WO2024098357A1 - 动力电池控制策略确定方法、设备控制方法和装置 - Google Patents

动力电池控制策略确定方法、设备控制方法和装置 Download PDF

Info

Publication number
WO2024098357A1
WO2024098357A1 PCT/CN2022/131261 CN2022131261W WO2024098357A1 WO 2024098357 A1 WO2024098357 A1 WO 2024098357A1 CN 2022131261 W CN2022131261 W CN 2022131261W WO 2024098357 A1 WO2024098357 A1 WO 2024098357A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
power battery
strategy
fault
fault type
Prior art date
Application number
PCT/CN2022/131261
Other languages
English (en)
French (fr)
Inventor
李瑞林
张堃
罗思佳
赵微
Original Assignee
宁德时代新能源科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 宁德时代新能源科技股份有限公司 filed Critical 宁德时代新能源科技股份有限公司
Priority to PCT/CN2022/131261 priority Critical patent/WO2024098357A1/zh
Publication of WO2024098357A1 publication Critical patent/WO2024098357A1/zh

Links

Images

Definitions

  • the present invention relates to the field of battery technology, and in particular to a method for determining a power battery control strategy, a device control method, a device, a computer device, a computer-readable storage medium, and a computer program product.
  • Power batteries refer to power sources that provide power for tools. With the development of science and technology and the continuous progress of society, the application areas of power batteries are constantly expanding. They are not only used in electric vehicles such as electric bicycles, electric motorcycles, and electric cars, but also in military equipment and aerospace and other fields. Traditional power battery control strategies are formulated based on laboratory simulation and testing of equipment used in power batteries, and it is difficult to simulate real working conditions. How to improve the control accuracy of power batteries is an urgent problem to be solved.
  • a power battery control strategy determination method a device control method, an apparatus, a computer device, a computer-readable storage medium, and a computer program product are provided.
  • the present application provides a method for determining a power battery control strategy, comprising:
  • the target characteristics of each fault type are analyzed and obtained;
  • the above power battery control strategy determination method obtains sampling data collected from the power battery and determines the features corresponding to different fault types according to the sampling data. According to the features corresponding to each fault type and the sampling data corresponding to each feature, the target features of each fault type are analyzed and obtained, and the power battery control strategy is generated according to the target features of each fault type, so as to update the power battery control strategy, thereby making the power battery control strategy more in line with the actual working conditions, making the control strategy more accurate, reducing the probability of fault occurrence, and extending the life of the power battery.
  • determining the features corresponding to different fault types according to the sampled data includes: performing a difference analysis on the sampled data collected from the power battery, and determining the features corresponding to the different fault types according to the difference analysis results. According to the difference analysis on the sampled data collected from the power battery, feature classification can be accurately performed to determine the features corresponding to different fault types.
  • the difference analysis includes at least one of statistical feature analysis, cross feature analysis between variables, and time series trend change feature analysis.
  • One or more of statistical feature analysis, cross feature analysis between variables, and time series trend change feature analysis can be selected for difference analysis based on actual needs.
  • the target features of each fault type are obtained by analyzing the features corresponding to each fault type and the sampling data corresponding to each of the features, including: screening the features corresponding to each fault type according to the sampling data of the features corresponding to each fault type to obtain the initial target features associated with each fault type; determining the target features of each fault type according to the sampling data of the initial target features associated with each fault type and the fault prediction model corresponding to each fault type.
  • the initial target features associated with each fault type are extracted by feature screening to avoid overlapping features with high correlation affecting the processing efficiency, and then the target features of each fault type are accurately obtained in combination with the fault prediction model.
  • the method of screening the features corresponding to each fault type according to the sampled data of the features corresponding to each fault type to obtain the initial target features associated with each fault type includes: preprocessing the sampled data of the features corresponding to each fault type to obtain the sampled data of each feature after preprocessing; screening the features corresponding to each fault type according to the sampled data of each feature after preprocessing to obtain the initial target features associated with each fault type.
  • the sampled data is first preprocessed to ensure the accuracy and completeness of the data, and then feature screening is performed to determine the initial target features associated with each fault type to improve the accuracy of feature association.
  • the preprocessing includes at least one of data cleaning and data completion.
  • One or more of data cleaning and data completion can be selected for data preprocessing according to actual needs.
  • the target feature of each fault type is determined based on the sampled data of the initial target feature associated with each fault type and the fault prediction model corresponding to each fault type, including: adding labels to the sampled data of the initial target feature associated with each fault type to generate training sample data of each fault type, wherein the labels include faulty equipment labels and normal equipment labels; performing model training based on the training sample data of each fault type to obtain the fault prediction model corresponding to each fault type; and determining the target feature of each fault type based on the fault prediction model corresponding to each fault type. Labels are added to the sampled data of the initial target feature associated with each fault type as training sample data for model training to ensure the accuracy of the fault prediction model.
  • the method further includes: performing model evaluation on the fault prediction model, and taking the fault prediction model that passes the evaluation as the final fault prediction model. Performing model evaluation on the trained fault prediction model and taking the fault prediction model that passes the evaluation as the final fault prediction model is more in line with actual needs.
  • the model evaluation of the fault prediction model and taking the fault prediction model that passes the evaluation as the final fault prediction model includes: performing business evaluation on the fault prediction model according to preset business standards; performing rationality evaluation on the fault prediction model that passes the business evaluation according to preset fault tree analysis results, and taking the fault prediction model that passes the rationality evaluation as the final fault prediction model.
  • the fault prediction model is simultaneously evaluated for business and rationality, and if both are evaluated, it is taken as the final fault prediction model, thereby improving the accuracy of model evaluation.
  • the control strategy of the power battery is determined according to the target characteristics of each fault type, including: generating a preliminary revised strategy of the power battery according to the target characteristics of each fault type; simulating the operation of the device according to the preliminary revised strategy to obtain simulated sampling data; and performing strategy evaluation according to the simulated sampling data and the fault prediction model to obtain the control strategy of the power battery.
  • the device operation is simulated to perform strategy evaluation to ensure that the control strategy of the power battery meets actual needs.
  • the strategy evaluation includes failure rate evaluation and/or service evaluation.
  • the revised strategy can be evaluated for failure rate and/or service according to actual needs.
  • the sampled data includes at least one of current data, temperature data, voltage data and insulation data.
  • One or more of current data, temperature data, voltage data and insulation data may be selected for collection to analyze the characteristics of each fault type.
  • the present application provides a device control method, comprising: determining a target revision strategy from a revision strategy of a power battery; the revision strategy of the power battery is obtained according to the above-mentioned power battery control strategy revision method; and sending the target revision strategy to a target device to update the control strategy of the power battery of the target device.
  • the above-mentioned equipment control method generates a revised strategy for the power battery according to the target characteristics of each fault type, so as to update the control strategy of the power battery, thereby making the control strategy of the power battery more in line with the actual working conditions, making the control strategy more accurate, reducing the probability of fault occurrence, and extending the life of the power battery.
  • the determination of the target revision strategy from the revision strategy of the power battery includes: determining the target device associated with each revision strategy according to the optimization target of the revision strategy; and taking all revision strategies associated with the target device as the target revision strategy. According to the optimization target of the revision strategy, taking all revision strategies associated with the target device as the target revision strategy to realize active update of the revision strategy of the target device.
  • determining the target revision policy from the revision policy of the power battery includes: according to an update request sent by the target device, obtaining the revision policy associated with the target device from the revision policy of the power battery as the target revision policy.
  • the update request can be sent through the target device to start the policy update operation according to actual needs.
  • the present application provides a device control method, comprising: obtaining a revised strategy for a power battery to be updated; the revised strategy for the power battery to be updated is obtained according to the above-mentioned power battery control strategy revision method; and updating the control strategy of the power battery of the device according to the revised strategy for the power battery to be updated.
  • the above-mentioned equipment control method generates a revised strategy for the power battery according to the target characteristics of each fault type, so as to update the control strategy of the power battery, thereby making the control strategy of the power battery more in line with the actual working conditions, making the control strategy more accurate, reducing the probability of fault occurrence, and extending the life of the power battery.
  • the obtaining of the revision policy of the power battery to be updated includes: receiving a target revision policy sent by a server as the revision policy of the power battery to be updated; the target revision policy is obtained by the server according to all revision policies associated with a target device, and the target device is determined according to an optimization target of the revision policy. According to the optimization target of the revision policy, all revision policies associated with the target device are used as the target revision policy to realize active update of the revision policy of the target device.
  • the obtaining of the revision strategy of the power battery to be updated includes: sending an update request to a server; receiving a target revision strategy returned by the server in response to the update request as the revision strategy of the power battery to be updated; the target revision strategy is obtained by the server obtaining the revision strategy associated with the optimization target and the target device from the revision strategy of the power battery according to the update request.
  • the update request can be sent through the target device to start the strategy update operation according to actual needs.
  • the present application provides a power battery control strategy determination device, comprising:
  • a data acquisition module used to acquire sampled data collected from the power battery, and determine features corresponding to different fault types according to the sampled data
  • a data analysis module used to analyze and obtain target features of each fault type according to features corresponding to each fault type and sampling data corresponding to each feature;
  • the strategy revision module is used to determine the control strategy of the power battery according to the target characteristics of each fault type.
  • the present application provides a device control apparatus, comprising:
  • a strategy determination module used to determine a target revision strategy from the revision strategy of the power battery; the revision strategy of the power battery is obtained according to the power battery control strategy determination device mentioned above;
  • the strategy sending module is used to send the target revision strategy to the target device to update the control strategy of the power battery of the target device.
  • the present application provides a device control apparatus, comprising:
  • a strategy acquisition module used to acquire a revised strategy of the power battery to be updated; the revised strategy of the power battery to be updated is obtained by the power battery control strategy determination device mentioned above;
  • the strategy updating module is used to update the control strategy of the power battery of the device according to the revised strategy of the power battery to be updated.
  • the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.
  • the present application provides a computer-readable storage medium having a computer program stored thereon, and the computer program implements the steps of the above method when executed by a processor.
  • the present application provides a computer program product, comprising a computer program, which implements the steps of the above method when executed by a processor.
  • FIG1 is a schematic diagram of an application environment of a method for determining a power battery control strategy in one embodiment
  • FIG2 is a flow chart of a method for determining a power battery control strategy in one embodiment
  • FIG3 is a flow chart of analyzing and obtaining target features of each fault type according to features corresponding to each fault type and sampling data corresponding to each feature in one embodiment
  • FIG4 is a flowchart of performing model training according to sampling data of initial target features associated with each fault type to obtain a fault prediction model corresponding to each fault type in one embodiment
  • FIG5 is a flow chart of generating a revised strategy for a power battery according to target characteristics of each fault type in one embodiment
  • FIG6 is a schematic diagram of step charging control under the original strategy in one embodiment
  • FIG7 is a schematic diagram of ranking faulty vehicles under an original strategy in one embodiment
  • FIG8 is a schematic diagram of step charging control under a new strategy in one embodiment
  • FIG9 is a schematic diagram showing a comparison of faulty vehicle rankings under the original strategy and the new strategy in one embodiment
  • FIG10 is a flow chart of a device control method in one embodiment
  • FIG11 is a flow chart of a device control method in another embodiment
  • FIG12 is a structural block diagram of a device for determining a power battery control strategy in one embodiment
  • FIG13 is a block diagram of a device control device in one embodiment
  • FIG14 is a block diagram of a device control apparatus in another embodiment
  • FIG. 15 is a diagram showing the internal structure of a computer device in one embodiment.
  • the term "and/or" is only a description of the association relationship of associated objects, indicating that three relationships may exist.
  • a and/or B can represent: A exists alone, A and B exist at the same time, and B exists alone.
  • the character "/" in this article generally indicates that the associated objects before and after are in an "or" relationship.
  • multiple refers to more than two (including two).
  • multiple groups refers to more than two groups (including two groups), and “multiple pieces” refers to more than two pieces (including two pieces).
  • the power battery is the power source that provides the power source for the tool. Most of them are valve-sealed lead-acid batteries, open-type tubular lead-acid batteries and lithium iron phosphate batteries, which have the characteristics of high energy, high power and high energy density.
  • the traditional power battery control strategy is formulated by combining the equipment used by the power battery for laboratory simulation and testing, and it is difficult to simulate the actual working conditions. Based on this, the present application provides a method for determining the power battery control strategy, by obtaining the sampled data obtained by collecting the power battery, and determining the characteristics corresponding to different fault types according to the sampled data.
