BATTERY PROGNOSIS METHOD AND SYSTEM

Abstract

An electric vehicle comprises an electric motor for driving the electric vehicle and a battery suitable for supplying the electric motor with electrical energy to drive the electric vehicle.The electric vehicle also includes one or more controllers programmed together with the following instructions: measuring battery data from the battery; using the battery data to create a model of the battery's storage capacity as a function of the distance traveled by the electric vehicle; evaluating the reliability of the model; if the model is assessed to be sufficiently reliable, performing a forecast for the battery using the model; and if the model is assessed to be insufficiently reliable, adopting a substitute model of the battery's storage capacity as a function of the distance traveled by the electric vehicle, the substitute model being based on battery data from one or more other vehicles, and performing a forecast for the battery using the substitute model.

Description

INTRODUCTIONThis disclosure relates to the field of battery prognosis.Batteries are a source of electrical energy for powering electric vehicles. Predicting premature or excessive degradation of the batteries is useful, among other things, to inform a vehicle owner of premature or excessive degradation of the batteries of the vehicle and to provide feedback to the design and manufacturing functions to enable them to make continuous improvements to the batteries.DESCRIPTIONAn electric vehicle comprises an electric motor for driving the electric vehicle and a battery for supplying the electric motor with electrical energy for driving the electric vehicle. The electric vehicle also includes one or more controllers programmed together with the following instructions: measurement of battery data of the battery; use the battery data to generate a model of storage capacity of the battery as a function of distance traveled by the electric vehicle; evaluate the reliability of the model; if the model is evaluated as sufficiently reliable, perform a prediction for the battery using the model; and if the model is evaluated as not sufficiently reliable, take a replacement model of storage capacity of the battery as a function of distance traveled by the electric vehicle, the replacement model being based on battery data of one or more other vehicles, and perform a prediction for the battery using the replacement model.The model may be judged not to be sufficiently reliable when the battery data has insufficient number of measurement data. Further, the model may be evaluated as not being sufficiently reliable if the model has an increasing storage capacity relative to the distance traveled. In addition, the model may be judged not to be sufficiently reliable when the model has noise.The one or more other vehicles may be selected based on the operating characteristics of the one or more other vehicles that are similar to the operating characteristics of the electric vehicle. These operating characteristics may include the average travel distance per day. These operating characteristics may alternatively or additionally comprise the average ambient temperature.The prediction may include a comparison of the anticipated storage capacities of the individual battery components. The prediction may alternatively or additionally include comparing the anticipated loss of storage capacity between the individual battery portions to a threshold. The prognosis may alternatively or additionally include using the model or the substitute model in combination with internal resistance measurements of the battery.A second electric vehicle comprises an electric motor for driving the electric vehicle and a battery which is suitable for supplying the electric motor with electrical energy for driving the electric vehicle. The electric vehicle additionally includes one or more controllers programmed in conjunction with the following instructions: measuring battery data of the battery; using the battery data to build a model of storage capacity of the battery as a function of operating time of the battery in the electric vehicle; evaluating reliability of the model; if the model is evaluated as sufficiently reliable, performing a prediction for the battery using the model; and if the model is evaluated as not sufficiently reliable, assuming a replacement model of storage capacity of the battery as a function of operating time of the electric vehicle, the replacement model being based on battery data of one or more other vehicles, and performing the prediction for the battery using the replacement model.In the second electric vehicle, the model may be judged not to be sufficiently reliable when the battery data has insufficient number of measurement data. Further, the model may be evaluated as not being sufficiently reliable when the model exhibits an increasing storage capacity in proportion to the operating time of the battery in the electric vehicle. Moreover, the model may be judged not to be sufficiently reliable when the model has noise.In the second electric vehicle, the one or more other vehicles may be selected based on the one or more other vehicles having similar operating characteristics to the electric vehicle. These operating characteristics may include the average driving performance per day. These operating characteristics may alternatively or additionally comprise the average ambient temperature.In the second electric vehicle, the forecast can comprise the comparison