CN117910148B

CN117910148B - BMS optimization method, system, medium and device based on shadow mode

Info

Publication number: CN117910148B
Application number: CN202410309589.7A
Authority: CN
Inventors: 柳华勤; 张微中; 柳扬
Original assignee: Daqin Digital Energy Technology Co ltd
Current assignee: Daqin Digital Energy Technology Co ltd
Priority date: 2024-03-19
Filing date: 2024-03-19
Publication date: 2024-05-24
Anticipated expiration: 2044-03-19
Also published as: CN117910148A

Abstract

BMS optimization method, system, medium and equipment based on shadow mode relate to BMS optimization technical field. The method comprises the following steps: triggering a shadow mode when the BMS detects a predicted event or a decision event; if the event is a predicted event, determining a first prediction model actually operated on battery equipment managed by the BMS, acquiring a first prediction result of the first prediction model, and simultaneously acquiring a second prediction result of a second prediction model virtually operated in a shadow mode; optimizing the first prediction model or the second prediction model based on the first prediction result and the second prediction result; if the decision event is the decision event, a first decision result of actually executing a first decision scheme on the battery equipment is obtained, and a second decision result of virtually executing a second decision scheme in a shadow mode is obtained at the same time; the first decision scheme or the second decision scheme is optimized based on the first decision result and the second decision result. By implementing the technical scheme provided by the application, the overall performance of the BMS can be improved.

Description

BMS optimization method, system, medium and device based on shadow mode

Technical Field

The application relates to the technical field of BMS optimization, in particular to a BMS optimization method, system, medium and device based on a shadow mode.

Background

With the popularization of electric vehicles, a Battery management system (Battery MANAGEMENT SYSTEM, BMS) is used as one of the core components of the electric vehicles, and the main function of the Battery management system is to accurately monitor and protect Battery equipment through a system algorithm, wherein the system algorithm mainly comprises a prediction model algorithm of the Battery equipment and a decision scheme algorithm when the Battery fails, so that safe and reliable operation of the Battery equipment is ensured.

The existing BMS has certain limitation in the aspect of decision management when battery state prediction and battery faults are carried out, and the performance of the BMS is directly affected, for example, a prediction model algorithm and a decision scheme algorithm in the existing BMS still operate in a relatively fixed and traditional mode, so that the feedback period of algorithm training is longer, the complex change condition of a battery is difficult to adapt, the accuracy of the algorithm in the BMS is lower, and the overall performance of the BMS is poorer.

Disclosure of Invention

The application provides a BMS optimization method, a system, a medium and equipment based on a shadow mode, which can rapidly optimize a prediction model and a decision scheme in the BMS and improve the accuracy of an algorithm in the BMS, thereby improving the overall performance of the BMS.

In a first aspect, the present application provides a BMS optimization method based on a shadow pattern, the method comprising:

Triggering a shadow mode corresponding to the BMS when the BMS detects a predicted event or a decision event;

if the BMS detects the predicted event, acquiring first target data corresponding to the predicted event, determining a first prediction model actually operated on battery equipment managed by the BMS, acquiring a first prediction result of the first prediction model for predicting based on the first target data, and simultaneously acquiring a second prediction result of a second prediction model virtually operated in the shadow mode for predicting based on the first target data;

Optimizing the first prediction model or the second prediction model based on the first prediction result and the second prediction result;

If the BMS detects the decision event, acquiring second target data corresponding to the decision event, determining a first decision scheme actually executed on the battery equipment, acquiring a first decision result of executing the first decision scheme based on the second target data, and simultaneously acquiring a second decision result of a second decision scheme virtually executed in the shadow mode based on the second target data;

optimizing the first decision scheme or the second decision scheme based on the first decision result and the second decision result.

By adopting the technical scheme, when the BMS detects the predicted event or the decision event, namely triggering the shadow mode, for the predicted event, the first predicted result is obtained by the first predicted model which is actually operated on the battery equipment, meanwhile, the second predicted result is obtained by the second predicted model which is different and is operated in parallel under the shadow mode, the feedback period of model training is shortened by the shadow mode, the BMS can be more rapidly adapted to the continuously changing battery state, the accuracy of the predicted model is improved, the rapid iterative optimization of the BMS predicted model is realized, for the decision event, the first decision result is obtained by the first decision scheme which is actually operated on the battery, meanwhile, the second decision result is obtained by simulating the second decision scheme which is different and is operated in parallel under the shadow mode, the scheme with better decision effect is found by comparison, the rapid iterative optimization of the decision scheme is completed, meanwhile, the more optimal scheme is provided rapidly under the decision, the comprehensive use of the algorithm is realized, the BMS is comprehensively optimized in the prediction algorithm and the decision algorithm, the accuracy of the algorithm in the BMS is improved, and the overall performance of the BMS is improved.

Optionally, the determining that the BMS detects the predicted event or the decision event includes: calculating a deviation value between the first prediction result and the second prediction result in real time, wherein the first prediction model and the second prediction model are in a real-time running state; when the deviation value is larger than a preset value, determining that the BMS detects a predicted event; and acquiring each piece of operation data of the battery equipment, and determining that the BMS detects a decision event when detecting that any piece of operation data exceeds a corresponding preset range.

By adopting the technical scheme, the prediction deviation and the abnormal condition of the battery in the BMS are captured in real time, so that the processing flow of the corresponding event is triggered, wherein the prediction model result deviation value judgment method can realize continuous monitoring of the prediction precision, verify the real-time performance and the accuracy of the prediction model, and ensure the reliability of a prediction system. The battery hyper-parameter detection method can rapidly reflect the abnormal condition of the battery, and timely start a decision scheme to protect and manage the battery.

