CN111142026B - Data processing method and device and electronic equipment - Google Patents
Data processing method and device and electronic equipment Download PDFInfo
- Publication number
- CN111142026B CN111142026B CN201911418635.2A CN201911418635A CN111142026B CN 111142026 B CN111142026 B CN 111142026B CN 201911418635 A CN201911418635 A CN 201911418635A CN 111142026 B CN111142026 B CN 111142026B
- Authority
- CN
- China
- Prior art keywords
- parameter
- data
- model
- battery
- health
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/392—Determining battery ageing or deterioration, e.g. state of health
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Secondary Cells (AREA)
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
Abstract
The application discloses a data processing method, a data processing device and electronic equipment.A health degree prediction model capable of predicting the health degree of a battery is trained by using a plurality of operation parameters such as environmental information, parking information, charging information and the like in a plurality of groups of operation data of the battery, and then a scoring model capable of predicting the influence degree of each operation parameter on the health degree of the battery is trained by using model parameters in a well-trained health degree prediction model as samples.
Description
Technical Field
The present application relates to the field of battery maintenance technologies, and in particular, to a data processing method and apparatus, and an electronic device.
Background
In order to save energy and protect environment, batteries have become a core new energy source. In practical applications, various factors affect the battery health, for example, the battery health is reduced due to too many charging and discharging times.
In order to alleviate the decrease of the battery health degree, it is important how to evaluate the influence degree of various environmental factors on the battery health degree.
Disclosure of Invention
In view of the above, the present application provides a data processing method, an apparatus and an electronic device, as follows:
a method of data processing, comprising:
obtaining training data of a battery, wherein the training data comprises at least two groups of operation data, and the operation data comprises parameter values of at least two operation parameters and corresponding battery health values;
taking the operation data in the training data as a training sample of a pre-constructed health degree prediction model, and training the health degree prediction model;
obtaining a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, wherein the parameter vector comprises: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model;
taking a parameter vector of a first operating parameter as input of a pre-constructed scoring model, taking a score value of the first operating parameter as output of the scoring model, and training the scoring model to obtain a trained scoring model, wherein the first operating parameter is an operating parameter with a score value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
In the method, preferably, the health degree prediction model is a model constructed based on an artificial neural network, and the health degree prediction model includes at least one hidden layer, and the hidden layer includes a plurality of neural network operation units;
wherein the parameter vector comprises: and the plurality of neural network operation units of the first layer of the hidden layer in the health degree prediction model correspond to the parameter weight values of the operating parameters.
Preferably, the above method, obtaining a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, includes:
obtaining model parameters of the health degree prediction model, wherein the model parameters comprise parameter weight values corresponding to the operation parameters of a plurality of neural network operation units in the at least one hidden layer;
obtaining the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters of the first layer of the hidden layer from the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters in the at least one layer of hidden layer to form the parameter vector of the operation parameters in the operation data.
Preferably, the method for training the health degree prediction model by using the operation data in the training data as a training sample of a pre-constructed health degree prediction model includes:
preprocessing the operating parameters in the operating data in the training data; the pretreatment comprises the following steps: any one or combination of any multiple of data normalization, data missing supplement, data redundancy removal, data noise reduction and data abnormal deletion;
and taking the operation parameters in the operation data in the training data as the input of a pre-constructed health degree prediction model, taking the battery health degree value in the operation data as the output of the health degree prediction model, and training the health degree prediction model to obtain a trained health degree prediction model.
The above method, preferably, further comprises:
obtaining verification data of the battery, wherein the verification data comprises at least two groups of operation data, and the operation data comprises parameter values of at least two operation parameters and corresponding battery health values; the validation data is different from the training data;
inputting the operation data in the verification data into the health degree prediction model to obtain a prediction result output by the health degree prediction model;
and modifying the model parameters according to the prediction result.
Preferably, the method for modifying the model parameter according to the prediction result includes:
comparing the prediction result with the battery health value in the operation data in the verification data to obtain a comparison result;
and modifying the model parameters according to the comparison result.
The above method, preferably, further comprises:
and optimizing the model parameters of the trained health degree prediction model by using a preset optimization algorithm.
The above method, preferably, further comprises:
and constructing the scoring model by using a preset regression algorithm.
