DETECTION OF ISOLATION IN A RECHARGEABLE ENERGY STORAGE SYSTEM

Abstract

An assessment of insulation loss in a rechargeable energy storage system (RESS) is planned. The assessment may include collecting insulation resistance data for a large number of energy storage systems operating on board a fleet of vehicles, applying statistical methods and engineering rules to remove erroneous spikes and duplicate values ​​from the insulation resistance data, generating one or more insulation rules based on the insulation resistance data to represent thresholds for determining whether insulation loss has occurred for the associated energy storage system, and identifying an insulation problem type for the energy storage systems where insulation loss has been detected.

Description

[0001] The present description relates to the detection of insulation of a rechargeable energy storage system (RESS), such as, but not necessarily limited to, the determination of insulation loss in an energy storage system included on board a vehicle to store and deliver electrical energy to a traction motor.

[0002] A rechargeable energy storage system (RESS) may be configured to store and deliver electrical energy, with one of the more common types of energy storage systems comprising a plurality of battery cells arranged in one or more battery packs. Such energy storage systems may be installed onboard a vehicle to store and deliver electrical energy to a traction motor, which converts the electrical energy into mechanical energy to propel the vehicle. To achieve desired operating levels, it may be advantageous to electrically isolate one or more of the battery packs from one or more other battery packs.Due to the advantages associated with maintaining electrical insulation, it is an object of the invention to detect when a loss of electrical insulation has occurred or is likely to occur so that remedial action can be taken.

[0003] US 2023 / 0 109 419 A1 describes an insulation monitoring method and system for a traction battery and an apparatus for solving the problem of how to accurately predict the risk of insulation deterioration of the traction battery before the insulation deterioration occurs in the traction battery, so as to provide early warning of failures in the traction battery. This method collects statistical data on a large number of insulation resistance values ​​over a long period of time, and determines whether the traction battery is at risk of insulation deterioration through analysis based on the statistical data result. If the traction battery is at risk of insulation deterioration, alarm information is output.This method can predict insulation deterioration before traction battery insulation deterioration occurs based on statistical data analysis performed on a large number of insulation resistance values ​​over a long period of time. This allows the user to perform timely battery maintenance before traction battery insulation deterioration occurs, thus preventing traction battery failures.

[0004] CN 1 17 154 249 A describes a safety management method and system for an energy storage battery cloud system, and the method comprises the following steps: constructing a battery cloud management system comprising a BMS protocol module for the energy storage battery, an EMS module for energy management, a communication module and a battery cloud platform, and activating the BMS protocol module for the energy storage battery to collect battery data information and upload the battery data information to the EMS module for energy management of a local energy storage, and the communication module uploads the data information to a battery cloud platform for internal short-circuit early warning and external short-circuit early warning.Energy storage battery data is uploaded to the cloud in real time, the battery cloud platform calculates in real time, once any calculation result is abnormal, a control command for the energy storage battery will be issued at the first time, and in the meantime, the monitoring personnel will be notified to intervene in the investigation; through the working condition that the charging and discharging of the energy storage battery is regular and stable, compared with a new energy car battery and other scenes, the monitoring calculation is accurate, and the safety of the battery and equipment in the charging and discharging process is guaranteed.

[0005] The description relates to the detection of insulation loss in a rechargeable energy storage system (RESS). The systems and methods described herein can be used to detect insulation loss in a variety of energy storage systems, including energy storage systems of the type having a plurality of battery cells arranged in one or more battery packs to store and deliver electrical energy to a traction motor of a vehicle. The insulation loss can be determined based on systematic processes for monitoring and tracking insulation resistance signals to predict insulation loss using statistical models, machine learning, and / or physics-based algorithms to predict insulation loss and identify an insulation type.

[0006] The description relates to a system for detecting / predicting insulation loss of high-voltage battery packs and identifying the insulation type. The system can use energy storage system insulation resistance, temperature, and / or a combination of features to enable accurate detection, identification, and robust determination of noise resistance based on a variety of factors. The system can monitor a fleet of vehicles to learn nominal limits and signature patterns sufficient to distinguish between coolant leak / water ingress and cell corrosion, to gain insights into insulation root causes, to monitor the progression of insulation loss, and to send proactive alerts / notifications to provide timely warnings to customers, and / or to manage vehicle operations.The system can optionally be implemented with a passive approach so that accompanying detection can take place during vehicle operation and optionally without affecting driving operations.

