Method and system for assessing the quality of a battery cell in a battery-cell-manufacturing process

Abstract

The present invention relates to a method for assessing the quality of a battery cell in a battery-cell-manufacturing process, the method comprising: passively measuring at least a voltage noise signal of at least a battery cell, processing the at least one voltage noise signal into a structured data set, analyzing the structured data set and extracting at least a characteristic descriptor from the structured data set, establishing at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell The present invention also relates to a corresponding system.

Description

[0001] Method and system for assessing the quality of a battery cell in a battery-cellmanufacturing process

[0002] Field

[0003] The present invention lies in the field of battery cell manufacturing and battery cell screening. More in particular, the present invention relates to a method and to a system for assessing the quality of a battery cell in a battery-cell-manufacturing process.

[0004] Background

[0005] The need for efficient and reliable battery-cell production has faced a substantial uplift in recent years. The market demand for Li-Ion batteries, Na-Ion batteries, x-Ion batteries or other rechargeable batteries, for example, has seen a substantial increase, at least and for the most part due to the current shift from conventional modes of mobility, including fossil fuel-based mobility, to electric mobility, or in short e-mobility.

[0006] The manufacturing of battery-cells typically includes three steps, namely the step of battery cell electrode preparation, the step of battery cell assembly, and the step of battery cell finishing. Modern battery cell manufacturing plants are for the most part automated and are adapted to manufacture and handle a number of battery cells of the order of tens to hundreds of million cells per year.

[0007] The present invention is directed at least in part to assessing the quality of a battery cell in terms of the "self-discharge" of the battery cell. It will be understood that the "selfdischarge", as referred to herein with regard to, e.g., a battery cell, can refer to the phenomenon wherein a battery cell, while not being actively discharged and / or connected to an external load, experiences a reduction in its state of charge over time. In other words, the battery-cell open circuit voltage (OCV) experiences a reduction over time. Said reduction can be attributed to internal mechanisms, including but not limited to, residual chemical reactions within the electrolyte, electrodes, and separator, as well as intrinsic leakage currents resulting from imperfections in the battery cell or, more generally, intrinsic properties and / or intrinsic activities of the battery cell. The rate of self-discharge can be influenced by factors, including but not limited to, the composition and quality of materials, the manufacturing process, the temperature at which the cell is stored, and the cell's age. It will be understood that a "self-discharge-like" metric or characteristic or effect, as referred to herein with regard to, e.g., a battery cell, can refer to a metric or characteristic or effect that can be obviously akin to, linked to, associated with or made dependent from the "self-discharge" of a battery cell.

[0008] The present invention is at least in part directed to the step of battery-cell finishing. The step of battery-cell finishing typically starts with filling the battery-cell with an electrolyte. The electrolyte can penetrate the electrodes of the battery-cell. The battery-cell is then generally subject to charging at progressively higher currents, with the aim of building up a robust yet thin solid electrolyte interphase (SEI) layer on the anode of the battery-cell. Following the step of building up an SEI, some transient self-discharge effects, such as continuous SEI growth and anode overhang equalization, can occur due, for example, to imperfect electrode alignment and lithium distribution. Said effects typically cannot be indicative of defects of the battery cell and are generally indicative of benign relaxation processes. However, different types of unwanted short-circuits can further increase the self-discharge of the battery cell: soft short-circuits reduce local resistance between the anode and cathode, while hard short-circuits create a highly conductive path that can lead to local separator damage when the path fuses. The two main causes of self-discharge, namely transient self-discharge effects and unwanted soft and / or hard short-circuits, can occur simultaneously and can be difficult to distinguish. To identify battery cells with selfdischarge not primarily due to benign relaxation processes, battery cells can usually undergo the step of battery-cell aging.

[0009] During the step of battery-cell aging, transient self-discharge effects can be allowed to stabilize. Battery cells can be stored in an automated, climate-controlled high-bay warehouse and periodically shuttled to stations for voltage measurements. The step of battery-cell aging can be conducted at varying temperatures: higher temperatures, e.g. higher than room temperature, can accelerate transient self-discharge effects, while lower temperatures, e.g. room temperature, can help identify self-discharge due to, e.g., unwanted soft and / or hard short-circuits. The OCV of a battery cell can be measured at least at the beginning and end of the of battery-cell aging to calculate the OCV-voltage drop of the battery cell. Said OCV-voltage drop is typically just a few millivolts, necessitating the use of high-accuracy multimeters with over 6 digits of resolution. The step of battery-cell aging can last typically between five and fifteen days, and up to several weeks. Each battery cell's OCV-voltage drop measurement typically takes about one second.

[0010] An alternative to measuring the OCV-voltage drop of a battery cell, can involve using an accurate voltage source to maintain an essentially constant battery cell voltage over the step of battery cell aging, which can be a prolonged period of time in the range of, e.g., more than five hours for small cells, and longer for large format cells. The current needed to maintain an essentially constant battery cell voltage over the step of battery cell aging can be taken as a self-discharge-like metric, since it can reflect the self-discharge current.

[0011] In addition to direct self-discharge measurement techniques, other methods may use indirect approaches to identify defective cells and / or to assess the quality of a battery cell based on data from additional techniques, including, for example, ultrasound, entropy, impedance, and calorimetry, optionally in combination with data from manufacturing execution systems (MES), machine learning algorithms (ML), and / or artificial intelligence (Al).

[0012] US 2024 / 0145797 Al relates to a cell charging and discharging tray, a cell ageing device and a cell ageing method, wherein the tray includes an insulating base component for accommodating cells and a conductive component for realizing sequential parallel connection of a type-A cell and type-B cells, and is adapted for simultaneously ageing one type-A cell and a plurality of type-B cells, where a potential of the type-A cell is higher than that of each type-B cell. The tray provided by the present application adopts a design that the base component is separated from the conductive component, so that the ageing of a high-potential cell will be completed after low-potential cells are aged, and the service life of the high-potential cell is not affected, so shipment may be performed after selfdischarge for selecting bad products is completed, which is conducive to improving the manufacturing efficiency.

[0013] CN 112130085 A provides a screening method and device for self-discharge performance of a lithium battery and a computer device. The method comprises the steps that firstly, the voltage of a cell of the lithium battery is adjusted to first voltage, then the temperature and surface pressure of the cell are adjusted to first preset conditions, and second voltage of the cell is recorded after the first preset time is maintained; then the cell is adjusted to third voltage, the temperature and surface pressure of the cell are adjusted to second preset conditions, and fourth voltage of the cell is recorded after the second preset time is maintained; a system calculates a self-discharge value of the cell according to the voltage value and the maintaining time; and the system judges whether the self-discharge value meets the screening standard or not, if yes, that the self-discharge value is qualified is judged, and if not, that the self-discharge value is unqualified is judged. According to the screening method and device for the self-discharge performance of the lithium battery and the computer device, the system can detect chemical self-discharge and physical selfdischarge of the lithium battery, and the detected self-discharge value is more comprehensive; and the corresponding screening standard is established for various types of self-discharge performance, and the screening precision of the self-discharge capacity of the lithium battery can be greatly improved. US 9,157,964 B2 discloses a method for producing a secondary battery, capable of selection of a secondary battery having a defect caused by a micro short-circuit with high accuracy. A step for producing a secondary battery includes an inspection step for selecting the secondary battery having the defect among multiple secondary batteries. In the inspection step, a short circuit resistance is calculated based on capacitances of the multiple secondary batteries calculated before self-discharge inspection, a time required for the self-discharge inspection, first and second reference voltages calculated from open circuit voltages measured before and after the self-discharge inspection in the multiple secondary batteries, and first and second voltages which are open circuit voltages of one selected from the multiple secondary batteries measured before and after the selfdischarge inspection. If the short circuit resistance is equal to or lower than a predetermined standard value, the selected secondary battery is determined to have the defect.

[0014] US 11,623,526 B2 discloses an electrical device including a battery, and a battery management system. The battery management system includes a controller in electrical communication with a pressure sensor to monitor the state of health of the battery. The controller applies a method for determining the state of health that uses a non-electrical (mechanical) signal of force measurements combined with incremental capacity analysis to estimate the capacity fading and other health indicators of the battery with better precision than existing methods. The pressure sensor may provide the force measurement signal to the controller, which may determine which incremental capacity curve based on force to use for the particular battery. The controller then executes a program utilizing the data from the pressure sensor and the stored incremental capacity curves based on force to estimate the capacity fading and signal a user with the state of health percentage.

[0015] US 10,794,960 B2 discloses a technique for effectively detecting a low voltage defect that may occur at a secondary battery. A method for detecting a low voltage defect of a secondary battery includes an assembling step of assembling a secondary battery by accommodating an electrode assembly, in which a positive electrode plate and a negative electrode plate are stacked with a separator being interposed therebetween, and an electrolytic solution in a battery case; a primary aging step of aging the assembled secondary battery at a temperature of 20 °C to 40 °C; a primary formation step of charging the aged secondary battery at a C-rate of 0.1 C to 0.5 C; a high-rate charging step of charging the secondary battery at a C-rate of 2 C or above , after the primary formation step ; and a detecting step of detecting a defect of the secondary battery , after the high- rate charging step. US 2021 / 0208208 Al discloses a method for detecting internal short circuit within an electrochemical cell, from on-line measuring and processing thermodynamics data and kinetics data on the electrochemical cell. The thermodynamics data comprises open-circuit voltage data, entropy variations and / or enthalpy variations data, and combinations thereof. The kinetics data comprises cell voltage, cell temperature, cell internal resistance and current, and combinations thereof.

[0016] US 2014 / 0354233 Al discloses a device for testing a battery. The device for testing a battery may include: a determination circuit configured to determine at least one of a differential enthalpy of the battery and a differential entropy of the battery; and an evaluation circuit configured to evaluate a health state of the battery based on the determined at least one of the differential enthalpy and the differential entropy.

[0017] US 11,860,130 B2 discloses a vehicle that can comprise a battery cell, a monitoring device, and a controller. The monitoring device can comprise an ultrasound source and an ultrasound sensor. The ultrasound source can direct ultrasound at the battery cell, and the ultrasound sensor can detect ultrasound transmitted through or reflected from at least a portion of an interior of the battery cell. The ultrasound sensor can generate one or more signals responsive to the detected ultrasound. The controller can process the one or more signals from the ultrasound sensor and can output an indication of an internal state of the first battery cell.

[0018] EP 3 998 487 Al relates to a battery management system and method for performing a battery health parameter observation, in particular cell impedance observation, with two redundant, independent and dissimilar lanes. Specifically, a cell impedance observation in a first one of the lanes is based on Electrochemical Impedance Spectroscopy, EIS. The other lane employs a different algorithm than EIS. In embodiments, a battery EP 3 998 487 Al state observation is further performed independently by the two lanes, wherein again the first lane employs EIS and the other lane a different (dissimilar) algorithm). On the basis of state and health observation, state (state of function) of the battery system can be predicted to determine a range of flight in accordance with a predetermined flight profile.

[0019] Existing technologies that are primarily directed to the step of battery-cell aging, can exhibit drawbacks.

[0020] The methods disclosed in the prior art may be time-consuming. The step of battery cell aging for quality control may take an extended period of time, which may span days or weeks, which may already be a disadvantage. This may also entail further drawbacks. Given the long time required, the aging process step may usually account for the largest floorspace of all process steps in battery cell manufacturing, the building for the aging process may require advanced climate control, fire detection and extinguishing equipment as well as a lot of logistics to shuttle trays around. This may increase the required capex. Given the long time required, the storage itself may create energy cost through the climate control and maintenance cost. The stored battery cells may further be idle capital and due to the risk of fires in the aging area insurance may be usually high. A technical problem to be solved can be, therefore, the realization of a shortening of the aging period in battery cell manufacturing.

[0021] Prior technologies may also relate to indirect approaches to identify defective cells and / or to assess the quality of a battery cell based on additional data sources, including, for example, ultrasound, entropy, impedance, and calorimetry techniques. Quality indicators of a battery cell, e.g. the self-discharge, may however have only a minor influence on the output of said indirect approaches. For instance, if an ultrasound method is used, the detection limit may be determined by the mechanical dimensions of the cell together with the ultrasound wavelength and, thus, be set to only a superficial layer of a battery cell. As another example, if an impedance method is used, and given that the typical internal resistance of a battery cell may be below 500 pQ, a soft short-circuit of the order of 100 kQ may not be detected by the impedance method.

[0022] Some of said indirect approaches may even not be primarily directed to the measurement of an electrical property, for example, in the case of entropy, calorimetry and ultrasound techniques. Said indirect approaches may further require expensive and extensive hardware, for example expensive and extensive hardware for thermal control in the case of calorimetry or current sources for each cell in the case of entropy.

[0023] Summary

[0024] The present invention alleviates at least some of the above-mentioned shortcomings. For example, the present invention may aim at shortening the aging period in battery cell manufacturing from many days to minutes by using data derived from cell voltage noise as a quality and / or performance indicator.

[0025] In a first aspect, the present invention relates to a method, comprising: measuring at least a voltage noise signal of at least a battery cell, processing the at least one voltage noise signal into a structured data set, analyzing the structured data set and extracting at least a characteristic descriptor from the structured data set, establishing at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell.

[0026] It will be understood that the word "noise", as referred to in this specification with regard to, e.g., a battery cell, can refer to the random fluctuations in an output of the battery cell, e.g. a voltage and / or current output of the battery cell, wherein the random fluctuations can arise from several sources, which can include, among others: thermal agitation of charge carriers in the battery cell's material, variations in chemical reaction rates and / or diffusion processes happening in the battery cell, structural variations and / or imperfections and / or inhomogeneities in the battery cell, external electromagnetic interference, variations in operational conditions of the battery cell such as load, temperature, state of charge and / or any kind of change in the state of the battery-cell, such as charge transfer, reactions, mass transport, mechanical deformation, and / or external influencing factors like, e.g., pressure. The noise can comprise an AC voltage component of the battery cell superimposed on a DC voltage of the battery cell, which is the open circuit voltage of the battery cell. It shall therefore be understood that word "noise", as used herein with regard to e.g. a battery cell, can be part of a concept which is familiar to the person skilled in the art.

[0027] The aging process in prior art technologies may lead to the problem that the amount of measurement data and thus insights generated may be insufficient or small. The OCV- voltage drop method and the method involving the use an accurate voltage source to maintain an essentially constant battery cell voltage over the step of battery cell aging may deliver only a self-discharge voltage and / or current value, lacking the possibilities for further data analysis. From a single value it may not be possible to identify different problems. The present invention may be advantageous in this regard, since it may generate a substantial amount of analyzable data, in the form, e.g., of eigenvalues and / or eigenvectors.

[0028] The method may comprise passively measuring at least a voltage noise signal of at least a battery cell.

[0029] It will be further understood that a "passive" measurement, as referred to in this specification with regard to, e.g., a voltage noise signal of a battery cell, can refer to a measurement that can achieve a measurement value on a system, wherein the operation and / or performance of the system can be substantially unchanged by the measurement. In other words, a "passive" measurement can typically be a non-intrusive measurement. Additionally, or alternatively, a "passive" measurement can refer to a measurement of a system, wherein the system is not subject to any active external excitation during the measurement. As an example, in the case of a "passive" measurement on a battery cell, key characteristics of the passive measurement can include, among others: absence of additional loads and / or currents during measurement, absence of a significant current drawn from the battery cell during measurement , utilization of a measurement device with a high input impedance in order not to load the battery cell, measurement of the battery cell in a resting state and / or during typical operation of the battery cell, absence of disturbance of the electrochemical processes happening in the battery cell during measurement. It shall therefore be understood that word "passive" measurement, as referred to herein with regard to, e.g., a voltage noise signal of a battery cell, can be part of a concept which is familiar to the person skilled in the art.

[0030] Passively measuring may be advantageous for at least the following reasons. Besides the OCV voltage-drop method, prior technologies may not involve passive measurements and may therefore disturb the natural operation of a battery cell. The noise measurement of the present invention may be purely passive measurement, meaning that no excitation (e.g., electrical, thermal, or mechanical) may be used. The passive nature of the method according to the present invention may mean little to no energy is injected into a possible faulty cell, thereby minimizing the risk of cells catching fire. Without an excitation signal, the required hardware effort for the measurement may further be reduced. This may make the method according to the present invention among the ones with lowest hardware effort. The highest effort may instead be needed for measurement methods that require stable thermal conditions.

[0031] The step of processing the at least one voltage noise signal into a structured data set may comprise determining at least a matrix based at least in part on the at least one voltage noise signal.

[0032] The step of processing the at least one voltage noise signal into a structured data set may comprise determining eigenvalues and / or eigenvectors of the at least one matrix.

[0033] The step of processing the at least one voltage noise signal into a structured data set may comprise determining singular values and / or singular vectors of the at least one matrix.

[0034] The structured data set may comprise at least an eigenvalue metric of said eigenvalues and / or at least an eigenvector metric of said eigenvectors.

[0035] The structured data set may comprise at least a singular value metric of said singular values and / or at least a singular vector metric of said singular vectors. The determination of matrices and eigenvalues and / or eigenvector may provide advantages over existing technologies. An intermediate result of the method according to the present invention, e. g., the eigenvectors and eigenvalues, may provide a higher data density compared to the classical OCV drop value. This amount of data together with different evaluation options may allow for much deeper data analysis. In methods involving an OCV-voltage drop, or a current needed to maintain an essentially constant battery cell voltage over the step of battery cell aging, or the like, only a limited amount of data may be generated in the step of battery-cell aging. In simpler words, in the exemplary case when an OCV-voltage drop method is used, the level quality of a battery cell is assessed based on one or more OCV-voltage drops. It may be difficult, from merely one or more OCV-voltage drops, to get insights into the cause of the level of quality of the battery-cell.