A data processing apparatus comprising:
the battery management system comprises a data obtaining unit, a battery management unit and a battery management unit, wherein the data obtaining unit is used for obtaining training data of a battery, the training data comprises at least two groups of operation data, and the operation data comprises parameter values of at least two operation parameters and corresponding battery health values;
the first training unit is used for taking operation data in the training data as a training sample of a pre-constructed health degree prediction model and training the health degree prediction model;
a vector obtaining unit, configured to obtain a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, where the parameter vector includes: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model;
the second training unit is used for taking a parameter vector of a first operating parameter as the input of a pre-constructed scoring model, taking a scoring value of the first operating parameter as the output of the scoring model, and training the scoring model to obtain a trained scoring model, wherein the first operating parameter is an operating parameter with a scoring value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
An electronic device, comprising:
the memory is used for storing the application program and data generated by the running of the application program;
a processor for executing the application to implement: obtaining training data of a battery, wherein the training data comprises at least two groups of operation data, and the operation data comprises parameter values of at least two operation parameters and corresponding battery health values; taking the operation data in the training data as a training sample of a pre-constructed health degree prediction model, and training the health degree prediction model; obtaining a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, wherein the parameter vector comprises: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model; taking a parameter vector of a first operating parameter as input of a pre-constructed scoring model, taking a score value of the first operating parameter as output of the scoring model, and training the scoring model to obtain a trained scoring model, wherein the first operating parameter is an operating parameter with a score value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
According to the technical scheme, the health degree prediction model capable of predicting the health degree of the battery is trained by using a plurality of operation parameters such as environmental information, parking information, charging information and the like in a plurality of groups of operation data of the battery, and then the rating model capable of predicting the influence degree of each operation parameter on the health degree of the battery is trained by using the model parameters in the well-trained health degree prediction model as samples, so that the influence degree of the operation parameters of the battery on various environmental factors on the health degree of the battery can be obtained by using the trained rating model, and the purpose of the application is achieved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a flowchart of a data processing method according to an embodiment of the present application;
FIG. 2 is a diagram illustrating hidden layers in a health prediction model according to an embodiment of the present disclosure;
FIG. 3 is a partial flow chart of a first embodiment of the present application;
fig. 4 is a schematic structural diagram of a data processing apparatus according to a second embodiment of the present application;
fig. 5 is a schematic structural diagram of an electronic device according to a third embodiment of the present application;
FIG. 6 is a diagram illustrating an exemplary process for a battery in a server according to the present application;
fig. 7 is a schematic diagram of a neural network model in an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Referring to fig. 1, a flowchart of an implementation of a data processing method provided in an embodiment of the present application is applied to an electronic device capable of performing data processing, such as a computer or a server. The method in the embodiment is mainly used for obtaining the influence degree of each operating parameter of the battery on the health degree of the battery so as to provide the influence degree for the battery user as reference data in the battery using process and slow down the reduction rate of the health degree of the battery in the battery using process.
Specifically, the method in this embodiment may include the following steps:
step 101: training data for the battery is obtained.
The training data comprises at least two groups of operation data, and each group of operation data comprises parameter values of at least two operation parameters and corresponding battery health values. Wherein, the operation parameters in each group of operation data may be: environmental information, parking information, automobile driving information, charging information, discharging information and other parameters. The environmental information can be understood as information such as weather and whether traffic is congested in the battery operating environment; the parking information may include: any one or more of information such as parking place (whether flat), parking mode and parking duration; the vehicle driving information may include: any one or more of information such as driving time length, driving speed and acceleration of the automobile in which the battery is located; the charging information and the discharging information may include: the number of charging and discharging times, the charging and discharging duration and the like.
As shown in table 1, the training data includes m sets of operating data of one or more batteries, m is a positive integer greater than or equal to 2, each set of operating data includes n operating parameters, n is a positive integer greater than or equal to 2, each set of operating data represents n operating parameters by X1, X2,. eta., Xn, etc., and each set of operating data represents a battery health value by Y.
TABLE 1 operating data
| X1 | X2 | ... | Xn | Y |
| x1_1 | X2_1 | ... | Xn_1 | Y_1 |
| X1_2 | X2_2 | ... | Xn_2 | Y_2 |
| X1_3 | X2_3 | ... | Xn_3 | Y_3 |
| X1_4 | X2_4 | ... | Xn_4 | Y_4 |
| ... | ... | ... | ... | ... |
| X1_m | X2_m | ... | Xn_m | Y_m |
Wherein, X1, X2, Xn and the like are factors influencing the battery health value, and the battery health value Y is calculated by other variables.
It should be noted that, in this embodiment, each operating parameter in the operating data is a parameter type determined after parameter screening, for example, it is determined according to experience or historical data which operating parameters are parameters related to the battery health value and which are parameters that do not affect the change of the battery health value, so that in obtaining the training data, the operating parameters in the obtained operating data are all parameters related to the battery health value, that is, parameters that can affect the change of the battery health value in a positive or negative direction.
Step 102: and taking the operation data in the training data as a training sample of a pre-constructed health degree prediction model to train the health degree prediction model.