  • the target characteristics of each fault type are analyzed, and the control strategy of the power battery is determined according to the target characteristics of each fault type, so as to update the control strategy of the power battery, so that the control strategy of the power battery is more in line with the actual working conditions, the control strategy is more accurate, the probability of faults is reduced, and the life of the power battery is extended.
  • the power battery control strategy determination method provided in the embodiment of the present application can be applied to the application environment shown in Figure 1.
  • the terminal 102 communicates with the server 104 through the network.
  • the data storage system can store the data that the server 104 needs to process.
  • the data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers.
  • the terminal 102 collects the sampling data obtained by the power battery and uploads it to the server 104.
  • the server 104 analyzes and obtains the target characteristics of each fault type according to the characteristics corresponding to each fault type and the sampling data corresponding to each characteristic; according to the target characteristics of each fault type, the control strategy of the power battery is determined to be used for updating the control strategy of the power battery.
  • the terminal 102 can be, but not limited to, various personal computers, laptops, smart phones, tablet computers, Internet of Things devices and portable wearable devices.
  • the Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc.
  • Portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc.
  • the server 104 can be implemented with an independent server or a server cluster composed of multiple servers. It should be noted that the power battery involved in the embodiment of the present application can be, but not limited to, applied to electrical equipment such as vehicles, ships or aircraft.
  • a method for determining a power battery control strategy is provided, wherein the control strategy may be a charging control strategy, a thermal management strategy, a charge and discharge balancing control strategy, etc. As shown in FIG2 , the method includes:
  • Step S100 acquiring sampling data collected from the power battery, and determining features corresponding to different fault types according to the sampling data.
  • the server can communicate with the device using the power battery to obtain the sampled data.
  • the device can be a vehicle, a ship or an aircraft, etc.
  • the server obtains the sampled data collected from the faulty vehicle and the normal vehicle, and determines the characteristics corresponding to the different fault types according to the sampled data.
  • the number of fault types is not unique, and may include high-charge and low-discharge faults, abnormal insulation faults, etc. It can be understood that the characteristics corresponding to different fault types will also be different, and the characteristics may be current, voltage, etc.
  • the sampled data may include current data, voltage data, etc.
  • the sampled data includes at least one of current data, temperature data, voltage data and insulation data. It may be actually necessary to select one or more of the current data, temperature data, voltage data and insulation data to analyze the characteristics of each fault type.
  • the server After acquiring the sampled data, the server performs data analysis in combination with the pre-saved mechanism analysis results and/or FTA (Fault Tree Analysis) analysis results to determine the features corresponding to different fault types.
  • the mechanism analysis results can characterize the cause of the vehicle failure.
  • a forward analysis is performed based on different types of faults to determine the cause of the fault.
  • the FTA analysis results are reversely deduced to analyze what kind of fault will be caused when a problem occurs in the vehicle.
  • the features corresponding to different fault types are determined based on the sampled data, including: performing a difference analysis on the sampled data collected from the power battery, and determining the features corresponding to different fault types based on the difference analysis results. Based on the difference analysis of the sampled data collected from the power battery, feature classification can be accurately performed to determine the features corresponding to different fault types.
  • the difference analysis may include at least one of statistical feature analysis, cross-feature analysis between variables, and time series trend change feature analysis.
  • One or more of statistical feature analysis, cross-feature analysis between variables, and time series trend change feature analysis can be selected for difference analysis in combination with actual needs.
  • statistical feature analysis can be to analyze whether there are significant differences in the statistical features (such as maximum value, mean, quantile, etc.) of each sampling data between the faulty vehicle and the normal vehicle;
  • cross-feature analysis between variables can be to analyze the cross-features between each variable, such as the difference in voltage when the current is greater than the current threshold;
  • time series trend change feature analysis can be to analyze voltage consistency, self-discharge differences, etc.
  • Step S200 Analyze and obtain the target features of each fault type based on the features corresponding to each fault type and the sampling data corresponding to each feature.
  • the target feature refers to the important feature with the highest correlation with the fault type, and the target feature can be one or more.
  • the server analyzes the actual sampling data of each feature to determine the target features of different fault types, which will be used as a reference for revising policies for different fault types.
  • Step S300 Determine the control strategy of the power battery according to the target characteristics of each fault type. After determining the target characteristics of each fault type, the server weakens or enhances the corresponding characteristics of the current control strategy of the vehicle according to the target characteristics, and obtains a revised strategy as the new control strategy of the power battery. The revised strategy is imported into the market vehicles to reduce the market failure rate. The obtained revised strategy can be used to directly update the control strategy of the power battery, or to perform a strategy evaluation on the revised strategy, and the revised strategy that passes the strategy evaluation is used to update the control strategy of the power battery.
  • the method of updating the control strategy of the power battery according to the revised strategy is not unique.
  • the server can determine the vehicle to which the revised strategy is applicable based on the optimization target of the revised strategy and the use area or service life of the vehicle, and use it as the target vehicle associated with the revised strategy. All revised strategies associated with the target vehicle are sent to the target vehicle to update the control strategy of the power battery. For example, if the optimization target of the revised strategy is charging current, it is considered that the revised strategy is applicable to all vehicles in the project, and the revised strategy is imported into all vehicles in the project; if the optimization target of the revised strategy is thermal management on, the revised strategy is only applicable to vehicles used in tropical areas.
  • the server can also screen the revised strategy associated with the target vehicle and send it to the target vehicle to update the control strategy of the power battery.
  • the above power battery control strategy determination method obtains sampling data collected from the power battery and determines the features corresponding to different fault types according to the sampling data. According to the features corresponding to each fault type and the sampling data corresponding to each feature, the target features of each fault type are analyzed and obtained, and the power battery control strategy is generated according to the target features of each fault type, so as to update the power battery control strategy, thereby making the power battery control strategy more in line with the actual working conditions, making the control strategy more accurate, reducing the probability of fault occurrence, and extending the life of the power battery.
  • step S200 includes step S210 and step S220 .
  • Step S210 According to the sampling data of the features corresponding to each fault type, the features corresponding to each fault type are screened to obtain the initial target features associated with each fault type. Specifically, after determining the features corresponding to each fault type, a machine learning algorithm (such as a random forest algorithm, an XGBoost algorithm, etc.) can be used in combination with the corresponding sampling data to screen the features, and obtain the core features associated with the fault type as the initial target features, so as to reduce the features with high linear correlation and avoid data overlap. It can be understood that for different fault types, the core features screened will also be different.
  • a machine learning algorithm such as a random forest algorithm, an XGBoost algorithm, etc.
  • the core features may include the changes in the ranking of the battery cell voltage during charging, the current and voltage fitting K value during discharge (characterizing the difference in the internal parameters of the battery cell), the voltage difference during charging, etc.
  • Step S220 Determine the target feature of each fault type according to the sampling data of the initial target feature associated with each fault type and the fault prediction model corresponding to each fault type.
  • the server determines the training sample data in combination with the initial target features to perform model training, and obtains the fault prediction model corresponding to each fault type.
  • the type of fault prediction model is not unique.
  • an integrated learning model can be selected but not limited to.
  • a specific model can be selected according to the amount of data.
  • the target features of the corresponding fault type can be determined in combination with the parameters of the fault prediction model.
  • the fault prediction model corresponding to each fault type can also be obtained by pre-training based on historical data.
  • the server combines the sampling data of the initial target features associated with each fault type and directly calls the corresponding fault prediction model to determine the target features of the corresponding fault type.
  • initial target features associated with each fault type are extracted through feature screening to avoid overlapping features with high correlation affecting processing efficiency, and then the target features of each fault type are accurately obtained in combination with the fault prediction model.
  • step S210 includes: preprocessing the sampling data of the features corresponding to each fault type to obtain the preprocessed sampling data of each feature; screening the features corresponding to each fault type according to the preprocessed sampling data of each feature to obtain the initial target features associated with each fault type.
  • the method of preprocessing the sampled data is not unique.
  • the preprocessing includes at least one of data cleaning and data completion.
  • One or more methods of data cleaning and data completion can be selected for data preprocessing according to actual needs. Taking data cleaning and data completion as an example, data cleaning is to remove abnormal values in the sampled data that exceed the set threshold. For example, if the sampled current data exceeds 2000A, it can be considered as an abnormal value.
  • Data completion is to perform analogy analysis in combination with other existing battery cell data, and to complete the invalid collected data through existing parameters.
  • a fault feature database is constructed according to the sampled data of each feature after preprocessing, and the features corresponding to each fault type are extracted from the fault feature database for screening to obtain the initial target features associated with each fault type.
  • the sampled data is first preprocessed to ensure the accuracy and completeness of the data, and then feature screening is performed to determine the initial target features associated with each fault type to improve the accuracy of feature association.
  • step S220 includes step S221 , step S222 , and step S224 .
  • Step S221 adding labels to the sampled data of the initial target features associated with each fault type to generate training sample data for each fault type.
  • the label includes a faulty device label and a normal device label.
  • the specific types of the faulty device label and the normal device label are not unique either, and can be numbers, symbols or codes.
  • the faulty device label is 0, and the normal device label is 1.
  • the sampled data of the initial target features are labeled in combination with the data obtained from the faulty vehicle and the normal vehicle, and training sample data of each fault type is generated.
  • 10 initial target features related to the fault type are extracted to form a data set, and the data set contains the sampled data of the 10 initial target features collected from the faulty vehicle and the normal vehicle.
  • the label 1 is added to the sampled data of the normal vehicle, and the label 0 is added to the sampled data of the faulty vehicle, and 11 training sample data of the fault type are obtained.
  • Step S222 Perform model training based on the training sample data of each fault type to obtain the fault prediction model corresponding to each fault type. Specifically, after obtaining the training sample data of each fault type, perform model training based on the training sample data of each fault type, so as to establish corresponding fault prediction models for different fault types.
  • the fault prediction model can be used to score the fault probability of vehicles on the market (such as the fault probability is 0.5, 0.8, etc.) and evaluate the vehicle fault probability.
  • Step S224 Determine the target feature of each fault type according to the fault prediction model corresponding to each fault type. After determining the fault prediction model corresponding to each fault type, directly determine the target feature of the corresponding fault type in combination with the parameters of the fault prediction model.
  • labels are added to the sampled data of the initial target features associated with each fault type as training sample data for model training to ensure the accuracy of the fault prediction model.
  • the fault prediction model is an integrated learning model
  • step S224 includes: determining the target feature of each fault type according to the feature importance of the integrated learning model corresponding to each fault type. According to the feature importance (Shapley value, etc.) given by the integrated learning model, one or more important features (such as current, voltage, etc.) related to the corresponding fault type can be obtained as the target feature of the corresponding fault type. Determining the target feature of each fault type directly through the feature importance of the integrated learning model is simple, fast and highly reliable.
  • step S220 also includes step S223: performing model evaluation on the fault prediction model, and using the fault prediction model that passes the evaluation as the final fault prediction model.
  • the method of model evaluation is not unique, and can be set according to actual needs.
  • model evaluation may include one or more of business evaluation and rationality evaluation.
  • the fault prediction model that passes the evaluation is used as the final fault prediction model, and is used as the target feature for subsequently determining the relevant fault type.
  • the target feature of the corresponding fault type is determined based on the final fault prediction model.
  • the trained fault prediction model is model evaluated, and the fault prediction model that passes the evaluation is used as the final fault prediction model, which is more in line with actual needs. It can be understood that in other embodiments, the fault prediction model may not be model evaluated, and then all the fault prediction models obtained by training are used to determine the target features of the corresponding fault type.
  • step S223 includes: performing a business evaluation on the fault prediction model according to a preset business standard; performing a rationality evaluation on the fault prediction model that passes the business evaluation according to a preset fault tree analysis result, and using the fault prediction model that passes the rationality evaluation as the final fault prediction model.
  • business evaluation standards can be set in advance to evaluate whether the fault prediction model meets business needs, such as whether it meets cost requirements.
  • the fault prediction model using an integrated learning model as an example, after the business evaluation of the integrated learning model is passed, the rationality of the target feature is confirmed by combining the feature importance given by the integrated learning model with the FTA analysis results. For example, if the FTA analysis results show that overcharging is more important than floating charging, the integrated learning model should also have a similar conclusion. After both the business evaluation and the rationality evaluation are passed, the fault prediction model is saved as the final fault prediction model.
