DE102021201166A1 - Method and device for providing operating data from a large number of devices - Google Patents

Abstract

The invention relates to a computer-implemented method for reducing a data volume of performance variable profiles of a multiplicity of devices (41), comprising:
- Providing (S1) a set of data sets from operating variable profiles;
- Merging (S4, S5) of similar data sets from the set of data sets to form a resulting data set and assigning a frequency value to the resulting data set;
- Selecting (S6) a predetermined number of the most frequent data records according to the frequency values in order to obtain a reduced set of data records;
- Using (S7) the reduced set of data records for further processing.

Description

technical field

The invention relates to methods for reducing an amount of data when providing operating data from a large number of devices for use in model creation, state determination or state prediction.

Technical background

To support the operation of technical devices such. B. energy storage systems for electrically powered vehicles, operating variable data from a variety of devices are often transmitted to a central unit, a so-called cloud. The data collected in the central unit enables an accurate modeling of the behavior of the device in question and a state determination and state prediction for the devices.

In the case of vehicle batteries as devices in motor vehicles, such models can support battery management systems in realizing the requirements for functional safety and, for example, in specifying battery-specific parameters such as the state of charge (SOC: State of Charge) or the aging state (SoH: State of Health). Although battery management systems perform a large number of functions and calculate the state of charge and state of health independently, the operating variable data collected in the central unit make it possible to specify the models used and/or to reduce the computing effort in the respective battery management system.

Collecting the performance data from a variety of devices, i. H. of the operating variable curves for the vehicle batteries of a large number of vehicles in a vehicle fleet, enable a detailed analysis of the behavior of the vehicle batteries and enable additional statements to be made about the service life of the vehicle battery or about the possibilities of battery-saving charging phases depending on the individual use of the vehicle batteries in the vehicle fleet under consideration. As a result, a significant gain in information can be used, which, for example, enables a general extension of the service life of the vehicle batteries in all vehicles.

The operating size data (fleet data) collected in the central unit is usually recorded in a very small time frame of less than 1 second, so that the amount of data generated becomes very high over time. As the amount of data increases, there is also a challenge in executing the corresponding algorithms, in particular algorithms for determining and predicting the state and/or for model creation, in a reasonable time, since the computing and memory requirements can often behave exponentially in relation to the amount of data. Depending on the complexity of the algorithms, the storage space for storage and/or the computing time required for evaluation can exceed the available limits in the event of strong growth in the data volume of the operating variable data.

Disclosure of Invention

According to the invention, a method for reducing a data volume of operating variable data from a large number of devices according to claim 1 and a central processing unit and an overall system with a central processing unit and a large number of devices communicating with it are provided according to the independent claims.

Further developments are specified in the dependent claims.

• - Bereitstellen einer Menge von Datensätzen aus Betriebsgrößenverläufen;
• - Zusammenfassen von ähnlichen bereitgestellten Datensätzen aus der Menge von Datensätzen zu einem resultierenden Datensatz und Zuordnung eines Häufigkeitswerts zu dem resultierenden Datensatz;
• - Auswählen einer vorgegebenen Anzahl der entsprechend den Häufigkeitswerten häufigsten Datensätze, um eine reduzierte Menge an Datensätzen zu erhalten; und
• - Verwenden der reduzierten Menge an Datensätzen für eine Weiterverarbeitung.

According to a first aspect, a method is provided for reducing a data volume of performance variable profiles of a multiplicity of devices, comprising:

• - Provision of a set of data sets from company variable profiles;
• - Combining similar provided data sets from the set of data sets to form a resulting data set and assigning a frequency value to the resulting data set;
• - selecting a predetermined number of the most frequent data records according to the frequency values in order to obtain a reduced set of data records; and
• - Use of the reduced amount of data sets for further processing.

In an overall system in which a large number of devices are in communication with a central unit, profiles of one or more operating variables of the devices can be recorded in the central unit. The performance variable profiles can then be used to characterize the behavior of the devices, such as electrical energy stores, in particular vehicle batteries, to model description models and/or to determine and/or predict states.

For example, for batteries in vehicles (as technical devices), the courses the state of charge can be analyzed during driving cycles, for example to track the aging of the vehicle batteries. The different aging behavior of the vehicle batteries depending on the use by the respective driver can be recognized in a suitable manner and taken into account for a prediction of aging or for the output of driving recommendations.

With a large number of recorded data sets from the progression of operating variables, the cumulative data set often includes redundant operating variable progressions that are identical or very similar to one another, so that many data sets are present in the cumulative data set that make no further contribution to the model creation, the status determination or the status prediction of individual ones devices can deliver.