The health degree prediction model can be constructed in advance through algorithms such as a neural network, and then after training data containing the health degree value of the battery are obtained, the health degree prediction model is trained through each group of operation data in the training data, and then the well-trained health degree prediction model is obtained.
Step 103: and obtaining a parameter vector of the operation parameter in the operation data according to the model parameter of the health degree prediction model.
Wherein the parameter vector includes: and a plurality of operation units in the health degree prediction model and parameter weight values corresponding to the operation parameters.
Specifically, since the model parameters in the health degree prediction model can represent the influence relationship between each operating parameter and the battery health degree value, in this embodiment, a parameter vector for each operating parameter is obtained according to the model parameters, the parameter vector includes a parameter weight value corresponding to the operating parameter in each operation unit, and each parameter weight value represents the contribution or influence degree that the corresponding operating parameter in the corresponding operation unit can make to the change of the battery health degree value when the battery health degree value is predicted.
Step 104: and taking the parameter vector of the first operating parameter as the input of a pre-constructed scoring model, taking the scoring value of the first operating parameter as the output of the scoring model, and training the scoring model to obtain the trained scoring model.
The first operation parameter as the training input is an operation parameter with a score value in at least two operation parameters in the operation data, and the score value of the first operation parameter represents the influence degree of the first operation parameter on the health value of the battery. That is, the first operating parameter in this embodiment is an operating parameter with a known value of credit, and the value of credit of the first operating parameter may be set manually by a user or calculated according to a certain algorithm, and the value of credit of the first operating parameter is considered to be accurate.
Accordingly, the scoring model trained in this embodiment can be used to output the scoring result of the second operating parameter with the parameter vector of the second operating parameter as input, where the second operating parameter is an operating parameter without a scoring value, i.e. the second operating parameter is an operating parameter of unknown score value, and the correspondingly obtained scoring result characterizes the degree of influence of the second operating parameter on the health value of said battery, it can be seen that in the present embodiment, the scoring model trained by the first operating parameter with known score is used to perform the prediction process on the second operating parameter with unknown score, to derive a score for the second operating parameter for the unknown degree of influence on the battery health value, to characterize the degree of influence of the second operating parameter on the battery health value, such as the extent to which the second operating parameter can contribute to or affect the change in the battery health value.
The scoring model in this embodiment may be created in advance based on a regression algorithm such as a decision tree or a random forest, and then model training is performed.
According to the above scheme, in the data processing method according to the embodiment of the present application, the health degree prediction model capable of predicting the health degree of the battery is trained by using the plurality of operation parameters, such as environmental information, parking information, charging information, and the like, in the plurality of sets of operation data of the battery, and then the score model capable of predicting the influence degree of each operation parameter on the health degree value of the battery is trained by using the model parameters in the well-trained health degree prediction model as samples, so that the influence degree of the operation parameters of the battery on various environmental factors on the health degree of the battery can be obtained by using the trained score model, and the purpose of the embodiment is achieved.
In one implementation, the health degree prediction model may be a model constructed based on an artificial neural network, and the constructed health degree prediction model includes at least one hidden layer, and the hidden layer includes a plurality of neural network operation units, as shown in fig. 2, the health degree prediction model may include at least two neural network hidden layers, each of the neural network hidden layers includes a plurality of neural network operation units, and each of the neural network operation units may also be referred to as a neuron or a neural network unit, and the like.
Accordingly, in this embodiment, the parameter vector obtained from the model parameter of the health degree prediction model includes: and a plurality of neural network operation units of a first layer of hidden layer in the health degree prediction model and parameter weight values corresponding to the operation parameters.
Specifically, in step 103 of this embodiment, when obtaining the parameter vector of the operating parameter in the operating data according to the model parameter of the health degree prediction model, the following method may be implemented:
first, model parameters of a health degree prediction model are obtained.
The model parameters comprise at least one layer of hidden layer, such as parameter weight values corresponding to a plurality of neural network operation units and operation parameters in each layer of hidden layer;
and then, obtaining the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters of the first layer of hidden layer from the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters in the hidden layers to form a parameter vector of the operation parameters in the operation data.
For example, the operation parameters, which are potential factors affecting the battery health degree, have n items, and there are k neural network operation units in the first hidden layer, then the parameter weight values in the parameter vectors corresponding to the n items of operation parameters form a weight matrix shown in table 2:
TABLE 2 weight matrix
| X1 | X2 | ... | Xn |
| W1_1 | W1_2 | ... | W1_n |
| W2_1 | W2_2 | ... | W2_n |
| ... | ... | ... | ... |
| Wi_1 | Wi_2 | ... | Wi_n |
| ... | ... | ... | ... |
| Wk_1 | Wk_2 | ... | Wk_n |
Wherein, X1, X2, the.. times, Xn are operation parameters (factors having potential influence on the battery health value), the i-th neural network arithmetic unit in the k neural network arithmetic units represents Wi _1, Wi _2, the.. times, Wi _ n for the n operation parameters, and then can be vectorized and represented as W1_1, W2_1, the.. times, Wk _1 for the influence factor X1.