  • the fault prediction model is subjected to business evaluation and rationality evaluation at the same time, and a model that passes both evaluations is used as the final fault prediction model, thereby improving the accuracy of model evaluation.
  • step S300 includes steps S310 to S330 .
  • Step S310 Generate a preliminary revised strategy for the power battery based on the target characteristics of each fault type.
  • the correspondence between the type of target characteristics and the adjustment method can be saved in advance. For example, if the target characteristic is current, it can be considered that the larger the current, the less safe it is, and the adjustment method is to reduce the current.
  • the strategy is revised on the basis of the original strategy of the vehicle in combination with the target characteristics, and a preliminary revised strategy for the power battery is generated by weakening or enhancing the corresponding characteristics (such as charging rate, charging switching point, thermal management start-up conditions, balancing management conditions, etc.).
  • the revision of the charging control strategy as an example, the target characteristic corresponding to the charging fault is the charging rate, then the parameters in the strategy are adjusted to reduce the charging rate to obtain a preliminary revised strategy.
  • Step S320 Simulate device operation according to the preliminary revised strategy to obtain simulated sampling data. Specifically, simulate vehicle operation according to the revised preliminary revised strategy, and obtain simulated sampling data related to the corresponding fault type to obtain a changed data set, so that the changed data set can simulate the changed preliminary revised strategy.
  • Step S330 Perform strategy evaluation based on the simulated sampling data and the fault prediction model to obtain the control strategy of the power battery. Specifically, the data in the change data set related to the fault type is imported into the corresponding fault prediction model, and the prediction results of the model are evaluated, for example, the cost of evaluating the change of the strategy, how much the failure rate in the market can be reduced, etc., and the impact of the preliminary revised strategy on the failure is analyzed to be reasonable, and the strategy that passes the evaluation is obtained as the control strategy of the power battery.
  • the operation of the device is simulated to perform strategy evaluation to ensure that the control strategy of the power battery meets actual needs.
  • the strategy evaluation includes a failure rate evaluation and/or a business evaluation.
  • the revised strategy may be selected to perform a failure rate evaluation and/or a business evaluation based on actual needs. Specifically, after the data in the change data set related to the fault type is imported into the corresponding fault prediction model, the failure rate is determined to be reduced based on the prediction results of the fault prediction model.
  • the preliminary revised strategy can be considered feasible and the failure rate evaluation has passed. Furthermore, a business needs analysis is performed on the preliminary revised strategy that has passed the failure rate evaluation in combination with preset business standards. If the preliminary revised strategy can reduce the failure rate and the cost is within a controllable range, the strategy revision is considered reasonable and can be used as a control strategy for the power battery.
  • the above-mentioned power battery control strategy determination method provided in this application revises the power battery control strategy of in-service new energy vehicles based on big data.
  • the control strategy is made more accurate, and the probability of failure is reduced without reducing the user experience, and the life of the power battery is extended.
  • it can also be used to guide the formulation of power battery control strategies in the incremental market, reduce the cost of simulation vehicle testing and shorten the cycle of vehicle development.
  • the control strategy is customized to achieve differentiated control of the vehicle.
  • the original charging strategy of electric vehicles assumes a step-by-step charging method, and the current design is shown in Figure 6.
  • the working condition characteristics are extracted and modeled according to the charging working condition. In a certain project, there are a total of 9 faulty vehicles.
  • the risk threshold of the fault occurrence is determined as score_threshold.
  • the final ranking of the faulty vehicles is shown in Figure 7.
  • the charging strategy is changed by using the power battery control strategy determination method provided by this application as shown in Figure 8.
  • the charging current before and after the change has almost no effect on the charging time (the current mean value remains almost unchanged).
  • new features are re-extracted and re-predicted.
  • the final ranking of faulty vehicles is shown in Figure 9.
  • the failure rate of most vehicles has decreased, and the number of vehicles that need to be checked in the market (exceeding the risk threshold score_threshold) has been greatly reduced, reducing the cost of troubleshooting.
  • a device control method including:
  • Step S410 Determine the target revision strategy from the revision strategies of the power battery; the revision strategy of the power battery is obtained according to the above power battery control strategy determination method. Specifically, after obtaining the revision strategy of each fault type, the server determines the target revision strategy from the revision strategy.
  • Step S420 Send the target revision strategy to the target device to update the control strategy of the power battery of the target device.
  • the server sends the target revision strategy to the target vehicle to update the control strategy of the power battery of the target vehicle.
  • step S410 includes: determining the target device associated with each revision policy according to the optimization target of the revision policy; and taking all revision policies associated with the target device as the target revision policy. According to the optimization target of the revision policy, taking all revision policies associated with the target device as the target revision policy to realize active update of the revision policy of the target device.
  • step S410 includes: according to the update request sent by the target device, obtaining the revision policy associated with the optimization target and the target device from the revision policy of the power battery as the target revision policy.
  • the update request can be sent through the target device to start the policy update operation according to actual needs.
  • the above-mentioned equipment control method generates a revised strategy for the power battery according to the target characteristics of each fault type, so as to update the control strategy of the power battery, thereby making the control strategy of the power battery more in line with the actual working conditions, making the control strategy more accurate, reducing the probability of fault occurrence, and extending the life of the power battery.
  • a device control method including:
  • Step S510 obtaining a revised strategy of the power battery to be updated; the revised strategy of the power battery to be updated is obtained according to the above-mentioned power battery control strategy determination method.
  • Step S520 according to the revised strategy of the power battery to be updated, the control strategy of the power battery of the device is updated. Specifically, the target vehicle obtains the revised strategy of the power battery to be updated issued by the server, and updates the control strategy of the power battery of the vehicle.
  • step S510 includes: receiving a target revision policy sent by a server as the revision policy of the power battery to be updated; the target revision policy is obtained by the server according to all revision policies associated with the target device, and the target device is determined according to the optimization target of the revision policy. According to the optimization target of the revision policy, all revision policies associated with the target device are used as the target revision policy to realize active update of the revision policy of the target device.
  • step S510 includes: sending an update request to a server; receiving a target revision policy returned by the server in response to the update request as the revision policy of the power battery to be updated; the target revision policy is obtained by the server according to the update request, obtaining the revision policy associated with the optimization target and the target device from the revision policy of the power battery.
  • the update request can be sent through the target device to start the policy update operation according to actual needs.
  • the above-mentioned equipment control method generates a revised strategy for the power battery according to the target characteristics of each fault type, so as to update the control strategy of the power battery, thereby making the control strategy of the power battery more in line with the actual working conditions, making the control strategy more accurate, reducing the probability of fault occurrence, and extending the life of the power battery.
  • steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
  • the embodiment of the present application also provides a power battery control strategy determination device for implementing the power battery control strategy determination method involved above.
  • the implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more power battery control strategy determination device embodiments provided below can refer to the limitations of the power battery control strategy determination method above, and will not be repeated here.
  • a power battery control strategy determination device is also provided, wherein the control strategy may be a charging control strategy, a thermal management strategy, a charge and discharge balance control strategy, etc.
  • the device includes: a data acquisition module 100, a data analysis module 200 and a strategy revision module 300, wherein:
  • the data acquisition module 100 is used to acquire sampled data collected from the power battery, and determine the characteristics corresponding to different fault types according to the sampled data.
  • the data analysis module 200 is used to analyze and obtain target features of each fault type according to the features corresponding to each fault type and the sampling data corresponding to each feature.
  • the strategy revision module 300 is used to determine the control strategy of the power battery according to the target characteristics of each fault type.
  • the data acquisition module 100 performs a difference analysis on the sampled data collected from the power battery, and determines the characteristics corresponding to different fault types according to the difference analysis results.
  • the data analysis module 200 screens the features corresponding to each fault type based on the sampling data of the features corresponding to each fault type to obtain the initial target features associated with each fault type; and determines the target features of each fault type based on the sampling data of the initial target features associated with each fault type and the fault prediction model corresponding to each fault type.
  • the data analysis module 200 preprocesses the sampling data of the features corresponding to each fault type to obtain the preprocessed sampling data of each feature; based on the preprocessed sampling data of each feature, the features corresponding to each fault type are screened to obtain the initial target features associated with each fault type.
  • the data analysis module 200 adds labels to the sampling data of the initial target features associated with each fault type to generate training sample data for each fault type, where the labels include faulty equipment labels and normal equipment labels; performs model training based on the training sample data for each fault type to obtain a fault prediction model corresponding to each fault type; and determines the target features of each fault type based on the fault prediction model corresponding to each fault type.
  • the data analysis module 200 performs model evaluation on the fault prediction model, and uses the fault prediction model that passes the evaluation as the final fault prediction model.
  • the data analysis module 200 performs a business evaluation on the fault prediction model according to preset business standards; performs a rationality evaluation on the fault prediction model that passes the business evaluation according to preset fault tree analysis results, and uses the fault prediction model that passes the rationality evaluation as the final fault prediction model.
  • the data analysis module 200 determines the target feature of each fault type according to the feature importance of the ensemble learning model corresponding to each fault type.
  • the strategy revision module 300 generates a preliminary revised strategy for the power battery according to the target characteristics of each fault type; simulates equipment operation according to the preliminary revised strategy to obtain simulated sampling data; and performs strategy evaluation based on the simulated sampling data and the fault prediction model to obtain the control strategy of the power battery.
  • Each module in the above-mentioned power battery control strategy determination device can be implemented in whole or in part by software, hardware and a combination thereof.
  • Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.
  • a device control apparatus including: a policy determination module 410 and a policy sending module 420, wherein:
  • the strategy determination module 410 is used to determine a target revision strategy from the revision strategies of the power battery; the revision strategy of the power battery is obtained according to the above-mentioned power battery control strategy determination device.
  • the strategy sending module 420 is used to send the target revised strategy to the target device to update the control strategy of the power battery of the target device.
  • the policy determination module 410 determines the target device associated with each revision policy according to the optimization target of the revision policy; and uses all revision policies associated with the target device as the target revision policy.
  • the policy determination module 410 obtains the revision policy associated with the optimization target and the target device from the revision policy of the power battery according to the update request sent by the target device, as the target revision policy.
  • a device control apparatus including: a policy acquisition module 510 and a policy update module 520, wherein:
  • the strategy acquisition module 510 is used to acquire the revised strategy of the power battery to be updated; the revised strategy of the power battery to be updated is obtained by the power battery control strategy determination device mentioned above.
  • the strategy updating module 520 is used to update the control strategy of the power battery of the device according to the revised strategy of the power battery to be updated.
  • the policy acquisition module 510 receives the target revision policy sent by the server as the revision policy of the power battery to be updated; the target revision policy is obtained by the server based on all revision policies associated with the target device, and the target device is determined based on the optimization target of the revision policy.
  • the policy acquisition module 510 sends an update request to the server; receives the target revision policy returned by the server in response to the update request as the revision policy of the power battery to be updated; the target revision policy is obtained by the server according to the update request, obtaining the revision policy associated with the optimization target and the target device from the revision policy of the power battery.
  • a computer device which may be a server, and its internal structure diagram may be shown in FIG15.
  • the computer device includes a processor, a memory, an input/output interface (Input/Output, referred to as I/O) and a communication interface.
  • the processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface.
  • the processor of the computer device is used to provide computing and control capabilities.
  • the memory of the computer device includes a non-volatile storage medium and an internal memory.
  • the non-volatile storage medium stores an operating system, a computer program and a database.
  • the internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium.
  • the database of the computer device is used to store data.
  • the input/output interface of the computer device is used to exchange information between the processor and an external device.
  • the communication interface of the computer device is used to communicate with an external terminal through a network connection.
  • FIG. 15 is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied.
  • the specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.
  • a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.
  • a computer-readable storage medium on which a computer program is stored.
  • the computer program is executed by a processor, the steps of the above method are implemented.
  • a computer program product comprising a computer program, which implements the steps of the above method when executed by a processor.
  • the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