Optionally, the obtaining the first target data corresponding to the predicted event includes: in the BMS detection process, acquiring the multi-mode data acquired by the BMS in real time; and if the BMS detects the predicted event, screening out first target data corresponding to the predicted event from the multi-mode data.

By adopting the technical scheme, the BMS can continuously collect data of various parameters such as voltage, current and temperature of the battery and environmental parameters in the running process, and store the data to form a multi-mode monitoring data set, and the first target data related to the predicted event is screened out from the multi-mode data, so that the accurate prediction of the predicted event can be improved, and meanwhile, the interference of other irrelevant data is avoided.

Optionally, after the obtaining the first target data corresponding to the predicted event, the method further includes: classifying the first target data based on a clustering algorithm to obtain classified first target data, and uploading the classified first target data to a cloud.

By adopting the technical scheme, the influence and the effect of the first target data of different types on the prediction result are different, the target data can be clustered into a plurality of classes according to the characteristics through the clustering algorithm, so that the data with larger and smaller influence on the prediction can be distinguished, the data with smaller influence on the prediction can be regarded as redundant data for filtering by the classification, the input quantity of a prediction model is reduced, the calculation complexity is reduced, meanwhile, the most critical target data can be determined, and the accuracy of the prediction is improved.

Optionally, the obtaining a second prediction result of the target data predicted by the second prediction model running virtually in the shadow mode includes: at least one second prediction model corresponding to the prediction event is called in a preset prediction model library; and inputting the first target data into the second prediction model to obtain a second prediction result of virtually running the second prediction model in the shadow mode.

By adopting the technical scheme, the prediction model library is built, because the optimal prediction models of the prediction model library are different for different prediction events in the BMS, the model library can provide storage registration of various prediction models and match the prediction models according to the types of the prediction events, so that the BMS can quickly acquire the matched prediction models, a second prediction result which is not influenced by the actual system environment can be obtained by combining accurate calling and shadow running of the prediction models, the result can be compared with the first prediction result, and the advantages and disadvantages of the two models are judged to select a better model or inventory model deviation reasons, so that model optimization is realized.

Optionally, the optimizing the first prediction model or the second prediction model based on the first prediction result and the second prediction result includes: acquiring an actual operation result of the battery equipment under the predicted event after a preset time; calculating a first deviation value of the first prediction result and the operation result, and a second deviation value of the second prediction result and the operation result; if the first deviation value is smaller than the second deviation value, marking the first prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the first prediction model reach preset times, taking the first prediction model as a standard prediction model corresponding to the prediction event and training and optimizing the second prediction model; and if the second deviation value is smaller than the first deviation value, marking the second prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the second prediction model reach the preset times, taking the second prediction model as a standard prediction model corresponding to the prediction event and training and optimizing the first prediction model.

By adopting the technical scheme, the actual operation data of the battery under the predicted event is obtained after a certain time as an accurate result. The actual result is respectively compared with the first prediction result and the second prediction result, the deviation value of the two prediction results is calculated, the relative prediction precision of the two prediction models is judged according to the magnitude of the deviation value, the ordering and judgment of the prediction capacities of the two prediction models can be realized, and the model selection, the optimization and the update are automatically carried out, so that the prediction system can be continuously self-perfected.

Optionally, the optimizing the first decision scheme or the second decision scheme based on the first decision result and the second decision result includes: comparing the first decision result with the second decision result to obtain a comparison result; if the comparison result is that the first decision result is better than the second decision result, marking the first decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the first decision scheme reach the preset times, taking the first decision scheme as a standard decision scheme corresponding to the decision event and optimizing the second decision scheme; if the comparison result is that the second decision result is better than the first decision result, marking the second decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the second decision scheme reach the preset times, taking the second decision scheme as a standard decision scheme corresponding to the decision event and optimizing the first decision scheme.

By adopting the technical scheme, the processing results of the two decision schemes under the same decision event are directly compared, if the processing effect of the first scheme is better, the first scheme is marked as the current target scheme, if the marking times reach the threshold value, the scheme is set as the standard scheme, and meanwhile, the further optimization adjustment of the second scheme is started to continuously approximate the processing effect of the first scheme, so that the ordering and judgment of the effects of the two decision schemes can be realized, the scheme selection and optimization updating are independently carried out, the decision system can be continuously self-perfected, and meanwhile, the setting of the standard scheme provides reliability guarantee for the subsequent treatment decision.

In a second aspect of the present application, there is provided a shadow pattern based BMS optimization system, the system comprising:

the shadow mode triggering module is used for triggering a shadow mode corresponding to the BMS when the BMS detects a predicted event or a decision event;

The BMS is used for acquiring a first target data corresponding to the predicted event, determining a first predicted model actually operated on battery equipment managed by the BMS, acquiring a first predicted result of the first predicted model predicted based on the first target data, and simultaneously acquiring a second predicted result of a second predicted model virtually operated in the shadow mode predicted based on the first target data;

a prediction model optimization module that optimizes the first prediction model or the second prediction model based on the first prediction result and the second prediction result;

The decision scheme executing module is used for acquiring second target data corresponding to the decision event if the BMS detects the decision event, determining a first decision scheme actually executed on the battery equipment, acquiring a first decision result of executing the first decision scheme based on the second target data, and simultaneously acquiring a second decision result of a second decision scheme virtually executed in the shadow mode based on the second target data;

And the decision scheme optimizing module is used for optimizing the first decision scheme or the second decision scheme based on the first decision result and the second decision result.

In a third aspect the application provides a computer storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the above-described method steps.

In a fourth aspect of the application there is provided an electronic device comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the above-mentioned method steps.