In order to simplify the processing of the data sets from the large number of devices, it is proposed to reduce the amount of data in such a way that the loss of information remains as small as possible when reduced data is used for modeling or for determining the state or predicting the state of individual devices. In this regard, it is provided according to the invention to combine data sets with performance variable profiles that are similar to one another to form the resulting data sets and to use only a number of the resulting data sets with the greatest frequencies for further evaluation. This enables a significant reduction in the amount of data on the course of operating variables from the large number of devices and a minimization of the loss of information in the reduced amount of data. As a result, the creation of the model or the determination of the state, in particular with regard to the required computing time and the required storage space, can be carried out in a more resource-saving manner.

Provision can be made for similar data sets to be determined as a function of a specified tolerance amount, with data sets being similar to one another if their corresponding operating variables do not deviate from one another by more than the specified tolerance amount.

Furthermore, the similar data sets can be combined by averaging or median formation of the corresponding values of the performance variable profiles of the data sets to be combined. The values that correspond to each other correspond to the values that correspond to each other in terms of progression of the course.

According to one embodiment, the data records can be received from a plurality of devices in a central unit (cloud), using the reduced set of data sets in the central unit or after retransmitting the reduced set of data sets to one or more of the plurality of devices in the respective device is done.

The extent of the data reduction can be flexibly adjusted by the amount of tolerance and the number of selected data sets with the highest frequency values, so that the reduced amount of data can be matched to the computing capacity available for model creation or state determination or state prediction.

• - Bereitstellen einer Menge von Datensätzen aus Betriebsgrößenverläufen;
• - Zusammenfassen von ähnlichen Datensätzen aus der Menge von Datensätzen zu einem resultierenden Datensatz und Zuordnung eines Häufigkeitswerts zu dem resultierenden Datensatz;
• - Auswählen einer vorgegebenen Anzahl der entsprechend den Häufigkeitswerten häufigsten Datensätze, um eine reduzierte Menge an Datensätzen zu erhalten;
• - Übermitteln einer Information an eines oder mehrere der Geräte, die von der reduzierten Menge an Datensätzen abhängt.

According to a further aspect, a central unit is provided which is in communication with a plurality of devices, the central unit being designed for:

• - Provision of a set of data sets from company variable profiles;
• - Combining similar data sets from the set of data sets to form a resulting data set and assigning a frequency value to the resulting data set;
• - selecting a predetermined number of the most frequent data records according to the frequency values in order to obtain a reduced set of data records;
• - Transmission of information to one or more of the devices that depends on the reduced amount of data sets.

character list

• 1 eine schematische Darstellung eines Systems mit einer Zentraleinheit und einer Vielzahl von Fahrzeugen einer Fahrzeugflotte;
• 2 ein Flussdiagramm zur Veranschaulichung eines Verfahrens zur Datenreduktion zur Reduzierung einer Datenmenge aus Datensätzen mit jeweils einem Betriebsgrößenverlauf für einen Fahrzyklus eines Fahrzeugs aus der Vielzahl von Fahrzeugen; und
• 3a bis 3e Tabellen mit mehreren Datensätzen, die jeweils einen realen oder gemittelten Betriebsgrößenverlauf darstellen, zur Veranschaulichung einer Datenmengenreduktion anhand eines Verlaufs eines fahrstreckenabhängigen Ladungszustands einer Fahrzeugbatterie.

Embodiments are explained in more detail below with reference to the accompanying drawings. Show it:

• 1 a schematic representation of a system with a central unit and a plurality of vehicles of a vehicle fleet;
• 2 a flowchart to illustrate a method for data reduction for reducing a data volume from data sets, each with an operating variable profile for a driving cycle of a vehicle from the plurality of vehicles; and
• 3a until 3e Tables with several data sets, each of which represents a real or averaged operating variable curve, to illustrate a reduction in the amount of data using a curve of a vehicle battery's state of charge as a function of the distance travelled.

Description of Embodiments

The method according to the invention is described below using vehicle batteries as an example of technical devices. The vehicle batteries are used in motor vehicles as traction batteries. The operation of the vehicle batteries or of a vehicle connected to it is to be improved and/or optimized by collecting operating variables of the vehicle batteries in a central unit and by evaluating them based on a large number of vehicles. For example, based on the collected operating variables of the vehicle batteries from the vehicle fleet, an aging state model can be continuously updated or retrained in the central unit. The aging state model can then be used to determine and predict aging states of the individual vehicle batteries.