Landscapes

  • Charge And Discharge Circuits For Batteries Or The Like (AREA)

Abstract

一种动力电池控制策略确定方法、设备控制方法、装置、计算机设备、计算机可读存储介质和计算机程序产品，动力电池控制策略确定方法包括：获取对动力电池进行采集得到的采样数据，根据采样数据确定不同故障类型对应的特征；根据各故障类型对应的特征，以及各特征对应的采样数据，分析获得各故障类型的目标特征；根据各故障类型的目标特征，确定动力电池的控制策略。使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。

Description

动力电池控制策略确定方法、设备控制方法和装置 技术领域
本发明涉及电池技术领域，特别是涉及一种动力电池控制策略确定方法、设备控制方法、装置、计算机设备、计算机可读存储介质和计算机程序产品。
背景技术
动力电池指为工具提供动力来源的电源。随着科技的发展和社会的不断进步，动力电池的应用领域不断扩展，不仅被应用于电动自行车、电动摩托车、电动汽车等电动交通工具，还被应用于军事装备和航空航天等多个领域。传统的动力电池控制策略，是结合动力电池所应用的设备进行实验室仿真和测试而制定，很难模拟真实工况。如何提高对动力电池的控制精准度，是一个亟待解决的问题。
发明内容
根据本申请的各种实施例，提供一种动力电池控制策略确定方法、设备控制方法、装置、计算机设备、计算机可读存储介质和计算机程序产品。
第一方面，本申请提供了一种动力电池控制策略确定方法，包括：
获取对动力电池进行采集得到的采样数据，根据所述采样数据确定不同故障类型对应的特征；
根据各故障类型对应的特征，以及各所述特征对应的采样数据，分析获得各故障类型的目标特征；
根据各故障类型的目标特征，确定动力电池的控制策略。
上述动力电池控制策略确定方法，通过获取对动力电池进行采集得到的采样数据，根据采样数据确定不同故障类型对应的特征。根据各故障类型对应的特征以及各特征对应的采样数据，分析得到各故障类型的目标特征，根据各故障类型的目标特征生成动力电池的控制策略，以供对动力电池的控制策略进行更新，从而使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。
在其中一个实施例中，所述根据所述采样数据确定不同故障类型对应的特征，包括：对动力电池采集得到的采样数据进行差异性分析，根据差异性分析结果确定不同故障类型对应的特征。根据对动力电池采集得到的采样数据进行差异性分析，可准确进行特征分类，确定不同故障类型所对应的特征。
在其中一个实施例中，所述差异性分析包括统计特征分析、变量间的交叉特征分析以及时序趋势变化特征分析中的至少一种。可结合实际需求选择统计特征分析、变量间的交叉特征分析以及时序趋势变化特征分析中的一种或多种进行差异性分析。
在其中一个实施例中，所述根据各故障类型对应的特征，以及各所述特征对应的采样数据，分析获得各故障类型的目标特征，包括：根据各故障类型对应的特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征；根据各故障类型关联的初始目标特征的采样数据，以及各故障类型对应的故障预测模型，确定各故障类型的目标特征。通过特征筛选提取与各故障类型关联的初始目标特征，避免相关性高的重叠特征影响处理效率，再结合故障预测模型准确获取各故障类型的目标特征。
在其中一个实施例中，所述根据各故障类型对应的特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征，包括：对各故障类型对应的特征的采样数据进行预处理，获得预处理后的各特征的采样数据；根据预处理后的各特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征。先对采样数据进行预处理，确保数据的准确和完整，然后进行特征筛选确定各故障类型关联的初始目标特征，提高特征关联准确性。
在其中一个实施例中，所述预处理包括数据清洗和数据补全中的至少一种。可根据实际需要选择数据清洗和数据补全中的一种或多种方式进行数据预处理。
在其中一个实施例中，所述根据各故障类型关联的初始目标特征的采样数据，以及各故障类型对应的故障预测模型，确定各故障类型的目标特征，包括：对各故障类型关联的初始目标特征的采样数据添加标签，生成各故障类型的训练样本数据，所述标签包括故障设备标签和正常设备标签；根据各故障类型的训练样本数据进行模型训练，获得各故障类型对应的故障预测模型；根据各故障类型对应的故障预测模型，确定各故障类型的目标特征。对各故障类型关联的初始目标特征的采样数据添加标签作为训练样本数据进行模型训练，确保故障预测模型的准确性。
在其中一个实施例中，所述根据各故障类型的训练样本数据进行模型训练，获得各故障类型对应的故障预测模型之后，还包括：对故障预测模型进行模型评估，将评估通过的故障预测模型作为最终获得的故障预测模型。对训练好的故障预测模型进行模型评估，评估通过的作为最终获得的故障预测模型，更加符合实际需求。
在其中一个实施例中，所述对故障预测模型进行模型评估，将评估通过的故障预测模型作为最终获得的故障预测模型，包括：根据预设的业务标准对故障预测模型进行业务评估；根据预设的故障树分析结果，对业务评估通过的故障预测模型进行合理性评估，将合理性评估通过的故障预测模型，作为最终获得的故障预测模型。同时对故障预测模型进行 业务评估和合理性评估，均评估通过则作为最终获得的故障预测模型，提高模型评估准确性。
在其中一个实施例中，所述故障预测模型为集成学习模型；所述根据各故障类型对应的故障预测模型，确定各故障类型的目标特征，包括：根据各故障类型对应的集成学习模型的特征重要性，确定各故障类型的目标特征。直接通过集成学习模型的特征重要性，确定各故障类型的目标特征，简便快捷且可靠性高。
在其中一个实施例中，根据各故障类型的目标特征，确定动力电池的控制策略，包括：根据各故障类型的目标特征生成动力电池的初步修订策略；根据所述初步修订策略模拟设备运行，获得模拟采样数据；根据所述模拟采样数据和所述故障预测模型进行策略评估，得到动力电池的控制策略。根据故障类型的目标特征生成动力电池的初步修订策略后，模拟设备运行进行策略评估，确保动力电池的控制策略符合实际需求。
在其中一个实施例中，所述策略评估包括故障率评估和/或业务评估。可根据实际需要选择对修订策略进行故障率评估和/或业务评估。
在其中一个实施例中，所述采样数据包括电流数据、温度数据、电压数据和绝缘数据中的至少一种。可实际需要选择采集电流数据、温度数据、电压数据和绝缘数据中的一种或多种对各故障类型的特征进行分析。
第二方面，本申请提供了一种设备控制方法，包括：从动力电池的修订策略中确定目标修订策略；所述动力电池的修订策略为根据上述动力电池控制策略修订方法得到；将所述目标修订策略向目标设备发送，以对所述目标设备的动力电池的控制策略进行更新。
上述设备控制方法，根据各故障类型的目标特征生成动力电池的修订策略，以供对动力电池的控制策略进行更新，从而使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。
在其中一个实施例中，所述从动力电池的修订策略中确定目标修订策略，包括：根据修订策略的优化目标，确定各所述修订策略关联的目标设备；将与所述目标设备关联的所有修订策略作为目标修订策略。根据修订策略的优化目标，将与目标设备关联的所有修订策略作为目标修订策略，实现对目标设备的修订策略主动更新。
在其中一个实施例中，所述从动力电池的修订策略中确定目标修订策略，包括：根据目标设备发送的更新请求，从动力电池的修订策略中获取优化目标与所述目标设备关联的修订策略，作为目标修订策略。可根据实际需要通过目标设备发送更新请求，启动策略更新操作。
第三方面，本申请提供了一种设备控制方法，包括：获取待更新的动力电池的修订策略；所述待更新的动力电池的修订策略为根据上述动力电池控制策略修订方法得到；根 据所述待更新的动力电池的修订策略，对设备的动力电池的控制策略更新。
上述设备控制方法，根据各故障类型的目标特征生成动力电池的修订策略，以供对动力电池的控制策略进行更新，从而使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。
在其中一个实施例中，所述获取待更新的动力电池的修订策略，包括：接收服务器发送的目标修订策略，作为待更新的动力电池的修订策略；所述目标修订策略为所述服务器根据与目标设备关联的所有修订策略得到，所述目标设备根据修订策略的优化目标确定。根据修订策略的优化目标，将与目标设备关联的所有修订策略作为目标修订策略，实现对目标设备的修订策略主动更新。
在其中一个实施例中，所述获取待更新的动力电池的修订策略，包括：发送更新请求至服务器；接收所述服务器响应所述更新请求返回的目标修订策略，作为待更新的动力电池的修订策略；所述目标修订策略为所述服务器根据所述更新请求，从动力电池的修订策略中获取优化目标与目标设备关联的修订策略得到。可根据实际需要通过目标设备发送更新请求，启动策略更新操作。
第四方面，本申请提供了一种动力电池控制策略确定装置，包括：
数据获取模块，用于获取对动力电池进行采集得到的采样数据，根据所述采样数据确定不同故障类型对应的特征；
数据分析模块，用于根据各故障类型对应的特征，以及各所述特征对应的采样数据，分析获得各故障类型的目标特征；
策略修订模块，用于根据各故障类型的目标特征，确定动力电池的控制策略。
第五方面，本申请提供了一种设备控制装置，包括：
策略确定模块，用于从动力电池的修订策略中确定目标修订策略；所述动力电池的修订策略为根据上述的动力电池控制策略确定装置得到；
策略发送模块，用于将所述目标修订策略向目标设备发送，以对所述目标设备的动力电池的控制策略进行更新。
第六方面，本申请提供了一种设备控制装置，包括：
策略获取模块，用于获取待更新的动力电池的修订策略；所述待更新的动力电池的修订策略为上述的动力电池控制策略确定装置得到；
策略更新模块，用于根据所述待更新的动力电池的修订策略，对设备的动力电池的控制策略更新。
第七方面，本申请提供了一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时实现上述方法的步骤。