In one implementation manner, when the health degree prediction model is trained by using the operation data in the training data as the training sample of the pre-constructed health degree prediction model in step 102, the method may specifically be implemented by:
firstly, preprocessing the operation parameters in the operation data in the training data in the embodiment; for example, the operation parameters are subjected to any one or combination of any multiple of preprocessing such as data normalization, data missing supplement, data redundancy removal, data noise reduction, data abnormal deletion and the like;
and then, taking the operation parameters in the operation data in the training data as the input of a pre-constructed health degree prediction model, taking the battery health degree value in the corresponding operation data as the output of the health degree prediction model, and training the health degree prediction model to obtain a trained health degree prediction model, wherein the model parameters in the trained health degree prediction model are optimized model parameters after training, so that the health degree prediction of the battery with unknown health degree value can be accurately carried out according to the operation parameters in the operation data of the battery.
Based on the realization, in the process of training the health degree prediction model, the model parameters of the trained health degree prediction model can be optimized by using a preset optimization algorithm such as a k-fold cross validation algorithm and the like, so that the optimized health degree prediction model can accurately predict the health degree value of the battery.
In one implementation, after training the fitness prediction model in step 102, the present embodiment may further include the following steps, as shown in fig. 3:
step 105: authentication data of the battery is obtained.
The verification data of the battery comprises at least two groups of operation data, the operation data in the verification data also comprises parameter values of at least two operation parameters and corresponding battery health values, and it needs to be stated that the verification data is different from the training data.
Step 106: and inputting the operation data in the verification data into the health degree prediction model to obtain a prediction result output by the health degree prediction model.
The prediction result comprises a prediction health value obtained by predicting operation data in the verification data.
Step 107: and modifying the model parameters according to the prediction result.
For example, in this embodiment, the predicted result is compared with the battery health value in the operation data in the verification data to obtain a comparison result, that is, the predicted health value is compared with the actual battery health value in the operation data, and then the model parameter is modified according to the comparison result, for example, if the comparison result indicates that the predicted health value is the same as the actual battery health value, it indicates that the model parameter is accurate, and the battery health value can be accurately predicted, and if the comparison result indicates that the predicted health value is different from the actual battery health value, it indicates that the model parameter has a parameter error, at this time, all or part of the parameters in the model parameter may be adjusted until the health prediction model with the adjusted parameters can accurately predict the battery health value.
Referring to fig. 4, a schematic structural diagram of a data processing apparatus provided in the second embodiment of the present application is shown, where the apparatus may be configured in an electronic device capable of performing data processing, such as a computer or a server. The device in the embodiment is mainly used for obtaining the influence degree of each operating parameter of the battery on the health degree of the battery so as to provide the influence degree for a battery user as reference data in the battery using process and slow down the reduction rate of the health degree of the battery in the battery using process.
Specifically, the apparatus in this embodiment may include the following units:
a data obtaining unit 401, configured to obtain training data of a battery, where the training data includes at least two sets of operation data, and the operation data includes parameter values of at least two operation parameters and corresponding battery health values;
a first training unit 402, configured to train the health degree prediction model by using operation data in the training data as a training sample of a pre-constructed health degree prediction model;
a vector obtaining unit 403, configured to obtain a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, where the parameter vector includes: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model;
a second training unit 404, configured to train a scoring model to obtain a trained scoring model by using a parameter vector of a first operating parameter as an input of a pre-constructed scoring model and using a scoring value of the first operating parameter as an output of the scoring model, where the first operating parameter is an operating parameter with a scoring value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
As can be seen from the foregoing solution, in the second data processing apparatus according to the embodiment of the present application, a health degree prediction model capable of predicting the health degree of a battery is trained according to a plurality of operation parameters, such as environmental information, parking information, and charging information, in a plurality of sets of operation data of the battery, and then a scoring model capable of predicting the degree of influence of each operation parameter on the health degree value of the battery is trained by using a model parameter in the well-trained health degree prediction model as a sample, so that the degree of influence of the operation parameter of the battery on various environmental factors on the health degree of the battery can be obtained by using the trained scoring model, thereby achieving the purpose of the present embodiment.