The above example is representative of a large number of stationary or mobile, in particular battery-operated, devices with a mains-independent energy supply, such as vehicles (electric vehicles, pedelecs, etc.), systems, machine tools, household appliances, IOT devices, building energy supplies, aircraft, in particular drones, autonomous robots and Consumer electronics devices, in particular mobile phones and the like, which are connected to a central unit (cloud) via a corresponding communication connection (e.g. LAN, Internet).

1 shows an example of a system 1 for collecting fleet data in a central unit 2 to create an aging model. The aging state model is used to determine the aging state of an electrical energy store in a motor vehicle. 1 shows a vehicle fleet 3 with several motor vehicles 4.

One of the motor vehicles 4 is in 1 shown in more detail. The motor vehicles 4 each have a vehicle battery 41 as a rechargeable electrical energy store, an electric drive motor 42 and a control unit 43 . The control unit 43 is connected to a communication module 44 which is suitable for transmitting data between the respective motor vehicle 4 and a central unit (a so-called cloud).

A data memory 21 is provided in the central unit 2 . The data records stored in the data memory 21 are processed in a processor unit 22 .

The motor vehicles 4 send the operating variables F to the central unit 2, which indicate at least variables on which the aging state of the vehicle battery 41 depends. Of course, any other type of history data for operating variables can also be transmitted, depending on which function is to be executed in the central unit 2 based on the operating variable trends.

In the case of a vehicle battery, the operating variables F can indicate an instantaneous battery current, an instantaneous battery voltage, an instantaneous battery temperature and an instantaneous state of charge (SOC: State of Charge), both at pack, module and/or cell level. The operating variables F are recorded in a fast time frame from 0.1 Hz to 100 Hz and can be transmitted regularly to the central unit 2 in uncompressed and/or compressed form. For example, the time series of the performance variables can be transmitted to the central unit 2 in blocks at intervals of 10 minutes to several hours. The course of operating variables can be assigned to operating cycles of the individual vehicles as data sets.

A modeling function can be implemented in the central unit 2, which determines or forms or trains a state of health model from the data sets of performance variable profiles. The aging state model can e.g. B. be designed to specify an aging condition from an empirical evaluation of the operating variables F. This state of health model can contain a state of health value e.g. B. derived from charge state curves over a distance traveled in a current driving cycle.

Due to the conventional functionality of battery management systems, operating variable profiles of vehicle batteries can be recorded in a simple manner with a high temporal resolution. The performance variables, in particular battery current, battery temperature, battery voltage, state of charge and the like, can now be forwarded to a central unit via the communication link or, for example, via charging stations or in workshops. The operating variable data contain, for example, the course of the state of charge over a distance traveled in connection with information about an aging state and/or the driving behavior of the driver.

Due to the high temporal resolution, the amount of data increases sharply over time (with an increasing number of driving cycles) when there are a large number of vehicles in the vehicle fleet, so that the effort for storing and evaluating the data sets increases accordingly. Since the effort for the evaluation of the Data, for example for model creation or state determination or prediction, is often quadratically, exponentially or progressively increased in some other way, the limits of the amount of data that can be processed in a reasonable amount of time are quickly reached.

Using the flowchart of 2 a method is described with which data sets from company size trends are collected and reduced. The data records each include profiles of a state of charge of a vehicle battery of a vehicle from vehicles in a vehicle fleet as a function of a route driven. These operating variable profiles can be used, for example, to determine the state of health of the individual vehicle batteries, in particular to create a state of health model for the vehicle batteries of the vehicles or to predict the state of health of the individual vehicles.

For this purpose, in step S1, charge state profiles are recorded as operating variable profiles in the vehicles and transmitted to the central unit 2 at regular or predetermined times. In this example, the performance variable curves include charge states, each of which is assigned information about the distance covered during the current driving cycle. This historical data for a driving cycle is summarized in a data set. In the central unit 2, the data sets with the operating variable data or the charge state profiles assigned to the routes are collected and added to previously transmitted data sets.

In step S2, the recorded data sets are normalized, i. H. each of the performance variable curves is brought into a series of values in which values reached one after the other indicate specific charge states in a specific grid of routes driven.

As for example in 3a shown, the data records of the operating variable curves indicate the state of charge after a predetermined route. In the example of 3a the states of charge at route points are specified after each 10 km driven as data sets, from which the method described herein is based. The resolution of the operating variable curves can be selected, and finer increments between the route points, such as between 1 km and 9 km, or larger increments, such as between 11 and 30 km, can also be used.