第八方面，本申请提供了一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上述方法的步骤。
第九方面，本申请提供了一种计算机程序产品，包括计算机程序，该计算机程序被处理器执行时实现上述方法的步骤。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本发明的其它特征、目的和优点将从说明书、附图以及权利要求书变得明显。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1为一个实施例中动力电池控制策略确定方法的应用环境示意图；
图2为一个实施例中动力电池控制策略确定方法的流程图；
图3为一个实施例中根据各故障类型对应的特征，以及各特征对应的采样数据，分析获得各故障类型的目标特征的流程图；
图4为一个实施例中根据各故障类型关联的初始目标特征的采样数据进行模型训练，获得各故障类型对应的故障预测模型的流程图；
图5为一个实施例中根据各故障类型的目标特征，生成动力电池的修订策略的流程图；
图6为一个实施例中原有策略下阶梯充电控制示意图；
图7为一个实施例中原有策略下故障车辆排名示意图；
图8为一个实施例中新策略下阶梯充电控制示意图；
图9为一个实施例中原有策略和新策略下故障车辆排名对比示意图；
图10为一个实施例中设备控制方法的流程图；
图11为另一个实施例中设备控制方法的流程图；
图12为一个实施例中动力电池控制策略确定装置的结构框图；
图13为一个实施例中设备控制装置的结构框图；
图14为另一个实施例中设备控制装置的结构框图；
图15为一个实施例中计算机设备的内部结构图。
具体实施方式
下面将结合附图对本申请技术方案的实施例进行详细的描述。以下实施例仅用于更加清楚地说明本申请的技术方案，因此只作为示例，而不能以此来限制本申请的保护范围。
除非另有定义，本文所使用的所有的技术和科学术语与属于本申请的技术领域的技术人员通常理解的含义相同；本文中所使用的术语只是为了描述具体的实施例的目的，不是旨在于限制本申请；本申请的说明书和权利要求书及上述附图说明中的术语“包括”和“具有”以及它们的任何变形，意图在于覆盖不排他的包含。
在本申请实施例的描述中，技术术语“第一”“第二”等仅用于区别不同对象，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量、特定顺序或主次关系。在本申请实施例的描述中，“多个”的含义是两个以上，除非另有明确具体的限定。
在本文中提及“实施例”意味着，结合实施例描述的特定特征、结构或特性可以包含在本申请的至少一个实施例中。在说明书中的各个位置出现该短语并不一定均是指相同的实施例，也不是与其它实施例互斥的独立的或备选的实施例。本领域技术人员显式地和隐式地理解的是，本文所描述的实施例可以与其它实施例相结合。
在本申请实施例的描述中，术语“和/或”仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中字符“/”，一般表示前后关联对象是一种“或”的关系。
在本申请实施例的描述中，术语“多个”指的是两个以上(包括两个)，同理，“多组”指的是两组以上(包括两组)，“多片”指的是两片以上(包括两片)。
在本申请实施例的描述中，技术术语“中心”“纵向”“横向”“长度”“宽度”“厚度”“上”“下”“前”“后”“左”“右”“竖直”“水平”“顶”“底”“内”“外”“顺时针”“逆时针”“轴向”“径向”“周向”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本申请实施例和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本申请实施例的限制。
在本申请实施例的描述中，除非另有明确的规定和限定，技术术语“安装”“相连”“连接”“固定”等术语应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或成一体；也可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通或两个元件的相互作用关系。对于本领域的普通技术人员而言，可以根据具体情况理解上述术语在本申请实施例中的具体含义。
动力电池即为工具提供动力来源的电源，多采用阀口密封式铅酸蓄电池、敞口式管式铅酸蓄电池以及磷酸铁锂蓄电池，具有高能量、高功率和高能量密度等特点。传统的动力电池控制策略，是结合动力电池所应用的设备进行实验室仿真和测试而制定，很难模拟真实工况。基于此，本申请提供一种动力电池控制策略确定方法,通过获取对动力电池进行 采集得到的采样数据，根据采样数据确定不同故障类型对应的特征。根据各故障类型对应的特征以及各特征对应的采样数据，分析得到各故障类型的目标特征，根据各故障类型的目标特征确定动力电池的控制策略，以供对动力电池的控制策略进行更新，从而使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。
本申请实施例提供的动力电池控制策略确定方法，可以应用于如图1所示的应用环境中。终端102通过网络与服务器104进行通信。数据存储系统可以存储服务器104需要处理的数据。数据存储系统可以集成在服务器104上，也可以放在云上或其他网络服务器上。通过终端102对动力电池进行采集得到采样数据上传至服务器104，服务器104根据各故障类型对应的特征，以及各特征对应的采样数据，分析获得各故障类型的目标特征；根据各故障类型的目标特征，确定动力电池的控制策略，以用作对动力电池的控制策略进行更新。其中，终端102可以但不限于是各种个人计算机、笔记本电脑、智能手机、平板电脑、物联网设备和便携式可穿戴设备，物联网设备可为智能音箱、智能电视、智能空调、智能车载设备等。便携式可穿戴设备可为智能手表、智能手环、头戴设备等。服务器104可以用独立的服务器或者是多个服务器组成的服务器集群来实现。需要说明的是，本申请实施例中涉及的动力电池，可以但不限应用于车辆、船舶或飞行器等用电设备中。
在一个实施例中，提供了一种动力电池控制策略确定方法，其中，控制策略可以是充电控制策略、热管理策略、充放电均衡控制策略等，如图2所示，该方法包括：
步骤S100：获取对动力电池进行采集得到的采样数据，根据采样数据确定不同故障类型对应的特征。
其中，可通过服务器与使用动力电池的设备进行通信，获取得到采样数据。设备可以是车辆、船舶或飞行器等，为便于理解，下文均以动力电池应用于车辆为例进行解释说明。服务器获取对故障车辆和正常车辆采集得到的采样数据，根据采样数据确定不同故障类型对应的特征。故障类型的数量并不唯一，可包括充高放低故障、绝缘异常故障等。可以理解，不同故障类型所对应的特征也会有所不同，特征可以是电流、电压等，对应地，采样数据可包括电流数据、电压数据等。在一个实施例中，采样数据包括电流数据、温度数据、电压数据和绝缘数据中的至少一种。可实际需要选择采集电流数据、温度数据、电压数据和绝缘数据中的一种或多种对各故障类型的特征进行分析。
服务器在获取到采样数据之后，结合预先保存的机理分析结果和/或FTA(Fault Tree Analysis，故障树分析)分析结果进行数据分析，确定不同故障类型对应的特征。其中，机理分析结果可表征车辆出现故障的原因，通过结合已有的市场车辆和车辆生产过程中的数据，根据不同类型的故障进行正向分析，确定故障存在的原因。对应地，FTA分析结果 则是通过反向推导，在车辆出现问题时分析会导致何种故障。在一个实施例中，根据采样数据确定不同故障类型对应的特征，包括：对动力电池采集得到的采样数据进行差异性分析，根据差异性分析结果确定不同故障类型对应的特征。根据对动力电池采集得到的采样数据进行差异性分析，可准确进行特征分类，确定不同故障类型所对应的特征。
其中，差异性分析的具体方式也并不唯一，差异性分析可包括统计特征分析、变量间的交叉特征分析以及时序趋势变化特征分析中的至少一种。可结合实际需求选择统计特征分析、变量间的交叉特征分析以及时序趋势变化特征分析中的一种或多种进行差异性分析。具体地，统计特征分析，可以是分析故障车辆与正常车辆在各采样数据的统计特征(如最值、均值、分位数等)中是否存在显著差异；变量间的交叉特征分析，可以是分析各变量间的交叉特征，如电流大于电流阈值时电压的差异等；时序趋势变化特征分析，可以是分析电压一致性、自放电差异等。根据分析出的显著特征确定和故障类型的关联性，对应不同的故障类型，确定故障车辆与正常车辆存在哪些数据差异，从而确定不同故障类型对应的特征。
步骤S200：根据各故障类型对应的特征，以及各特征对应的采样数据，分析获得各故障类型的目标特征。其中，目标特征即指与故障类型关联性最高的重要特征，目标特征可以是一个也可以是多个。服务器在确定各故障类型对应的特征之后，根据各特征的实际采样数据进行分析，确定不同故障类型的目标特征，后续作为针对不同故障类型进行策略修订的参考依据。
步骤S300：根据各故障类型的目标特征，确定动力电池的控制策略。其中，在确定各故障类型的目标特征之后，服务器根据目标特征对车辆当前的控制策略进行减弱或增强相应特征处理，得到修订策略作为动力电池新的控制策略。通过导入修订策略到市场车辆，以降低市场故障率。得到的修订策略可以是直接用来对动力电池的控制策略进行更新，还可以是对修订策略进行策略评估，将通过策略评估的修订策略对动力电池的控制策略进行更新。
进一步地，根据修订策略对动力电池的控制策略进行更新的方式也不是唯一的，服务器可以是根据修订策略的优化目标，结合车辆的使用地区或使用年限确定修订策略适用的车辆，作为修订策略关联的目标车辆，将与目标车辆关联的所有修订策略下发给目标车辆对动力电池的控制策略进行更新。例如，若修订策略的优化目标为充电电流，则认为该修订策略适用于项目内所有车辆，对项目内所有车辆全部导入该修订策略；若修订策略的优化目标为热管理开启，则该修订策略仅适用于热带地区使用的车辆。此外，服务器也可以是在接收到目标车辆发送的更新请求后，筛选与目标车辆关联的修订策略下发给目标车辆对动力电池的控制策略进行更新。
上述动力电池控制策略确定方法，通过获取对动力电池进行采集得到的采样数据，根据采样数据确定不同故障类型对应的特征。根据各故障类型对应的特征以及各特征对应的采样数据，分析得到各故障类型的目标特征，根据各故障类型的目标特征生成动力电池的控制策略，以供对动力电池的控制策略进行更新，从而使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。
