG.
- occurrences x may be, for example, a value of a power of an actuator x xi , a value of an accuracy of a sensor x, ⁇ 2, or a value of an aging state of a catalyst x ⁇ .
- each device j has several properties X,.
- the simplifying assumption is made that each device j has only one property X 1.
- the described property X would mathematically be a tuple or vector with an additional run variable.
- the artificial neural network KNN of the transformation model TM therefore preferably produces a functional relationship between a specific configuration x x of a realization x of a genus of technical units X and the recorded data matrix Y x .
- the particular characteristic or expressions x i of the at least one device j of the motor x to be analyzed is preferably output 308, in particular via a data interface 34.
- an entire configuration x x of the motor x to be tested can also be output if the corresponding values of all Characteristics x, of which facilities j are available or assumptions made.
- the data set Y x of the unit under test x is temporally aligned with the transformation model TM 306-1.
- the sample rate of the data set Y x of the unit to be tested x is adapted to the transformation model TM 306-1.
- the respective data sets Y x are shifted such that the sequence of the defined operating cycle B in the data record Y x recorded by means of the unit under test x is assigned to the transformation model TM in a timely manner.
- the transformation model TM also has time-dependent assignment function functions that are location-dependent or defined by the defined operating cycle B in relation to different points in time of the defined operating cycle B.
- matching the sample rate 306-2 also serves to make the data set Y x of the unit to be tested x processable by the transformation model TM.
- the presence of a data region q is checked 1 10-1, 302-1, for which it is known that a correlation to the expression x nj of a property X, at least one device j exists or is known.
- Such a data area q or section of a data area q preferably corresponds to a driving state or operating state of the vehicle 1 or of the engine x.
- At least one trigger condition is monitored with respect to the recorded data set Y x of the motor x to be tested 1 10-1, 302-1.
- the operation of recording the record Y x of the unit under test x 304 or also determining 307 an expression x, of this unit x can be carried out only if such a driving state or operating state are detected.
- a computational effort in the data evaluation can be significantly reduced.
- a unit under test x is not operated 301 over an entire defined operating cycle B, but selectively only those maneuvers are driven in which driving conditions of the vehicle 1 or the engine x occurrences or occurrences of which it is known that a correlation between observation magnitudes r and the occurrences x nj of a property X.
- a test operation for analyzing the device j of the engine to be tested x can be significantly shortened.
- OBD event on-board diagnosis
- it can be made dependent on whether records Y x are recorded 304 or whether it is attempted to determine an expression x 1, or configuration x x of the engine to be tested 307.
- OBD Systems of a vehicle 1 to be tested for their function or their reliability can make sense above all if OBD Systems of a vehicle 1 to be tested for their function or their reliability.
- a data set Y x of observation variables r of the motor to be tested x is transmitted to a central computer 305.
- the transformation model TM is preferably stored on this central computer, so that the expression x ,, or even x x of the configuration to be tested Unit x can be determined on the central computer 307.
- the transformation model TM itself can be calculated on the central computer.
- data of a multiplicity of vehicles 1 can be selected on such a central computer, which may be advantageous in particular for determining the transformation model TM.
- inventive method according to Figures 1 and 16 can be used for a variety of application examples. At least partially, the teaching described in this regard is also valid for the method 100 according to FIGS. 2, 3 and 14.
- learning parameters may be set based on the determined characteristic or occurrences x i of the motor x to be tested or its configuration x x 1 13-1. 309-1. These learning parameters are preferably stored in the engine control unit ECU of the engine x and can be used, for example, to compensate for aging effects on the engine x.
- a replacement date of a component of a device j or even the device j itself can be determined on the basis of the output characteristic x ,, 1 13-2, 309-2.
- this can also be an automatic message sent to a server, in which, for example, a garage or the manufacturer is informed that the corresponding component or the device j itself must be replaced soon.
- a service date of the engine to be tested x or a device j can be determined on the basis of the output characteristic x ,, 1 13-3, 309-3.
- an OBD indication for example a warning lamp or a field in a display
- the method according to the invention can also be used to simultaneously monitor and analyze a plurality of devices 1 ⁇ j ⁇ k of an engine to be tested.