In summary, one or more technical solutions provided in the embodiments of the present application at least have the following technical effects or advantages:

When the BMS detects a prediction event or a decision event, namely triggering a shadow mode, for the prediction event, a first prediction model actually operated on battery equipment obtains a first prediction result, and simultaneously, a second prediction model which is different in parallel operated under the shadow mode obtains a second prediction result, the feedback period of model training is shortened through the shadow mode, so that the BMS can be more rapidly adapted to the continuously changed battery state, the accuracy of the prediction model is improved, the rapid iterative optimization of the BMS prediction model is realized, for the decision event, a first decision result is obtained by a first decision scheme actually operated on a battery, and simultaneously, a second decision result is obtained by simulating a second decision scheme which is different in parallel operated under the shadow mode, the rapid iterative optimization of the decision scheme is completed by comparing and finding a scheme with better decision effect, and meanwhile, the method is beneficial to rapidly providing a better scheme in decision time, and the shadow algorithm is comprehensively used, so that the BMS is comprehensively optimized in the prediction algorithm and the decision algorithm, the accuracy of the algorithm in the BMS is improved, and the overall performance of the BMS is improved.

Drawings

Fig. 1 is a schematic flow chart of a BMS optimization method based on a shadow mode according to an embodiment of the present application;

fig. 2 is a schematic block diagram of a BMS optimization system based on a shadow mode according to an embodiment of the present application;

fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Reference numerals illustrate: 300. an electronic device; 301. a processor; 302. a communication bus; 303. a user interface; 304. a network interface; 305. a memory.

Detailed Description

In order that those skilled in the art will better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments.

In describing embodiments of the present application, words such as "for example" or "for example" are used to mean serving as examples, illustrations, or descriptions. Any embodiment or design described herein as "such as" or "for example" in embodiments of the application should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "or" for example "is intended to present related concepts in a concrete fashion.

In the description of embodiments of the application, the term "plurality" means two or more. For example, a plurality of systems means two or more systems, and a plurality of screen terminals means two or more screen terminals. Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating an indicated technical feature. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless expressly specified otherwise.

The following description of the embodiments of the present application will be given in detail with reference to the accompanying drawings, and it is apparent that the embodiments described are only some, but not all embodiments of the present application.

Referring to fig. 1, a schematic flow chart of BMS optimization based on a shadow mode is specifically provided, the method may be implemented by a computer program, may be implemented by a single chip microcomputer, may also be run on a BMS optimization system based on a shadow mode, and the computer program may be integrated in an intelligent terminal or may also be run as an independent tool application, and specifically, the method includes steps 10 to 50, where the steps are as follows:

step 10: when the BMS is determined to detect a predicted event or a decision event, a shadow pattern corresponding to the BMS is triggered.

The embodiment of the application can be applied to all fields needing to optimize the battery management system, including but not limited to electric automobiles, storage batteries, electronic equipment and other battery systems using BMS, for example, the battery state can be evaluated, the power management can be performed, and the like.

The Battery management system (Battery MANAGEMENT SYSTEM, BMS) is an electronic system for managing, protecting and monitoring Battery equipment, is a core component for ensuring safe and reliable operation of a Battery, and the BMS can monitor parameters such as voltage, current, temperature and the like of the Battery equipment in real time and can perform functions such as charge and discharge management, battery equalization, state evaluation and the like.

The shadow mode refers to an emerging system modeling method, and the core idea is to dynamically simulate various states and operation scenes of a real system by establishing a virtual 'shadow' system so as to realize performance evaluation and optimization of the real system. The shadow system can execute multiple schemes in parallel, and normal work of the real system is not interfered. In the embodiment of the application, the shadow mode is that when the BMS operates normally, a virtual BMS corresponding to the real BMS is established.

In the embodiment of the application, a predicted event and a decision event can be called as a trigger event, wherein the predicted event refers to an event which is judged to be abnormal in prediction when a large deviation exists between two predicted results of a plurality of predicted models during battery state prediction; the decision event is a decision event when the actual operation parameters of the battery are detected to be out of the normal range during the charge and discharge management of the battery.

Specifically, in order to implement optimization and upgrade of the BMS, it is necessary to be able to quickly iterate the prediction model and decision scheme of the system. According to the scheme, in the BMS monitoring process, when abnormal events such as abnormal prediction results or abnormal decision effects are detected, the shadow BMS is triggered to perform parallel calculation, so that the situation is achieved, because the prediction deviation is overlarge, the space which can be optimized exists in the BMS is reflected by decision failure, and the shadow mode can obtain result data under different running conditions in parallel, so that better prediction models or decision schemes can be found.

On the basis of the above embodiments, as an alternative embodiment, the step of determining that the BMS detects the predicted event or the decision event may further include the steps of:

Step 101: and calculating a deviation value between the first prediction result and the second prediction result in real time, wherein the first prediction model and the second prediction model are in real-time running states.

Specifically, the prediction models in the BMS are all real-time running states, and the prediction models can be a first prediction model actually running on the battery equipment or at least one second prediction model virtually running in a shadow mode, so that the system can make the prediction models use the same group of real-time collected battery data to predict, and therefore battery health state results output by the prediction models are obtained in parallel, and the prediction capability of the models can be directly compared by making a plurality of models predict in parallel. In the embodiment of the application, all the prediction models corresponding to the prediction event are regarded as being composed of the first prediction model and the second prediction model, and a plurality of other prediction models can be included, so that in order to judge whether the prediction models are abnormal, the difference between the prediction results of different models needs to be quantitatively analyzed, after the prediction results of each prediction model on the same data set are obtained, namely after the first prediction result corresponding to the first prediction model and the second prediction result corresponding to the second prediction model are obtained, the system calculates the deviation value between the first prediction result and the second prediction result in real time by adopting methods such as root mean square error and the like, and the deviation of the prediction results is calculated because the prediction values of different prediction models are directly compared, whether the prediction effects of the models have significant differences can be quantitatively judged, and the situation of prediction abnormality is reflected.