[0036] The method may comprise using a characteristic quantitative descriptor.

[0037] The method may comprise using a characteristic qualitative descriptor.

[0038] The method may comprise using a statistical descriptor as the characteristic descriptor.

[0039] The method may comprise using a quality metric of the at least one battery cell as the performance indicator of the at least one battery cell.

[0040] The method may comprise using a self-discharge-like metric as the performance indicator of the at least one battery cell.

[0041] Prior use of noise analysis methods may be focused only on the effects of state of charge (SoC) and state of health (SoH) of batteries. In this regard, the present invention can provide an advantage by extending the scope of noise analysis to self-discharge.

[0042] The method may comprise building at least a map based at least in part on the at least one relation between the at least one characteristic descriptor and at least a performance indicator.

[0043] The map may link the at least one characteristic descriptor to at least an equivalent performance indicator and / or an updated performance indicator.

[0044] The method may comprise using the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on the at least one characteristic descriptor. In other words, the map may transform the at least one characteristic descriptor into at least one among: a performance parameter equivalent or a new performance indicator. If, for example, the performance indicator was a self-discharge-like metric, then the performance parameter equivalent may also be a self-discharge-like metric. The performance parameter may be known before the method according to the present invention is performed. Moreover, if, for example, the performance indicator was a self- discharge-like metric, then the new performance indicator may not be a self-discharge-like metric. The new performance indicator may not be known before the method according to the present invention is performed. On the contrary, new performance indicator may be established thanks to the data analyses performed in the method according to the present invention.

[0045] The method may comprise: measuring at least a voltage noise signal of at least a battery cell with a measuring module, processing the at least one voltage noise signal into a structured data set with a processing module, analyzing the structured data set and extracting at least a characteristic descriptor from the structured data set with an analyzing module, establishing at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell with an establishing module.

[0046] The method may comprise using a data and control interface module.

[0047] The method may comprise using a transform module, comprised in the processing module.

[0048] The method may comprise communicating between the measuring module and the processing module.

[0049] The method may comprise communicating between the measuring module and the processing module at least the at least one measured voltage noise signal.

[0050] The method may comprise communicating between the processing module and the transform module.

[0051] The method may comprise communicating between the transform module and the analyzing module.

[0052] The method may comprise communicating between the analyzing module and the establishing module. The method may comprise communicating between the analyzing module and the establishing module at least the at least one characteristic descriptor.

[0053] The method may comprise communicating between the data and control interface module and any combination of modules configured to perform the method according to any of the preceding embodiments, said combination of modules comprising at least one module.

[0054] The method may comprise changing, with the data and control interface module, at least one parameter related to any combination of steps of the method, said combination of steps comprising at least a step of the method.

[0055] It will be understood that a combination of modules can, for example, comprise the processing module, the analyzing module, and the establishing module. The combination may be referred to as an evaluation unit. The evaluation unit may perform, without limitation, the steps of the method performed by the processing module, the analyzing module, and the establishing module.

[0056] The method may comprise communicating, with the data and control interface module, to a user at least any combination between: the at least one voltage noise signal of at the least one battery cell, the structured data set, the at least one characteristic descriptor, the at least one relation between the at least one characteristic descriptor, and the at least one performance indicator of the at least one battery cell.

[0057] The method may comprise arranging locally or remotely one or more among the transform module, processing module, the analyzing module and the establishing module.

[0058] In other words, one or more modules among, for example, the transform module, processing module, the analyzing module and the establishing module may not be physically located in the vicinity of the measuring module.

[0059] The method may comprise using a computing device.

[0060] Any module according to any of the preceding embodiments may comprise the computing device.

[0061] The method may comprise utilizing different software for different purposes within a module according to any of the preceding embodiments.

[0062] The method may comprise communicating between the data and control interface module and a manufacturing execution system (MES) of a battery manufacturing plant. The method may comprise using a high-resolution AC voltage detecting set-up.

[0063] The method may comprise using a card of the high-resolution AC voltage detecting set-up to measure a voltage noise signal from the at least one battery cell.

[0064] The method may comprise using multiple cards of the high-resolution AC voltage detecting set-up to measure a plurality of voltage noise signal from a plurality of battery cells, in parallel.

[0065] The method may comprise using the high-resolution AC voltage detecting set-up is with a resolution below 10 nV, preferably in a range between 0.1 nV and 10 nV, preferably below 0.1 nV.

[0066] The method may comprise using an AC-coupling device, embedded in the card of the high- resolution AC voltage detecting set-up.

[0067] The method may comprise using a coupling capacitor, comprises in the AC-coupling device.

[0068] The method may comprise using an AC-coupling device with a lower cut-off frequency in a range between 10 Hz and 50 Hz, preferably in a range between 1 Hz and 10 Hz, more preferably below 1 Hz.

[0069] The method may comprise utilizing a high-gain low-noise amplifier embedded in the card of the high-resolution AC voltage detecting set-up.

[0070] The method may comprise using the high-gain low-noise amplifier with a gain in a range between 100 and 2000, preferably in a range between 2000- and 10000, most preferably above 10000.

[0071] The method may comprise using the high-gain low-noise amplifier with a noise figure at 1 Hz below 50 nV / sqrt(Hz), preferably below 20 nV / sqrt(Hz), more preferably below 5 nV / sqrt(Hz).

[0072] The method may comprise utilizing a high-resolution analog-to-digital (ADC) converter embedded in the card of the high-resolution AC voltage detecting set-up. The method may comprise utilizing the ADC with a sampling rate in a range between 1 Sa / s and 100 MSa / s, preferably between 100 Sa / s and 1 MSa / s, most preferably between 0.5 kSa / s and 30 kSa / s.

[0073] The method may comprise utilizing the ADC with a resolution greater than or equal to 16 bit, preferably greater than or equal to 20 bit, more preferably greater than or equal to 24 bit.

[0074] The method may comprise comprises utilizing a control and communication unit, comprised in the card of the high-resolution AC voltage detecting set-up.

[0075] The method may comprise isolating an alternate coupling (AC) voltage noise signal of the at least one battery cell from the voltage signal of at the at least one battery, said voltage signal comprising an AC component and a direct coupling (DC) component.

[0076] In existing technologies, the noise measurement features are not described in detail, but rather generically, thus potentially ignoring the problem of a very small AC component on top of a large DC component.

[0077] The method may comprise amplifying the AC voltage noise signal into an amplified voltage noise signal.

[0078] The method may comprise sampling the amplified voltage noise signal and obtaining a digitized voltage noise signal.

[0079] The method may comprise communicating the digitized voltage noise signal to the data and control interface module.

[0080] The method may comprise communicating, via ethernet, the digitized voltage noise signal to the data and control interface module.

[0081] It will be understood that the data and control interface module may comprise a control PC.

[0082] The method may comprise controlling the gain and / or the sampling rate of the high- resolution ADC with the control and communication unit.

[0083] The method may comprise controlling a polarization compensation of the coupling capacitor with the control and communication unit. The method may comprise enabling at least a temperature measurement of the at least one battery cell with the control and communication unit.

[0084] The method may comprise enabling at least a measurement of the at least one voltage noise signal.

[0085] The method may comprise pre-processing the at least one voltage noise signal into at least one pre-processed voltage noise signal.

[0086] The pre-processing may comprise cleaning the at least one voltage noise signal from artifacts.

[0087] For instance, artifacts can be spikes in the at least one voltage noise signal.

[0088] The pre-processing may comprise detrending the at least one voltage noise signal.

[0089] The method may comprise generating the at least one matrix from the at least one voltage noise signal.

[0090] The method may comprise generating the at least one matrix from the at least one pre- processed voltage noise signal.

[0091] The method may comprise establishing a measurement signal, wherein the measurement signal comprises a section of the pre-processed voltage noise signal.

[0092] The section of the pre-processed voltage noise signal may have a temporal duration between 1 minute and 240 minutes, preferably between 5 minutes and 60 minutes, more preferably between 5 minutes and 15 minutes.

[0093] The method may comprise establishing at least a processing window, wherein the processing window comprises a section of the measurement signal.

[0094] The method may comprise sliding, in time, the at one least processing window over the measurement signal.

[0095] The section of the measurement signal, comprised in the processing window at a first time, may overlap in time with the section of the measurement signal, comprised in the processing window at a second time. Said overlap may in a range between 70% and 99.9%, preferably in a range between 90% and 99%, more preferably in a range between 97% and 99% of a temporal length of the processing window.

[0096] The method may comprise using said overlap as a parameter.

[0097] The method may comprise using a temporal length of the processing window as a parameter.

[0098] The temporal length of the processing window may be in a range between 30 s and 360 s, preferably between 60 s and 240 s, more preferably between 120 s and 180 s.

[0099] The method may comprise establishing non-overlapping segments.

[0100] Each of the non-overlapping segments may comprise a non-overlapping section of the at least one processing window.

[0101] The non-overlapping segments may uniform.

[0102] The method may comprise using the number of the non-overlapping segments as a parameter.

[0103] The method may comprise deriving the at least one matrix from the non-overlapping segments.

[0104] The at least one matrix may comprise a Pearson correlation matrix.

[0105] The method may comprise deriving the Pearson correlation matrix from a measurement matrix.

[0106] The method may comprise deriving the measurement matrix from the non-overlapping segments.

[0107] The method may comprise generating the entries of a row of the measurement matrix from one of the non-overlapping segments.

[0108] The method may comprise generating the Pearson correlation matrix based on the following formula: C = AAT / (n-l) wherein C may represent the Pearson correlation matrix, wherein A may represent the measurement matrix, n may be the number of columns of the measurement matrix, wherein each element in a row of A may have the mean of said row of A removed, and wherein each element in a row of A may be divided by the standard deviation of said row of A.

[0109] The at least one matrix may comprise a cross-correlation matrix.

[0110] The method may comprise deriving the cross-correlation matrix from the non-overlapping segments.

[0111] The method may comprise generating the entries of the cross-correlation matrix based on the absolute cross correlation function between any of the uniform non-overlapping segments.

[0112] The at least one matrix may comprise a mutual information matrix.

[0113] The method may comprise deriving the mutual information matrix from the nonoverlapping segments.

[0114] The method may comprise generating the entries of the mutual information matrix from a mutual information between any of the uniform non-overlapping segments.

[0115] The at least one matrix may comprise a square matrix.

[0116] The at least one matrix may comprise a matrix-shaped representation of the at least one voltage noise signal.

[0117] The method may comprise method comprises determining eigenvalues and / or eigenvectors of at least one among: the Pearson correlation matrix, the cross-correlation matrix, the mutual information matrix, the matrix-shaped representation of the at least one voltage noise signal.

[0118] The at least one matrix may comprise a measurement matrix. The method may comprise deriving the measurement matrix from the non-overlapping segments. It will be understood that, instead of determining eigenvalues and / or eigenvectors of the Pearson correlation matrix, the method may alternatively comprise determining the singular values and / or singular vectors of the measurement matrix.

[0119] It will be understood that the calculation of the eigenvalues and eigenvectors based on singular value decomposition of the measurement matrix may result in left and right singular vectors of the measurement matrix. The left singular vectors may represent the eigenvectors of the correlation matrix, the right singular vectors may represent the time signals that form the basis of the measurement matrix. The squared singular values may represent the eigenvalues of the correlation matrix. It will be understood, additionally or alternatively to metrics such as statistical moments, spectral properties in the frequency domain, signal entropy and / or the like used for time domain signals, that any metric according to any embodiment of the present invention, such as qualitative and / or quantitative metrics mentioned herein, may also be applied, mutatis mutandis, to the right and / or left singular vectors. Advantages and / or details discussed in the context of, e.g., eigenvalues and eigenvectors may apply also in the context of singular values and / or singular vectors.

[0120] It will be further understood that there might be a plurality of matrices that are used simultaneously for eigenvalue and / or eigenvector calculation.

[0121] The method may comprise determining the at least one eigenvalue metric and / or at least one eigenvector metric with the transform module.

[0122] The method may comprise determining the at least one singular value metric and / or at least one singular vector metric with the transform module.

[0123] The at least one eigenvalue metric may comprise a spectral radius and / or a trace and / or a condition number and / or distributions of the eigenvalues and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0124] It will be understood that the spacings of the eigenvalues and / or the overlap of the eigenvalues can be of arbitrary order.

[0125] The at least one eigenvector metric may comprise a localization metric and / or an entropy of components of the eigenvector and / or a localization length and / or the like. For instance, the localization metric may be, without limitation, an inverse participation ratio of the eigenvectors and / or a participation ratio of the eigenvectors. The entropy of components of the eigenvector can be, for example and without limitation, a Shannon entropy, a Structural entropy, and / or a Renyi entropy.

[0126] The method may comprise selecting a set of the at least one eigenvalue and / or at least one eigenvector based on limits provided by the random matrix theory (RMT).

[0127] In simpler words, the knowledge from RMT may be used to separate information from purely random behavior. RMT may relate to the ideal case for pure random noise without information; any deviation from it may be caused by information within the noise which may allow a first preselection.

[0128] The at least one singular value metric may calculated on the squares of the singular values, which are eigenvalues, and may comprise: a spectral radius and / or a trace and / or condition number and / or distributions of the eigenvalues values and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0129] The at least one singular vector metric may be calculated on the left singular vectors and may comprise: a localization metric and / or an entropy of components of the singular vector and / or a localization length and / or the like.

[0130] The method may comprise selecting a set of the at least one singular value and / or at least one singular vector based on limits provided by the random matrix theory (RMT).

[0131] Embodiments of the present invention may be at least in part directed to voltage noise measurements of a battery cell and to the analysis of voltage noise measurements of a battery cell at least in part with methods from random matrix theory. This may lead to advantages.

[0132] For example, although prior technologies may be directed to voltage noise measurements of a battery cell, they may not analyze the voltage noise measurements of a battery cell at least in part with methods from random matrix theory, thus potentially missing, at least in part, the temporal structure of the voltage noise. Noise data analysis for battery voltage noise measurements may be done in prior technologies by, for example, analysis of frequency spectra, analysis of statistical modes, fractal analysis and / or recurrence analysis. With the exception of recurrence analysis, the methods may use a whole dataset at once to analyze the data, thus potentially ignoring the temporal structure of the noise.

[0133] Further, the evaluation of eigenvalues and eigenvectors may be rooted in RMT and may have found applications in other fields where noise may be analyzed, like financial markets or network analysis. However, it may not be used in prior technologies for noise analysis in either corrosion or battery noise measurements.

[0134] The method may comprise enabling an evaluation of the structured data set based on the at least one characteristic descriptor.

[0135] The method may comprise extracting at least a characteristic descriptor from the structured data set derived at least from the processing window at the first time and from the processing window at the second time.

[0136] In other words, as the processing window may slide, in time, over the measurement signal and comprise different sections of the measurement signal at different times, an analysis can be done based on the processing window at each of said different times.

[0137] It will be understood that descriptors may be built from the structured data set of all evaluated processing windows.

[0138] The at least one characteristic descriptor may comprise one or more statistical moments of the structured data set and / or one or more divergence values for one or more densities in the structured data set and / or one or more residual squared sum values for one or more differences between the structured data set and theoretical values and / or the like.

[0139] For example, the at least one characteristic descriptor can comprise, without limitation, a mean of the largest eigenvalue for all processing windows in a measurement. As another example, the at least one characteristic descriptor can comprise, without limitation, a standard deviation of the largest eigenvalue for all processing windows in a measurement.

[0140] The method may comprise using a remaining useful life-time and / or a power handling capability as the performance indicator of the at least one battery cell.

[0141] The method may comprise performing a sensitivity analysis of the at least one characteristic descriptor with respect to the performance indicator of the at least one battery cell. The method may comprise selecting a set of the at least one characteristic descriptor based on the sensitivity analysis.

[0142] In other words, the sensitivity analysis may look for a high correlation between the characteristic descriptor and the performance indicator (e.g. self-discharge). As a result, the characteristic descriptors with high correlation to the performance indicator may be used to build the map in other steps of the method, while other characteristic descriptors that may not show high correlation may be discarded.

[0143] The method may comprise establishing the at least one relation based, at least in part, on the sensitivity analysis.

[0144] The method may comprise performing a statistical analysis of the at least one characteristic descriptor.

[0145] The method may comprise utilizing the statistical analysis to identify clusters of behavior and / or trends and / or tendencies among the at least one characteristic descriptor.

[0146] The method may comprise establishing the at least one relation based, at least in part, on the statistical analysis.

[0147] The at least one relation may be a functional relation.

[0148] The map may coincide with the at least one relation.

[0149] It will be understood that there can be a plurality of maps that are used simultaneously.

[0150] The at least one equivalent performance indicator and / or one updated performance indicator may be a continuous performance indicator.

[0151] A continuous performance indicator can, for instance, be a self-discharge, since the selfdischarge can have continuous values.

[0152] The at least one equivalent performance indicator and / or one updated performance indicator may be a discrete performance indicator.

[0153] A discrete performance indicator can, for instance, be grades of battery cell quality, since the grading of battery cell quality can comprise discrete grades of battery cell quality. The method may comprise using the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on at least a characteristic descriptor of at least a newly measured battery cell.