- the transformation model according to the invention can be identified from this plurality of devices 1 ⁇ j ⁇ k based on a recorded defined operating cycle of the engine to be tested x that device j, which is faulty 1 13-5, 309-5. In particular, this can be determined on the basis of the specific characteristic x.sub.xj of the devices j of the engine to be tested x, in particular if the engine to be tested x lies outside a predetermined value range with respect to a property X.
- Figures 2 and 3 illustrate functional block diagrams relating to the methods 100, 200 of the first and third aspects of the invention. Both Figures relate to only a portion of the methods 100, 200, respectively.
- the methods 100, 200 are described in relation to the corresponding systems 10, 20.
- a separate representation of the exemplary sequence of the method 100 according to the first aspect of the invention is shown again in block diagram in FIG. 14, an exemplary sequence of the method 200 according to the third aspect of the invention in FIG.
- FIGS. 2 and 3 will first be described in relation to the computer-based method 200 for training a transformation model TM for characterizing units x of a class of units X comprising a plurality of different devices j, in particular a vehicle 1 or a vehicle Motors, described.
- a transformation model TM for characterizing units x of a class of units X comprising a plurality of different devices j, in particular a vehicle 1 or a vehicle Motors, described.
- units X of the type of units are described by way of example only.
- the method 200 for training a transformation model TM is used, in particular, to improve the transformation model TM in such a way that it can be used for the analysis of individual engines x to be tested.
- the method 200 is neither a pure simulation method nor the application of a mathematical theory. Rather, due to the inventive method 200 according to the second aspect of the invention, a methodology may be implemented to provide an individualized model of a unit, a so-called digital twin.
- a single transformation model TM is formed, in which all units X of a genus are preferably mapped.
- the basis of the methods 100, 200 according to the invention preferably forms a defined operating cycle B, which is used both to simulate the operation of units X of the type, and to obtain operating data of a unit to be tested x, as already explained with reference to FIG ,
- Such a defined operating cycle B is purely by way of example an arbitrary driving cycle within the scope of the described exemplary embodiments.
- this is a so-called identification cycle, which contains maneuvers that occur particularly frequently in the real operation of a vehicle.
- this driving cycle B can be specified artificially on the basis of various criteria in order to be able to carry out a simulation 203 for training the transformation model TM.
- Such a drive cycle B consists of individual maneuvers or driving states, which are processed one after the other, but do not have to result in a meaningful route course of a vehicle 1.
- a drive cycle B can also be a closed drive cycle, as would occur when a real route is traveled.
- a real driving cycle is to be used as the operating cycle B, a specific unit x, preferably the unit x to be tested later, is operated in a real ferry operation on the road 101, 201. During this operation 101, 201, a real driving cycle is recorded 102-1, 202. In particular, a speed profile over the time or the distance covered or a route profile over the time covered is recorded 102-1, 202. This real driving cycle can then as a defined driving cycle B for generating the Transformation model TM are used.
- additional data is recorded which characterizes the operation of the engine to be tested 102-2.
- data may, for example, as already described with reference to FIG. 1, be those which are available at the engine control unit of the engine to be tested x, for example a speed curve, a torque curve, courses of the mass flows, courses of the pressures and temperatures, fuel consumption, Emission curves, OBD indications, parameter curves and learning parameter settings.
- These data are preferably stored in a data set Y x of the motor to be tested x, which in particular forms a multidimensional data matrix.
- This data matrix Y x is preferably suitable for characterizing the operating behavior during the real ferrying operation of the engine to be tested x.
- such a real data set Y x can be used in various application examples of the analysis method 100, 300 according to the invention.
- data of a plurality of motors X of the same kind are provided.
- Motors of the same type are preferably identical engines of a single series, which are preferably also installed in vehicles 1 of the same type. This ensures that deviations of the operating behavior of the individual motors X of the genus are statistically evaluable.
- each of the motors X is identical, they generally have a different configuration x n, respectively.
- a configuration can be defined, for example, by different powers of an actuator, an accuracy of a sensor, or also by a state of aging of a catalytic converter.
- a property X may also be a production tolerance of a device j.
- the properties X, of the individual devices of a configuration x n of a unit X of the genus are characterized by the respective expression x nj of the respective characteristic X or the device j.