Step 102: when the existing deviation value is greater than a preset value, it is determined that the BMS detects the predicted event.

Specifically, after obtaining the deviation value between the prediction outputs of each prediction model, comparing the deviation value with a preset value, and when the deviation between two prediction models exceeds the preset value, judging that the BMS detects a prediction event at the moment, wherein the fact that the deviation between different prediction models is overlarge indicates that the prediction model which needs to be optimized exists in the system.

Step 103: and acquiring each operation data of the battery equipment, and determining that the BMS detects a decision event when detecting that any operation data exceeds a corresponding preset range.

Specifically, to optimize the decision scheme of the BMS, it is necessary to monitor whether the real-time operation data of the battery device is within a normal range. When detecting that a certain battery operation parameter exceeds a corresponding preset range, the method can determine that the battery operation parameter is a decision event, and the decision scheme needs to be optimized. The BMS of the present application continuously detects the operation index of the battery device, such as data of voltage, current, temperature, etc., and compares it with a preset range set correspondingly. These operating parameters of the battery reflect the effects of the power management decision strategy. When the detection finds that the parameters such as the voltage of the single battery or the working temperature of the battery exceed the normal preset range configured in advance, the abnormal running state of the battery can be judged, the abnormal state of the battery needs to be processed by starting a decision scheme, and the BMS can judge that a decision event is detected.

On the basis of the foregoing embodiment, as an optional embodiment, after determining that the BMS detects the predicted event or the decision event and triggering the shadow mode corresponding to the BMS, the method further includes a process of collecting and uploading data corresponding to the predicted event or the decision event, which specifically includes the following steps:

Specifically, in order to make full use of data collected during the operation of the BMS, multi-mode monitoring data of the BMS needs to be acquired in real time. These multimodal data can provide additional useful information for analyzing the cause of anomalies, aiding in decision making, when a predicted event or decision event is detected. The multi-mode data refers to data which is collected by the BMS in the running process and contains multiple modes or forms, mainly comprises, but is not limited to, environment data and battery monitoring data, wherein the environment data refers to monitoring data of the working environment of the battery, such as sensor data of parameters of temperature, humidity, vibration and the like, and the environment data can reflect the environment condition of the battery. The battery data refers to data of directly monitoring the performance of the battery, such as voltage, current, temperature, internal resistance and other parameters, which can directly reflect the working condition of the battery.

When it is determined that the BMS detects the predicted event or the decision event, a shadow mode needs to be started to perform model or policy optimization, and the BMS will first screen target data corresponding to the predicted event or the decision event time period from collected historical multi-mode data, where the data are more valuable and include abnormal richer information, for example, when it is determined that the BMS detects the predicted event, first target data corresponding to the predicted event time period is screened from the multi-mode data, where the first target data includes real-time battery data and environment data corresponding to the predicted event time period. When the BMS detects the decision event, second target data of a time period corresponding to the decision event is screened out from the multi-mode data.

Further, the BMS may use a clustering algorithm, such as K-Means, to perform classification analysis on the screened abnormal related data, for example, to divide the events with types of temperature anomalies together, because the data patterns corresponding to different types of anomalies are different, and classification can help to analyze the data features of each type of anomalies more clearly. The classified data are uploaded to a cloud server, so that deep data training and mining are carried out by utilizing mass storage and computing resources of cloud computing, an optimal prediction model or decision scheme is determined quickly in an auxiliary mode, online optimization and upgrading of the BMS system are completed, finer data support is provided for the system, and data uploading efficiency is optimized. It should be noted that, the multimodal data in the normal state does not need to be uploaded to the cloud, which also improves the data uploading efficiency.

Step 20: if the BMS detects a predicted event, acquiring first target data corresponding to the predicted event, determining a first prediction model actually operated on battery equipment managed by the BMS, acquiring a first prediction result of the first prediction model for predicting based on the first target data, and simultaneously acquiring a second prediction result of a second prediction model virtually operated in a shadow mode for predicting based on the first target data.

Specifically, when the BMS detects a predicted event, in order to evaluate the effects of different prediction models, it is necessary to obtain a first prediction model output, that is, a first prediction result, running on a real battery device, and simultaneously, to obtain a second prediction result by running a different second model in a shadow system, so as to perform a comparative analysis. In specific implementation, when a prediction abnormality occurs after a prediction event is detected by an actual BMS, in order to locate a problem source, a first prediction model of actual operation of battery equipment managed by the current BMS needs to be defined, where the current first prediction model may be a prediction model selected by default or manually in advance for the battery equipment, for example, a model based on an LSTM network, and then the system may screen data at a time point when the prediction abnormality occurs from stored multi-mode data, as first target data corresponding to the abnormality, where the first target data includes battery operating parameters such as voltage and temperature. Meanwhile, corresponding input data, namely first target data, used by the first prediction model when the abnormality occurs is needed, and the first prediction model predicts the first target data to obtain a first prediction result, wherein the first prediction result can comprise the service life of a battery, the residual electric quantity, the state of the battery and the like.

Meanwhile, the BMS in the shadow mode calls a configured preset second prediction model, such as a model based on a random forest algorithm, uses the same group of first target data as input data, and predicts in the shadow system to obtain a second prediction result. The prediction output of the two models on the same data set is obtained in parallel, the prediction effect difference of the two models can be directly compared, and the prediction result of which model is more accurate is judged, so that a model which is better in the current state is determined, and the model is used for adjusting and optimizing the prediction model of the real BMS system in the follow-up process so as to correct the prediction deviation and improve the prediction precision of the model.