In one implementation manner, the health degree prediction model is a model constructed based on an artificial neural network, and the health degree prediction model includes at least one hidden layer, and the hidden layer includes a plurality of neural network operation units;
wherein the parameter vector comprises: and the plurality of neural network operation units of the first layer of the hidden layer in the health degree prediction model correspond to the parameter weight values of the operating parameters.
Correspondingly, the vector obtaining unit 403 obtains the parameter vector of the operating parameter in the operating data according to the model parameter of the health degree prediction model, including:
obtaining model parameters of the health degree prediction model, wherein the model parameters comprise parameter weight values corresponding to the operation parameters of a plurality of neural network operation units in the at least one hidden layer;
obtaining the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters of the first layer of the hidden layer from the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters in the at least one layer of hidden layer to form the parameter vector of the operation parameters in the operation data.
In one implementation, the training unit 402 trains the health degree prediction model with operation data in the training data as a training sample of a pre-constructed health degree prediction model, including:
preprocessing the operating parameters in the operating data in the training data; the pretreatment comprises the following steps: any one or combination of any multiple of data normalization, data missing supplement, data redundancy removal, data noise reduction and data abnormal deletion;
and taking the operation parameters in the operation data in the training data as the input of a pre-constructed health degree prediction model, taking the battery health degree value in the operation data as the output of the health degree prediction model, and training the health degree prediction model to obtain a trained health degree prediction model.
In one implementation, the first training unit 402 is further configured to:
obtaining verification data of the battery, wherein the verification data comprises at least two groups of operation data, and the operation data comprises parameter values of at least two operation parameters and corresponding battery health values; the validation data is different from the training data;
inputting the operation data in the verification data into the health degree prediction model to obtain a prediction result output by the health degree prediction model;
and modifying the model parameters according to the prediction result.
Optionally, the first training unit 402 modifies the model parameters according to the prediction result, including:
comparing the prediction result with the battery health value in the operation data in the verification data to obtain a comparison result;
and modifying the model parameters according to the comparison result.
In one implementation, the first training unit 402 is further configured to: and optimizing the model parameters of the trained health degree prediction model by using a preset optimization algorithm.
Optionally, the scoring model is constructed by using a preset regression algorithm.
It should be noted that, for the specific implementation of each unit in the present embodiment, reference may be made to the corresponding content in the foregoing, and details are not described here.
Referring to fig. 5, a schematic structural diagram of an electronic device according to a third embodiment of the present disclosure is provided, where the electronic device may be an electronic device capable of performing data processing, such as a computer or a server. The electronic device in this embodiment is mainly used to obtain the degree of influence of each operating parameter of the battery on the health degree of the battery, so as to provide the battery user with the degree of influence as reference data in the battery using process, and slow down the rate of decrease of the health degree of the battery in the battery using process.
Specifically, the electronic device in this embodiment may include the following structure:
a memory 501 for storing applications and data generated by the application operations;
a processor 502 for executing the application to implement: obtaining training data of a battery, wherein the training data comprises at least two groups of operation data, and the operation data comprises parameter values of at least two operation parameters and corresponding battery health values; taking the operation data in the training data as a training sample of a pre-constructed health degree prediction model, and training the health degree prediction model; obtaining a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, wherein the parameter vector comprises: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model; taking a parameter vector of a first operating parameter as input of a pre-constructed scoring model, taking a score value of the first operating parameter as output of the scoring model, and training the scoring model to obtain a trained scoring model, wherein the first operating parameter is an operating parameter with a score value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
According to the scheme, the electronic device in the third embodiment of the present application trains the health degree prediction model capable of predicting the health degree of the battery by using multiple operation parameters in multiple sets of operation data of the battery, such as environmental information, parking information, charging information, and the like, and further trains the scoring model capable of predicting the influence degree of each operation parameter on the health degree value of the battery by using the model parameters in the well-trained health degree prediction model as samples, so that the influence degree of the operation parameters of the battery on various environmental factors on the health degree of the battery can be obtained by using the trained scoring model, and the purpose of the present embodiment is achieved.
Taking an electronic device as an example as a server, referring to the flow shown in fig. 6, the steps of obtaining the influence degree of various factors on the health value of the battery are as follows:
step one, data acquisition and processing:
first, in the data acquisition stage, environment information, parking information, vehicle driving information, charging information, discharging information (represented by: X1, X2,.. multidot.xn, etc.), and battery health degree information (represented by Y) are mainly acquired from the existing information systems, wherein X1, X2,.. multidot.xn are mainly potential factors (operating parameters) affecting the battery health degree.
Then, in the data processing stage, the obtained data is mainly processed in the step, so that high-quality data is obtained. The method mainly comprises abnormal value processing, missing value processing and data standardization (normalization) to ensure the correctness and validity of the data.