- These occurrences x nj indicate values, value ranges or value constellations of the property X, of the device j.
- the expression x nj can specify a tolerance band of a production tolerance of a device j, in particular of a component, of an engine X of the type.
- an operation in the defined operating cycle B is simulated 103, 203 by means of a model.
- this model preferably at least one device j is depicted as a virtual device.
- all devices 1 ⁇ j ⁇ k are represented by the model.
- the units X are all modeled in the model.
- the model used for simulating is preferably a physical model in which the association between input variables and output variables is modeled by physical relationships.
- the models can also be empirical in nature, that is, the relationships between input variables and output variables are based on empirical empirical values.
- the models can be semi-physical, that is to say both physical and empirical relationships.
- the simulation is preferably carried out on a simulation system 21, as offered, for example, by the applicant within the scope of the software tool Cameo ⁇ .
- the simulation can be carried out as a so-called hardware-in-the-loop simulation, wherein at least the real motor control of the respective motor X is included in the simulation.
- the data of observation quantities r generated by the simulations 103, 203 are recorded as data matrices 104, 204.
- values y rt are assigned to the individual observation variables r for each sampling time in which a measurement could be taken, where r is the respective observation size and t indicates the time.
- the number of all observation variables r and T is preferably the length of the recording of the defined driving cycle B, which simulates the operation of the motors X.
- the individual data sets Yi,..., Y n ,..., Y N are preferably stored in a data memory 12, 22. There, from the individual data records Y n, a total data record Y is preferably formed over all configurations and observation variables as well as over the entire observation period T, as shown in FIG. 2.
- the individual recorded data sets Y n are preferably adjusted in time 105-1, 205-1, so that identical time periods in the drive cycle B in the overall matrix Y are superimposed in each case. Furthermore, for the same purpose, the sample rate of the recorded data sets Y n is preferably adjusted 105-2, 205-2.
- the recording rows are arranged one after another in a row with respect to the individual observation quantities. The values y rt of the same observation quantity r at the same point in time are therefore arranged in the total matrix Y in each case in one column.
- the recorded data sets Y n are preferably divided into a conditioning section and an identification section based on a criterion 106, 206.
- the conditioning section serves the same purpose
- a particulate filter may be set to the same initial condition, for example by performing or simulating regeneration.
- data areas q can then be selected from the overall data record Y 108, 208, which are suitable for training the transformation model TM.
- the use of a transformation model TM is based on the knowledge that the respective realization x, a property X, or the device j, which have this property or properties X, precipitate in the respective values y rt of the observation quantities r during operation of a unit X with a configuration x n .
- Each value constellation y rt of the observation variables 1 ⁇ r ⁇ R of each configuration x n contains information and by comparing the value constellations y rt generated by the individual configuration x n , statements about the respective configuration x n of a unit X of a genus can be derived.
- a so-called feature selection method is preferably used according to the invention.
- This method is a machine learning approach only a subset of available features or characteristics are used for a learning algorithm.
- Examples of unsupervised feature selection methods include Infinite Selection, Laplacian Score, Local Learning-Based Clustering, Multi-classe / Cluster Data, Regularized Discriminative Selection.
- Examples of Supervised Feature Selection methods include the ReliefF method, the Eigenvector Centrality method, Concave Minimization, Infinite Latent Selection, Infinite Selection, and Robust Selection.
- data regions q are selected 108, 208 which correspond to so-called features of the ReliefF method.
- This selection 108, 208 preferably takes place in four working steps: First, individual instances, which in the present case correspond to rows of the total matrix Y, that is to say of the total data set across all configurations x n and observation variables r, are plotted as data points in a multidimensional data space.
- Standardized values x f which correspond to the respective value constellation of the respective instance, as shown in the uppermost box of FIG. 3, serve as the ordinate of the data space. These are preferably arranged in two groups with respect to a nominal value of all occurrences x f. In the present case, this nominal value is zero.
- the ReliefF method shown were used as a feature selection method for a temperature sensor as unit j, the data points above the nominal value in the graph would correspond to positive temperature deviation data points in the temperature measurement by the sensor j, the values below the nominal deviations of Temperature with a lower value than the nominal value.
- an instance is randomly selected.