[0154] In simpler words, the method may comprise a phase of the method, which may be addressed as "set-up" phase, wherein the method comprises measuring at least one battery-cell and building the map based, at least in part, thereon. The method may further comprise another phase of the method, which may be referred to as "use" phase, wherein the method comprises measuring at least a new battery cell and using the map based, at least in part, thereon.

[0155] The method may comprise improving a battery-cell manufacturing process based at least in part on said the at least one equivalent performance indicator and / or one updated performance indicator.

[0156] As an example, steps of the method according to the present invention may involve measurements on a first battery cells of a first type of battery cells and on a second battery cells of a second type of battery cells. An equivalent performance indicator and / or an updated performance indicator for the first battery cells and for the second battery cells can be evaluated according to steps of the method of the present invention. It may then be possible to see or infer how the first type of battery cells and the second type of battery cells generally performs. This piece of information may be useful to tune the battery-cell manufacturing process in order to make better types of battery cells. In other words, for example, this piece of information may be useful to tune a process parameter of the battery-cell-manufacturing process on the basis on how features of a battery cell, i.e. something defining the type of a cell, influences its performance. The method according to the present invention may therefore comprise improving a battery-cell manufacturing process.

[0157] In light of the above, the present invention may provide advantages as compared to existing technologies. Conventional battery cell aging processes may take prolonged amounts of time, of the order of day or weeks. Therefore, it may be difficult to provide swift feedback to the process for battery-cell-manufacturing, aimed at improving the process for battery-cell-manufacturing. The present invention may, at least in part, do that.

[0158] The present invention may therefore, at least in part, limit the generally high scrap-rate of a battery-cell-manufacturing process, especially at ramp-up of the battery-cellmanufacturing process. Besides directly identifying weak cells in the early phases of ageing, the method may have the potential of providing fast feedback for the production process and additional data to predict the lifecycle performance of the cells. For fast production process feedback, a measurement method may need to provide data, that allows for discrimination off different causes of problems. For example, the basic OCV voltage drop method may only provide a limited range of values. From a limited range of values it may not possible to identify different problems.

[0159] It may generally be a goal for cell manufacturers to have access to a method for providing a fast feedback loop for process optimization und tuning.

[0160] The method may be a method for assessing the quality of a battery cell in a battery-cellmanufacturing process.

[0161] The method may be is a method for assessing the intrinsic activity of a battery cell.

[0162] The step of measuring at least a voltage noise signal of at least a battery cell may take a range of time from 1 hour to 2 hours, preferably from 30 minutes to 1 hour, more preferably from 10 minutes to 30 minutes.

[0163] The step of measuring at least a voltage noise signal of at least a battery cell may take a range of time from 1 minute to 2 hours, preferably from 5 minutes to 1 hour, more preferably from 10 minutes to 30 minutes.

[0164] This may be particularly advantageous in speeding up the aging phase, as compared to existing technologies. Existing technologies may take a time going from a couple of days to weeks for the aging process. Embodiments of the present invention, instead, may allow to quantify, e. g., the self-discharge in a time that may be mainly taken up by the step of measuring at least a voltage noise signal of at least a battery cell, said step taking considerably less time than the quantification of self-discharge via, e.g., aging according to existing technologies.

[0165] The method may be a method for manufacturing at least a battery cell.

[0166] The method may be a method for identifying anomalies in at least a battery cell.

[0167] The method may be a method for identifying at least a cause of anomalies in at least a battery cell.

[0168] The method may be is a method for testing at least a battery cell for self-discharge in a battery-cell-manufacturing process. The method may be is a method for speeding up the cell finishing step in a battery-cellmanufacturing process.

[0169] The method may be a method for speeding up the aging phase in a battery-cell manufacturing process.

[0170] The aging phase in a battery-cell manufacturing process using the method according to any embodiment of the present invention may last between 10 minutes and 5 days, preferably between 20 minutes and 10 hours, more preferably between 30 minutes and 3 hours.

[0171] The aging phase in a battery-cell manufacturing process using the method according to any of the embodiments of the present invention may last between 1 and 5 days, preferably between 5 hours and 10 hours, more preferably between 1 hour and 3 hours.

[0172] The method may be a method for directly identifying at least a weak and / or no good and / or outlier and / or scrap battery cell during the aging phase of a battery-cell-manufacturing process.

[0173] The method may be is a method for testing incoming goods inspection for self-discharge and / or anomalies, wherein the incoming goods comprise a battery cell.

[0174] For example, older cells may need to be measured for a possible second life use case. For this application, cells may need to be tested for increased self-discharge and possible anomalies.

[0175] The method may be a method for testing at least a fuel-cell for anomalies.

[0176] Fuel-cells may be difficult to measure due to the complex nature of the electrochemical reactions at the membrane of fuel-cells. Using a passive measurement technique, according to embodiments of the present invention, may allow a measurement without substance conversion, thus removing this influence.

[0177] The method may be a method for testing a performance indicator of a chemical sensor.

[0178] Chemical sensor may benefit from a noise-based measurement. The method may be a method for improving at least a battery-cell-manufacturing process of at least a battery cell.

[0179] The method may be a method a method for improving at least a step of a battery-cellmanufacturing process of at least a battery cell.

[0180] The method may be a method a method for optimizing the costs of at least a battery-cellmanufacturing process of at least a battery cell.

[0181] The method may be a method a method for increasing the yield of at least a battery-cellmanufacturing process of at least a battery cell.

[0182] In a second aspect, the present invention relates to a system, comprising: at least a measuring module configured to measure at least a voltage noise signal of at least a battery cell, at least a processing module configured to process the at least one voltage noise signal into a structured data set, at least an analyzing module configured to analyze the structured data set and to extract at least a characteristic descriptor from the structured data set, at least an establishing module configured to establish at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell.

[0183] The aging process in prior art technologies may lead to the problem that the amount of measurement data and thus insights generated is very small. The OCV-voltage drop method and the method involving the use an accurate voltage source to maintain an essentially constant battery cell voltage over the step of battery cell aging may deliver only a self-discharge voltage and / or current value, lacking the possibilities for further data analysis. From a single value it may not be possible to identify different problems. The system according to present invention may be advantageous in this regard, since it may generate a substantial amount of analyzable data, in the form, e.g., of eigenvalues and / or eigenvectors.

[0184] The measuring module may be configured to passively measure at least a voltage noise signal of at least a battery cell.

[0185] Passively measuring may be advantageous for at least the following reasons. Besides the OCV voltage-drop method, prior technologies may not involve passive measurements and may therefore disturb the natural operation of a battery cell. The noise measurement of the present invention may be purely passive measurement, meaning that no excitation (electrical or mechanical) may be used. The passive nature of the measurement performed by the system according to the present invention may mean little to no energy is injected into a possible faulty cell, thereby reducing the risk of cells catching fire. Without an excitation signal, the required hardware effort for the measurement may further be reduced. This may make the system according to the present invention among the ones with lowest hardware effort. The highest effort may instead be needed for measurement systems that require stable thermal conditions.

[0186] The processing module may be configured to determine at least a matrix based at least in part on the at least one voltage noise signal.

[0187] The processing module may be configured to determine eigenvalues and / or eigenvectors of the at least one matrix.

[0188] The processing module may be configured to determine singular values and / or singular vectors of the at least one matrix.

[0189] The determination of matrices and eigenvalues and / or eigenvector may provide advantages over existing technologies. An intermediate result the system according to the present invention can produce, e. g. the eigenvectors and eigenvalues, may provide a higher data density compared to the classical OCV drop value. This amount of data together with different evaluation options may allow for much deeper data analysis. In systems configured for performing methods involving an OCV-voltage drop, or involving a current needed to maintain an essentially constant battery cell voltage over the step of battery cell aging, or the like, only a limited amount of data may be generated in the step of batterycell aging. In simpler words, in the exemplary case when system adapted for the OCV- voltage drop method is used, the level quality of a battery cell is assessed based on one or more OCV-voltage drops. It may be difficult, from merely one or more OCV-voltage drops, to get insights into the cause of the level of quality of the battery-cell.

[0190] The structured data set may comprise at least an eigenvalue metric of said eigenvalues and / or at least an eigenvector metric of said eigenvectors.

[0191] The structured data set may comprise at least a singular value metric of said singular values and / or at least a singular vector metric of said singular vectors.

[0192] The analyzing module may be configured to use a characteristic quantitative descriptor.

[0193] The analyzing module may be configured to use a characteristic qualitative descriptor. The analyzing module may be configured to use a statistical descriptor as the characteristic descriptor.

[0194] The establishing module may be configured to use a quality metric of the at least one battery cell as the performance indicator of the at least one battery cell.

[0195] The establishing module may be configured to use a self-discharge-like metric as the performance indicator of the at least one battery cell.

[0196] The establishing module may be configured to build at least a map based at least in part on the at least one relation between the at least one characteristic descriptor and at least a performance indicator.

[0197] The map may link the at least one characteristic descriptor to at least an equivalent performance indicator and / or an updated performance indicator

[0198] The establishing module may be configured to use the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on the at least one characteristic descriptor.

[0199] In other words, the map may transform the at least one characteristic descriptor into at least one among: a performance parameter equivalent or a new performance indicator. If, for example, the performance indicator was a self-discharge-like metric, then the performance parameter equivalent may also be a self-discharge-like metric. The performance parameter may be known before the method according to the present invention is performed. Moreover, if, for example, the performance indicator was a self- discharge-like metric, then the new performance indicator may not be a self-discharge-like metric. The new performance indicator may not be known before the method according to the present invention is performed. On the contrary, new performance indicator may be established thanks to the data analyses performed in the method according to the present invention.

[0200] The system may comprise a data and control interface module.

[0201] The processing module may comprise a transform module.

[0202] The measuring module may be configured to communicate with the processing module. The measuring module may be configured to communicate to the processing module at least the at least one measured voltage noise signal.

[0203] The processing module may be configured to communicate with the transform module.

[0204] The transform module may be configured to communicate with the analyzing module.

[0205] The analyzing module may be configured to communicate with the establishing module.

[0206] The analyzing module may be configured to communicate to the establishing module at least the at least one characteristic descriptor.

[0207] The data and control interface module may be configured to communicate with any combination of modules according to any of the preceding system embodiments, said combination of modules comprising at least one module.

[0208] The data and control interface module may be configured to change at least one parameter related to any combination of modules according to any of the preceding system embodiments, said combination of modules comprising at least one module.

[0209] It will be understood that a combination of modules can, for example, comprise the processing module, the analyzing module, and the establishing module. The combination may be referred to as an evaluation unit. The evaluation unit may be configured to perform, without limitation, what the processing module, the analyzing module, and the establishing module are configured to perform.

[0210] It will further be understood that, in this specification, "module may be configured to ..." or analogous terms may be construed as "system may be configured to ..." or analogous terms, and vice versa.

[0211] The data and control interface module may be configured to communicate to a user at least any combination between: the at least one voltage noise signal of at the least one battery cell, the structured data set, the at least one characteristic descriptor, the at least one relation between the at least one characteristic descriptor, and the at least one performance indicator of the at least one battery cell.

[0212] One or more among the transform module, processing module, the analyzing module and the establishing module may be arranged locally or remotely. In other words, one or more modules among, for example, the transform module, processing module, the analyzing module and the establishing module may not be physically located in the vicinity of the measuring module.

[0213] The system may comprise a module comprising a combination of any of the modules according to any of the preceding system embodiments.

[0214] The system may be configured to utilize different software for different purposes within a module according to any of the preceding system embodiments.

[0215] Any module according to any of the preceding embodiments may comprise a computing device.

[0216] The data and control interface module may be configured to communicate with a manufacturing execution system (MES) of a battery manufacturing plant.

[0217] The measuring module may comprise a high-resolution AC voltage detecting set-up.

[0218] The high-resolution AC voltage detecting set-up may comprise at least a card of the high- resolution AC voltage detecting set-up, and wherein said card may be configured to measure a voltage noise signal from the at least one battery cell.

[0219] The high-resolution AC voltage detecting set-up c may comprise multiple cards of the high- resolution AC voltage detecting set-up, and wherein said multiple cards may be configured to measure a plurality of voltage noise signal from a plurality of battery cells, in parallel.

[0220] The high-resolution AC voltage detecting set-up may have a resolution below 10 nV, preferably in a range between 0.1 nV and 10 nV, preferably below 0.1 nV.

[0221] The at least one card may comprise an AC-coupling device.

[0222] The AC-coupling device may comprise a coupling capacitor.

[0223] The AC-coupling device may have a lower cut-off frequency in a range between 10 Hz and 50 Hz, preferably in a range between 1 Hz and 10 Hz, more preferably below 1 Hz.

[0224] The at least one card may comprise a high-gain low-noise amplifier. The high-gain low-noise amplifier may have a gain in a range between 100 and 2000, preferably in a range between 2000- and 10000, most preferably above 10000.

[0225] The high-gain low-noise amplifier may have a noise figure at 1 Hz below 50 nV / sqrt(Hz), preferably below 20 nV / sqrt(Hz), more preferably below 5 nV / sqrt(Hz).

[0226] The at least one card may comprise a high-resolution analog-to-digital (ADC) converter.

[0227] The high-resolution ADC may have a sampling rate in a range between 1 Sa / s and 100 MSa / s, preferably between 100 Sa / s and 1 MSa / s, most preferably between 0.5 kSa / s and 30 kSa / s.

[0228] The high-resolution ADC may have a resolution greater than or equal to 16 bit, preferably greater than or equal to 20 bit, more preferably greater than or equal to 24 bit.

[0229] The at least one card may comprise a control and communication unit.

[0230] The AC coupling device may be configured to isolate an alternate coupling (AC) voltage noise signal of the at least one battery cell from a voltage signal of at the at least one battery, said voltage signal comprising an AC component and a direct coupling (DC) component.

[0231] In existing technologies, the noise measurement features are not described in detail, but rather generically, thus potentially ignoring the problem of a very small AC component on top of a large DC component.

[0232] The high-gain low-noise amplifier comprises may be configured to amplify the AC voltage noise signal into an amplified voltage noise signal.

[0233] The high-resolution ADC may be configured to sample the amplified voltage noise signal and obtain a digitized voltage noise signal.

[0234] The measuring module may be configured to communicate the digitized voltage noise signal to the data and control interface module.

[0235] The measuring module may be configured to communicate, via ethernet, the digitized voltage noise signal to the data and control interface module. It will be understood that the data and control interface module may comprise a control PC or the like.

[0236] The control and communication unit may be configured to control the gain and / or the sampling rate of the high-resolution ADC.

[0237] The control and communication unit may be configured to control a polarization compensation of the coupling capacitor.

[0238] The control and communication unit may be configured to enable at least a temperature measurement of the at least one battery cell with the control and communication unit.

[0239] The control and communication unit may be configured to enable at least a measurement of the at least one voltage noise signal.

[0240] The processing module may be configured to pre-process the at least one voltage noise signal into at least one pre-processed voltage noise signal.

[0241] The processing module may be configured to pre-process the at least one voltage noise signal into at least one pre-processed voltage noise signal by cleaning the at least one voltage noise signal from artifacts.

[0242] For instance, artifacts can be spikes in the at least one voltage noise signal.

[0243] The processing module may be configured to pre-process the at least one voltage noise signal into at least one pre-processed voltage noise signal by detrending the at least one voltage noise signal.

[0244] The processing module may be configured to generate the at least one matrix from the at least one voltage noise signal.

[0245] The processing module may be configured to generate the at least one matrix from the at least one pre-processed voltage noise signal.

[0246] The processing module may be configured to establish a measurement signal, wherein the measurement signal comprises a section of the pre-processed voltage noise signal. The section of the pre-processed voltage noise signal may have a temporal duration between 1 minute and 240 minutes, preferably between 5 minutes and 60 minutes, more preferably between 5 minutes and 15 minutes.

[0247] The processing module may be configured to establish at least a processing window, wherein the processing window comprises a section of the measurement signal.

[0248] The processing module may be configured to slide, in time, the at one least processing window over the measurement signal.

[0249] The section of the measurement signal, comprised in the processing window at a first time, may overlap in time with the section of the measurement signal, comprised in the processing window at a second time.

[0250] Said overlap may be in a range between 70% and 99.9%, preferably in a range between 90% and 99%, more preferably in a range between 97% and 99% of a temporal length of the processing window.

[0251] The processing module may be configured to use said overlap as a parameter.

[0252] The processing module may be configured to use a temporal length of the processing window as a parameter.

[0253] The temporal length of the processing window may be in a range between 30 s and 360 s, preferably between 60 s and 240 s, more preferably between 120 s and 180 s.

[0254] The processing module may be configured to establish non-overlapping segments.

[0255] Each of the non-overlapping segments may comprise a non-overlapping section of the at least one processing window.

[0256] The non-overlapping segments may be uniform.

[0257] The processing module may be configured to use the number of the non-overlapping segments as a parameter.

[0258] The processing module may be configured to derive the at least one matrix from the nonoverlapping segments. The at least one matrix may comprise a Pearson correlation matrix.

[0259] The processing module may be configured to derive the Pearson correlation matrix from a measurement matrix.

[0260] The processing module may be configured to derive the measurement matrix from the non-overlapping segments.

[0261] The processing module may be configured to generate the entries of a row of the measurement matrix from one of the non-overlapping segments.