Step two, influence factor representation:
1) neural network model construction
The construction of the neural network is completed in the step. By analyzing the characteristics of the acquired data, a proper neural network model is constructed, the basic structure is shown in fig. 7, the input layer is a neural network unit on the left side in the figure, the output layer is a battery health value (on the right side), and the hidden layer is generally formed by a plurality of layers of neural networks.
2) Model training
The step is mainly to train the neural network model constructed in the foregoing. First, the data acquired in step one is split, and the data is split into a training set (training data) and a verification set (verification data). And taking parameter values of potential influence factors such as environment information, parking information, automobile driving information, charging information and discharging information as the input of the model, and taking the battery health value as the output of the model to train the battery health prediction model. And then, searching and optimizing to obtain a better model parameter by using a K-fold cross validation method according to the characteristics of the constructed model, or adjusting the parameter by using an empirical setting method. And obtaining an optimized neural network model M (health degree prediction model) through multiple iterations.
3) Vectorized representation of influencing factors
The method mainly realizes vectorization representation of the influencing factors. For the health degree prediction model M of the battery obtained in 2), extracting a parameter weight matrix of a first layer of neural unit (neural network operation unit) in the model M, as shown in table 2.
Step three, generating a score:
(1) data annotation
According to the priori knowledge, marking scores on part of the influence factors in a manual marking mode, wherein marking results are shown in a table 3, and the score values represent the influence degrees of the corresponding factors on the battery health value:
TABLE 3 influencing factor Scoring notation
| Influencing factor | Value of credit |
| Factor_1 | 0.3 |
| Factor_2 | 0.1 |
| ... | ... |
| Factor_i | 0 |
| ... | ... |
| Factor_n | 0 |
Wherein Factor _ i is a vectorized representation of the i-th influencing Factor obtained through step two, i.e., W1_ i, W2_ i,. and Wk _ i, as shown in table 2. A score of 0 indicates that the influencing factor cannot be labeled.
(2) Scoring model construction
Constructing a training set according to the annotation data obtained in the step three (1), wherein the vectorization of the influence factors is represented as the input of the model, and the annotated score value is the output of the model; with respect to the model, it can be obtained by classical algorithm training, such as Lasso, decision tree, random forest, etc.; through training, a model P is obtained.
(3) Outputting the score
And (3) obtaining vectorization representation of the factors by the step two, and outputting scores of the influencing factors by the model obtained by the step three (2).
In summary, in a real situation, the types of factors affecting the battery health degree are complex, and the relationships between different factors and the battery health degree are different, for example, the relationships between some factors and the battery health degree are linear, and the relationships between some factors and the battery health degree are nonlinear, and the technical scheme of the application is expressed in a vectorization manner by the relationship between the model learning factors and the battery health degree, so that the fact is better met; compared with the traditional influence factor analysis, the artificial experience information is introduced through the artificial marking data, the fusion of data driving and industry through judgment is realized, and the reliability of the analysis result of the influence factor is improved.
In addition, the technical scheme of the application can be suitable for analyzing the complex factors and can also be tried in analysis work of influence factors in other industries.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A method of data processing, comprising:
obtaining training data of a battery, wherein the training data comprises at least two groups of operation data, the operation data comprises parameter values of at least two operation parameters and corresponding battery health values, and the operation parameters in the operation data comprise: environmental information, parking information, car information of traveling, charge information and discharge information, environmental information includes: weather information in the battery operating environment or/and whether traffic is congested; the parking information includes: any one or more of parking place, parking mode and parking duration; the automobile driving information includes: any one or more of driving time length, driving speed and acceleration of the automobile in which the battery is located; the charging information and the discharging information include: the number of times or/and the duration of charge and discharge;
taking the operation data in the training data as a training sample of a pre-constructed health degree prediction model, and training the health degree prediction model;
obtaining a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, wherein the parameter vector comprises: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model;
taking a parameter vector of a first operating parameter as input of a pre-constructed scoring model, taking a score value of the first operating parameter as output of the scoring model, and training the scoring model to obtain a trained scoring model, wherein the first operating parameter is an operating parameter with a score value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
2. The method according to claim 1, wherein the health prediction model is a model constructed based on an artificial neural network, and the health prediction model comprises at least one hidden layer, and the hidden layer comprises a plurality of neural network operation units;
wherein the parameter vector comprises: and the plurality of neural network operation units of the first layer of the hidden layer in the health degree prediction model correspond to the parameter weight values of the operating parameters.