- a third step by means of a calculation of the Euclidean distance, the respective next data points or instances for the selected instance are determined in each class of data points.
- the next data point of an instance in the same class, that is with the same sign, is called Near Hit y NH .
- the next data point or instance located in the other class is called Near Miss y NM .
- the selected instance is named y n .
- the process described is repeated several times. The process described is preferably repeated until the weight order of the individual features no longer changes.
- the individual features which correspond to columns of the total matrix Y of the overall data set, are weighted. In doing so, important features are generally expected to have a greater Euclidean distance to the near miss instance y NM than to the near hit instance y NH .
- the assignment of the weighting and the joining in a whole Y, the selected data areas for the device j or the property X corresponds to a fourth step of the selection process 108, 208 and is shown in the lower right box in FIG.
- a matrix Y of the selected data areas, the most important features are sorted according to their respective weighting W, j from left to right.
- the selected data areas q and the corresponding correlating characteristics are output via a data interface 24 and read into an artificial neural network KNN 209.
- the selected data areas q preferably contain information about the values of the observed variables r and about the associated time t of the defined driving cycle B.
- the artificial neural network KNN recognizes patterns between the selected data areas corresponding features q and the occurrences x nj , so that on the basis of the artificial neural network KNN an assignment rule between observation magnitudes r and constellations of observation variables r and the properties X, or their give different values x nj .
- a transformation model TM is generated, which is preferably output 210.
- the transformation model TM is preferably output as a data sequence.
- a system 10 for analyzing at least one device j of a unit x to be tested x of a type of units X is illustrated on the basis of the functional block diagram of FIGS. 2 and 3.
- the selected data areas q are not only output to a compensation calculation which forms the basis for a transformation model, but the compensation calculation is based on the selected data areas q and the correlating Characteristic x nj also performed 109.
- the transformation model TM formed as a result of the compensation calculation has an assignment rule between observation variables r and one or a plurality of properties X, at least one device j.
- the transformation model TM preferably contains an assignment rule between the observation variables r and properties X of all devices j.
- the transformation model TM is finally also used in the method 100 for analysis 11 1.
- Applying in the sense of the invention means providing or outputting the transformation model TM for further data processing and / or determining an expression x nj of at least one device j by means of the Transformation model TM, as shown in Fig. 3.
- the transformation model TM is preferably provided as a data sequence.
- the transformation model TM is preferably set up to output a characteristic of the at least one device j to be analyzed 112 on the basis of at least one data set Y x recorded on the unit under test x in the operating cycle B.
- values x.sub.j are determined on the basis of the method 100 for analysis, different method steps, which are correspondingly also part of the method 300 for analyzing at least one device j, can additionally be carried out for this purpose.
- FIG. 4 is a graph of engine speed over time of an engine x, showing the original recorded data and the moving average during a drive cycle B.
- FIG. 4 is a graph of engine speed over time of an engine x, showing the original recorded data and the moving average during a drive cycle B.
- the moving average can be calculated using the formula from FIG. 5a.
- the value of the respective moving average is calculated for each individual value of the total matrix Y or of the total data set 107-1, 207-1, which leads to a total matrix Y 1 ⁇ of the moving average.
- a feature selection method as described in FIG. 3 can also be carried out.
- additional weighted features q result as a result.
- the data areas q selected in this way can be combined in a totality of selected moving average data areas for respective devices j and properties X, preferably as matrix Y j MA .
- the central difference quotient can be calculated as by a formula from FIG. 7a.
- the central difference quotient can accordingly be calculated for all values of the overall data matrix Y of the overall data record 107-2, 207-2, whereby an overall matrix of the central difference quotients is obtained. Also on the basis of this total matrix Y DQ , a feature selection method as described with reference to FIG. 3 can be carried out.
- the data regions q selected in this way can be combined in a totality of selected data regions of the difference quotient for respective devices j or properties X, preferably as matrix Y j DQ .
- FIG. 8 shows a diagram in which selected data areas q or features are represented as data points, the origin of the respective data areas q or features being plotted over their respective importance W, j according to the weighting of the feature selection method.
- the plurality of important selected data areas q are part of the totality of the selected average data areas Y ⁇ MA .
- the accuracy of a T2 sensor was examined, which measures the temperature in the intake manifold of an engine.