[0262] The processing module may be configured to generate the Pearson correlation matrix based on the following formula: C = AAT / (n-l) wherein C may represent the Pearson correlation matrix, wherein A represent the measurement matrix, n may be the number of columns of the measurement matrix, wherein each element in a row of A may have the mean of said row of A removed, and wherein each element in a row of A may be divided by the standard deviation of said row of A.

[0263] The at least one matrix may comprise a cross-correlation matrix.

[0264] The processing module may be configured to derive the cross-correlation matrix from the non-overlapping segments.

[0265] The processing module may be configured to generate the entries of the cross-correlation matrix based on the absolute cross correlation function between any of the uniform nonoverlapping segments.

[0266] The at least one matrix may comprise a mutual information matrix.

[0267] The processing module may be configured to derive the mutual information matrix from the non-overlapping segments.

[0268] The processing module may be configured to generate the entries of the mutual information matrix from a mutual information between any of the uniform non-overlapping segments.

[0269] The at least one matrix may comprise a square matrix. The at least one matrix may comprise a matrix-shaped representation of the at least one voltage noise signal.

[0270] The processing module may be configured to determine eigenvalues and / or eigenvectors of at least one among: the Pearson correlation matrix, the cross-correlation matrix, the mutual information matrix, the matrix-shaped representation of the at least one voltage noise signal.

[0271] The at least one matrix may comprise a measurement matrix.

[0272] The processing module may be configured to derive the measurement matrix from the non-overlapping segments.

[0273] It will be understood that, instead of determining eigenvalues and / or eigenvectors of the Pearson correlation matrix, the processing module may alternatively be configured to determine the singular values and / or singular vectors of the measurement matrix.

[0274] It will be understood that the calculation of the eigenvalues and eigenvectors based on singular value decomposition of the measurement matrix may result in left and right singular vectors of the measurement matrix. The left singular vectors may represent the eigenvectors of the correlation matrix, the right singular vectors may represent the time signals that form the basis of the measurement matrix. The squared singular values may represent the eigenvalues of the correlation matrix. It will be understood, that additionally or alternatively to metrics such as statistical moments, spectral properties in the frequency domain, signal entropy and / or the like used for time domain signals, any metric according to any embodiment of the present invention, such as qualitative and / or quantitative metrics mentioned herein, may also be applied, mutatis mutandis, to the right and / or leftsingular vectors. Advantages and / or details discussed in the context of, e.g., eigenvalues and eigenvectors may apply also in the context of singular values and / or singular vectors.

[0275] It will be further understood that there might be a plurality of matrices that are used simultaneously for eigenvalue and / or eigenvector calculation.

[0276] The transform module may be configured to determine the at least one eigenvalue metric and / or at least one eigenvector metric.

[0277] The transform module may be configured to determine the at least one singular value metric and / or at least one singular vector metric. The at least one eigenvalue metric may comprise a spectral radius and / or a trace and / or a condition number and / or distributions of the eigenvalues and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0278] It will be understood that the spacings of the eigenvalues and / or the overlap of the eigenvalue spacing can be of arbitrary order.

[0279] The at least one eigenvector metric may comprise a localization metric and / or an entropy of components of the eigenvector and / or a localization length and / or the like.

[0280] For instance, the localization metric may be, without limitation, an inverse participation ratio of the eigenvectors and / or a participation ratio of the eigenvectors. The entropy of components of the eigenvector can be, for example and without limitation, a Shannon entropy, a Structural entropy, and / or a Renyi entropy.

[0281] The transform module may be configured to select a set of the at least one eigenvalue and / or at least one eigenvector based on limits provided by the random matrix theory (RMT).

[0282] In simpler words, the knowledge from RMT may be used to separate information from purely random behavior. RMT may relate to the ideal case for pure random noise without information; any deviation from it may be caused by information within the noise which may allow a first preselection.

[0283] The at least one singular value metric may calculated on the squares of the singular values, which are eigenvalues, and may comprise: a spectral radius and / or a trace and / or condition number and / or distributions of the eigenvalues values and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0284] The at least one singular vector metric may be calculated on the left singular vectors and may comprise: a localization metric and / or an entropy of components of the singular vector and / or a localization length and / or the like. The transform module may be configured to select a set of the at least one singular value and / or at least one singular vector based on limits provided by the random matrix theory (RMT).

[0285] Embodiments of the present invention may be at least in part directed to a system configured to perform voltage noise measurements of a battery cell and to a system configured to analyze the voltage noise measurements of a battery cell at least in part with methods from random matrix theory. This may lead to advantages.

[0286] For example, although prior technologies may be directed to systems configured to perform voltage noise measurements of a battery cell, these systems may not be configured to analyze the voltage noise measurements of a battery cell at least in part with methods from random matrix theory, thus potentially missing, at least in part, the temporal structure of the voltage noise. Noise data analysis for battery voltage noise measurements may be done in prior technologies by systems configured to perform, for example, analysis of frequency spectra, analysis of statistical modes, fractal analysis and / or recurrence analysis. With the exception of system configured to perform recurrence analysis, such systems may use a whole dataset at once to analyze the data, thus potentially ignoring the temporal structure of the noise.

[0287] The analyzing module may be configured to enable an evaluation of the structured data set based on the at least one characteristic descriptor.

[0288] The analyzing module may be configured to extract at least a characteristic descriptor from the structured data set derived at least from the processing window at the first time and from the processing window at the second time.

[0289] In other words, as the processing window may slide, in time, over the measurement signal and comprise different sections of the measurement signal at different times, the system may be configured to perform an analysis on the processing window at each of said different times.

[0290] It will be understood that descriptors may be built from the structured data set of all evaluated processing windows.

[0291] The at least one characteristic descriptor may comprisesone or more statistical moments of the structured data set and / or one or more divergence values for one or more densities in the structured data set and / or one or more residual squared sum values for one or more differences between the structured data set and theoretical values and / or the like. For example, the at least one characteristic descriptor can comprise, without limitation, a mean of the largest eigenvalue for all processing windows in a measurement. As another example, the at least one characteristic descriptor can comprise, without limitation, a standard deviation of the largest eigenvalue for all processing windows in a measurement.

[0292] The establishing module may be configured to use a remaining useful life-time and / or a power handling capability as the performance indicator of the at least one battery cell.

[0293] The establishing module may be configured to perform a sensitivity analysis of the at least one characteristic descriptor with respect to the performance indicator of the at least one battery cell.

[0294] The establishing module may be configured to select a set of the at least one characteristic descriptor based on the sensitivity analysis.

[0295] In other words, the system may look, thanks to the sensitivity analysis, for a high correlation between the characteristic descriptor and the performance indicator (e.g. selfdischarge). As a result, the system may be configured to use characteristic descriptors with high correlation to the performance indicator to build the map, while other characteristic descriptors that may not show high correlation may be discarded by the system.

[0296] The establishing module may be configured to establish the at least one relation based, at least in part, on the sensitivity analysis.

[0297] The establishing module may be configured to perform a statistical analysis of the at least one characteristic descriptor.

[0298] The establishing module may be configured to utilize the statistical analysis to identify clusters of behavior and / or trends and / or tendencies among the at least one characteristic descriptor.

[0299] The establishing module may be configured to establish the at least one relation based, at least in part, on the statistical analysis.

[0300] The at least one relation may be a functional relation.

[0301] The map may coincide with the at least one relation. It will be understood that there can be a plurality of maps that are used simultaneously.

[0302] The at least one equivalent performance indicator and / or one updated performance indicator may be a continuous performance indicator.

[0303] A continuous performance indicator can, for instance, be a self-discharge, since the selfdischarge can have continuous values.

[0304] The at least one equivalent performance indicator and / or one updated performance indicator may be a discrete performance indicator.

[0305] A discrete performance indicator can, for instance, be grades of battery cell quality, since the grading of battery cell quality can comprise discrete grades of battery cell quality.

[0306] The establishing module may be configured to use the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on at least a characteristic descriptor of at least a newly measured battery cell.

[0307] In simpler words, the system may be configured to perform a phase, which may be addressed as "set-up" phase, wherein in the "set-up" phase the system may be configured to measure at least a battery-cell and to build the map based, at least in part, thereon. The system may further be configured to be perform another phase, which may be referred to as "use" phase, wherein in the "use" phase the system may be configured to measure at least a new battery cell and using the map based, at least in part, thereon.

[0308] The establishing module may be configured to improve a battery-cell manufacturing process based at least in part on said the at least one equivalent performance indicator and / or one updated performance indicator.

[0309] As an example, the system may be configured to perform measurements on first battery cells of a first type of battery cells and on second battery cells of a second type of battery cells. The system, according to the present invention, may be configured to evaluate an equivalent performance indicator and / or an updated performance indicator for the first battery cells and for the second battery cells. It may then be possible to see or infer how the first type of battery cells and the second type of battery cells generally performs. This piece of information may be useful to tune the battery-cell manufacturing process in order to make better types of battery cells. In other words, for example, this piece of information may be useful to tune a process parameter of the battery-cell-manufacturing process on the basis on how features of a battery cell, i.e. something defining the type of a cell, influences its performance. The system according to the present invention may therefore be configured to improve a battery-cell manufacturing process.

[0310] In light of the above, the present invention may provide advantages as compared to existing technologies. Conventional battery cell aging processes may take prolonged amounts of time, of the order of day or weeks. Therefore, it may be difficult to provide swift feedback to the process for battery-cell-manufacturing, aimed at improving the process for battery-cell-manufacturing. The present invention may, at least in part, do that. The present invention may therefore, at least in part, limit the generally high scrap-rate of a battery-cell-manufacturing process, especially at ramp-up of the battery-cellmanufacturing process. Besides directly identifying weak cells in the early phases of ageing, the system according to the present invention may have the potential of providing fast feedback for the production process and additional data to predict the lifecycle performance of the cells. It may generally be a goal for cell manufacturers to have access to a system configured to provide a fast feedback loop for process optimization und tuning.

[0311] The system may be a system for assessing the quality of a battery cell in a battery-cellmanufacturing process.

[0312] The system may be a system for assessing the intrinsic activity of a battery cell.

[0313] The measuring module may be configured measure the at least one voltage noise signal in a time from 1 minute to 2 hours, preferably from 5 minutes to 1 hour, more preferably from 10 minutes to 30 minutes.

[0314] The system may be a system for manufacturing at least a battery cell.

[0315] The system may be a system for identifying anomalies in at least a battery cell.

[0316] The system may be a system for identifying at least a cause of anomalies in at least a battery cell.

[0317] The system may be a system for testing at least a battery cell for self-discharge in a battery-cell-manufacturing process.

[0318] The system may be a system for speeding up the cell finishing step in a battery-cellmanufacturing process. The system may be a system for speeding up the aging phase in a battery-cell manufacturing process.

[0319] The aging phase in a battery-cell manufacturing process using the system according to any embodiment of the present invention may last between 10 minutes and 5 days, preferably between 20 minutes and 10 hours, more preferably between 30 minutes and 3 hours.

[0320] The system may be a system for directly identifying at least a weak and / or no good and / or outlier and / or scrap battery cell during the aging phase of a battery-cell-manufacturing process.

[0321] The system may be a system for testing incoming goods inspection for self-discharge and / or anomalies, wherein the incoming goods comprise a battery cell.

[0322] For example, older cells may need to be measured for a possible second life use case. For this application, cells may need to be tested for increased self-discharge and possible anomalies.

[0323] The system may be a system for testing at least a fuel-cell for anomalies.

[0324] Fuel-cells may be difficult to measure due to the complex nature of the electrochemical reactions at their membrane and / or solid electrolyte. Using a system configured for passive measurement technique, according to embodiments of the present invention, may allow a measurement without substance conversion, thus removing this influence.

[0325] The system may be a system for testing a performance indicator of a chemical sensor. Chemical sensor may benefit from a noise-based measurement.

[0326] The system may be a system for improving at least a battery-cell-manufacturing process of at least a battery cell.

[0327] The system may be a system for improving at least a step of a battery-cell-manufacturing process of at least a battery cell.

[0328] The system may be a system for optimizing the costs of at least a battery-cellmanufacturing process of at least a battery cell.

[0329] The system may be a system for increasing the yield of at least a battery-cellmanufacturing process of at least a battery cell. The system may be adapted to carry out the method recited in any of the method embodiments of the present invention.

[0330] The system may be adapted to carry out any given step of the method recited in any of the method embodiments of the present invention.

[0331] The method may comprise carrying out the method according to any method embodiment of the present invention via the system according any of the system embodiment of the present invention.

[0332] In another aspect, the present invention relates to a use of the system according to any of the system embodiments of the present invention.

[0333] The use may be for carrying out the method according to any of the method embodiments of the present invention.

[0334] In another aspect, the present invention relates to a computer program product comprising instructions which, when executed by a processing component, cause the component to carry out the method according to any of the method embodiments of the present invention.

[0335] In another aspect, the present invention relates to a data carried signal carrying said computer program product.

[0336] In another aspect, the present invention relates to a computer-readable medium comprising instructions which, when executed by a processor, cause the computer to carry out the method according to any of the method embodiments of the present invention.

[0337] The present technology is also described by the following numbered embodiments.

[0338] Below, method embodiments are presented. Method embodiments are abbreviated by the letter "M" followed by a number. Whenever reference is made herein to "method embodiments", these embodiments are meant.

[0339] Ml. A method comprising: measuring at least a voltage noise signal of at least a battery cell, processing the at least one voltage noise signal into a structured data set, analyzing the structured data set and extracting at least a characteristic descriptor from the structured data set, establishing at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell.

[0340] M2. The method according to the preceding embodiment, wherein the method comprises passively measuring at least a voltage noise signal of at least a battery cell.

[0341] M3. The method according to any of the preceding embodiments, wherein the step of processing the at least one voltage noise signal into a structured data set comprises determining at least a matrix based at least in part on the at least one voltage noise signal.

[0342] M4. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the step of processing the at least one voltage noise signal into a structured data set comprises determining eigenvalues and / or eigenvectors of the at least one matrix.

[0343] M5. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the step of processing the at least one voltage noise signal into a structured data set comprises determining singular values and / or singular vectors of the at least one matrix.

[0344] M6. The method according to any of the preceding embodiments, with the features of embodiment M4, wherein the structured data set comprises at least an eigenvalue metric of said eigenvalues and / or at least an eigenvector metric of said eigenvectors.

[0345] M7. The method according to any of the preceding embodiments, with the features of embodiment M5, wherein the structured data set comprises at least a singular value metric of said singular values and / or at least a singular vector metric of said singular vectors.

[0346] M8. The method according to any of the preceding embodiments, wherein the method comprises using a characteristic quantitative descriptor.

[0347] M9. The method according to any of the preceding embodiments, wherein the method comprises using a characteristic qualitative descriptor. MIO. The method according to any of the preceding embodiments, wherein the method comprises using a statistical descriptor as the characteristic descriptor.

[0348] Mil. The method according to any of the preceding embodiments, wherein the method comprises using a quality metric of the at least one battery cell as the performance indicator of the at least one battery cell.

[0349] M12. The method according to any of the preceding embodiments, wherein the method comprises using a self-discharge-like metric as the performance indicator of the at least one battery cell.

[0350] M13. The method according to any of the preceding embodiments, wherein the method comprises building at least a map based at least in part on the at least one relation between the at least one characteristic descriptor and at least a performance indicator.

[0351] M14. The method according to any of the preceding embodiments, with the features of embodiment M13, wherein the map links the at least one characteristic descriptor to at least an equivalent performance indicator and / or an updated performance indicator.

[0352] M15. The method according to any of the preceding embodiments, with the features of embodiment M13, wherein the method comprises using the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on the at least one characteristic descriptor.

[0353] M16. The method according to any of the preceding embodiments, wherein the method comprises: measuring at least a voltage noise signal of at least a battery cell with a measuring module, processing the at least one voltage noise signal into a structured data set with a processing module, analyzing the structured data set and extracting at least a characteristic descriptor from the structured data set with an analyzing module, establishing at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell with an establishing module. M17. The method according to any of the preceding embodiments, wherein the method comprises using a data and control interface module.

[0354] M18. The method according to any of the preceding embodiments, with the features of embodiment M16, wherein the method comprises using a transform module, comprised in the processing module.

[0355] M19. The method according to any of the preceding embodiments, with the features of embodiment M16, wherein the method comprises communicating between the measuring module and the processing module.

[0356] M20. The method according to any of the preceding embodiments, with the features of embodiment M16, wherein the method comprises communicating between the measuring module and the processing module at least the at least one measured voltage noise signal.

[0357] M21. The method according to any of the preceding embodiments, with the features of embodiments M16 and M18, wherein the method comprises communicating between the processing module and the transform module.

[0358] M22. The method according to any of the preceding embodiments, with the features of embodiment M16 and M18, wherein the method comprises communicating between the transform module and the analyzing module.

[0359] M23. The method according to any of the preceding embodiments, with the features of embodiment M16, wherein the method comprises communicating between the analyzing module and the establishing module.

[0360] M24. The method according to any of the preceding embodiments, with the features of embodiment M16, wherein the method comprises communicating between the analyzing module and the establishing module at least the at least one characteristic descriptor.

[0361] M25. The method according to any of the preceding embodiments, with the features of embodiment M17, wherein the method comprises communicating between the data and control interface module and any combination of modules configured to perform the method according to any of the preceding embodiments, said combination of modules comprising at least one module. M26. The method according to any of the preceding embodiments, with the features of embodiment M17, wherein the method comprises changing, with the data and control interface module, at least one parameter related to any combination of steps of the method, said combination of steps comprising at least a step of the method.