The state of charge is given as a percentage and indicates a proportion of the battery capacity that is still available, compared to a fully charged vehicle battery 41. The operating variable curves of the 3a three vehicles are specified for example, with the first vehicle Fz1 operating variable curves Fz1.1, Fz1.2, Fz1.3, Fz1.4 for four trips, the second vehicle Fz2 operating variable curves Fz2.1, Fz2.2, Fz2.3 for three Trips and the third vehicle Fz3 operating variable profiles Fz3.1, Fz3.2, Fz3.3 provides for three trips.

In a subsequent step S3, the recorded operating variable profiles are sorted according to the length of the route of the respective driving cycle and according to the state of charge at the start of the driving cycle. this is in 3b shown.

In step S4, it is analyzed which driving cycles of the same length show a similar progression. The similarity of two operating variable profiles is given when the profiles of the charge states of several of the data sets at the individual support points, ie at each route point, in terms of amount by less than a predetermined tolerance distance, such. B. between 1-5%, preferably 2%, deviate from each other. Such datasets are, for example, in 3c marked by the corresponding reinforced border.

In step S5, the similar performance variable profiles are combined to form an averaged performance variable profile and a frequency value is assigned to the resulting data set, which indicates the number of data records taken into account in the averaged performance variable profile. 3d shows the corresponding table of the data sets or the course of the company variables including the associated frequency values. The example given is for illustration only. In the case of large amounts of data from the course of recorded company variables, histograms are formed with several thousand similar data sets.

In a next step S6, a number N of data records is selected which have the highest frequency values. In the embodiment shown 3e selects the first three records (heavy border) that have the highest frequency values. In the example shown, 10 company variable curves or data sets were initially assumed, which were reduced to three data sets using the method presented. These resulting datasets represent 5 of the original datasets with only a small loss of information.

The creation of the model or the determination or prediction of the state can now be carried out in step S7 based on the reduced data sets of the selected performance variable curves. For this purpose, the model creation or the status determination can either be carried out in the central unit 2 or the resulting data sets can be transmitted to one or more of the vehicles 4 in which the corresponding computing unit 43 can take over the processing of the data sets.

The degree of data reduction or the quantity of data sets to be further processed in step S7 can be determined by selecting the tolerance amount for the determination of similar performance variable profiles/data sets or the number of the most frequent data sets to be selected.

Claims (10)

Computer-implemented method for reducing a data volume of performance variable profiles of a multiplicity of devices (41), comprising: - Providing (S1) a set of data sets from operating variable profiles; - Merging (S4, S5) of similar data sets from the set of data sets to form a resulting data set and assigning a frequency value to the resulting data set; - Selecting (S6) a predetermined number of the most frequent data records according to the frequency values in order to obtain a reduced set of data records; - Using (S7) the reduced set of data records for further processing. procedure after claim 1 , wherein the reduced set of data sets is used for model creation, state determination and/or state prediction. procedure after claim 1 or 2 , wherein similar data sets are determined as a function of a specified tolerance amount, wherein data sets are similar to one another if their corresponding operating variables do not deviate from one another by more than the specified tolerance amount. Procedure according to one of Claims 1 until 3 , wherein the combining of the similar data sets is carried out by averaging or median formation of the respectively mutually corresponding values of the performance variable curves of the data sets to be combined. Procedure according to one of Claims 1 until 4 , wherein the data records are received from a plurality of devices in a central unit, the reduced set of data sets being used in the central unit or after the reduced set of data sets has been retransmitted to one or more of the plurality of devices in the respective device. Procedure according to one of Claims 1 until 5 , wherein the devices correspond to energy storage devices, in particular batteries, in motor vehicles, with the performance variable profiles specifying a state of charge of the energy storage device over a distance traveled in a driving cycle. Central processing unit (3) which is in communication with a large number of devices (41), the central processing unit (3) being designed for: - Provision of a set of data sets from company variable profiles; - Combining similar data sets from the set of data sets to form a resulting data set and assigning a frequency value to the resulting data set; - selecting a predetermined number of the most frequent data records according to the frequency values in order to obtain a reduced set of data records; - Transmission of information to one or more of the devices (41) which depends on the reduced amount of data records. use of the central unit claim 7 that are connected to a large number of stationary or mobile, in particular battery-operated, devices with a mains-independent energy supply, such as vehicles, systems, machine tools, household appliances, IOT devices, building energy supplies, aircraft, autonomous robots and consumer electronics devices, via a corresponding communication link, with the devices (41) the reduced set of data sets for model creation, status determination and/or status prediction are provided. Computer program product comprising instructions which, when the program is executed by a computer, cause the latter to carry out the steps of the method according to one of Claims 1 until 6 to execute. A machine-readable storage medium comprising instructions which, when executed by a computer, cause the computer to perform the steps of the method of any one of Claims 1 until 6 to execute.

Images