of the anticipated storage capacities of the individual battery parts. The prediction may alternatively or additionally include comparing the anticipated loss of storage capacity between the individual battery portions to a threshold. The prognosis may alternatively or additionally comprise the use of the model or the substitute model in combination with internal resistance measurements of the battery.A method for predicting a battery of an electric vehicle includes, by one or more controllers, measuring battery data for the battery. The method additionally includes, by one or more controllers, using the battery data to generate a battery storage capacity model as a function of distance traveled by the electric vehicle or time of operation of the battery in the electric vehicle. Further, the method includes selecting one or more other vehicles based on the one or more other vehicles having operating characteristics similar to the operating characteristics of the electric vehicle. Additionally, the method includes creating a replacement model of the storage capacity of the battery as a function of distance traveled by the electric vehicle or time of operation of the battery in the electric vehicle based on the battery data of the one or more other vehicles. The method also includes using the model or replacement model in combination with battery internal resistance measurements as a tool for predicting the battery and identifying one or more causes of failure for predicted degradation of the battery.The summary above does not represent every embodiment or aspect of this disclosure. The above features and advantages of the present disclosure, as well as other possible features and advantages, will be readily apparent from the following detailed description of the embodiments and best modes for carrying out the disclosure when taken in conjunction with the accompanying drawings and claims. Moreover, this disclosure expressly includes combinations and sub-combinations of the elements and features set forth above and below.BRIEF DESCRIPTION OF THE DRAWINGSFIG. 1 shows an electric vehicle and a backoffice connected to this electric vehicle. FIG. 2 shows an electrical system of the motor vehicle of FIG. 1. FIG. 3 shows models of capacity versus drive power for the battery pack of the electric vehicle and for the battery packs of a representative fleet of electric vehicles. FIG. 4 illustrates a method for modeling battery capacity. FIG. 5 illustrates clusters into which a fleet of electric vehicles may be divided based on the operational characteristics of the electric vehicles. FIG. 6 shows a method for monitoring the internal resistance and capacity of the battery over time for a plurality of electric vehicles. FIG. 7 shows various causes of battery failures.DETAILED DESCRIPTIONThe present disclosure may be embodied in many different forms. Representative examples of the disclosure are illustrated in the drawings and will be described herein in detail as non-limiting examples of the disclosed principles. To this end, elements and limitations described in the sections "Summary", "Introduction", "Summary", and "Detailed Description", but not expressly recited in the claims, should not be included in the claims, either alone or in common, either by introduction or by inference, or otherwise.For purposes of the present specification, the use of the singular includes the plural and vice versa, unless expressly omitted, the terms "and" and "or" apply in both the conjunctiva and disjunction, "each" and "all" mean "each and all", and the words "including", "containing", "comprising", "with" and the like mean "including without limitation". Moreover, words of approximation such as "about", "fast", "substantially", "generally", "about", etc. may be used herein in the sense of "at, near, or near", or "within 0-5% of", or "within acceptable manufacturing tolerances", or logical combinations thereof.First, reference is made to FIG. 1, which illustrates an electric vehicle 10. FIG. 2 additionally shows that the electric vehicle 10 has an electrical system 12. The electric system 12 includes an electric powertrain for the electric vehicle 10. the electric powertrain for the electric vehicle 10 includes an electric motor 14 for powering the electric vehicle 10. the electric motor 14 may be an alternating current (AC) motor. The electrical system 12 also includes a battery pack 16. the battery pack 16 is connected to an inverter module (PIM) 18. The battery pack 16 supplies motive power to the electric motor 14 by switching within the PIM 18, which is controlled by a powertrain control unit (PCU) 20. The PCU 20 may be a stand-alone controller, may be integrated with the PIM 18, or may be integrated with other controllers in the electrical system 12 of the electric vehicle 10.The electric vehicle 10 may be any vehicle that is fully or partially powered with electrical energy, and may include all-electric vehicles and hybrid electric vehicles. Moreover, the electric vehicle 10 may be any type of vehicle, such as passenger cars, trucks, vans, sport utility vehicles, motor bicycles, motor scooters, boats, aircraft, and the like.The battery pack 16 may include a plurality of battery cells arranged in battery cell groups. Four such cell groups, cell group 30, cell group 32, cell group 34, and cell group 36, are depicted. The battery pack 16 may include any number of battery cells and cell groups. The cell groups may be grouped into battery modules, e.g., the battery module 38 