[0362] M27. The method according to any of the preceding embodiments, with the features of embodiment M17, wherein the method comprises communicating, with the data and control interface module, to a user at least any combination between: the at least one voltage noise signal of at the least one battery cell, the structured data set, the at least one characteristic descriptor, the at least one relation between the at least one characteristic descriptor, and the at least one performance indicator of the at least one battery cell.

[0363] M28. The method according to any of the preceding embodiments, with the features of embodiments M16 and M18, wherein the method comprises arranging locally or remotely one or more among the transform module, processing module, the analyzing module and the establishing module.

[0364] M29. The method according to any of the preceding embodiments, wherein the method comprises utilizing a module comprising a combination of any of the modules configured to perform the method according to any of the preceding embodiments.

[0365] M30. The method according to any of the preceding embodiments, wherein the method comprises utilizing different software for different purposes within a module according to any of the preceding embodiments.

[0366] M31. The method according to any of the preceding embodiments, wherein the method comprises using a computing device.

[0367] M32. The method according to any of the preceding embodiments, with the features of embodiment M31, wherein any module according to any of the preceding embodiments comprises the computing device.

[0368] M33. The method according to any of the preceding embodiments, with the features of embodiment M17, wherein the method comprises communicating between the data and control interface module and a manufacturing execution system (MES) of a battery manufacturing plant. M34. The method according to any of the preceding embodiments, wherein the method comprises using a high-resolution AC voltage detecting set-up.

[0369] M35. The method according to any of the preceding embodiments, with the features of embodiment M34, wherein the method comprises using a card of the high- resolution AC voltage detecting set-up to measure a voltage noise signal from the at least one battery cell.

[0370] M36. The method according to any of the preceding embodiments, with the features of embodiment M34, wherein the method comprises using multiple cards of the high- resolution AC voltage detecting set-up to measure a plurality of voltage noise signal from a plurality of battery cells, in parallel.

[0371] M37. The method according to any of the preceding embodiments, with the features of embodiment M34, wherein the method comprises using the high-resolution AC voltage detecting set-up is with a resolution below 10 nV, preferably in a range between 0.1 nV and 10 nV, preferably below 0.1 nV.

[0372] M38. The method according to any of the preceding embodiments, with the features of embodiment M35, wherein the method comprises using an AC-coupling device, embedded in the card of the high-resolution AC voltage detecting set-up.

[0373] M39. The method according to any of the preceding embodiments, with the features of embodiment M38, wherein the method comprises using a coupling capacitor, comprised in the AC-coupling device.

[0374] M40. The method according to any of the preceding embodiments, with the features of embodiment M38, wherein the method comprises using an AC-coupling device with a lower cut-off frequency in a range between 10 Hz and 50 Hz, preferably in a range between 1 Hz and 10 Hz, more preferably below 1 Hz.

[0375] M41. The method according to any of the preceding embodiments, with the features of embodiment M35, wherein the method comprises utilizing a high-gain low-noise amplifier embedded in the card of the high-resolution AC voltage detecting setup.

[0376] M42. The method according to any of the preceding embodiments, with the features of embodiment M41, wherein the method comprises using the high-gain low-noise amplifier with a gain in a range between 100 and 2000, preferably in a range between 2000- and 10000, most preferably above 10000.

[0377] M43. The method according to any of the preceding embodiments, with the features of embodiment M41, wherein the method comprises using the high-gain low-noise amplifier with a noise figure at 1 Hz below 50 nV / sqrt(Hz), preferably below 20 nV / sqrt(Hz), more preferably below 5 nV / sqrt(Hz).

[0378] M44. The method according to any of the preceding embodiments, with the features of embodiment M35, wherein the method comprises utilizing a high-resolution analog-to-digital (ADC) converter embedded in the card of the high-resolution AC voltage detecting set-up.

[0379] M45. The method according to any of the preceding embodiments, with the features of embodiment M44, wherein the method comprises utilizing the ADC with a sampling rate in a range between 1 Sa / s and 100 MSa / s, preferably between 100 Sa / s and 1 MSa / s, most preferably between 0.5 kSa / s and 30 kSa / s.

[0380] M46. The method according to any of the preceding embodiments, with the features of embodiment M44, wherein the method comprises utilizing the ADC with a resolution greater than or equal to 16 bit, preferably greater than or equal to 20 bit, more preferably greater than or equal to 24 bit.

[0381] M47. The method according to any of the preceding embodiments, with the features of embodiment M35, wherein the method comprises utilizing a control and communication unit, comprised in the card of the high-resolution AC voltage detecting set-up.

[0382] M48. The method according to any of the preceding embodiments, wherein the method comprises isolating an alternate coupling (AC) voltage noise signal of the at least one battery cell from the voltage signal of at the at least one battery, said voltage signal comprising an AC component and a direct coupling (DC) component.

[0383] M49. The method according to any of the preceding embodiments, with the features of embodiment M41 and M48, wherein the method comprises amplifying the AC voltage noise signal into an amplified voltage noise signal. M50. The method according to any of the preceding embodiments, with the features of embodiment M44 and M49, wherein the method comprises sampling the amplified voltage noise signal and obtaining a digitized voltage noise signal.

[0384] M51. The method according to any of the preceding embodiments, with the features of embodiment M47, M50 and M17, wherein the method comprises communicating the digitized voltage noise signal to the data and control interface module.

[0385] M52. The method according to any of the preceding embodiments, with the features of embodiment M47, M50 and M17, wherein the method comprises communicating, via ethernet, the digitized voltage noise signal to the data and control interface module.

[0386] M53. The method according to any of the preceding embodiments, with the features of embodiment M44 and M47, wherein the method comprises controlling the gain and / or the sampling rate of the high-resolution ADC with the control and communication unit.

[0387] M54. The method according to any of the preceding embodiments, with the features of embodiment M39 and M47, wherein the method comprises controlling a polarization compensation of the coupling capacitor with the control and communication unit.

[0388] M55. The method according to any of the preceding embodiments, with the features of embodiment M47, wherein the method comprises enabling at least a temperature measurement of the at least one battery cell with the control and communication unit.

[0389] M56. The method according to any of the preceding embodiments, with the features of embodiment M47, wherein the method comprises enabling at least a measurement of the at least one voltage noise signal.

[0390] M57. The method according to any of the preceding embodiments, wherein the method comprises pre-processing the at least one voltage noise signal into at least one pre-processed voltage noise signal.

[0391] M58. The method according to any of the preceding embodiments, with the features of embodiment M57, wherein the pre-processing comprises cleaning the at least one voltage noise signal from artifacts. M59. The method according to any of the preceding embodiments, with the features of embodiment M57, wherein the pre-processing comprises detrending the at least one voltage noise signal.

[0392] M60. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the method comprises generating the at least one matrix from the at least one voltage noise signal.

[0393] M61. The method according to any of the preceding embodiments, with the features of embodiment M3 and M57, wherein the method comprises generating the at least one matrix from the at least one pre-processed voltage noise signal.

[0394] M62. The method according to any of the preceding embodiments, with the features of embodiment M57, wherein the method comprises establishing a measurement signal, wherein the measurement signal comprises a section of the pre-processed voltage noise signal.

[0395] M63. The method according to any of the preceding embodiments, with the features of embodiment M62, wherein the section of the pre-processed voltage noise signal has a temporal duration between 1 minute and 240 minutes, preferably between 5 minutes and 60 minutes, more preferably between 5 minutes and 15 minutes.

[0396] M64. The method according to any of the preceding embodiments, with the features of embodiment M62, wherein the method comprises establishing at least a processing window, wherein the processing window comprises a section of the measurement signal.

[0397] M65. The method according to any of the preceding embodiments, with the features of embodiment M64, wherein the method comprises sliding, in time, the at one least processing window over the measurement signal.

[0398] M66. The method according to any of the preceding embodiments, with the features of embodiment M65, wherein the section of the measurement signal, comprised in the processing window at a first time, overlaps in time with the section of the measurement signal, comprised in the processing window at a second time.

[0399] M67. The method according to any of the preceding embodiments, with the features of embodiment M66, wherein said overlap is in a range between 70% and 99.9%, preferably in a range between 90% and 99%, more preferably in a range between 97% and 99% of a temporal length of the processing window.

[0400] M68. The method according to any of the preceding embodiments, with the features of embodiment M66, wherein the method comprises using said overlap as a parameter.

[0401] M69. The method according to any of the preceding embodiments, with the features of embodiment M64, wherein the method comprises using a temporal length of the processing window as a parameter.

[0402] M70. The method according to any of the preceding embodiments, with the features of embodiment M64, wherein a temporal length of the processing window is in a range between 30 s and 360 s, preferably between 60 s and 240 s, more preferably between 120 s and 180 s.

[0403] M71. The method according to any of the preceding embodiments, with the features of embodiment M64, wherein the method comprises establishing non-overlapping segments.

[0404] M72. The method according to any of the preceding embodiments, with the features of embodiment M71, wherein each of the non-overlapping segments comprises a non-overlapping section of the at least one processing window.

[0405] M73. The method according to any of the preceding embodiments, with the features of embodiment M71, wherein the non-overlapping segments are uniform.

[0406] M74. The method according to any of the preceding embodiments, with the features of embodiment M71, wherein the wherein the method comprises using the number of the non-overlapping segments as a parameter.

[0407] M75. The method according to any of the preceding embodiments, with the features of embodiments M3 and M71, wherein the method comprises deriving the at least one matrix from the non-overlapping segments.

[0408] M76. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the at least one matrix comprises a Pearson correlation matrix. M77. The method according to any of the preceding embodiments, with the features of embodiment M76, wherein the method comprises deriving the Pearson correlation matrix from a measurement matrix.

[0409] M78. The method according to any of the preceding embodiments, with the features of embodiment M71 and M77, wherein the method comprises deriving the measurement matrix from the non-overlapping segments.

[0410] M79. The method according to any of the preceding embodiments, with the features of embodiment M71 and M77, wherein the method comprises generating the entries of a row of the measurement matrix from one of the non-overlapping segments.

[0411] M80. The method according to any of the preceding embodiments, with the features of embodiment M77, wherein the method comprises generating the Pearson correlation matrix based on the following formula: C = AAT / (n-l) wherein C represents the Pearson correlation matrix, wherein A represents the measurement matrix, n is the number of columns of the measurement matrix, wherein each element in a row of A has the mean of said row of A removed, and wherein each element in a row of A is divided by the standard deviation of said row of A.

[0412] M81. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the at least one matrix comprises a cross-correlation matrix.

[0413] M82. The method according to any of the preceding embodiments, with the features of embodiment M71 and M81, wherein the method comprises deriving the crosscorrelation matrix from the non-overlapping segments.

[0414] M83. The method according to any of the preceding embodiments, with the features of embodiment M71 and M81, wherein the method comprises generating the entries of the cross-correlation matrix based on the absolute cross correlation function between any of the uniform non-overlapping segments.

[0415] M84. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the at least one matrix comprises a mutual information matrix. M85. The method according to any of the preceding embodiments, with the features of embodiment M71 and M84, wherein the method comprises deriving the mutual information matrix from the non-overlapping segments.

[0416] M86. The method according to any of the preceding embodiments, with the features of embodiment M71 and M84, wherein the method comprises generating the entries of the mutual information matrix from a mutual information between any of the uniform non-overlapping segments.

[0417] M87. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the at least one matrix comprises a square matrix.

[0418] M88. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the at least one matrix comprises a matrix-shaped representation of the at least one voltage noise signal.

[0419] M89. The method according to any of the preceding embodiments, with the features of embodiments M76, M81, M84, and M88, wherein the method comprises determining eigenvalues and / or eigenvectors of at least one among: the Pearson correlation matrix, the cross-correlation matrix, the mutual information matrix, the matrix-shaped representation of the at least one voltage noise signal.

[0420] M90. The method according to any of the preceding embodiments, with the features of embodiment M3, wherein the at least one matrix comprises a measurement matrix.

[0421] M91. The method according to any of the preceding embodiments, with the features of embodiment M71 and M90, wherein the method comprises deriving the measurement matrix from the non-overlapping segments.

[0422] M92. The method according to any of the preceding embodiments, with the features of embodiment M6 and M18, wherein the method comprises determining the at least one eigenvalue metric and / or at least one eigenvector metric with the transform module.

[0423] M93. The method according to any of the preceding embodiments, with the features of embodiment M7 and M18, wherein the method comprises determining the at least one singular value metric and / or at least one singular vector metric with the transform module. M94. The method according to any of the preceding embodiments, with the features of embodiment M6, wherein the at least one eigenvalue metric comprises a spectral radius and / or a trace and / or a condition number and / or distributions of the eigenvalues and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0424] M95. The method according to any of the preceding embodiments, with the features of embodiment M6, wherein the at least one eigenvector metric comprises a localization metric and / or an entropy of components of the eigenvector and / or a localization length and / or the like.

[0425] M96. The method according to any of the preceding embodiments, with the features of embodiment M6, wherein the method comprises selecting a set of the at least one eigenvalue and / or at least one eigenvector based on limits provided by the random matrix theory (RMT).

[0426] M97. The method according to any of the preceding embodiments, with the features of embodiment M7, wherein the at least one singular value metric is calculated on the squares of the singular values, which are eigenvalues, and comprises: a spectral radius and / or a trace and / or condition number and / or distributions of the eigenvalues values and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0427] M98. The method according to any of the preceding embodiments, with the features of embodiment M7, wherein the at least one singular vector metric is calculated on the left singular vectors and comprises: a localization metric and / or an entropy of components of the singular vector and / or a localization length and / or the like.

[0428] M99. The method according to any of the preceding embodiments, with the features of embodiment M7, wherein the method comprises selecting a set of the at least one singular value and / or at least one singular vector based on limits provided by the random matrix theory (RMT). M100. The method according to any of the preceding embodiments, wherein the method comprises enabling an evaluation of the structured data set based on the at least one characteristic descriptor.

[0429] M101. The method according to any of the preceding embodiments, with the features of embodiment M66, wherein the method comprises extracting at least a characteristic descriptor from the structured data set derived at least from the processing window at the first time and from the processing window at the second time.

[0430] M102. The method according to any of the preceding embodiments, wherein the at least one characteristic descriptor comprises one or more statistical moments of the structured data set and / or one or more divergence values for one or more densities in the structured data set and / or one or more residual squared sum values for one or more differences between the structured data set and theoretical values and / or the like.

[0431] M103. The method according to any of the preceding embodiments, wherein the method comprises using a remaining useful life-time and / or a power handling capability as the performance indicator of the at least one battery cell.

[0432] M104. The method according to any of the preceding embodiments, wherein the method comprises performing a sensitivity analysis of the at least one characteristic descriptor with respect to the performance indicator of the at least one battery cell.

[0433] M105. The method according to any of the preceding embodiments, with the features of embodiment M104, wherein the method comprises selecting a set of the at least one characteristic descriptor based on the sensitivity analysis.

[0434] M106. The method according to any of the preceding embodiments, with the features of embodiment M104, wherein the method comprises establishing the at least one relation based, at least in part, on the sensitivity analysis.

[0435] M107. The method according to any of the preceding embodiments, wherein the method comprises performing a statistical analysis of the at least one characteristic descriptor. M108. The method according to any of the preceding embodiments, with the features of embodiment M107, wherein the method comprises utilizing the statistical analysis to identify clusters of behavior and / or trends and / or tendencies among the at least one characteristic descriptor.

[0436] M109. The method according to any of the preceding embodiments, with the features of embodiment M107, wherein the method comprises establishing the at least one relation based, at least in part, on the statistical analysis.

[0437] MHO. The method according to any of the preceding embodiments, wherein the at least one relation is a functional relation.

[0438] Mill. The method according to any of the preceding embodiments, with the features of embodiment M15, wherein the map coincides with the at least one relation.

[0439] M112. The method according to any of the preceding embodiments, with the features of embodiment M15, wherein the at least one equivalent performance indicator and / or one updated performance indicator is a continuous performance indicator.

[0440] M113. The method according to any of the preceding embodiments, with the features of embodiment M15, wherein the at least one equivalent performance indicator and / or one updated performance indicator is a discrete performance indicator.

[0441] Ml 14. The method according to any of the preceding embodiments, with the features of embodiment M15, wherein the method comprises using the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on at least a characteristic descriptor of at least a newly measured battery cell.

[0442] M115. The method according to any of the preceding embodiments, with the features of embodiment M15, wherein the method comprises improving a battery-cell manufacturing process based at least in part on said the at least one equivalent performance indicator and / or one updated performance indicator.

[0443] M116. The method according to the preceding embodiment, wherein the method is a method for assessing the quality of a battery cell in a battery-cell-manufacturing process. M117. The method according to the preceding embodiment, wherein the method is a method for assessing the intrinsic activity of a battery cell.

[0444] M118. The method according to any of the preceding embodiments, wherein the step of measuring at least a voltage noise signal of at least a battery cell takes a range of time from 1 hour to 2 hours, preferably from 30 minutes to 1 hour, more preferably from 10 minutes to 30 minutes.

[0445] M119. The method according to any of the preceding embodiments, wherein the step of measuring at least a voltage noise signal of at least a battery cell takes a range of time from 1 minute to 2 hours, preferably from 5 minutes to 1 hour, more preferably from 10 minutes to 30 minutes.