[0007] The object of the invention is achieved by a system for assessing insulation loss for a rechargeable energy storage system, RESS. The system comprises a data acquisition module configured to determine insulation resistance data for a plurality of energy storage systems operating on board a vehicle fleet, wherein the energy storage systems are configured to store and supply electrical energy for a traction motor of the associated vehicle. The system further comprises an intelligent filtering and reprocessing module configured to apply statistical methods and engineering rules to remove erroneous peaks and duplicate values ​​from the insulation resistance data.The system further comprises a prediction module configured to generate insulation rules from healthy data contained in the insulation resistance data, wherein the insulation rules serve to evaluate whether the insulation resistance data is representative of an insulation loss. The system further comprises an evaluation module configured to characterize each insulation loss.

[0008] According to one embodiment, the prediction module is configured to generate the isolation rules that include thresholds to determine whether the loss of isolation has occurred.

[0009] According to another embodiment, the evaluation module is configured to apply the insulation rules to the insulation resistance data to identify an insulation problem type for the energy storage systems in which insulation loss has been determined.

[0010] According to another embodiment, the evaluation module is configured to determine the insulation problem type to correspond to a water ingress event when the insulation resistance data exceeds a water ingress resistance signature included as one of the insulation rules.

[0011] According to another embodiment, the evaluation module is configured to determine the insulation problem type to correspond to a cell corrosion event when the insulation resistance data exceeds a cell corrosion resistance signature included as one of the insulation rules.

[0012] According to another embodiment, the intelligent filtering and reconditioning module is configured to perform multi-packet processing, wherein the multi-packet processing removes the insulation resistance data associated with each energy storage system having multiple packs with similar faulty behavior.

[0013] According to another embodiment, the prediction module is configured to perform distribution fitting and threshold learning to generate control limits and prediction thresholds for the fleet.

[0014] According to another embodiment, the assessment module is configured to use a severity index and a severity score to determine a severity level and provide a customer with early warning for each vehicle with an insulation loss.

[0015] According to another embodiment, the system includes a warning module configured to send a warning to an operating system on board each vehicle to which the early warning is to be provided.

[0016] According to another embodiment, the warning module is configured to generate the warning containing instructions to prompt an operator to service the associated vehicle.

[0017] According to another embodiment, the warning module is configured to generate the warning to inform an operator of the associated vehicle that a shutdown command has been issued to prevent further use of the associated energy storage system.

[0018] In one example, a method for evaluating insulation loss for a rechargeable energy storage system (RESS) is described, which serves to better understand the system.The method includes collecting insulation resistance data for a plurality of energy storage systems operating onboard a fleet of vehicles, optionally wherein each energy storage system operates onboard a vehicle to store and provide electrical energy to a traction motor, applying statistical methods and engineering rules to remove erroneous peaks and duplicate values ​​from the insulation resistance data, generating one or more insulation rules based on the insulation resistance data, the insulation rules optionally representing thresholds to determine whether an insulation loss has occurred, applying the insulation rules to the insulation resistance data to identify one or more energy storage systems having an insulation problem, and identifying an insulation problem type for the energy storage systems identified as having an insulation problem.

[0019] In one example, the method includes deriving graphical representations of the insulation resistance data as signatures that represent the operation of the energy storage system in healthy and unhealthy states, respectively.

[0020] In one example, the method includes performing peak filtering of the insulation resistance data according to engineering rules to remove erroneous peaks from the graphical representations.

[0021] In one example, the method includes communicating an alert to an operating system on board one or more vehicles in which the isolation problem has been detected.

[0022] In one example, the method includes instructions in the alert message to notify the operator that a shutdown command has been issued to prevent further use of the associated energy storage system.

[0023] These features and advantages, as well as other features and advantages of the present teachings, will become apparent from the following detailed description of the embodiments of the present teachings, taken in conjunction with the accompanying drawings. It should be understood that, although the following figures and embodiments may be described separately, individual features thereof may be combined to form additional embodiments.

[0024] The accompanying drawings, which may be incorporated in and constitute a part of this specification, illustrate embodiments of the description and, together with the description, serve to explain the principles of the description. Fig. 1 shows a system for detecting insulation loss. Fig. Figure 2 shows a flowchart of a method for detecting insulation loss. Fig. 3 shows a flowchart of a filtering process. Fig. Figure 4 shows a flowchart of a forecasting procedure. Fig. Figure 5 shows a flowchart of the offline learning process. Fig. Figure 6 shows a flowchart of an evaluation process. Fig. 7 shows a graph of a water intrusion type. Fig. Figure 8 shows a graph of a cell corrosion type.

[0025] As appropriate, detailed embodiments of the present description may be described herein; however, it is to be understood that the described embodiments are merely exemplary of the description, which may be embodied in various and alternative forms. The figures are not necessarily to scale; some features may be exaggerated or reduced to show details of particular components. Therefore, specific structural and functional details described herein are not to be interpreted as limiting, but merely as a representative basis for showing one skilled in the art how to variously apply the present description.