including the cell group 30 and the cell group 32 and the battery module 40 including the cell group 34 and the cell group 36. Although two cell groups are shown in each battery module 38 and battery module 40, this is only for clarity. The battery modules may include any number of cell groups, and the battery pack 16 may have any number of battery modules. The battery cells, cell groups, and battery modules referred to herein may be referred to as components of the battery pack 16.The electrical system 12 may also include a battery monitoring system (BMS) controller 22 that monitors the battery pack 16. Specifically, the BMS controller 22 may monitor the state of charge of the battery pack 16 as well as the charging and discharging of the battery pack 16. In addition, the BMS controller 22 may monitor the state of charge of each individual battery module, e.g., the battery module 38, as well as its charging and discharging. In addition, the BMS controller 22 may monitor the state of charge of each cell group, e.g., the cell group 30, as well as its charge and discharge. The BMS controller 22 can also monitor the internal resistances of the battery pack 16 and / or the cell groups and / or the battery modules, for example by monitoring the charging and discharging currents and voltages of the battery pack 16 and the cell groups and battery modules, wherein the internal resistance is the quotient of voltage and current.The BMS controller 22 may be integrated with the battery pack 16 or may be separate from the battery pack 16. The BMS controller 22 is understood to have the necessary connections for monitoring the entire battery pack 16 as well as for monitoring each cell group and battery module monitored by the BMS controller 22. The BMS controller 22 must have sufficient resources (e.g., microcontroller, software, inputs, outputs, memory, and the like) to accomplish the tasks assigned thereto.The electrical system 12 may include a temperature sensor 23 that measures the ambient temperature proximate the electric vehicle 10. The temperature sensor 23 may be connected to the BMS controller 22 or to another controller in the electrical system 12.In addition to the powertrain controller 20 and the BMS controller 22, the electric vehicle 10 may include a plurality of other controllers in the electrical system 12. For example, the electric vehicle 10 may include a telematics control unit (TCU) 26 that enables the electric vehicle 10 to participate in cellular telecommunications, including cellular data communication. This telecommunication may comprise the telecommunication with a backoffice 11 operated by the manufacturer of the electric vehicle 10. (OnStar® operated by General Motors is an example of such a backoffice.) The TCU 26 may have or be connected to a suitable antenna 27 which facilitates such communication. The electric vehicle 10 may also have any number of other controllers (generally shown as controller 28 a, controller 28 b, and controller 28 nin FIG. 2 ) to control other functions of the electric vehicle 10, e.g., brakes, infotainment, steering, lighting, and the like. The various controllers of the electric vehicle 10 may be networked together via one or more data buses 29 or via individual circuits, such that the controllers may share data when performing their various tasks. Further, the electrical system 12 of the electric vehicle 10 may be split differently than shown in FIG. 2, with the tasks being performed or shared by different or different controllers.The various computations, comparisons, and other tasks described below may be performed by one controller or collectively by multiple controllers that may be networked together in the electrical system 12 of the electric vehicle 10. Each of these controllers, alone and / or in common, is believed to have sufficient electronic resources (microprocessor, memory, inputs, outputs, software, cellular network access device (NAD)) to perform the functions described in this disclosure. Some of the functions described herein may be performed by computers external to the electric vehicle 10, such as computers in the backoffice 11 or in the cloud.The one or more controllers may be responsive to "instructions" that may include one or more software instructions. Additionally, an instruction may include one or more additional instructions.Referring now to FIGS. 3 and 4, an algorithm for modeling battery capacity is shown. One function of this algorithm may be to build a battery storage capacity model represented by curve 200 as a function of distance traveled (i.e., mileage in miles or kilometers, for example) for battery pack 16.In block 300, the battery data, i.e., the storage capacity of the battery pack 16, is measured or calculated. The BMS controller 22, which monitors the charging and discharging of the battery pack 16, may detect the current storage capacity of the battery pack 16 in terms of amp-hours, kilowatts-hours, or other suitable unit of storage capacity. This capacity can be calculated regularly and continuously. The storage capacities of the cell groups and battery modules in the battery pack 16 can also be measured or calculated continuously.At block 301, telematics data may be acquired via the TCU 26. Telematics data may be provided to backoffice 11. Telematics data may include the measured capacity of the battery pack 16 or its components (cell groups, battery modules). Telematics data may also include good-fit data (from block 320). In block 302, a determination is made as