Based on the foregoing embodiment, as an optional embodiment, the step of obtaining a second prediction result of predicting the target data by the second prediction model that virtually operates in the shadow mode may further include the steps of:

step 201: and retrieving at least one second prediction model corresponding to the prediction event from a preset prediction model library.

Step 202: and inputting the first target data into the second prediction model to obtain a second prediction result of virtually running the second prediction model in the shadow mode.

Specifically, in order to evaluate the effect of different prediction models under the current condition, in addition to obtaining the result of the first prediction model, other models need to be called to perform prediction so as to obtain a second prediction result for comparison analysis. In specific implementation, the system establishes a prediction model library of various prediction models in advance, wherein the prediction model library comprises various prediction models based on LSTM, random forests and the like. When a prediction abnormality occurs, the system selects one or more alternative prediction models related to the predicted event from the prediction model library as a second prediction model. Taking the current battery health prediction as an example, another battery health state prediction model is called from a prediction model library to serve as a second prediction model, then the obtained first target data for triggering the current prediction event is input, the second prediction model is operated in a shadow BMS system, and the prediction output of the second prediction model on the same input is obtained to serve as a second prediction result. The prediction output of the two models to the same group of input data can directly compare the prediction robustness and effect of the two models under the abnormal condition, and can assist in judging whether the first prediction model needs to be adjusted and optimized or not, so that the prediction capability of the BMS is improved.

Step 30: and optimizing the first prediction model or the second prediction model based on the first prediction result and the second prediction result.

Specifically, after two sets of prediction results of the first prediction model and the second prediction model on the same input data are obtained, the comparison condition of the two sets of prediction results can be utilized to determine and optimize the prediction model in the BMS, so that the prediction performance is improved. The first prediction result may reflect problems and defects of the prediction model used in the current BMS, and a prediction error or abnormality occurs, while the second prediction result may be regarded as a comparison reference under the same input, and it may be determined whether the prediction of the first model is excessively biased. In practice, the system can count and compare the error magnitude, trend deviation and the like of the two results to judge the problems of the first prediction model, such as insufficient training, excessively complex model and the like. If the error of the first prediction result is significantly larger than that of the second prediction result, the structure of the first prediction model is required to be adjusted or training parameters are required to be adjusted for further optimization, such as increasing a training data set, adopting regularization to avoid overfitting and other measures. If the two result errors are similar, abnormal points may exist in the data, abnormal data may be marked at this time, and the first prediction model is retrained. By comparing the second prediction result with the second prediction result as a reference, the effect of the first prediction model on the current situation can be clarified, the model optimization can be performed in a targeted manner, the model error is reduced, and the battery state prediction precision and the robustness of the BMS are improved.

On the basis of the above embodiments, as an alternative embodiment, the step of optimizing the prediction model in the BMS based on the first prediction result and the second prediction result may further include the steps of:

Step 301: and acquiring an actual operation result of the battery equipment under a predicted event after the preset time.

Step 302: a first deviation value of the first predicted result and the operational result and a second deviation value of the second predicted result and the operational result are determined.

Specifically, in order to evaluate the effects of different prediction models, it is necessary to obtain a deviation between the model prediction result and the actual operation result of the battery. After the prediction of the two models is performed, the system can continuously detect actual operation data of the battery equipment in a preset time period, such as parameters of voltage, current, temperature and the like, and the operation parameters are used as actual operation results under a prediction event. And then comparing the data in the operation result and the first prediction result, calculating a first deviation value, simultaneously calculating a second deviation value between the second prediction result and the operation result, and intuitively reflecting the prediction effect of each prediction model by calculating the deviation value between the model prediction result and the actual result.

Step 303: if the first deviation value is smaller than the second deviation value, marking the first prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the first prediction model reach the preset times, taking the first prediction model as a standard prediction model corresponding to the prediction event and training and optimizing the second prediction model.

Specifically, in order to continuously optimize the prediction capability of the BMS, it is necessary to determine an optimal standard prediction model according to different models and prediction effects of actual situations. The system counts the comparison condition of the deviation values of the prediction results of each prediction model, for example, after comparison, the first deviation value is smaller than the second deviation value, so that the first prediction model is more accurate in correspondence, the first prediction model is marked as a target prediction model corresponding to the prediction event, and when the number of marking times of the first model, which is superior to that of the second model, reaches a preset threshold value, for example, exceeds 5 times, the first model is confirmed to be a standard prediction model corresponding to the current prediction event. The comparison of the first deviation value and the second deviation value can directly judge which model has more accurate and reliable prediction results under the condition of a certain detected prediction abnormality, and if the deviation of the first prediction model is continuously smaller, the first prediction model is fully proved to be more suitable for the prediction events. So that in the following practical application, when similar prediction events are detected again, the system directly calls the standard prediction model to predict. Meanwhile, the first prediction model is used as an optimization training target, more relevant data are adopted to conduct targeted optimization on the second prediction model, the prediction effect of the standard model is gradually approximated, meanwhile, the feedback period of model training is shortened, the system can adapt to continuously changing battery states more rapidly, and the accuracy of the model is improved. The adaptability of the BMS to specific prediction scenes can be continuously improved, so that the prediction result is more accurate and reliable.

Step 304: and if the second deviation value is smaller than the first deviation value, marking the second prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the second prediction model reach the preset times, taking the second prediction model as a standard prediction model corresponding to the prediction event and carrying out training optimization on the first prediction model.