[0026] Fig. 1 shows a system 10 for detecting insulation loss according to one non-limiting aspect of the present description. The system 10 may operate with a fleet of vehicles, which may include a plurality of related and / or unrelated vehicles similar to a vehicle 12. The vehicle 12, which may be interchangeably referred to as an electric or hybrid vehicle 12, may include an electric motor 14 that converts electrical energy into mechanical energy to perform work, for example, to mechanically drive a driveline 16 to propel the vehicle. The vehicle 12 is depicted as a hybrid vehicle because the driveline 16 optionally includes an internal combustion engine (ICE) 18 for generating mechanical energy. The driveline 16 may include components that facilitate the transfer of rotational power from the electric motor 14 to one or more of the wheels 20, 22, 24, 26.The vehicle 12 may include a rechargeable energy storage system (RESS) 30 for storing and delivering electrical energy to the electric motor 14 and / or other systems, buses, etc., connected to one or more electric buses 32 on board the vehicle 12. The RESS 30 may include a plurality of battery cells or other energy storage units arranged into a plurality of battery packs or other unit groupings. It may be desirable to detect the insulation between one or more battery packs and one or more other battery packs of the RESS 30, which may be based on assessing whether the electrical insulation between one or more battery packs is degraded, lost, or likely to be disrupted in a manner that may cause the RESS 30 and / or the vehicle 12 to operate differently than intended.

[0027] The vehicle 12 may include a vehicle controller 34 to facilitate monitoring, controlling, measuring, and otherwise managing the operation, performance, etc., onboard the vehicle 12, which may include taking measurements, taking readings, or otherwise collecting RESS data from the RESS 30. The system 10 may include an insulation detection controller 36 in a back office or other location that can communicate with the vehicle fleet 12, for example, via wireless signaling with the respective controllers 34. The insulation detection controller 36 may be configured to perform a variety of operations, processes, etc., to facilitate the insulation loss detection contemplated herein.The insulation detection controller 36 and the vehicle controller 34 may each operate based on and / or in accordance with one or more corresponding processors executing a corresponding plurality of non-transitory instructions stored on an associated computer-readable storage medium. The non-transitory instructions may be executed for the purpose of transmitting messages, executing software or other algorithms, performing calculations, and otherwise facilitating data exchange with the vehicles 12 to perform the various insulation loss detection activities described herein. The operation of the insulation detection controller 36 may be assigned to a data collection module 40, an intelligent filtering and reprocessing module 42, a prediction module 44, and an evaluation module 46.

[0028] Fig. 2 shows a flowchart 50 of a method for detecting insulation loss according to one non-limiting aspect of the present description. Block 52 refers to a collection process where the data collection module 40 may determine RESS data, or more specifically, insulation resistance data, for a plurality of RESSs operating onboard the vehicle fleet 12. The insulation resistance data may be based on the vehicle controllers 34 using sensors or other features onboard the vehicle 12 to measure or otherwise evaluate the insulation resistance associated with the RESS 30 and / or the battery packs. For example, the insulation resistances may be based on measuring the input and output voltages of the battery packs and / or the entire RESS 30 and utilizing the characteristics of the associated battery pack to estimate the insulation resistance from the measured voltages.This may include the vehicle controller 34 generating insulation-related signals for transmission to the insulation detection controller 36, which may optionally include adjustments to the insulation resistance estimates to account for temperature and / or other influencing factors. The data collection module 40 may be configured to calculate the insulation resistances in addition to, or instead of, the vehicle controller 34. The resulting insulation resistance data may include vehicle identifiers, RESS identifiers, and / or other information to associate the individual insulation resistances with a particular vehicle so that a customer, facility, etc., associated with the vehicle can be determined.

[0029] Block 54 refers to a filtering process in which the intelligent filtering and reprocessing module 42 may apply statistical methods and engineering rules to remove erroneous peaks and / or duplicate values ​​from the insulation resistance data. The filtering process may include the intelligent filtering and reprocessing module 42 receiving the insulation resistance data from the data collection module 40 and processing the insulation resistance data into filtered insulation resistance data from which duplicate values, erroneous peaks, etc., have been removed. Fig. 3 shows a flowchart 56 of the filtering process according to one non-limiting aspect of the present description. Block 62 refers to a learning process in which the intelligent filter reprocessing module 42 may apply statistical techniques to learn acceptable noise levels for the insulation resistance data. Block 64 refers to a cleaning process in which the intelligent filter reprocessing module 42 may remove standard, erroneous, and duplicate values ​​from the insulation resistance data. Block 66 refers to an engineering rules process in which the intelligent filter and reprocessing module 42 may apply engineering rules to remove erroneous peak values. Block 68 refers to a refinement process for applying signal processing techniques and the learned noise levels to filter the signals, i.e., the insulation resistance data.Block 70 refers to a multi-package method in which the insulation resistance data associated with each RESS with multiple packages exhibiting similar faulty behavior can be removed. Block 72 refers to the generation of filtered insulation resistance data, i.e., the output of the insulation resistance data after removing faulty peaks, duplicate values, etc.