to whether the battery pack 16 has been replaced. In block 304, the operating months and the driving performance of the electric vehicle 10 for the battery pack 16 are updated when the battery pack 16 has been replaced with a new battery pack. At block 306, any distortion arising from the replacement of the battery pack 16 with a new battery pack (the new battery pack having a beginning-of-life (BOL) capacity that is within a narrow tolerance range) is eliminated and the algorithms described herein are reset / re-learned to account for the replacement of the battery pack.The measured or calculated storage capacity is then used at the beginning of a data classification portion 307 of the algorithm in which the data for the capacity of the battery pack 16 can be processed, filtered and classified. The calculated or measured capacity data for the battery pack 16 is fitted into a regression model of the storage capacity as a function of the travel distance of the electric vehicle 10 in block 308. The statistics of the regression model are calculated in block 310. These statistics may include slope, mean square error, and r-square statistics. At block 312, it is then determined whether the number of data points in the regression model is sufficient (i.e., above a predetermined threshold) to conclude that the mileage-based capacitance model for the battery pack 16 can be considered reliable for characterizing the battery pack 16. As a non-limiting example, the number of data points evaluated as "sufficient" may be two to three data points per month over a period of five or six months. If the determination in block 312 is NO (i.e., the number of data points is not sufficient), then it is determined in block 314 that the model of capacity as a function of mileage for the battery pack 16 is based on insufficient data and is labeled as a model with "Premature-Fit data.". If it is concluded in block 312 that there are sufficient data points in the regression model, the method continues in block 316. In block 316, the statistics of the model are tested to determine the reliability of the model, e.g., by comparison to thresholds suitable for the specific statistics to be tested. For example, the r-square for the model may be compared to a threshold value that may be 95%. At block 316, it may also be determined whether the slope of the modeled curve is negative or whether the slope of the modeled curve or portions thereof is positive; batteries have a decreasing capacity with increasing mileage so that a negative slope of the curve would be expected for "good" data. If the reliability tested statistics fail their respective tests (e.g., the r-square for the data from which the model was created is not above the threshold, or the slope of the model curve is not negative, or the slope of the modeled curve, or portions thereof, is positive), it may be inferred that the data from which the model was created is poorly fit (e.g., the data may be considered "noisy"). The data may accordingly be labeled as "bad fit data" in block 318. It will be appreciated that the tests in blocks 312 and 316 may determine whether a sufficiently reliable model may be generated using the battery data.However, if it is determined in block 316 that the statistics tested pass their respective tests (e.g., the r-square for the data is above the threshold and the slope of the curve representing the model is reliably negative), it may be inferred that the data that generated the model is good-fit data (i.e., the data is quantitative and qualitative to generate a reliable model of the battery pack 16) and the data may be marked as such in block 320. Thus, the final prediction of storage capacity over the travel distance of the battery pack 16 during the life of the battery pack 16 may be determined at block 322. This final prediction may be used for predicting the battery pack 16 (block 324). These may include predictions of whether the capacity of the battery pack 16 is likely to be less than expected during its life (block 326) and to predict the likelihood of whether the battery pack 16 is likely to be subject to a guarantee claim (block 328). This information may be sent to the backoffice 11 and used to notify the owner of the electric vehicle 10 that the life of the battery pack 16 may be shorter than expected.If the answer in block 316 is NO, the data may be labeled bad-fit data in block 318.Once the data of the electric vehicle 10 has been characterized as "good-fit data", this data can be forwarded from block 320 to block 332.With continued reference to FIG. 4, a portion 330 (clustering and regression prediction) of the algorithm disclosed herein may then be input. There, starting at block 332, "good-fit data" can be clustered from a fleet of vehicles. (An automobile manufacturer who has access to the data of a fleet of vehicles via the backoffice 11 can use this data with considerable advantage here). This good-fit data may then be clustered based on the average distance traveled per day (e.g., miles or kilometers) on the x-axis for the vehicles in that fleet and the average ambient temperatures on the y-axis at which the vehicles are operated in that fleet. The formation of clusters from a vehicle population is illustrated in FIG. 5, where clusters 350, 352, 354, and 356 are depicted. "Regression", as used herein, may be linear regression or other regression analysis.At block 334, regression models are created for the good-fit data of each cluster. At block 336, the model may be selected