- the total number of simulated engines using the Feature Selection (ReliefF) method to train the TM transformation model was 481 engines.
- a ReliefF method as described with reference to FIG main features or data areas q from the total data set Y, the total difference data set Y DQ and the overall average data set Y MA .
- FIG. 9 shows in tabular form the 15 most important features W to Wi 5 j , ie the highest weighted features, wherein the device j is again the T2 sensor.
- observation variables designated by "T2" and "TC speed” in FIG. 9 according to the respective engine control unit channels belonging to the observation variables r, in which the most important selected data regions q or features are identified, are the two observation variables of the temperature T2 measured by the T2 sensor, as well as the turbo speed TC Speed. In the other 34 observation variables, no data areas are selected.
- the respective point in time in the defined driving cycle B is indicated, which belongs to the respectively most important selected data areas q. From this column it can be seen that the temperature values of the observation quantity T2 in the first 13 seconds of the defined drive cycle B were identified as important features and the turbocharger speed values of the observation variable TC Speed after 389 and 390 seconds of the defined drive cycle B were important features.
- FIG. 10 shows the beginning of a diagram on which the course of a simulated T2 temperature at the beginning of the defined driving cycle B is plotted for three different motors 1, 2, 3, that is to say the observation quantity T2.
- Each of the motors is one of the ten retained motors, that is, one of those motors X, which were not used to form the transformation model TM.
- Fig. 11 shows a diagram corresponding to the diagram of Fig. 10, wherein now not the course of the observation size T2 of the three retained engines 1, 2, 3 is plotted, but the course of the turbocharger speed, so the observation size TC Speed.
- the selected features as well as the time points or data taken into account for the selected features are drawn on the three courses.
- the selected features are at about 390 seconds, as shown in FIG. 9. Since the selected feature is one of the selected average data areas Y ⁇ MA , the bias area as in FIG. 10 is the moving average ⁇ 30
- the data taken into account in this case is from about 360 seconds to 420 seconds of the defined driving cycle B.
- Fig. 12 shows a diagram of the prediction accuracy with respect to the observation quantity T2, that is, the temperature measured by the T2 sensor for the ten retained motors, that is, those motors which were not used to form the transformation model TM.
- FIG. 13 shows a diagram of the engine speed over the time t in seconds for the defined drive cycle B, which also includes the feature selection method according to FIG. 8 and 9, as well as a diagram for the course of the T2 temperature and the turbocharger speed according to FIGS. 10 and 11.
- the influencing region which was identified for the observation variable T2-temperature in FIG. 10 is shown as a thick box in the diagram according to FIG.
- This period is an identification maneuver or an identification driving state.
- Feature Selection method which was carried out with respect to the measurement accuracy of the temperature sensor T2 and the results of which are shown in Figures 8 and 9, it is known that a large part of the important for the modeling of the transformation model TM characteristics or Data areas q lie in this identification maneuver.
- a defined driving cycle B for a vehicle 1 can be predefined.
- the operation of units X of the genus, each having a different configuration x n, can be simulated and the transformation model TM can be formed on the basis of the overall data record Y obtained in this way by means of a feature selection method.
- the unit x to be tested must also be operated as accurately as possible in the same predetermined defined operating cycle B in the ferry mode or on the test bench.
- the transformation model TM or the inherent assignment rule validity can be provided.
- This procedure may make sense in particular if the unit to be tested or the engine to be tested x is to be checked or analyzed in a so-called workshop maneuver, for example standstill regeneration, warm-up at standstill, NEDC cycle, etc.
- a fundamentally different procedure is based on a driving cycle B actually defined by a test unit X or an engine.
- a simulation 103, 203 of a plurality of units X of the same kind as the unit under test x is performed after operating 101, 101, 301 of the unit under test x in the defined operation cycle B. Also with the recorded data records Y n is by means of a feature
- Selection method trains a transformation model TM.
- This transformation model TM is in turn applied to a record Y x of the unit under test x, which was also recorded during operation of the unit under test x 102-2, 304.
- Operation of the unit to be tested x can be carried out in the field or on the road, on a chassis dynamometer, powertrain test bench or engine test bench.