[0030] Back to Fig. 2, block 76 may refer to a prediction process in which the prediction module 44 may generate insulation rules from healthy data included in the filtered insulation resistance data. The insulation rules may be used to assess whether the insulation resistance data represents a loss of insulation for the associated RESS 30. The prediction process may be operated to identify healthy data and unhealthy data included in the filtered insulation resistance data based on identifying the healthy data with healthy operation of the RESS 30 and the unhealthy data with unhealthy operation of the RESS 30. The prediction process may include generating the insulation rules to include thresholds for determining whether the loss of insulation has occurred in block 78. Fig. Figure 4 shows a flowchart 82 illustrating further details of the prediction process according to a non-limiting aspect of the present description. Block 84 may refer to an additional feature process, where the prediction module 44 may derive additional features from the insulation resistances (e.g., moving window-based average resistance value (r a ), slope, rate of change of slope, etc.), optionally with additional temperature compensation if necessary. Block 86 may refer to an offline learning process where the prediction module may use the healthy data (for example, the RESS data without leaks / loss of insulation) and the unhealthy data (for example, the RESS data with leaks / loss of insulation) to learn the nominal model and the detection threshold or the outlier decision boundary.

[0031] Fig. 5 shows a flowchart 88 of the offline learning process according to one non-limiting aspect of the present description. Block 90 refers to a data process for identifying the healthy RESS data, for example, based on the filtered insulation resistance data. Block 92 refers to an additional feature process, wherein the prediction module 44 may derive additional features from smooth resistances, slope, rate of change of slope, etc., for the insulation resistance data. Block 94 refers to a distribution fitting process in parameter learning, where the prediction module may determine when little or no bimodality is observed to use bootstrapping to generate distributions, fit using kernel density estimation in the case of distributions with multimodality, and / or establish control limits using population statistics.Block 96 refers to data processing to identify the unhealthy RESS data, for example, based on the filtered insulation resistance data. Block 98 refers to a threshold learning process, where the prediction module can find the thresholds used to determine insulation problems through an optimization problem, for example, maximizing true positives and minimizing false positives (for example, Gini impurity score) or engineering rules based on the best unacceptable part. Alternatively or additionally, the interquartile range (IQR) could be used to set the thresholds, or a machine learning (ML) algorithm could be trained to detect anomalous patterns.In this way, within the scope of the present description, nominal insulation or insulation loss detection models for the battery packs may be created based on several approaches, which may include the use of one or more of the following insulation loss detection methods: physically based methods (rules derived from engineering), statistical methods, distribution fitting and application of 6-sigma rules, signal-based outlier detection (statistical IQR), and / or artificial intelligence (AI) / ML-based (for example, a class of support vector machine (SVM)). Block 100 refers to an output process where the prediction module may generate thresholds or other rules for comparison with the vehicle fleet to detect insulation.

[0032] Back to Fig. 4, block 104 refers to an application process where the prediction module can apply learned nominal models to detect and confirm an insulation loss and / or identify the insulation problem type. The process can include determining an insulation loss when resistive characteristics are above the threshold or outside the nominal limit and / or when the problem is present for 'n' consecutive samples / trips. Additionally or alternatively, the method can include performing moving window-based aggregation to determine if a problem exists for m consecutive windows. The process can be performed for each individual vehicle, where the RESS data collected from each vehicle can be analyzed so that a decision can be made at block 78 as to whether a problem has been detected and confirmed.Block 106 may relate to determining the insulation is adequate so that monitoring can continue for the corresponding vehicle. Block 108 may relate to determining an insulation problem, such as an existing or predicted loss of insulation. Back to . Fig. 2, a problem determination may result in the prognostic module forwarding relevant information, RESS data, insulation data, etc. to the assessment module 46 for further analysis.