from the corresponding cluster as a "place holder" or "substitute" for the actual model of the battery pack 16 if the battery pack 16 has a model with "Premature Fit Data" or "Bad Fit Data.". The method continues to block 338 when the battery pack 16 has a "Pressure-Fit Data" model and the regression model for the corresponding cluster is preliminarily used as a model for the capacity relative to the running performance of the battery pack 16. However, from block 336, if the battery pack 16 has a "bad-fit data" model and the regression model for the corresponding cluster is preliminarily used as a predicted capacity-to-distance model for the battery pack 16, the method proceeds to block 340. Since the battery 16 does not yet have its own model with "fit data", it can be provisionally assumed that the model of the battery 16 is comparable to the models of the other vehicles in the same cluster. That is, it may be preliminarily assumed that the model of capacity relative to kilometers traveled for the battery pack 16 of the electric vehicle 10 generally matches the "good-fit data" models of other vehicles operating under similar conditions.FIG. 3 shows an example of a model with "bad-fit data" or "premature-fit data" and the corresponding correction. (The x-axis of the graph in FIG. 3 may be the distance traveled by the electric vehicle 10, which may be represented by miles or kilometers registered on the odometer of the electric vehicle 10. The y-axis of the graph of FIG. 3 may be the electrical storage capacity of the battery pack 16 in suitable units, such as amp-hours, kilowatts-hours, or other suitable units). First, it can be assumed that a new battery, at the beginning of its life ("BOL"), labeled "0K" (or zero thousand) miles or kilometers on the x-axis of FIG. 3, has a capacity known within a close tolerance, as this battery has just been manufactured and has not yet experienced any use or aging. One example: the BOL capacity 210 of the battery pack 16 may be about 181 amp-hours, as shown in the example example of FIG. 3. A regression model using good-fit data from vehicles in the same cluster as electric vehicle 10 may be represented as curve 212 in FIG. 3. When a model for the battery pack 16 is calculated and the model is based on "bad-fit data" or "premature-fit data", the model may have distortion. A "raw" capacitance curve segment with distortion is represented as curve segment 214. This distortion 215 may be corrected by knowing that curve 212 from BOL 210 would have to have a decrease in capacitance that would generally result in the curve for battery pack 16 being in curve segment 216. The calculated model for the battery pack 16 is adjusted accordingly.Additionally, curve segment 214 (a segment of the "raw" capacity model for battery pack 16 before bias correction) and curve segment 216 (a segment of the "raw" capacity model for battery pack 16 after bias correction) may have noise. It is naturally expected that the storage capacity of the battery decreases with time. However, portions of the curve 200 illustrated by portions 217 aand 217 bthat have increasing (or positive) slope indicate noise in the capacity data for the battery pack 16. Such noise may be a reason for the conclusion that the model for the battery pack 16 is based on "bad fit data" (see also block 316 of FIG. 4 ).Calculating a model for the capacity as a function of the driving power based on "bad-fit data" or "premature-fit data" for the battery pack 16 of the electric vehicle 10 may result in the curve 200. Using an alternative model with good-fit data based on a similarly stored group of vehicles may result in curve 212. The use of curve 200 may result in a premature prediction of a shortened life of the battery pack 16 (i.e., a misestimate of the life of the battery pack 16 by an amount of 230, e.g., because the long-term capacity degradation predicted by curve 200 is greater than that predicted by curve 212). Or, the use of curve 200 may result in a premature prognosis that the battery pack 16 will have a greater capacity degradation (i.e., a misestimate of the amount of degradation by an amount 232, e.g., because the long-term capacity degradation predicted by curve 200 is greater than that predicted by curve 212). Thus, for battery prediction, the model resulting from curve 212 may be used as a replacement model for the actual model for battery pack 16 until battery pack 16 has a model created from "good-fit data.".While the graph in FIG. 3 depicts storage capacity (y-axis) versus distance traveled (x-axis), the models may alternatively or additionally be created based on storage capacity versus operating time of the battery pack 16 in the electric vehicle 10 (e.g., the number of months the battery pack 16 was installed in the electric vehicle 10). The models may alternatively or additionally be created based on storage capacity as a function of the operating time of the electric vehicle 10 with the battery pack 16 installed (i.e., the cumulative elapsed time that the electric vehicle 10 was actually powered by the battery pack 16). The embodiments in this disclosure are to be understood accordingly. Moreover, the predictions described herein based on a model of storage capacity as a function of distance traveled and the predictions based on a model of storage capacity as a function of one or both of the time measures discussed in this section may be used in combination; they are not mutually exclusive.Once a model has been developed for the battery pack 16 