[0446] M120. The method according to any of the preceding embodiments, wherein the method is a method for manufacturing at least a battery cell.

[0447] M121. The method according to any of the preceding embodiments, wherein the method is a method for identifying anomalies in at least a battery cell.

[0448] M122. The method according to any of the preceding embodiments, wherein the method is a method for identifying at least a cause of anomalies in at least a battery cell.

[0449] M123. The method according to any of the preceding embodiments, wherein the method is a method for testing at least a battery cell for self-discharge in a battery-cellmanufacturing process.

[0450] M124. The method according to any of the preceding embodiments, wherein the method is a method for speeding up the cell finishing step in a battery-cell-manufacturing process.

[0451] M125. The method according to any of the preceding embodiments, wherein the method is a method for speeding up the aging phase in a battery-cell manufacturing process.

[0452] M126. The method according to any of the preceding embodiments, wherein the aging phase in a battery-cell manufacturing process using the method according to any of the preceding embodiments lasts between 10 minutes and 5 days, preferably between 20 minutes and 10 hours, more preferably between 30 minutes and 3 hours. M127. The method according to any of the preceding embodiments, wherein the method is a method for directly identifying at least a weak and / or no good and / or outlier and / or scrap battery cell during the aging phase of a battery-cell-manufacturing process.

[0453] M128. The method according to any of the preceding embodiments, wherein the method is a method for testing incoming goods inspection for self-discharge and / or anomalies, wherein the incoming goods comprise a battery cell.

[0454] M129. The method according to any of the preceding embodiments, wherein the method is a method for testing at least a fuel-cell for anomalies.

[0455] M130. The method according to any of the preceding embodiments, wherein the method is a method for testing a performance indicator of a chemical sensor.

[0456] M131. The method according to any of the preceding embodiments, wherein the method is a method for improving at least a battery-cell-manufacturing process of at least a battery cell.

[0457] M132. The method according to any of the preceding embodiments, wherein the method is a method for improving at least a step of a battery-cell-manufacturing process of at least a battery cell.

[0458] M133. The method according to any of the preceding embodiments, wherein the method is a method for optimizing the costs of at least a battery-cell-manufacturing process of at least a battery cell.

[0459] M134. The method according to any of the preceding embodiments, wherein the method is a method for increasing the yield of at least a battery-cell-manufacturing process of at least a battery cell.

[0460] Below, system embodiments are presented. System embodiments are abbreviated by the letter "S" followed by a number. Whenever reference is made herein to "system embodiments", these embodiments are meant.

[0461] SI. A system comprising: at least a measuring module configured to measure at least a voltage noise signal of at least a battery cell, at least a processing module configured to process the at least one voltage noise signal into a structured data set, at least an analyzing module configured to analyze the structured data set and to extract at least a characteristic descriptor from the structured data set, at least an establishing module configured to establish at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell.

[0462] 52. The system according to the preceding system embodiment, wherein the measuring module is configured to passively measure at least a voltage noise signal of at least a battery cell.

[0463] 53. The system according to any of the preceding system embodiments, wherein the processing module is configured to determine at least a matrix based at least in part on the at least one voltage noise signal.

[0464] 54. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the processing module is configured to determine eigenvalues and / or eigenvectors of the at least one matrix.

[0465] 55. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the processing module is configured to determine singular values and / or singular vectors of the at least one matrix.

[0466] 56. The system according to any of the preceding system embodiments, with the features of embodiment S4, wherein the structured data set comprises at least an eigenvalue metric of said eigenvalues and / or at least an eigenvector metric of said eigenvectors.

[0467] 57. The system according to any of the preceding system embodiments, with the features of embodiment S5, wherein the structured data set comprises at least a singular value metric of said singular values and / or at least a singular vector metric of said singular vectors.

[0468] 58. The system according to any of the preceding system embodiments, wherein the analyzing module is configured to use a characteristic quantitative descriptor.

[0469] 59. The system according to any of the preceding system embodiments, wherein the analyzing module is configured to use a characteristic qualitative descriptor. 510. The system according to any of the preceding system embodiments, wherein the analyzing module is configured to use a statistical descriptor as the characteristic descriptor.

[0470] 511. The system according to any of the preceding system embodiments, wherein the establishing module is configured to use a quality metric of the at least one battery cell as the performance indicator of the at least one battery cell.

[0471] 512. The system according to any of the preceding system embodiments, wherein the establishing module is configured to use a self-discharge-like metric as the performance indicator of the at least one battery cell.

[0472] 513. The system according to any of the preceding system embodiments, wherein the establishing module is configured to build at least a map based at least in part on the at least one relation between the at least one characteristic descriptor and at least a performance indicator.

[0473] 514. The system according to any of the preceding system embodiments, with the features of embodiment S13, wherein the map links the at least one characteristic descriptor to at least an equivalent performance indicator and / or an updated performance indicator

[0474] 515. The system according to any of the preceding system embodiments, with the features of embodiment S13, wherein the establishing module is configured to use the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on the at least one characteristic descriptor.

[0475] 516. The system according to any of the preceding system embodiments, wherein the system comprises a data and control interface module.

[0476] 517. The system according to any of the preceding system embodiments, wherein the processing module comprises a transform module.

[0477] 518. The system according to any of the preceding system embodiments, wherein the measuring module is configured to communicate with the processing module. 519. The system according to any of the preceding system embodiments wherein the measuring module is configured to communicate to the processing module at least the at least one measured voltage noise signal.

[0478] 520. The system according to any of the preceding system embodiments, with the features of embodiment S17, wherein the processing module is configured to communicate with the transform module.

[0479] 521. The system according to any of the preceding system embodiments, with the features of embodiment S17, wherein the transform module is configured to communicate with the analyzing module.

[0480] 522. The system according to any of the preceding system embodiments, wherein the analyzing module is configured to communicate with the establishing module.

[0481] 523. The system according to any of the preceding system embodiments, wherein the analyzing module is configured to communicate to the establishing module at least the at least one characteristic descriptor.

[0482] 524. The system according to any of the preceding system embodiments, wherein the data and control interface module is configured to communicate with any combination of modules according to any of the preceding system embodiments, said combination of modules comprising at least one module.

[0483] 525. The system according to any of the preceding system embodiments, with the features of embodiment S16, wherein the data and control interface module is configured to change at least one parameter related to any combination of modules according to any of the preceding system embodiments, said combination of modules comprising at least one module.

[0484] 526. The system according to any of the preceding system embodiments, with the features of embodiment S16, wherein the data and control interface module is configured to communicate to a user at least any combination between: the at least one voltage noise signal of at the least one battery cell, the structured data set, the at least one characteristic descriptor, the at least one relation between the at least one characteristic descriptor, and the at least one performance indicator of the at least one battery cell. 527. The system according to any of the preceding system embodiments, with the features of embodiment S17, wherein one or more among the transform module, processing module, the analyzing module and the establishing module may be arranged locally or remotely.

[0485] 528. The system according to any of the preceding system embodiments, wherein the system comprises a module comprising a combination of any of the modules according to any of the preceding system embodiments.

[0486] 529. The system according to any of the preceding system embodiments, wherein the system is configured to utilize different software for different purposes within a module according to any of the preceding system embodiments.

[0487] 530. The system according to any of the preceding system embodiments, wherein any module according to any of the preceding embodiments comprises a computing device.

[0488] 531. The system according to any of the preceding system embodiments, with the features of embodiment S16, wherein the data and control interface module is configured to communicate with a manufacturing execution system (MES) of a battery manufacturing plant.

[0489] 532. The system according to any of the preceding system embodiments, wherein the measuring module comprises a high-resolution AC voltage detecting set-up.

[0490] 533. The system according to any of the preceding system embodiments, with the features of embodiment S32, wherein the high-resolution AC voltage detecting set-up comprises at least a card of the high-resolution AC voltage detecting setup, and wherein said card is configured to measure a voltage noise signal from the at least one battery cell.

[0491] 534. The system according to any of the preceding system embodiments, with the features of embodiment S32, wherein the high-resolution AC voltage detecting set-up comprises multiple cards of the high-resolution AC voltage detecting setup, and wherein said multiple cards are configured to measure a plurality of voltage noise signal from a plurality of battery cells, in parallel.

[0492] 535. The system according to any of the preceding system embodiments, with the features of embodiment S32, wherein the high-resolution AC voltage detecting set-up has a resolution below 10 nV, preferably in a range between 0.1 nV and 10 nV, preferably below 0.1 nV.

[0493] 536. The system according to any of the preceding system embodiments, with the features of embodiment S33, wherein the at least one card comprises an AC- coupling device.

[0494] 537. The system according to any of the preceding system embodiments, with the features of embodiment S36, wherein the AC-coupling device comprises a coupling capacitor.

[0495] 538. The system according to any of the preceding system embodiments, with the features of embodiment S36, wherein the AC-coupling device has a lower cut-off frequency in a range between 10 Hz and 50 Hz, preferably in a range between 1 Hz and 10 Hz, more preferably below 1 Hz.

[0496] 539. The system according to any of the preceding system embodiments, with the features of embodiment S33, wherein the at least one card comprises a high-gain low-noise amplifier.

[0497] 540. The system according to any of the preceding system embodiments, with the features of embodiment S39, wherein the high-gain low-noise amplifier has a gain in a range between 100 and 2000, preferably in a range between 2000- and 10000, most preferably above 10000.

[0498] 541. The system according to any of the preceding system embodiments, with the features of embodiment S39, wherein the high-gain low-noise amplifier has a noise figure at 1 Hz below 50 nV / sqrt(Hz), preferably below 20 nV / sqrt(Hz), more preferably below 5 nV / sqrt(Hz).

[0499] 542. The system according to any of the preceding system embodiments, with the features of embodiment S33, wherein the at least one card comprises a high- resolution analog-to-digital (ADC) converter.

[0500] 543. The system according to any of the preceding system embodiments, with the features of embodiment S42, wherein the high-resolution ADC has a sampling rate in a range between 1 Sa / s and 100 MSa / s, preferably between 100 Sa / s and 1 MSa / s, most preferably between 0.5 kSa / s and 30 kSa / s. 544. The system according to any of the preceding system embodiments, with the features of embodiment S42, wherein the high-resolution ADC has a resolution greater than or equal to 16 bit, preferably greater than or equal to 20 bit, more preferably greater than or equal to 24 bit.

[0501] 545. The system according to any of the preceding system embodiments, with the features of embodiment S33, wherein the at least one card comprises a control and communication unit.

[0502] 546. The system according to any of the preceding system embodiments, with the features of embodiment S36, wherein the AC coupling device is configured to isolate an alternate coupling (AC) voltage noise signal of the at least one battery cell from a voltage signal of at the at least one battery, said voltage signal comprising an AC component and a direct coupling (DC) component.

[0503] 547. The system according to any of the preceding system embodiments, with the features of embodiments S39 and S46, wherein the high-gain low-noise amplifier comprises is configured to amplify the AC voltage noise signal into an amplified voltage noise signal.

[0504] 548. The system according to any of the preceding system embodiments, with the features of embodiments S42 and S47, wherein the high-resolution ADC is configured to sample the amplified voltage noise signal and obtain a digitized voltage noise signal.

[0505] 549. The system according to any of the preceding system embodiments, with the features of embodiments S45, S48 and S16, wherein the measuring module is configured to communicate the digitized voltage noise signal to the data and control interface module.

[0506] 550. The system according to any of the preceding system embodiments, with the features of embodiments S45, S48 and S16, wherein the measuring module is configured to communicate, via ethernet, the digitized voltage noise signal to the data and control interface module.

[0507] 551. The system according to any of the preceding system embodiments, with the features of embodiments S42 and S45, wherein the control and communication unit is configured to control the gain and / or the sampling rate of the high- resolution ADC. 552. The system according to any of the preceding system embodiments, with the features of embodiments S37 and S45, wherein the control and communication unit is configured to control a polarization compensation of the coupling capacitor.

[0508] 553. The system according to any of the preceding system embodiments, with the features of embodiment S45, wherein the control and communication unit is configured to enable at least a temperature measurement of the at least one battery cell with the control and communication unit.

[0509] 554. The system according to any of the preceding system embodiments, with the features of embodiment S45, wherein the control and communication unit is configured to enable at least a measurement of the at least one voltage noise signal.

[0510] 555. The system according to any of the preceding system embodiments, wherein the processing module is configured to pre-process the at least one voltage noise signal into at least one pre-processed voltage noise signal.

[0511] 556. The system according to any of the preceding system embodiments, with the features of embodiment S55, wherein the processing module is configured to pre- process the at least one voltage noise signal into at least one pre-processed voltage noise signal by cleaning the at least one voltage noise signal from artifacts.

[0512] 557. The system according to any of the preceding system embodiments, with the features of embodiment S55, wherein the processing module is configured to pre- process the at least one voltage noise signal into at least one pre-processed voltage noise signal by detrending the at least one voltage noise signal.

[0513] 558. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the processing module is configured to generate the at least one matrix from the at least one voltage noise signal.

[0514] 559. The system according to any of the preceding system embodiments, with the features of embodiments S3 and S55, wherein the processing module is configured to generate the at least one matrix from the at least one pre-processed voltage noise signal. The system according to any of the preceding system embodiments, with the features of embodiment S55, wherein the processing module is configured to establish a measurement signal, wherein the measurement signal comprises a section of the pre-processed voltage noise signal. The system according to any of the preceding system embodiments, with the features of embodiment S60, wherein the section of the pre-processed voltage noise signal has a temporal duration between 1 minute and 240 minutes, preferably between 5 minutes and 60 minutes, more preferably between 5 minutes and 15 minutes. The system according to any of the preceding system embodiments, with the features of embodiment S60, wherein the processing module is configured to establish at least a processing window, wherein the processing window comprises a section of the measurement signal. The system according to any of the preceding system embodiments, with the features of embodiment S62, wherein the processing module is configured to slide, in time, the at one least processing window over the measurement signal. The system according to any of the preceding system embodiments, with the features of embodiment S63, wherein the section of the measurement signal, comprised in the processing window at a first time, overlaps in time with the section of the measurement signal, comprised in the processing window at a second time. The system according to any of the preceding system embodiments, with the features of embodiment S64, wherein said overlap is in a range between 70% and 99.9%, preferably in a range between 90% and 99%, more preferably in a range between 97% and 99% of a temporal length of the processing window. The system according to any of the preceding system embodiments, with the features of embodiment S64, wherein the processing module is configured to use said overlap as a parameter. The system according to any of the preceding system embodiments, with the features of embodiment S62, wherein the processing module is configured to use a temporal length of the processing window as a parameter. 568. The system according to any of the preceding system embodiments, with the features of embodiment S62, wherein a temporal length of the processing window is in a range between 30 s and 360 s, preferably between 60 s and 240 s, more preferably between 120 s and 180 s.

[0515] 569. The system according to any of the preceding system embodiments, with the features of embodiment S62, wherein the processing module is configured to establish non-overlapping segments.

[0516] 570. The system according to any of the preceding system embodiments, with the features of embodiment S69, wherein each of the non-overlapping segments comprises a non-overlapping section of the at least one processing window.

[0517] 571. The system according to any of the preceding system embodiments, with the features of embodiment S69, wherein the non-overlapping segments are uniform.

[0518] 572. The system according to any of the preceding system embodiments, with the features of embodiment S69, wherein the processing module is configured to use the number of the non-overlapping segments as a parameter.

[0519] 573. The system according to any of the preceding system embodiments, with the features of embodiments S3 and S69, wherein the processing module is configured to derive the at least one matrix from the non-overlapping segments.

[0520] 574. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the at least one matrix comprises a Pearson correlation matrix.

[0521] 575. The system according to any of the preceding system embodiments, with the features of embodiment S74, wherein the processing module is configured to derive the Pearson correlation matrix from a measurement matrix.

[0522] 576. The system according to any of the preceding system embodiments, with the features of embodiments S69 and S75, wherein the processing module is configured to derive the measurement matrix from the non-overlapping segments.

[0523] 577. The system according to any of the preceding system embodiments, with the features of embodiments S69 and S75, wherein the processing module is configured to generate the entries of a row of the measurement matrix from one of the non-overlapping segments. The system according to any of the preceding system embodiments, with the features of embodiment S75, wherein the processing module is configured to generate the Pearson correlation matrix based on the following formula :

[0524] C = AAT / (n-l) wherein C represents the Pearson correlation matrix, wherein A represents the measurement matrix, n is the number of columns of the measurement matrix, wherein each element in a row of A has the mean of said row of A removed, and wherein each element in a row of A is divided by the standard deviation of said row of A. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the at least one matrix comprises a crosscorrelation matrix. The system according to any of the preceding system embodiments, with the features of embodiments S69 and S79, wherein the processing module is configured to derive the cross-correlation matrix from the non-overlapping segments. The system according to any of the preceding system embodiments, with the features of embodiments S69 and S79, wherein the processing module is configured to generate the entries of the cross-correlation matrix based on the absolute cross correlation function between any of the uniform non-overlapping segments. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the at least one matrix comprises a mutual information matrix. The system according to any of the preceding system embodiments, with the features of embodiments S69 and S82, wherein the processing module is configured to derive the mutual information matrix from the non-overlapping segments. The system according to any of the preceding system embodiments, with the features of embodiments S69 and S82, wherein the processing module is configured to generate the entries of the mutual information matrix from a mutual information between any of the uniform non-overlapping segments. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the at least one matrix comprises a square matrix. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the at least one matrix comprises a matrixshaped representation of the at least one voltage noise signal. The system according to any of the preceding system embodiments, with the features of embodiments S74, S79, S82, and S86, wherein the processing module is configured to determine eigenvalues and / or eigenvectors of at least one among: the Pearson correlation matrix, the cross-correlation matrix, the mutual information matrix, the matrix-shaped representation of the at least one voltage noise signal. The system according to any of the preceding system embodiments, with the features of embodiment S3, wherein the at least one matrix comprises a measurement matrix. The system according to any of the preceding system embodiments, with the features of embodiment S69 and S88, wherein the processing module is configured to derive the measurement matrix from the non-overlapping segments. The system according to any of the preceding system embodiments, with the features of embodiment S6 and S17, wherein the transform module is configured to determine the at least one eigenvalue metric and / or at least one eigenvector metric. The system according to any of the preceding system embodiments, with the features of embodiment S7 and S17, wherein the transform module is configured to determine the at least one singular value metric and / or at least one singular vector metric. The system according to any of the preceding system embodiments, with the features of embodiment S6, wherein the at least one eigenvalue metric comprises a spectral radius and / or a trace and / or a condition number and / or distributions of the eigenvalues and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0525] 593. The system according to any of the preceding system embodiments, with the features of embodiment S6, wherein the at least one eigenvector metric comprises a localization metric and / or an entropy of components of the eigenvector and / or a localization length and / or the like.