在一个实施例中，如图3所示，步骤S200包括步骤S210和步骤S220。
步骤S210：根据各故障类型对应的特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征。具体地，在确定各故障类型对应的特征之后，可利用机器学习算法(如随机森林算法、XGBoost算法等)结合相应的采样数据对特征进行筛选，得到与故障类型关联的核心特征作为初始目标特征，以减少线性相关性高的特征，避免数据重叠。可以理解，针对不同故障类型，筛选得到的核心特征也会对应有所不同，故障类型以充高放低故障为例，则核心特征可包括充电过程中电芯电压排名的变化、放电过程中电流电压拟合K值(表征电芯内部参数的差异)、充电过程中的压差等。
步骤S220：根据各故障类型关联的初始目标特征的采样数据，以及各故障类型对应的故障预测模型，确定各故障类型的目标特征。
其中，可以是在确定各故障类型关联的初始目标特征之后，服务器结合初始目标特征确定训练样本数据进行模型训练，获得各故障类型对应的故障预测模型。可以理解，故障预测模型的类型也不是唯一的，例如可选择但不限于集成学习模型，实际可根据数据量的大小选择具体的模型。在确定各故障类型对应的故障预测模型后，结合故障预测模型的参数便可确定对应故障类型的目标特征。可以理解，在其他实施例中，也可以是预先根据历史数据训练得到各故障类型对应的故障预测模型，服务器结合各故障类型关联的初始目标特征的采样数据，直接调用相应的故障预测模型，从而确定对应故障类型的目标特征。
本实施例中，通过特征筛选提取与各故障类型关联的初始目标特征，避免相关性高的重叠特征影响处理效率，再结合故障预测模型准确获取各故障类型的目标特征。
在一个实施例中，步骤S210包括：对各故障类型对应的特征的采样数据进行预处理，获得预处理后的各特征的采样数据；根据预处理后的各特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征。
可以理解，对采样数据进行预处理的方式也不是唯一的，本实施例中，预处理包括数据清洗和数据补全中的至少一种。可根据实际需要选择数据清洗和数据补全中的一种或多种方式进行数据预处理。以同时进行数据清洗和数据补全为例，数据清洗即是对采样数据中超过设定阈值的异常值进行剔除，如采样到的电流数据超过2000A，则可认为属于异常值。数据补全即是结合其他已有的电芯数据进行模拟类推分析，通过已有参数对无效采集 的数据进行补全。通过同时进行数据清洗和数据补全之后，根据预处理后的各特征的采样数据构建故障特征数据库，从故障特征数据库中提取各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征。
本实施例中，先对采样数据进行预处理，确保数据的准确和完整，然后进行特征筛选确定各故障类型关联的初始目标特征，提高特征关联准确性。
在一个实施例中，如图4所示，步骤S220包括步骤S221、步骤S222和步骤S224。
步骤S221：对各故障类型关联的初始目标特征的采样数据添加标签，生成各故障类型的训练样本数据。
其中，标签包括故障设备标签和正常设备标签。故障设备标签和正常设备标签的具体类型也不是唯一的，可以是数字、符号或编码。本实施例中，故障设备标签为0，正常设备标签为1。具体地，在确定各故障类型关联的初始目标特征之后，结合从故障车辆和正常车辆获得的数据，对初始目标特征的采样数据添加标签，生成各故障类型的训练样本数据。以某个故障类型为例，提取得到该故障类型相关的10项初始目标特征构成数据集，数据集中包含对故障车辆和正常车辆采集到的这10项初始目标特征的采样数据，对正常车辆的采样数据增加标签1，故障车辆的采样数据增加标签0，得到该故障类型的11项训练样本数据。
步骤S222：根据各故障类型的训练样本数据进行模型训练，获得各故障类型对应的故障预测模型。具体地，在得到各故障类型的训练样本数据之后，分别根据各故障类型的训练样本数据进行模型训练，从而针对不同故障类型建立对应的故障预测模型。故障预测模型可用作对市场上车辆的故障概率进行打分(如故障概率为0.5、0.8等)，评估车辆故障概率。
步骤S224：根据各故障类型对应的故障预测模型，确定各故障类型的目标特征。在确定各故障类型对应的故障预测模型后，结合故障预测模型的参数直接确定对应故障类型的目标特征。
本实施例中，对各故障类型关联的初始目标特征的采样数据添加标签作为训练样本数据进行模型训练，确保故障预测模型的准确性。
在一个实施例中，故障预测模型为集成学习模型；步骤S224包括：根据各故障类型对应的集成学习模型的特征重要性，确定各故障类型的目标特征。根据集成学习模型给出的特征重要性(Shapley value(沙普利值)等)，可得到与对应故障类型相关的一种或多种重要特征(例如电流、电压等)，作为对应故障类型的目标特征。直接通过集成学习模型的特征重要性，确定各故障类型的目标特征，简便快捷且可靠性高。
进一步地，在一个实施例中，继续参照图4，步骤S222之后，步骤S220还包括步骤 S223：对故障预测模型进行模型评估，将评估通过的故障预测模型作为最终获得的故障预测模型。同样的，模型评估的方式也并不是唯一的，可根据实际需要进行设置，例如模型评估可包括业务评估和合理性评估中的一种或多种。将评估通过的故障预测模型作为最终获得的故障预测模型，用作后续确定相关故障类型的目标特征。对应地，步骤S224中则是根据最终获得的故障预测模型，确定相应故障类型的目标特征。本实施例中，对训练好的故障预测模型进行模型评估，评估通过的作为最终获得的故障预测模型，更加符合实际需求。可以理解，在其他实施例中，也可以不对故障预测模型进行模型评估，则后续利用训练得到的所有故障预测模型确定对应故障类型的目标特征。
在一个实施例中，步骤S223包括：根据预设的业务标准对故障预测模型进行业务评估；根据预设的故障树分析结果，对业务评估通过的故障预测模型进行合理性评估，将合理性评估通过的故障预测模型，作为最终获得的故障预测模型。
具体地，可预先设置业务评估标准，以评估故障预测模型是否满足业务需求，例如是否符合成本要求等。同样以故障预测模型采用集成学习模型为例，在对集成学习模型进行业务评估通过后，结合集成学习模型给出的特征重要性与FTA分析结果确认目标特征的合理性，如FTA分析结果分析认为超充比浮充更重要，集成学习模型也应该有相似的结论。在业务评估和合理性评估都通过后，则将故障预测模型作为最终获得的故障预测模型进行保存。
本实施例中，同时对故障预测模型进行业务评估和合理性评估，均评估通过则作为最终获得的故障预测模型，提高模型评估准确性。
在一个实施例中，如图5所示，步骤S300包括步骤S310至步骤S330。
步骤S310：根据各故障类型的目标特征生成动力电池的初步修订策略。具体地，可预先保存目标特征的类型与调整方式(减弱或增强)的对应关系，例如目标特征为电流，可认为电流越大越不安全，则调整方式为减小电流。在确定故障类型的目标特征之后，结合目标特征在车辆的原有策略基础上进行策略修订，通过减弱或增强相应特征(如充电倍率、充电切换点、热管理开启条件、均衡管理条件等)，生成动力电池的初步修订策略。以充电控制策略修订为例，充电故障所对应的目标特征为充电倍率，则对策略中的参数进行调整减小充电倍率，得到初步修订策略。
步骤S320：根据初步修订策略模拟设备运行，获得模拟采样数据。具体地，根据修订后的初步修订策略模拟车辆运行，并获取与对应故障类型相关的模拟采样数据得到变更数据集，使得变更数据集可以模拟变更后的初步修订策略。
步骤S330：根据模拟采样数据和故障预测模型进行策略评估，得到动力电池的控制策略。具体地，将故障类型相关的变更数据集中的数据，导入到对应的故障预测模型中， 对模型的预测结果进行评估，例如可以是评估策略的变更需要多少成本，能降低多少市场上的故障率等，分析该初步修订策略对故障的影响是否合理，获取评估通过的策略作为动力电池的控制策略。
本实施例中，根据故障类型的目标特征生成动力电池的初步修订策略后，模拟设备运行进行策略评估，确保动力电池的控制策略符合实际需求。
在一个实施例中，策略评估包括故障率评估和/或业务评估。可根据实际需要选择对修订策略进行故障率评估和/或业务评估。具体地，将故障类型相关的变更数据集中的数据，导入到对应的故障预测模型之后，根据故障预测模型的预测结果确定故障率降低，则可认为初步修订策略可行，故障率评估通过。进一步地，结合预设的业务标准对故障率评估通过的初步修订策略进行业务需求分析，若初步修订策略能够降低故障率且成本在可控范围，则认为策略修订合理，可作为动力电池的控制策略。
传统的动力电池控制策略，结合动力电池所应用的设备进行实验室仿真和测试而制定，这类方法存在如下问题：
1、实验室仿真和整车测试很难模拟真实工况；
2、仅对动力电池初期进行模拟，待动力电池老化后该类方法很难模拟；
3、若对老化后的动力电池做测试，该类方法成本极高；
4、同一项目采用统一的策略，无法针对单一车辆进行策略修订。
本申请提供的上述动力电池控制策略确定方法，基于大数据修订在役新能源车的动力电池控制策略，通过对在役新能源车进行策略修订，使控制策略更加精准，在不降低用户体验的情况下，降低故障发生概率，延长动力电池寿命。此外，还可用于指导增量市场动力电池控制策略的制定，减少仿真整车测试的成本和缩短整车开发的周期。以及针对存量市场的车辆，定制化地修订控制策略，实现对车辆差异化地控制。
下面以充电倍率修订进行举例说明：电动车辆的原充电策略假定采用阶梯充电的方式，电流设计如图6所示。根据充电工况对工况特征提取并建模，在某项目里一共有9台故障车，在故障预测模型中，确定故障发生的风险阈值为score_threshold，最终故障车辆的排名情况如图7所示。
采用本申请提供的动力电池控制策略确定方法更改充电策略如图8所示，变更前后充电电流对充电时长几乎无影响(电流均值几乎不变)。对于新策略重新提取新的特征，并对新的特征重新预测，最终故障车辆的排名情况如图9所示。通过策略变更使得大多数车故障发生率下降，且需要市场排查的车辆(超过风险阈值score_threshold)的车辆大幅减小，降低了故障排查成本。
在一个实施例中，如图10所示，还提供了一种设备控制方法，包括：
步骤S410：从动力电池的修订策略中确定目标修订策略；动力电池的修订策略为根据上述动力电池控制策略确定方法得到。具体地，服务器得到各故障类型的修订策略后，从修订策略中确定目标修订策略。