Specifically, if the second deviation value is smaller than the first deviation value, the training and optimization of the first prediction model may also be implemented by referring to the process of step 303, which is not described herein.

Step 40: if the BMS detects a decision event, second target data corresponding to the decision event is obtained, a first decision scheme actually executed on the battery equipment is determined, a first decision result of executing the first decision scheme based on the second target data is obtained, and a second decision result of a second decision scheme virtually executed in a shadow mode based on the second target data is obtained.

When the BMS detects a decision event, in order to evaluate the effect of different decision schemes, the result of the decision scheme executed on the real system, that is, the first decision result, needs to be obtained, and meanwhile, different decision schemes are operated in the shadow system to obtain the second decision result so as to perform comparison analysis. The management effect of the two decision schemes on the same battery is directly compared, which decision strategy is more reasonable can be judged, and therefore the existing decision mechanism is optimized pertinently.

In particular, the BMS needs to determine a first decision scheme, such as a rule-based charging policy, currently running on the battery device in order to locate the root cause of a problem after detecting a decision event, i.e., when the battery device has a certain fault. And then the system screens out the data at the moment of abnormality from the stored multi-mode data as second target data, wherein the second target data may comprise data such as battery voltage, SOC and the like. The first decision scheme is actually operated on the battery equipment, the second target data is taken as input, decision output for controlling the battery is generated, for example, a charging instruction with a certain current is output, such as battery pressure change, and the changed battery pressure is taken as a first decision result, such as the health state, voltage change and the like of the battery.

In order to evaluate the effect of different decision schemes under the current condition, other decision schemes are required to be called for comparison in addition to the first decision scheme, so as to obtain a second decision result. The system presets a decision scheme library of various decision schemes, including different decision strategies based on rules, reinforcement learning and the like. When a decision event is detected, the system selects at least one alternative decision scheme related to the decision event from the decision scheme library as a second decision scheme. Taking the current battery charging decision as an example, the system may call another charging decision scheme from the decision library, for example, an adaptive charging strategy based on Q learning optimization, as a second decision scheme. Then, the second target data that has been acquired is input, and the second decision scheme is simulated to be executed in the shadow BMS, i.e., the charging according to the policy is simulated, and the data result of the virtual execution of the second decision scheme is taken as a second decision result. By comparing the effects of the two decision schemes on the same group of input data, which decision strategy is more suitable for the current battery state can be judged, so that whether the first decision scheme needs to be optimally adjusted or not is determined, and the decision effect of the BMS is improved.

Step 50: and optimizing the first decision scheme or the second decision scheme based on the first decision result and the second decision result.

After the execution results of the first decision scheme and the second decision scheme are obtained, the decision scheme in the optimized BMS can be determined according to the effect comparison situation of the first decision scheme and the second decision scheme, and the decision capability of the system is improved. The first decision result may reflect problems and limitations of the decision strategy used in the current BMS, which may have decision deviation or abnormality, and the second decision result may be regarded as a comparison reference, which may determine whether the decision of the first decision scheme is reasonable.

Specifically, the system can count and compare the difference of the effects of the battery state of health change, battery life loss and the like caused by the two decision results to judge the defects of the first decision scheme, such as improper treatment of a certain situation and the like. If the first result is significantly worse than the second result, the first decision scheme needs to be adjusted, its decision log or rule is modified, and the decision result is optimized for that situation. If the two results are similar, the data is further analyzed, the decision basis is optimized, and beneficial decision factors in the second scheme can be added into the decision model. By comparing the effect of the second decision scheme with that of the second decision scheme as a reference, the decision mechanism of the BMS can be continuously perfected, so that the BMS is more widely suitable for various battery health state conditions, and the management effect of the battery is optimized.

On the basis of the above embodiments, as an alternative embodiment, the step of optimizing the decision scheme in the BMS based on the first decision result and the second decision result may further include the steps of:

Step 501: and comparing the first decision result with the second decision result to obtain a comparison result.

Specifically, result data generated by the first decision scheme and the second decision scheme under the same second target data input is obtained. The first decision result may include battery health, voltage curve, etc., and the second decision result may also include corresponding battery status data. Then, the system quantitatively or qualitatively compares the two decision results to generate a comparison analysis report. For example, the influence of two decision results on the battery health state is compared, the health degree is reduced by 10% by the first scheme, and the health degree is reduced by 5% by the second scheme, or the sorting conclusion of the two advantages and disadvantages is directly given, so that the comparison result can be intuitively displayed in the current scene, which decision scheme can bring better battery management effect, and a reference is provided for the follow-up determination of the optimization target.

Step 502: if the comparison result is that the first decision result is better than the second decision result, marking the first decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the first decision scheme reach the preset times, taking the first decision scheme as a standard decision scheme corresponding to the decision event and optimizing the second decision scheme.

Specifically, in order to continuously optimize the decision capability of the BMS, it is necessary to determine a standard decision scheme with the best effect in the current decision scene according to the comparison results of different decision schemes. The system can count the comparison situation of two decision results each time, for example, after comparison, the first decision scheme is better than the second decision scheme, the management capability of the battery equipment corresponding to the first decision scheme is better, the first decision scheme is marked as a target decision scheme corresponding to a decision event, when the marking times of the first decision scheme which is better than the second decision scheme reach a preset threshold value, for example, more than 5 times, the first decision scheme is confirmed to be a standard decision scheme under the decision abnormal event. When similar decision events are detected in the actual application, the system directly calls the standard decision scheme for decision management. Meanwhile, the system uses the first decision scheme as an optimization target, uses more relevant data training to correct the second decision scheme, and gradually approximates the effect of the first decision scheme. Therefore, the adaptability of the BMS to specific decision scenes can be continuously improved, so that the decision result of the BMS is more in line with the actual requirement of the current battery, and better battery management is realized.