3. The method of claim 2, obtaining a parameter vector for the operating parameter in the operating data from a model parameter of the health prediction model, comprising:
obtaining model parameters of the health degree prediction model, wherein the model parameters comprise parameter weight values corresponding to the operation parameters of a plurality of neural network operation units in the at least one hidden layer;
obtaining the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters of the first layer of the hidden layer from the parameter weight values corresponding to the plurality of neural network operation units and the operation parameters in the at least one layer of hidden layer to form the parameter vector of the operation parameters in the operation data.
4. The method according to claim 1 or 2, wherein the training of the health degree prediction model is performed by using the operation data in the training data as a training sample of a pre-constructed health degree prediction model, and comprises the following steps:
preprocessing the operating parameters in the operating data in the training data; the pretreatment comprises the following steps: any one or combination of any multiple of data normalization, data missing supplement, data redundancy removal, data noise reduction and data abnormal deletion;
and taking the operation parameters in the operation data in the training data as the input of a pre-constructed health degree prediction model, taking the battery health degree value in the operation data as the output of the health degree prediction model, and training the health degree prediction model to obtain a trained health degree prediction model.
5. The method of claim 1 or 2, further comprising:
obtaining verification data of the battery, wherein the verification data comprises at least two groups of operation data, and the operation data comprises parameter values of at least two operation parameters and corresponding battery health values; the validation data is different from the training data;
inputting the operation data in the verification data into the health degree prediction model to obtain a prediction result output by the health degree prediction model;
and modifying the model parameters according to the prediction result.
6. The method of claim 5, modifying the model parameters based on the prediction, comprising:
comparing the prediction result with the battery health value in the operation data in the verification data to obtain a comparison result;
and modifying the model parameters according to the comparison result.
7. The method of claim 3, further comprising:
and optimizing the model parameters of the trained health degree prediction model by using a preset optimization algorithm.
8. The method of claim 1, further comprising:
and constructing the scoring model by using a preset regression algorithm.
9. A data processing apparatus comprising:
a data obtaining unit, configured to obtain training data of a battery, where the training data includes at least two sets of operation data, the operation data includes parameter values of at least two operation parameters and corresponding battery health values, and the operation parameters in the operation data include: environmental information, parking information, car information of traveling, charge information and discharge information, environmental information includes: weather information in the battery operating environment or/and whether traffic is congested; the parking information includes: any one or more of parking place, parking mode and parking duration; the automobile driving information includes: any one or more of driving time length, driving speed and acceleration of the automobile in which the battery is located; the charging information and the discharging information include: the number of times or/and the duration of charge and discharge;
the first training unit is used for taking operation data in the training data as a training sample of a pre-constructed health degree prediction model and training the health degree prediction model;
a vector obtaining unit, configured to obtain a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, where the parameter vector includes: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model;
the second training unit is used for taking a parameter vector of a first operating parameter as the input of a pre-constructed scoring model, taking a scoring value of the first operating parameter as the output of the scoring model, and training the scoring model to obtain a trained scoring model, wherein the first operating parameter is an operating parameter with a scoring value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
10. An electronic device, comprising:
the memory is used for storing the application program and data generated by the running of the application program;
a processor for executing the application to implement: obtaining training data of a battery, wherein the training data comprises at least two groups of operation data, the operation data comprises parameter values of at least two operation parameters and corresponding battery health values, and the operation parameters in the operation data comprise: environmental information, parking information, car information of traveling, charge information and discharge information, environmental information includes: weather information in the battery operating environment or/and whether traffic is congested; the parking information includes: any one or more of parking place, parking mode and parking duration; the automobile driving information includes: any one or more of driving time length, driving speed and acceleration of the automobile in which the battery is located; the charging information and the discharging information include: the number of times or/and the duration of charge and discharge; taking the operation data in the training data as a training sample of a pre-constructed health degree prediction model, and training the health degree prediction model; obtaining a parameter vector of the operating parameter in the operating data according to a model parameter of the health degree prediction model, wherein the parameter vector comprises: parameter weight values corresponding to the operation parameters of a plurality of operation units in the health degree prediction model; taking a parameter vector of a first operating parameter as input of a pre-constructed scoring model, taking a score value of the first operating parameter as output of the scoring model, and training the scoring model to obtain a trained scoring model, wherein the first operating parameter is an operating parameter with a score value in the at least two operating parameters; the score value of the first operating parameter characterizes a degree of influence of the first operating parameter on the health value of the battery;