- the embodiments are merely examples that are not intended to limit the scope, applications, and construction in any way. Rather, the expert is given by the preceding description a guide for the implementation of at least one embodiment, with various changes, in particular with regard to the function and arrangement of the components described, can be made without departing from the scope, as it is apparent from the claims and gives these equivalent feature combinations.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computer Hardware Design (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Computational Mathematics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Automation & Control Theory (AREA)
- Mathematical Optimization (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Analysis (AREA)
- Medical Informatics (AREA)
- Pure & Applied Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Combined Controls Of Internal Combustion Engines (AREA)
- Testing Of Engines (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE102018201933.9A DE102018201933A1 (de) | 2018-02-07 | 2018-02-07 | Verfahren und System zur Analyse wenigstens einer Einrichtung einer Einheit, welche eine Mehrzahl an verschiedenen Einrichtungen aufweist |
PCT/AT2019/060046 WO2019153026A1 (de) | 2018-02-07 | 2019-02-07 | Verfahren und system zur analyse wenigstens einer einrichtung einer einheit, welche eine mehrzahl an verschiedenen einrichtungen aufweist |
Publications (1)
Publication Number | Publication Date |
---|---|
EP3750089A1 true EP3750089A1 (de) | 2020-12-16 |
Family
ID=65686650
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
EP19709362.8A Pending EP3750089A1 (de) | 2018-02-07 | 2019-02-07 | Verfahren und system zur analyse wenigstens einer einrichtung einer einheit, welche eine mehrzahl an verschiedenen einrichtungen aufweist |
Country Status (3)
Country | Link |
---|---|
EP (1) | EP3750089A1 (de) |
DE (1) | DE102018201933A1 (de) |
WO (1) | WO2019153026A1 (de) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102018222519A1 (de) * | 2018-12-20 | 2020-06-25 | Zf Friedrichshafen Ag | Verfahren zum Bestimmen von Bauteil - oder Baugruppentoleranzen |
DE102019128655B4 (de) | 2019-10-23 | 2021-11-25 | Technische Universität Ilmenau | Verfahren zur Bereitstellung einer rechnergestützten Steuerung für ein technisches System |
AT523093A1 (de) * | 2019-11-12 | 2021-05-15 | Avl List Gmbh | Verfahren und System zum Analysieren und/oder Optimieren einer Konfiguration einer Fahrzeuggattung |
JP2023510404A (ja) * | 2020-01-16 | 2023-03-13 | インベンテイオ・アクテイエンゲゼルシヤフト | 人員輸送設備の部品のデジタルドキュメンテーション及びシミュレーションのための方法 |
AT525949B1 (de) | 2022-02-22 | 2024-05-15 | Avl List Gmbh | Verfahren zum Bestimmen einer Konzentration eines Gases in einem Gaspfad einer Brennkraftmaschine |
DE102022134771A1 (de) | 2022-12-23 | 2024-07-04 | Rolls-Royce Solutions GmbH | Verfahren und Steuereinrichtung zum Betreiben einer technischen Einrichtung sowie technische Einrichtung mit der Steuereinrichtung |
Family Cites Families (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7635043B2 (en) * | 1991-07-09 | 2009-12-22 | Automotive Technologies International, Inc. | Crash sensor arrangement for controlling deployment of an occupant restraint device |
US7549327B2 (en) * | 2001-02-16 | 2009-06-23 | Automotive Technologies International, Inc. | Tire-mounted energy generator and monitor |
US7499777B2 (en) * | 2005-04-08 | 2009-03-03 | Caterpillar Inc. | Diagnostic and prognostic method and system |
US7328625B2 (en) * | 2005-04-28 | 2008-02-12 | Caterpillar Inc. | Systems and methods for determining fatigue life |
US7483774B2 (en) * | 2006-12-21 | 2009-01-27 | Caterpillar Inc. | Method and system for intelligent maintenance |
JP6323121B2 (ja) | 2014-03-31 | 2018-05-16 | 株式会社デンソー | 未知データ分析装置 |