[0033] Fig. 6 shows a flowchart 112 of an evaluation process according to one non-limiting aspect of the present description. The evaluation process may be performed by the evaluation module 46 in response to the information provided by the prediction module for one or more detected isolation problems. Block 114 refers to a severity evaluation process in which the evaluation module performs a severity evaluation for the detected isolation problem. The process may include determining how far the signals of any of the vehicles may be from the fleet median and, based thereon, relating this difference relative to the threshold to quantify a low, medium, or high severity, or counting the number of cases i in which the problem is detected in a window size w. For example, if i=10, the worst case may correspond to i=10 and the least concern may correspond to i=1.Block 116 relates to a signature pattern evaluation process, where the evaluation module may apply the insulation rules to the insulation resistance data to identify an insulation problem type for the RESSs detected to have insulation loss. The evaluation module may be configured to determine the insulation problem type to correspond to a water intrusion event if the insulation resistance data exceeds a water resistance signature included as one of the insulation rules, such as according to a water intrusion resistance signature chart 120 shown in FIG. Fig. 7. The diagram 120 may include a vertical axis 122 representing insulation resistance and a horizontal axis 124 representing time, so that a water intrusion type may correspond to a signal 126 representing insulation resistance decreasing rapidly or gradually and remaining below a normal or nominal level 128. The evaluation module 46 may be configured to determine the insulation problem type corresponding to a cell corrosion event when the insulation resistance data exceeds a cell corrosion resistance signature included as one of the insulation rules, such as according to a graph 132 of the cell corrosion insulation resistance signature included in Fig. 8. The graph 132 may include a vertical axis 134 representing insulation resistance and a horizontal axis 136 representing time, so that a cell corrosion problem may correspond to a signal 138 representing a rapid drop in insulation resistance and a frequent return to a normal or nominal value 140.

[0034] Back to Fig.2, block 144 may refer to a notification process where the alerting module may generate a notification for transmission to engineering or other entities responsible for monitoring or otherwise working with the vehicle fleet. Block 146 may refer to a review process where the notification may be analyzed to determine whether a customer, i.e., the person who ordered the vehicle, and / or a manufacturer should be notified of the isolation issue. Block 148 may refer to the alerting module that is currently responding sending an alert to an operating system onboard each vehicle that is to receive the early warning, which may include instructions to request an operator to service the associated vehicle and / or to notify an operator of the associated vehicle that a shutdown command has been issued to prevent further use of the associated RESS.

Claims

[1] System (10) for assessing insulation loss in rechargeable energy storage systems, RESS, (30) comprising: a data acquisition module (40) configured to determine insulation resistance data for a plurality of energy storage systems (30) operating on board a fleet of vehicles (12), the energy storage systems (30) being configured to store and deliver electrical energy to a traction motor of the associated vehicle (12); an intelligent filtering and reprocessing module (42) configured to apply statistical methods and engineering rules to remove erroneous peaks and duplicate values ​​from the insulation resistance data; a prediction module (44) configured to generate insulation rules from healthy data included in the insulation resistance data, the insulation rules being used to evaluate whether the insulation resistance data is representative of an insulation loss; and an evaluation module (44) configured to characterize each insulation loss. [2] System (10) according to claim 1, wherein: the prediction module (44) is configured to generate the insulation rules that include thresholds to determine whether the insulation loss has occurred. [3] System (10) according to claim 2, wherein: the evaluation module (44) is configured to apply the insulation rules to the insulation resistance data to identify an insulation problem type for the energy storage systems (30) in which an insulation loss has been determined. [4] System (10) according to claim 3, wherein: the evaluation module (44) is configured to determine the insulation problem type to correspond to a water ingress event when the insulation resistance data exceeds a water ingress resistance signature included as one of the insulation rules. [5] System (10) according to claim 3, wherein: the evaluation module (44) is configured to determine the insulation problem type to correspond to a cell corrosion event when the insulation resistance data exceeds a cell corrosion resistance signature included as one of the insulation rules. [6] System (10) according to claim 1, wherein: the intelligent filtering and reconditioning module (42) is configured to perform multi-pack processing, wherein the multi-pack processing removes the insulation resistance data associated with each energy storage system (30) having multiple packs with similar faulty behavior. [7] System (10) according to claim 1, wherein: the forecast module (44) is configured to perform distribution fitting and threshold learning to generate control limits and forecast thresholds for the fleet. [8] System (10) according to claim 1, wherein: the assessment module (44) is configured to use a severity index and a severity rating to determine a severity level and provide a customer with early warning for each vehicle (12) with the insulation loss. [9] System (10) according to claim 8, further comprising: a warning module configured to send a warning to an operating system on board each vehicle (12) to which the early warning is to be provided. [10] System (10) according to claim 9, wherein: the warning module is configured to generate the warning containing instructions to prompt an operator to service the associated vehicle (12).

Images