the scoring model is used for outputting a scoring result of a second operation parameter by taking a parameter vector of the second operation parameter as input, wherein the second operation parameter is an operation parameter without a scoring value; the scoring result characterizes a degree of influence of the second operating parameter on a health value of the battery.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911418635.2A CN111142026B (en) | 2019-12-31 | 2019-12-31 | Data processing method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911418635.2A CN111142026B (en) | 2019-12-31 | 2019-12-31 | Data processing method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111142026A CN111142026A (en) | 2020-05-12 |
| CN111142026B true CN111142026B (en) | 2021-12-24 |
Family
ID=70523046
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911418635.2A Active CN111142026B (en) | 2019-12-31 | 2019-12-31 | Data processing method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111142026B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114325454B (en) * | 2021-12-30 | 2023-07-04 | 东软睿驰汽车技术（沈阳）有限公司 | Method, device, equipment and medium for determining influence of multiple characteristics on battery health |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1691209B1 (en) * | 2005-02-14 | 2008-10-01 | Denso Corporation | Method and apparatus for detecting charged state of secondary battery based on neural network calculation |
| CN104827928B (en) * | 2014-09-09 | 2017-07-11 | 北汽福田汽车股份有限公司 | The control system of electric automobile and the electric automobile with it |
| CN107069122A (en) * | 2017-04-01 | 2017-08-18 | 山东省科学院自动化研究所 | A kind of Forecasting Methodology of electrokinetic cell remaining life |
| CN108320043B (en) * | 2017-12-19 | 2021-12-14 | 江苏瑞中数据股份有限公司 | Power distribution network equipment state diagnosis and prediction method based on electric power big data |
| CN108711008B (en) * | 2018-05-17 | 2022-03-29 | 北京中软国际信息技术有限公司 | Multi-dimensional software project health degree evaluation method based on convolutional neural network |
| US10983167B2 (en) * | 2018-06-14 | 2021-04-20 | Huayuan Semiconductor (Shenzhen) Limited Company | Method and device for gauging an electronic apparatus |
| CN109283469A (en) * | 2018-09-21 | 2019-01-29 | 四川长虹电器股份有限公司 | Battery management system failure prediction method, device and readable storage medium storing program for executing |
| CN109447531A (en) * | 2018-12-25 | 2019-03-08 | 国网福建省电力有限公司 | A kind of intelligent electric meter method for evaluating state based on the evaluation of neural network health degree |
| KR102650175B1 (en) * | 2019-06-17 | 2024-03-22 | 엘지전자 주식회사 | Method for charging battery included in robot and apparatus thereof |
-
2019
- 2019-12-31 CN CN201911418635.2A patent/CN111142026B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111142026A (en) | 2020-05-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107122594B (en) | New energy vehicle battery health prediction method and system | |
| KR102233589B1 (en) | Modeling method and device for evaluation model | |
| CN111237988B (en) | Control method and system for subway vehicle-mounted air conditioning unit | |
| CN110738564A (en) | Post-loan risk assessment method and device and storage medium | |
| CN106022521A (en) | Hadoop framework-based short-term load prediction method for distributed BP neural network | |
| CN112270546A (en) | Risk prediction method and device based on stacking algorithm and electronic equipment | |
| Oliinyk et al. | The decision tree construction based on a stochastic search for the neuro-fuzzy network synthesis | |
| CN111177655B (en) | Data processing method and device and electronic equipment | |
| CN107704883A (en) | A kind of sorting technique and system of the grade of magnesite ore | |
| CN111142026B (en) | Data processing method and device and electronic equipment | |
| CN114154252B (en) | Risk assessment method and device for failure mode of power battery system of new energy automobile | |
| CN111860971A (en) | Method and device for predicting residual life of turnout switch machine | |
| CN115099326A (en) | Behavior prediction method, behavior prediction device, behavior prediction equipment and storage medium based on artificial intelligence | |
| Mandal et al. | Selecting best attributes for software defect prediction | |
| CN115081609A (en) | Acceleration method in intelligent decision, terminal equipment and storage medium | |
| CN114492967A (en) | Urban rail station passenger flow prediction method and medium based on CEEMDAN and BLSTM combined model | |
| CN114692888A (en) | System parameter processing method, device, equipment and storage medium | |
| Wang et al. | An enhanced software defect prediction model with multiple metrics and learners | |
| Chernyshev et al. | Digital Object Detection of Construction Site Based on Building Information Modeling and Artificial Intelligence Systems. | |
| CN112035942B (en) | Driving behavior-based energy consumption simulation method and device | |
| CN112434839B (en) | Distribution transformer heavy overload risk prediction method and electronic equipment | |
| CN117010283B (en) | Method and system for predicting deformation of steel pipe column structure of PBA station | |
| Tsiporkova et al. | A novel gene-centric clustering algorithm for standardization of time series expression data | |
| CN116542355A (en) | Transaction uplink prediction model construction method and device, electronic equipment and storage medium | |
| CN117526302A (en) | Short-term power load prediction method and system based on STL-VMD-COAHS-MTS-Mixers combination model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |