Method for detecting an alternative use of a vehicle battery

Abstract

The invention relates to a method for detecting the alternative usability of a battery (B) of a vehicle (F), wherein at least one battery characteristic (BK) is recorded and monitored by the vehicle (F), wherein the occurrence of at least one trigger condition (T) for replacing the battery (B) is checked, wherein a decision algorithm (EA) comprising artificial intelligence, in particular a neural network, determines, based on the at least one battery characteristic (BK), the trigger condition (T) and at least one dynamic threshold (DSW) from an external source, whether the battery (B) is qualified for a second-life application and, in this case, generates a purchase offer (KA) for the battery (B) and submits it to a customer (K), and proposes a possible allocation of the battery (B) for a specific second-life application.

Description

[0001] The invention relates to methods for detecting an alternative use of a vehicle battery according to the preamble of claim 1.

[0002] The battery of a vehicle, such as an electric vehicle, that is older and / or has high mileage may have reduced performance. This could render the battery unsuitable for use as a vehicle battery. However, the battery could potentially be used in various second-life applications. For example, several batteries that have reached the end of their service life could be connected to form a larger energy storage system. Currently, however, it is not possible to automatically detect whether a battery is suitable for a second-life application.

[0003] DE 20 2021 105 689 U1 describes an intelligent system for improving the condition of electric vehicle batteries using machine learning and artificial intelligence, wherein the intelligent system comprises: an electric vehicle battery pack, consisting of: several cell modules arranged in series and parallel; a battery monitoring device typically used to monitor the voltage of each battery cell in the system, the temperature of various points in the battery packs, and other vehicle conditions; a battery management system, the battery management system comprising: Monitoring components are located near the battery cells themselves; one or more power conversion stages are selected according to the vehicle's needs; and an intelligent control unit is placed at strategic points in the architecture to manage various aspects of the vehicle's energy subsystem, with data being reported to a battery cell management controller and, depending on the system's complexity, to higher-level processing elements. the intelligent control unit is used to control the efficient charging and discharging of the battery, as this avoids thermal outliers or other conditions that would reduce either the capacity or the lifespan of the battery, the control unit using a machine learning algorithm that is used to manage and improve the condition of electric vehicle batteries.

[0004] Furthermore, DE 10 2020 201 697 discloses a B3 method for categorizing a battery with regard to its suitability for further handling, such a battery, a battery recycling system, and an associated motor vehicle. DE 10 2016 220 860 A1 describes a method, a device, and a system for evaluating a traction battery. EP 4 123 321 A1 describes a high-precision coulometry measurement for used batteries to obtain an estimate of their residual value, in particular their suitability for a second-life application in a stationary energy storage system.

[0005] The invention is based on the objective of providing a novel method for detecting an alternative use of a vehicle battery.

[0006] The problem is solved according to the invention by a method for detecting an alternative use of a vehicle battery with the features of claim 1.

[0007] Advantageous embodiments of the invention are the subject of the dependent claims.

[0008] A method for detecting the alternative use of a vehicle battery, for example, a traction battery of an electric vehicle, is proposed. According to the invention, at least one battery characteristic is detected and monitored by the vehicle, in particular by at least one control unit of the vehicle, and the occurrence of at least one trigger condition for battery replacement is checked. A decision algorithm, which includes artificial intelligence, in particular a neural network, uses the at least one battery characteristic, the trigger condition, and at least one dynamic threshold from an external source to determine whether the battery is qualified for a second-life application. If so, a purchase offer for the battery is generated and submitted to a customer, and a possible allocation of the battery for a specific second-life application is proposed.

[0009] In one embodiment, a battery characteristic value is recorded and monitored, including a state of health, a remaining capacity, a number of switching cycles of at least one contactor of the battery, an internal resistance, an age and / or a technical property of the battery, in particular a capacity and / or information on the cell chemistry of the battery.

[0010] In one embodiment, the trigger condition is checked to determine whether at least one threshold relating to the battery has been reached, exceeded or fallen below, for example a minimum age, a remaining capacity and / or a minimum number of switching cycles of a contactor of the battery, the presence of irreparable damage or a customer request for replacement.

[0011] In one embodiment, a market demand, market availability and / or market price for the battery and / or for at least one component of the battery and / or for at least one resource required to manufacture the battery and / or a type of possible and / or demanded second-life application for the battery is taken into account as a dynamic threshold.

[0012] In one embodiment, the decision algorithm takes into account certain requirements for battery characteristics, in particular the state of health, for some or all types of second-life applications.

[0013] In one embodiment, the decision algorithm takes into account the temporal evolution of the dynamic thresholds over a minimum period of time.

[0014] In one embodiment, the decision algorithm classifies the battery as qualified for a second-life application if: - the battery parameters have reached predetermined thresholds and the market demand and / or the market price for the battery exceeds a certain threshold, and / or - if at least one trigger condition for replacing the battery based on battery characteristics is met, and / or - if the customer requests a battery replacement.

[0015] In one embodiment, the decision algorithm is located in a backend that communicates with the vehicle wirelessly or via a wired connection.

[0016] In one embodiment, the decisions of the decision algorithm are monitored by a battery manager, which trains the artificial intelligence by confirming or rejecting the decision.

[0017] In one embodiment, the use of the battery as a buffer storage device or as a traction battery for another vehicle model, or the recycling of the battery, is considered as a type of second-life application.

[0018] The solution according to the invention combines the determination of a battery health status (SOH), resource allocation on demand, and the training of a neural network for process optimization.

[0019] The solution according to the invention makes it possible to identify batteries that need replacing, to automatically determine whether a battery is suitable for a second-life application, to increase efficiency in distribution, and to enable or simplify the purchase of old batteries.

[0020] By monitoring battery parameters (e.g., state of health, internal resistance, contactor switching cycles) throughout the vehicle's entire lifecycle, it can be determined when the battery is no longer optimal for driving. Vehicle parameters, current prices, the availability of relevant resources, and potential second-life applications are used in an intelligent system to decide whether the battery qualifies for a second-life application. If so, the customer can then be offered a purchase of the battery. The decision algorithm can be implemented, for example, using machine learning based on a neural network. This neural network can be trained with vehicle and market data, as well as information on potential second-life applications.

[0021] Using machine learning, real-time time series analyses are used to optimize the prediction of the development of the health status of batteries and their availability for various second-life applications based on different SOH parameters depending on the different battery types.

[0022] The matching algorithm is trained using a neural network that rewards prediction accuracy via a reward mechanism, provided the battery is successfully reused for the Second Life application. Thresholds are established as matching decisions by the initial decision and are iteratively and automatically adjusted based on subsequent decisions.

[0023] Here, SOH value tolerances are defined for various Second Life applications, within which a battery qualifies for a given Second Life application. These value tolerances are iteratively adjusted to the actual decisions through manual central interventions in the allocation system. The neural network is trained to approximate the matching preference as closely as possible with its suggested values. The reward function is adjusted accordingly.

[0024] Exemplary embodiments of the invention are explained in more detail below with reference to drawings.

[0025] This shows: Fig. 1. A schematic view of a method for detecting an alternative use of a vehicle battery, and Fig. 2 a schematic view of an architecture for carrying out the procedure.

[0026] Corresponding parts are marked with the same reference symbols in all figures.

[0027] Fig. Figure 1 is a schematic view of a method for detecting an alternative use of a battery B of a vehicle F.

[0028] Fig. Figure 2 is a schematic view of an architecture for carrying out the procedure.

[0029] Battery parameters BK are recorded and monitored by vehicle F, for example a state of health (SOH), a residual capacity, a number of switching cycles of at least one contactor of battery B, an internal resistance, an age and / or technical properties of battery B, for example the capacity and / or information on cell chemistry.

[0030] Furthermore, the occurrence of at least one trigger condition T is checked, for example, exceeding or falling below at least one threshold value relating to battery B, such as a minimum age of, for example, five years, a remaining capacity of, for example, 70%, and / or a minimum number of contactor switching cycles of, for example, 10,000. Alternatively or additionally, the trigger condition T for replacing battery B can be whether a battery replacement is necessary, for example, because battery B is irreparably damaged, has a remaining capacity of 50% or less, or because a customer K requests a replacement.

[0031] Furthermore, dynamic threshold values ​​(DSW) from external sources can be considered, for example, at least one key figure for quantifying a market situation regarding battery B, resource prices, key figures for resource availability, possible second-life applications for battery B, such as buffer storage (PS), recycling (RC), for another vehicle model (FM), or similar. The buffer storage (PS) can be used, for example, by an energy supplier, a solar power plant operator, or a private consumer. For some or all second-life applications, specific requirements for the battery key figures (BK), such as the state of health (SOH), may exist, which can be considered by a decision algorithm (EA).

[0032] The dynamic thresholds DSW, for example, are considered over a minimum period, such as one year, to assess the market situation for battery B.

[0033] A decision algorithm EA determines, based on the battery characteristics BK, the trigger condition T and the dynamic thresholds DSW, whether the battery B is qualified for a second-life application.

[0034] The decision algorithm EA can, for example, decide that battery B is qualified for a second-life application if: - the battery parameters BK have reached predetermined threshold values ​​and the market (the demand and / or the price for battery B) is currently high, for example, and / or - if the threshold values ​​of battery B indicate that a battery replacement is necessary, and / or - if there is a customer request to replace battery B.

[0035] If battery B qualifies for a second-life application, a purchase offer KA for battery B is generated and submitted to customer K. Furthermore, a possible allocation of battery B for a specific second-life application is proposed.

[0036] If none of the above criteria apply, then battery B is classified as not qualified for a second-life application.

[0037] The decision algorithm EA can consist of a neural network and be located in a backend or in the cloud, which, for example, communicates wirelessly with the vehicle F to receive the battery parameters BK and / or send a purchase offer KA. The decisions of the decision algorithm EA can be monitored, for example, by a battery manager BM, in particular a person, who trains the neural network by confirming or rejecting the decision. Reference symbol list B Battery BM Battery Manager BK Battery Characteristic Value DSW dynamic threshold EA Decision Algorithm F vehicle FM vehicle model K Customer KA purchase offer PS buffer storage RC Recycling T Trigger Condition

Claims

[1] Method for detecting an alternative use of a battery (B) of a vehicle (F), characterized by , that at least one battery characteristic value (BK) is recorded and monitored by the vehicle (F), furthermore checking the occurrence of at least one trigger condition (T) for replacing the battery (B), wherein a decision algorithm (EA) comprising artificial intelligence, in particular a neural network, determines, based on the at least one battery characteristic value (BK), the trigger condition (T) and at least one dynamic threshold (DSW) from an external source, whether the battery (B) is qualified for a second-life application and, in this case, generates a purchase offer (KA) for the battery (B) and submits it to a customer (K), and proposes a possible allocation of the battery (B) for a specific second-life application. [2] Method according to claim 1, characterized by, that the battery characteristic value (BK) includes a health status, a residual capacity, a number of switching cycles of at least one contactor of the battery (B), an internal resistance, an age and / or a technical characteristic of the battery (B), in particular a capacity and / or information on the cell chemistry of the battery (B), and is recorded and monitored. [3] Method according to claim 1 or 2, characterized by , that as a trigger condition (T) the reaching, exceeding or falling below at least one threshold value relating to the battery (B), for example a minimum age, a residual capacity and / or a minimum number of switching cycles of a contactor of the battery (B), the presence of irreparable damage or a customer request for replacement is checked. [4] Method according to any one of the preceding claims, characterized by, that a market demand, market availability and / or market price for the battery (B) and / or for at least one component of the battery (B) and / or for at least one resource required to manufacture the battery (B) and / or a type of possible and / or demanded second-life application for the battery (B) is taken into account as a dynamic threshold (DSW). [5] Method according to claim 4, characterized by , that for some or all types of second-life applications, certain requirements regarding battery characteristics (BK), in particular the state of health, are taken into account by the decision algorithm (EA). [6] Method according to any one of the preceding claims, characterized by , that a temporal progression of the dynamic thresholds (DSW) over a minimum period is taken into account by the decision algorithm (EA). [7] Method according to any one of the preceding claims, characterized by, that the decision algorithm (EA) classifies the battery (B) as qualified for a second-life application if: - the battery parameters (BK) have reached predetermined thresholds and the market demand and / or the market price for the battery (B) exceeds a certain threshold, and / or - if at least one trigger condition (T) for replacing the battery (B) with battery characteristics (BK) is met, and / or - if there is a customer request to replace the battery (B). [8] Method according to any one of the preceding claims, characterized by , that the decision algorithm (EA) is located in a backend which communicates wirelessly or via a wired connection with the vehicle (F). [9] Method according to any one of the preceding claims, characterized bythat the decisions of the decision algorithm (EA) are monitored by a battery manager (BM) which trains the artificial intelligence by confirming or rejecting the decision. [10] Method according to any one of the preceding claims, characterized by , that as a type of second-life application, the use of the battery (B) as a buffer storage (PS) or as a traction battery for another vehicle model (FM) or the recycling (RC) of the battery (B) is taken into consideration.

Images