that uses either good-fit data, or a provisional model or a replacement model (in the case of "premature-fit data" or "bad-fit data"), it may be used for predicting the battery pack 16 and / or the cell groups and / or battery modules. Here, the term "prognosis" is used to refer to a prediction of the future performance or state of the battery pack 16 and / or the cell groups and / or the battery modules. Of course, a first prediction is provided by the capacity-to-mileage model itself (whether it be a model such as curve 200 if it results from "good-fit data", or a replacement model such as the model represented by curve 212). The model predicts the storage capacity of the battery pack 16 over the distance traveled by the electric vehicle 10."Prognosis" in this disclosure may also refer to predicting the difference in capacity between the cell groups of the battery pack 16 and / or between the battery modules of the battery pack 16; this may be used, for example, to isolate one or more battery cells or battery modules that are predicted to behave differently than the others, possibly due to one or more expected failure modes or failure mechanisms. "Prognosis" in this disclosure may also refer to predicting which failure modes or failure mechanisms might develop or develop in the battery pack 16 and / or the cell groups and / or the battery modules.The prediction can be performed via one or more algorithms for capacity monitoring. Such an algorithm may compare capacity models of various cell groups in the battery pack 16 at a particular mileage. Various comparisons can be made, such as the following:Table 1 Table 1(a)max[C CGi, i ∈[0, n]] - min[C CGi, i ∈[0, n]](b)max[C CGi, i ∈[0, n]] - avg[C CGi, i ∈[0, n]] or max[C CGi, i ∈[0, n]] - median[C CGi, i ∈[0, n]](c)avg[C CGi, i ∈[0, n]] - min[C CGi, i ∈[0, n]] or median[C CGi, i ∈[0, n]] - min[C CGi, i ∈[0, n]]where CG i is a cell group number ("CG") of the battery pack 16, n is the total number of cell groups in the battery pack 16, "max" is the maximum capacity among the "n" cell groups, "min" is the minimum capacity among the "n" cell groups, "avg" is the average capacity of all "n" cell groups, and "median" is the average capacity of all "n" cell groups. If one or more of the values in the above table are above a threshold, a predicted loss storage capacity flag of the battery pack 16 may be set and the suspect cell group(s) may be predicted to become / become faulty (e.g., a shorter than design lifetime or a greater than design loss of storage capacity over time). The suspect cell group(s) may / may be identified for predictions aligned with the root cause(s) for the predicted loss of storage capacity. This algorithm may be applied with respect to a certain future mileage. Illustratively, line (a) in Table 1 shows the calculation of the difference between the maximum estimated capacity of a cell group and the minimum estimated capacity of a cell group, and compares this difference with a threshold value. Line (b) in Table 1 illustrates the calculation of the difference between the maximum estimated capacity of one cell group and the average or median of the estimated capacities of all "n" cell groups and compares this difference with a threshold value. Line (c) in Table 1 illustrates the calculation of the difference between the average or median of the estimated capacities of all "n" cell groups and the estimated minimum capacity of a cell group and compares this difference with a threshold value.Instead of carrying out the above prediction at the cell group level, it can also be carried out at the battery module level (e.g. between battery module 38 and battery module 40 and all other battery modules within battery pack 16). For purposes of cell group level prognosis, the BMS controller 22 should be understood to have electrical connections within the cell group level battery pack 16. For purposes of module-level prognosis, the BMS controller 22 should have electrical connections within the module-level battery pack 16.Another algorithm that may be employed may use the predicted capacity loss (which may also be referred to as estimated capacity decay) for individual cell groups. In this case, the percentage capacity decay over time can be calculated as follows: where CG i _capacity(t) is the capacity of the cell group CG i at any time t and CG i _capacity(0) is the capacity of the cell group CG i at time 0. If one of the calculations of the capacity decay is above a threshold, a flag for the suspect cell group(s) may / may be set and the suspect cell group(s) identified. It is to be expected that this cell group(s) will become / become faulty (e.g., a shorter than the structurally predefined lifetime or a greater than the structurally predefined loss of storage capacity over time).Instead of performing the above prediction at the cell group level, it may also be performed at the module level (e.g., between the battery module 38 and the battery module 40 and all other battery modules within the battery pack 16). It is understood that the BMS controller 22 has connections within the battery pack 16 to establish electrical connections within the battery pack 16 at the module level.Monitoring the capacity of the battery pack 16 as described above may also be coupled with monitoring the resistance of the battery pack 16 over time, which is an additional prognosis instrument. (Such resistance monitoring may be done at the cell group, module, or battery pack level, may be performed at regular intervals, and