US20170249788A1 (en) * | 2016-01-13 | 2017-08-31 | Donald Remboski | Accurate application approval |
-
2018
- 2018-02-07 DE DE102018201933.9A patent/DE102018201933A1/de active Pending
-
2019
- 2019-02-07 EP EP19709362.8A patent/EP3750089A1/de active Pending
- 2019-02-07 WO PCT/AT2019/060046 patent/WO2019153026A1/de unknown
Also Published As
Publication number | Publication date |
---|---|
DE102018201933A1 (de) | 2019-08-08 |
WO2019153026A1 (de) | 2019-08-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3750089A1 (de) | Verfahren und system zur analyse wenigstens einer einrichtung einer einheit, welche eine mehrzahl an verschiedenen einrichtungen aufweist | |
AT518850B1 (de) | Verfahren zur simulationsbasierten Analyse eines Kraftfahrzeugs | |
DE102020103768A1 (de) | Überwachung und Diagnose von Fahrzeugsystemproblemen mit Maschinenlern-Klassifikatoren | |
DE112008001654T5 (de) | System und Verfahren für virtuelle Sensoren | |
EP3721199B1 (de) | Prüfstand und verfahren zur durchführung eines prüfversuchs | |
AT520827B1 (de) | Verfahren zum Bestimmen eines Fahrzeugparameters eines Fahrzeugdatensatzes eines Fahrzeugs und Verwendung des Fahrzeugparameters an einem Prüfstand | |
DE102015008754A1 (de) | Zustandsüberwachung eines Stellantriebs in einem Fluggerät | |
WO2013131836A2 (de) | Verfahren zur emissionsoptimierung von verbrennungskraftmaschinen | |
DE102017105469A1 (de) | Hardware-in-the-loop System und Verfahren zur Bedatung einer Steuerung eines Verbrennungsmotors | |
DE102005020821A1 (de) | Verfahren zur vereinfachten Echtzeitdiagnose unter Verwendung adaptiver Modellierung | |
DE102020106880A1 (de) | Verfahren und Vorrichtung zum Kalibrieren eines Steuer- und Regelvorrichtungssystems zur Steuer- und Regelung eines Betriebszustandes | |
WO2006056355A2 (de) | Diagnose- und servicesystem für ein kraftfahrzeug | |
DE102016125538A1 (de) | Verfahren zum Verifizieren von Aktuator-Steuerungsdaten | |
DE102008057494A1 (de) | Verfahren zum Betrieb eines Verbrennungsmotors | |
WO2018178196A1 (de) | Verfahren zur bestimmung einer schädigungsmassunsicherheit eines kraftfahrzeugs | |
DE102017106943A1 (de) | Verfahren und Anordnung zur Simulation von Fahrversuchen | |
DE102006045785A1 (de) | Verfahren zur Selbstdiagnose von Versuchsanordnungen sowie Versuchsanordnung, insbesondere Prüfstand | |
WO2018177526A1 (de) | Robustheitsanalyse bei fahrzeugen | |
EP1878900A2 (de) | Verfahren zur Diagnose von fehlerhaften Betriebszuständen einer Brennkraftmaschine | |
DE102022210276A1 (de) | Verfahren zur Erkennung von Defekten in einem System | |
DE102017111455A1 (de) | Verfahren zur Prüfung einer durch ein Steuergerät gesteuerten Maschine | |
DE102019131722B3 (de) | Verfahren zur Steuerung oder Regelung einer Klimatisierung eines Fahrzeugs | |
EP3132322B1 (de) | Verfahren zur diagnose eines kraftfahrzeugsystems, diagnosegerät für ein kraftfahrzeugsystem, steuergerät für ein kraftfahrzeugsystem und kraftfahrzeug | |
DE102005014308A1 (de) | Verfahren und Anordnung zur Nutzerführung bei der Diagnose komplexer Systeme sowie ein entsprechendes Computerprogramm und ein entsprechendes computerlesbares Speichermedium | |
DE102018212801A1 (de) | Diagnose komplexer Systeme |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
17P | Request for examination filed |
Effective date: 20200904 |
|
AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
AX | Request for extension of the european patent |
Extension state: BA ME |
|
DAV | Request for validation of the european patent (deleted) | ||
DAX | Request for extension of the european patent (deleted) | ||
P01 | Opt-out of the competence of the unified patent court (upc) registered |
Effective date: 20230503 |
|
STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |
|
17Q | First examination report despatched |
Effective date: 20230627 |