Step 503: if the comparison result is that the second decision result is better than the first decision result, marking the second decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the second decision scheme reach the preset times, taking the second decision scheme as a standard decision scheme corresponding to the decision event and optimizing the first decision scheme.

Specifically, if the comparison result is that the second decision result is better than the first decision result, the training and optimization of the first decision scheme may be implemented by referring to the process of step 502, which is not described herein.

Referring to fig. 2, a schematic diagram of a BMS optimization system module based on a shadow mode according to an embodiment of the present application may include: the system comprises a shadow mode triggering module, a prediction model running module, a prediction model optimizing module, a decision scheme executing module and a decision scheme optimizing module, wherein:

Optionally, the shadow mode triggering module is further configured to calculate, in real time, a deviation value between the first prediction result and the second prediction result, where the first prediction model and the second prediction model are real-time running states; when the deviation value is larger than a preset value, determining that the BMS detects a predicted event; and acquiring each piece of operation data of the battery equipment, and determining that the BMS detects a decision event when detecting that any piece of operation data exceeds a corresponding preset range.

Optionally, the prediction model operation module is further configured to obtain, in real time, multi-mode data collected by the BMS during the BMS detection process; and if the BMS detects the predicted event, screening out first target data corresponding to the predicted event from the multi-mode data.

Optionally, the BMS optimization system based on the shadow mode further includes a data uploading module, where the data uploading module is configured to classify the first target data based on a clustering algorithm, obtain classified first target data, and upload the classified first target data to the cloud.

Optionally, the prediction model operation module is further configured to invoke at least one second prediction model corresponding to the prediction event in a preset prediction model library; and inputting the first target data into the second prediction model to obtain a second prediction result of virtually running the second prediction model in the shadow mode.

Optionally, the prediction model optimization module is further configured to obtain an actual operation result of the battery device under the prediction event after a preset time; calculating a first deviation value of the first prediction result and the operation result, and a second deviation value of the second prediction result and the operation result; if the first deviation value is smaller than the second deviation value, marking the first prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the first prediction model reach preset times, taking the first prediction model as a standard prediction model corresponding to the prediction event and training and optimizing the second prediction model; and if the second deviation value is smaller than the first deviation value, marking the second prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the second prediction model reach the preset times, taking the second prediction model as a standard prediction model corresponding to the prediction event and training and optimizing the first prediction model.

Optionally, the decision scheme optimization module is further configured to compare the first decision result with the second decision result to obtain a comparison result; if the comparison result is that the first decision result is better than the second decision result, marking the first decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the first decision scheme reach the preset times, taking the first decision scheme as a standard decision scheme corresponding to the decision event and optimizing the second decision scheme; if the comparison result is that the second decision result is better than the first decision result, marking the second decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the second decision scheme reach the preset times, taking the second decision scheme as a standard decision scheme corresponding to the decision event and optimizing the first decision scheme.

It should be noted that: in the system provided in the above embodiment, when implementing the functions thereof, only the division of the above functional modules is used as an example, in practical application, the above functional allocation may be implemented by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules, so as to implement all or part of the functions described above. In addition, the system and method embodiments provided in the foregoing embodiments belong to the same concept, and specific implementation processes of the system and method embodiments are detailed in the method embodiments, which are not repeated herein.

The embodiment of the present application further provides a computer storage medium, where the computer storage medium may store a plurality of instructions, where the instructions are suitable for being loaded by a processor and executed by the processor, and a specific execution process may refer to a specific description of the foregoing embodiment, and details are not repeated herein.

Referring to fig. 3, the application also discloses an electronic device. Fig. 3 is a schematic structural diagram of an electronic device according to the disclosure. The electronic device 300 may include: at least one processor 301, at least one network interface 304, a user interface 303, a memory 305, at least one communication bus 302.

Wherein the communication bus 302 is used to enable connected communication between these components.

The user interface 303 may include a Display screen (Display), a Camera (Camera), and the optional user interface 303 may further include a standard wired interface, and a wireless interface.

The network interface 304 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface), among others.

Wherein the processor 301 may include one or more processing cores. The processor 301 utilizes various interfaces and lines to connect various portions of the overall server, perform various functions of the server and process data by executing or executing instructions, programs, code sets, or instruction sets stored in the memory 305, and invoking data stored in the memory 305. Alternatively, the processor 301 may be implemented in at least one hardware form of digital signal Processing (DIGITAL SIGNAL Processing, DSP), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA), programmable logic array (Programmable Logic Array, PLA). The processor 301 may integrate one or a combination of several of a central processing unit (Central Processing Unit, CPU), an image processor (Graphics Processing Unit, GPU), and a modem, etc. The CPU mainly processes an operating system, a user interface, an application program and the like; the GPU is used for rendering and drawing the content required to be displayed by the display screen; the modem is used to handle wireless communications. It will be appreciated that the modem may not be integrated into the processor 301 and may be implemented by a single chip.

The Memory 305 may include a random access Memory (Random Access Memory, RAM) or a Read-Only Memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer readable medium (non-transitory computer-readable storage medium). Memory 305 may be used to store instructions, programs, code, sets of codes, or sets of instructions. The memory 305 may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-described respective method embodiments, etc.; the storage data area may store data or the like involved in the above respective method embodiments. Memory 305 may also optionally be at least one storage device located remotely from the aforementioned processor 301. Referring to fig. 3, an operating system, a network communication module, a user interface module, and an application program of a BMS optimization method based on a shadow pattern may be included in the memory 305 as a computer storage medium.