步骤S420：将目标修订策略向目标设备发送，以对目标设备的动力电池的控制策略进行更新。服务器将目标修订策略发送至目标车辆，从而对目标车辆的动力电池的控制策略进行更新。
在一个实施例中，步骤S410包括：根据修订策略的优化目标，确定各修订策略关联的目标设备；将与目标设备关联的所有修订策略作为目标修订策略。根据修订策略的优化目标，将与目标设备关联的所有修订策略作为目标修订策略，实现对目标设备的修订策略主动更新。
在另一个实施例中，步骤S410包括：根据目标设备发送的更新请求，从动力电池的修订策略中获取优化目标与目标设备关联的修订策略，作为目标修订策略。可根据实际需要通过目标设备发送更新请求，启动策略更新操作。
可以理解，上述设备控制方法的具体步骤在上述动力电池控制策略确定方法中进行了详细解释说明，在此不再赘述。
上述设备控制方法，根据各故障类型的目标特征生成动力电池的修订策略，以供对动力电池的控制策略进行更新，从而使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。
在一个实施例中，如图11所示，还提供了一种设备控制方法，包括：
步骤S510：获取待更新的动力电池的修订策略；待更新的动力电池的修订策略为根据上述动力电池控制策略确定方法得到。
步骤S520：根据待更新的动力电池的修订策略，对设备的动力电池的控制策略更新。具体地，目标车辆获取服务器下发的待更新的动力电池的修订策略，对车辆的动力电池的控制策略更新。
在一个实施例中，步骤S510包括：接收服务器发送的目标修订策略，作为待更新的动力电池的修订策略；目标修订策略为服务器根据与目标设备关联的所有修订策略得到，目标设备根据修订策略的优化目标确定。根据修订策略的优化目标，将与目标设备关联的所有修订策略作为目标修订策略，实现对目标设备的修订策略主动更新。
在另一个实施例中，步骤S510包括：发送更新请求至服务器；接收服务器响应更新请求返回的目标修订策略，作为待更新的动力电池的修订策略；目标修订策略为服务器根据更新请求，从动力电池的修订策略中获取优化目标与目标设备关联的修订策略得到。可根据实际需要通过目标设备发送更新请求，启动策略更新操作。
可以理解，上述设备控制方法的具体步骤在上述动力电池控制策略确定方法中进行了详细解释说明，在此不再赘述。
上述设备控制方法，根据各故障类型的目标特征生成动力电池的修订策略，以供对动力电池的控制策略进行更新，从而使得动力电池的控制策略更加符合实际工况，使控制策略更加精准，降低故障发生概率，延长动力电池的寿命。
应该理解的是，虽然如上所述的各实施例所涉及的流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，这些步骤可以以其它的顺序执行。而且，如上所述的各实施例所涉及的流程图中的至少一部分步骤可以包括多个步骤或者多个阶段，这些步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，这些步骤或者阶段的执行顺序也不必然是依次进行，而是可以与其它步骤或者其它步骤中的步骤或者阶段的至少一部分轮流或者交替地执行。
基于同样的发明构思，本申请实施例还提供了一种用于实现上述所涉及的动力电池控制策略确定方法的动力电池控制策略确定装置。该装置所提供的解决问题的实现方案与上述方法中所记载的实现方案相似，故下面所提供的一个或多个动力电池控制策略确定装置实施例中的具体限定可以参见上文中对于动力电池控制策略确定方法的限定，在此不再赘述。
在一个实施例中，还提供了一种动力电池控制策略确定装置，其中，控制策略可以是充电控制策略、热管理策略、充放电均衡控制策略等，如图12所示，该装置包括：数据获取模块100、数据分析模块200和策略修订模块300，其中：
数据获取模块100，用于获取对动力电池进行采集得到的采样数据，根据采样数据确定不同故障类型对应的特征。
数据分析模块200，用于根据各故障类型对应的特征，以及各特征对应的采样数据，分析获得各故障类型的目标特征。
策略修订模块300，用于根据各故障类型的目标特征，确定动力电池的控制策略。
在一个实施例中，数据获取模块100对动力电池采集得到的采样数据进行差异性分析，根据差异性分析结果确定不同故障类型对应的特征。
在一个实施例中，数据分析模块200根据各故障类型对应的特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征；根据各故障类型关联的初始目标特征的采样数据，以及各故障类型对应的故障预测模型，确定各故障类型的目标特征。
在一个实施例中，数据分析模块200对各故障类型对应的特征的采样数据进行预 处理，获得预处理后的各特征的采样数据；根据预处理后的各特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征。
在一个实施例中，数据分析模块200对各故障类型关联的初始目标特征的采样数据添加标签，生成各故障类型的训练样本数据，标签包括故障设备标签和正常设备标签；根据各故障类型的训练样本数据进行模型训练，获得各故障类型对应的故障预测模型；根据各故障类型对应的故障预测模型，确定各故障类型的目标特征。
在一个实施例中，数据分析模块200对故障预测模型进行模型评估，将评估通过的故障预测模型作为最终获得的故障预测模型。
在一个实施例中，数据分析模块200根据预设的业务标准对故障预测模型进行业务评估；根据预设的故障树分析结果，对业务评估通过的故障预测模型进行合理性评估，将合理性评估通过的故障预测模型，作为最终获得的故障预测模型。
在一个实施例中，数据分析模块200根据各故障类型对应的集成学习模型的特征重要性，确定各故障类型的目标特征。
在一个实施例中，策略修订模块300根据各故障类型的目标特征生成动力电池的初步修订策略；根据初步修订策略模拟设备运行，获得模拟采样数据；根据模拟采样数据和故障预测模型进行策略评估，得到动力电池的控制策略。
上述动力电池控制策略确定装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
在一个实施例中，如图13所示，还提供了一种设备控制装置，包括：策略确定模块410和策略发送模块420，其中：
策略确定模块410，用于从动力电池的修订策略中确定目标修订策略；动力电池的修订策略为根据上述的动力电池控制策略确定装置得到。
策略发送模块420，用于将目标修订策略向目标设备发送，以对目标设备的动力电池的控制策略进行更新。
在一个实施例中，策略确定模块410根据修订策略的优化目标，确定各修订策略关联的目标设备；将与目标设备关联的所有修订策略作为目标修订策略。
在一个实施例中，策略确定模块410根据目标设备发送的更新请求，从动力电池的修订策略中获取优化目标与目标设备关联的修订策略，作为目标修订策略。
可以理解，上述设备控制装置的具体原理在上述动力电池控制策略确定装置中进行了解释说明，在此不再赘述。
在一个实施例中，如图14所示，还提供了一种设备控制装置，包括：策略获取模块510和策略更新模块520，其中：
策略获取模块510，用于获取待更新的动力电池的修订策略；待更新的动力电池的修订策略为上述的动力电池控制策略确定装置得到。
策略更新模块520，用于根据待更新的动力电池的修订策略，对设备的动力电池的控制策略更新。
在一个实施例中，策略获取模块510接收服务器发送的目标修订策略，作为待更新的动力电池的修订策略；目标修订策略为服务器根据与目标设备关联的所有修订策略得到，目标设备根据修订策略的优化目标确定。
在一个实施例中，策略获取模块510发送更新请求至服务器；接收服务器响应更新请求返回的目标修订策略，作为待更新的动力电池的修订策略；目标修订策略为服务器根据更新请求，从动力电池的修订策略中获取优化目标与目标设备关联的修订策略得到。
可以理解，上述设备控制装置的具体原理在上述动力电池控制策略确定装置中进行了解释说明，在此不再赘述。
在一个实施例中，提供了一种计算机设备，该计算机设备可以是服务器，其内部结构图可以如图15所示。该计算机设备包括处理器、存储器、输入/输出接口(Input/Output，简称I/O)和通信接口。其中，处理器、存储器和输入/输出接口通过系统总线连接，通信接口通过输入/输出接口连接到系统总线。其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质和内存储器。该非易失性存储介质存储有操作系统、计算机程序和数据库。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该计算机设备的数据库用于存储数据。该计算机设备的输入/输出接口用于处理器与外部设备之间交换信息。该计算机设备的通信接口用于与外部的终端通过网络连接通信。该计算机程序被处理器执行时以实现一种动力电池控制策略确定方法，或设备控制方法。
本领域技术人员可以理解，图15中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的计算机设备的限定，具体的计算机设备可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
在一个实施例中，还提供了一种计算机设备，包括存储器和处理器，存储器存储有计算机程序，处理器执行计算机程序时实现上述方法的步骤。
在一个实施例中，还提供了一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现上述方法的步骤。
在一个实施例中，还提供了一种计算机程序产品，包括计算机程序，该计算机程序 被处理器执行时实现上述方法的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，该计算机程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，前述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)等非易失性存储介质，或随机存储记忆体(Random Access Memory，RAM)等。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进，这些都属于本发明的保护范围。因此，本发明专利的保护范围应以所附权利要求为准。