the data may be extrapolated to the future to produce a model of resistance versus mileage or a model of resistance versus time.). For example, taking the depiction of three vehicles, vehicle 400, vehicle 402, and vehicle 404 in FIG. 6, predictions via a combined algorithm 410 that examines the projected cell group resistances over time and the projected cell group capacitances over time may show that over comparable time periods, one or more cell groups of the battery pack of the electric vehicle 400 have a more pronounced increase in resistance over time and / or a more pronounced decrease in capacitance over time (see plots 450 and 452 of resistance (R) over time and storage capacitance (C) over time compared to the battery packs of vehicle 402 (plot 454 and plot 456) and vehicle 404 (plot 458 and plot 460). This may be used as a prediction tool, where the curves shown in FIG. 6 are used by physical-based analysis or machine learning to isolate the causes of the predicted battery degradation. The curves shown in FIG. 6 may be compared to the curves of FIG. 7 generated based on an actual laboratory diagnosis of the causes of fault in batteries from a fleet of vehicles. Such causes may be, for example, a high moisture content in the battery electrolyte (curves 500 aand 500 bin FIG. 7 ), a metal contamination of the battery (curves 510 aand 510 bin FIG. 7 ), a loss of battery electrolyte (curves 520 aand 520 bin FIG. 7 ) as compared to a baseline (curves 530 aand 530 bin FIG. 7 ) detected over a vehicle and battery population. The lithium plating of a battery electrode may also be a cause of the deterioration of the battery. It will be recalled that a vehicle manufacturer having a large fleet of customer vehicles on the road and access data about the vehicles via a telecommunications network also including the backoffice 11 may have access to a large amount of data. These analyses of the causes of battery quality degradation may be passed to the development and manufacturing departments responsible for the battery pack 16 to make improvements in design and manufacture as appropriate.Embodiments of the present disclosure are described herein. It is to be understood, however, that the disclosed embodiments are merely examples and other embodiments may take various and alternative forms. The figures are not necessarily to scale; some features could be exaggerated or minimized to show details of particular components. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a representative basis for teaching one skilled in the art to variously employ the present disclosure.Moreover, the embodiments shown in the drawings or the features of various embodiments mentioned in the present description are not necessarily to be understood as embodiments which are independent of one another. Rather, it is possible that each of the features described in one of the embodiments may be combined with one or more other desired features of other embodiments, resulting in other embodiments that are not described in words or by reference to the drawings. Accordingly, such other embodiments are within the scope of the appended claims. Moreover, this disclosure expressly includes combinations and sub-combinations of the elements and features set forth above and below.

Claims

An electric vehicle comprising: an electric motor for powering the electric vehicle; a battery adapted to provide electrical power to the electric motor to power the electric vehicle, and one or more controllers programmed in conjunction with the following instructions: measure battery data of the battery; use the battery data to generate a model of storage capacity of the battery as a function of distance travelled by the electric vehicle; evaluate reliability of the model; if the model is evaluated to be sufficiently reliable, a prediction for the battery is performed using the model; If the model is judged not to be sufficiently reliable, a replacement model of the storage capacity of the battery depending on the distance travelled by the electric vehicle is used, wherein the replacement model is based on battery data of one or more other vehicles, and a prediction is made for the battery using the replacement model.The electric vehicle of claim 1, further comprising selecting the one or more other vehicles based on the one or more other vehicles having operational characteristics similar to those of the electric vehicle.The electric vehicle of claim 2, wherein the operating characteristics comprise the average travel distance per day.The electric vehicle of claim 2, wherein the operating characteristics comprise the average ambient temperature.The electric vehicle of claim 1, wherein the model is evaluated as not sufficiently reliable if the battery data has an insufficient number of data measurements.The electric vehicle according to claim 1, wherein the model is judged not to be sufficiently reliable when the model has a storage capacity increasing with a travel distance or when the model has noise.The electric vehicle of claim 1, wherein the prediction comprises comparing the predicted storage capacities between the individual parts of the battery.The electric vehicle of claim 1, wherein the prediction comprises comparing the predicted storage capacity loss between the individual parts of the battery to a threshold.The electric vehicle of claim 1, wherein the prognosis comprises using the model or the replacement model in combination with internal resistance measurements of the battery.

Images