In the electronic device 300 shown in fig. 3, the user interface 303 is mainly used for providing an input interface for a user, and acquiring data input by the user; and the processor 301 may be configured to invoke an application program in the memory 305 that stores a shadow-mode-based BMS optimization method that, when executed by the one or more processors 301, causes the electronic device 300 to perform the method as described in one or more of the embodiments above. It should be noted that, for simplicity of description, the foregoing method embodiments are all described as a series of acts, but it should be understood by those skilled in the art that the present application is not limited by the order of acts described, as some steps may be performed in other orders or concurrently in accordance with the present application. Further, those skilled in the art will also appreciate that the embodiments described in the specification are all of the preferred embodiments, and that the acts and modules referred to are not necessarily required for the present application.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.

In the several embodiments provided by the present application, it should be understood that the disclosed apparatus may be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative, such as a division of units, merely a division of logic functions, and there may be additional divisions in actual implementation, such as multiple units or components may be combined or integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some service interface, device or unit indirect coupling or communication connection, electrical or otherwise.

The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable memory. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in whole or in part in the form of a software product stored in a memory, comprising several instructions for causing a computer device (which may be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method of the various embodiments of the present application. And the aforementioned memory includes: various media capable of storing program codes, such as a U disk, a mobile hard disk, a magnetic disk or an optical disk.

The foregoing is merely exemplary embodiments of the present disclosure and is not intended to limit the scope of the present disclosure. That is, equivalent changes and modifications are contemplated by the teachings of this disclosure, which fall within the scope of the present disclosure. Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure.

This application is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a scope and spirit of the disclosure being indicated by the claims.

Claims

1. A method for BMS optimization based on shadow patterns, the method comprising:

Calculating a deviation value between a first prediction result and a second prediction result in real time, wherein the first prediction model and the second prediction model are in a real-time running state; when the deviation value is larger than a preset value, determining that the BMS detects a predicted event; acquiring each operation data of battery equipment, and determining that the BMS detects a decision event when detecting that any operation data exceeds a corresponding preset range;

If the BMS detects the predicted event, acquiring the multi-mode data acquired by the BMS in real time in the BMS detection process; screening first target data corresponding to the predicted event from the multi-mode data, determining a first predicted model actually operated on battery equipment managed by the BMS, acquiring a first predicted result of the first predicted model for predicting based on the first target data, and simultaneously calling at least one second predicted model corresponding to the predicted event from a preset predicted model library; inputting the first target data into the second prediction model to obtain a second prediction result of virtually running the second prediction model in the shadow mode;

2. The BMS optimization method based on shadow mode according to claim 1, wherein after the first target data corresponding to the predicted event is screened out from the multi-mode data, further comprising:

classifying the first target data based on a clustering algorithm to obtain classified first target data, and uploading the classified first target data to a cloud.

3. The shadow-mode-based BMS optimization method of claim 1, wherein optimizing the first prediction model or the second prediction model based on the first prediction result and the second prediction result comprises:

Acquiring an actual operation result of the battery equipment under the predicted event after a preset time;

Calculating a first deviation value of the first prediction result and the operation result, and a second deviation value of the second prediction result and the operation result;

If the first deviation value is smaller than the second deviation value, marking the first prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the first prediction model reach preset times, taking the first prediction model as a standard prediction model corresponding to the prediction event and training and optimizing the second prediction model;

and if the second deviation value is smaller than the first deviation value, marking the second prediction model as a target prediction model corresponding to the prediction event, and when the marking times of the second prediction model reach the preset times, taking the second prediction model as a standard prediction model corresponding to the prediction event and training and optimizing the first prediction model.

4. The shadow pattern based BMS optimization method of claim 1, wherein the optimizing the first decision scheme or the second decision scheme based on the first decision result and the second decision result comprises:

Comparing the first decision result with the second decision result to obtain a comparison result;

If the comparison result is that the first decision result is better than the second decision result, marking the first decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the first decision scheme reach the preset times, taking the first decision scheme as a standard decision scheme corresponding to the decision event and optimizing the second decision scheme;

If the comparison result is that the second decision result is better than the first decision result, marking the second decision scheme as a target decision scheme corresponding to the decision event, and when the marking times of the second decision scheme reach the preset times, taking the second decision scheme as a standard decision scheme corresponding to the decision event and optimizing the first decision scheme.

5. A BMS optimization system based on shadow patterns, the system comprising:

the shadow mode triggering module is used for calculating a deviation value between a first prediction result and a second prediction result in real time, and the first prediction model and the second prediction model are in a real-time running state; when the deviation value is larger than a preset value, determining that the BMS detects a predicted event; acquiring each operation data of battery equipment, and determining that the BMS detects a decision event when detecting that any operation data exceeds a corresponding preset range; triggering a shadow mode corresponding to the BMS when the BMS detects a predicted event or a decision event;

The prediction model operation module is used for acquiring the multi-mode data acquired by the BMS in real time in the BMS detection process if the BMS detects the prediction event; screening first target data corresponding to the predicted event from the multi-mode data, determining a first predicted model actually operated on battery equipment managed by the BMS, acquiring a first predicted result of the first predicted model for predicting based on the first target data, and simultaneously calling at least one second predicted model corresponding to the predicted event from a preset predicted model library; inputting the first target data into the second prediction model to obtain a second prediction result of virtually running the second prediction model in the shadow mode;

6. A computer readable storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the method of any one of claims 1-4.

7. An electronic device comprising a processor, a memory, a user interface, and a network interface, the memory for storing instructions, the user interface and the network interface for communicating to other devices, the processor for executing the instructions stored in the memory to cause the electronic device to perform the method of any of claims 1-4.