[0526] 594. The system according to any of the preceding system embodiments, with the features of embodiment S6 and S17, wherein the transform module is configured to select a set of the at least one eigenvalue and / or at least one eigenvector based on limits provided by the random matrix theory (RMT).

[0527] 595. The system according to any of the preceding system embodiments, with the features of embodiment S7, wherein the at least one singular value metric is calculated on the squares of the singular values, which are eigenvalues, and comprises: a spectral radius and / or a trace and / or condition number and / or distributions of the eigenvalues values and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like.

[0528] 596. The system according to any of the preceding system embodiments, with the features of embodiment S7, wherein the at least one singular vector metric is calculated on the left singular vectors and comprises: a localization metric and / or an entropy of components of the singular vector and / or a localization length and / or the like.

[0529] 597. The system according to any of the preceding system embodiments, with the features of embodiment S7, wherein the transform module is configured to select a set of the at least one singular value and / or at least one singular vector based on limits provided by the random matrix theory (RMT).

[0530] 598. The system according to any of the preceding system embodiments, wherein the analyzing module is configured to enable an evaluation of the structured data set based on the at least one characteristic descriptor. S99. The system according to any of the preceding system embodiments, with the features of embodiment S64, wherein the analyzing module is configured to extract at least a characteristic descriptor from the structured data set derived at least from the processing window at the first time and from the processing window at the second time.

[0531] 5100. The system according to any of the preceding system embodiments, wherein the at least one characteristic descriptor comprises one or more statistical moments of the structured data set and / or one or more divergence values for one or more densities in the structured data set and / or one or more residual squared sum values for one or more differences between the structured data set and theoretical values and / or the like.

[0532] 5101. The system according to any of the preceding system embodiments, wherein the establishing module is configured to use a remaining useful life-time and / or a power handling capability as the performance indicator of the at least one battery cell.

[0533] 5102. The system according to any of the preceding system embodiments, wherein the establishing module is configured to perform a sensitivity analysis of the at least one characteristic descriptor with respect to the performance indicator of the at least one battery cell.

[0534] 5103. The system according to any of the preceding system embodiments, with the features of embodiment S102, wherein the establishing module is configured to select a set of the at least one characteristic descriptor based on the sensitivity analysis.

[0535] 5104. The system according to any of the preceding system embodiments, with the features of embodiment S102, wherein the establishing module is configured to establish the at least one relation based, at least in part, on the sensitivity analysis.

[0536] 5105. The system according to any of the preceding system embodiments, wherein the establishing module is configured to perform a statistical analysis of the at least one characteristic descriptor.

[0537] S106. The system according to any of the preceding system embodiments, with the features of embodiment S105, wherein the establishing module is configured to utilize the statistical analysis to identify clusters of behavior and / or trends and / or tendencies among the at least one characteristic descriptor.

[0538] 5107. The system according to any of the preceding system embodiments, with the features of embodiment S105, wherein the establishing module is configured to establish the at least one relation based, at least in part, on the statistical analysis.

[0539] 5108. The system according to any of the preceding system embodiments, wherein the at least one relation is a functional relation.

[0540] 5109. The system according to any of the preceding system embodiments, with the features of embodiment S15, wherein the map coincides with the at least one relation.

[0541] SI 10. The system according to any of the preceding system embodiments, with the features of embodiment S15, wherein the at least one equivalent performance indicator and / or one updated performance indicator is a continuous performance indicator.

[0542] Sill. The system according to any of the preceding system embodiments, with the features of embodiment S15, wherein the at least one equivalent performance indicator and / or one updated performance indicator is a discrete performance indicator.

[0543] SI 12. The system according to any of the preceding system embodiments, with the features of embodiment S15, wherein the establishing module is configured to use the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on at least a characteristic descriptor of at least a newly measured battery cell.

[0544] SI 13. The system according to any of the preceding system embodiments, with the features of embodiment S15, wherein the establishing module is configured to improve a battery-cell manufacturing process based at least in part on said the at least one equivalent performance indicator and / or one updated performance indicator.

[0545] SI 14. The system according to any of the preceding system embodiments, wherein the system is a system for assessing the quality of a battery cell in a battery-cellmanufacturing process. SI 15. The system according to any of the preceding system embodiments, wherein the system is a system for assessing the intrinsic activity of a battery cell.

[0546] SI 16. The system according to any of the preceding system embodiments, wherein the measuring module is configured measure the at least one voltage noise signal in a time from 1 minute to 2 hours, preferably from 5 minutes to 1 hour, more preferably from 10 minutes to 30 minutes.

[0547] SI 17. The system according to any of the preceding system embodiments, wherein the system is a system for manufacturing at least a battery cell.

[0548] SI 18. The system according to any of the preceding system embodiments, wherein the system is a system for identifying anomalies in at least a battery cell.

[0549] SI 19. The system according to any of the preceding system embodiments, wherein the system is a system for identifying at least a cause of anomalies in at least a battery cell.

[0550] 5120. The system according to any of the preceding system embodiments, wherein the system is a system for testing at least a battery cell for self-discharge in a batterycell-manufacturing process.

[0551] 5121. The system according to any of the preceding system embodiments, wherein the system is a system for speeding up the cell finishing step in a battery-cellmanufacturing process.

[0552] 5122. The system according to any of the preceding system embodiments, wherein the system is a system for speeding up the aging phase in a battery-cell manufacturing process.

[0553] 5123. The system according to any of the preceding system embodiments, wherein the aging phase in a battery-cell manufacturing process using the system according to any of the preceding embodiments lasts between 10 minutes and 5 days, preferably between 20 minutes and 10 hours, more preferably between 30 minutes and 3 hours.

[0554] 5124. The system according to any of the preceding system embodiments, wherein the system is a system for directly identifying at least a weak and / or no good and / or outlier and / or scrap battery cell during the aging phase of a battery-cellmanufacturing process.

[0555] 5125. The system according to any of the preceding system embodiments, wherein the system is a system for testing incoming goods inspection for self-discharge and / or anomalies, wherein the incoming goods comprise a battery cell.

[0556] 5126. The system according to any of the preceding system embodiments, wherein the system is a system for testing at least a fuel-cell for anomalies.

[0557] 5127. The system according to any of the preceding system embodiments, wherein the system is a system for testing a performance indicator of a chemical sensor.

[0558] 5128. The system according to any of the preceding system embodiments, wherein the system is a system for improving at least a battery-cell-manufacturing process of at least a battery cell.

[0559] 5129. The system according to any of the preceding system embodiments, wherein the system is a system for improving at least a step of a battery-cell-manufacturing process of at least a battery cell.

[0560] 5130. The system according to any of the preceding system embodiments, wherein the system is a system for optimizing the costs of at least a battery-cell-manufacturing process of at least a battery cell.

[0561] 5131. The system according to any of the preceding system embodiments, wherein the system is a system for increasing the yield of at least a battery-cell-manufacturing process of at least a battery cell.

[0562] 5132. The system according to any of the preceding system embodiments, wherein the system is adapted to carry out the method recited in any of the preceding method embodiments.

[0563] 5133. The system according to any of the preceding system embodiments, wherein the system is adapted to carry out any given step of the method recited in any of the preceding method embodiments.

[0564] M135. The method according to any of the preceding method embodiments, wherein the method comprises carrying out the method according to any of the preceding method embodiments via the system according to any of the preceding system embodiments.

[0565] Below, use embodiments are presented. Use embodiments are abbreviated by the letter "U" followed by a number. Whenever reference is made herein to "use embodiments", these embodiments are meant.

[0566] Ul. Use of the system according to any of the preceding system embodiments.

[0567] U2. Use according to the preceding embodiment for carrying out the method according to any of the preceding method embodiments.

[0568] Below, computer program embodiments are presented. Computer program embodiments are abbreviated by the letter "C" followed by a number. Whenever reference is made herein to "computer program embodiments", these embodiments are meant.

[0569] Cl. A computer program product comprising instructions which, when executed by a processing component, cause the component to carry out the method according to any of the preceding method embodiments.

[0570] C2. A computer-readable medium comprising instructions which, when executed by a processor, cause the computer to carry out the method according to any of the preceding method embodiments.

[0571] C3. A data carried signal carrying the computer program product of embodiment Cl.

[0572] Brief description of the figures

[0573] Fig. 1 illustrates, as an example, a high-level overview of a method according to a preferred embodiment of the present invention.

[0574] Fig. 2 illustrates, as an example, an overview of a system according to a preferred embodiment of the present invention.

[0575] Fig. 3 illustrates, as an example, a detailed overview of a sub-step of the processing step of a method according to a preferred embodiment of the present invention.

[0576] Fig. 4 illustrates, as an example, a detailed overview of a sub-step of the processing step of a method according to a preferred embodiment of the present invention. Fig. 5 illustrates, as an example, a detailed overview of a sub-step of the establishing step of a method according to a preferred embodiment of the present invention.

[0577] Fig. 6 illustrates, as an example, a data and control interface module according to a preferred embodiment of the present invention

[0578] Detailed description of the figures

[0579] It is noted that not all the drawings carry all reference signs. Instead, in some of the drawings, some of the reference signs have been omitted for sake of brevity and simplicity of illustration.

[0580] Hereafter, exemplary embodiments of the present invention will be described in detail, referring to the accompanying figures.

[0581] Fig. 1 illustrates, as an example, a high-level overview of a method according to a preferred embodiment of the present invention. The method can be a method for assessing the quality of a battery cell in a battery-cell-manufacturing process.

[0582] According to the embodiment of Fig. 1, the method can comprise the step of measuring 1 a voltage signal of at least a battery cell. The voltage signal can be a voltage noise signal. The voltage noise signal can be measured passively. The step of measuring a voltage signal can be performed by a measuring component.

[0583] The method can further comprise the step of processing 2 the measured voltage noise signal into a structured data set. The step of processing 2 can comprise the determination of a mathematical structure, including, among others, vectors, matrices, lists and the like. In a preferred embodiment, the step of processing 2 can comprise the determination of a matrix. The step of processing 2 the measured voltage noise signal into a structured data set can comprise the determination of, for examples, eigenvalues and / or the eigenvectors of said matrix. The structured data set can, for instance, comprise a metric for said eigenvalues and / or a metric for said eigenvectors, such as a distribution of the eigenvalues or a participation ratio of the eigenvectors.

[0584] The method can further comprise the step of analyzing 3 the structured data set to extract one or more characteristic descriptor 22 from the structured data set. In simpler terms, the structured data set defined in the preceding step is examined and, after examination, a characteristic descriptor 22 of the structured data set is identified. In a preferred embodiment, the characteristic descriptor can be a characteristic quantitative descriptor. For example, the characteristic descriptor 22 can comprise statistical moments of the structured data set.

[0585] The method can further comprise the step of establishing a relation 4 between the one or more characteristic descriptor 22 and at least a performance indicator 42 of the at least one battery cell. In a preferred embodiment, the at least one performance indicator 42 can be a quality metric of the at least one battery cell. For instance, the at least one performance indicator 42 can be a self-discharge-like metric. The at least one performance indicator 42 of the at least one battery cell can further comprise, for example, one or more characteristics and / or features of a battery cell defining a type of the battery cell. For instance, the at least one performance indicator 42 of the at least one battery cell can comprise a performance metric that allows to group a plurality of battery cells into a group of battery cells of the same type during a battery-cell-manufacturing process.

[0586] The method can further comprise building a map based at least in part on the at least one relation between the at least one characteristic descriptor 22 and at least a performance indicator 42.

[0587] In a more practical example, the method can allow for the quantification of the selfdischarge of the least one battery cell and the formulation of a hypothesis about the causes of the self-discharge in the at least one battery cell and / or in the group of battery cell comprising battery cells of the same type as the at least one battery cell.

[0588] Fig. 2 illustrates, as an example, an overview of a system according to a preferred embodiment of the present invention. The system can be a system for assessing the quality of a battery cell in a battery-cell-manufacturing process

[0589] The system can comprise a measuring module 5. The measuring module can be configured to measure a voltage signal, preferably a voltage noise signal, from a battery cell 6. The battery cell 6 can be the product of a battery-cell-manufacturing process 11. The measuring module 5 can comprise at least a card 7. Each card can be configured to measure a voltage signal, preferably a voltage noise signal, from a battery cell 6. In other words, multiple cards (indicated by 8) can allow the measurement of a voltage signal from multiple batteries (indicated by 9), in parallel.

[0590] A cell voltage signal 10 can generally comprise a DC and an AC component. The card 5 can comprise a coupling capacitor 12, which can be configured to separate the AC noise component from the cell voltage noise signal. Put differently, the coupling capacitor may output a cell voltage noise signal 13. The card 5 may further comprise a high-gain low-noise amplifier 14, which can amplify the cell voltage noise signal 13 into an amplified cell voltage noise signal 15. The card 5 can further comprise a high-resolution ADC 16, which can be configured to sample the amplified cell voltage noise signal 15 to obtain a signal in the digital domain, in other words a digitized cell voltage noise signal 17. It will be understood that the digitized cell voltage noise signal 17 may also be referred to as voltage noise signal 17. The card 5 may further comprise a control and communication unit 18, which may communicate the digitized cell voltage noise signal 17 to further units of the system, for example to an evaluation unit 19 or to any modules comprises in the evaluation unit 19.

[0591] The control and communication unit 18 can be configured to control the gain and sample rate of the high-resolution ADC 16, to control a polarization compensation of the coupling capacitor 12, to enable a temperature measurement of the battery cell 6, and / or to enable the measurement of the cell voltage signal 10.

[0592] The measurement module 5 can be configured to carry out the measuring step 1 of the embodiment of Fig. 1.

[0593] Generally speaking, the evaluation unit 19 can be configured to communicate with and / or control the card 5. Additionally, or alternatively, the evaluation unit 19 can be adapted to provide a user with an interface to control and set-up the measurement module 5, to provide a manufacturing system in a factory with an interface for manufacturing execution, and / or to provide a data storage system with an interface. Further, the evaluation unit 19 can be configured to perform calculations.

[0594] In the evaluation unit the following functions can be performed in a calculation pipeline. In other words, the evaluation unit 19 may perform, without limitation, the following steps. Cleaning and / or detrending the voltage noise signal 17. Determining eigenvalues and / or eigenvectors 20 of a matrix, wherein the matrix may be constructed based, at least in part, on the voltage noise signal 17. Determining at least a structured data set 21, which can be metrics for the eigenvalues and / or metrics for the eigenvectors, based on RMT, i.e. using limits provided by RMT to filter out pure random noise. Determining at least a characteristic descriptor 22 from the at least one structured data set 21. Using a sensitivity analysis 43 of the at least one characteristic descriptor 22 with respect to at least a performance indicator 42 of the battery cell 6 to determine a set of the at least one characteristic descriptor 22, said set showing a large correlation to the at least one performance indicator 42 of the battery cell 6. Establishing at least a relation between the at least one characteristic descriptor 22 and the at least one performance indicator 42. Establishing a map 23 based at least in part on the at least one relation, the map linking the at least one characteristic descriptor 22 to at least an equivalent performance indicator and / or an updated performance indicator 24. Using the map to determine the equivalent performance indicator and / or the updated performance indicator 24 for newly measured cells. Communicating at least the equivalent performance indicator and / or an updated performance indicator 24 to a user and / or to other system, such as, for instance, a manufacturing execution system for battery cells.

[0595] The evaluation unit 19 can comprise a processing module 25. The processing module 25 can be configured to carry out the processing step 2 of the embodiment of Fig. 1. The processing module 25 can comprise a transform module 26, which can be configured to determine the at least one structured data set 21 from the eigenvalues and / or eigenvectors

[0596] 20. The processing module can be adapted to output the at least one structured data set

[0597] 21.

[0598] The evaluation unit 19 can comprise an analyzing module 27. The analyzing module 27 can be configured to carry out the analyzing step 3 of the embodiment of Fig. 1. The analyzing module 27 can be configured to determine the at least one characteristic descriptor 22 from the at least one structured data set 21 and to output the at least one characteristic descriptor 22.