Claims (25)

  1. 一种动力电池控制策略确定方法，其特征在于，包括：
    获取对动力电池进行采集得到的采样数据，根据所述采样数据确定不同故障类型对应的特征；
    根据各故障类型对应的特征，以及各所述特征对应的采样数据，分析获得各故障类型的目标特征；
    根据各故障类型的目标特征，确定动力电池的控制策略。
  2. 根据权利要求1所述的方法，其特征在于，所述根据所述采样数据确定不同故障类型对应的特征，包括：
    对动力电池采集得到的采样数据进行差异性分析，根据差异性分析结果确定不同故障类型对应的特征。
  3. 根据权利要求2所述的方法，其特征在于，所述差异性分析包括统计特征分析、变量间的交叉特征分析以及时序趋势变化特征分析中的至少一种。
  4. 根据权利要求1所述的方法，其特征在于，所述根据各故障类型对应的特征，以及各所述特征对应的采样数据，分析获得各故障类型的目标特征，包括：
    根据各故障类型对应的特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征；
    根据各故障类型关联的初始目标特征的采样数据，以及各故障类型对应的故障预测模型，确定各故障类型的目标特征。
  5. 根据权利要求4所述的方法，其特征在于，所述根据各故障类型对应的特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征，包括：
    对各故障类型对应的特征的采样数据进行预处理，获得预处理后的各特征的采样数据；
    根据预处理后的各特征的采样数据，对各故障类型对应的特征进行筛选，得到与各故障类型关联的初始目标特征。
  6. 根据权利要求5所述的方法，其特征在于，所述预处理包括数据清洗和数据补全中的至少一种。
  7. 根据权利要求4所述的方法，其特征在于，所述根据各故障类型关联的初始目标特征的采样数据，以及各故障类型对应的故障预测模型，确定各故障类型的目标特征，包括：
    对各故障类型关联的初始目标特征的采样数据添加标签，生成各故障类型的训练样本 数据，所述标签包括故障设备标签和正常设备标签；
    根据各故障类型的训练样本数据进行模型训练，获得各故障类型对应的故障预测模型；
    根据各故障类型对应的故障预测模型，确定各故障类型的目标特征。
  8. 根据权利要求7所述的方法，其特征在于，所述根据各故障类型的训练样本数据进行模型训练，获得各故障类型对应的故障预测模型之后，还包括：对故障预测模型进行模型评估，将评估通过的故障预测模型作为最终获得的故障预测模型。
  9. 根据权利要求8所述的方法，其特征在于，所述对故障预测模型进行模型评估，将评估通过的故障预测模型作为最终获得的故障预测模型，包括：
    根据预设的业务标准对故障预测模型进行业务评估；
    根据预设的故障树分析结果，对业务评估通过的故障预测模型进行合理性评估，将合理性评估通过的故障预测模型，作为最终获得的故障预测模型。
  10. 根据权利要求7所述的方法，其特征在于，所述故障预测模型为集成学习模型；所述根据各故障类型对应的故障预测模型，确定各故障类型的目标特征，包括：
    根据各故障类型对应的集成学习模型的特征重要性，确定各故障类型的目标特征。
  11. 根据权利要求4所述的方法，其特征在于，所述根据各故障类型的目标特征，确定动力电池的控制策略，包括：
    根据各故障类型的目标特征生成动力电池的初步修订策略；
    根据所述初步修订策略模拟设备运行，获得模拟采样数据；
    根据所述模拟采样数据和所述故障预测模型进行策略评估，得到动力电池的控制策略。
  12. 根据权利要求11所述的方法，其特征在于，所述策略评估包括故障率评估和/或业务评估。
  13. 根据权利要求1-12任意一项所述的方法，其特征在于，所述采样数据包括电流数据、温度数据、电压数据和绝缘数据中的至少一种。
  14. 一种设备控制方法，其特征在于，包括：
    从动力电池的修订策略中确定目标修订策略；所述动力电池的修订策略为根据权利要求1-13任意一项所述的方法得到；
    将所述目标修订策略向目标设备发送，以对所述目标设备的动力电池的控制策略进行更新。
  15. 根据权利要求14所述的方法，其特征在于，所述从动力电池的修订策略中确定目标修订策略，包括：
    根据修订策略的优化目标，确定各所述修订策略关联的目标设备；
    将与所述目标设备关联的所有修订策略作为目标修订策略。
  16. 根据权利要求14所述的方法，其特征在于，所述从动力电池的修订策略中确定目标修订策略，包括：
    根据目标设备发送的更新请求，从动力电池的修订策略中获取优化目标与所述目标设备关联的修订策略，作为目标修订策略。
  17. 一种设备控制方法，其特征在于，包括：
    获取待更新的动力电池的修订策略；所述待更新的动力电池的修订策略为根据权利要求1-13任意一项所述的方法得到；
    根据所述待更新的动力电池的修订策略，对设备的动力电池的控制策略更新。
  18. 根据权利要求17所述的方法，其特征在于，所述获取待更新的动力电池的修订策略，包括：
    接收服务器发送的目标修订策略，作为待更新的动力电池的修订策略；所述目标修订策略为所述服务器根据与目标设备关联的所有修订策略得到，所述目标设备根据修订策略的优化目标确定。
  19. 根据权利要求17所述的方法，其特征在于，所述获取待更新的动力电池的修订策略，包括：
    发送更新请求至服务器；
    接收所述服务器响应所述更新请求返回的目标修订策略，作为待更新的动力电池的修订策略；所述目标修订策略为所述服务器根据所述更新请求，从动力电池的修订策略中获取优化目标与目标设备关联的修订策略得到。
  20. 一种动力电池控制策略确定装置，其特征在于，包括：
    数据获取模块，用于获取对动力电池进行采集得到的采样数据，根据所述采样数据确定不同故障类型对应的特征；
    数据分析模块，用于根据各故障类型对应的特征，以及各所述特征对应的采样数据，分析获得各故障类型的目标特征；
    策略修订模块，用于根据各故障类型的目标特征，确定动力电池的控制策略。
  21. 一种设备控制装置，其特征在于，包括：
    策略确定模块，用于从动力电池的修订策略中确定目标修订策略；所述动力电池的修订策略为根据权利要求20所述的装置得到；
    策略发送模块，用于将所述目标修订策略向目标设备发送，以对所述目标设备的动力电池的控制策略进行更新。
  22. 一种设备控制装置，其特征在于，包括：
    策略获取模块，用于获取待更新的动力电池的修订策略；所述待更新的动力电池的修订策略为根据权利要求20所述的装置得到；
    策略更新模块，用于根据所述待更新的动力电池的修订策略，对设备的动力电池的控制策略更新。
  23. 一种计算机设备，包括存储器和处理器，所述存储器存储有计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1至19中任一项所述的方法的步骤。
  24. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1至19中任一项所述的方法的步骤。
  25. 一种计算机程序产品，包括计算机程序，其特征在于，该计算机程序被处理器执行时实现权利要求1至19中任一项所述的方法的步骤。
PCT/CN2022/131261 2022-11-11 2022-11-11 动力电池控制策略确定方法、设备控制方法和装置 WO2024098357A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/131261 WO2024098357A1 (zh) 2022-11-11 2022-11-11 动力电池控制策略确定方法、设备控制方法和装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/131261 WO2024098357A1 (zh) 2022-11-11 2022-11-11 动力电池控制策略确定方法、设备控制方法和装置

Publications (1)

Publication Number Publication Date
WO2024098357A1 true WO2024098357A1 (zh) 2024-05-16

Family

ID=91031754

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/131261 WO2024098357A1 (zh) 2022-11-11 2022-11-11 动力电池控制策略确定方法、设备控制方法和装置

Country Status (1)

Country Link
WO (1) WO2024098357A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110091748A (zh) * 2019-05-21 2019-08-06 广州小鹏汽车科技有限公司 电动车辆充电异常的处理方法、装置和车辆
CN110492557A (zh) * 2019-07-30 2019-11-22 深圳易马达科技有限公司 一种电池的故障处理方法及设备
CN110712560A (zh) * 2019-09-27 2020-01-21 东软睿驰汽车技术（沈阳）有限公司 异常处理方法、装置、设备及介质
CN113406438A (zh) * 2021-06-23 2021-09-17 安徽南瑞中天电力电子有限公司 一种适用于低压台区的故障智能诊断方法及其运维系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110091748A (zh) * 2019-05-21 2019-08-06 广州小鹏汽车科技有限公司 电动车辆充电异常的处理方法、装置和车辆
CN110492557A (zh) * 2019-07-30 2019-11-22 深圳易马达科技有限公司 一种电池的故障处理方法及设备
CN110712560A (zh) * 2019-09-27 2020-01-21 东软睿驰汽车技术（沈阳）有限公司 异常处理方法、装置、设备及介质
CN113406438A (zh) * 2021-06-23 2021-09-17 安徽南瑞中天电力电子有限公司 一种适用于低压台区的故障智能诊断方法及其运维系统

Similar Documents

Publication Publication Date Title
CN114154429B (zh) 一种储能系统的数字孪生体构建方法、装置及存储介质
Bernardon et al. Multicriteria distribution network reconfiguration considering subtransmission analysis
CN110658460B (zh) 一种电池包的电池寿命预测方法及装置
CN112380630B (zh) 基于车联网数据的新能源汽车电池热失控风险的预测方法
CN114371409B (zh) 电池状态预测模型的训练方法、电池状态预测方法及装置
EP4055399A1 (en) Battery performance prediction
Xiang et al. Scale evolution of electric vehicles: A system dynamics approach
CN112633316A (zh) 一种基于边界估值理论的负荷预测方法和装置
CN113822481A (zh) 基于多任务学习策略和深度学习的综合能源负荷预测方法
CN116911076B (zh) 多微电网对配电网的韧性支撑仿真方法、装置和电子设备
CN116821660A (zh) 一种电力管理方法、系统、终端设备及存储介质
Qiu et al. A conditional generative adversarial network-based synthetic data augmentation technique for battery state-of-charge estimation
CN117174899B (zh) 一种氟化碳正极材料制备方法
Zeng et al. Forecasting the sustainable classified recycling of used lithium batteries by gray Graphical Evaluation and Review Technique
WO2024098357A1 (zh) 动力电池控制策略确定方法、设备控制方法和装置
CN105514988A (zh) 一种计及动态时空特征的微电网电源规划方案优选方法
Morison et al. Identification of voltage control areas and reactive power reserve; an advancement in on-line voltage security assessment
CN110059840B (zh) 一种受端电网中电池储能系统选址方法及系统
CN110457448A (zh) 问题推送方法、装置、计算机可读存储介质和计算机设备
CN116128279A (zh) 一种电网薄弱环节改造方法
CN115144757A (zh) 一种获取电池产热的方法、装置和电子设备
CN114091243A (zh) 一种产品概念设计可靠性的预测方法及系统
Marchegiani et al. Li-ion battery aging model robustness: An analysis using univariate and multivariate techniques
Hametner et al. Model based lithium ion cell ageing data analysis
CN116579121B (zh) 用于回收动力电池模组的人机协作拆卸线平衡方法