[0599] The evaluation unit 19 can comprise an establishing module 28. The establishing module 28 can be configured to carry out the establishing step 4 of the embodiment of Fig. 1. The establishing module 28 can be configured to establish at least a relation between the at least one characteristic descriptor 22 and at least one performance indicator 42 of the battery cell 6. The establishing module 28 can further be configured to establish a map 23 based at least in part on the at least one relation, the map 23 linking the at least one characteristic descriptor 22 to at least an equivalent performance indicator and / or an updated performance indicator 24. The map 23 may coincide with the at least one relation. The establishing module 28 can be configured to output the map 23. This can be the case, for example, during a set-up phase 45, wherein the output of the establishing module 28 can be the map 23. The establishing module 28 can further be configured to determine the equivalent performance indicator and / or the updated performance indicator 24 for newly measured cells. The establishing module 28 can be configured to output the equivalent performance indicator and / or the updated performance indicator 24 for newly measured cells. This can be the case, for example, during a use phase 46, wherein the output of the establishing module 28 can be the output the equivalent performance indicator and / or the updated performance indicator 24 for newly measured cells.

[0600] The evaluation unit 19 or any combination of the modules comprised in the evaluation unit 19 can be arranged locally or remotely, such as in a cloud solution. The processing module 25, the analyzing module 27, and the establishing module 28, or any combination thereof, can be combined into a single module.

[0601] The evaluation unit 19 can further comprise a data and control interface module 29. The data and control interface module 29 can be configured to communicate with any module comprised in the system and / or any combination of modules comprised in the system (this is indicated by arrows 30). The data and control interface module 29 can be configured to communicate the output of any module comprised in the system to a user and to provide, to an authorized user, access to any of the modules of the system. Furthermore, the data and control interface module 29 can be configured to modify one or more parameters related to the functioning of any module comprised in the system. For instance, the data and control interface module 29 can modify the measurement parameters used in measuring module 5, such as the sampling rate of the high-resolution ADC 16. As yet another example, the data and control interface module 29 can modify the type of structured data set 21 outputted by the processing module 25 and / or the type of characteristic descriptor 22 output by the analyzing module 27.

[0602] Any module comprised in the system may comprise a computing device.

[0603] Fig. 3 illustrates, as an example, a detailed overview of a sub-step of the processing step 2 of a method according to a preferred embodiment of the present invention.

[0604] The sub-step can comprise a pre-processing 31 of the voltage noise signal 17. The preprocessing may include removing artifacts from the voltage noise signal 17, such as spikes, and detrending the voltage noise signal 17. The pre-processing may generate a pre- processed voltage noise signal 32.

[0605] The sub-step can comprise a windowed evaluation 33 of the pre-processed voltage noise signal 32. The windowed evaluation 33 can comprise the generation of one or more matrices from the pre-processed voltage noise signal 32. The one or more matrices can comprise a measurement matrix 34, where the windowed evaluation may further comprises generating a Pearson correlation matrix 35 from the measurement matrix 34. The one or more matrices can comprise a cross correlation matrix 36. The one or more matrices can comprise a mutual information matrix 37. The one or more matrices can further comprise any matrix shaped representation 38 of the pre-processed voltage noise signal 32.

[0606] The sub-step can comprise calculating the eigenvalues and / or eigenvectors 20 of the of one or more matrices.

[0607] Fig. 4 illustrates, as an example, a detailed overview of a sub-step of the processing step 2 of a method according to a preferred embodiment of the present invention.

[0608] The sub-step can comprise the definition of a measurement signal 39. The measurement signal 39 essentially comprises a section, in time, of the noise voltage signal 17.

[0609] The sub-step can comprise the windowed evaluation 33. The windowed evaluation 33 can comprise the definition of a processing window 40. The processing window 40 essentially comprises a section, in time, of the measurement signal 39. The processing window 40 can slide in time, as indicated by the arrow, such that at a later time the processing window 40' comprises a section, in time, of the measurement signal 39 that is different from the section, in time, of the measurement signal 39 comprised by the processing window 40.

[0610] The processing window 40 and the processing window 40' may be the same processing window, at different times. The windowed evaluation 33 can comprise using the time duration of the processing window 40 and / or of the processing window 40' and / or the overlap of the processing window 40 with the processing window 40' as parameters.

[0611] The windowed evaluation 33 can comprise defining non-overlapping segments 41, 41', 41" from the processing window 40. The same applies to any time-slided version of the processing window 40, such as the processing window 40'. The non-overlapping segments 41, 41', 41" may be uniform, i.e. they may have the same length. In other words, they may contain the same number of data points. Each of the non-overlapping segments 41, 41', 41" can comprise a section of the processing window 40 that does not overlap with other sections of the processing window 40. The windowed evaluation 33 can comprise using the number of non-uniform overlapping segments as a parameter.

[0612] The cross-correlation matrix 36 and / or the mutual information matrix 37 and / or any matrix shaped representation 38 of the pre-processed voltage noise 32 can be determined based on the non-uniform overlapping segments. For example, the windowed evaluation 33 can comprise generating the entries of the cross-correlation matrix 36 based on the absolute cross correlation function between any of the non-overlapping segments 41, 41', 41". The windowed evaluation 33 can comprise generating the measurement matrix 34 based on the non-overlapping segments 41, 41', 41". In particular, the entries of a row of the measurement matrix 34 can generated from one of the non-overlapping segments 41, 41', 41". In other words, for example, if the non-overlapping segment 41 has n data points, then the first element of the first row of the measurement matrix 34 may be equal to the first of the n data points, the second element of the first row of the measurement matrix 34 may be equal to the second of the n data points, et cetera. The second row of the measurement matrix 34 may be determined, according to the same logic, from the nonoverlapping segment 41', and the third row of the measurement matrix 34 may be determined, according to the same logic, from the non-overlapping segment 41".

[0613] Fig. 5 illustrates, as an example, a detailed overview of a sub-step of the establishing step 4 of a method according to a preferred embodiment of the present invention.

[0614] The sub-step can comprise a set-up phase 45. The set-up phase 45 can comprise performing a sensitivity analysis 43 of the at least one characteristic descriptor 22 with regard to the at least one performance indicator 42. The sensitivity analysis 43 may look for a highest correlation between the at least one characteristic descriptor 22 and the at least one performance indicator 42. As a result, a set of the at least one characteristic descriptor 22 with high correlation to the at least one performance indicator 42 may be selected. Preferably based on said set, the sub-step can comprise establishing the at least one relation between the at least one performance indicator 42 and the at least one characteristic descriptor 22. The relation may be a functional relation. For example, the relation may be a functional relation expressing a performance indicator, such as the selfdischarge of a battery cell 6, as a function of a plurality of characteristic descriptors. Preferably based on said set, the sub-step can comprise establishing a map 23 based at least in part on the at least one relation. The map may link the at least one characteristic descriptor 22 to at least an equivalent performance indicator and / or an updated performance indicator 24. The at least one equivalent performance indicator may essentially coincide with the at least one performance indicator 42. As an example, if the sensitivity analysis is performed with regards to, e.g., the self-discharge, then the map may link the self-discharge itself to the at least one characteristic descriptor 22. The at least one updated performance indicator may be essentially different from the at least one performance indicator 42. For example, if the sensitivity analysis is performed with regards to, e.g., the self-discharge, then the map may link a grade of cell quality to the at least one characteristic descriptor 22. The map 23 may coincide with the at least one relation.

[0615] The sub-step can comprise a use phase 46. The use phase 46 can comprise an application 44 of the map 23. In other words, the map 23 may be used, using as inputs the characteristic descriptors 22, to determine the equivalent performance indicator and / or the updated performance indicator 24. The determination of the equivalent performance indicator and / or the updated performance indicator 24 may be done for newly measured cells. In other words, during the set-up phase a map 23 may be built. Subsequently, during the use phase 46, new battery cells can be measured and the map 23 may be applied, using as inputs the characteristic descriptors of the newly measured cells, said descriptors determined by applying steps of the method according to the present invention.

[0616] Fig. 6 illustrates, as an example, a data and control interface module 29 according to a preferred embodiment of the present invention. The data and control interface module 29 can be configured to bi-directionally communicate with any module 49 comprised in the system according to any embodiment of the present invention. The data and control interface module 29 can be configured to display, e.g. to a user, one or more outputs of any module 30 comprised in the system according to any embodiment of the present invention.

[0617] For example, the data and control interface module 29 can be configured to display a first data stream 47 and a second data stream 47'. The first data stream 47 can be streamed from a first card, measuring a first battery cell 48, in the measurement module 5. The second data stream 47' can be streamed from a second card, measuring a second battery cell 48', in the measurement module 5. The first data stream 47 and the second data stream 47' can comprise streams of any signal handled within the measurement module 5. For example, the first data stream 47 can comprise a stream of the voltage noise signal of the first battery cell 48 and the second data stream 47' can comprise a stream of the voltage noise signal of the second battery cell 48'.

[0618] Furthermore, the data and control interface module 29 can be configured to display a first classification 50 of the first battery cell 48 and a second classification 50' of the second battery cell 4'. The first classification 50 and the second classification 50' can relate to one or more classes and / or types and / or parameters of the first battery cell 48 and of the second battery cell 48'.

[0619] While in this specification preferred embodiments of the present invention are described, the person skilled in the art will understand that the preferred embodiments are provided for illustrative purposes only and to render the disclosure of the present invention complete, and should by no means be construed to limit the scope of the present invention, which is defined by the claims. Whenever a relative term, such as "about", "substantially", "essentially" or "approximately" is used in this specification, such a term should also be construed to also include the exact term. That is, e.g., "substantially straight" should be construed to also include "(exactly) straight".

[0620] Whenever steps were recited in the above or also in the appended claims, it should be noted that the order in which the steps are recited in this text may be accidental. That is, unless otherwise specified or unless clear to the skilled person, the order in which steps are recited may be accidental. That is, when the present document states, e.g., that a method comprises steps (A) and (B), this does not necessarily mean that step (A) precedes step (B), but it is also possible that step (A) is performed (at least partly) simultaneously with step (B) or that step (B) precedes step (A). Furthermore, when a step (X) is said to precede another step (Z), this does not imply that there is no step between steps (X) and (Z). That is, step (X) preceding step (Z) encompasses the situation that step (X) is performed directly before step (Z), but also the situation that (X) is performed before one or more steps (Yl), ..., followed by step (Z). Corresponding considerations apply when terms like "after" or "before" are used.

Claims

83Claims1. A method for assessing the quality of a battery cell in a battery-cell-manufacturing process, the method comprising: passively measuring at least a voltage noise signal of at least a battery cell, processing the at least one voltage noise signal into a structured data set, analyzing the structured data set and extracting at least a characteristic descriptor from the structured data set, establishing at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell.

2. The method according to the preceding claim, wherein the method comprises isolating an AC voltage noise signal of the at least one battery cell from the voltage signal of at the at least one battery, said voltage signal comprising an AC component and a DC component, wherein the method comprises amplifying the AC voltage noise signal into an amplified voltage noise signal, wherein the method comprises sampling the amplified voltage noise signal and obtaining a digitized voltage noise signal.

3. The method according to any of the preceding claims, wherein the step of processing the at least one voltage noise signal into a structured data set comprises determining at least a matrix based at least in part on the at least one voltage noise signal, and / or determining eigenvalues and / or eigenvectors of the at least one matrix.

4. The method according to any of the preceding claims, wherein the method comprises: pre-processing the at least one voltage noise signal into at least one pre-processed voltage noise signal; establishing a measurement signal, wherein the measurement signal comprises a section of the pre-processed voltage noise signal; establishing at least a processing window, wherein the processing window comprises a section of the measurement signal; sliding, in time, the at one least processing window over the measurement signal; establishing non-overlapping segments, wherein each of the non-overlapping segments comprises a nonoverlapping section of the at least one processing window; deriving the at least one matrix from the non-overlapping segments.

5. The method according to claim 3, wherein the structured data set comprises at least an eigenvalue metric of said eigenvalues and / or at least an eigenvector metric of said eigenvectors, wherein at least one eigenvalue metric comprises a84 spectral radius and / or a trace and / or a condition number and / or distributions of the eigenvalues and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like, wherein the at least one eigenvector metric comprises a localization metric and / or an entropy of components of the eigenvector and / or a localization length and / or the like, the method comprises selecting a set of the at least one eigenvalue and / or at least one eigenvector based on limits provided by the random matrix theory.

6. The method according to any of the preceding claims, wherein the method comprises using a quality metric of the at least one battery cell as the performance indicator of the at least one battery cell and / or wherein the method comprises using a self-discharge-like metric as the performance indicator of the at least one battery cell.

7. The method according to any of the preceding claims wherein the at least one characteristic descriptor comprises one or more statistical moments of the structured data set and / or one or more divergence values for one or more densities in the structured data set and / or one or more residual squared sum values for one or more differences between the structured data set and theoretical values and / or the like.

8. The method according to any of the preceding claims, wherein the method comprises building at least a map based at least in part on the at least one relation between the at least one characteristic descriptor and at least a performance indicator, wherein the map links the at least one characteristic descriptor to the at least one equivalent performance indicator and / or one updated performance indicator9. The method according to the preceding claim, wherein the method comprises using the map to determine the at least one equivalent performance indicator and / or one updated performance indicator based at least in part on the at least one characteristic descriptor and / or based at least in part on at least a characteristic descriptor of at least a newly measured battery cell.

10. The method according to claim 8, wherein the method comprises improving a battery-cell manufacturing process based at least in part on said the at least one equivalent performance indicator and / or one updated performance indicator.8511. A system for assessing the quality of a battery cell in a battery-cell-manufacturing process, the system comprising: at least a measuring module configured to passively measure at least a voltage noise signal of at least a battery cell, at least a processing module configured to process the at least one voltage noise signal into a structured data set, at least an analyzing module configured to analyze the structured data set and to extract at least a characteristic descriptor from the structured data set, at least an establishing module configured to establish at least a relation between the at least one characteristic descriptor and at least a performance indicator of the at least one battery cell.

12. The system according to the preceding system claim, wherein the measuring module comprises a high-resolution AC voltage detecting set-up, comprising at least at least a card of the high-resolution AC voltage detecting set-up, wherein said card is configured to measure a voltage noise signal from the at least one battery cell, wherein the at least one card comprises an AC-coupling device, a high-gain low-noise amplifier, and a high-resolution ADC converter, wherein the AC coupling device is configured to isolate an AC voltage noise signal of the at least one battery cell from a voltage signal of at the at least one battery, said voltage signal comprising an AC component and a DC component, wherein the high-gain low-noise amplifier is configured to amplify the AC voltage noise signal into an amplified voltage noise signal, wherein the high-resolution ADC is configured to sample the amplified voltage noise signal and obtain a digitized voltage noise signal.

13. The system according to any of the preceding system claims, wherein the processing module is configured to determine at least a matrix based at least in part on the at least one voltage noise signal, and / or to determine eigenvalues and / or eigenvectors of the at least one matrix.

14. The system according to the preceding system claim, wherein the processing module is configured to: pre-process the at least one voltage noise signal into at least one pre-processed voltage noise signal; establish a measurement signal, wherein the measurement signal comprises a section of the pre-processed voltage noise signal; establish at least a processing window, wherein the processing window comprises a section of the measurement signal; slide, in time, the at one least processing window over the measurement signal; establish non-overlapping86 segments, wherein each of the non-overlapping segments comprises a nonoverlapping section of the at least one processing window; derive the at least one matrix from the non-overlapping segments.

15. The system according to system claim 13, wherein the processing module comprises a transform module, wherein the structured data set comprises at least an eigenvalue metric of said eigenvalues and / or at least an eigenvector metric of said eigenvectors, wherein the at least one eigenvalue metric comprises a spectral radius and / or a trace and / or a condition number and / or distributions of the eigenvalues and / or one or more distributions of a spacing of the eigenvalues and / or one or more distributions of the ratios of spacings of the eigenvalues and / or a spectral rigidity of the eigenvalues and / or a number variance of the eigenvalues and / or the like, wherein the at least one eigenvector metric comprises a localization metric and / or an entropy of components of the eigenvector and / or a localization length and / or the like, wherein the transform module is configured to select a set of the at least one eigenvalue and / or at least one eigenvector based on limits provided by the random matrix theory (RMT).

16. The system according to any of the preceding system claims, wherein the establishing module is configured to use a quality metric of the at least one battery cell as the performance indicator of the at least one battery cell, and / or to use a self-discharge-like metric as the performance indicator of the at least one battery cell.

17. The system according to any of the preceding system claims, wherein the at least one characteristic descriptor comprises one or more statistical moments of the structured data set and / or one or more divergence values for one or more densities in the structured data set and / or one or more residual squared sum values for one or more differences between the structured data set and theoretical values and / or the like.

18. The system according to any of the preceding system claims, wherein the establishing module is configured to build at least a map based at least in part on the at least one relation between the at least one characteristic descriptor and at least a performance indicator, wherein the map links the at least one characteristic descriptor to at least an equivalent performance indicator and / or an updated performance indicator.

19. The system according to the preceding system claim, wherein the establishing module is configured to use the map to determine at least an equivalent performance indicator and / or an updated performance indicator based at least in part on the at least one characteristic descriptor and / or based at least in part on at least a characteristic descriptor of at least a newly measured battery cell.

20. The system according to system claim 18, wherein the establishing module is configured to improve a battery-cell manufacturing process based at least in part on said the at least one equivalent performance indicator and / or one updated performance indicator.

Images